The Climate of Past Interglacials, Volume 7 (Developments in Quaternary Science)

DEVELOPMENTS IN QUATERNARY SCIENCE, 7 SERIES EDITOR: JAAP J .M . VAN DER MEER THE CLIMATE OF PAST INTERGLACIALS Deve...

Author: F. Sirocko | M. Claussen | T. Litt | M.F. Sanchez-Goni

29 downloads 844 Views 55MB Size Report

This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!

Report copyright / DMCA form

DOWNLOAD PDF

DEVELOPMENTS IN QUATERNARY SCIENCE, 7 SERIES EDITOR: JAAP J .M . VAN DER MEER

THE CLIMATE OF PAST INTERGLACIALS

Developments in Quaternary Science (Series editor: Jaap J.M. van der Meer) Volumes in this series 1. The Quaternary Period in the United States Edited by A.R. Gillespie, S.C. Porter, B.F. Atwater, 0-444-51470-8 (hardbound); 0-444-51471-6 (paperback) – 2003 2. Quaternary Glaciations – Extent and Chronology Edited by J. Ehlers, P.L. Gibbard Part I: Europe ISBN 0-444-51462-7 (hardbound) – 2004 Part II: North America ISBN 0-444-51592-5 (hardbound) – 2004 Part III: South America, Asia, Australasia, Antarctica ISBN 0-444-51593-3 (hardbound) – 2004 3. Ice Age Southern Andes – A Chronicle of Paleoecological Events Authored by C.J. Heusser 0-444-51478-3 (hardbound) – 2003 4. Spitsbergen Push Moraines – Including a translation of K. Gripp: Glaciologische und geologische Ergebnisse der Hamburgischen Spitzbergen-Expedition 1927 Edited by J.J.M. van der Meer 0-444-51544-5 (hardbound) – 2004 5. Iceland – Modern Processes and Past Environments ´ . Knudsen Edited by C. Caseldine, A. Russell, J. HarDardo´ttir, O 0-444-50652-7 (hardbound) – 2005 6. Glaciotectonism Authored by J.S. Aber, A. Ber 0-444-52943-8 (hardbound) – 2006 7. The Climate of Past Interglacials Edited by F. Sirocko, M. Claussen, M.F. Sa´nchez Gon˜i, T. Litt 0-444-52955-1 (hardbound) – 2007 For further information as well as other related products, please visit the Elsevier homepage (http://www.elsevier.com)

Developments in Quaternary Science, 7 Series editor: Jaap J.M. van der Meer

THE CLIMATE OF PAST INTERGLACIALS edited by

Frank Sirocko Institute for Geoscience, University of Mainz, Mainz, Germany

Martin Claussen Meteorological Institute, University Hamburg, and Max Planck Institute for Meteorology, Hamburg, Germany

Marı´a Fernanda Sa´nchez Gon˜i Department of Geology and Oceanography, EPHE-UMR-CNRS 5805, EPOC, University Bordeaux 1, Talence, France

Thomas Litt Institute for Paleontology, University of Bonn, Bonn, Germany

Amsterdam – Boston – Heidelberg – London – New York – Oxford – Paris San Diego – San Francisco – Singapore – Sydney – Tokyo

Elsevier Radarweg 29, PO Box 211, 1000 AE Amsterdam, The Netherlands The Boulevard, Langford Lane, Kidlington, Oxford OX5 1GB, UK First edition 2007 Copyright ª 2007 Elsevier B.V. All rights reserved No part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means electronic, mechanical, photocopying, recording or otherwise without the prior written permission of the publisher Permissions may be sought directly from Elsevier’s Science & Technology Rights Department in Oxford, UK: phone ðþ44Þ (0) 1865 843830; fax ðþ44Þ (0) 1865 853333; email: [email protected]. Alternatively you can submit your request online by visiting the Elsevier web site at http://elsevier.com/locate/permissions, and selecting Obtaining permission to use Elsevier material 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 verification of diagnoses and drug dosages should be made Library of Congress Cataloging-in-Publication Data A catalog record for this book is available from the Library of Congress British Library Cataloguing in Publication Data A catalogue record for this book is available from the British Library ISBN-13: 978-0-444-52955-8 ISBN-10: 0-444-52955-1 ISSN: 1571-0866 For information on all Elsevier publications visit our website at books.elsevier.com Printed and bound in Italy. 07 08 09 10 11 10 9 8

7 6 5 4 3

2 1

Working together to grow libraries in developing countries www.elsevier.com | www.bookaid.org | www.sabre.org

In fond memory of Nicholas Shackleton (1937–2006)

This page intentionally left blank

Table of Contents Preface Thorsten Kiefer and Christoph Kull ..................................................................... xi Acknowledgements ..........................................................................................................xv Section 1 Forcing Mechanisms (ed. Martin Claussen) 1. Introduction to Climate Forcing and Climate Feedbacks....................................................3 Martin Claussen 2. Insolation During Interglacial ................................................................................................13 A. Berger, M.F. Loutre, F. Kaspar and S.J. Lorenz 3. A Survey of Hypotheses for the 100-kyr Cycle...................................................................29 Martin Claussen, Andre´ Berger and Hermann Held 4. Modelling the 100-kyr Cycle – An Example From LLN EMICs ......................................37 Andre´ Berger and Marie-France Loutre Section 2 Methods of Palaeoclimate Reconstruction and Dating (ed. Frank Sirocko) 5. Introduction – Palaeoclimate Reconstructions and Dating...............................................47 Frank Sirocko 6. Late Quaternary Interglacials in East Antarctica From Ice-Core Dust Records ............53 Barbara Delmonte, Jean Robert Petit, Isabelle Basile-Doelsch, Emil Jagoutz and Valter Maggi 7. Eustatic Sea Level During Past Interglacials .......................................................................75 M. Siddall, J. Chappell and E.-K. Potter 8. Uranium-Series Dating of Peat from Central and Northern Europe...............................93 Manfred Frechen, Melanie Sierralta, Deniz Oezen and Brigitte Urban 9. U-Redistribution in Fossil Reef Corals from Barbados, West Indies, and Sea-Level Reconstruction for MIS 6.5 .................................................................................119 Denis Scholz, Augusto Mangini and Dieter Meischner 10. Holocene and Eemian Varve Types of Eifel Maar Lake Sediments...............................141 Bert Rein, Knut Ja¨ger, Yvonne Kocot, Kirsten Grimm and Frank Sirocko 11. Dating of Interglacial Sediments by Luminescence Methods.........................................157 D. Degering and M.R. Krbetschek 12. Neanderthal Presence and Behaviour in Central and Northwestern Europe During MIS 5e ..........................................................................................................173 Stefan Wenzel

viii

Table of Contents

Section 3 Climate and Vegetation in Europe During MIS 5 (ed. Maria Fernanda Sa´nchez Gon˜i) 13. Introduction to Climate and Vegetation in Europe During MIS 5.................................197 Marı´a Fernanda Sa´nchez Gon˜i 14. Abrupt Cooling Events at the Very End of the Last Interglacial....................................207 Klemens Seelos and Frank Sirocko 15. Estimates of Temperature and Precipitation Variations During the Eemian Interglacial: New Data From the Grande Pile Record (GP XXI) ...............231 Denis-Didier Rousseau, Christine Hatte´, Danielle Duzer, Patrick Schevin, George Kukla and Joel Guiot 16. Quantitative Time-Series Reconstructions of Holsteinian and Eemian Temperatures Using Botanical Data .............................................................239 Norbert Ku¨hl and Thomas Litt 17. Comparative Analysis of Vegetation and Climate Changes During the Eemian Interglacial in Central and Eastern Europe ..................................................255 A.A. Velichko, E.Y. Novenko, E.M. Zelikson, T. Boettger and F.W. Junge 18. Indications of Short-Term Climate Warming at the Very End of the Eemian in Terrestrial Records of Central and Eastern Europe.......................................265 T. Boettger, F.W. Junge, S. Knetsch, E.Y. Novenko, O.K. Borisova, K.V. Kremenetski and A.A. Velichko 19. Vegetation Dynamics in Southern Germany During Marine Isotope Stage 5 ( 130 to 70 kyr Ago) .................................................................................277 Ulrich C. Mu¨ller and Maria F. Sa´nchez Gon˜i 20. Subtropical NW Atlantic Surface Water Variability During the Last Interglacial .......289 M.J. Vautravers, G. Bianchi and N.J. Shackleton 21. Abrupt Change of El Nin˜o Activity off Peru During Stage MIS 5e-d...........................305 Rein, Bert, Frank Sirocko, Andreas Lu¨ckge, Lutz Reinhardt, Anja Wolf and Wolf-Christian Dullo 22. Interglacial and Glacial Fingerprints from Lake Deposits in the Gobi Desert, NW China ........................................................................................................323 Bernd Wu¨nnemann, Kai Hartmann, Norbert Altmann, Ulrich Hambach, Hans-Joachim Pachur and Hucai Zhang Section 4 Climate, Vegetation and Mammalian Faunas in Europe during Middle Pleistocene Interglacials (MIS 7, 9, 11) (ed. Thomas Litt) 23. Introduction: Climate, Vegetation and Mammalian Faunas in Europe during Middle Pleistocene Interglacials (MIS 7, 9, 11)................................................................................351 Thomas Litt 24. Fine-Tuning the Land–ocean Correlation for the Late Middle Pleistocene of Southern Europe...........................................................................................359 K.H. Roucoux, P.C. Tzedakis, L. de Abreu and N.J. Shackleton

Table of Contents

ix

25. Climate Variability of the Last Five Isotopic Interglacials: Direct Land–Sea–Ice Correlation from the Multiproxy Analysis of North-Western Iberian Margin Deep-Sea Cores ........................................................................................................375 S. Desprat, M.F. Sa´nchez Gon˜i, F. Naughton, J.-L. Turon, J. Duprat, B. Malaize´, E. Cortijo and J.-P. Peypouquet 26. Palynological and Geochronological Study of the Holsteinian/Hoxnian/Landos Interglacial...............................................................................................................................387 Mebus A. Geyh and Helmut Mu¨ller 27. A New Holsteinian Pollen Record From the Dry Maar at Do¨ttingen (Eifel) ...............397 Markus Diehl and Frank Sirocko 28. Interglacial Pollen Records from Scho¨ningen, North Germany.....................................417 Brigitte Urban 29. Mammalian Faunas From the Interglacial Periods in Central Europe and Their Stratigraphic Correlation....................................................................................445 Wighart von Koenigswald 30. MIS 5 to MIS 8 – Numerically Dated Palaeontological Cave Sites of Central Europe .........................................................................................................455 Wilfried Rosendahl, Doris Do¨ppes and Stephan Kempe 31. The Last and the Penultimate Interglacial as Recorded by Speleothems From a Climatically Sensitive High-Elevation Cave Site in the Alps .........................................471 Christoph Spo¨tl, Steffen Holzka¨mper and Augusto Mangini Section 5 Modelling Past Interglacial Climates (ed. Martin Claussen) 32. Climate System Models – A Brief Introduction................................................................495 Martin Claussen 33. Simulations of the Eemian Interglacial and the Subsequent Glacial Inception with a Coupled Ocean–Atmosphere General Circulation Model..................................499 Frank Kaspar and Ulrich Cubasch 34. Simulated Teleconnections During the Eemian, the Last Glacial Inception and the Preindustrial Period ................................................................................................517 Martin Widmann, Nikolaus Groll and Julie M. Jones 35. Orbital Forcing on Atmospheric Dynamics During the Last Interglacial and Glacial Inception ............................................................................................................527 Gerrit Lohmann and Stephan J. Lorenz 36. Interglacials as Simulated by the LLN 2-D NH and MoBidiC Climate Models..........547 M.F. Loutre, A. Berger, M. Crucifix, S. Desprat and M.F. Sa´nchez Gon˜i 37. Vegetation–Climate Feedbacks in Transient Simulations Over the Last Interglacial (128 000–113 000 yr BP) .....................................................................................563 M. Gro¨ger, E. Maier-Reimer, U. Mikolajewicz, G. Schurgers, M. Vizcaino and A. Winguth

x

Table of Contents

38. Mechanisms Leading to the Last Glacial Inception over North America: Results From the CLIMBER-GREMLIN Atmosphere–Ocean–Vegetation Northern Hemisphere Ice-Sheet Model .............................................................................573 Masa Kageyama, Sylvie Charbit, Catherine Ritz, Myriam Khodri and Gilles Ramstein 39. Modelling the End of an Interglacial (MIS 1, 5, 7, 9, 11)..................................................583 Claudia Kubatzki, Martin Claussen, Reinhard Calov and Andrey Ganopolski Section 6 Synthesis 40. Chronology and Climate Forcing of the Last Four Interglacials....................................597 Frank Sirocko, Martin Claussen, Thomas Litt, Maria Fernanda Sa´nchez Gon˜i, Andre´ Berger, Tatjana Boettger, Markus Diehl, Ste´phanie Desprat, Barbara Delmonte, Detlev Degering, Manfred Frechen, Mebus A. Geyh, Matthias Groeger, Masa Kageyama, Frank Kaspar, Norbert Ku¨hl, Claudia Kubatzki, Gerrit Lohmann, Marie-France Loutre, Ulrich Mu¨ller, Bert Rein, Wilfried Rosendahl, Katy Roucoux, Denis-Didier Rousseau, Klemens Seelos, Mark Siddall, Denis Scholz, Christoph Spo¨tl, Brigitte Urban, Maryline Vautravers, Andrei Velichko, Stefan Wenzel, Martin Widmann and Bernd Wu¨nnemann Index .............................................................................................................................615

Preface: Climates of Past Interglacials – a PAGES Perspective Thorsten Kiefer and Christoph Kull PAGES International Project Office, Sulgeneckstrasse 38, CH-3007- Berne, Switzerland

The cultural evolution of humans has accelerated considerably during the Holocene interglacial. This explosion of civilisation has probably only been possible under the mild and relatively stable climatic conditions that have prevailed for the last 11 000 years. However, these conditions cannot be taken for granted. This is one of the rather simple but unequivocal and important lessons we have learned from the palaeoclimate record. All other interglacials terminated after a few thousands to a few tens of thousands of years. In fact, interglacial states similar to those of today, with little land ice and largely elevated temperatures at mid-high latitudes, prevailed during no more than 15% of the last half million years. These simple empirics already give clear evidence that interglacials are rather unstable on a 10 000-year timescale. Also on shorter timescales of millennia to decades, late Holocene climate fluctuations such as the Little Ice Age and those associated with the Maunder Minimum proove that interglacial climate is not entirely stable on a regional scale but responds to even subtle changes in radiative forcing. Moreover, the discovery of the 8.2 kyr cooling event made it clear that even the worst-case scenario (socioeconomically speaking) of an abrupt change of climate within years is not just a theoretical possibility but has in fact happened in the prehistoric past. It is therefore clear that in principle it could happen again, once some perturbation exceeds a critical threshold of the climate system. Given that humans are indeed becoming increasingly effective at perturbing the climate–environment system, learning

about its sensitivity, thresholds and feedbacks should be in our best interest. An obvious way to do this is to study climate and environment in a suite of experiments where boundary conditions are similar but not identical to those of today. Through the quasi-cyclic reoccurrence of interglacials during the late Pleistocene, we have several such experiments at hand. The palaeoclimate community therefore holds an important key to scientific information on climate change that provides a basis for appropriate adaptation and mitigation strategies. The authors of this book have taken up this challenge and summarise their results in this special volume on climates of past interglacials. It presents state-ofthe-art science on new reconstructions from all spheres of the Earth System and on their synthesis, on methodological advances and on the current ability of numerical models to simulate low- and high-frequency changes of climate, environment and chemical cycling related to interglacials. Most of the authors had been involved in the German DEKLIM programme (www.deklim.de) and have attended some of the five DEKLIM workshops and conferences between 2001 and 2005. The discussions were quite controversial in the beginning, but step by step the dating, reconstruction and modelling of interglacial climate evolution came to a convincing synthesis. Not all open questions are settled, and there is not full consensus on all subjects, but the picture of past climates is now much clearer than it was five years ago. Beyond the pure scientific findings, DEKLIM will leave more (secondary) footprints

xii

Preface

in palaeoscience which are of equally high value in the perspective of PAGES because of their integrative character. Two fields of palaeoscience that did not often mix well with classic palaeoclimatology were deliberately and successfully incorporated as genuine counterparts of the project. Firstly, studies on palaeovegetation added important information on the regional expression of, and regional feedbacks to, global change. Secondly, numerical modelling of climatic and environmental and geochemical parameters was an integral part of the project. It seems that a generation of modellers and data people has evolved who are capable of communicating effectively on mutual needs and complementary results. A great number of papers of this volume reflect this successful data-model symbiosis. In addition, in spite of being funded by the German Federal Ministry of Education and Research, DEKLIM managed to cooperate successfully beyond national boundaries. The science community and the science consequently benefited greatly from the involvement of many renowned international scientists. The lasting value of this project comes not only from its scientific results, which have led to an improved understanding of the operation of the climate system, but also from the identification of current scientific limitations and open questions. These provide orientation as to where researchers, funders and programmes like PAGES might most effectively focus their resources and efforts for future research. For example, one fundamental limitation is given by the uncertainties in the age control on pre-Holocene interglacials. They are out of reach of radiocarbon dating, not anywhere near the next palaeomagnetic reversal, and the relative smoothness of the climatic records leaves only minor characteristics for event stratigraphy. However, to enable modellers to prescribe forcings with a realistic phasing will require a submillennial chronological accuracy. Creative new approaches are needed, some of which are discussed in this volume.

A similarly high standard is demanded from the quality of the reconstruction of climatic and environmental parameters. The proxies used to reconstruct past interglacial conditions need to detect variations that are usually rather subtle compared to glacial or semi-glacial scenarios. This task imposes higher requirements on methods in terms of precision and accuracy. Hence, the ongoing improvement of analytics, and the development and refinement of proxies and their calibrations, is not just an academic obsession but a fundamental need to advance our understanding of past climate and ultimately the accuracy of prediction. One limitation on model simulations is records of climate forcing, both natural (orbital, solar irradiance, volcanic and greenhouse gases) and anthropogenic (greenhouse gases, aerosols and land use). The validity of any climate simulation depends on the accuracy of the input parameters that generate climate change and environmental response. The abovementioned improvements of chronological tools and proxies will contribute to the advancement in the accuracy of forcing records. Raising the awareness of its importance may be another measure to foster progress. In previous years and decades, a lot of effort was concentrated on understanding the large-amplitude climate variations during glacials, and between glacials and interglacials, thereby disregarding interglacial global change issues. As a result, the spatial density and temporal resolution of records are presently insufficient to allow for firm conclusions on high-frequency or regionalscale climate variability during past interglacials. This becomes more and more acute the further back in time one goes due to the increased difficulty in obtaining good archives. The ambitious target for the coming years will be to reconstruct and simulate centennial-to-decadal scale variability of climate models, not only for the Holocene, but also for the slightly differing boundary conditions of pre-Holocene interglacials.

Preface

Since climate models reveal themselves to be relatively patchy dynamic climate patterns, a critical density of data is necessary to map them. This will require concerted efforts focussing on particular regions and time intervals. The largest region identified as being underrepresented in the climate record and therefore a prime target of future research is the entire southern hemisphere. To be effective, concerted efforts, such as those described above, will require an infrastructure for data compilation, storage and accessibility. This will be important in order to exploit the full potential of numerical climate models. Another fundamental future challenge will be to further increase

xiii

professionalism in both data management and the attitude towards data dissemination. The challenges described above are among those also adopted by PAGES as being particularly worthy of support. These objectives will not only require excellent research by individual scientists and groups but also a high level of interdisciplinarity and organisational structure. While PAGES cannot influence individual scientific excellence, it will continue to facilitate integrative palaeoscience along the lines of the DEKLIM project. Meanwhile, the community can build on many scientific results and datasets, an international network and a comprehensive concept for an integrative global change research programme.

This page intentionally left blank

Acknowledgements Our sincere thanks go to Ulrich Katenkamp (Federal Ministry of Education and Research, Germany) for planning and organizing the DEKLIM program, which generously funded research, workshops and a large conference on the climate of past interglacials.

This book could not have been put together without the continuous, meticulous care of Saskia Rudert, who took responsibility for the DEKLIM-EEM project office and for this book in particular.

This page intentionally left blank

Section 1 Forcing Mechanisms (ed. Martin Claussen)

This page intentionally left blank

1. Introduction to Climate Forcing and Climate Feedbacks Martin Claussen Meteorological Institute, University Hamburg, and Max Planck Institute for Meteorology, Bundesstr. 53, D-20146 Hamburg, Germany

1.1 WHAT IS CLIMATE? When interpreting palaeoclimate archives, the question arises as to what is the driver, or forcing, of climate. For example, when analysing reconstructions of temperature and atmospheric CO2 concentrations during the past interglacials and during the transition from an interglacial to a glacial, it is often asked whether temperature drives changes in atmospheric CO2 concentrations or vice versa – of course with a predisposed assumption of relevance for present-day climate change. Hence the perception of climate forcing, of what is external or internal, is of outmost importance in climate research. To provide some guidance on the discussion in this book, this introductory paper provides a definition of climate, with a discussion on climate forcing and climate feedback and its relevance to the interpretation of palaeoclimate archives. In a classical definition, climate is viewed as the sum of all meteorological phenomena that characterize the mean state of the atmosphere (Hann, 1883), or in more colloquial terms, as mean weather. With the notion of climate variations, not only the mean state of the atmosphere, but also its variability, including the statistics of extreme events, has been incorporated into the definition of climate. Hence in meteorological terms – meteorological, because this definition is made in terms of meteorological state variables – climate is briefly known as the statistics of weather (Hantel et al., 1987 ). The World Meteorological Organization has defined a time span of 30 years over which the statistics of weather is to be calculated, and the climate in the 30-year period of 1931 to 1960 is sometimes referred to as ‘normal climate’. But shorter and longer

time spans can be found in the literature as well – there is nothing special about choosing 30 years. It is required that the time span to define climate has to be choesn such that the statistical moments are stable. However, when browsing through climate archives, it becomes immediately evident that climate is not a stationary process. Climate changes on all timescales. The classical definition of climate has proven to be useful for climatology, the descriptive view of climate. However, for understanding climate dynamics, that is, the processes which govern the mean state of the atmosphere, the classical definition appears to be too restrictive since the mean state of the atmosphere is affected by more than just atmospheric phenomena. It has been realized that the mean state of the atmosphere as well as its variations depends on the dynamics of the climate system and on the interaction between the various components of this system (see Fig. 1.1). von Humboldt (1845) was probably the first who explicitly stated that problem. Therefore in climate dynamics and climate physics, a wider definition of climate has been proposed in terms of state and statistics of the climate system (e.g. Kraus, 2000; Annex in: Houghton et al., 2001). The definition of a climate system is not deduced from first principles. It is done ad hoc, in a pragmatic way to isolate the focus of research, the system, from its external environment. One could define a hierarchy of systems depending on timescales considered (see Peixoto and Oort, 1992). For investigation of seasonal climate variations, for example, the dynamics of the atmosphere, the upper ocean, the terrestrial biosphere

4

Martin Claussen

The climate system

Energy

ATMOSPHERE TERRESTRIAL VEGETATION

Water Momentum

Carbon

CRYOSPHERE

OCEAN

MARINE BIOTA

PEDOSPHERE, LITHOSPHERE

© CLIMBER

Fig. 1.1 Sketch of the climate system and its components: the atmosphere, the hydrosphere (mainly the ocean, but also rivers, lakes, rain and groundwater are included), the cryosphere (i.e. ice sheets, glacier, sea ice, snow, permafrost and clathrates), the marine and terrestrial biosphere, the pedosphere (soils and rocks) and – not explicitly shown here – the lithosphere and the upper Earth’s mantle. The latter have to be considered when long times over which the bedrock of ice sheets and the continent changes are considered. The climate system is driven by the solar energy flux and the geothermal heat flux. If Pleistocene climate variations are discussed, volcanic activity and outgasing at oceanic ridges are considered as external forcing. The climate system components are coupled through fluxes of energy, momentum and matter.

(excluding migration of vegetation), sea ice and the upper metres of soil need to be considered. The location of continents, the state of the deep ocean and of the ice sheets can be assumed to be constant in time. Hence, processes in the atmosphere, upper ocean, terrestrial biosphere, sea ice and soil as well as exchange processes, or feedbacks, between these components are internal processes. The heat and mass fluxes from the deep ocean into the upper ocean and from the ice sheets to the atmosphere as well as the geothermal heat flux appear as external forcing. Although a reduction of the climate system to fewer components is often done in climate system modelling, it is problematic; the dynamics of the reduced system can be quite different from the dynamics of the more complete system. In Section 5 of this book (see Chapter 39 in particular) in which evidence from palaeoclimate archives is interpreted by climate system models, these problems will be revisited.

1.2 WHY DOES CLIMATE VARY? If one would sketch climate variability, for example in terms of temperature variance, as a function of frequency, one would obtain a spectrum like the one shown in Fig. 1.2. This figure reveals a so-called red spectrum, that is, an increase in variability with decreasing frequency. Moreover, the red spectrum eventually levels off and merges into a flat, or white, spectrum at low frequencies. Superimposed on the spectrum are a few spikes some of which appear as a direct response to forcing as, for example, the diurnal or the annual cycle. Such a spectrum indicates that climate variability emerges as a direct response to changes in forcing and to some internal processes seemingly independent of a direct forcing. This behaviour has been described as forced or external variability and free or internal variability, respectively (Lorenz, 1979).

Surface temperature variance

Introduction to Climate Forcing and Climate Feedbacks

ORBITAL CYCLES Annual cycle and harmonics

Diurnal cycle and harmonics

Cryosphere Deep-ocean circulation Mixed layer ocean Atmosphere

106

105

104

Thermal relaxation Inertial relaxation 103

102

101

100

10–1

10–2

10–3

10–4

10–5

Period (years)

Fig. 1.2 Sketch of a temperature variance in the nearsurface atmosphere as function of frequency. This figure is taken from Crowley and North (1991) with the permission of Oxford University Press.

1.3 CLIMATE FORCING The most important external forcing is the solar energy flux. This energy flux increases slowly by about 10% per billion years as the Sun becomes steadily hotter. But there are also variations of the solar energy flux at periods of 11, 22, 80 and roughly 200 years. Presumably, these variations amount to only 0.1 to 0.3% of the mean flux, which, however, is only known for the shortest 11-year cycle from satellite measurements. For longer-term variations, estimates are derived from proxy data in combination with theories of solar dynamics (e.g. Lean et al., 1995; Bard et al., 2000). Albeit small, a direct response of the near-surface atmosphere to estimates in solar forcing is found in many climate models, in particular in simulations of the last 1000 years before the industrial revolution (Jones and Mann, 2004). It has also been hypothesized that the atmosphere could react in a nonlinear way to insolation changes by amplifying the weak forcing via wave interaction between the troposphere, the lower 10-km deep ‘weather’ layer of the atmosphere, and the upper atmosphere where the much larger variations in UV radiation modify atmospheric chemistry and hence, the energy budget (e.g. Shindell et al., 1999).

5

Solar wind and the magnetic field of the Earth shield the Earth from cosmic radiation. There are a number of studies that reveal correlations between changes in cosmic rays and atmospheric variables (e.g. Svensmark and Friis-Christensen, 1997). In most cases, these correlations appear as artefacts because the correlations vanish if new data appear, or data are simply not treated properly (Laut, 2003). There are several hypotheses of how cosmic rays could affect climate; none of them, however, is commonly accepted (Ramaswamy et al., 2001). The Earth rotates and it spins like a top around the Sun. Since the Sun, the Earth’s moon and the larger planets exert torques on the spinning Earth, the Earth wobbles – the eccentricity of the Earth’s orbit, the obliquity of the Earth’s axis and the perihelion (the position of the Earth closest to the Sun) change. These orbital changes affect the meridional and seasonal distribution of insolation (see Chapter 2). This way, the orbital signal is seen in many climate archives (see also Chapters 3 and 4). Volcanic explosions affect the chemistry of the atmosphere such that a global-scale cooling of some tenths of a degree over two to three years after the explosion can be detected. For example, the Pinatubo eruption cooled the near-surface Earth by some 0.3 K. Hence, it is not the individual volcanic eruption that influences long-term climate variation, but the frequency of eruptions. For example, the lack of eruptions at the beginning of the twentieth century contributed to the observed global warming at that time (Briffa et al., 1998). Volcanic forcing belongs to the class of tectonic forcing. Tectonic forcing is driven by mantle convection, that is, by processes in the Earth’s interior. Whether these processes are defined as forcing or feedback depends on the timescales under consideration. In climate system models that aim to describe the evolution of the Earth’s climate over billions of years, tectonic processes such as spreading of oceanic crust and subduction of continental plates have to be

6

Martin Claussen

considered as internal processes (e.g. von Bloh et al., 2003). For simulations of Pleistocene climate variations, tectonic processes can be considered as external. In some books, humankind is considered as part of the biosphere and, thus, included in the climate system (Kraus, 2000). This approach is, however, not commonly accepted. Simulation of the natural components of the climate system relies on the physics of motion and chemistry of gases and liquids. Human activity, economics, culture and values are, however, not accessible by these tools. Therefore, anthropogenic activity is taken as external forcing in climate system models. The so-called integrated models of the Earth system (including human dynamics as well as the climate system, e.g. Alcamo, 1994; Schellnhuber, 1999) require new approaches. Humankind affects the climate system mainly by altering the chemical composition of the atmosphere and by changing the land-surface structure. Currently, anthropogenic emissions of CO2 by burning fossil fuel and harvesting forests is 50 to 100 times stronger than CO2 emission by outgasing from the Earth’s interior, and the atmospheric CO2 concentration reaches levels of 380 ppmv (parts of CO2 volume per million parts of atmospheric volume) which is well above the preindustrial value of approximately 280 ppmv and the late Pleistocene average of approximately 210 ppmv (Prentice et al., 2001); the atmospheric CH4 concentration rose from approximately 700 ppbv (parts per billion) in the preindustrial to approximately 1750 ppbv, with a late Pleistocene average of approximately 450 ppbv (Prather et al. 2001). Furthermore, a number of new chemical substances with a much stronger potential as a greenhouse gas than CO2 and CH4 have been created by humankind. One-third to one-half of the land is directly affected by land use (Vitousek et al., 1997). Besides CO2 emissions, this leads to changes in the near-surface energy budget, because a deforested area – in particular, when snow

covered – reflects more sunlight than forests and affects transpiration. The brightening of the Earth’s surface due to deforestation has presumably contributed to the cooling during the last centuries by approximately 0.35 K over the last 1000 years (Bauer et al., 2003). 1.4 INTERNAL VARIABILITY Even if there were no changes in external forcing, climate would vary anyhow because of internal instabilities embedded in the climate system components. An illustrative example of this free, internal variability is the motion of a thin layer of oil in a fry pan which is heated from below. If some ingredients are added to visualize the oil flow (pepper, for example), then one can observe regular structures like hexagonal cells or parallel streaks although the heating from below is (ideally) perfectly homogeneous. The structure of flow patterns depends on the strength of the (constant) forcing, that is, on the strength of the imposed heating. With a strong heating, the pattern could exhibit a wavy motion albeit the heating itself is kept steady. Finally, there is a possibility of a turbulent motion of boiling oil – again with a heating which is constant in space and time. Similarly, even with a steady external forcing, the atmosphere would heat at the equator and cool at the poles, and, if the Earth does not rotate, there would be a strong temperature gradient between the day and the night side of the Earth. An atmospheric motion could exist in response to this temperature gradient. Superimposed on a mean circulation between the hot side and the cold side of a resting Earth, atmospheric variability would exist as it does in the case of a homogeneously and constantly heated oil in a fry pan. The weather pattern, of course, would look rather different from our current patterns of zonally moving lowand high-pressure systems. Because the climate system components have different response times or, in other

Introduction to Climate Forcing and Climate Feedbacks

words, are more of less sluggish, the interaction between them leads to a red spectrum (see Hasselmann, 1976, and Fig. 2 from Crowley and North, 1991). Random variations in a fast climate system component, say the atmosphere, could by chance add up to an anomaly which affects a more sluggish component, the ocean for example. The oceanic random motion generated in this way may also reveal anomalies which then are felt by an even more sluggish ice sheets. Hence, anomalies add up randomly with larger amplitude in the slower variations. Wunsch (2003) argues that this process is the main source of climate variations, while external forcing would add only a little bit (see also Chapter 3).

7

phenomenon of the ‘green’ Sahara. According to palaeoclimatic evidence (e.g. Prentice et al., 2000), the Sahara was much greener in the early and mid-Holocene approximately 9000 to 6000 years ago. The greening has been explained by an increase of the palaeomonsoon in response to changes in the Earth’s orbit which led to a stronger warming of continents (Kutzbach and Guetter, 1986). The resulting increase in monsoonal precipitation appeared to be insufficient to cause any Saharan greening in all climate models. Subsequently, it was shown that a biogeophysical feedback, originally proposed by Charney (1975), could amplify the precipitation to generate enough rain for a substantial northward shift of savannah and steppes (Claussen, 1997).

1.5 CLIMATE AMPLIFIER 1.6 CLIMATE CHANGE TRIGGERS Internal processes and feedbacks between climate system components may not be caused by internal climate variability only. Feedbacks can also amplify external forcing in a disproportional, or nonlinear, way. For example, emission of greenhouse gases in present-day climate would increase the global and annual mean temperature by approximately 1 K if the concentration of greenhouse gases is doubled. This warming leads to larger evaporation and hence an increase in the most important greenhouse gas, water vapour. The so-called moist greenhouse effect amplifies the initial warming. Other feedbacks can further amplify or attenuate the warming. The final warming lies in the range of 1.5–4.5 K. This uncertainty of warming to a doubling in greenhouse gas concentration reflects the uncertainty in understanding feedback processes. Interestingly, the range of uncertainty has not changed very much over the last two decades (McAvaney et al., 2001), but there are attempts underway to reduce this uncertainty by analysing information from palaeoclimate archives. A further illustrative example of positive, that is, amplifying, feedbacks is the

In climate archives, many examples of rapid, abrupt climate changes are found. These could be interpreted as random fluctuations or random jumps of the climate system from one state to another which is one type of internal variability. In many cases, however, it seems plausible to assume that the abrupt climate change is triggered by some external driver. Again the Sahara, in particular the abrupt expansion of the Sahara approximately 5500 years ago (de Menocal, 2000), can serve as an illustrative example. In a coupled atmosphere–vegetation model, Claussen (1997) found that in present-day climate, Northern Africa can exhibit two states: an arid state like today and a ‘green’ state like in the early Holocene. For early Holocene conditions, only the green state seems to exist (Claussen and Gayler, 1997). As the external forcing, in this case the summer insolation over the Northern Hemisphere, declines during the Holocene, the atmosphere–vegetation system over Northern Africa moves from a equilibrium state in the early and middle Holocene to a state with two equilibria today. In theory, it would be possible that

8

Martin Claussen

Northern Africa could still be green, because the green state is an equilibrium, or stable, state. However, dynamically speaking, the green state today is much less stable than the arid state. Hence, any perturbation by some ubiquitous changes in the temperature of the tropical North Atlantic, for example, lets the atmosphere-vegetation system jump from the green into the arid state (Brovkin et al., 1998). Hence, the rapid aridification is not directly driven by orbital forcing; rather, it is triggered. Likewise, the rapid Dansgaard/Oeschger fluctuations are interpreted as transitions between two states of the inter-hemispheric Atlantic Ocean circulation which is driven by large-scale gradients of heating and freshwater fluxes (Ganopolski and Rahmstorf, 2001). Interestingly, the Dansgaard/Oeschger fluctuations occur with a remarkably regular pace. Therefore, it has been suggested that some external driver, which itself is too weak to generate any noticeable change in the climate system, would synchronize these fluctuations (Rahmstorf, 2003). Again, this – yet unknown – pacemaker cannot be considered as forcing, but rather as trigger. 1.7 RELEVANCE TO THE INTERPRETATION OF PALAEOCLIMATE ARCHIVES Climate can vary in response to changes in climate forcing, but it also can vary because of internal instabilities independently of any change in forcing. The part of climate variability which is forced by some external process is easily predictable once the forcing is known. On the contrary, internal climate variability is much less assessible to prediction. Only some statistical properties of internal climate variability can be derived. Hence, the knowledge of what is external and of what is internal becomes important when interpreting palaeoclimate archives in view of climate predictability. In geology and geography (Sirocko, personal communication), climate variations

on timescales much longer than the classical averaging period of 30 years to define climate are sometimes considered as determined by some forcing. This is a fallacy because also the sluggish components of the climate system, such as ice sheets and the deep ocean, exhibit internal, and hence less predictable, variability at longer timescales. Furthermore, internal climate variability could, by chance, emerge as seemingly regular, deterministic oscillation; nonetheless it would be futile, in this case, to search for any oscillating forcing behind it. A problem related to the question of external and internal climate variability is the ‘chicken and egg problem’ of temperature and atmospheric CO2 concentration. It has been suggested that the problem of whether temperature forces atmospheric CO2 concentration or vice versa, whether atmospheric CO2 concentration affects temperature, can be solved by determining the leads and lags between changes in temperature and in greenhouse gases found in palaeoclimate archives. From the dynamical point of view, this approach is futile. Firstly, in oscillating coupled systems the chicken and egg problem cannot be solved by analysing leads and lags. There are numerous examples in which an apparently lagging system drives a seemingly leading system. Secondly and more importantly, temperature and greenhouse gases co-evolve in the climate system by affecting each other. An increase in greenhouse gases leads to a rise in atmospheric temperature, and, in turn, an increase in temperature affects vegetation and upper ocean and thereby the carbon fluxes between the climate system components. Hence, one has to understand temperature and atmospheric CO2 fluctuations as feedback processes within the climate system. Only outgasing of CO2 by tectonic processes and anthropogenic greenhouse gas emissions are considered as climate forcing from this point of view. In Section 1.1.3, several climate forcings such as changes in solar energy flux, in

Introduction to Climate Forcing and Climate Feedbacks

volcanic activity and changes in the Earth’s orbit around the Sun are discussed. Because there are not yet any robust estimates of solar activity and of volcanic activity during the last interglacials, this forcing has not been considered in numerical simulations presented in this book. Hence, orbital forcing is the main agent in the current discussion of the dynamics of past interglacials and glacial inception, while the role of solar and volcanic activity has still to be explored. In general, climate variations appear as both externally driven and internally generated. In particular, internal feedbacks can amplify external forcing which itself would be too weak to cause any detectable climate change. In such cases, forcing should be viewed as trigger, and the relation between trigger and climate change is likely to be highly nonlinear. For example, the idea of a climate change trigger is used to explain glacial inceptions (Calov et al., 2004; see also Chapter 39). In the numerical model of the climate system used by Calov et al. (2004), a glacial inception appears as an instability of the atmosphere–ice sheet system. Once this instability is triggered, then the resulting climate changes are amplified by fast climate feedbacks such as changes in atmospheric water vapour and eventually by slower feedbacks such as vegetation shift, shift in oceanic circulation and in the carbon cycle which, in turn, affects atmospheric CO2 concentration and other greenhouse gases. The analysis of external and internal climate processes is complicated by the fact that, partly due to the lack of computational resources and partly due to the gaps in understanding the climate system, many climate system models do not simulate the interaction between all climate system components explicitly. Instead, only the dynamics of some subsystems, such as atmosphere and ocean for example, is explicitly considered, while the state of the other subsystems, such as the big ice sheets, is prescribed from palaeoclimate archives or other theoretical estimates. As mentioned in Section 1.1.1, the reduction of models of the

9

climate system has to be done with caution, and hence, this method will be critically assessed in Section 5.8. Finally, it is understood that the study of past interglacials and glacial inceptions by both excavating palaeoclimate archives and interpreting palaeoclimatic evidence by a spectrum of climate system models (see Chapter 32) – as done in this book – will considerably advance our understanding of the dynamics of the climate system in response to natural as well as anthropogenic forcing.

ACKNOWLEDGEMENT The author wishes to thank Claudia Kubatzki, Alfred-Wegener Institute, Bremerhaven, and Frank Sirocko, University of Mainz, for constructive discussion, and Saskia Rudert, University Mainz, Ursula Werner, Potsdam Institute for Climate Impact Research, and Barbara Zinecker, Max Planck Institute for Meteorology, for technical and editorial assistance.

REFERENCES Alcamo, J. (ed.), 1994. IMAGE 2.0: Integrated modeling of global climate change. Special issue of Water Air Soil Pollution, 76, 1–321. Bard, E., Raisbeck, G., Yiou, F., Jouzel, J., 2000. Solar irradiance during the last 1200 years based on cosmogenic nuclides. Tellus, 52B, 985–992. Bauer, E., Claussen, M., Brovkin, V., Hu¨nerbein, A., 2003. Assessing climate forcings of the Earth system for the past millennium. Geophysical Research Letter, 30, 1276–1279. von Bloh, W., Bounama, C., Franck, S., 2003. Cambrian explosion triggered by geosphere-biosphere feedbacks. Geophysical Research Letter, 30, 1963–1966. Briffa, K.R., Jones, P.D., Schweingruber, F.H., Osborn, T.J., 1998. Influence of volcanic eruptions on Northern Hemisphere summer temperature over the past 600 years. Science, 393, 450–455. Brovkin, V., Claussen, M., Petoukhov, V., Ganopolski, A., 1998. On the stability of the atmospherevegetation system in the Sahara/Sahel region. Journal of Geophysical Research, 103, 31613–31624.

10

Martin Claussen

Calov, R., Ganopolski, A., Petoukhov, V., Claussen, M., Greve, R., 2004. Transient simulation of the last glacial inception. Part I: Glacial inception as a bifurcation in the climate system. Climate Dynamics, 24(6), 545–562. Charney, J.G., 1975. Dynamics of deserts and droughts in the Sahel. Quarterly Journal of the Royal Meteorological Society, 101, 193–202. Claussen, M., 1997. Modelling biogeophysical feedback in the African and Indian Monsoon region. Climate Dynamics, 13, 247–257. Claussen, M., Gayler, V., 1997. The greening of Sahara during the mid-Holocene: results of an interactive atmosphere – biome model. Global Ecology and Biogeography Letters, 6, 369–377. Crowley, T., North, G., 1991. Paleoclimatology. Oxford Monographs on Geology and Geophysics, 18. New York: Oxford University Press, 339 pp. Ganopolski, A., Rahmstorf, S., 2001. Simulation of rapid glacial climate changes in a coupled climate model. Nature, 409, 153–158. Hann, J. 1883. Handbuch der Klimatologie. Engelhorn, Stuttgart. Hantel, M., Kraus, H., Scho¨nwiese, C.-D., 1987. Climate definition. In: Fischer, G. (Hrsg.): Climatology. Landolt-Bo¨rnstein, Functional Relationships in Science and Technology V/4/c1. Berlin: Springer. Hasselmann, K., 1976. Stochastic models. I. Theory. Tellus, 28, 473–485. Houghton, J.T., Ding, Y., Griggs, D.J., Noguer, M., van der Linden, P., Dai, X., Maskell, K., Johnson, C. I. (eds.), 2001. 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, 881 pp. von Humboldt, A., 1845. Kosmos, Entwurf einer physischen Weltbeschreibung. Band I, J.G. Cotta, Stuttgart and Tu¨bingen, 493 pp. Jones, P.D., Mann, M.E., 2004. Climate over past millennia. Reviews of Geophysics, 42, 1–42. Kraus, H., 2000. Die Atmospha¨re der Erde. Eine Einfu¨hrung in die Meteorologie. Vieweg. Braunschweig/Wiesbaden, 470 pp. Kutzbach, J.E., Guetter, P.J., 1986. The influence of changing orbital parameters and surface boundary conditions on climate simulations for the past 18 000 years. Journal of Atmospheric and Oceanic Science, 43, 1726–1759. Laut, P., 2003. Solar activity and terrestrial climate: some dubious correlations. Journal of Atmospheric & Solar-Terrestrical Physics, 65, 801–812. Lean, J., Beer, J., Bradley, R., 1995. Reconstruction of solar irradiance since 1610: implications for climate change. Geophysical Research Letter, 22, 3195–3198.

Lorenz, E.N., 1979. Forced and free variations of weather and climate. Journal of Atmospheric and Oceanic Science, 36, 1367–1376. de Menocal, P.B., Ortiz, J., Guilderson, T., Adkins, J., Sarnthein, M., Baker, L., Yarusinski, M., 2000. Abrupt onset and termination of the African Humid Period: Rapid climate response to gradual insolation forcing. Quaternary Science Review, 19, 347–361. McAvaney, B.J., Covey, C., Joussaume, S., Kattsov, V., Kitoh, A., Ogana, W., Pitman, A.J., Weaver, A.J., Wood, R.A., Zhao, Z.-C., 2001. Model evaluation. In: Houghton, J.T., Ding, Y., Griggs, D.J., Noguer, M., van der Linden, P., Dai, X., Maskell, K., Johnson, C.I. (eds.), 2001: 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, 881 pp. Peixoto, J.P., Oort, A.H., 1992. Physics of Climate. New York: American Institute of Physics, 111– 125. Prather, M., Ehhalt, D., Dentener, F., Derwent, R., Dlugokencky, E., Holland, E., Isaksen, I., Katima, J., Kirschoff, V., Matson, P., Midgley, P., Wang, M., 2001. Atmospheric chemistry and greenhouse gases. In: Houghton, J.T., Ding, Y., Griggs, D.J., Noguer, M., van der Linden, P., Dai, X., Maskell, K., Johnson, C.I. (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, 881 pp. Prentice, I.C., Jolly, D., BIOME 6000 members, 2000. Mid-Holocene and glacial-maximum vegetation geography of the northern continents and Africa. Journal of Biogeography, 27, 507–519. Prentice, I.C., Farquhar, G.D., Fasham, M.J.R., Goulden, M.L., Heimann, M., Jaramillo, V.J., Kheshgi, H.S., Le Que´re´, C., Scholes, R.J., Wallace, D.W.R., 2001. The carbon cycle and atmospheric carbon dioxide. In: Houghton, J.T., Ding, Y., Griggs, D.J., Noguer, M., van der Linden, P., Dai, X., Maskell, K., Johnson, C.I. (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, 881 pp. Rahmstorf, S., 2003. Timing of abrupt climate change: a precise clock. Geophysical Research Letter, 30, 1510–1513. Ramaswamy, Y., Boucher, O., Jaigh, J., Hauglustaine, D., Haywood, J., Myhre, G., Nakajima, T., Shi, G. Y., Solomon, S., 2001. Radiative forcing of climate change. In: Houghton, J.T., Ding, Y., Griggs, D.J., Noguer, M., van der Linden, P.,

Introduction to Climate Forcing and Climate Feedbacks Dai, X., Maskell, K., Johnson, C.I. (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, 881 pp. Schellnhuber, H.J., 1999. ‘Earth system’ analysis and the second Copernican revolution. Nature, 402, C19–C26. Shindell, D., Rind, D., Balachandran, N., Lean, J., Lonergan, P., 1999. Solar cycle variability, ozone, and climate. Science, 284, 305–308.

11

Svensmark, H., Friis-Christensen, E., 1997. Variation of cosmic ray flux and global could coverage – a missing link in solar–climate relationships. Journal of Atmospheric and Solar-Terrestrial Physics, 59, 1225–1232. Vitousek, P.M., Mooney, H.A., Lubchenco, J., Melillo, J.M., 1997. Human domination of Earth’s ecosystems. Science 277, 494–499. Wunsch, C., 2003. The spectral description of climate change including the 100ky energy. Climate Dynamics, 20, 353–363.

This page intentionally left blank

2. Insolation During Interglacial A. Berger1, M.F. Loutre1, F. Kaspar2 and S.J. Lorenz3 1

Universite´ catholique de Louvain, Institut d’Astronomie et de Ge´ophysique G. Lemaıˆtre, 1348 Louvain-la-Neuve, Belgium 2 Institut fu¨r Meteorologie, Freie Universita¨t Berlin, 12165 Berlin, Germany 3 Meteorologisches Institut, Universita¨t Hamburg, 20146 Hamburg, Germany

ABSTRACT The main insolation parameters are reviewed, in particular the energy received by the whole Earth over a full year and the 24-hour mean irradiance. Their spectral characteristics are underlined, and some remarks are made about the differences between insolation at the top of the atmosphere and at the Earth’s surface, the caloric and the astronomical seasons, the insolations in the tropical and in the high latitudes, the mid-month and the calendar insolations. The insolations characterizing interglacials marine isotope stage (MIS) 1 to MIS 11 are described, and their common features are discussed. 2.1 INTRODUCTION A link between climate and insolation forcing is indicated by numerous records of past environmental changes. Adhe´mar (1842) suggested that variations in the orbital configuration of the Earth, more specifically the precession of the equinoxes, could be responsible for glacial–interglacial cycles. In the second half of the nineteenth century, Croll (1875) approached the glaciation problem from the synergistic standpoint of the combined effects of all three major astronomical factors on seasonal insolation during perihelion and aphelion. A specific characteristic of his model essentially lies in his hypothesis that the critical season for the initiation of glacial stages is northern hemisphere (NH) winter. He argued that a decrease in the amount of

sunlight received during the winter favours the accumulation of snow and that any small initial increase in the size of the area covered by snow would be amplified by the snowfields themselves (positive feedback). Almost at the same time, Murphy (1876) adopted the opposite view that a long, cool summer and a short, mild winter provide the most favourable conditions for glaciation, a theory later recognized by Bru¨ckner et al. (1925). Today, the astronomical theory of palaeoclimates is mainly associated with the name of Milankovitch. Milankovitch (1941) was among the firsts to argue that the insolation during the NH summer could be the critical parameter and to take into account the effects of all astronomical parameters (eccentricity, obliquity and position of the perihelion). He integrated their effects to compute the insolation at the top of the atmosphere. Actually, Milankovitch (1941, p. 542 original edition) cites Ko¨ppen who ‘after an exhaustive discussion of all the possibility, . . . answered the question by indicating that it is the diminution of heat during the summer half-year which is the decisive factor in glaciation’ (p. 544 English edition). This discussion can be found in Ko¨ppen and Wegener (1924) where Ko¨ppen acknowledges the work of Penck and Bru¨ckner (1909) who suggested that glaciation is a problem not of enhanced precipitation, but of reduced ablation. An overview of the development of astronomical theories of palaeoclimate is given by Berger (1988) and by Paillard (2001). Algorithms for the precise calculation of the insolation patterns have been published previously (e.g. Berger,

14

A. Berger et al.

1978). Here, we present some fundamental reflections on the underlying mechanisms and discuss differences and similarities between the interglacials of the past 500 000 years. The article is supplemented with digital material containing tables with time series of insolation and plots for the interglacials.

solar constant, S0, the semi-major axis, a, of the Earth’s orbit around the Sun (the ecliptic), of its eccentricity, e, its obliquity, ", and the longitude of the perihelion measured ~ (Fig. 2.1). from the moving equinox, ! S0 is calculated at the mean distance, rm , from the Earth to the Sun. On the energy point of view, rm is given by: pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ r 2m ¼ a2 1 e2

2.2 ENERGY RECEIVED BY THE WHOLE EARTH OVER A FULL YEAR Assuming a perfectly transparent atmosphere, the energy, W, available at any given latitude on the Earth at any time during the year is a single-valued function of the

Therefore: Sa S0 ¼ pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ 1 e2

Sun Earth

Variation of the ‘eccentricity’ Today Summer solstice 22 June

Spring equinox 21 March 92.8 89.0

23°27′

Sun

93.6 89.8 Fall equinox 23 September

Winter solstice 22 December

Fig. 2.1 Top: the shape of the Earth’s orbit around the Sun is given by the eccentricity (it must be kept in mind that the semi-major axis is an invariant). Bottom left: The climatic precession parameter and the present-day configuration of the Earth’s orbit, with the winter solstice occurring very close to the perihelion. The length of the seasons is also indicated in days. Bottom right: the tilt of the Earth’s axis of rotation is called obliquity, and its present-day value is 23 279. These figures are reproduced from ‘European Science Foundation European Latsis Prize 2001’ and extended.

Insolation During Interglacial

where Sa is the energy received by unit of time on a unit area perpendicular to the Sun’s rays situated at a distance a from the Sun (Berger and Loutre, 1994). As a has no purely secular variations, Sa is only a function of the solar output. In the calculation of the long-term climatic variations over the Quaternary, Sa is supposed to be a constant equal to 1365 Wm2 in this paper.* e is a measure of the shape of the Earth’s orbit around the Sun. Its long-term variations are characterized by a mean period of about 100 kyr superimposed on a longer period of about 400 kyr, although their spectral structure is more complex. Its present-day value is equal to 0.016 and it varies between a maximum of 0.075 and a minimum of zero (the orbit is in this case circular) over the last millions of years. As a consequence of such a range of variation, S0 varies between 1365 and 1369 Wm2 for respectively the maximum and minimum values of e. From this, it follows that the energy received per unit area of the Earth’s surface and per unit of time over a full year is given by

WE ¼

S0 Sa ¼ pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ 4 4 1 e2

and varies by 1 Wm2 between the extrema of e. e is actually the only astronomical parameter which can change the total energy received by the whole Earth over one year. The other two parameters are redistributing the energy among latitudes and seasons. 2.3 PRECESSION AND OBLIQUITY Although the direct impact of e is rather small, e plays another much more A constant value of 1365 Wm2 is suggested by the Paleo Model Intercomparison Project (PMIP, phase 2) and is used in various modelling studies. The observed value during the last decades was slightly higher, but latest publications concluded that solar activity during the last decades was exceptionally high compared to the previous centuries (Fro¨hlich and Lean, 2004) and even millennia (Solanki et al., 2004). *

15

important role by modulating the amplitude of the climatic precession parameter, ~ This parameter is a measure of the e sin !. Earth–Sun distance at a fixed given time ~ in the year. The present-day value of !, following the calculation by Berger (1978) ~ is measured from the vernal where ! point, is 102 which means that the winter (December) solstice occurs close to the perihelion. This particular feature makes that the winter (summer) in the northern latitudes receives more (less) energy than usual which pleads in favour of using a long-term average for the insolation values instead of the present-day ones as reference. The average period of the long~ is 21 kyr resultterm variations of e sin ! ing from four main spectral components situated around 23 and 19 kyr. This parameter has clearly an opposite effect in the two hemispheres during their similar local seasons. About 12 kyr ago, when the summer solstice was at the perihelion, all the latitudes over the Earth received daily more energy than now during the astronomical NH summer (a season defined between the spring and fall equinoxes). This means that the NH latitudes received more energy during their summer and less during their winter, the reverse being equally true for the southern hemisphere (SH) latitudes during the SH seasons. The seasonal gradient was therefore magnified in the NH and reduced in the SH. " is the tilt of the Earth’s rotational axis relative to a perpendicular drawn to the plane of the ecliptic. Its present-day value is 23 279. It varies steadily with an average period of 41 kyr. An increase of " leads to an insolation increase in all latitudes during their respective summer and the reverse during winter. The seasonal gradient is therefore larger everywhere over the Earth. A decrease of " leads similarly to a decrease in insolation in the summer hemisphere and an increase in the winter hemisphere. As the strength of this effect is small in the tropics and maximum at the poles, it damps the

16

A. Berger et al.

seasonal cycle in the high latitudes of both hemisphere simultaneously. As glaciations occur also more or less in phase in both hemispheres, this is most probably one of the reasons for which the obliquity signal appears so clearly in most palaeoclimatic records in high latitudes. The spectral characteristics of the astronomical elements and of the insolations date back from the 1970s only. Emiliani (1955) was one of the firsts to estimate the mean periods of the astronomical parameters by counting the number of peaks from the Milankovitch curves. This led him with 92, 40 and 21 kyr for respec~ Hays et al. (1976) tively e, " and e sin !. used a spectral analysis technique which they applied on the numerical values of the astronomical parameters re-calculated by Brouwer and van Woerkom (1950) and Vernekar (1972). They found 125 and 96 kyr for e, 41 for " and 23 and 19 kyr for precession. In the mean time, Berger (1978) had completed his calculation, providing for the first time a full list of the periods characterizing the theoretical expansion of e (with periods of 400, 125, 95, 99, 131 and 2035 kyr), of " (with peri~ ods of 41, 54 and 29 kyr) and of e sin ! (with periods of 23.7, 22.4, 18.9 and 19.2 kyr) (see Berger, 1978 for a full list). It is the splitting of the 21-kyr precessional period into 23 and 19 kyr found independently in geological and astronomical data which was one of the first most delicate and impressive tests of the Milankovitch theory.

2.4 24-h MEAN IRRADIANCE ~ on the From these impacts of e, " and e sin ! total and the latitudinal and seasonal distributions of the energy received by the Earth from the Sun, the Milankovitch condition for entering into glaciation (1941) can be expressed in terms of orbital parameters. A minimum summer insolation in high

northern latitudes is indeed reached when (i) " is minimum, the seasonal gradient ~ is maxbeing then reduced in NH, (ii) e sin ! imum which implies that both the summer solstice occurs at aphelion (as it is more or less the case now) and e is maximum making the Earth–Sun distance at summer solstice, rA , even larger. As we have rA ¼ að1 þ eÞ; rP ¼ að1 eÞ; rP rA ¼ 2ae; rP being the Earth–Sun distance at perihelion, the difference in the total energy received by the Earth between perihelion and aphelion is proportional to four times the eccentricity. For a maximum of e (0.075), this means a difference reaching 30%; it is presently 6.4%. The Milankovitch hypothesis is also coherent with a warm winter allowing more precipitations over the continental high latitudes. However, a maximum eccentricity has a counter effect, although small, in increasing the total energy received by the Earth. All these considerations can be mathematically deduced from the theoretical calculation of the energy received from the Sun on a horizontal surface for each day and each latitude, . For the latitudes where there is a sunrise and a sunset every day ðjj < 2 jjÞ, the average energy over a time period of 24 h centred on solar noon (the 24-h mean irradiance) and further called incoming solar radiation (insolation) is given by: Sa a 2 W¼ ðH0 sin sin þ cos cos sin H0 Þ r where , the declination of the Sun, is related to its longitude, , by sin ¼ sin sin " being the angular distance from the spring equinox in the ecliptic;

Insolation During Interglacial

H0 is its hour angle, H, at sunset (at sunrise H ¼ H0 and at solar noon H ¼ 0). It is given by: cos H0 ¼ tan tan r is the distance from the Earth to the Sun given by r¼

að1 eÞ ; 1 þ e cosð !Þ

the numerical value of ! being equal to ~ þ 180 as discussed in Berger et al. (1993). ! r, and are assumed to be constant over one day. For all latitudes jj 2 jj (a) where there is a long polar night: < 0 W ¼0 Length of day ¼ 0 (b) where there is a long polar day: > 0 2 W ¼ S a sin sin r Length of day ¼ 24 hours Analysis of the spectral characteristics of W shows that for a given , (a/r)2 is a function of precession only and the third factor ðH0 sin sin þ cos cos sin H0 Þ is a function of " only through the declination ðÞ and the length of the day (2 H0). The mathematical expression of W allows to conclude (Berger et al., 1993) that its variations are (i) dominated by variations in precession (except close to the polar night), with the obliquity playing a secondary role but more important in high than in low latitudes; (ii) at the equinoxes, insolation is only a function of precession for all latitudes; (iii) at the solstices, both precession and obliquity

17

influence W , but precession is dominating at all latitudes (except close to the polar night). 2.5 INSOLATION AT THE TOP OF THE ATMOSPHERE AND AT THE EARTH’S SURFACE The values of W are usually referred to as ‘insolation at the top of the atmosphere’ because the perturbations by the atmosphere are not taken into account. Their pattern differs however from the pattern of radiation reaching the Earth’s surface, particularly in high latitudes where the surface albedo is large. Using a radiative, convective model, Tricot and Berger (1988) showed that the atmospheric attenuation essentially reduces the absolute variations of the incident solar radiation at the Earth’s surface as compared with the variations of the insolation at the top of the atmosphere. These variations are maximum in high latitudes, but the variations of the absorbed radiation at the Earth’s surface are maximum in tropical and middle latitudes related to the increase of the surface albedo with latitude. In the summer hemisphere, the largescale gradient of insolation between the tropics and the polar regions shows deviations from its present-day value the characteristic frequencies of which depend upon the type of insolation considered: (i) for the extraterrestrial insolation, the main frequency of the variations of the large-scale latitudinal gradient is about 40 kyr, whereas (ii) for the incident and mainly the absorbed insolation at the surface the large-scale gradient shows, in addition, quasi-periodicity of about 23 kyr; this difference is related to the atmospheric attenuation which reduces more strongly the variations of insolation at the surface in high latitudes than in tropics, preventing the obliquity signal to appear.

18

A. Berger et al.

2.6 CALORIC SEASONS This behaviour differs from those of the caloric insolations (i.e. the total solar energy received during a half-year caloric season) introduced by Milankovitch (1941). Such a caloric season is exactly half-a-year long, and the caloric summer counts all days for which insolation is larger than for any day of the caloric winter. These caloric insolations are mainly a function of precession in low latitudes and obliquity in high latitudes. Milankovitch’s idea of introducing the caloric seasons was to attempt at solving the difficulty of having to deal with both the total energy received during a given season and its length, which both vary in time. However, it is far from being that easy. The caloric seasons by Milankovitch raise indeed two problems: first, their start changes with time; second it is impossible to define them in the equatorial regions where the seasonal march of insolation has two minima and two maxima. 2.7 TROPICAL LATITUDES In the intertropical belt, the seasonal cycle of insolation is particular because the Sun comes overhead twice a year at each latitude (Berger and Loutre, 1997). At the equator for example, the maxima and the minima are reached at about the equinoxes and the solstices respectively (this is not exactly true because the Earth–Sun distance modulates the effect of the declination, but we may assume that the approximation is acceptable). Moreover, because of precession, the 24-h mean irradiance at the spring equinox will alternatively be larger and lower than at the fall equinox. The same holds for the solstice minima. If we assume that the climate responds to the equatorial absolute maximum, whether it occurs at spring or fall equinox, a 100-kyr (related to eccentricity) and a 11-kyr (half the precession period) appear very strongly in the long-term

variations of the equatorial insolation (Berger et al., 2004). Moreover, the ‘seasonal gradient’ (difference between the absolute maximum and the absolute minimum) introduces a 5.5-kyr periodicity related to one-fourth of the precessional cycle. 2.8 THE ANNUAL AND SEASONAL IRRADIATION The annual irradiation, that is, the total amount of solar energy received during one year at a given latitude does not depend on climatic precession. By adding all the daily values, one must take into account the fact that the Earth–Sun distance varies over the year (by up to 15% for the largest eccentricity value, see Section 2.3). But on the other hand, according to Kepler’s laws, the Earth travels faster at the perihelion than at the aphelion. Integrating over time, these two effects exactly compensate one another such that the total annual irradiation no longer depends on the Earth–Sun distance, but only on the inclination of the Earth’s axis of rotation with respect to the ecliptic, it means on obliquity. In both the northern and southern high latitudes, the total annual irradiation varies in phase with obliquity. Over the last 500 kyr, at 80 N, the maximum amplitude of the variation is 500 106 J m2 around a mean value of 5600 106 J m2 . At low latitudes, the total annual irradiation and obliquity are exactly out of phase. The phase reversal between high and low latitudes varies in time, occurring actually between 43 and 44 (N and S). At the equator, the maximum amplitude of the signal over the last 500 kyr is only 120 106 J m2 , and the mean value is 13 100 106 J m2 . In both the high and low latitude cases, the obliquity variation explains nearly 99% of the variance, the rest being related to eccentricity. Likewise, the total irradiation (in J m2 ) received at a given latitude between two given orbital positions of the Earth, defined

Insolation During Interglacial

by their true longitude, depends only on obliquity and eccentricity. Seasonal irradiations (astronomical or meteorological) are particular cases. For example, the total summer (JJA) irradiation at 80 N varies over the last 500 kyr in phase with obliquity with an amplitude of 250 106 J m2 , i.e. 9% of its mean value. At the equator, the insolation signal is in phase opposition with obliquity and its amplitude is only 30 106 J m2 , i.e., 0:9% of its mean value. However, the duration of a time interval through the year, a season in particular, is a function of precession. The astronomical seasons are the periods during which the Sun traverses the quadrants of the ecliptic, counted from the vernal equinox. Since the areas of the corresponding sectors are not equal, and since the radius vector of the Sun sweeps out equal areas in equal time, the four seasons have different durations. As an example, over the last 500 kyr, the length of astronomical summer varies between 83.6 and 99.8 days. Consequently, the average insolation over a season (in W m2 ) depends both on precession and on obliquity. However, " being developed around a constant value, the ratio between obliquity and precession is dominated by precession to the first order of magnitude. This leads to a summer mean irradiance at the equator strongly dominated by precession. The power of the obliquity component increases towards the high latitudes, but precession remains the dominant component of the summer mean irradiance for all latitudes. 2.9 MID-MONTH AND CALENDAR INSOLATIONS The longitude varies from 0 (assumed to define the time of spring equinox, which is arbitrarily fixed at March 21 if a calendar has to be used) to 360 , with ¼ 90 , 180 and 270 defining respectively the summer solstice, fall equinox and winter solstice.

19

¼ 0, 30 , 60 . . . define the so-called midmonth values which are actually for around the 20th day of each month. Because the length of the astronomical seasons varies in time (Berger and Loutre, 1994), these midmonth values are not related to a fixed calendar date. If such calendar dates need to be used, the mean longitude m must be used which is linked to through a formula ~ In order to illustrate involving e and !. the difference between the mid-month and the calendar values, a comparison is first made between the 60 N daily insolation at present and 10 000 years before present (BP), for ¼ 210. The difference between these ‘‘mid-month’’ insolations amounts to 5 Wm2 , but ¼ 210 presently refers to 24 October ð109 Wm2 Þ and, at 10 000 years BP, it referred to 16 October ð104 Wm2 Þ. Thus, the difference reflects mainly the secular changes of both obliquity and shape of the ecliptic. Second, if a calendar date insolation is considered, the long-term variations of the length of the astronomical seasons is explicitly recognized. For the same latitude and years, if daily insolations at 16 October are compared, a difference of 29 Wm2 is found. This is a result of the fact that on 16 October the true longitude of the Earth is presently 202 ð133 Wm2 Þ and, at 10 000 years BP, 210 (104 Wm2 ). Another example, Fig. 2.2c and 2.2d illustrates the insolation anomaly at 125 kyr BP relative to today’s conditions. In Fig. 2.2c, anomalies are calculated for the same position on the orbit, that is, the same true solar longitude. In Fig. 2.2d, the anomalies are calculated based on the classical calendar. For both figures, vernal equinox has been used as a reference. Therefore, differences between both figures are small around March. Differences are largest in autumn. At 125 kyr BP, the autumnal equinox is at 12 September, whereas it is at 23 September today. This difference of 11 days leads to significantly shorter summers (85 days 125 kyr ago against 94 today). The largest differences

20

A. Berger et al. 90N

90N

480

60N

60N 3

480

30N

30N

460 440 EQ

420 400 375 350 300 250 200 150 100 50 5

440

30S

60S

440 –15 EQ

440

–9

30S –3

–3

60S

90S

90S –60 Jan

(a)

–30 Feb

0 30 60 90 120 150 180 (VE) Apr May Jun Jul Aug Sep True longitude on Earth’s orbit from vernal equinox (VE)

210 Oct

240 Nov

(b)

50 100 150 200 250 300 350 375 400 420 440 460 480 500 520 540 560 580 600

5

–60 Jan

270 Dec

90N

–30 Feb

0 30 60 90 120 150 180 (VE) Apr May Jun Jul Aug Sep True longitude on Earth’s orbit from vernal equinox (VE)

–42–39 –36–33–30–27–24–21–18–15–12–9–6–3 3

6

210 Oct

240 Nov

270 Dec

9 12 15 18 21 24 27 30 33 36 39 42

90N

60

–10

60N

–10

60N

–40

–20 –20

30N

–30

30N –30

50 EQ

EQ

40 30

30S –50

30S

20

60S

90S

–60 Jan

(c)

–30 Feb

0 30 60 90 120 150 180 (VE) Apr May Jun Jul Aug Sep True longitude on Earth’s orbit from vernal equinox (VE)

210 Oct

240 Nov

270 Dec

90S

60

–60

60S

10

0

30 Jan

(d)

–65 –60 –55 –50 –45 –40 –35 –30 –25 –20 –15 –10 –5 5 10 15 20 25 30 35 40 45 50 55 60 65

60 Feb

90 Mar

120 150 180 210 240 270 300 330 360 Apr May Jun Jul Aug Sep Oct Nov Dec (annual cycle)

–65 –60 –55 –50 –45 –40 –35 –30 –25 –20 –15 –10 –5 5 10 15 20 25 30 35 40 45 50 55 60 65

90N

90N

–40 60N

60N

–10

10 –30

60

30N

30N

50 EQ

–20 EQ

40 30

30S

20

10

10

60S

60S

90S

90S –60 Jan

–30 Feb

0 (VE)

30 Apr

60 May

90 Jun

120 Jul

150 Aug

180 Sep

210 Oct

240 Nov

270 Dec

–65 –60 –55 –50 –45 –40 –35 –30 –25 –20 –15 –10 –5 5 10 15 20 25 30 35 40 45 50 55 60 65

10

–60 Jan

–30 Feb

0 (VE)

30 Apr

60 May

90 Jun

120 Jul

150 Aug

180 Sep

210 Oct

240 Nov

270 Dec

True longitude on Earth’s orbit from vernal equinox (VE)

True longitude on Earth’s orbit from vernal equinox (VE)

(e)

–10

30S

20

(f)

–65 –60 –55 –50 –45 –40 –35 –30 –25 –20 –15 –10 –5 5 10 15 20 25 30 35 40 45 50 55 60 65

Fig. 2.2-a (left) Present-day insolation as absolute values. Fig. 2.2-b (right): Present-day insolation as deviation from the mean of the last 800 000 years. Fig. 2.2-c (left) Insolation at 125 kyr as deviation from the present-day value calculated by comparing same positions on the orbit with a fixed vernal equinox. Fig. 2.2-d (right): Same as (2-c), but calculated by comparing same calendar dates with a fixed vernal equinox. Insolation anomaly at 128 kyr Fig. 2.2-e (left) and at 115 kyr Fig. 2.2-f (right) as deviation from the present-day values. Fig. 2.2-a to Fig. 2.2-f Distribution of insolation [Wm2] over latitudes and date in the year. Except for Fig. 2.2-d, anomalies are calculated based on a comparison of the same position on the Earth’s orbit (measured from vernal equinox). Anomalies in Fig. 2.2-c to Fig. 2.2-f refer to today’s values.

Insolation During Interglacial

2.10 CALENDAR DEFINITION AND PALAEOCLIMATE SIMULATIONS The decision how monthly or seasonal differences are calculated is of particular importance when results of climate simulations are analysed as anomalies from today’s climate. Joussaume and Braconnot (1997) discussed this problem for simulations of the climate at 126 000 yr BP with an atmosphere general circulation model (GCM). They compared simulated monthly temperatures over Europe and precipitation over central Africa using the calendar and mid-month definitions with a vernal equinox fixed at 21 March. The annual cycle showed similar features for both definitions, but with important differences in autumn. These differences between the two definitions were as large as the differences between 126 kyr BP and the present. As autumn started earlier at 126 kyr BP, it was already colder for the same calendar date over Europe and precipitation was weaker over Africa. Large September temperature anomalies due to the definition of the calendar were visible over large parts of NH’s continents, reaching a maximum of more than 10 C over eastern Asia. This difference was larger than the internal variability of the model and exceeds the differences between 126 kyr BP and today significantly. The differences observed by Joussaume and Braconnot (1997) are not fully caused

by direct consequences of the calendar definition, but are partly attributable to a calendar hidden in boundary conditions. As they used an atmosphere-only GCM, boundary conditions (sea surface temperature and sea-ice temperature) had to be prescribed. They had been fixed to the present-day cycle. By performing experiments in which the prescribed sea-ice temperature was replaced by a simple sea-ice model, they showed that the ‘hidden calendar’ contributed to a bias in the results. As climate models of the current generation typically contains at least simple ocean and sea-ice models, the problem of a calendar hidden in the boundary conditions is reduced, but can still occur due to other parameters, e.g. related to vegetation. The magnitude of this ‘calendar problem’ varies in time and is mainly related to the climatic precession. To illustrate the effect, Fig. 2.3 shows the day in the year (counted from 1 January) of the equinox in boreal autumn (true solar longitude is equal to 180 , with vernal equinox fixed at 21 March). The largest deviation for autumnal equinox during the last 500 000 years occurred at 198 kyr BP (8.9.: 15 days) and 209 kyr BP (1.10.: þ13 days). The difference between these two extremes is therefore 29 days. During this period, eccentricity had its Day of autumnal equinox within the year

between the two definitions are visible on those latitudes, where absolute values show strong differences within a few days. This is mainly the case for high northern and southern latitudes (see Fig. 2.2a). On these latitudes, anomalies of more than 60 Wm2 occur when they are calculated on the basis of the classical calendar definition. If anomalies are calculated according to the astronomical longitude, they are close to zero for September and October in the high northern and southern latitudes.

21

275

270

265

260

255

250 0

–50 –100 –150 –200 –250 –300 –350 –400 –450 –500

Time (kyr) from 1950 AD

Fig. 2.3 Day at which the autumnal equinox occurs (i.e. true longitude of the Sun is 180 ), when vernal equinox is fixed at 21 March. Days are counted from 1 January.

22

A. Berger et al.

maximum. Owing to this variation in time, the consequences also have to be considered when interpreting transient simulations (see also discussion in Kubatzki et al., Chapter 39, this volume). 2.11 CHARACTERISTICS OF INSOLATION DURING INTERGLACIALS In this section, we will discuss the insolation during the interglacials, with a special emphasis on the interglacial prior to the Holocene, i.e. marine isotope substage (MIS) 5e (also referred to as 5.5.) in the marine records, more or less equivalent to the Eemian in the continental records. It is characterized by large values of the eccentricity (up to 0.0414 at 115 kyr BP), although the largest values of the last 500 kyr are reached during MIS 7 (0.0503 at 207 kyr BP). As eccentricity modulates the amplitude of the precession signal, this is also large. Moreover, according to climatic precession, NH summer occurred at perihelion at 127 kyr BP and at aphelion at 116 kyr BP. Over the same time, obliquity varies from a maximum value (24.259 ) at 131 kyr BP to a minimum value (22.316 ) at 112 kyr BP.

Consequently, the insolation at 65 N at the summer (June) solstice is maximum at 128 kyr BP ð547 Wm2 Þ. It is actually the largest value reached over the last 200 000 yr. Nevertheless, it must be kept in mind that the climate system is driven by the time evolution of the distribution of insolation during the year and along the latitude, so other latitudes and days deserve to be accounted for (see Fig. 2.2a to 2.2f). Some features of the interglacials of the last 500 kyr are summarized in Table 2.1. Figure 2.2e, 2.2c and 2.2f show the distribution of insolation over latitudes and seasons for respectively 128 kyr BP, 125 kyr BP and 115 kyr BP as anomaly from today. 125 kyr BP is often used for equilibrium experiments with GCMs (e.g. chapters 33, 34, 38). Although 3000 years after the maximum of insolation at 65 N, this date is selected because conventionally the minimum global ice cover occurs approximately at that time, which is considered as the climate optimum (Kukla et al., 2002). For example, insolation at 65 N decreases by 2.4% between 128 and 125 kyr BP. Compared to these dates, the angle of perihelion at 115 kyr BP was almost opposite, but similar to today. This date represents approximately the end of the warm phase (Kukla

Table 2.1 Orbital features and mid-month June insolation at selected dates during interglacials of the last 500 000 years

MIS 1 MIS 5 MIS 7 MIS 9 MIS 11

Date1 (ka)

Eccentricity

Climatic precession2

Obliquity (degrees)

Mid-month June insolation at 65 N (Wm2 )

6 11 122 128 216 219 330 334 405 411

0.018682 0.019529 0.040744 0.039017 0.048858 0.047839 0.034387 0.031539 0.019535 0.019057

FE SS FE SS FE (end summer) SS (early summer) FE (end summer) SS FE (end summer) SS (early summer)

24.105 24.201 23.336 24.130 24.308 23.947 24.099 24.238 23.208 24.023

506.60 528.45 498.23 548.32 529.86 551.96 514.61 541.95 496.22 522.68

1 For each marine isotope stage (MIS), the first date corresponds to the peak of the interglacial according to SPECMAP (Martinson et al., 1987) and the second one to the maximum of mid-month June insolation prior to the peak of the interglacial. 2 Here we give the approximate time in the year at perihelion (FE stands for fall equinox and SS for summer solstice). The exact time of passage at perihelion can be obtained from the values in the supplementary tables.

Insolation During Interglacial

et al., 2002) and is therefore used in experiments examining the start of the glaciation. It must be kept in mind however that it is difficult to have an absolute dating of palaeoclimate records. At best, the radioactive methods are accurate within a few per cent. For stage 5, it means a few thousands years which is therefore more or less the same order of magnitude as the assumed lag between insolation and proxy records. Let us first consider the long-term variations of the deviation from present day of the insolation at the summer (June) solstice for time slices of 30 kyr centred at the peak of the interglacials from the northern to the southern poles (Fig. 2.4 and figures in the supplementary material). This latitudinal distribution around MIS 5e (Fig. 2.4) displays strong positive deviations going from 137 to 121 kyr BP, over all the latitudes. The largest deviation is in the northern polar regions at 128 kyr BP (larger than 70 Wm2 ). Negative deviations follow from 121 to 111 kyr, with the most negative value at the North Pole at 116 kyr BP (lower than 40 Wm2 ). Between these dates, the rate of change at the North Pole is about 10 Wm2

23

per kyr. It is of the same order for the other interglacials, except for MIS 11 and MIS 1 where it is slightly lower. This pattern of large positive deviations before the peak of the interglacial, much smaller deviations at the peak of the interglacial (even negative deviations) and large negative deviations after the peak of the interglacial is similar for the different interglacials over the last 500 kyr (supplementary figures). However, the most negative deviation after MIS 7.3 does not appear in the polar region but rather in the equatorial one. The peak of the interglacials lags behind the insolation maximum by 3 to 6 kyr. To follow, the deviations from present day of the insolation at 65 N will be investigated all through the year for the same 30-kyr time intervals (Fig. 2.5 and supplemental material). Over the time interval 137–107 kyr BP, the maximum positive deviation occurs in June at 128 kyr BP (Fig. 2.5). Between 137 and 128 kyr BP, the largest deviation occurs earlier in the year; it occurs later in the year after 128 kyr (Fig. 2.5). This is related to the 2 kyr phase shift of the insolation, from one month to the next. Moreover, the

80N –40

60N –30

40N

60

Latitude

20N –20

EQ 40 40

–10

20S

30

0

40S 60S

20 10

20

10

80S –108

–111

–114

–117

–120

–123

–126

–129

–132

–135

Time (kyr) from 1950 AD –70 –60 –50 –40 –35 –30 –25 –20 –15 –10 –5

5

10

15

20

25

30

35

40

50

60

70

Fig. 2.4 Latitudinal distribution of the time evolution of mid-month insolation (Wm2) at June solstice from 137 to 107 kyr (i.e. including MIS 5.5) as deviation from the present-day values.

24

A. Berger et al.

–108

–10

–111 10

–114

Time (kyr) from 1950 AD

–20

20

–117 10

–120

20 –20 40

–123 –10

–126 –129 –10

60

–132 40

–135

20 10

–60 Jan

–30 Feb

0 (VE)

30 Apr

60 May

90 Jun

120 Jul

150 Aug

180 Sep

210 Oct

240 Nov

270 Dec

True longitude on earth’s orbit from vernal equinox (VE) –80 –70 –60 –50 –40 –30 –25 –20 –15 –10 –5

5

10 15 20 25

30 40 50 60 70 80

Fig. 2.5 Time evolution of the annual cycle of mid-month insolation at 65 N (Wm2) from 137 to 107 kyr as deviation from the present-day value.

following largest negative deviation after the peak of the interglacial also occurs in June. The general pattern is similar for the other interglacials (supplementary figures), although the amplitude is smaller then, except for MIS 7. The positive deviation is larger at MIS 7.3 than at MIS 5e. Moreover, the negative deviation at MIS 7.2 does not display one single maximum in June but rather a double one, in April and August. Finally, the long-term changes in the latitudinal and seasonal distributions of the deviations from present-day insolations can be documented over these time slices (Fig. 2.2a to 2.2f, supplementary figures and movies). For MIS 5e, starting from 137 kyr BP, the deviations become positive at the beginning of the year. The largest values are initially in the January–February SH and propagate very quickly (132 kyr BP) to the NH in the NH spring. The maximum deviation is then reached (128 kyr BP) in June at the North Pole. At that time, the deviations are positive in NH spring and NH summer over the whole Earth. Then, the positive deviation moves towards NH winter and it fades away. At the same time

(starting from 128 kyr BP), a large negative deviation appears, first in the tropical SH in January. At 122 kyr BP (the peak of the interglacial according to SPECMAP; Martinson et al., 1987), the deviation from the present day is negative in NH winter and NH spring, with a maximum negative value in the southern tropical area in February. It is positive in NH summer and NH autumn, with a maximum value in the northern tropical area in August. The negative deviations propagate towards the North with smaller amplitude and then culminate in the northern polar regions in June (116 kyr BP). The other interglacials display a similar behaviour, although the amplitude might be different. As discussed theoretically, the seasonal and annual irradiations (Fig. 2.6) at a given latitude only depend on the obliquity (as well as very slightly on eccentricity). Moreover, the length of the season only depends upon precession. Consequently, the seasonal mean irradiance is a function of both obliquity and precession. For MIS 5e, the total irradiation during the astronomical NH summer at 65 N displays a maximum

Insolation During Interglacial

maximum at 127 kyr BP. It must be underlined that, for all the interglacials, the timing of maximum of NH summer mean irradiance is driven by the timing of the minimum of the length of the season and not by the timing of the total seasonal irradiation. Moreover, the interglacial peaks only a few thousand years after the maximum in NH summer mean irradiance. The strong imprint of obliquity on the total NH summer and annual irradiations also shows up very clearly in the latitudinal distribution of their deviations from the present day (Fig. 2.7 and supplementary material). For the NH summer, there is a latitudinal reversal in the sign of the deviation at 12 N. This reversal is double, at 42 N and 42 S, for the annual value, that is, the deviation has the same sign polewards of these latitudes. During the last interglacial, a reversal in time in the total NH summer irradiation occurs at 122 kyr BP, that is, at the peak of the interglacial. However, it is not a general feature of all the interglacials. Indeed, the peak of the interglacial occurs either during a positive deviation in the NH (MIS 1, MIS 7 and

BER78

3000 2950 2900 2850

100

2800

95 90 85

420 400 380 360 340

215

210 0

–100

–200

–300

–400

Fig. 2.6 From top to bottom: time evolution between 450 and 50 kyr AP of the total summer irradiation (106Jm2) at 65 N, the length of the summer season (day), the summer mean irradiance (Wm2) at 65 N and the annual mean irradiance (Wm2) at 65 N.

value at 131 kyr BP, NH summer is the shortest at 127 kyr BP, and consequently, NH summer mean irradiance reaches a 80N

25

–140

100

–100

60N

60 –60

40N 20

–20

Latitude

20N EQ

20 –20

20S 40S 60S –20

20

80S –108

–111

–114

–117

–120

–123

–126

–129

–132

–135

Time (kyr) from 1950 AD –160 –140 –120 –100 –80 –60 –40 –20

–5

0

5

20

40

60

80

100 120 140 160

Fig. 2.7 Variation between 137 and 107 kyr BP of the deviation from the present-day value of the summer irradiation (JJA, 106Jm2).

26

A. Berger et al.

80N 60N

Latitude

40N 20N

–20

EQ

40 –10

20S

–10

20

40S

20 0

60S

10

10

80S –108

–111

–114

–117

–120

–123

–126

–129

–132

–135

Time (kyr) from 1950 AD –60

–50

–40

–30

–20

–10

–5

0

5

10

20

30

40

50

60

Fig. 2.8 Variation between 137 and 107 kyr BP of the deviation from the present-day value of the summer mean irradiance (JJA, Wm2).

MIS 9), during a negative deviation in the Northern Hemisphere (MIS 11) or at the transition (MIS 5). There is a better agreement between the pattern of the deviation for the JJA mean irradiance for different interglacials (Fig. 2.8 and supplementary figures). The positive deviation over the Northern Hemisphere prior to the peak of the interglacial is replaced by a negative deviation following it. The largest amplitudes are during MIS 7.3.

REFERENCES Adhe´mar, J.A., 1842. Re´volutions de la Mer: De´luges Pe´riodiques, Carilian-Goeury et V. Dalmont, Paris. Berger, A., 1978. Long-term variations of daily insolation and Quaternary climatic changes. Journal of Atmospheric Sciences, 35(12), 2362–2367. Berger, A., 1988. Milankovitch theory and climate. Reviews of Geophysics 26(4), 624–657. Berger, A., Loutre, M.F., 1994. Long-term variations of the astronomical seasons. In: Topics in Atmospheric and Interstellar Physics and Chemistry, Cl. Boutron (ed.), 33–61, Les Editions de Physique, Les Ulis, France.

Berger, A., Loutre, M.F., 1997. Intertropical latitudes and precessional and half precessional cycles. Science, 278, 1476–1478. Berger, A., Loutre, M-F., Tricot, C., 1993. Insolation and Earth’s orbital periods. Journal of Geophysical Research, 98 (D6), 10.341–10.362. Berger, A., Loutre, M.F., Me´lice, J.L., 2004. 100-kyr and 5.5-kyr periods in tropical insolation. AGU, PP16-A Tropical Perspectives on the Ice Ages, section Paleoceanography and Paleoclimatology, San Francisco, 14 December 2004. Brouwer, D., van Woerkom, A.J.J., 1950. Secular variations of the orbital elements of principal planets. Astron. Papers Am. Ephem., 13(2), 81–107. ¨ ber Bru¨ckner, E., Ko¨ppen, W., Wegener, A., 1925. U die Klimate der geologischen Vorzeit, Zeitschrift fu¨r Gletscherkunde, 14. Croll, J., 1875. Climate and Time in Their Geological Relations. Appleton, New York. Emiliani C.R.W., 1955. Pleistocene temperatures. Journal of Geology, 63(6), 538–578. Fro¨hlich, C., Lean, J., 2004. Solar radiative output and its variability: evidence and mechanisms. Astron. Astrophys. Rev., 12, 273–320. Hays, J.D., Imbrie, J., Shackleton, N.J., 1976. Variations in the Earth’s orbit : Pacemaker of the ice ages. Science, 194, 1121–1132. Joussaume, S., Braconnot, P., 1997. Sensitivity of paleoclimate simulation results to season definition. Journal of Geophysical Research, 102(D2), 1943–1956.

Insolation During Interglacial Ko¨ppen V., Wegener, A., 1924. Die Klimate der geologischen Vorzeit. Verlag Gebru¨der Borntraeger, Berlin, 255 pp. Kukla, G.J., Bender, M.L., de Beaulieu, J.L., Bond, G., Broecker, W.S., Cleveringa, P.J., Gavin, J.E., Herbert, T.E., Imbrie, J., Jouzel, J., Keigwin, L.D., Knudsen, K.-L., McManus, J.F., Merkt, J., Muhs, D.R., Mu¨ller, H., Poore, R.Z., Porter, S.C., Seret, G., Shackleton, N.J., Turner, C., Tzedakis, P.C., Winograd, I.J., 2002. Last interglacial climates. Quaternary Research, 58, 2–13. Martinson, D.G., Pisias, N.G., Hayes, J.D., Imbrie, J., Moore, T.C., Shackleton, N.J., 1987. Age dating and the orbital theory of the ice ages: development of a high-resolution 0 to 300,000-year chronostratigraphy. Quaternary Research 27, 1–29. Milankovitch, M., 1941. Kanon der Erdbestrahlung und seine Anwendung auf das Eiszeitenproblem. Royal Serbian Sciences, Spec.pub.132, Section of Mathematical and Natural Sciences, Vol. 33, pp. 633, Belgrade (‘‘Canon of Insolation and the Ice Age problem’’, English Translation by Israe¨l Program for Scientific Translation and published for the U.S. Department of Commerce and the

27

National Science Foundation, Washington D.C., 1969, and by Zavod za Udzbenike I nastavna Sredstva in cooperation with muzej nauke I technike Srpske akademije nauka I umetnosti, Beograd, 1998). Murphy, J.J., 1876. The glacial climate and the polar ice-cap. Q.J. Geol. Soc. London, 32, 400–406. Paillard, D., 2001. Glacial cycles: Toward a new paradigm. Reviews of Geophysics, 39(3), 325–346. Penck, A., Bru¨ckner, E., 1909. Die Alpen im Eiszeitalter, Tauchnitz, Leipzig. Solanki, S.K., Usoskin, I.G., Kromer, B., Schu¨ssler, M., Beer, J., 2004. Unusual activity of the Sun during recent decades compared to the previous 11,000 year. Nature, 431, 1084–1087. Tricot, Ch., Berger, A., 1988. Sensitivity of present day climate to astronomical forcing. In: Long and Short Term Variability of Climate, H. Wanner, U. Siegenthaler (eds), 132–152, Earth Science Series, Springer Verlag. Vernekar, A.D., 1972. Long-period global variations of incoming solar radiation. Meteorol. Monograph, 12(34), 130 p.

This page intentionally left blank

3. A Survey of Hypotheses for the 100-kyr Cycle Martin Claussen1, Andre´ Berger2 and Hermann Held3 1

Meteorological Institute, University Hamburg, and Max Planck Institute for Meteorology, Bundesstr. 53, D-20146 Hamburg, Germany 2 Universite´ catholique de Louvain, Institut d’Astronomie et de Ge´ophysique G. Lemaıˆtre, 1348 Louvain-la-Neuve, Belgium 3 Potsdam Institute for Climate Impact Research, Telegrafenberg A31, D-14473 Potsdam, Germany

ABSTRACT Theories and mathematical models of longterm Quaternary climate variations are briefly summarized and revisited. We conclude that the problem of Quaternary climate variations, in particular the existence of a dominant 100-kyr ice-age cycle is, even approximately 160 years after first geological evidence of ice ages was found, not yet solved. However, we have some clues on what elements a theory of Quaternary Earth system dynamics should consist of. Assessment of a number of conceptual models – ranging from models in which forcing is necessary to yield observed climate variability to models of free climate oscillations – cannot favour any model over its competitor on the grounds of tuning each model to the time series of global ice volume. Hence, geographically explicit fully coupled climate system, or natural Earth system, models are required to analyse the system’s response to geographically varying forcing and internal feedbacks. Evidence emerges that much of Quaternary climate variability arises due to internal feedbacks, with ice sheets and biogeochemical cycles as critical elements and orbital forcing as pacemaker. 3.1 A BRIEF HISTORY OF THEORIES OF ICE AGES In 1842, briefly after L. Agassiz proposed the existence of ice ages on the grounds of

geological evidence, J.A. Adhe´mar suggested the first astronomical theory of climate change on the basis of the known precession of equinoxes. During the following decades, glacial geology became strongly tied to the astronomical theory which was advanced by J. Croll in the 1860s (references to Agassiz, Adhe´mar and Croll, see Paillard, 2001). Since Croll’s theory appeared to be more and more at variance with emerging geological evidence, his theory was eventually refuted. In 1896, Arrhenius concluded that ‘it seems that the great advantage which Croll’s hypothesis promised to geologists, viz. of giving them a natural chronology, predisposed them in favour of its acceptance. But this circumstance, which at first appeared advantageous, seems with the advance of investigation rather to militate against the theory, because it becomes more and more impossible to reconcile the chronology demanded by Croll’s hypothesis with the facts of observation’ (Arrhenius 1896, p. 274). The astronomical theory was modified and advanced by Milankovitch (1941) and by Ko¨ppen and Wegener (1924) at the beginning of the last century; however, it was disputed because it did not seem to be supported by geological data (Crowley and North, 1991). The orbital theory saw a strong revival after new geological evidence presented by J.D. Hays, J. Imbrie and N.J. Shackleton in 1976 (Hays et al., 1976) corroborated many of its predictions advanced and refined by A. Berger in the 1970s (Berger, 1977, 1978).

Martin Claussen, Andre´ Berger and Hermann Held

The theory of ice ages in which geochemical reactions and CO2 play a major role is perhaps as old as the astronomical theory. Early work goes back to J.J. Ebelmen in the 1840s, J. Tyndall in 1861 and S. Arrhenius in 1896. Arrhenius, for example, was convinced that changes in atmospheric transparency (due to changes in atmospheric CO2) would ‘prove useful in explaining some points in geological climatology which have hitherto proved most difficult to interpret’ (Arrhenius 1896, p. 275). Also today, there are models, e.g. by G. Shaffer and M.E. Raymo, developed in the 1990s (Shaffer, 1990; Raymo et al., 1997 ), which could be described as biogeochemical oscillators in which ocean biogeochemistry is the key player. 3.2 THE 100-kyr PARADOX Climate archives of the Pleistocene reveal variations in oxygen isotopes, with a dominant periodicity of approximately 40 kyr in the early Pleistocene, approximately 2 to 1 million years before present. In the late Pleistocene, the dominant periodicity of climate variations shifts to 100 kyr (e.g. Ruddiman et al., 1986). The amplitude of climate variations – interpreted as change in temperature and ice volume – seemingly increased. In particular, a tendency towards warmer, or more ice free, interglacials was found for the last 500 kyr in ice cores (see EPICA community Members, 2004, and figures in Chapter 4) as well as in marine isotopes (Fig. 3.1). At a first glance, the periodicities found in proxy data seem to coincide with those of the three astronomical parameters which characterize the Milankovitch theory. Milankovitch (1941) showed that meridional and seasonal changes in insolation consist of changes in the eccentricity of the Earth’s orbit, in the obliquity and in the precession of equinoxes. Berger (1977, 1978) demonstrated from analytical

–0.5

δ18O (‰)

30

–1.0 –1.5 –2.0 2.0

1.5

1.0

0.5

0.0

Age (million years)

Fig. 3.1 Changes in 18 O reconstructed from marine sediment cores for the last 2 million years. These changes are interpreted as changes in global ice volume where increasing 18 O-values indicate increasing ice volume. The curve represents data from the tropical western Pacific sedimentary core ODP 806. This figure is taken from Saltzmann (2002) with permission of Academic Press.

developments of these astronomical parameters that the dominant periods are 400, 125 and 95 kyr for eccentricity, a dominant period at 41 kyr for obliquity and a bimodal period at 23 and 19 kyr for precession. However, the amplitudes of insolation changes due to changes in eccentricity appear to be much weaker than the amplitudes associated with obliquity and precessional variations, while proxy data show just the opposite. Hence, all theories employing orbital forcing in some way or the other need to resolve this so-called ‘100-kyr paradox’: the coexistence of a weak 100-kyr component in the astronomical forcing and a strong one in the response. The paradox also applies to theories involving other astronomical parameters than those used in the traditional Milankovitch theory. Again, the radiative forcing implied by these other astronomical forcings appears to be rather small, and thus, some additional amplifier is required to explain the dominance of the 100-kyr period. 3.3 FREE MODELS A wealth of models have been suggested to explain Pleistocene climate variations, in particular the 100-kyr cycle. According to Saltzman (2002), these models can be categorized into ‘forced’ and ‘free’ models. In

A Survey of Hypotheses for the 100-kyr Cycle

forced models, astronomical forcing (which explicitly includes the frequencies of orbital variations) is necessary to obtain the observed frequencies of climate variations. Free models do not rely on astronomical forcing at all, but the 100-kyr period appears as an internal (free) oscillation perhaps due to an internal instability of the system. In terms of dynamical systems analysis, the 100-kyr oscillation is the consequence of a bifurcation* to a self-sustained oscillation which is driven by an instationarity of the Earth system. Such free oscillators can address ice masses interacting with another variable: either ice-sheet location, bedrock depression or the thermohaline circulation. The free oscillator is forced by steady long-term variations, for example by a tectonic forcing associated with a slow variation in CO2 outgasing. An illustrative example of a free model is given by Paul and Berger (1997). They construct a nonlinear oscillator, which mimics ice-sheet dynamics as function of insolation, accumulation and bedrock response:

31

glacial – interglacial sea-level variations of the late Pleistocene when orbital forcing is applied – of course, because it is designed to do so. A conceptual model like this is not designed as a predictive tool, but an exploratory tool. An interesting result, however, is that even with a random, white noise forcing, the model recaptures the asymmetric, sawtooth like 100-kyr cycle. This result suggests that the 100-kyr cycle could emerge as an internal property of the dyna-mics of some climate system component – for example the ice sheet – independent of any specific cyclicity in the forcing. Another example of a (more elaborate) freely oscillating model that includes ice-sheet dynamics with basal melting and sliding is given by Saltzman and Verbitzky (1993). An essential point is that the free models do not require amplification of small forcing and explain climate changes over the Pleistocene as simply the response of an internal oscillator that could in some cases even resist additional forcing over a variety of timescales.

dV ¼ a þ ib V c ðmðVÞÞd dt where t is the time, V is the nondimensional ice volume, a is a constant accumulation, i is the normalized ð0 < i < 1Þ insolation and m is a memory function m ¼ Rt m0 þ 1=T tT Vðt9 Þdt9 which could be interpreted as bedrock response. The forcing i(t) is chosen as summer insolation, either monthly mean or summer mean or annual maximum insolation at 65 N. That boreal summer insolation could the decisive forcing parameter of Pleistocene ice ages was suggested by Wladimir Ko¨ppen based on suggestions by Albrecht Penck and Eduard Bru¨ckner (see Ko¨ppen and Wegener, 1924; Milankovitch, 1941). The unknown parameters a, b, c, d, m0 and T are tuned to data. This model perfectly reproduces the *

(i.e. a qualitative change of the system’s state caused by a change in a control parameter, may it be a key, internally generated slow variable or an external forcing).

3.4 FORCED MODELS Saltzman (2002) provides a table in which he summarizes the various types of forced models. Here we would like to highlight only a few types. The simplest forced models are linear models of global ice mass in which ice-sheet physics and feedback processes with the atmosphere or surficial physics are neglected. For example, Imbrie (1980, see Paillard, 2001) explores the consequences of a simple low-pass filter that relates the dynamic change of nondimensional ice volume V to insolation i by dV ði VÞ ¼ dt where is a response time which differs for ablation (i.e. when dV=dt < 0) and

32

Martin Claussen, Andre´ Berger and Hermann Held

accumulation (when dV=dt 0). Imbrie’s model underestimates the 100-kyr cycle. Similar is valid for the simple response/ threshold model by Calder (1974, see Paillard, 2001) in which dV ¼ kði i0 Þ dt where i0 is a threshold of insolation and k is a parameter which differs for accumulation and ablation. Calder’s idea of a threshold-dominated system was more consequently formulated by Paillard (1998). Paillard assumes that the climate system exhibits multiple equilibria. If a certain threshold in the forcing is reached, then the system would jump from one equilibrium to another. In Paillard’s model, these equilibria are associated with an interglacial state, a glacial maximum and a weak glacial, such as the early Weichselian, for example. Again, the parameters (in this case the threshold values in insolation in Paillard’s model) are tuned to data. Paillard’s model provides a new view on the climate system as a threshold system. In particular, it solves the paradox of long interglacials occurring at seemingly low insolation: it suggests that not the insolation itself is the important driver, but the variation of insolation which would cause certain thresholds to be crossed more or less often. During long interglacials, like that at MIS (marine isotope stage) 11, insolation is much smaller than compared to the Eemian interglacial, but, because of low eccentricity, it varies only marginally for a longer period of time. Threshold models can become even more sophisticated, if in addition, stochasticity is present. Then the periodic forcing may be less pronounced and even too weak to cause transitions between equilibria in the noisefree case, and yet, with noise being added, it can induce synchronized jumps due to stochastic resonance (e.g. Benzi et al., 1982). Again, it is important for this theory to work that the system reveals multiple equilibria.

Hence, the external driver is not really a forcing, but rather a trigger. In general, stochastic climate models explore the consequences of random walk processes. This idea was originally proposed by Hasselmann (1976). Hasselmann assumed that the annual cycle of insolation generates variability in the fast climate components which could be randomly accumulated by the more sluggish climate components. Recently, Wunsch (2003) pursued this idea by demonstrating that most long-term climate archives reveal a rednoise spectrum, that is, a spectrum with amplitude of variance decaying with larger frequencies. Superimposed on the red spectrum are weak structures corresponding to the frequency bands of an orbital forcing. To explain the dominant 100-kyr climate variability, Wunsch suggests a stochastic forcing of a system with a collapse threshold. In such a system, a timescale can be generated without having to assume an external frequency component or an internal resonance, but the average time of a Brownian random walk to reach the threshold is sufficient to explain an average period of 100 kyr. Furthermore, Wunsch’s model yields a transition in the spectral domain from red to white, that is, a flat spectrum with amplitudes independent of frequency. This way, variability on the 20–100-kyr timescale, as well as on shorter timescales can be explained in a combined way. The categorization of models of Quaternary climate system dynamics as ‘forced’ and ‘free’ can be complemented by ordering them by means of complexity in terms of physical processes involved. According to Saltzman (2002, p. 276), climate system models aim at ‘ever more complete representation of the full slow-response climate system’, hence the centre manifold.* *

In the ‘vicinity’ of a bifurcation (see above), one can identify slow and fast processes. Then, the centre manifold characterizes the interplay of the few slowest variables with the collective effect of all the other variables, and therefore allows to study a complex system’s long-time behaviour by just a few degrees of freedom, hence by a conceptual model.

A Survey of Hypotheses for the 100-kyr Cycle

Saltzman (2002, p. 276) suggests that ‘ice sheets and their bedrock and basal properties, coupled with forced and free variations of carbon dioxide, operating on an Earth characterized by a high-inertia deep thermohaline ocean that can store carbon and heat . . .’ encompass the centre manifold. 3.5 WHAT TYPE OF MODEL DO WE NEED? So far, only conceptual, or inductive, models have been discussed. These models are designed to demonstrate the plausibility of processes; they are based on a gross understanding of feedbacks that are likely to be involved (Saltzman, 1985). Roe and Allen (1999) investigated the performance of six representatives of inductive models. They tuned each model to the time series (the last 900 kyr) of global ice volume and also the time rate of change for the ice volume while modelling the residuals as first- (and second) order autoregressive process. They found that within 95% error bars, one cannot favour any model over its competitors. Obviously, some assumptions on which conceptual models are based on are oversimplified and thus, could even be misleading. For example, the assumption that orbital forcing is identified with summer insolation at high northern latitudes overemphasizes the ice albedo feedback at high northern latitudes and neglects the fact that other components of the climate system can react to varying insolation in a different way than Northern Hemisphere ice sheets do. Therefore, it seems sensible to explore the role of geographically varying forcing and feedback processes in geographically explicit models. The degree of spatial and temporal resolution necessary for palaeoclimate simulations is disputed. Comprehensive, ‘state-of-the-art’ coupled models describing the general circulation of the atmosphere and the ocean, the dynamics of the terrestrial biosphere and the ice sheet as well as

33

biogeochemistry are supposed to be the most realistic laboratory of the natural Earth system. However, their applicability to long-term simulations is limited by high computational costs. Therefore it was proposed to use Earth system models of intermediate complexity (EMICs; Claussen et al., 2002) which operate at a higher level of spatial and temporal aggregation. Many processes resolved in AOGCMs have to be parameterized in EMICs. The advantage gained by this reduction is computational efficiency which makes them a useful tool for integrated palaeoclimate modelling. There are already numerous palaeoclimatic studies using EMICs and AOGCMs, and some examples will be given in Section 5 of this book. So far, however, even with EMICs, a realistic simulation of glacial – interglacial cycles by using orbital forcing only remains difficult to be done. 3.6 PERSPECTIVE Even approximately 160 years after first geological evidence, the ice-age riddle is not yet fully solved. However, we have some clues on what elements a theory of Quaternary Earth system dynamics should consist of – regarding concepts and model structure. Saltzman (2002) has proposed a unified theory of Quaternary Earth system dynamics. The term ‘unified’ is used, because it combines theories based on orbital forcing and on greenhouse gas forcing, respectively. Saltzman supposes that the slow part of the climate system involves the ice sheets and their bedrock and basal properties, coupled with tectonically forced and free variations of carbon dioxide, and a high-inertia deep thermohaline ocean. While it is very likely that slow variables exist, it is still not obvious that those can be identified with particular physical entities just mentioned. Quite the contrary, in a highly resolved spatiotemporal dynamics, such slow variables may emerge as complex patterns across physical entities,

34

Martin Claussen, Andre´ Berger and Hermann Held

which strongly support the use of spatially resolved climate models. Validation of inductive models appears to be an almost futile task: assessment of a number of inductive models cannot favour any model over its competitor on the grounds of tuning each model to the time series of global ice volume. In the range of more comprehensive, quasi-deductive models, only EMICs have been used for long-term studies. Some of these numerical experiments (e.g. Galle´e et al., 1992; Calov et al., 2005) support the idea of Hays et al. (1976) that orbital forcing may act as a pacemaker of glacial–interglacial cycles. However, the situation is rather complex. Obviously the response of the Earth system to a given forcing is a function of the actual state of the Earth system as well as meridional and seasonal changes of the forcing. Presumably, changes in insolation associated with changes in orbital parameters trigger fast internal feedbacks such as the water vapour – temperature feedback and the snow – albedo feedback which then are further amplified by slower feedbacks such as biogeochemical and biogeophysical feedback and the isostatic response of the lithosphere to ice-sheet loading. Some of these feedbacks even change sign during the course of a glacial–interglacial cycle. In Chapter 4, an example of exploring the feedbacks involved in the dynamics of glacial–interglacial climate change will be presented, while in chapters 5.1 to 5.8, the discussion is focused on the forcing and feedbacks at the end of an interglacial.

REFERENCES Arrhenius, S., 1896. On the influence of carbonic acid in the air upon the temperature of the ground. Philosophical Magazine and Journal of Science, 41, 237–276. Benzi, R.A., Parisi, G., Sutera, A., Vulpiani, A., 1982. Stochastic resonance in climatic change. Tellus, 34, 10–16. Berger, A., 1977. Support for the astronomical theory of climatic change. Nature, 268, 44–45.

Berger, A., 1978. Long-term variations of daily insolation and Quaternary climatic changes. Journal of Atmospheric and Oceanic Science, 35, 2362–2367. Calov, R., Ganopolski, A., Petoukhov, V., Claussen, M., Greve, R., 2005. Transient simulation of the last glacial inception. Part I: Glacial inception as a bifurcation in the climate system. Climate Dynamics, 25(6), 545–562. Claussen, M., Mysak, L.A., Weaver, A.J., Crucifix, M., Fichefet, T., Loutre, M.-F., Weber, S.L., Alcamo, J., Alexeev, V.A., Berger, A., Calov, R., Ganopolski, A., Goosse, H., Lohman, G., Lunkeit, F., Mokhov, I.I., Petoukhov, V., Stone, P., Wang, Zh., 2002. Earth system models of intermediate complexity: Closing the gap in the spectrum of climate system models. Climate Dynamics, 18, 579–586. Crowley, T.J., North, G.R., 1991. Paleoclimatology. Oxford University Press, 339 pp. EPICA Community Members, 2004. Eight glacial cycles from an Antarctic ice core. Nature, 429, 623–628. Galle´e, H., van Ypersele, J.-P., Fichefet, T., Marsiat, I., Tricot, C., Berger, A., 1992. Simulation of the last glacial cycle by a coupled, sectorially averaged climate-ice sheet model. Part II: Response to insolation and CO2 variation. Journal of Geophysical Research, 97, 15,713–15,740. Hasselmann, K., 1976. Stochastic models. I. Theory. Tellus, 28, 473–485. Hays, J.D., Imbrie, J., Shackleton, N.J., 1976. Variations in the earth’s orbit: pacemaker of the ice ages. Science, 194, 1121–1132. Ko¨ppen, W., Wegener, A., 1924. Die Klimate der geologischen Vorzeit. Borntraeger, Berlin, 255 pp. Milankovitch, M., 1941. Kanon der Erdbestrahlung und seine Anwendung auf das Eiszeitenproblem. Ko¨niglich Serbische Akademie, Belgrad, 633 pp. Paillard, D., 1998. The timing of Pleistocene glaciations from a simple multiple-state climate model. Nature, 391, 378–381. Paillard, D., 2001. Glacial cycles: towards a new paradigm. Review of Geophysics, 39, 325–346. Paul, A., Berger, W.H., 1997. Modellierung der Eiszeiten: Klimazyklen und Klimau¨berga¨nge. Geowissenschaften, 15, 20–27. Raymo, M.E., Oppo, D.W., Curry, W., 1997. The midPleistocene climate transition: a deep sea carbon isotopic perspective. Paleoceanography, 12, 546–559. Roe, G.H., Allen, M.R., 1999. A comparison of competing explanations for the 100,000-yr ice age cycle. Geophysical Research Letter, 26, 2259–2262. Ruddiman W.F., Raymo, M.E., McIntyre, A., 1986. Matuyama 41,000-year cycles: North Atlantic Ocean and Northern Hemisphere ice sheets. Earth and Planetary Science Letters, 80, 117–129.

A Survey of Hypotheses for the 100-kyr Cycle Saltzman, B., 1985. Paleoclimatic modeling. In: Hecht, A.D. (ed.), Paleoclimate analysis and modeling. Wiley, 341–396. Saltzman, B., 2002. Dynamical Paleclimatology: Generalized Theory of Global Climate Change. International Geophsics Series, Vol. 80, Academic Press, San Diego, 354 pp. Saltzman, B., Verbitzky, M.Y., 1993. Multiple instabilities and modes of glacial rhythmicity in the Plio-

35

Pleistocene: a general theory of late Cenozoic climatic change. Climate Dynamics, 9, 1–15. Shaffer, G., 1990. A non-linear climate oscillator controlled by biogeochemical cycling in the ocean: an alternative model of Quaternary ice age cycles. Climate Dynamics, 4, 127–143. Wunsch, C., 2003. The spectral description of climate change including the 100ky energy. Climate Dynamics, 20, 353–363.

This page intentionally left blank

4. Modelling the 100-kyr Cycle – An Example From LLN EMICs Andre´ Berger and Marie-France Loutre Universite´ catholique de Louvain, Institut d’Astronomie et de Ge´ophysique G. Lemaıˆtre, 1348 Louvain-la-Neuve, Belgium

ABSTRACT The glacial–interglacial cycles have been reproduced by the LLN-2D Northern Hemisphere model over the whole Quaternary. A short description of the model is presented before the main results are presented for the last cycle, the last 200 kyr, the last 800 kyr and the transition between the 41-kyr and the 100-kyr world. The quality and the deficiencies of these results are discussed in relationship with proxy geological records originating from some deep-sea and ice cores. 4.1 INTRODUCTION The 100-kyr cycles in eccentricity result from combination of the periods associated with the first terms in the expansion of the climatic precession parameter (e.g. Berger, 1994). The signal of these 100-kyr cycles is very weak in the insolation spectra (Berger et al., 1993c). This is why, in the previous chapter, it was mentioned that, although the eccentricity cycle seems in phase with the 100-kyr cycle in most climatic records, it needs a nonlinear amplification by mechanisms such as those related to the ice sheets, the ice albedoand the water vapour–temperature feedbacks, the carbon cycle, the isostatic rebound, the deep-ocean circulation and/ or the ocean–ice interactions. This is the reason for which a model of the fully coupled climate system needs to be constructed to allow at least some of the interactions between the atmosphere, the hydrosphere, the cryosphere, the

biosphere and the lithosphere to be taken into account. Such a climate model was built in Louvain-la-Neuve in the late 1980s. It links the Northern Hemisphere atmosphere, ocean mixed layer, sea ice, ice sheets and continents (Galle´e et al., 1991). It is a twodimension (latitude–altitude) sectorially averaged model. In each latitudinal belt, the surface is divided into at most seven oceanic or continental surface types, each of which interacts separately with the subsurface and the atmosphere. Special attention is paid to the albedo of snow, of vegetation in the northern high latitudes and of sea ice. The atmosphere–ocean model is asynchronously coupled to a model of the three main Northern Hemisphere ice sheets and their underlying bedrock. The coupled climate model is then forced by the astronomically derived insolation for each day and latitude and by the atmospheric CO2 concentration (this model did not contain an interactive carbon cycle). More details on the model are given in Galle´e et al. (1991) and also in Berger et al. (1990) for the ice sheet–lithosphere model, in Berger et al. (1989) for the upper ocean and in Berger et al. (1994) for the radiative convective scheme. The model is able to reproduce the main features of the present-day atmospheric general circulation and seasonal cycles of the oceanic mixed layer, of the sea ice and of the snow cover (Galle´e et al., 1991). This model has been further enlarged, leading to the MoBidiC model which has both the Northern and the Southern Hemispheres and a much more elaborated ocean (Crucifix et al., 2002).

38

Andre´ Berger and Marie-France Loutre

4.2 LAST GLACIAL–INTERGLACIAL CYCLE A first set of modelling experiments (Galle´e et al., 1991) showed that the variations in the Earth’s insolation alone induce feedbacks in the climate system which are sufficient to amplify the direct radiative impact and generate large climatic changes, provided CO2 is kept below 240 ppmv. This result confirms the (Hays et al., 1976) idea that the orbital forcing acts as a pacemaker of the ice ages, but initiation and termination of glacial cycles cannot be explained without invoking both the fast feedbacks associated with atmospheric processes (water vapour, cloud, snow and sea ice) and the slower feedbacks associated with other parts of the climate system, in particular build-up and disintegration of ice sheets. Moreover, taking into account the Vostok CO2 variations (Barnola et al., 1987) allows to better shape the last glacial–interglacial cycle and in particular the air temperature (Galle´e et al., 1992; Loutre et al., 1994). In a similar experiment with MoBidiC, there is a generally increasing trend in the continental ice volume from the last interglacial until the last glacial maximum. This maximum in the Northern Hemisphere is reached at 18 kyr BP with a volume of more than 40 106 km3 , a value probably slightly underestimated when compared to empirical reconstructions and other simulations (Hagdorn, in preparation). The simulated annually averaged surface temperature varies by up to 2.1 C in global average. These variations are greater for the Northern Hemisphere (2.8 C) than for the Southern Hemisphere (1.5 C), but larger variations can occur in higher latitudes. In the 50–60 N latitude band, the annual mean surface temperature of the Eurasian continent increases by 3.7 C between 25 kyr BP and the present, with an amplitude of the variations much larger in monthly than in annual means. The high northern latitudes of the Atlantic Ocean exhibit abrupt temperature changes of large amplitude at

around 100 and 10 kyr BP. Both events are characterized by a rapid cooling of 4 to 5 C followed by a period of slow warming, although temperatures remain low. Climate then warms rapidly by 6–7 C. In both cases, the export of NADW (North Atlantic Deep Water) is reduced by 5 Sv (1 Sv ¼ 106 m3 s1 ). These events are also characterized by a southern shift of the convection zone in the Atlantic and a cooling of the upper and intermediate waters of the Atlantic Ocean. The total area and volume of Arctic sea ice also increase during these events. 4.3 LAST 200 kyr Because of the sensitivity of the nonlinear climate model, it was crucial to see whether it can sustain more than one glacial– interglacial cycle. This was at the origin of a second simulation covering the last 200 kyr (Galle´e et al., 1993). Both the insolation forcing and the CO2 variations reconstructed from deep-sea (Shackleton et al., 1992) and ice (Jouzel et al., 1993) cores were used (Fig. 4.1 top). Broadly speaking, the response of the model to the Vostok CO2 and to the insolation (Berger, 1978) forcings reproduces quite well the low frequency part of the geological record over the last 200 kyr (Fig. 4.2 top) (Berger and Loutre, 1996; Berger et al., 1998). The timing of the stacked, smoothed oxygen isotope record of SPECMAP (Imbrie et al., 1984; Martinson et al., 1987) compares favourably with the simulation, although it might be argued that this result is biased by the astronomical tuning of the SPECMAP record. Moreover, there are discrepancies in the magnitude of the simulated ice volume. The largest one is probably the too large ice melting simulated by the model around 170 kyr BP and induced by the large values of insolation around 175 kyr BP, although the ice volume maximum at 182 kyr BP seems to be well captured by the model. Either it is a deficiency of the model or we have to look for a

Modelling the 100-kyr Cycle 0

–100

–200

–300

0

–100

–200

–300

39

–400

–500

–600

–700

–800

–400

–500

–600

–700

–800

300 275

Atmospheric CO2 concentration (ppmv)

250 225 200 300 275 250 225 200 300 275 250 225 200

Time (kyr BP)

Fig. 4.1 CO2 concentration from: (top) Vostok record (Jouzel et al., 1993), (middle) linear regressions between the marine isotopic and CO2 records used in Berger et al. (1996) (red curve), in Li et al. (1998) (green curve), in Loutre and Berger (2003) (dark blue curve) and Vostok–CO2 (Petit et al., 1999) (light blue curve), (bottom) Vostok–CO2 (Petit et al., 1999; Parrenin et al., 2004) (Barnola J. M., personal communication) (light blue curve) and linear regressions derived from the EPICA Dome C Deuterium record (EPICA, 2004). –100

–200

–300

–400

–500

–600

–700

–800 –2

0

–1

10 20

0

30 1

40

2

50 0 10 20 30 40 50 0

δ18O (standard deviation units)

Northern Hemisphere continental ice volume (106 km3)

0

10 20 30 40 50

0

–100

–200

–300

–400

–500

–600

–700

–800

Time (kyr BP)

Fig. 4.2 Simulated ice volume of the Northern Hemisphere using the LLN 2-D NH climate model forced by insolation (Berger, 1978) and the different scenarios of atmospheric CO2 concentration of Fig. 4.1. Black curve is the ice volume of SPECMAP (Imbrie et al., 1984).

40

Andre´ Berger and Marie-France Loutre

significant change in the Southern Hemisphere continental ice at that time. At the end of stage 6, the simulated glacial maximum occurs at 135 kyr BP, while the 18 O ice volume maximum occurs at 151 kyr, but remains large from 156 to 133 kyr BP. The model simulates very well the transitions between isotopic stages 6 and 5 and between the isotopic substages 5e (Eemian interglacial) and 5d. This is due mainly to the insolation changes, but it is reinforced by important changes in the CO2 concentration and all the feedbacks in the model including those related to land surface cover (Berger, 2001). Although the timing of what may correspond to isotopic substages 5c, 5b and 5a is well reproduced, the model simulates a total melting of the Northern Hemisphere ice sheets between 126 and 117 kyr BP, 100 and 97 kyr BP and 83 and 74 kyr BP. Although this is not realistic (the Greenland ice sheet having survived at least over the last two to three glacial–interglacial cycles Dansgaard et al. (1993)), it does not prevent the ice sheets to grow again, leading to a 100-kyr quasi-cyclicity similar to the one seen in the geological data. The entrance into stage 4, starting at 70 kyr BP, is not rapid enough when compared to geological record and especially to the SPECMAP reconstruction, a failure that can be solved by taking into account both the volumes of snow and ice. Then there is a brief reversal between 60 and 50 kyr BP, followed by a re-growth of the ice sheets and leading to the last glacial maximum. The maximum amount of ice reaches 47 106 km3 at 15 kyr BP. Finally, the model reproduces the deglaciation from 15 to 3 kyr BP. It simulates correctly the total disappearance of the Eurasian ice sheet around 7 kyr BP, followed by the melting of the North American one, only the Greenland ice sheet being left in the Northern Hemisphere with roughly 2:96 106 km3 of ice. Since 3 kyr BP, the simulated Greenland ice volume in the absence of human perturbations is increasing slightly, reaching today 3:07 106 km3 , which represents

about the actual present-day value. In parallel, our simulated natural climate is slightly cooling since the peak of the Holocene. These experiments confirm that variations in the Earth’s orbit and related insolation act as a pacemaker of the ice ages (Hays et al., 1976). CO2 variations help to shape the 100-kyr cycle and mainly improve the simulated surface air temperature (Galle´e et al., 1992; Loutre et al., 1994). A deeper analysis of these experiments also confirms the importance of the processes governing the response of the modelled climate system to insolation and/or CO2 changes. These are fundamentally related to the albedo- and water vapour feedbacks (Berger et al., 1993b), to the taigatundra direct and indirect impacts on high latitudes surface albedo (Kubatzki and Claussen, 1998; Berger, 2001), to the altitude and continental effects on the precipitations over the ice sheets, to the lagging lithospheric response to the ice-sheet loading (Crucifix et al., 2001) and to the mechanical destabilization of the ice sheets through the rapid melting of their southern front as compared with the northern one (Berger et al., 1992; Berger et al., 1993a). 4.4 LAST 800 kyr To further test the capacity of the model to sustain the 100-kyr cycle over a long period of time, four CO2 scenarios were actually used for simulating the past 575-kyr (Fig. 4.1 middle for 440 kyr) (Loutre and Berger, 2003). The first one is based on a multiple regression between the deep-sea record from Ontong Java Plateau in the western equatorial Pacific and the ice core CO2 from Antarctica (Berger W. et al., 1996). The second one is generated from a regression between the Vostok CO2 concentration and the SPECMAP oxygen isotope values calculated over the last 218 kyr and extended over the past 575-kyr (Li et al., 1998). The same procedure was applied for the third one, using the low latitude stacked 18 O record from marine core MD 900963 of site 677 (Bassinot et al.,

Modelling the 100-kyr Cycle

1994) instead of the SPECMAP record. At last, the CO2 concentration reconstructed from Vostok over the last four glacial–interglacial cycles was used (Petit et al., 1999). As this Vostok record extends only to 414 kyr BP, the CO2 from Li et al. (1998) was used from 575 to 440 kyr BP, and a linear interpolation between 440 and 414 kyr BP ensured the transition towards the Vostok record. The broad features of the results obtained from these different CO2 scenarios are pretty well similar (Fig. 4.2 middle). The spectra of the simulated ice-sheet volume and of the oxygen isotope proxy record are highly coherent in the frequency bands associated with the periods of 100, 41, 23 and 19 kyr. In these experiments, as in previous ones, the albedo- and water vapour– temperature feedbacks play a significant role in amplifying the forcings. According to a simulation of the last glacial maximum in response to insolation and CO2 , the water vapour feedback would explain 40% of the cooling (Berger et al., 1993b). Such a change in the water vapour content of the atmosphere has also been simulated by Li et al. (1998) over the last 500 kyr, with a vertically integrated value over the whole Northern Hemisphere in July varying from a little bit more than 30 kg/m2 during glacial times to about 55 during the interglacials. For the land surface–climate interaction, when insolation and CO2 , decrease (insolation leading CO2), both the snow fields and tundra lead to an increase of the surface albedo creating a positive feedback, which is reinforced by the subsequent decrease in the water vapour content of the atmosphere (see also Loutre et al., this volume). These feedbacks lead finally to the building up of the continental ice sheets which in turn enter the feedback loops. The major difference between these experiments covering the last four glacial– interglacial cycles arises during isotopic stages 11 and 10. From 400 to 350 kyr BP, the ice volume simulated with Vostok CO2 concentrations remains lower than 5 106 km3 over the whole interval

41

(Fig. 4.2), while in the other experiments, it remains at its minimum for only 10 kyr. This feature of the experiment using Vostok CO2 concentrations, if confirmed, is similar to what might happen to our stage 1, even more if the future CO2 level is kept high over a sufficiently long period of time. From 330 kyr BP (peak of MIS 9 interglacial) to 250 kyr BP (glacial maximum in MIS 8), SPECMAP displays a regular trend towards glaciation, while the simulation displays stadials–interstadials with full interglacial conditions. From stage 5e onwards, there are no real significant differences between these experiments and the experiment discussed in the previous section. This can be expected from the rather good agreement between the different CO2 reconstructions, except during the early part of stage 5e. However, this difference in the CO2 series does not lead to any significant difference between the simulated ice volumes, contrary to what happened at stage 11 because the amplitude in the insolation variation is much larger during MIS 5e than during MIS 11. The similarity in the CO2 concentration of stage 5e and stage 11 (long lasting maximum) and the difference in their insolation stress gain the importance of better understanding the relatively long interglacials (Loutre et al., this volume) during which high values of CO2 are sustained, whether the ice volume remains rather low (stage 11) or not (stage 5e) during the whole stage. Several other attempts were made to provide CO2 values prior to 400 kyr BP, it means prior to the end of the Vostok ice core. The recent EPICA core is expected to solve this problem. Before its CO2 became available, a linear regression was built between Vostok CO2 (Petit et al., 1999) and EPICA deuterium (EPICA, 2004) over the last 414 kyr and used in order to extend the CO2 record to the lower end of the EPICA core (710 kyr BP, Fig. 4.1 bottom). The chronology of Vostok–CO2 includes corrections made by Barnola (personal communication), (Parrenin et al., 2004; Raynaud et al., 2005).

42

Andre´ Berger and Marie-France Loutre

In this regression, from 710 to 430 kyr BP, CO2 values vary between 200 and 250 ppmv, a range that is much less than in the Vostok record. For this reason, another scenario was created with essentially a decrease of the CO2 minima to 180 ppmv. In both scenarios, the model continues to simulate the 100-kyr cycle before 400 kyr BP. However, over these earlier times, the Northern Hemisphere ice sheets are still disappearing during interglacials, and their size during glacials is similar to what is simulated for MIS 8 and MIS 10 using the Vostok–CO2 record. Although for the ice volume, both the phase and amplitude compare pretty well with the reconstructions over the last 400 kyr, none of the simulations reproduces a reduction in the amplitude of the glacial–interglacial cycles before MIS 11 (Fig. 4.2 bottom) as it is seen in both deep-sea and EPICA records (EPICA, 2004), that is, cool interglacial–cold glacial. That might be related to processes lacking in our model and which play an important role before MIS 11 or to a Southern Hemisphere much cooler than the Northern one (we must recall that our model is only simulating NH), although evidence for this must still be found. 4.5 FROM 41- TO 100-kyr CYCLE A final test which needed to be performed is related to the transition between the 41-kyr and the 100-kyr world. Can this be simulated by the same model without any re-tuning and what might be a possible cause for it? To analyse this, a linearly concentration from decreasing CO2 320 ppmv at 3 Myr BP to 200 ppmv at the last glacial maximum was used as a scenario (Berger et al., 1999). Actually, during the late Pliocene, the simulated ice volumes are small and interglacials are long, while long glacials with a large amount of ice prevail during the late Pleistocene. The model is thus able to reverse from a late Pliocene/

early Pleistocene climate dominated by warm interglacials to a late Pleistocene cold climate where glacials prevail. At the transition, CO2 is sufficiently low (240 ppmv in our model) to allow the ice sheets to start developing fully. A few sensitivity tests confirmed the need of crossing such a CO2 threshold to generate the 100-kyr cycles. If the CO2 concentration is above the threshold, no glacial–interglacial cycle can be generated because no ice sheet can develop. The spectral analysis performed on the simulated ice volume between 2 and 1 Myr BP shows that the most important periodicity is related to obliquity. For the last 1 Myr, the global spectrum is characterized by the obliquity and precession frequencies, but it is the variance components near 100 kyr which dominate. This is confirmed in an evolutive spectrum from 1.5 Myr BP to present, where in addition one can see that the obliquity signal remains more or less constant and the periods in the precessional band start to strengthen around 1.3 Myr BP, especially the 23-kyr cycle which splits into the 23.7- and 22.4-kyr periods as it is the case in the precession expansion (Berger, 1977). As a conclusion, these model results stress the role of the CO2 concentration crossing some threshold value to allow the climate system and its feedback mechanisms to respond nonlinearly to the astronomical forcing and hence to create the 100-kyr cycle as a result of a beat between the main precessional frequencies. 4.6 CONCLUSIONS The LLN-2D model succeeds to reproduce the 41-kyr cycle of the early and middle Pleistocene, up to about 1 Myr BP and the progressive transition towards the 100-kyr cycle which dominates the spectrum of the last 400 kyr. The transition coincides with the crossing of a CO2 threshold of 240 ppmv. This value is indeed sufficiently low to allow ice sheets to start building up and to be further sustained. The same simulation

Modelling the 100-kyr Cycle

shows also that stages 11 and 1 cannot be interglacials if CO2 is low. After MIS 11, the model simulates pretty well the last 400 kyr when forced with the long-term variations of insolation and CO2. Sensitivity analyses stress the role of CO2 during times of very low eccentricity. As a consequence, an entrance into glaciation now can be simulated only if CO2 remains below 240 ppmv. Between 400 and 900 kyr BP, the model simulates the 100-kyr cycles with reduced amplitude, but is globally too warm, not allowing cool interglacials and cold glacials.

REFERENCES Barnola, J.M., Raynaud, D., Korotkevitch, V.S., Lorius, C., 1987. Vostok ice core: a 160,000 year record of atmospheric CO2. Nature 329(6138), 408–414. Bassinot, F.C., Labeyrie, L.D., Vincent, E., Quidelleur, S., Shackleton, N.J., Lancelot, L.Y., 1994. The astronomical theory of climate and the age of the Brunhes–Matuyama magnetic reversal, Earth and Planetary Science Letters 126(1–3), 91–108. Berger, A., 1977. Support for the astronomical theory of climatic change. Nature 268, 44–45. Berger, A., 1978. Long-term variations of daily insolation and Quaternary climatic changes, Journal Atmospheric Science 35(12), 2362–2367. Berger, A., 1994. Astronomical theory of palaeoclimates, in: Cl. Boutron (Ed.), Topics in Atmospheric and Interstellar Physics and Chemistry, Les Ulis, France, 411–452. Berger, A., 2001. The role of CO2 , sealevel and vegetation during the Milankovitch forced glacialinterglacial cycles, in: C.U.H.L. Bengtsson (Ed.), Geosphere–Biosphere Interactions and Climate, Cambridge University Press, New York, 119–146. Berger, A., Loutre, M.F., 1996. Modelling the climate response to astronomical and CO2 forcings, C.R. Acad. Sci. Paris, t. 323(se´rie IIa), 1–16. Berger, A., Fichefet, T., Galle´e, H., Tricot, C., Marsiat, L., van Ypersele, J.-P., 1989. Astronomical forcing of the last glacial–interglacial cycle, in: P. Crutzen, J.C. Ge´rard and R. Zander (Eds.), Our Changing Atmosphere, Universite´ de Lie`ge, Institut d’Astrophysique, Cointe-Ougre´e, 353–382. Berger, A., Fichefet, T., Galle´e, H., Tricot, C., Marsiat, L., van Ypersele, J.-P., 1990. Physical interactions within a coupled climate model over the last glacial-interglacial cycle, Philosophical Transactions of the Royal Society of Edinburgh: Earth Sciences 81(4), 357–369.

43

Berger, A., Fichefet, T., Galle´e, H., Tricot, C., van Ypersele, J.-P., 1992. Entering the glaciation with a 2-D coupled climate model, Quaternary Science Reviews 11(4), 481–493. Berger, A., Galle´e, H., Tricot, C., 1993a. Glaciation and deglaciation mechanisms in a coupled 2-D climate-ice sheet model, Journal of Glaciology 39(131), 45–49. Berger, A., Tricot, C., Galle´e, H., Loutre, M.F., 1993b. Water-vapor, CO2 and insolation over the last glacial-interglacial cycles, Philosophical Transactions of the Royal Society of London Series BBiological Sciences 341(1297), 253–261. Berger, A.M., Loutre, M.F., Tricot, Ch., 1993c. Insolation and the Earths orbital periods, Journal of Geophysical Research-Atmosphere 98(D6), 10 341–10 362. Berger, A., Tricot, C., Galle´e, H., Fichefet, T., Loutre, M.F., 1994. The last two glacial–interglacial cycles simulated by the LLN model, in: J.-C. Duplessy, M.-T. Spyridakis (Eds.), Long Term Climatic Variations, Data and Modelling 22, Springer, Berlin, 411–452. Berger, A., Loutre, M.F., Galle´e, H., 1998. Sensitivity of the LLN climate model to the astronomical and CO2 forcings over the last 200 ky, Climate Dynamics 14(9), 615–629. Berger, A., Li, X.S., Loutre, M.F., 1999. Modelling northern hemisphere ice volume over the last 3 Ma, Quaternary Science Reviews 18, 1–11. Berger, W.H., Bickert, T., Yasuda, M.K., Wefer, G., 1996. Reconstruction of atmospheric CO2 from ice-core data and the deep-sea record of Ontong Java plateau: the Milankovitch chron., Geol. Rundsch. 85, 466–495. Crucifix, M., Loutre, M.F., Lambeck, K., Berger, A., 2001. Effect of isostatic rebound on modelled ice volume variations during the last 200 kyr, Earth and Planetary Science Letters 184(3–4), 623–633. Crucifix M., Loutre M.F., Tulkens Ph., Fichefet T., Berger A., 2002. Climate evolution during the Holocene: A study with an Earth system model of intermediate complexity, Climate Dynamics 19, 43–60. Dansgaard, W., Johnsen, S.J., Clausen, H.B., DahlJensen, D., Gundestrup, N.S., Hammer, C.U., Hvldborg, C.S., Steffensen, J.P., Sveinbjo¨rnsdottir, A.E., Jouzel, J., Bend, G., 1993. Evidence for general instability of past climate from a 250kyr ice-core record, Nature 364, 218–220. EPICA Community Members, 2004. Eight glacial cycles from an Antarctic ice core. Nature 429 (6992), 623–628. Galle´e, H., van Ypersele, J.-P., Fichefet, T., Tricot, C., Berger, A., 1991. Simulation of the last glacial cycle by a coupled, sectorially averaged climate-ice sheet model. Part I: The climate model, Journal

44

Andre´ Berger and Marie-France Loutre

of Geophysical Research-Atmospheres 96, 13 139–13 161. Galle´e, H., van Ypersele, J.-P., Fichefet, T., Marsiat, I., Tricot, C., Berger, A., 1992. Simulation of the last glacial cycle by a coupled, sectorially averaged climate ice sheet model. Part II: Response to insolation and CO2 variation, Journal of Geophysical Research-Atmospheres 97, 15 713–15 740. Galle´e, H., Berger, A., Shackleton, N.J., 1993. Simulation of the climate of the last 200 kyr with the LLN 2-D model, in: W.R. Peltier (Ed.), Ice in the Climate System, NATO ASI Series I, Global Environmental Change 12, 321–341, Springer, Berlin. Hays, J.D., Imbrie, J., Shackleton, N.J., 1976. Variations in the earth’s orbit: pacemaker of the ice ages, Science 194, 1121–1132. Imbrie, J., Hays, J., Martinson, D.G., McIntyre, A., Mix, A.C., Morley, J.J., Pisias, N.G., Prell, W.L., Shackleton, N.J., 1984. The orbital theory of Pleistocene climate: support for revised chronology of the Marine 18O record, in: A. Berger, J. Imbrie, J.D. Hays, G. Kukla, B. Saltzman (Eds.), Milankovitch and Climate, Reidel, Dordrecht, 269–305. Jouzel, J., Barkov, N.I., Barnola, J.M., Bender, M., Chappellaz, J., Genthon, C., Kotlyakov, V.M., Lorius, C., Petit, J.R., Raynaud, D., Raisbeck, G., Ritz, C., Sowers, T., Stievenard, M., Yiou, F., Yiou, P., 1993. Extending the Vostok ice-core record of paleoclimatic to the penultimate glacial period, Nature 364(6436), 407–412. Kubatzki, C., Claussen, M., 1998. Simulation of the global bio-geophysical interactions during the Last Glacial maximum, Climate Dynamics 14, 461–471. Li, X.S., Berger, A., Loutre, M.F., 1998. CO2 and northern hemisphere ice volume variations over the

middle and late quaternary, Climate Dynamics 14(7–8), 537–544. Loutre, M.F., Berger, A., 2003. Stage 11 as an analogue for the present interglacial, Global and Planetary Change 36, 209–217. Loutre, M.F., Berger, A., Dutrieux, A., Galle´e, H., 1994. The response of the LLN climate model to the astronomical forcing over the last glacial– interglacial cycle, Terra Nostra, Schriften der Alfred-Wegener-Stiftung 1/94, 11–15. Martinson, D.G., Pisias, N.G., Hays, J.D., Imbrie, J., Moore, J.T.C., Shackleton, N.J., 1987. Age dating and the orbital theory of the ice ages: development of a high-resolution 0 to 300 000-year chronostratigraphy, Quaternary Research 27, 1–29. Parrenin, F., Re´my, F., Ritz, C., Siegert, M.J., Jouzel, J., 2004. New modeling of the Vostok ice flow line and implication for the glaciological chronology of the Vostok ice core. Journal of Geophysical Research 109, D20102, DOI: 10.1029/ 2004JD004561. Petit, J.R., Jouzel, J., Raynaud, D., Barkov, N.I., Barnola, J.M., Basile, I., Bender, M., Chappellaz, J., Davis, M., Delaygue, G., Delmotte, M., Kotlyakov, V.M., Legrand, M., Lipenkov, V.Y., Lorius, C., Pe´pin, L., Ritz, C., Saltzman, E., Stievenard, M., 1999. Climate and atmospheric history of the past 420 000 years from the Vostok ice core, Antarctica, Nature 399(6735), 429–436. Raynaud, D., Barnola, J.M., Souchez, R., Lorrain, R., Petit, J.R., Duval, P., Lipenkov, V.Y., 2005. Revisiting the Vostok record: the CO2 paradox of marine isotope stage 11. Shackleton, N.J., Le, J., Mix, A., Hall, M.A., 1992. Carbon isotope records from Pacific surface waters and atmospheric carbon dioxide. Quaternary Science Reviews 11(4), 387–400.

Section 2 Methods of Palaeoclimate Reconstruction and Dating (ed. Frank Sirocko)

This page intentionally left blank

5. Introduction – Palaeoclimate Reconstructions and Dating Frank Sirocko Institute for Geosciences, Johannes Gutenberg-University, Becherweg 21, 55099 Mainz, Germany

This book on the ‘Climate of Past Interglacials’ concentrates on the last four interglacials preceding the Holocene, that is, during the time from about 100 to 450 kyr. The first knowledge on this time has been extracted almost 40 years ago from marine deep sea cores with a sampling resolution of several thousand of years. Deepocean sediments can be rather easily dated, because 18 O stratigraphy (Shackleton and Opdyke, 1973) can be applied all over the ocean basins and tuned directly to the ice volume/sea-level master curve of SPECMAP (Martinson et al., 1987), which is based on the beat of the orbital insolation cyclicities (Berger, 1978; Berger et al., this volume). Sediments of the ocean are today investigated with a much higher time resolution (see for example the papers by Vautravers et al., Roucoux et al. and Rein et al. in this volume) and have revealed century-scale abrupt events of changing wind directions, sea-surface temperatures, strength of ocean currents, ice-sheet stability and deep-water formation, most of which are processes that can directly affect climate in the regions under their direct influence or even worldwide (e.g. El Nin˜o or the thermohaline circulation). The stratigraphy of such rapid events is, however, derived from a correlation with ice core chronologies. In particular, the GRIP, GISP2 and NorthGRIP drillings in Greenland (Johnsen et al., 1992; Grootes et al., 1993; NorthGRIP, 2004) have shaped our knowledge about the very high speed of climate change on the northern hemisphere and serve now as master chronology for MIS 1–5. The long records of Vostok (Petit et al., 1999) and EPICA (EPICA Community Members, 2004) from Antarctica

are the most highest resolution source of information for the southern hemisphere during the last 800 000 years. The stratigraphy of these long cores is, however, at some tie points tuned to SPECMAP. Thus, the stratigraphy of ice cores and marine sediments is based on the robustness of the oxygen isotope stratigraphy which was first developed by Emiliani (1955) and established by Nicholas Shackleton (1937–2006), to whom this volume on the ‘Climate of Past Interglacials’ is dedicated. Latest research of this topic goes even further by modelling oxygen isotope variations. See also the compilation of several different approaches for sea-level reconstructions by Siddall (this volume). The reliability of the oxygen isotope stratigraphy was established by numerous radiometric dates, mainly 14C for the last 55 000 years, U/Th for the time back to 350 000 years and K/Ar or Ar/Ar datings well back into the early Pleistocene. All of these methods have specific demands on the amount of datable material and are prone to analytical error. The time of interest for this book is from 100 to 450 kyr, thus mainly in the range of U/Th dating, and we have thus two papers in this chapter that deal with this isotope system. Scholz et al. (this volume) developed a refinement of coral ages for a sea-level highstand during isotopic stage 7a, and Frechen (this volume) applied the U/Th technique to terrestrial peats in North Germany, where the age of the Holsteinian (MIS 9 or 11) has come under strong debate (see also the paper by Geyh and Mu¨ller, this volume). Both of these applications depend on the immobility of U and Th in peat or coral aragonite.

48

Frank Sirocko

Whether these prerequisites for reliable dating are indeed given is a serious matter of concern; see for example the U/Th values for the classical Eemian profile of Gro¨bern. The U/Th values presented by Frechen et al. (this volume) would date the Gro¨bern site into MIS 7e. This is highly unlikely, and Frechen et al. themselves claim that Gro¨bern should be MIS 5e. Accordingly, U/Th dates on peat have to be interpreted with some suspicion, which has to be kept in mind before the U/Th dating of the Holsteinian into MIS 9 can be indeed taken for granted (Geyh and Mu¨ller, this volume). The high demand for dating techniques for terrestrial sediments of the last 400 000 years led Krbetschek and Degering (this volume) to develop the new luminescence dating technique ‘radioluminescene’, which can be applied to clastic feldspar minerals from well-bleached sedimentary environments. They apply this technique to the sediments of Munster (North Germany) and arrive at an age of 334 kyr for the top of the Kieselgur, again indicating an age of MIS 9 for the Holsteinian. The dating of the diatomite at Munster is particularly important, because this is the only clearly Holsteinian record which is varved and has been used to determine the length of this interglacial to about 15 000 years (Mu¨ller, 1974). It is not only of academic interest whether the Holsteinian belongs to MIS 9 or MIS 11, because MIS 11 is the only interval during the late Pleistocene when the insolation forcing was similar to that during the Holocene. MIS 11 was also the longest of the last five interglacials (Fig. 5.1, Petit et al., 1999, see also McManus et al., 2002). The Munster record shows two spikes of Betula and Pinus dominance during the interglacial, which was interpreted as evidence for the occurrence of extreme cold events during an interglacial (Mu¨ller, 1974). Thus, it is of specific importance whether Munster indeed represents MIS 11, because it would be the only annually varved record for an older interglacial with climate forcing comparable to the Holocene and thus could inform us

about the processes and rates of climate change, which may lie ahead of us at the end of the Holocene. The most established continental record of the late Pleistocene interglacials is the VOSTOK ice core (Petit et al., 1999). We use these data for Fig. 5.1, even if the absolute dates of the interglacials have slightly changed in the new EPICA chronology. Based on the Vostok chronology, all the five interglacials have apparent differences in length, with MIS 11 of more than 20 000 years being the longest, and MIS 7e with less than 5000 years being the shortest. MIS 1, 5e and 9e are all about 12 000 years long. MIS 7e is spike shaped, MIS 5e and 9e are sawtooth shaped, quite in contrast to MIS 1 and MIS 11 which show no asymmetric shape. Another feature of importance for the forcing of past interglacials is documented in the trace gas content of the Vostok ice core (Fig. 5.1). Methane parallels the D (temperature) trend almost perfect. CO2 in contrast lags the temperature record by 4000 years during the last glacial inception (MIS 5e–5d transition), and a lag is also clearly visible for the glacial inception at the end of MIS 11 and MIS 9e. Accordingly, some parts of the global interglacial climate systems remained on an interglacial level much longer than the temperature over Antarctica. This pattern was the same during all past interglacials except MIS 7e. We do not want to speculate on possible reasons for this observation, but it matches the observation of several authors in this book, that the floral changes on the northern hemisphere developed time transgressive during the last interglacial. The CO2 lag as well as its relation to CH4 changes is a particularly important point, because it hints at interaction processes between various climate system components (see Claussen, this volume), in particular between the terrestrial biosphere, wetlands, marine biosphere and oceanic chemistry, which are yet not understood. Presumably, CO2 changes amplify climate changes – a process which becomes interesting in the light of the present-day climate

Palaeoclimate Reconstructions and Dating

49 –410

775

545

–450

430

–470

–490 >24.2 kyr

315

>12.2 kyr

14.8 kyr

4.2 kyr

14.7 kyr

>11.1 kyr

18.3 kyr

2.7 kyr

16 kyr

>20.3 kyr –410

300

CO2 (ppmv)

270

–430

260 ppmv

240

–450

210

–470 termination onsets

180

0

50

δD [‰SMOW]

–430

118 kyr

100

150

200

250

300

350

δD [‰SMOW]

CH4 (ppbv)

660

–490

400

Ice/gas age (kyr), GT4 timescale after Petit et al. (1999)

Fig. 5.1 D (temperature) and the greenhouse gases CH4 and CO2 for the Vostok ice core, data from Petit et al. (1999). The Vostok chronology of past interglacials differs substantially from the chronology of the new EPICA core. The Vostok core, however, faithfully documents the lead/lag relation between local temperature over Antarctica (D) and the global trace gas content of CH4 and CO2.

changes in which anthropogenic CO2 emissions have to be regarded as forcing of the climate system. The CO2 lag has, however, also implications for the dating and duration of past interglacials. The duration of the interglacial CO2 maximum above the value of 118 kyr, which marks the inception of the last glacial (see Seelos and Sirocko, this volume; Kubatzki et al., this volume), is 18 300 years, the length of the interglacial temperature maximum only 14 800 years (Fig. 5.1). Apparently, the duration of an interglacial depends on a definition, which is classically based on the oxygen isotope stratigraphy and maximum sea-level highstands (see above). Martinson et al. (1987) used the maximum gradients of sea-level increase or decrease to define the boundaries of the SPECMAP stages. This approach is still used for all marine sediment cores. Terrestrial palynologists, however,

prefer the expression of a ‘thermomere’ to address a phase with abundant thermophilous pollen, and this can be an interglacial or an interstadial. The ice core community, in contrast, has recently presented a new approach to define an interglacial, i.e. they used the lowest Holocene D values in the EPICA core (EPICA Community Members, 2004) and address all past intervals with D values above this Holocene minimum value as ‘interglacial’. For Fig. 5.1, we used the midHolocence CO2 concentration of 260 ppm to mark a typical interglacial value, although peak values of 280 ppmv or 300 ppmv might be the maximum natural values during the Holocene. The above examples show that the name ‘interglacial’ is indeed used in very different ways. The CO2 lag (see above) indicates clearly that different parts of the climate system stay on interglacial values for

50

Frank Sirocko

different amounts of time, and these differences can last up to several thousand years (Fig. 5.1). The controversial discussion about the length of past interglacials (Kukla et al., 1997) versus (Turner, 2002) may thus indeed be obsolete, because an interglacial has apparently different durations at different locations even on the same hemisphere. The principal asynchroneity (see-saw) of the MIS 3 interstadials between the northern and southern hemisphere was demonstrated very convincingly by Blunier et al. (1998). Asynchroneities between the low latitudes and the northern hemisphere have been reported from an early sea-level rise (Henderson and Slowey, 2000) and early highstands of African lakes (Trauth et al., 2003) at the beginning of Termination II. Steep gradients in the vegetation belts across Europe during the last glacial inception are just about to be detected. Sa´nchez Gon˜i et al. (2005) and Sirocko et al. (2005) demonstrated on the basis of highresolution records from the Portugal margin and Eifel maar lakes in soutwest Germany respectively that the development of vegetation at the end of the Eemian in Portugal/ France/southwest Germany is out of phase with the floral evolution in north/east Germany and Scandinavia. Both authors claim for steep gradients in the SST fields of the Atlantic Ocean, which could decouple the climate of northern Europe from the climate in southern Europe. Mu¨ller and Sa´nchez Gon˜i (this volume) discuss the vegetation evolution of southern Germany in the same context. Seelos and Sirocko (this volume) develop a correlation between cold events in the North Atlantic; a north German dust record, and dust records from Eifel maar sediments. Again, the floral evolution between north Germany and the Eifel is offset by several thousand years when records are correlated on the basis of dust events. The C-events are probably the best indicator for the existence of the North American ice sheet during the early Weichselian glaciation, but they can only occur after the ice has been already build up; thus they

cannot be used to date the last glacial inception at the end of the Eemian. The most direct indicator for the beginning of the last glaciation at 118 kyr is probably the beginning of sea-level regression as documented by U/Th dating of reef terraces (Lambeck and Chappell, 2001). The same age of 118 kyr is given by the end of the late Eemian speleothem growth in the Spannagel cave of the Alps (Holzka¨mper et al., 2004; Spo¨tl et al., this volume). This cave presents probably the best record for absolute dating of past interglacials in general, because it lies at the altitude of the modern snow line, which implies that the dripping water necessary for the formation of the speleothems freezes as soon as the snow line (representing an annual average temperature of 0 C) drops below the modern value. Time transgressive climate evolution is also documented by Rein et al. (this volume) for the ENSO region off Peru. Low-latitude climate shifts at the beginning and end of the last interglacial most probably lead the climate of the high northern latitudes, because the ENSO system is more directly forced by the seasonal gradient between spring and autumn climate (Clement et al., 1999) and not directly linked to the summer insolation forcing of the northern hemisphere. Such lead/lag relations are often used to depict causal mechanisms in the climate system, but it must not necessarily imply a forcing of high latitudes by lowlatitude processes (Claussen et al., 2003). The above evaluation follows the records presented in this book and is not a summary of the global available information on timing and forcing of past interglacials. There are many other important and excellent records worldwide, and a full review of the state of art on past interglacial research is beyond the scope of this book. The records presented are, however, completely sufficient to draw a few general inferences: 1. The beginning, end and duration of the past interglacials were not synchronous all over the world, i.e. parts of the climate

Palaeoclimate Reconstructions and Dating

system have been longer in an interglacial state than others. 2. Beginning and end of interglacials in the low latitudes and in the Antarctic lead respective changes on the northern hemisphere. 3. Time transgressive climate shifts are also strong over Europe, where the SST patterns of the North Atlantic drift presumably cause a stepwise shift of the vegetation zones at least during the end of the past interglacial, with a longer interglacial in southern Europe and shorter interglacial in the north. This dynamic climate evolution of the past interglacial must have been of high importance also for the evolution of mankind. Neanderthal hominids lived and hunted in Europe during the Eemian (Wenzel, this volume), but were replaced during MIS 3 by the modern humans. Genetic data indicate that these early humans migrated out of Africa during MIS 5 and lived in the Mediterranean during MIS 4 and the early MIS 3. Why and when they exactly moved into Northern Europe and Asia has certainly no relation to the past interglacial climate. The reason why they left Africa during the early MIS 5, however, might well have to been seen in this context. REFERENCES Berger, A. (1978). Long-term variations of daily insolation and Quaternary climatic changes. Journal of Atmospheric Sciences 35(12), 2362–2367. Blunier, T., Chappellaz, J., Schwander, J., Da¨llenbach, A., Stauffer, B., Stocker, T.F., Raynaud, D., Jouzel, J., Clausen, H.B., Hammer, C.U., and Johnsen, S.J. (1998). Asynchrony of Antarctic and Greenland climate change during the last glacial period. Nature 394, 739–743. Claussen, M., Ganopolski, A., Brovkin, V., Gerstengarbe, F.-W., and Werner, P. (2003). Simulated global-scale response of the climate system to Dansgaard-Oeschger and Heinrich events. Climate Dynamics 21, 361–370. Clement, A.C., Seager, R., and Cane, M.A. (1999). Orbital controls on the El Nino/southern

51

oscillation and tropical Pacific climate during the last millennium. Nature 424, 271–276. Emiliani, C. (1955). Pleistocene temperatures. Journal of Geology 63, 538–578. EPICA Community Members (2004). Eight glacial cycles from an Antarctic ice core. Nature 429, 623–628. Grootes, P.M., Stuiver, M., White, J.W.C., Johnsen, S., and Jouzel, J. (1993). Comparison of oxygen isotope records from the GISP2 and GRIP Greenland ice cores. Nature 366, 552–554. Henderson, G.M., and Slowey, N. (2000). Evidence from U–Th dating against Northern Hemisphere forcing of the penultimate deglaciation. Nature 404, 61–66. Holzka¨mper, S., Mangini, A., Spo¨tl, C., and Mudelsee, M. (2004). Timing and progression of the last interglacial derived from a high alpine stalagmite. Geophysical Research Letters 31, L07201, doi:10.1029/2003GL019112. Johnsen, S.J., Clausen, H.B., Dansgaard, W., Fuhrer, K., Gundestrup, N., Hammer, C.U., Iversen, P., Jouzel, J., Staufer, B., and Steffensen, J.P. (1992). Irregular glacial interstadials recorded in a new Greenland ice core. Nature 359, 311–313. Kukla, G., McManus, J.F., Rousseau, D.-D., and Chuine, I. (1997). How long and how stable was the last interglacial? Quaternary Science Reviews 16, 605–612. Lambeck, K., and Chappell, J. (2001). Sealevel change through the last glacial cycle. Science 292, 679–685. Martinson, D.G., Pisias, N.G., Hays, J.D., Imbrie, J., Moore Jr., T.C., and Shackelton, N.J. (1987). Age dating and the orbital theory of the ice ages: Development of a high-resolution 0 to 300,000-year chronostratigraphy. Quaternary Research 27, 1–29. McManus, J.F., Oppo, D.W., Keigwin, L.D., Cullen, J.L., and Bond, G.C. (2002). Thermohaline circulation and prolonged interglacial warmth in the North Atlantic. Quaternary Research 58, 17–21. Mu¨ller, H. (1974). Pollenanalytische Untersuchungen mit Jahresschichtza¨hlungen an der holststeinzeitlichen Kieselgur von Munster Brehloh. Geologisches Jahrbuch A 21, 107–140. NorthGRIP. (2004). High resolution record of Northern Hemisphere climate extending into the last interglacial period. Nature 43, 147–151. Petit, J.R., Jouzel, J., Raynaud, D., Barkov, N.I., Barnola, J.M., Basile, I., Bender, M., Chappellaz, J., Davis, J., Delaygue, G., Delmotte, M., Kotlyakov, V.M., Legrand, M., Lipenkov, V.M., Lorius, C., Pe`pin, L., Ritz, C., Saltzman, E., and Stievenard, M. (1999). Climate and atmospheric history of the past 420,000 years from the Vostok ice core, Antarctica. Nature 399, 429–436. Raynaud, D., Barnola, J.-M., Souchez, R., Lorrain, R., Petit, J.-R., Duval, P., and Lipenkov, V. (2005). The record of marine isotopic stage 11. Nature 436, 39–40.

52

Frank Sirocko

Sa´nchez Gon˜i, M.F., Loutre, M.F., Peyron, O., Santos, L., Duprat, J., Malaize, B., Turon, J.-L., and Peypouquet, J.-P. (2005). Increasing vegetation and climate gradient in Western Europe over the last glacial inception (122–110 ka): data model comparison. Earth Planetary Science Letters 231, 111–130. Shackleton, N.J., and Opdyke, N. (1973). Oxygene isotope and paleomagnetic stratigraphy of equatorial Pacific core V28-238: Oxygene isotope temperatures and ice volume on a 105 and 106 * 10 year scale. Quaternary Research 3, 39–55.

Sirocko, F., Seelos, K., Schaber, K., Rein, B., Dreher, F., Diehl, M., Lehne, R., Ja¨ger, K., Krbetschek, M., and Degering, D. (2005). A Late Eemian Aridity Pulse in central Europe during the last glacial inception. Nature 436, 833–836. Trauth, M.H., Deino, A.L., G.N., B., and Strecker, M.R. (2003). East African climate change and orbital forcing during the last 175 kyr BP. Earth and Planetary Science Letters 206, 297–313. Turner, C. (2002). Problems of the duration of the Eemian interglacial in Europe North of the Alps. Quaternary Research 58, 45–48.

6. Late Quaternary Interglacials in East Antarctica From Ice-Core Dust Records Barbara Delmonte1,2, Jean Robert Petit1, Isabelle Basile-Doelsch3, Emil Jagoutz4 and Valter Maggi2 1

LGGE-CNRS, BP96, 38402, Saint Martin d’He`res, France University of Milano-Bicocca, DISAT, Piazza della Scienza 1, 20126 Milano, Italy 3 CNRS-CEREGE, UMR 6635, Europole Me´diterrane´en de l’Arbois, BP 80, 13545, Aix en Provence, France 4 Max Planck Institute of Chemistry, Kosmochemistry Department 55020 Mainz, Germany 2

ABSTRACT Aeolian dust records from deep East Antarctic ice cores evidence extremely low dust fluxes during the last five interglacials (10 to 25 times lower than in glacial periods), related to reduced primary production and mobilization on the Southern Hemisphere continents, to changes in atmospheric transport and hydrological cycle. The Sr–Nd isotope fingerprint of aeolian dust in Antarctica suggests a dominant southern South America provenance during Quaternary glacial times, but the first geochemical data for Stage 5.5 and the Holocene presented in this work show significant differences and open the possibility for a different source mixing. Dust-size variability in the EPICA-Dome C ice core suggests shorter transport time for dust or more direct air mass penetration to the site during interglacials with respect to cold periods and a clear multisecular scale mode of atmospheric circulation variability during the Holocene. 6.1 INTRODUCTION Mineral aerosol (dust) deflated from continental areas and transported in the atmosphere is of importance for the climate system (e.g. Harrison et al., 2001; Houghton et al., 2001 and references therein) for biogeochemical cycles and can be used as tracer

for depicting atmospheric circulation patterns and variability. Dust affects the solar radiation. The radiative impact of small (< 20 mm in diameter) dust particles is nonlinearly related to a number of factors such as their atmospheric concentration and vertical distribution, optical and aerodynamic properties as well as mineralogical composition (e.g. Sokolik and Toon, 1996; Claquin et al., 1998). Dust plays a role in atmospheric chemistry (see Harrison et al., 2001 for a general overview) and processes as formation of cloud droplets (Zhang and Carmichael, 1999) and ice nuclei (e.g. Rogers and Yau, 1989), as well as in biogeochemical cycles through nutrients supply to terrestrial (e.g. Swap et al., 1992) and marine ecosystems (e.g. Falkowski et al., 1998; Hutchins and Brunland, 1998; Fung et al., 2000), potentially influencing the global carbon cycle and the atmospheric concentration of greenhouse gases. Soil dust emissions at present time are estimated around 2150 Mt=yr on global scale (Houghton et al., 2001), with high spatial and temporal variability. In particular, emissions in the Southern Hemisphere, which is devoid of major dust sources, are less than one-fifth of those from the Northern Hemisphere, where North Africa, the Middle East, central Asia and the Indian subcontinent play a major role (Prospero et al., 2002). The mineral particles deflated from the austral continental landmasses are of

Barbara Delmonte et al. 0°

So

AN

uth

CE CO

I NT

Afr

ica

45°S

A TL

IAN IND

A

60°S

EA OC N

75°S

EAST ANTARCTICA DB

AN W TA ES RC T TIC

90°W

A

90°E

KMS

VK

EDC

4000 km

tra Au s

Ze a Ne w

AN

CE

O

lan d

C FI CI

lia

PA

interest for air mass tracking as they can be transported long range into the mid-to-high troposphere through the zonal westerly wind flux, ultimately reaching the interior of Antarctica, where they are archived in the ice layers. The reconstruction of Quaternary ‘background’ dust (of nonvolcanic origin) variability and provenance from deep ice cores recovered from low accumulation sites of the high East Antarctic Plateau can therefore provide detailed information about palaeoenvironmental conditions at the dust-source regions, air mass exchanges and variability between mid and high latitudes of the southern hemisphere and past atmospheric circulation changes on timescales from glacial–interglacial cycles (e.g. Petit et al., 1999) to submillennial, or centennial periods (Delmonte et al., 2004b, 2005). Amongst ice cores, the Vostok ice core (Fig. 6.1) (78 S, 106 E, 3480 m a.s.l.) first provided climate records spanning the last 420 kyr (Petit et al., 1999), followed by the 340 kyr long Dome Fuji record (Watanabe et al., 2003). Today, the EPICA (European Project for Ice Coring in Antarctica) deep ice core from Dome C (East Antarctica, 75 069S, 123 219E, 3233 m a.s.l.) allowed the extension of the climate record back to 740 kyr BP, corresponding to the last eight glacial cycles (EPICA Community Members, 2004). At shorter timescale, the investigation of the timing and magnitude of Holocene climatic variability has today become a major focus for the comprehension of present-day climatic trends and the assessment of anthropogenic contributions on natural climate change (Houghton et al., 2001). Over the last decade, a number of marine sediment cores from the North Atlantic (e.g. Bond et al., 1997; Bianchi and McCave, 1999), from the Indian Ocean (e.g. Sakar et al., 2000) and the South-East Pacific (e.g. Lamy et al., 2001), as well as atmospheric proxies from tropical latitudes (Moy et al., 2002) and northern polar areas (Mayewski et al., 1997) highlighted a pronounced millennial scale mode of variability of atmospheric and oceanic indicators over the

So Am uth eric a

54

180°

Fig. 6.1 Southern Hemisphere polar stereographic map of Antarctica with the location of ice cores presented here. EDC: EPICA–Dome C (75 06’S, 123 21’E, 3233 m a.s.l.); VK: Vostok (78 289S, 106 489E, 3480 m a.s.l.); DB: Dome B (77 059S, 94 559E, 3650 m a.s.l.); KMS: Komsomolskaya (74 059S, 97 299E, 3500 m a.s.l.).

Holocene epoch. High temporal resolution analyses followed and indicated that the millennial scale structures from North Atlantic deep-sea sediment records are actually composed by high-frequency oscillations (Bond et al., 2001). An 1500-year oscillation of climate was suggested and possibly associated with solar activity. At the high latitudes of the Southern Hemisphere, secular scale periodicities (of 200and 400-year duration) were detected (Leventer et al., 1996; Domack and Mayewski, 1999; Domack et al., 2001) in Holocene biogeochemical sediment records from the Antarctic Peninsula. The changes in sediment properties were interpreted in terms of local changes in upper ocean conditions. Aeolian dust preserved in Antarctic ice cores has the potential to document climate variability at different timescale; also, it can provide relevant information that cannot be inferred from other proxies. Indeed, for the

Late Quaternary Interglacials in East Antarctica

mineral aerosol spreads over Antarctica, the narrow and geographically fixed location of their main source regions (southern South America see §3), their relatively longdistance transport and their low chemical reactivity together make this proxy an almost passive tracer worth capturing the changes in atmospheric circulation patterns around Antarctica. Moreover, the size distribution of the dust is sensitive to the transport conditions, and it could provide an additional information on the pathway. This property is likely unique by comparison to other climatic indicators (e.g. stable isotope composition of ice or total concentration of chemical component) having extensive and variable sources (e.g. austral ocean) or more sensitive or reactive to other factors (e.g. temperature, vapour saturation pressure and other chemicals) during their pathway. In the following sections, we first present the broad patterns of aeolian dust variability in East Antarctica during the late Quaternary (over the last four glacial cycles and last 0.4 million years) within the global context of glacial/interglacial cycles (§2) and with particular focus on the characteristics of aeolian dust during interglacials. Then, we summarize the present-day knowledge about geographic provenance of dust to the Antarctic snow during glacial times and provide evidence for a possible source differentiation between cold glacials and warm interglacials (section 6.3). Afterwards, we present some recent findings on the spatiotemporal variability of dust transport in East Antarctica for two periods of interest: the last glacial maximum (LGM) to Holocene climatic transition (section 6.4) and the Holocene (section 6.5).

6.2 LATE QUATERNARY AEOLIAN DUST VARIABILITY IN ANTARCTICA 6.2.1 Glacial/interglacial cycles The late Quaternary stable isotope record from EPICA-Dome C (hereinafter EDC) ice

55

core (EPICA Community Members, 2004, Fig. 6.2a), taken as proxy for temperature variations, shows a sawtooth pattern of warm interglacial stages (and corresponding to marine isotopic stages – MIS – 11.3, 9.3, 7.5, 5.5 and to the Holocene) followed by glacial periods increasingly colder and punctuated by cool interstadials. The EDC isotope profile is very similar to the previous Vostok (Petit et al., 1999) and Dome Fuji (Watanabe et al., 2003) records, demonstrating a rather good uniformity of glacial– interglacial climate changes across Antarctica. Temperature changes highlighted also the correlation with atmospheric greenhouse gases (CO2 and CH4) content, the most rapid changes occurring during glacial–interglacial transitions (Petit et al., 1999; EPICA Community Members, 2004). The pattern of dust concentration variability from EDC (Fig. 6.2b) and Vostok (Fig. 6.2c) ice cores measured by Coulter Counter technique on discrete ice samples is very similar and is anti-correlated at first order with the isotope record. Interglacial periods are characterized by extremely low dust inputs ð0:40:6 mg m2 yr1 Þ, while 10 to 25 times higher fluxes characterize cold glacial stages (up to 50 times and even higher for concentrations, due to the reduced precipitation rate). This evidence points out that the major mode of glacial/interglacial aeolian dust variations in central Antarctica is remarkably uniform. Indeed, such pattern is of global significance. There is considerable evidence for enhanced atmospheric dust load during Quaternary cold periods, and consequently higher deposition in oceans and on continents, the evidence coming from northern polar records (e.g. Steffensen et al., 1997; Ruth et al., 2003), from terrestrial (e.g. Kukla et al., 1990) and from marine (e.g. Rea, 1994) sequences. A few regional exceptions however do exist (see Kohfeld and Harrison, 2001 for an update overview), especially from tropical and subtropical latitudes, generally associated with opposite climate changes or with the variability of atmospheric pathways. For

56

Barbara Delmonte et al.

–360 –390

200

300

5.5

400

9.3

7.5

11.3

3

–450

2

4

6

(b)

8

12

10

1000 100 10

1000 100 5.5

10 2

7.5

(d)

7.4

6 4

9.3 10

8

11.3 12

0 5.1 5.3

7.1 7.3

3

Chinese loess Magnetic susceptibility

(e)

7.5

–2 9.3

5.5

10

2

11.3 L4

L2

L1

δ18O‰

Vostok dust (ppb)

(c)

EPICA–DOME C dust (ppb)

–420

7.4

δD ‰ (V-SMOW)

(a)

100 Holocene

0

L3

100 S0

0

S1

100

S2

200

S3

300

S4

400

Age (kyr BP)

Fig. 6.2 Late Quaternary glacial/interglacial changes: comparison of ice cores, marine and terrestrial sequences. (a) – EDC isotope record (EPICA Community Members, 2004), proxy for temperature variations; (b) – EDC dust record, measured by Coulter Counter technique (Delmonte et al., 2002a, 2004a; EPICA Community Members, 2004); (c) – Vostok dust record (from Petit et al., 1999), with timescale adjusted on EDC; (d) – Marine sediment 18 O record (Bassinot et al., 1994), proxy for global ice volume; (e) – Magnetic susceptibility record of Chinese Loess (L) and palaeosol (S) sequences (from Kukla et al., 1990). Even and odd numbers indicate glacial and interglacial stages respectively. The two dust records (b) and (c) are reported in logarithmic scale since they derive from nonlinear (synergic) interaction among different environmental factors (see text). The black dashed lines provide a link among the different records with respect to cold periods.

example, a 300-kyr long marine record from southwestern Africa, off the Namib Desert (20–25 S), highlighted humid

glacials and dry interglacials occurring in response to equatorward shifts of precipitation belts during cold stages (Stuut et al., 2002). Climate and atmospheric variability over the Quaternary is mainly related to the periodicity of the Earth’s orbital parameters (e.g. Imbrie et al., 1993) modulating the amount, as well as the seasonal and the latitudinal distribution of solar radiation (Berger and Loutre, 1991) and likely initiating climate changes. As for most climate proxies recorded in Antarctic ice cores (Petit et al., 1999 and EPICA Community Members, 2004), a large part of the dust and isotope variance is concentrated in the 100- and 41-kyr spectral bands, while their links with similar periodicities of the eccentricity and the obliquity respectively still need to be assessed. The overall higher global atmospheric dust load characterizing cold climatic stages can be explained by a number of synergetic factors: like (1) the widespread continental aridity and the lower atmospheric vapour content imposed by cooler temperatures, leading to changes in soil moisture (hence resistance to erosion) and vegetation cover; (2) the enhanced primary production of dust through physical weathering processes in periglacial areas (e.g. frost and thawing cycles), which became particularly important during cold periods because of the glacier cover extension in mountain area and large ice sheet cover on continental landmasses; (3) the substantial enlargement of the dust-source areas due to sea-level lowering; (4) the reduced intensity of the hydrological cycle, leading to less efficient scavenging processes by precipitations and consequently to an increase in the atmospheric dust residence time (Yung et al., 1996); and finally (5) a generally more vigorous atmospheric circulation associated with steeper latitudinal thermal gradients (e.g. COHMAP Project Members, 1988; Kohfeld and Harrison, 2001). Under the generally windier conditions prevailing in the continents in glacial times, processes

Late Quaternary Interglacials in East Antarctica

like soil deflation, dust injection into the atmosphere and long-range transport were very efficient. Under mild interglacial climates, in contrast, the dust production, mobilization and transport decreased on average because of damper environmental conditions, dense vegetation cover, enhanced evaporation, precipitation and aerosol scavenging. Dust and isotope records from Greenland and Antarctic ice cores differ significantly. Large and very rapid dust changes occurred in Greenland during the Dansgaard– Oeschger (D/O) events of the last glacial period, and during the last climatic transition, where the Younger Dryas event was accompanied by a rapid return of dust concentration to glacial levels (e.g. Ruth et al., 2003). Aeolian dust archived in Greenland mostly originates from central Asia (Biscaye et al., 1997), travels for several thousands of kilometres through the high troposphere over the Pacific Ocean and finally reach northern latitudes. As suggested by the dust/isotope correlation, the dust input to Greenland is strongly anti-correlated with local temperature. For East Antarctic records, the southern South America is the dominant source region of dust during cold periods (see section 6.3). Indeed, the climate of this region is sensitive to the sea ice cover, and it was greatly influenced by the extended sea ice cover in the South Atlantic Ocean during glacial times, modulating the intensity and average position of the polar front (e.g. Heusser, 1989). Altogether, this makes the East Antarctic dust records more sensitive to the source and to the southern South American climate and environment. Moreover, this peculiar area under the westerlies is close to the Drake Passage and the Weddell Sea, which are key ocean locations well connected to large-scale deep-ocean currents (e.g. North Atlantic Deep Water, circum polar current) and sensitive in turn to the global ocean circulation and climate. The apparent correlation between the global ice volume (Fig. 6.2c) and the dust records from East Antarctica

57

supports such a view, with possibilities for large-scale teleconnections. Interestingly, the Antarctic dust records display a glacial/interglacial pattern that is similar to the magnetic susceptibility record of loess–palaeosoil sequences (Kukla et al., 1990) from the Chinese Loess Plateau (Fig. 6.2d). Despite the records yet having to be synchronized, the similarity suggests a global character or at least an intercontinental (central Asia–southern South America) connection of climate and environment with respect to aeolian dust production, mobilization, transport and accumulation. 6.2.2 Characteristics of dust during interglacials Figure 6.3 displays the records from the EDC ice core for stable isotopes and aeolian dust during the last five interglacials (Holocene, MIS 5.5, MIS 7.5, MIS 9.3 and MIS 11.3). A sharp rise of the isotope content with a variable pattern marks the climatic transitions (Terminations) from glacial stages to interglacials of variable duration. If the time period spent by the isotope record above the 403 per mil level (300-year average minimum value observed during the full Holocene epoch) is considered as threshold for interglacial temperature conditions (EPICA Community Members, 2004), the duration of the last five interglacials (Fig. 6.3a and 6.3d) looks very variable around 16 9 kyr, with a maximum of 29 kyr (MIS 11.3) and a minimum of 5 kyr (MIS 7.5) During these warm periods, the dust concentration in the EDC ice core is always very low (Fig. 6.3b and 6.3e), typically lower than 30 ppb, corresponding to about 0:5 mg m2 yr1 in flux on average (Table 6.1). From one interglacial to the other, dust concentration and fluxes are very similar, and it seems that in all cases the low (interglacial) dust levels are reached a few thousand years before the culmination of the isotope to the interglacial value. Sometimes a two-step increase of isotope content occurs afterwards as in the case of the last climatic

58

Barbara Delmonte et al. Termination I 0 1

–360

80 –360

–400 2

4

Termination II

Termination III

Termination IV

Termination V

100 120 140 160 180

220 240 260 280 300

300 320 340 360 380

380 400 420 440 460

5.5

–360

–360

7.5

–360

–400

–400

–400

9.3

12

–440

–440

–440

(b)

1000

–1000

–1000

–1000

–1000

100

Holocene

–440

100

1

CPP (%)

(c)

25

25

20

20

15

15

10

10

20 40 60 80

25

7.5

25 8

20 6

100 120 140 160 180

30 ppb

10 11.3

9.3

25

20

20

15

15

15

10

10

10

220 240 260 280 300

12

100

10

2

0

5.5

10

100

10

10

10

10

8

100

6

–403‰

–400

–440

2

11.3

10

8

6

(ppb)

δD‰ (V-SMOW)

(a)

20 40 60

300 320 340 360 380

Coarse dust 12

Fine dust 380 400 420 440 460

Age (kyr BP)

Fig. 6.3 EDC records of stable isotope, dust concentration and dust-size changes during the last five climatic transitions. (a) – The deuterium record with indication (horizontal dashed line) of the minimum 300-year average value observed during the full Holocene epoch (403 per mil, EPICA Community Members, 2004). The D values above this level are taken as indicators for interglacial temperature conditions, and interglacials are highlighted by the yellow bands. (b) – EDC dust concentration changes (log scale). Typical interglacial levels lie below the 30 ppb level (dashed horizontal line). (c) – Dust size (expressed as ‘coarse particle percentage’, CPP, corresponding to the proportion of particles having a diameter between 3 and 5 mm with respect to the total dust mass, typically included in the interval of diameter between 1 and 5 mm). At the time of each climatic transition, the mean particle size progressively decreases (blue arrows) during glacials and until the very beginning of the climatic transition, then sharply increases (red arrows) at the onset of warm conditions. It is interesting to point out that Termination III displays two drastic changes of dust concentration accompanied by a minimum of dust size. One occurs around 260–270 kyr BP, during stage 8, the other around 250 kyr BP during the climatic transition.

transition and the Antarctic Cold Reversal (ACR) during Termination I (Fig. 6.3), a pattern possibly associated with the timing of final sea ice retreat in the South Atlantic (Jouzel et al., 1995). At the end of each glacial period, a systematic change in the particle grain size occurs at Dome C (Fig. 6.3c and 6.3f). The particle grade is given here by the coarse particle percentage, CPP parameter (see section 6.4). CPP always increases at the onset of warm periods and on average is higher during interglacials than during glacials (Table 6.1). During the last five terminations,

particle size displays a decreasing trend at the end of each glacial period, and in spite of some variability with spikes of large particles, it displays a minimum value at the very beginning of the climatic transition at all times (Fig. 6.3c). Such behaviour for EDC dust suggests that similar phenomena characterize the end of each cold period and the onset of each interglacial. This can ultimately provide information on atmospheric circulation patterns over Dome C. The occurrence of the same pattern during the last 400 kyr suggests this is a typical feature of climatic transitions in East Antarctica and

Late Quaternary Interglacials in East Antarctica 0

δD‰ (V-SMOW)

(d)

–360

10

Holocene

–380

120 –360

130 –360

5.5

–380

240

–380

250

320

–360

7.5

330

340

390 400 410 420 –360

9.3

–380

59

–400

–400

–400

–400

–420

–420

–420

–420

–420

–440

–440

–440

–440

100

100

80

80

Dust concentration (ppb, linear scale)

(e)

60

~16 kyr

100

40

40

20 0 30

60 16 ± 10

~29 kyr

100

80 22 ± 18

–403‰

–440 ~15 kyr

100

80

60

15 ± 8

~5 kyr

11.3

–380

–400

80 16 ± 7

60

60

40

40

40

20

20

20

20

0 30

0 30

0 30

0 30

25

25

25

25

25

20

20

20

20

20

15

15

15

15

15

10

10

10

10

10

19 ± 17

Coarse dust

CPP (%)

(f)

0

10

120

130

240

250

320

330

340

Fine dust 390 400 410 420

Age (kyr BP)

Fig. 6.3 (Continued) (d) – Isotope record (same as (a)) for interglacials. Horizontal blue arrows indicate the broad duration of each interglacial, calculated as the time period spent by the isotope record above the 403 per mil level. (e) – EDC dust concentration changes during interglacials reported in linear scale. For the Holocene and MIS 5.5, the high-resolution data from Delmonte et al. (2004a) are reported. For MIS 7.5, MIS 9.3 and MIS 11.3, the records are the same as reported in (b). Average concentration levels (red dot) are similar among interglacials. (f) – Dust-size (CPP) changes during interglacials. For the Holocene and MIS 5.5, the high-resolution data are reported; for older interglacials, a zoom of the profile (c) is reported.

likely reflects conditions of the ocean surrounding Antarctica, sea ice extent and its distribution as well as peculiar atmospheric circulation patterns. A closer inspection of mineral dust-size variability at Dome C (Fig. 6.3f), moreover, shows a marked mode of variability

throughout each interglacial. This can be better appreciated during MIS 5.5 and particularly during the Holocene (as it will be discussed in §5), which are documented at higher temporal resolution. As far as dust-size variability is concerned, it is very difficult to extrapolate

Table 6.1 Aeolian dust concentration and flux in East Antarctica during the last five interglacial stages from EDC ice core, with indication of dust-size changes Dust concentration (ppb) Dust flux ðmg=m2 yr1 Þ Glacial–interglacial dust size change

Holocene

MIS 5.5

MIS 7.5

MIS 9.3

MIS 11.3

15 8

16 10

22 18

16 7

19 17

0.4 (0.2–0.6)

0.4 (0.2–0.7)

0.6 (0.2–1)

0.4 (0.2–0.6)

0.5 (0.1–0.9)

Increase CPP (ca. þ10%)

Increase CPP (ca. þ8%)

Increase CPP (ca. þ8%)

Increase CPP (ca. þ9%)

Increase CPP (ca. þ7%)

The average dust levels reported in the table are calculated over the period when the isotope curve lies above the minimum 300-year average value observed during the Holocene epoch (403 per mil, EPICA Community Members, 2004, see Fig. 6.3). Taking such level as indicator for interglacial temperature conditions, the duration of warm periods looks very variable around 16 9 kyr on average between MIS 5.5 ð 16 kyrÞ, MIS 7.5 ð 5 kyrÞ, MIS 9.3 ð 15 kyrÞ and MIS 11.3 ð 29 kyrÞ.

60

Barbara Delmonte et al.

results from EDC to the whole Antarctic since regional effects may play a fundamental role both during deglaciations (section 6.4.2) and during interglacials (section 6.5.2). Climate and dust records from other sites from the Plateau will be very instructive in this respect. 6.3 GEOGRAPHIC DUST PROVENANCE DURING GLACIALS AND INTERGLACIALS The source regions providing the bigger dust fluxes today are characterized, in general terms, by little or no ground cover, easily wind-erodible soils and seasonal wetness (Mahowald et al., 1999). A recent satellitebased worldwide geographical mapping of major atmospheric dust sources provided by Prospero et al. (2002) highlighted arid areas centred over topographical lows or on lands adjacent to topographical relief to constitute the major sources for dust transported long range. Indeed, terrains where the recent geomorphologic history has favoured the concentration of fine-grained material on low roughness surfaces are much more active suppliers than coarse-grained old sandy deserts already impoverished in the fine fraction (Tegen et al., 2002). Atmospheric transport exerts a strong size and mineralogical selection on particles, making small quartz

grains and plate clays, having the best aerodynamic properties, the most diffuse aeolian minerals around the globe (Gaudichet et al., 1992). The mineral dust deposited in ice on the East Antarctic Plateau after long-range transport is very fine grained (< 5 mm in diameter) and consists mostly of clays (mostly illite), crystalline silica, feldspars and minor amounts of pyroxenes and amphiboles, metallic oxides and volcanic glasses (Gaudichet et al., 1988, 1992). In order to depict the geographical provenance for aeolian dust at polar latitudes, Grousset et al. (1992) first proposed a successful geochemical approach, already used in oceanography (e.g. Grousset et al., 1988, Revel et al., 1996), consisting in the comparison of the 87Sr/86Sr versus 143Nd/144Nd isotopic signature of mineral particles extracted from Antarctic ice cores to that from potential source area (PSA) samples from the Southern Hemisphere (Table 6.2). The authors suggested a possible southern South American provenance for dust in East Antarctica during glacial stage 2 (MIS 2). Investigations on Antarctic dust provenance were largely developed later by Basile et al. (1997) and by Delmonte et al. (2004a, 2004b) on four East Antarctic sites from the plateau (Dome C, Vostok, Dome B and Komsomolskaya, Fig. 6.1). In parallel, the authors analysed a number of size-selected samples of loesses, aeolian deposits, sands

Table 6.2 87Sr/86Sr ð2 106 Þ, 143Nd/144Nd ð2 106 Þ and "Nd ð0Þ isotopic values for interglacial (Holocene and stage 5.5) samples from EDC and Vostok ice cores (1020 mg total dust per sample) Sample Holocene 1 – EDC Holocene 2 – EDC Holocene 3 – VK Holocene 4 – VK Stage 5.5 – EDC Stage 5.5 – VK Stage 5.5 – Volcanic

87

Sr/86Sr ð2 106 Þ 0.710013 (55) 0.709435 (37) 0.711200 (35) 0.709289 (50) 0.710213 (26) N.M. 0.704983 (36)

143

Nd/144Nd ð2 106 Þ 0.512407 (101) 0.512347 (95) 0.512126 (17) 0.512379 (44) 0.512211 (172) 0.512261 (75) 0.512823 (53)

"Nd ð0Þ 4:51 5:68 9:99 5:05 8:29 7:35 þ3:61

The whole analytical procedure is reported in Delmonte (2003) and in Delmonte et al. (2004a). Neodymium ratios are calculated as follows: "Nd ð0Þ ¼ ðð143 Nd=144 NdÞmeas =ð143 Nd=144 NdÞCHUR 1Þ 104 using the present-day CHUR (chondritic uniform reservoir) value for 143 Nd/144Nd ratio of 0.512638 (Jacobsen and Wasserburg, 1980). The measured 143Nd/144Nd were corrected for mass fractionation by normalizing to 146 Nd=144 Nd ¼ 0:7219, while the 87Sr/86Sr ratios were normalized to 86 Sr=88 Sr ¼ 0:1194.

Late Quaternary Interglacials in East Antarctica

and fluvioglacial sediments from key areas of the Southern Hemisphere and compared these to the ice core dust signature taking into account (1) size-dependent isotopic fractionation effects and (2) a possible 87 Sr=86 Sr shift for carbonate contribution (Delmonte et al., 2004a). The results of analysis are shown on Fig. 6.4. It can be observed that the isotopic signature for southern South America, constructed by collecting samples from the Argentine Pampas, Patagonia and other areas southward of 32 S, nicely fits that of ice core dust during glacial stages, suggesting southern South America was likely the dominant dust source during cold climatic stages for all the East Antarctic sites investigated. In particular, the South America samples from the Pampas region as well as those from the Cordoba loess region and from Patagonia south of 41 S show the best fit with glacial dust. The Sr–Nd isotopic fields for New Zealand and for the Antarctic Dry Valleys, however, also show a partial overlap with the South American and glacial dust ones (Fig. 6.4a), but complementary arguments allowed estimating their possible contribution as being negligible (Delmonte et al., 2004a). As far as New Zealand is concerned, the absence of tephra layers from the active Taupo Volcanic Zone in the Vostok ice core for the last 420 000 years (Basile et al., 2001), the very limited surface of both the North and South Island and the absence of significant contributions from the nearby large deserts of Australia suggest this region was unlikely to be the major dust source during cold periods. The Dry Valleys of coastal Antarctica were also unlikely the principal dust sources in glacial times: icefree areas were less extended than today during cold periods and the glacier’s fronts were closer to the coast (Denton and Hughes, 2002). Moreover, the primary production of fine-grained mineral particles was low in that region and limited to the mechanical alteration of rocks as the seasonal temperature variations and the hydrological cycle were reduced. Also, salts

61

deriving from the first stages of chemical weathering of rocks, like gypsum and carbonates, are very common in the Antarctic Dry Valleys (Campbell and Claridge, 1987), but they were not found in Antarctic ice (De Angelis et al., 1992). Finally, the strong catabatic winds blowing off the East Antarctic Plateau would lead to carry the mineral aerosol towards the ocean (Schwerdtfeger, 1984), and therefore a transport from the coast to the high Plateau would imply uplift of dust into the middle-high troposphere and a backward transport into the Continent, which seems quite unlikely. Using the same laboratory procedure as in Delmonte et al. (2004a), the Sr–Nd isotopic composition of EDC and Vostok dust from interglacial ice sections was measured and is reported in this work (Fig. 6.4a, 6.4b and Table 6.2). As the total amount of dust extracted from interglacial samples was extremely low (about 10 to 20 mg per sample), the error of measurement for Nd is relatively important. Despite this, and the limited number of samples analysed, the isotopic signature for interglacial dust looks significantly less radiogenic in 143Nd/144Nd ð10 < "Nd ð0Þ < 4:5Þ than glacial dust (Fig. 6.4b), with the exception of one sample from stage 5.5 containing volcanic ash, which shows obviously the typical fingerprint of volcanic rocks. The ice core dust isotopic data appear nicely aligned along a mixing line between two hypothetical volcanic and crustal end-members. The isotopic difference between samples from glacial and interglacial periods could be hardly accounted for by the South America dust collected south of 32 S and is reported in Fig. 6.4a. Indeed, environmental changes occurring within a given region and related to chemical weathering processes and/or pedogenesis would likely affect the Rb–Sr system (which is open and allows different elemental fractionation of Sr and Rb during chemical weathering leading to 87Sr/86Sr enrichment in the clay fraction) rather than the Sm–Nd system (Basile et al., 1997; Delmonte, 2003).

62

Barbara Delmonte et al.

II

VK glacial stages 4 to 12 EDC glacial stages 2, 4 and 6 DB glacial stage 2 KMS Termination I Old Dome C stage 2 EDC and VK interglacials [Holocene, Stage 5.5] EDC volcanic dust sample

I

South America

ε

8

Sr = 0

12

4

East Antarctic aeolian dust (glacials) 9

0

GLACIAL DUST New Zealand (South Island)

–12

Adelie land

Dry Valleys

0

NVL (Antarctica)

Af

–3

–9

Hol-EDC

Hol-VK Hol-EDC

–6

a ric

East Antarctic aeolian dust (interglacials)

–20

3

(Antarctica)>>>

h ut So

–16

Australia

εNd (0)

–8

New Zealand (North Island)

εNd (0)

–4

6

5.5-VK

INTERGLACIAL DUST

5.5-EDC Hol-VK

–24

IV

III 0.700

(a)

0.710

0.720

0.730

0.740

–12

0.750

87Sr/86Sr

0.704

(b)

0.706

0.708

0.710

0.712

0.714

87Sr/86Sr

Fig. 6.4a 87Sr/ 86Sr versus "Nd ð0Þ isotopic signature of East Antarctic ice core dust from glacials (white open circles, from Grousset et al., 1992; Basile et al., 1997; Delmonte et al., 2004a) and interglacials (white open squares, this work) and comparison with the signature of samples from the Southern Hemisphere potential source area (PSA). The PSA samples were selected for their < 5 m size fraction (with the exception of Australian samples, selected for their <10 m fraction), equivalent to the grain size of ice core dust. Isotopic fields were arbitrarily defined on the basis of the available points. The Antarctic sample from Ade´lie Land falls outside the graph. PSA data are from Grousset et al. (1992), Basile et al. (1997) and Delmonte et al. (2004a). Fig. 6.4b Glacial (Grousset et al., 1992; Basile et al., 1997; Delmonte et al., 2004a) and interglacial (Holocene and stage 5.5, this work) 87Sr/ 86Sr versus "Nd ð0Þ isotopic ratios (with error bars) of aeolian dust extracted from the four East Antarctic ice cores marked in Fig. 6.1 (VK: Vostok; EDC: EPICA–Dome C; DB: Dome B; KMS: Komsomolskaya). The arrow indicates an additional sample from stage 5.5 for which only Nd isotopes were measured (see Table 6.2). The volcanic dust sample was extracted from one EDC ice core section corresponding to stage 5.5. The dashed lines are mixing hyperbolae traced as example between hypothetical volcanic and crustal end-members (from Basile et al., 1997; Delmonte, 2003).

Therefore, basically two possibilities must be considered. The first is that interglacial dust still originates mostly from South America, but from an area which has not been captured by the source sampling south of 32 S used to construct the isotopic field of Fig. 6.4a. This first hypothesis opens the possibility for other latitudinal bands of

South America (north of 32 S, in agreement with the Nd isotopic gradient observed by Smith et al., 2003) to contribute during interglacials. However, this seems quite unlikely as glacial/interglacial atmospheric circulation changes were associated with a mean southward rather than northward displacement of westerly winds and the polar front.

Late Quaternary Interglacials in East Antarctica

A more reasonable possibility is that a somewhat different mixture of sources to the East Antarctic occurred in interglacial times. On the basis of the Sr and Nd concentrations and isotopic composition of the end-members (Faure, 1986), different mixing hyperbolae can be obtained. At this stage, no definite conclusions can be drawn as candidate end-members as Australia for instance needs to be documented at higher detail. Preliminary results with source data available today (not shown) suggest a theoretical mixing hyperbola which could be traced between South America (defined on the basis of the average values of South America samples) and a south Africa endmember (from the data reported in Fig. 6.3a) does not encompass the interglacial dust field. This points to this mixture as an unlikely candidate for interglacial dust in Antarctica. Conversely, a mixture of South American and Australian dust seems very likely, but further Australian data are needed at this step. 6.4 THE LGM TO HOLOCENE TRANSITION 6.4.1 Dust flux changes A recent compilation of literature data (DIRTMAP database, Kohfeld and Harrison, 2001) assembling ice core, terrestrial and marine evidence for deposition rates and transport paths of aeolian lithogenic material during the LGM to Holocene transition ( 21 to 10 kyr BP) pointed out that the magnitude and sign of atmospheric dust flux changes were not globally uniform. Significant differences exist between marine sediment records from the three ocean basins and, regionally, among different latitudes. While the mean global LGM/Holocene dust flux ratio was estimated around 2.5 (Mahowald et al., 1999), for polar areas, the first estimates from Greenland (Steffensen et al., 1997) and Antarctic (Petit et al., 1999) ice cores already revealed considerably higher values.

63

Recent investigations (Delmonte et al., 2004a, 2004b) on East Antarctic ice cores drilled at different locations (DB, EDC, KMS, see Fig. 6.1) on the interior Plateau pointed out a remarkable homogeneity of dust flux variability over the last climatic transition (20 to 10 kyr BP), as can be observed in Fig. 6.5b. The very high levels characterizing the LGM started to decrease sharply around 18 kyr BP, reaching typical Holocene values at approximately 14.5 kyr BP. The refined estimate of LGM/Holocene dust input to East Antarctica suggested by these studies is 2628 in flux and 53–55 in total concentration. An interesting feature characterizing Termination I (i.e. LGM to Holocene transition) in different ice cores from the Plateau is a shallow re-increase of dust concentration levels throughout the ACR phase of the isotope record (14.5 to 12.2 kyr BP, see Fig. 6.5a, 6.5b) and a well-marked, 800–1000 years long pre-Holocene dust minimum (ca. 11.3 to 12.1 kyr BP). These features suggest a tight link between Antarctic aeolian dust and the environmental conditions at the dust source regions, the ACR dust event likely representing a possible return to cooler conditions during the last climatic transition, the pre-Holocene dust minimum reflecting hydrological changes, in terms of either increased humidity at the source or increased scavenging of dust en route (Delmonte et al., 2002a). The pre-Holocene dust minimum, moreover, represents a very useful stratigraphic marker for Antarctic ice cores. The rather poor temporal resolution of dust measurements during Terminations II to V (see Fig. 6.3b and 6.3e) does not allow excluding the hypothesis that a similar dust event occurred also during older climatic transitions. 6.4.2 Dust-size changes: evidence for regional diversity of dust transport patterns inside the Plateau The mass (volume)–size distribution of aeolian particles in Antarctic ice is

Barbara Delmonte et al. Kyr BP 10

12

14

16

18

20

δD ‰ (V-SMOW)

(a) 1000 –400

100 –450

(b)

Dust mass (ppb)

64

10 Fine

(c) 55

DB

DC KMS

45

40

–60° 60° –65°

(d)

–75° –80° –85°

DB

DB Coarse

DC 90°

–70°

Dust size

FPP (%)

50

[I]

KMS

[II] –60° 60°

–60° 60° –65° –70° –75° –80° –85°

90 °

–65° –70° –75° –80° –85°

° 90

–65° –70° –75° –80° –85°

[IV] –60° 60°

90°

EDC

–65° –70° –75° –80° –85°

90°

120°

VK

[III] –60° 60°

120°

120°

120°

120°

° 180

18 kyr BP

0° 18

0° 18

15 kyr BP

15 0°

18

0°

10 kyr BP

15 0°

15 0°

15 0°

0°

15

° 180

= SUBSIDENCE ZONE

20 kyr BP

Fig. 6.5 Climate and dust records for the last climatic transition (20 to 10 kyr BP). (a) – Stable isotope profile from EDC ice core (Jouzel et al., 2001). (b) – Dust concentration records from DB (black line), EDC (red) and KMS (blue) smoothed with a 200-yr running average. The grey band marks the pre-Holocene dust minimum period. (c) – DB (black), EDC (red) and KMS (blue) records of dust size, expressed as fine particle percentage (FPP). FPP corresponds to the proportion of particles having a diameter between 1 and 2 m with respect to the total dust mass, which is typically between 1 and 5 m (Delmonte et al., 2004b). (d) – Location of the ice core drilling sites (left) and sketch of the polar vortex migration hypothesized for the last climatic transition. The shaded area marks the average position of preferential upper-air convergence and subsidence areas over central East Antarctica. After the last glacial maximum (LGM), around 18 kyr BP, the vortex expanded or shifted southwards towards the DB region. At about 15 kyr BP, all sites were almost under the same influence or circulation regime. The vortex displacement continued until 10 kyr BP, corresponding to the beginning of the Holocene.

approximately log-normally distributed around a mean modal value of 2 mm, a value that is typical for long-range dust transports (Schulz et al., 1998). However, a slight but significant variability of dust size among different drilling sites and climatic

stages was recently observed (Delmonte et al., 2004b), as it can be observed in Fig. 6.6. Behind the overall uniformity of dust concentration variability characterizing the LGM to Holocene transition in East Antarctica (Fig. 6.5b), the timing and pattern of

Late Quaternary Interglacials in East Antarctica

dust-size variability actually shows an unequivocal opposite behaviour (Fig. 6.5c, where the dust size is expressed as the ‘fine particle percentage’, FPP, i.e. the proportion of fine particles to the total mass) between the EDC–KMS and the DB (and probably Vostok) regions. While EDC and KMS change from very fine glacial to relatively coarse interglacial dust throughout

Termination I (Figs 6.5c, 6.6b, 6.6c, 6.6e and 6.6f), changes were clearly opposite in sign for DB (Figs 6.5c, 6.6a and 6.6d). Dust-size fractionation observed among East Antarctic sites sharing common sources was attributed to atmospheric transport pathways, in terms of either horizontal trajectories or altitude of advections (Delmonte et al., 2004b). During LGM, preferential

DB

KMS

(a) 16 kyr BP

65

EDC

(b)

(c)

1.6

FPP

CPP

FPP

CPP

FPP

CPP

1.2

0.8

(%)

0.4

0.0

2.3 μm

(d) 12 kyr BP 1.6

FPP

1.8 μm

(e)

CPP

FPP

1.7 μm

(f) CPP

FPP

CPP

1.2

0.8

1.8 μm

2.2 μm Particle diameter (μm)

5

4

3

0.7 0.8 0.9 1

5

4

3

0.7 0.8 0.9 1

5

4

3

0.0 0.7 0.8 0.9 1

(%)

0.4

1.9 μm

Fig. 6.6 Examples of normalized volume (mass)–size distributions of mineral dust in East Antarctic ice samples. Each distribution corresponds to an average of three to four adjacent samples selected around 16 and 12 kyr BP. The mode of the four-parameter Weibull function used to fit the raw data is also indicated (from Delmonte et al., 2004b, modified). The mode of a simple log-normal fit is also suitable for particle size characterization (e.g. Steffensen et al., 1997). The size distributions are cut at the first zero value of the distribution. The horizontal arrows indicate the typical intervals where the percentage of fine and coarse particles is calculated (FPP and CPP respectively). This approach is closer to raw data with respect to the Weibull (or lognormal) fit, and this is particularly important when concentrations are very low as in the case of interglacials. The FPP (e.g. Fig. 6.5c), the CPP (Fig. 6.7c and 6.7d) and the modal value of particle mass–size distribution (Fig. 6.6) are comparable indicators for dust size, and therefore they can be used interchangeably. Each of them captures the overall pattern of dust-size variability, an increase of 0:50 m in the mode corresponding to ’an 1015 % increase (decrease) in CPP (FPP) for EDC and Vostok dust. Obviously, with respect to FPP variations, the mode and CPP variations are opposite.

66

Barbara Delmonte et al.

upper-air subsidence (associated with fine dust) likely characterized the EDC–KMS region, while DB and Vostok were involved by more direct penetration of relatively lower altitude air masses. A possible explanation for the opposite changes observed is a progressive poleward migration of a hypothetical, localized centre of upper-air subsidence within the Antarctic polar vortex (Fig. 6.5d) during the last glacial/interglacial transition (20 to 10 kyr BP) along with the general reorganization of the atmospheric circulation and Southern Ocean conditions. Such hypothesis is also supported by the glacial to Holocene regional changes of snow precipitation and reflected by an increase of the ratio of accumulation rate between EDC and Vostok (Udisti et al., 2004). Indeed, the Antarctic vortex is a permanent pattern of the Earth’s atmospheric circulation since the Quaternary period (e.g. King and Turner, 1997 and references therein), and among its characteristics, the variability of its eccentricity with respect to the geographic south pole is spread over a large spectrum of timescales, from daily to glacial/interglacial cycles. Therefore, the variability superposed on a progressive reduction of the eccentricity of the polar vortex may account for the asynchronous submillennial dust-size oscillation and for the main trend observed for the three records during Termination I (Fig. 6.5c). Such a variability is also clearly expressed in records covering the Holocene epoch (see section 6.5). It has been observed (section 6.2.2) that a glacial/interglacial dust-size increase at EDC is a typical feature for the last five Terminations. Given the regional variability of atmospheric patterns during the LGM to Holocene transition, a comparison between the EDC dust-size record and other sites would be very helpful also for older deglaciations. This would allow appreciating whether the migration of the Antarctic polar vortex is a typical feature of late Quaternary glacial/interglacial transition.

6.5 HOLOCENE EPOCH 6.5.1 Evidence for multisecular scale changes The low-resolution profiles of dust-size variability during the last five interglacials reported in Fig. 6.3c and 6.3f show that warm periods are characterized by significant dustsize changes; in East Antarctica these have been investigated in great detail for the Holocene epoch. The first record from EDC spanning the period from 13 to 2 kyr BP highlighted a pronounced multisecular scale mode of particle size variability (Delmonte et al., 2002b), linked to atmospheric circulation changes in the Antarctic and circumAntarctic regions. Further investigations at high temporal resolution (Fig. 6.7c and 6.7d) definitely revealed that multisecular and submillennial scale periodicities indeed are imprinted on the Holocene dust-size records both from EDC and from Vostok. The stable isotope (deuterium) profile (Fig. 6.7a) shows short-term low-amplitude oscillations superposed on the main trend of changes and an early Holocene climatic optimum between 11.5 and 9 kyr BP. Millennial and centennial scale cycles are only weak and aperiodic and have been detected only from stacked Holocene Antarctic ice core isotopic records (Masson et al., 2000). Isotope variability is less evident than that of the dust-size record, likely because the source for vapour and precipitation is the ocean surrounding Antarctica, while areas supplying mineral dust are geographically fixed and narrow. Also, the record of total dust concentration from EDC and Vostok ice cores (Fig. 6.7b) does not show clear evidence for periodic variability but shows instead a long-term Holocene decrease starting just after the end of the pre-Holocene dust minimum (section 6.4.1). Indeed, the total concentration depends on several parameters (section 6.2), while particle grading depends mainly on transport conditions. The dust-size changes observed during the Holocene at EDC and Vostok sites are

Late Quaternary Interglacials in East Antarctica 4

(a)

6

8

10

12

Holocene

–380

(b)

10

(c)

EDC

30

20

CPP (%)

Dust mass (ppb)

–400 100

δD ‰ (V-SMOW)

2

10

(d) CPP (%)

40

VK

30

20 2

4

6

8

10

12

Age (kyr BP)

Fig. 6.7 EDC and Vostok Holocene records. (a) – Deuterium record (D expressed in ‰ with respect to Vienna–SMOW) from the EDC ice core (Jouzel et al., 2001). (b) – Total dust mass concentration (ppb or 109 g=g) for EDC (dark grey line) and Vostok (light grey line) ice cores (moving average over 70 years). (c) – EDC and (d) Vostok records of coarse particle percentage (CPP, %). For EDC, coarse particles were calculated within the 3 to 5 mm interval of diameter, while for Vostok, displaying smaller dust on average during the Holocene, coarse particles were calculated within the 2.5 to 5 m interval. For (c) and (d), the thin black lines represent the 70-year moving average, while the thick lines superposed represent the records filtered in order to keep only periodicities > 1000 years.

less pronounced than the observed glacialto-Holocene differences. Indeed, dust-size variability highlights mesoscale diversity of climatic and atmospheric regimes existing among East Antarctic sites, even when relatively close and located at similar altitude and distance from the coast. The

67

phenomenon that was already observed for the last glacial/interglacial climatic transition, therefore, likely holds also for the Holocene epoch. As already noted, the resolution of measurements does not allow appreciating if a similar mode of variability was operative in earlier interglacial times. Secular and multisecular scale modes of variability are clearly embedded in the two East Antarctic Holocene dust-size records from EDC and Vostok, as it can be observed from the smoothed dust-size records reported in Fig. 6.7c and 6.7d. These shortterm periodicities, moreover, are embedded within millennial scale structures, particularly evident in the early and mid-Holocene part of the EDC record (13 to 5 kyr BP). Spectral analyses indicate that a 200-year periodicity is common to EDC and Vostok Holocene dust-size records. These shortterm particle size oscillations of secular and millennial scale duration are similar to those superposed on the long-term see-saw trend of the changes during the last climatic transition (section 6.4.2) and originate probably also from the variable location of the polar vortex. 6.5.2 The 200-year dipolar oscillations Taking advantage of a number of stratigraphic markers representing strong chronological bridges between the EDC and the Vostok ice cores, the cross-spectral analysis performed on the records highlighted a common band of variability around 200 years, but a phase opposition (or, in other terms, an 100-year lead or lag) at this frequency for the two sites. The phase shift can be also qualitatively appreciated in Fig. 6.8, where EDC and Vostok series were filtered in the 200-year periodicity band. The coherence of the filtered components of EDC and Vostok (inverted scale) is evident over more than 5000 years (i.e. from 9.8 to 4.2 kyr BP). Such opposite behaviour between the two sites may reflect opposite pathways of air mass advection poleward, varying

68

Barbara Delmonte et al.

according to the position of the average centre of upper cyclonic vorticity (Fig. 6.9b). A linear combination of the dust-size indices (CPP) from the two sites provided useful indicators for characterizing the atmospheric variability between sites. The composite difference ( parameter) and the composite sum ( parameter) are taken here as indicators for the asymmetrical and the symmetrical modes respectively of the atmospheric circulation variability over the sector of the East Antarctic Plateau (Fig. 6.9). As a result, a sketch of atmospheric circulation changes associated with the EDC and Vostok dust-size variability is reported in Fig. 6.9b, where x and y axes are the dustsize parameters (CPP), while the diagonals are the two composite indicators for regional differences ðÞ and for the overall air mass advection over Antarctica ðÞ. Since the symmetric mode is likely related to the meridional pressure gradient with respect to lower latitudes, the parameter was suggested (Delmonte et al., 2005) as proxy for the long-term variability of the Antarctic oscillation (AAO). The power spectrum of (Fig. 6.9a), which is smooth and overprinted by a continuum of multicentury (150 to 500 years) 4

5

6

7

9

10

180- to 220-year filter

6 EPICA CPP (%) filtered

8

–6

4

–4

2

–2

0

0

–2

2

–4

4

–6

6 3

4

5

6

7

Age (years BP)

8

9

Vostok CPP (%) filtered – inverted scale

3

10

Stratigraphic markers

Fig. 6.8 EDC and Vostok CPP records (Vostok scale inverted) filtered around 200 years (180- to 220-year window). EDC: dark grey line; Vostok: light grey. Arrows indicate the position of stratigraphic markers which tighten the relative chronology of the two ice records.

periodicities likely represents the dynamics of the air masses exchanges between Antarctica and lower latitudes or that from the coupled ocean–atmosphere–sea ice system which governs such exchanges. On the other hand, the asymmetric mode would correspond to the oscillation of the centre of the polar vortex, with opposite influence on the two sites. This could be interpreted as a dipolar oscillation, with alternating subsidence of high altitude air over one site and advection from relatively lower troposphere levels over the other. The spectral analysis of the parameter (Fig. 6.9a) indicates a pronounced 200-year mode of variability (172- to 240-year band) as well as a shorter one around 125–160 years. Such mode of variability is similar to the secular variability of the solar activity with the 205year DeVries solar cycle which is still poorly documented. Finally, the secular variation of dust size may throw new light on the possible influence of solar activity on the Antarctic climate. 6.6 CONCLUSIONS AND PERSPECTIVES Continental dust preserved in ice core brings new insights for palaeoclimate studies and represents a proxy for source changes and a reliable tracer of atmospheric circulation. Today, the EDC and the Vostok ice cores document the last five climatic cycles in East Antarctica and evidence a glacial/interglacial decrease of continental aeolian dust flux by a factor higher than 20, while global estimates are about 2.5. The last five interglacial periods in East Antarctica are characterized by similar and very low dust fluxes representing only 0:5 mg m2 per year (about 15 to 30 ppb of total concentration). The Antarctic dust records display a first-order pattern similar to loess and palaeosol sequences from China, suggesting possible climate connections between Asia and South America.

Late Quaternary Interglacials in East Antarctica

69

Frequency (10–3 cycles/yr) (a)

Power

106

2

0

4

8

10

Δ (EDC – VK) CPP 240–172

160–125

105

104 106

Power

6

Σ (EDC + VK) CPP

105

104 0

2

4

6

8

10

Frequency (10–3 cycles/yr) (b)

[II] Δ

EDC cpp

6 –60° 0° –65° –70° –75° –80° –85°

[I] Σ ° 90

° 90

–60° 60° –65° –70° –75° –80° –85°

120°

120°

15

0°

0°

15

0° 18

0° 18

Σ = (EDC + Vostok) Δ = (EDC – Vostok)

120°

120°

15 0°

15 0°

18

0°

Σ 18

0°

[III]

VOSTOK cpp

° 90

–60° 60° –65° –70° –75° –80° –85°

° 90

–60° 60° –65° –70° –75° –80° –85°

Δ [IV]

Fig 6.9 Symmetric and asymmetric modes of variability of atmospheric circulation over East Antarctica during the Holocene. (a) – Spectral analysis (multitaper method or MTM, Dettinger et al., 1995) of (top) and (bottom) with indication of the 95 and 99% confidence levels with respect to a red noise signal. (b) – Vostok CPP and EDC CPP dust parameters are represented by the x and y axes respectively. The 60 E– 180 E sector of the East Antarctic Plateau with location of drilling sites is reported as in Fig. 6.5d. DB and KMS sites are assumed to have a behaviour similar to the Vostok and EDC regions respectively, as during the climatic transition (Fig. 6.5c). The (sum) and (difference) composite dust parameters are represented by the first and second diagonals. The arbitrarily rounded grey area represents the major centre of subsidence. Along the axis, a symmetric mode of variation from generally enhanced (quadrant III) to reduced (quadrant I) subsidence may possibly be associated with the AAO variability. Along the axis (asymmetric mode), the area of subsidence varies from an eccentric location (quadrant IV) to a more poleward position (quadrant II). This may correspond to an atmospheric dipole over East Antarctica having a pronounced 200year periodicity (Fig. 6.9a).

70

Barbara Delmonte et al.

Geochemical (Sr–Nd isotopes) fingerprinting allows identifying a southern South American provenance for aeolian dust in Antarctica during glacial periods. Owing to the very low amount of material, isotopic measurements of dust from interglacials require specific analytical care, and therefore data are limited. For the Holocene and MIS 5.5, the isotopic composition of dust appears slightly different from that for glacial periods and suggests a possible mixture with other sources. Contribution of Australian dust is likely, but additional analyses of this potential source are needed at this step. The size distribution of dust archived in Antarctic ice cores provides pertinent and new documentation of the variability of atmospheric transport onto the East Antarctic Plateau. The assumption is based on the grading of the aeolian dust with transport time. For central Antarctica, we associate relatively large particles with advections from low atmospheric levels and, conversely, fine dust to upper-air subsidence likely corresponding to the centre of the polar vortex. The EDC ice core shows a systematic dust-size increase during each glacial/interglacial climatic transition for the last 500 kyr, as well as a pronounced mode of variability during interglacials. Comparing East Antarctic dust-size records, moreover, indicates clear regional differences between sites and an opposite air mass influence during Termination I and the Holocene. The regional variability is expressed by secular and submillennial asynchronous dust-size oscillations which are clear in all records. Dustsize changes are associated with the variable regional influence of a hypothetical, localized centre of upper-air subsidence within the Antarctic polar vortex. During the climatic transition, a progressive change in the location of the centre of the polar vortex towards the geographic South Pole is suggested. For the Holocene, a variable strength of the subsidence (or in the AAO) as well as a secular wobbling of the centre of the polar

vortex is likely. The 200-year band of variability evokes a possible link with solar activity and calls for further developments. Further work is necessary to better assess the location of geographic sources for dust during late Quaternary interglacials and to document the modes of variability of the southern climatic system at different timescale. Similarly, modelling efforts are needed to assess the dust variability as a tracer of the atmospheric transports over Antarctica, and to support our hypothesis. The ongoing EPICA projects at Dome C and Dronning Maud land represent opportunities for extending the record back in time (up to 800 000 years) and to provide new records from different locations in Antarctica. The study of past interglacial periods from Antarctic ice cores as well as the documentation of the most recent period for which considerable information is available globally is of importance and together could contribute to the debate about the anthropogenic impact relative to the natural variability of the climate system, in the context of global warming. ACKNOWLEDGEMENTS The work of B. Delmonte was supported by a 2003 Prince of Asturias Fellowship awarded by the Scientific Committee for Antarctic Research (SCAR). We thank B. Narcisi (ENEA, Rome) for her help. This work is a contribution to the ‘European Project for Ice Coring in Antarctica’ (EPICA), a joint ESF (European Science Foundation)/EC scientific programme, funded by the European Commission under the Environmental and Climate Programme (1994–1998) contract ENV4-CT95-0074 and by national contributions from Belgium, Denmark, France, Germany, Italy, the Netherlands, Norway, Sweden, Switzerland and the United Kingdom. The Vostok ice core was obtained through the Russia–US– France tripartite collaboration and benefited from the support of Russian Antarctic

Late Quaternary Interglacials in East Antarctica

Expeditions, the US National Foundation and the French Paul Emile Victor Institute (IPEV).

REFERENCES Basile, I., Grousset, F.E., Revel, M., Petit, J.R., Biscaye, P.E., Barkov, N.I., 1997. Patagonian origin of glacial dust deposited in East Antarctica (Vostok and Dome C) during glacial stages 2, 4 and 6. Earth and Planetary Science Letters 146, 573–589. Bassinot, F.C., Labeyrie, L.D., Vincent, E., Quidelleur, X., Shackleton, N.J., Lancelot, Y., 1994. The astronomical theory of climate and the age of the Brunhes-Matuyama magnetic reversal. Earth and Planetary Science Letters 126, 91–108. Berger, A., Loutre, M.F., 1991. Insolation values for the climate of the last 10 million years. Quaternary Science Reviews 10, 297–317. Bianchi, G.G., McCave, I.N., 1999. Holocene periodicity in North Atlantic climate and deep-ocean flow south of Iceland. Nature 397, 515–517. Biscaye, P.E., Grousset, F.E., Revel, M., Van der Gaast, S., Zielinski, G.A., Vaars, A., Kukla, G., 1997. Asian provenance of glacial dust (stage 2) in the Greenland Ice Sheet Project 2 ice core, summit, Greenland. Journal of Geophysical Research 102, 26765– 26781. Bond, G., Showers, W., Cheseby, M., Lotti, R., Almasi, P., DeMenocal, P., Priore, P., Cullen, H., Hajadas, I., Bonani, G., 1997. A pervasive millennial scale cycle in the North Atlantic Holocene and glacial climates. Science 278, 1257–1266. Bond, G., Kromer, B., Beer, J., Muscheler, R., Evans, M.N., Showers, W., Hoffmann, S., Lotti-Bond, R., Hajadas, I., Bonani, G., 2001. Persistent solar influence on North Atlantic climate during the Holocene. Science 294, 2130–2136. Campbell, I.B., Claridge, G.G.C., 1987. Antarctica: soils, weathering processes and environment. Elsevier Amsterdam, 368 pp. Claquin, T., Schulz, M., Balkanski, Y., Boucher, O., 1998. Uncertainties in assessing radiative forcing by mineral dust. Tellus B50, 491–505. COHMAP Project Members, 1988. Climatic changes of the last 18 000 years: observations and model simulations. Science 241, 1043–1052. De Angelis, M., Barkov, N.I., Petrov, V.N., 1992. Sources of continental dust over Antarctica during the last glacial cycle. Journal of Atmospheric Chemistry 14, 233–244. Delmonte, B., 2003. Quaternary variations and origin of continental dust in East Antarctica. Ph.D.

71

Thesis, Department of Earth Science, University of Siena, Italy. 274 pp. Delmonte, B., Petit, J.R., Maggi, V., 2002a. Glacial to Holocene implications of the new 27,000-year dust record from the EPICA Dome C (East Antarctica) ice core. Climate Dynamics 18, 647–660. Delmonte, B., Petit, J.R., Maggi, V., 2002b. LGMHolocene changes and Holocene millennial-scale oscillations of dust particles in the EPICA Dome C ice core, East Antarctica. Annals of Glaciology 35, 306–312. Delmonte, B., Basile-Doelsch, I., Petit, J.R., Maggi, V., Revel-Rolland, M., Michard, A., Jagoutz, E., Grousset, F., 2004a. Comparing the EPICA and Vostok dust records during the last 220 000 years: stratigraphical correlation and provenancein glacial periods. Earth Science Reviews 66, 63–87. Delmonte, B., Petit, J.R., Andersen, K.K., BasileDoelsch, I., Maggi, V., Lipenkov, V., 2004b. Dust size evidence for opposite regional atmospheric circulation changes over East Antarctica during the last climatic transition. Climate Dynamics 23, 427–438. Delmonte, B., Petit, J.R., Krinner, G., Maggi, V., Jouzel, J., Udisti, R., 2005. Ice core evidence for secular variability and 200-year dipolar oscillations in atmospheric circulation over East Antarctica during the Holocene. Climate Dynamics 24, 641–654. Denton, G.H., Hughes, T.J., 2002. Reconstructing the Antarctic ice sheet at the last glacial maximum. Quaternary Science Reviews 21, 193–202. Dettinger, M.D., Ghil, M., Strong, C.M., Weibel, W., Yiou, P., 1995. Software expedites singularspectrum analyses of noisy time series. Eos Trans 76, 12. Domack, E.W., Mayewski, P.A., 1999. Bi-polar ocean linkages: evidence from late-Holocene Antarctic marine and Greenland ice core records. The Holocene 9, 247–251. Domack, E., Leventer, A., Dunbar, R., Taylor, F., Brachfeld, S., Sjunneskog, C., Party, O.L.S.., 2001. Chronology of the Palmer Deep site, Antarctic Peninsula: a Holocene palaeoenvironmental reference for the circum-Antarctic. The Holocene 11, 1–9. EPICA Community Members, 2004. Eight glacial cycles from an Antarctic ice core. Nature 429, 623–628. Falkowski, P.G., Barber, R.T., Smetacek, V., 1998. Biogeochemical controls and feedbacks on ocean primary production. Science 281, 200–206. Faure, G., 1986. Principles of isotope geology. John Wiley and Sons, New York. Fung, I., Meyn, S., Tegen, I., Doney, S.C., John, J., Bishop, J.K.B.., 2000. Iron supply and demand in the upper ocean. Global Biogeochemical Cycles 14, 281–296.

72

Barbara Delmonte et al.

Gaudichet, A., Angelis, M.D., Lefevre, R., Petit, J.R., Korotkevitch, Y.S., Petrov, V.N., 1988. Mineralogy of insoluble particles in the Vostok Antarctic ice core over the last climatic cycle (150 kyr). Geophysical Research Letters 15, 1471–1474. Gaudichet, A., Angelis, M.D., Jossasume, S., Petit, J.R., Korotkevitch, Y.S., Petrov, V.N., 1992. Comments on the origin of dust in East Antarctica for present and ice age conditions. Journal of Atmospheric Chemistry 14, 129–142. Grousset, F.E., Biscaye, P.E., Zindler, A., Prospero, J.M., Chester, R., 1988. Neodymium isotopes as tracers in marine sediments and aerosols: North Atlantic. Earth and Planetary Science Letters 87, 367–378. Grousset, F.E., Biscaye, P.E., Revel, M., Petit, J.R., Pye, K., Jossaume, S., Jouzel, J., 1992. Antarctic (Dome C) ice-core dust at 18 k.y. B.P.: isotopic constraints and origins. Earth and Planetary Science Letters 111, 175–182. Harrison, S.P., Kohfeld, K., Roelandt, C., Claquin, T., 2001. The role of dust on climate changes today, at the last glacial maximum and in the future. Earth Science Reviews 54, 43–80. Heusser, C., 1989. Polar perspective of late-Quaternary climates in the Southern Hemisphere. Quaternary Research 32, 60–71. Houghton, J.T., Ding, Y., Griggs, D.J., Noguer, M., Van der Linden P.J., Dai, X., Maskell, K., Johnson, C.A., 2001. 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 , 944 pp. Hutchins, D.A., Brunland, K.W., 1998. Iron-limited diatom growth and Si:N uptake ratios in a coastal upwelling regime. Nature 393, 561–564. Imbrie, J., Berger, A., Boyle, E.A., Clemens, S.C., Duffy, A., Howard, W.R., Kukla, G., Kutzbach, J., Martinson, D.G., McIntyre, A., Mix, A.C., Molfino, B., Morley, J.J., Peterson, L.C., Pisias N.G., Prell, W.L., Raymo, M.E., Shackleton, N.J., Toggweiler, J.R., 1993. On the structure and origin of major glaciation cycles 2. The 100 000-year cycle. Paleoceanography 8, 699–736. Jacobsen, S.B., Wasserburg, G.J., 1980. Sm-Nd isotopic evolution of chondrites. Earth and Planetary Science Letters 50, 139–155. Jouzel, J., Vaikmae, R., Petit, J.R., Martin, M., Duclos, Y., Stievenard, M., Lorius, C., Toots, M., Melie`res, M.A., Burckle, L.H., Barkov, N.I., Kotlyakov, V.M., 1995. The two-step shape and timing of the last deglaciation in Antarctica. Climate Dynamics 11, 151–161. Jouzel, J., Masson, V., Cattani, O., Falourd, S., Stievenard, M., Stenni, B., Longinelli, A., Johnsen, S.J., Steffensen, J.P., Petit, J.R., Schwander, J., Souchez, R., 2001. A new 27 kyr high resolution East

Antarctic climate record. Journal of Geophysical Research 28, 3199–3202. King, J.C., Turner, J., 1997. Antarctic meteorology and climatology. Cambridge University Press, Cambridge, UK, 425 pp. Kohfeld, K., Harrison, S.P., 2001. DIRTMAP: the geological record of dust. Earth Science Reviews 54, 81–114. Kukla, G., An, Z.S., Melice, J.L., Gavin, J., Xiao, J.L., 1990. Magnetic susceptibility record of Chinese loess. Transactions Royal Society of Edinburgh Earth Sciences 81, 263–288. Lamy, F., Hebbeln, D., Rohl, U., Wefer, G., 2001. Holocene rainfall variability in southern Chile: a marine record of Southern Westerlies. Earth and Planetary Science Letters 185, 369–382. Leventer, A., Domack, E.W., Ishman, S.E., Brachfeld, S., McClennen, C.E., Manley, P., 1996. Productivity cycles of 200–300 years in the Antarctic Peninsula region: understanding linkages among the Sun, atmosphere, ocean, sea ice and biota. GSA Bulletin 108, 1626–1644. Mahowald, N., Kohfeld, K., Hansson, M., Balkanski, Y., Harrison, S.P., Prentice, I.C., Schulz, M., Rodhe, H., 1999. Dust sources and deposition during the last glacial maximum and current climate: a comparison of model results with paleodata from ice cores and marine sediments. Journal of Geophysical Research 104, 15 895–15 916. Masson, V., Vimeux, F., Jouzel, J., Morgan, V., Delmotte, M., Cias, P., Hammer, C., Johnsen, S., Lipenkov, V.Y., Mosley-Thompson, E., Petit, J.R., Steig, E.J., Stievenard, M., Vaikmae, R., 2000. Holocene climate variability in Antarctica based on 11 ice core isotopic records. Quaternary Research 54, 348–358. Mayewski, P., Meeker, L.D., Twickler, M.S., Whitlow, S., Yang, Q., Lyons, W.B., Prentice, M., 1997. Major features and forcing of high-latitude northern hemisphere atmospheric circulation using a 110,000 year-long glaciochemical series. Journal of Geophysical Research 102, 26,345– 326,366. Moy, C.M., Seltzer, G.O., Rodbell, D.T., Anderson, D.M., 2002. Variability of El Nin˜o/Southern Oscillation activity at millennial timescales during the Holocene epoch. Nature 420, 162–165. Petit, J.R., Jouzel, J., Raynaud, D., Barkov, N.I., Barnola J.M., Basile, I., Bender, M., Chappellaz, J., Davis, M., Delaygue, G., Delmotte, M., Kotlyakov, V.M., Legrand, M., Lipenkov, V.Y., Lorius, C., Pe´pin, L., Ritz, C., Saltzman, E., Stievenard, M., 1999. Climate and atmospheric history of the past 420000 years from the Vostok ice core, Antarctica. Nature 399, 429–436. Prospero, J.M., Ginoux, P., Torres, O., Nicholson, S.E., Gill. T.E., 2002. Environmental characteristics of

Late Quaternary Interglacials in East Antarctica global sources of atmospheric soil dust derived from the NIMBUS-7 TOMS absorbing aerosol product. Reviews of Geophysics 40, 2–31. Rea, D.K., 1994. The paleoclimatic record provided by eolian deposition in the deep sea: the geologic history of wind. Reviews of Geophysics 32, 159–195. Revel, M., Sinko, J.A., Grousset, F.E., Biscaye, P.E., 1996. Sr and Nd isotopes as tracers of North Atlantic lithic particles: paleoclimatic implications. Paleoceanography 11, 95–113. Rogers, R.R., Yau, M.K., 1989. A short course in cloud physics. Pergamon Press, 290 pp. Ruth, U., Wagenbach, D., Steffensen, J.P., Biggler, M., 2003. Continuous record of microparticle concentration and size distribution in the central Greenland NGRIP ice core during the last glacial period. Journal of Geophysical Research 108, 4098–4110. Sakar, A., Ramesh, R., Somayajulu, B.L.K.., Agnihotri, R., Jull, A.J.T.., Burr, G.S., 2000. High resolution Holocene monsoon record from the eastern Arabian Sea. Earth and Planetary Science Letters 177, 209–218. Schulz, M., Balkanski, Y.J., Guelle, W., Dulac, F., 1998. Role of aerosol size distribution and source location in a three-dimensional simulation of a Sahara dust episode tested against satellite-derived optical thickness. Journal of Geophysical Research 103, 10 579–10 592. Schwerdtfeger, W., 1984. Weather and climate in the Antarctic. Elsevier Amsterdam, 261 pp. Smith, J., Vance, D., Kemp, R.A., Archer, C., Toms, P., King, M., Zarate, M., 2003. Isotopic constraints on the source of Argentinian loess – with implications for atmospheric circulation and the provenance of Antarctic dust during recent glacial maxima. Earth and Planetary Science Letters 212, 181–196. Sokolik, I.N., Toon, O.B., 1996. Direct radiative forcing by anthropogenic airborne mineral aerosols. Nature 381, 681–683.

73

Steffensen, J.P., 1997. The size distribution of microparticles from selected segments of the GRIP ice core representing different climatic periods. Journal of Geophysical Research 102(C12), 26 755– 26 763. Stuut, J.-B., Prins, M.A., Schneider, R.R., Weltje, G.J., Jansen, J.H.F.., Postma, G., 2002. A 300-kyr record of aridity and wind strength in southwestern Africa: inferences from grain size distributions of sediments on Walvis Ridge, SE Atlantic. Marine Geology 180, 221–233. Swap, R., Garstang, M., S.Greco, Talbot, R., Kallberg, P., 1992. Saharan dust in the Amazon Basin. Tellus B44(2), 133–149. Tegen, I., Harrison, S.P., Kohfeld, K., Prentice, I.C., Coe, M., Heimann, M. 2002. The impact of vegetation and preferential source areas on global dust aerosol: Results from a model study. Journal of Geophysical Research 107(D21), 4576–4603. Udisti, R., Becagli, S., Castellano, E., Delmonte, B., Jouzel, J., Petit, J.R., Schwander, J., Stenni, B., Wolff, E., 2004. Stratigraphic correlations between the European Project for ice coring in Antarctica (EPICA) Dome C and Vostok ice cores showing the relative variations of snow accumulation over the past 45 kyr. Journal of Geophysical Research 109, D08101 (doi:10.1029/2003JD004180). Watanabe, O., Jouzel, J., Johnsen, S., Parrenin, F., Shoji, H., Yoshida, N., 2003. Homogeneous climate variability across East Antarctica over the past three glacial cycles. Nature 422, 509–512. Yung, Y.L., Lee, T., Wang, C.H., Shieh, Y.T., 1996. Dust: a diagnostic of the hydrologic cycle during the last glacial maximum. Science 271, 962–963. Zhang, Y., Carmichael, G.R., 1999. The role of mineral aerosol in tropospheric chemistry in East Asia – a model study. Journal of Applied Meteorology 38, 353–366.

This page intentionally left blank

7. Eustatic Sea Level During Past Interglacials M. Siddall1, J. Chappell2 and E.-K. Potter3 1

Climate and Environmental Physics, Physics Institute, University of Bern, Switzerland Research School of Earth Sciences, The Australian National University, Canberra 0200, Australia 3 Institute for Isotope Geology and Mineral Resources, Department of Earth Sciences, ETH, Zurich, Switzerland 2

ABSTRACT Eustatic sea-level variation is the primary index of global ice volume over glacial cycles. Here, we review recent studies of eustatic sea level during interglacial periods. This review includes in-depth discussion of the variability, magnitude and duration of the last five interglacial periods and a summary of evidence for the last nine interglacial periods. The last nine interglacial periods differ not only in height and variability of sea level, but also in timing relative to northern summer insolation peaks. Some interglacials have a single peak and others have several, while MIS 11 persisted with little variation for at least 30 kyr. Estimates of interglacial sea levels remain subject to uncertainties. There is an outstanding need for glacio-hydro-isostatic modelling for all glacial cycles of interest, as well as improvements in dating and dating-correction techniques. 7.1 INTRODUCTION Global climate is currently in an interglacial state, and sea level has been relatively constant for the last 7000 years since melting of the northern continental ice sheets of the last ice age was completed. However, the exact nature of the ‘interglacial state’ is unclear, because interglacial sea levels are thought to have varied by the order of 10 m over the last 800 kyr (Pirazzoli et al., 1991), suggesting that the state of the climate and volume

of the icecaps have varied between and perhaps during interglacials. Sealevel may be used to define the duration and timing of interglacial periods (Stirling et al., 1998), and can show whether the volume of ice and stability of the climate varied from one interglacial to another (Chen et al., 1991; Neumann and Hearty, 1996). On Quaternary timescales, sea-level changes are attributable to eustatic and isostatic effects. Eustatic changes principally are global variations of sea level, arising from changes in the volume of water in the oceans, which is strongly affected by the size of the icecaps. Thermal expansion of water, in response to global climate changes, has a small effect on the volume of ocean water, amounting to only tens of centimetres of sea-level variation within or between interglacials (Wigley and Raper, 1987). Eustatic sea level is also affected by the geometry of the ocean basins which changes on timescales of millions of years and is associated with plate tectonics (Lambeck and Chappell, 2001). Isostatic effects are regional and relatively localised movements of the Earth’s crust caused by changing surface loads of ocean water (hydro-isostasy) and icecaps (glacioisostasy). As these movements are relatively slow, owing to the highly viscous nature of the Earth’s mantle, isostatic adjustment to the changes of ice and ocean water that marked the end of the last ice age is not yet complete. Because many observations of past sea level are made relative to local benchmarks which themselves are rising or

76

M. Siddall, J. Chappell and E.-K. Potter

falling isostatically, observations of postglacial sea-level rise show a highly variable spatial pattern. Changes in the gravitational equipotential level of the Earth due to redistribution of internal and surface (ice and water) masses during glacial cycles are also an important component of isostasy. Finally, spatial variation of the height of the ocean due to ocean circulation, known as the dynamic height of the ocean, affects sea level on local to global scales on the order of several (for example in the vicinity of the Gulf Stream) (Levitus and Boyer, 1994). Owing to its dependence on global ice volume, eustatic sea level is a primary index for comparing the conditions of one interglacial to another. However, to arrive at an estimate of eustatic sea-level from local observations of the relevant ancient shorelines, the global isostatic displacements need to be reconstructed, which in turn requires a record of sea-level and global ice distribution during the preceding glacial periods. The procedure for determining eustatic and isostatic contributions to sea level thus is iterative and calls into play proxy records of eustatic sea level based on oxygen isotope variations in deep-sea sediments, together with local observations based on dated shorelines and coral reefs as well as on reconstructions of past ice sheets (Lambeck and Chappell, 2001). In general, reconstruction of Holocene isostatic variations provides a partial guide to regional variations that may be expected in past interglacials. The present (Holocene) interglacial sea-level highstand has persisted for the last 7000 years (6200 radiocarbon years BP): setting aside regions close to former icecaps, together with tectonically active areas, global postglacial isostatic adjustments have caused shorelines aged 7000 years to vary in elevation from several metres above the present sea level to several metres below it. Thus, considering the eustatic implication of a shoreline of a given past interglacial, it is necessary to establish its location in time within that

interglacial as well as the local isostatic response. Moreover, when inferring eustatic changes by comparing shorelines of previous interglacials with Holocene shorelines, it must be recalled that there is some debate on whether eustatic variations have occurred within the last 7000 years. The occurrence of mid-Holocene shorelines up to several metres above the present sea level does not imply global eustatic fall or increasing ice volume: on the contrary, global isostatic modelling suggests that ice has melted from Antarctica in the last 7000 years, equivalent to about 3 m of sea level (see Lambeck and Chappell, 2001, and references therein). More debatable than the trends of global isostatic and eustatic changes in the last 7000 years are possible local higher-frequency fluctuations of sea level, which in some places have been claimed to range to several metres [e.g. southern Sweden (Mo¨rner, 1971) and the northwestern Black Sea (Chepalyga, 1984)] but have not been detected in detailed shoreline records elsewhere (e.g. the Great Barrier Reef: Chappell et al., 1983). This volume lies below the limit of resolution for proxy sea-level data such as oxygen isotopes and, if Holocene sea-level fluctuations did occur, they would be far less easy to resolve than variability of Holocene climate (Rohling et al., 2002). 7.2 EVIDENCE AND CHRONOLOGY Evidence of palaeo-sea levels comes in many forms, each with specific limitations. Two general classes exist. In the first, reconstructions make use of markers of sea-level position at or close to the coast, including erosional features such as wave cut notches (Hearty, 1998) or depositional features including sand-barrier shorelines (MurrayWallace, 2002), coral reefs (Szabo et al., 1994), speleothems (Bard et al., 2002) and other deposits that can only form above sea level. These methods usually generate discontinuous records of sea level that rely heavily on

Eustatic Sea Level During Past Interglacials

dating techniques (Gallup et al., 1994; Stirling et al., 1998). Some forms of evidence are stronger than others. For example, drill cores from coral reefs can provide semicontinuous series of precisely dated points (Edwards et al., 1993), but erosional features by their very nature rarely provide dateable material. In general, the interpretation of serial changes from individual dated points is strengthened by understanding the surrounding stratigraphy. Radiocarbon (14C) is widely used for dating sea-level deposits of the present interglacials, but previous interglacials lie beyond the 14C range. Amongst various techniques for dating past interglacial sea levels, U/Th methods applied to coral formations have yielded the most accurate results. At timescales around 100 kyr, the precision of a U/Th date typically is better than 1%, and reproducibility amongst a set of samples from the same site is almost as good (Stirling et al., 1998). To achieve this precision, due attention must be paid to sample quality and a large percentage [typically more than 50% (Stirling et al., 1998)] of corals in any given study may be rejected having failed tests for diagenetic effects. Recently several authors have made attempts to correct diagenetic effects in corals, but their results remain controversial (see Scholz and Mangini, this volume for a review of this topic). Note also that this precision may be optimistic since it does not take into account uncertainties in the spike calibration – a unique calibration standard is needed but is not currently available. Precision decreases as the age increases and typically is 7% at 700 kyr (Stirling et al., 2001). The uncertainty also increases with age for other, less precise methods that are used when U/Th dating is not applicable, such as electron spin resonance (ESR) (Schellmann and Radtke, 2004). Amino acid racemisation methods also are used when no other options are available but are only reliable to within one glacial cycle (Hearty, 1998; Murray-Wallace, 2002).

77

Where dated interglacial shorelines occur at sites that can be shown to have been tectonically stable on 106-year timescales, in principle it only remains to establish the isostatic corrections, to reduce the observations to estimates of eustatic sea levels. However, evidence of tectonic stability can be elusive and, when interglacial shorelines themselves are introduced as proof of stability, can become circular. Moreover, sealevel estimates from stable sites may be difficult to resolve because shorelines of different interglacials tend to be superimposed, if their eustatic levels were similar. This difficulty may be circumvented on tectonically rising coasts, where successive interglacial shorelines tend to be preserved as series of coastal terraces, but as the height of a given shoreline position depends on both its age and the uplift history, the accurate estimation of interglacial sea levels from such sites is no less problematic than elsewhere. Nonetheless, useful interglacial sea levels have been derived from welldated coral reef terraces at tectonically rising sites, such as Barbados (Schellmann and Radtke, 2004) and are included in our review. The second major class of sea-level reconstructions is based on the interpretation of varying oceanic oxygen isotope ratios recorded by calcareous foraminifera in marine sediments. The light isotope 16O is preferentially evaporated from the ocean and precipitated in snow (Dansgaard, 1984); in glacial periods, growth of the large continental ice sheets leads to increase of the 18 O/16O ratio in ocean water. The oxygen isotope ratio in foraminifera thus is sensitive to global ice volume, but also is affected by isotopic variability within the oceans and by temperature of the water in which the foraminifera live (e.g. Shackleton and Opdyke, 1973; Lea et al., 2002; Waelbroeck et al., 2002). If the mean isotopic composition of the ice sheets remained constant while they changed in size, and if the temperature variations were known and the water mass structure of the oceans also was constant,

78

M. Siddall, J. Chappell and E.-K. Potter

sea level could be accurately estimated from marine isotope records. In practice, the ice composition and ocean structure usually are assumed to be constant, and sea levels are estimated after subtracting the temperature effect (Shackleton, 1987). Oxygen isotope ratios in marine sediment cores provide continuous records, resolvable at millennial timescales, and are the basis of a standard marine sedimentary stratigraphy in which glacial cycles are resolved (Emiliani, 1955, 1966; Shackleton and Opdyke, 1973; Martinson et al., 1987; Shackleton et al., 1990). Each major peak and trough in this record is assigned to a numbered marine isotope stage (MIS), with interglacials generally given odd numbers and glacials given even numbers. The advent of higher-resolution isotope records has lead to the identification of substages, which are either decimalised (5.1, 5.2 etc.) or lettered (5a, 5b etc.) so that the first number refers to the glacial/interglacial cycle and the second symbol refers to the substage. Minima in the oxygen isotope ratios during these substages correspond to relatively high sea level and are known as interstadials. In many cases, the true interglacial, or sea-level highstand, is labelled in the same way as an interstadial: for example MIS 5e is the interglacial for MIS 5. A powerful chronologic tool has been established for dating marine records, based on slow variations of the Earth’s orbit, which control the seasonal distribution of solar radiation (insolation) and have been shown to be linked to icecap growth and retreat (Hays et al., 1976; Imbrie et al., 1984) (the theory relating climate to orbital variations is widely known as the Milankovitch theory, after its originator Milutin Milankovitch: see Imbrie et al., 1984). Based on highly precise calculations of planetary motions within the solar system, these variations have been precisely determined for the last several million years (Berger and Loutre, 1991). Making certain assumptions about the phasing of glacials, interglacials and summer insolation at 65 N, the marine

isotope record has been assigned a timescale defined by the orbital variations (Martinson et al., 1987; Shackleton et al., 1990). Orbitally based age estimates for the interglacial periods are more precise than U/Th ages for all interglacials except the last (MIS 5e) and possibly the previous one (MIS 7). However, as most interglacial sea-level deposits are discontinuous, these procedures do not resolve questions about their durations or relative heights. Importantly, there is apparent disagreement about the timing of MIS 5e based on U/Th dating of fossil coral reefs and orbitally based estimates of the timing of MIS 5e. As discussed later in the text, this may challenge the simplest model of orbitally forced glacial–interglacial sea-level variation. Many have contributed to the resolution of questions about past interglacial sea levels, with improvements in dating (e.g. Bard et al., 1990; Gallup et al., 1994; Stirling et al., 1995, 1998, 2001), stratigraphic control (e.g. Pirazzoli et al., 1991; Murray-Wallace, 2002) and oxygen isotope sea-level estimates (Shackleton, 2000; Lea et al., 2002; Waelbroeck et al., 2002; Siddall et al., 2003). We focus on the magnitude and timing of eustatic sea-level variability during the interglacial periods of the last 800 kyr. This chapter is not intended to be an exhaustive list of all of the work on interglacial sea level carried out over the past several decades, but concentrates on more recent work. 7.3 CONTINUOUS RECORDS OF SEA-LEVEL CHANGE Continuous records of sea-level variation can be used to compare sea-level variability to solar forcing and to identify which of the insolation maxima are associated with sealevel highstands. This provides a framework in which to examine reconstructions of each individual interglacial. Figure 7.1 shows sea-level curves based on six different techniques for deriving sea level from oxygen isotope ratios (Shackleton

Sea level (m)

5e

0

11

9

7

−50 −100 −150

Sea level (m)

50

100

11

0

150

200

250

300

15

13

350

400

Eustatic Sea Level During Past Interglacials

0

450

19

17

−50 −100 −150 350

400

450

500

550

600

650

700

750

800

Age (kyr BP) Murray-Wallace (2002) sea-level estimate

Pirazzoli et al. (1991) sea-level estimate

Lea et al. (2002)

Waelbroeck et al. (2002)

Siddall et al. (2003)

Labeyrie et al. (1987)

Shackleton (2000) Scaled benthic isotopes from V19-30, after Cutler et al. (2003)

Continued 79

80

M. Siddall, J. Chappell and E.-K. Potter

Fig. 7.1 Continuous sea-level estimates on the SPECMAP timescale from a variety of sources, together with other sea-level indicators, compared with the benthonic isotope record of core V19-30 (Shackleton and Pisias, 1985) from the east equatorial Pacific (scaled after Cutler et al., 2003). Sources: Lea et al. (2002) – planktonic temperatures derived from foraminiferal Mg/Ca were subtracted from planktonic oxygen isotopes to give the oxygen isotope ratio of water, assumed to be proportional to ice volume; Waelbroeck et al. (2002) – North Atlantic and equatorial Pacific benthonic oxygen isotope records were scaled to sea level using regression models for glacial and deglacial periods calibrated with fossil coral reef estimates of sea level; Shackleton (2000) – benthonic oxygen isotope records from the equatorial Pacific (V19-30), the Vostock air oxygen isotope ratio and assumptions about the Dole effect and deep-water temperatures were combined to generate a sea-level curve; Siddall et al. (2003) – the limited exchange of water across the strait between the Red Sea and the open ocean is strongly affected by sea-level variation, which dramatically changes the oxygen isotope ratio in the basin. Using a model, sea level is inferred from Red Sea oxygen isotope records; Labeyrie et al. (1987) – combined benthonic oxygen isotope records from the Norwegian Sea and equatorial Pacific (V19-30) are considered to be subject to minimal temperature fluctuations and respond primarily to ice volume. Dating of the Murray-Wallace (2002) and Pirazzoli et al. (1991) sea-level estimates is by correlation with the benthonic isotope record (V19-30). Marine isotope stages associated with interglacials are numbered. Boxed areas are shown in the other figures. Note that a 30 m envelope covers the variation in the scaled V19-30 benthonic isotope curve and other estimates.

and Pisias, 1985; Labeyrie et al., 1987; Shackleton et al., 1990; Shackleton, 2000; Lea et al., 2002; Waelbroeck et al., 2002; Cutler et al., 2003; Siddall et al., 2003). A brief summary describing the origins of each record is given in the figure caption, but a more complete discussion of each individual method for deriving sea level from oxygen isotope records is beyond the scope of this book chapter. Interested readers should refer directly to the references given for a more detailed discussion. A benthonic oxygen isotope record from the east equatorial Pacific (V19-30, Shackleton and Pisias, 1985) that extends to over 800 kyr is used as a reference with which to compare the other records and sea-level estimates. This has been rescaled after making adjustments for glacial–interglacial temperature shifts (Chappell and Shackleton, 1986; Shackleton, 1987; Cutler et al., 2003). All curves in Fig. 7.1 share similar, orbitally based chronologies, and it is to be expected that their peaks coincide. More importantly, the six records show similar amplitudes of glacial–interglacial sea-level changes and they broadly agree on the magnitude of variability between different isotope stages, despite differences in methods used to construct them: for example, there is greater variability during stage 3 and stages 5a–d

than in stage 9. Disagreements between the sea-level reconstructions and benthonic isotope records largely reflect temperature variations that probably are not fully accounted for in our rescaling of V19-30. Examined more closely, there are quite substantial variations amongst interglacial sea-level estimates in Fig. 7.1, exceeding 20 m. We will attempt to resolve this when examining each interglacial in turn, below, by using estimates based on ancient shorelines, which in the best cases are more accurate. Despite the uncertainties, the continuous records provide insights into complexities that remain unresolved, such as the occurrence of multiple peaks. For example, in some records MIS 7 appears to have two major highstands (Labeyrie et al., 1987; Lea et al., 2002; Siddall et al., 2003), but others show three (Shackleton, 2000; Waelbroeck et al., 2002). Similar variation appears in MIS 13 and 15 (Fig. 7.1). In summary, despite the difficulties in estimating eustatic sea level from marine oxygen isotope sequences, there are lessons to be gained from such studies: (1) Some interglacials apparently have multiple highstands of similar magnitude, which should be taken into account when assessing other sea-level information;

Eustatic Sea Level During Past Interglacials 20

10

Sea level (m)

(2) There are fewer data for earlier interglacials, but benthonic isotope records indicate that interglacial sea levels generally were lower before MIS 11 than they have been since; (3) Disagreements between oxygen isotopebased sea-level estimates highlight the importance of data from ancient shorelines.

0

−10

−20 110

7.4 SEA LEVEL DURING STAGE 5e

81

115

120

125

130

135

Age (kyr BP) Stirling et al. (1998) highstand estimate

The last interglacial (MIS 5e) has been the subject of considerable study, and its interpretation has evolved since the initial studies several decades ago (e.g. Stearns, 1976; Chappell and Veeh, 1978). Disagreements surrounding sea level during this interglacial might reflect complexity during this period or may be due to there being different interpretations from different sites. For example, for some authors the onset of MIS 5e is complicated by an apparent reversal in the rising sea-level trend during the MIS 6/5e deglaciation (Stein et al., 1993; Stirling et al., 1998; Esat et al., 1999; Gallup et al., 2002), while others consider that the interglacial itself was punctuated by significant sea-level changes (Aharon et al., 1980; Neumann and Hearty, 1996; Thompson and Goldstein, 2005). Stirling et al. (1998) present detailed estimates of the timing and duration of the MIS 5e highstand based on coral reef formations in Western Australia, which is a tectonically very stable region. The data comprise 70 U/Th ages (Fig. 7.2), derived from carefully screened samples of in situ unaltered corals which show no evidence of uptake or loss of uranium or thorium. This study found that MIS 5e lasted from 128 1 kyr BP to 116 1 kyr BP. The onset date of 128 1 kyr BP agrees closely with the oldest 5e corals sampled in growth position from raised reefs on Oahu (131 3 kyr BPÞ (Szabo et al., 1994), Sumba Island ð131 1 kyr BPÞ (Bard et al., 1996), the New Hebrides Arc ð129 1 kyr BPÞ

Oxygen isotope-based sea-level estimates (see caption for detail) Neumann and Hearty (1996) estimate of variability

Fig. 7.2 MIS 5e sea-level estimates as described in the text. The estimates of several authors converge on a highstand lasting from 128 1 kyr BP to 116 1 kyr BP (Chen et al., 1991; Stirling et al., 1998; Antonioli et al., 2004), with sealevels around þ 2 to þ 4 m (Hearty and Kindler, 1995; Stirling et al., 1998; Schellmann and Radtke, 2004). Sea-level variations suggested by some authors are also shown, including a mid interglacial dip, and a more controversial peak at the conclusion of the interglacial inferred by Hearty and Kindler (1995). Oxygen isotope-based estimates are as follows: Lea et al. (2002) (light blue); Waelbroeck et al. (2002) (purple); Siddall et al. (2003) (red); Shackleton (2000) (green); Labeyrie et al. (1987) (dark blue); V19-30, scaled after Cutler et al. (2003) (black). A summary of the origins of the oxygen isotope-based estimates is given the in the Fig. 7.1 caption.

(Edwards et al., 1986, 1987) and the Bahamas (132 to 129 kyr BP) (Chen et al., 1991). We note that this is several thousand years earlier than predicted by the theory of orbitally driven climate change and suggests that the Earth system may include processes not fully accounted for by the orbital theory (Henderson and Slowey, 2000). Finally, the termination date of MIS 5e of 116 1 kyr BP (Stirling et al., 1998) agrees with results from Barbados raised reefs (117 to 120 kyr BP: Gallup et al., 1994), but is later than the estimate from Bahamas raised reefs (120 to 123 kyr BP: Chen et al., 1991) (see Table 7.1 for summary).

82

M. Siddall, J. Chappell and E.-K. Potter

Table 7.1 Summary of dates, sea level and variability for previous interglacial periods as discussed in text Age MIS 5e, SPECMAP age 122 ka BP 129 1 ka BP – U/ Th dates from a raised coral reef on the New Hebrides Arc (Edwards et al., 1986, 1987). 132 to 129 ka BP onset – from a raised coral reef on the Bahamas (Chen et al., 1991). 131 3 ka BP onset – U/Th dates from a raised coral reef on Oahu (Szabo et al., 1994). 131 3 ka BP onset – U/Th dates from a raised coral reef on Oahu (Szabo et al., 1994). 131 1 ka BP onset – U/Th dates from a raised coral reef on Sumba Island (Bard et al., 1996). 120–123 ka BP termination – from a raised coral reef on the Bahamas (Chen et al., 1991). 117 to 120 ka BP termination – from a raised coral reef on Barbados (Gallup et al., 1994). 117 to 120 ka BP termination – from a raised coral reef on Barbados (Gallup et al., 1994). 128 1 ka BP to 116 1 ka BP duration – U/Th dates on fossil corals from a tectonically stable part of Western Australia (Stirling et al., 1998) MIS 7a, SPECMAP age 194 ka BP 201 1:2 to 193:5 2:8 ka BP – U/ Th-dated corals from a raised coral reef on Barbados (Gallup et al., 1994). 201 2 to 190 2 ka BP – U/Thdated speleothems, Argenterola Cave (Bard et al., 2002; Antonioli et al., 2004). MIS 7c, SPECMAP age 216 ka BP 212 to 220 ka – U/Th-dated growth hiatuses in the Bahama flowstone (Li et al., 1989). MIS 7e, SPECMAP age 238 ka BP 230 to 235 ka – U/Th-dated growth hiatuses in the Bahama flowstone (Li et al., 1989).

Sea level

Evidence of variability

þ2 to þ4 m, fossil corals from a tectonically stable part of Western Australia (Stirling et al., 1998). þ0 to þ6 m, re-evaluation of Barbados uplift rates and fossil coral reef evidence (Schellmann and Radtke, 2004).

Mid-interglacial sea-level dip soon after 121 ka BP lasting approximately 2000 years (Aharon et al., 1980; Chen et al., 1991; Szabo et al., 1994; Stirling et al., 1995, 1998; Thompson and Goldstein, 2005). Spike in sea level at the end of interglacial (Hearty and Kindler, 1995).

5 to 15 m, mixed evidence with some disagreement – see text for detail (Li et al., 1989; Lundberg and Ford 1994; Toscano and Lundberg, 1999; Antonioli et al., 2004).

Same as for MIS 7a.

Same as for MIS 7a.

Eustatic Sea Level During Past Interglacials

83

Table 7.1 Continued Age MIS 9c, SPECMAP age 331 ka BP 324 3 to 318 3 ka BP, U/Th dated corals from a raised terrace at Henderson Island (Pitcairn Group, subequatorial south Pacific) (Stirling et al., 2001).

MIS 11 Duration of 30 to 40 ka – analysis of isotope records from north Atlantic ODP cores 980 and 983 (McManus et al., 2003). Two ESR-dated uplifted Barbados reefs at 398 and 410 ka BP (Schellmann and Radtke, 2004).

Sea level

Evidence of variability

þ4 m – erosional features, Bermuda and the Bahamas (Hearty and Kindler, 1995). 3 m – erosional features, Grand Cayman (3 to þ0:5 m: Ve´zina et al., 1999). 1 m – sedimentological features, Coorong Coastal plain, Australia (Murray-Wallace, 2002). 3 and þ8 m – uplifted coral reefs in the Bahamas (Schellmann and Radtke, 2004). þ4 m – erosional features, Bermuda and the Bahamas (Hearty and Kindler, 1995). 3 m – erosional features, Grand Cayman (3 to þ0:5 m: Ve´zina et al., 1999). 1 m – sedimentological features, Coorong Coastal plain, Australia (Murray-Wallace, 2002). One reef at 2 m and þ11 m and one reef at þ5 m to þ18 m – uplifted coral reefs in the Bahamas (Schellman and Radtke, 2004).

Sea-level spike of 20 m at the end of the interglacial on the basis of perched beach deposits at 22 m on Eleuthera Island in the Bahamas (Hearty, 1998). Stable MIS 11 sea level within 10 m of present – analysis of isotope records from north Atlantic ODP cores 980 and 983 (McManus et al., 2003).

MIS 13, 15, 17, 19 Generally below present – evaluation of benthic oxygen isotope records (Shackleton, 1987). 0 m, 0 m, 20 m and 15 m, respectively, within 10 m Cape Laundi, Sumba Island (Pirazzoli et al., 1991). Sea level similar to the present level – Coorong Coastal Plain, south Australia (Murray-Wallace, 2002).

Most of the dated coral sites mentioned above lie in the isostatic far field, remote from former ice sheets, but all of them are tectonically uplifted except for the Western Australia and Bahamas sites. Holocene reefs at the Western Australia sites are at the present sea level, indicating that isostatic corrections are minor, and we consider that

global eustatic sea level for MIS 5e to generally lie within the range of þ2 to þ4 m reported by Stirling et al. (1998). This is in agreement with a recent re-analysis of Barbados uplift rates which placed the MIS 5e interglacial at between 0 m and þ6 m based on evidence from raised coral reefs (Schellmann and Radtke, 2004). Whether

84

M. Siddall, J. Chappell and E.-K. Potter

sea level fluctuated within MIS 5e is less clear. Variations are not ruled out by the oxygen isotope-based records (Fig. 7.2), and discontinuities within Bahaman MIS 5e reefs may represent a fall of sea level followed by a rise (Aharon et al., 1980; Chen et al., 1991). A comparable interruption occurs at other intermediate and farfield sites, soon after 121 kyr BP, and may reflect the effects of either a cold episode or sea-level fluctuation (Stirling et al., 1995, 1998). However, the duration of this episode is unlikely to have been more than 2000 years, according to U/Th-series ages of MIS 5e corals on Oahu reported by Szabo et al. (1994). The existence of such a brief dip in sea level during MIS 5e is supported by recent work by Thompson and Goldstein (2005) who attempted to correct U/Th age estimates on corals which had undergone diagenetic alteration. Thompson and Goldstein (2005) also suggest substantial sealevel variation at the start of the interglacial. However, controversy surrounds the correction procedure applied by Thompson and Goldstein (2005) (Scholz and Mangini, this volume), and until these issues are dealt with, there will be considerable ambiguity as to the existence of sea-level variation within and at the start of MIS 5e. In addition to a possible sea-level lowering around 121 kyr, some authors infer that the interglacial ended with a sharp sea-level rise from þ6 to þ8:5 m, inferred from wave cut notches, rubble benches and other shoreline sediments (Neumann and Hearty, 1996; Hearty and Kindler, 1995; Hearty and Neumann, 2001) (Fig. 7.2). The highstand is inferred to have been short-lived – perhaps no more than 600 years (Neumann and Hearty, 1996; Hearty and Neumann, 2001). The 6 m rise is similar in magnitude to a collapse of the West Antarctic ice sheet and, if it occurred, the influx of fresh water would disturb the thermohaline circulation and affect global climate (Stocker and Wright, 1991). As a collapse of Antarctic ice could be a possible outcome of present global warming (Stocker et al., 2001), it is important that the

claimed sudden rise at the end of the last interglacial be further investigated (see Table 7.1 for summary). 7.5 INTERGLACIAL SEA LEVELS IN MIS 7, 9 AND 11 As the last interglacial highstand was above modern sea level, its deposits are readily accessible. In comparison, the task of defining sea levels for several of the previous interglacials is more difficult on tectonically stable coasts, as deposition and erosion during later highstands have masked the traces of earlier highstands. The evidence tends to be better preserved on tectonically rising coasts, which often show coral and sedimentary terraces that can be matched with interglacials with great consistency (Pirazzoli et al., 1991; Murray-Wallace, 2002). At such sites, the calculation of relative sea level for each highstand relies heavily on accurate determination of uplift rate and shoreline ages, as well as on the assumption of a constant uplift rate. The uplift rate (RU) in most studies has been estimated from the height H5e of the MIS 5e shoreline at a given site, calculated as RU ¼ ðH5e S5e Þ=t5e where S5e is the sea level in MIS 5e at time t5e. Different studies adopt different figures for S5e and t5e but few attempts to include isostatic corrections: extrapolation of the resulting uncertainties in RU produces substantial uncertainties in interglacial sea-level estimates. For example, sea levels estimated for MIS 9 and MIS 11 from coral terraces at Barbados are between 8 and 18 m if the uplift rate is based on a þ6 m sea level for MIS 5e, but is close to the present sea level if a value of þ2 m is used (Schellmann and Radtke, 2004). With these cautions in mind, we now review the observations from MIS 7, 9 and 11. 7.5.1 MIS 7 Oxygen isotope records of sufficiently high resolution show that MIS 7 may include

Eustatic Sea Level During Past Interglacials

85

10

Sea level (m)

0

↓

↓

↓

↓

−10

↓

↑

↓

↑

−20

−30 180

190

200

210

220

230

240

Age (kyr BP) Gallup et al. (1994) highstand estimate (Barbados corals) Li et al. (1989) Bahamas speleothem growth periods [corrected after Toscano and Lundberg (1999)] Murray-Wallace (2002) sea-level estimate (Coorong coastal plain) Hearty and Kindler (1995) sea-level estimate (Bahamas geological indicators) Schellmann and Radtke (2004) MIS 9c sea-level estimate (Barbados coral) Gallup et al. (1994) sea-level upper estimate (Barbados coral) Bard et al. (2002) marine layers on speleothem (Argenterola Cave) Oxygen isotope-based sea-level estimates (see caption for detail)

Fig. 7.3 MIS 7 sea-level estimates as described in the text. Despite the complexity of this interglacial, there is agreement amongst several authors on the timing of MIS 7a at 193 – 201 kyr (Li et al., 1989; Gallup et al., 1994; Bard et al., 2002; Antonioli et al., 2004). Estimates based on raised shorelines and corals (Hearty and Kindler, 1995; Murray-Wallace, 2002; Schellmann and Radtke, 2004) assigned to the interstadials by their authors using Amino Acid Racemisation, ESR and stratigraphic methods. Arrows indicate upper or lower limits. Uplift rate uncertainty affects the Bahamas speleothems (between 0 and 0.02 m kyr-1). Where known, age uncertainty is shown by grey bands or extended markers. Oxygen isotope-based estimates are as follows: Lea et al. (2002) (light blue); Waelbroeck et al. (2002) (purple); Siddall et al. (2003) (red); Shackleton (2000) (green); V19-30, scaled after Cutler et al. (2003) (black). A summary of the origins of the oxygen isotope-based estimates is given the in the Fig. 7.1 caption. Note that the Murray-Wallace (2002) and Hearty and Kindler (1995) sea-level estimates for MIS 7e are both similar, and the respective symbols are plotted over each other.

up to three sea-level peaks separated by episodes of substantially lower sea level (Fig. 7.3). In order of increasing age, these have been referred to as MIS 7a, 7c and 7e or, alternatively, as MIS 7.1, 7.3 and 7.5 (Martinson et al., 1987). Ideally, the sea-level uncertainties could be reduced using evidence from MIS 7 shoreline deposits with well-resolved stratigraphy, assisted by accurate U/Th dating. This has not been achieved for all three peaks at a single site, and the MIS 7 sea level can only be reconstructed in a piecemeal manner, at present. The youngest peak shown in Fig. 7.3 (MIS 7a) extends from 201 1:2 to 193:5 2:8 kyr BP, according to dated coral from a raised reef at Barbados (Gallup et al.,

1994). Support comes from U/Th dates from a growth hiatus in submerged speleothems from 18.5 m below present sea level in Argenterola Cave, Italy, which indicate that the cave was drowned after 201 2 kyr BP and re-emerged before 190 2 kyr BP (Bard et al., 2002; Antonioli et al., 2004). U/Thdated growth hiatuses in the Bahamas flowstone place the age of MIS 7c (MIS 7.3) between 212 and 220 kyr, and 7e (MIS 7.5) is between 230 and 235 kyr (Li et al., 1989). Results from Argenterola speleothems are compatible with these figures (Antonioli et al., 2004) (see Table 7.1 for summary). Sea-level estimates for MIS 7a range from 6 to þ9 m at Barbados, which is subject to uncertainties in the uplift correction (Gallup

86

M. Siddall, J. Chappell and E.-K. Potter

et al., 1994; Schellmann and Radtke, 2004). Similar uncertainties appear in MIS 7a sea levels from other raised coral reefs, estimated by Pirazzoli et al. (1991) to be 6 10 m. An upper limit can be estimated from a U/Th-dated flowstone found in a Bahamas cave, which shows no hiatus in growth between 209 and 139 kyr and has an elevation of 10 m. Note that the original references assume that the speleothem grew on the floor of the cave and give its depth as –15 m (Li et al., 1989; Lundberg and Ford, 1994). However, this speleothem was not found in growth position but was lying on the floor of the cave (Toscano and Lundberg, 1999) and is presumed to have fallen from the roof, or a shelf close to the roof. Toscano and Lundberg (1999) take the roof of the cave ð10 mÞ as the speleothem growth position. Allowing for a local subsidence rate of up to 1 m per 50 kyr relative sea level (Carew and Mylroie, 1995), the upper limit of sea level in MIS 7a relative to this site is 6 m but could be lower (Li et al., 1989; Lundberg and Ford, 1994; Toscano and Lundberg, 1999). Collectively, these results put MIS 7a sea level close to 6 m in the Caribbean region, but differences arising from isostatic movements are to be expected. In the Mediterranean, however, speleothem evidence suggests that sea level was as low as 18 m at Argentarola Cave during MIS 7a (Antonioli et al., 2004). We note that MIS 7a is the last of three peaks in MIS 7: by comparison, MIS 5a is the last of three peaks in stage MIS 5, and there is 20 m isostatic difference between MIS 5a relative sea levels in the northern and southern Caribbean (Potter and Lambeck, 2003). Until this factor has been determined for the substages of MIS 7, a eustatic sea level for MIS 7a cannot be finalised but probably lies between 5 and 15 m. Similar considerations apply to the previous highstands in MIS 7 (MIS 7c and MIS 7e). Sea level at the Bahamas during these hiatuses must have been higher than 10 m (without uplift correction) or higher than 6 m (correcting for uplift as before). On

the basis of ancient coastal deposits at the Bahamas, Hearty and Kindler (1995) consider that sea level was within 5 m of present in one or more stages of MIS 7. As this is compatible with the flowstone data, we conclude that Bahaman sea level was above 10 m in both MIS 7c and 7e and may have been within 5 m of present in at least one of these highstands. Estimates of MIS 7 sea levels elsewhere, although subject to comparable uncertainties, do not conflict with these results. Stratigraphically wellresolved formations in South Australia, where very slow regional uplift is well constrained, contain shoreline deposits representing two sea-level peaks, which have been assigned to MIS 7a (East Dairy Formation: sealevel 6 m) and MIS 7e (Reedy Creek Formation: sealevel 0 m) (MurrayWallace, 2002). The correlations are subject to dating uncertainties ranging from 11 to 25 kyr, but the sea-level estimates will not be greatly affected if these were changed by future work (for example, if the East Dairy Formation proves to be MIS 7c, the sea-level estimate would shift from 6 m to 8 m) (see Table 7.1 for summary).

7.5.2 MIS 9 Oxygen isotope records for MIS 9 show a single, dominant peak at 331 kyr (MIS 9c), followed by a secondary peak at 310 (MIS 9a) (Fig. 7.4). U/Th ages from corals in a raised MIS 9 reef at Henderson Island (Pitcairn Group, subequatorial south Pacific) suggest that this stage lasted from 334 4 to 306 4 kyr BP and that the MIS 9c peak began 324 3 and ended 318 3 kyr BP, which is closely in accordance with age predicted by orbital forcing (Stirling et al., 2001). Sea level during MIS 9c appears to have been close to the present level, according to estimates from South Australia (1 m; Murray-Wallace, 2002), Bermuda and the Bahamas (þ4 m; Hearty and Kindler, 1995) and Grand Cayman (3 to þ0:5 m; Ve´zina et al., 1999). Although less closely dated than

Eustatic Sea Level During Past Interglacials

the other results. As with MIS 7, further studies together with global isostatic modelling are required, before a global eustatic figure for the MIS 9c interglacial is established (see Table 7.1 for summary).

10

0

Sea level (m)

87

−10

−20

7.5.3 MIS 11

−30

−40 300

310

320

330

340

Age (kyr BP) Stirling et al. (2001) MIS 9c highstand duration Stirling et al. (2001) MIS 9 highstand duration Murray-Wallace (2002) sea-level estimate Hearty and Kindler (1995) sea-level estimate Schellmann and Radtke (2004) MIS 9c sea-level estimate Oxygen isotope-based sea-level estimates (see caption for detail)

Fig. 7.4 MIS 9 sea-level estimates as described in the text. U / Th dates from Henderson Island (Stirling et al., 2001) constrain the timing but not the sealevel during MIS 9, and results are shown as vertical grey bars: pale grey for the entire stage and darker grey for the MIS 9c highstand, which lags the orbitally tuned age for the highstand derived from oxygen isotopes by around 5 kyr. Estimates from raised shorelines (Hearty and Kindler 1995; Murray-Wallace 2002; Schellmann and Radtke 2004) are assigned to the MIS 9 by the authors using amino acid racemisation, ESR and stratigraphic methods. Oxygen isotope-based estimates are as follows: Lea et al. (2002) (light blue); Waelbroeck et al. (2002) (purple); Siddall et al. (2003) (red); Shackleton (2000) (green); V19-30, scaled after Cutler et al. (2003) (black). A summary of the origins of the oxygen isotope-based estimates is given the in the Fig. 7.1 caption.

MIS 9c at Henderson Island, uplift or subsidence rates at these sites are low and are constrained by other data (we note that Stirling et al., 2001, do not give an estimate of MIS 9c sealevel, because the uplift rate at Henderson Island has not been determined). Estimates from more rapidly rising sites are less well constrained, owing to uncertainties in both dating and uplift rate, but recent work at Barbados indicates that MIS 9c sealevel was between 3 and þ8 m (Schellmann and Radtke, 2004), which is consistent with

Beyond MIS 9, the uncertainties in radiometric dates from ancient shorelines are relatively large, and diagenetic effects are likely to be a major issue in older samples. Although it is likely that this will be improved by ongoing technical developments (Andersen et al., 2004) and interglacial dating at present rests on the orbital chronology. Importantly, new methods for correcting diagenetic effects to allow U/Th dating of older corals may result in larger age uncertainties than standard U/Th dating on unaltered younger corals (see Scholz and Mangini, this volume, for a review of this topic), thereby restricting the usefulness of U/Th techniques to the last few glacial cycles. A number of oxygen isotope records suggest that high sealevels of the MIS 11 interglacial commenced at 415 kyr and persisted to at least 395 kyr, although the resolution is not high and the degree of variation within this interval has been uncertain (Fig. 7.5). Recently, using very detailed isotope records from north Atlantic ODP cores 980 and 983 ( 200-year resolution), McManus et al. (2003) found that the interglacial persisted for 30 to 40 kyr, depending on the age model used for the cores, while the isotope ratios show little variation throughout. MIS 11 sealevels close to the present level have been inferred from shoreline sites with independent estimates of uplift or subsidence, including South Australia (3 m; Murray-Wallace, 2002), Bermuda and the Bahamas (þ4 m; Hearty and Kindler, 1995) and Grand Cayman (9 to 5:5 m; Ve´zina et al., 1999). Two fossil coral reef terraces in Barbados have been assigned to MIS 11 on the basis of ESR dating; these represent sea levels of between 2 and þ11 m and þ5 to þ18 m

88

M. Siddall, J. Chappell and E.-K. Potter

summary). The possibility of such a sharp sea-level rise at the conclusion of MIS 11 needs further investigation.

20

Sea level (m)

10

0 −10

7.6 INTERGLACIAL SEA LEVEL PREVIOUS TO MIS 11

−20 −30 −40 380

390

400

410

420

430

Age (kyr BP) McManus et al. (2001) sea-level estimates Murray-Wallace (2002) sea-level estimate Hearty and Kindler (1995) sea-level estimate Schellmann and Radtke (2004) MIS 11 sea-level estimates Oxygen isotope-based sea-level estimates (see caption for detail)

Fig. 7.5 MIS 11 sea-level estimates as described in the text. High-resolution benthonic isotope records show little variation through MIS 11, which lasted at least 30 000 years (McManus et al. 2003). Sealevel appears to have been within 10 m of the present, according to estimates from raised shorelines and corals (Hearty and Kindler 1995; Murray-Wallace 2002; Schellmann and Radtke 2004). Oxygen isotope-based estimates are as follows: Waelbroeck et al. (2002) (purple); Siddall et al. (2003) (red); Shackleton (2000) (green); V19-30, scaled after Cutler et al. (2003) (black). A summary of the origins of the oxygen isotope-based estimates is given the in the Fig. 7.1 caption.

(Schellmann and Radtke, 2004). A further terrace on Barbados might be of MIS 11 origin but equally may have been formed during MIS 9 (Schellmann and Radtke, 2004). In addition, on the basis of perched beach deposits at 22 m on Eleuthera Island in the Bahamas, Hearty (1998) has inferred that MIS 11 ended with a sharp sea-level rise of 20 m, followed by rapid decline into the glacial of MIS 10. No hint of this sea-level spike has so far been discovered in high-resolution isotope records such as those reported by McManus et al. (2003) which imply that sea level has been within 10 m of the present and effectively stable through this long interglacial (see Table 7.1 for

There is little definitive evidence for the position of sealevel during interglacials prior to MIS 11. Reviewing the oxygen isotope data from planktonic and benthonic foraminifera in a number of marine sediment cores, Shackleton (1987) concluded that sealevels during MIS 7, 13, 15, 17 and 19 may not have reached the present level. In contrast, Murray-Wallace (2002) concludes that ancient shorelines in South Australia indicate that sealevels during the MIS 13, 15, 17 and 19 interglacials may have been close to the present level, depending on whether the slow tectonic uplift of this region has been uniform. Somewhat different conclusions for MIS 17 and 19 were reached by Pirazzoli et al. (1991), on the basis of raised coral reefs at Cape Laundi, Sumba Island. Making the assumption that sealevels during MIS 5, 9, 11 and 25 were similar to today (within 5 m) these authors ‘tuned’ the local uplift rate and concluded that the sealevels of MIS 13, 15, 17 and 19 were 0, 0, 20 and 15 m, respectively, within 10 m. Uncertainties in the isotopic data are such that these issues cannot be resolved, but generally support Shackleton’s (1987) view that sealevels prior to MIS 11 were lower than today, at least during MIS 17 and 19 (Fig. 7.1) (see Table 7.1 for summary). 7.7 DISCUSSION As the summaries presented here show, the last nine interglacials differ in not only in height and variability of sealevel, but also in timing relative to northern summer insolation peaks. Independent dating shows that peak sea levels in MIS 1, 7 and 9 followed insolation maxima, as the theory of orbital

Eustatic Sea Level During Past Interglacials

forcing predicts, whereas the onset of high sea level in MIS 5e coincides with an insolation maximum. Moreover, some interglacials have a single peak of less than 10-kyr duration (e.g. MIS 9) and others have several (MIS 7), while MIS 11 persisted with little variation for at least 30 kyr (see Table 7.1 for summary). An obvious consequence of the early onset of MIS 5e is that it calls into question the assumptions behind a purely orbital timing of the interglacials. Instead, factors other than northern hemisphere summer insolation forcing are implicated in the MIS 6/5e deglaciation. Henderson and Slowey (2000) review possible alternative mechanisms which may have driven the early onset of MIS 5e. These may include summer insolation in the Southern Hemisphere, which peaked at 138 kyr BP and might have driven an increase of atmospheric CO2 that in turn enhanced the glacial termination (Broecker and Henderson, 1998; Henderson and Slowey, 2000). Other factors may include the behaviour of southern Ocean sea ice (Kim et al., 1998) or the effect of tropical insolation, which can influence the average number of El Nin˜o/Southern Oscillation (ENSO) warm events and may promote early deglaciation (Clement et al., 1999). Doubtless, these possibilities will eventually be explored in climate models which include insolation forcing as well as the alternative means to alter ice volume mentioned here – ultimately we will require deterministic approaches in order to understand the mechanisms of climate change at orbital timescale. We note that there may be a link between the relative magnitudes of MIS 5e and MIS 7 highstands and the phasing of the high stands with respect to orbital forcing. The contrasted phasing of MIS 5e and MIS 7.1, relative to northern summer insolation maxima, led Bard et al. (2002) to write: ‘A tempting conclusion is that second order sea-level highstands such as MIS 7.1 may be purely driven by astronomical changes, in contrast with major terminations such as Termination II. Atmospheric CO2 levels may partly explain this contrast since the two sea-level

89

transitions (MIS 7.2–7.1 and MIS 6–5.5) are characterised by very different CO2 rises (20 vs 100 ppm)’. Finally, although dating techniques have improved considerably in the last decade, estimates of interglacial sealevels remain subject to uncertainties. Some of these issues will be helped by improved models of diagenetic effects to help correct altered corals. Further work will also help bring into existence better controls on spike calibration and half lives used in U/Th dating. Debate also surrounds the magnitude and variability of the high stands. Whether MIS 5e and MIS 11 terminated with sharp sea-level spikes is controversial; variations amongst sea-level estimates for each interglacial prior to MIS 5e are considerable, and there is conflict amongst sea-level proxies, particularly for MIS 7. There is an outstanding need for glacio-hydro-isostatic modelling for all glacial cycles of interest, which even if driven by an isotope curve in lieu of sea level would provide useful estimates of regional sea-level differences for each interglacial substage. Lastly, further work at expanded sequences of ancient shorelines with slow, uniform tectonics, such as those discussed by MurrayWallace (2002), should be invaluable. ACKNOWLEDGEMENTS Thanks are due to Laurent Labeyrie, Claire Waelbroeck, David Lea and David Richards who provided data and useful advice. M. Siddall and E.-K. Potter are supported by the STOPFEN European Network Research Project (HPRN-CT-2002–00221). REFERENCES Aharon, P., Chappell, J., Compston, W., 1980. Stable isotope and sealevel data from New Guinea supports Antarctic ice-surge theory of ice ages. Nature 283, 649–651. Andersen, M.B., Stirling, C.H., Potter, E.-K., Halliday, A.N., 2004. Toward epsilon levels of measurement precision on 234U/238U by using

90

M. Siddall, J. Chappell and E.-K. Potter

MC-ICPMS. International Journal of Mass Spectrometry 237, 106–117. Antonioli, F., Bard, E., Potter, E.-K., Silenzi, S., Improta, S., 2004. 215-ka History of sealevel oscillations from marine and continental layers in Argenterola Cave speleothems (Italy). Global and Planetary Change 43(1–2), 57–78. Bard, E., Hamelin, B., Fairbanks, R.G., Zindler, A., Arnold, M., Mathieu, G., 1990. U/Th and 14C ages of corals from Barbados and their use for calibrating the 14C timescale beyond 9000 years BP. Nuclear Instruments and Methods B 52, 461–468. Bard, E., Jouannic, C., Hamelin, B., Pirazzoli, P., Arnold, M., Faure, G., Sumosusastro, P., Syaefudin, 1996. Pleistocene sea levels and tectonic uplift based on dating of corals from Sumba Island, Indonesia. Geophysical Research Letters 23, 1473–1476. Bard, E., Antonioli, F., Silenzi S., 2002. Sealevel during the penultimate interglacial period based on a submerged stalagmite from Argenterola Cave, Italy. Earth and Planetary Science Letters 196, 135–146. Berger, A., Loutre, M.F., 1991. Insolation values for the climate of the last 1 0000 000 years. Quaternary Science Reviews 10(4), 297–317. Broecker, W.S., Henderson, G.M., 1998. The sequence of events surrounding Termination II and their implications for, the cause of glacial–interglacial CO2 changes. Paleoceanography 13, 352–364. Carew, J.L., Mylroie, J.E., 1995. Quaternary tectonic stability of the Bahamian Archipelago – evidence from fossil coral-reefs and flank Margin Caves. Quaternary Science Reviews 14(2), 145–153. Chappell J., 2002. Sea level changes forced ice breakouts in the last glacial cycle: new results from coral terraces. Quaternary Science Reviews 21, 1229–1240. Chappell, J., Veeh, H.H., 1978. Late Quaternary tectonic movements and sealevel changes at Timor and Atauro Island. Geological Society of America Bulletin 89(3), 356–368. Chappell, J., Shackleton, N.J., 1986. Oxygen isotopes and sealevel. Nature 324, 137–140. Chappell, J., Chivas, A., Wallensky, E., Polach, H.A., Aharon, P., 1983. Holocene palaeo-environmental changes, central to north Great Barrier Reef inner zone, BMR. Journal of Australian Geology and Geophysics 8, 223–235. Chen, J.H., Curran, H.A., White, B., Wasserburg, G.J., 1991. Precise chronology of the last interglacial period: 234U/230Th data from fossil coral reefs in the Bahamas. Geological Society of America Bulletin 103, 82–97. Chepalyga, A.L., 1984. Inland Sea Basins, In: Velichko, A.A. (Ed.), Late Quaternary Environments of the Soviet Union. University of Minnesota Press, Minneapolis, pp. 229–250.

Clement, A., Seager, R., Cane, M., 1999. Orbital controls on ENSO and the tropical climate. Paleoceanography 14, 441–456. Cutler, K.B., Edwards, R.L., Taylor, F.W., Cheng, H., Adkins, J., Gallup, C.D., Cutler, P.M., Burr, G.S., Bloom, A.L., 2003. Rapid sealevel fall and deepocean temperature change since the last interglacial period. Earth and Planetary Science Letters 206, 253–271. Dansgaard, W., Johnsen S.J., Clausen H.B., DahlJensen D., Gundestrup N., Hammer C.U., Oeschger H., 1984. North Atlantic climatic oscillations revealed by deep Greenland ice cores. Geophysical Monograph Series 29, 288–298. Edwards, R.L., Chen, J.H., Wasserburg, G.J., 1986. 238 U-234U-230Th-232Th systematics and the precise measurement of time over the past 500 000 years. Earth and Planetary Science Letters 81, 175–192. Edwards R.L., Chen J.H., Ku T.-L., Wasserburg G.J., 1987. Precise timing of the last interglacial period from mass spectrometric determination of Thorium-230 in corals. Science 236, 1547–1553. Edwards, R.L., Beck, J.W., Burr, G.S., Donahue, D.J., Chappell, J., Bloom, A.L., Druffel, E.R.M., Taylor, F.W., 1993. A large drop in atmospheric 14C/12C and reduced melting in the Younger Dryas, documented with 230Th ages of corals. Science 260, 962–968. Emiliani, C., 1955. Pleistocene temperatures, Journal of Geology 63, 585–599. Emiliani, C., 1966. Paleotemperature analysis of Caribbean cores P 6304–8 and P 6304–9 and a generalised curve for the last 425,000 years. Journal of Geology 74, 109–126. Esat, T.M., McCulloch, M.T., Chappell, J., Pillans, B., Omura, A., 1999. Rapid fluctuations in sealevel at Huon Peninsula during the penultimate deglaciation. Science 283, 197–201. Gallup, C.D., Edwards, R.L., Johnson, R.G., 1994. The timing of high sealevels over the past 200,000 years. Science 263, 796–800. Gallup, C.D., Cheng, H., Edwards, R.L., Taylor, F.W., 2002. Direct determination of the timing of sealevel change during termination II. Science 295, 310–313. Hayes, J.D., Imbrie, J., Shackleton, N.J., 1976. Variations in the Earth’s orbit: Pacemaker of the ice ages. Science 194, 1121–1132. Hearty, P.J., 1998. The geology of Eleuthera Island, Bahamas: a Rosetta Stone of Quaternary stratigraphy and sealevel history. Quaternary Science Reviews 17, 333–355. Hearty, P.J., Kindler, P., 1995. Sealevel highstand chronology from stable carbonate platforms (Bermuda and the Bahamas). Journal of Coastal Research 11(3), 675–689. Hearty, P.J., Neumann, A.C., 2001. Rapid sealevel and climate change at the close of the last interglaciation

Eustatic Sea Level During Past Interglacials (MIS 5e): evidence from the Bahama Islands. Quaternary Science Reviews 20(18), 1881–1895. Henderson, G.M., Slowey, N.C., 2000. Evidence from U–Th dating against northern hemisphere forcing of the penultimate deglaciation. Nature 404, 61–65. Imbrie, J., Hays, J.D., Martinson, D.G., McIntyre, A., Mix, A.C., Morley, J.J., Pisias, N.G., Prell, W.L., Shackleton, N.J., 1984. The orbital theory of Pleistocene climate: support from a revised chronology of the marine 18 O record. In: Berger, A.L., Imbrie, J., Hays, J., Kukla, G., Saltzman, B. (Eds.), Milankovitch and Climate, Part 1. D. Reidel, pp. 269–305. Kim, S.J., Crowley, T.J., Stossel, A., 1998. Local orbital forcing of Antarctic climate change during the Last Interglacial. Science 280, 728–730. Labeyrie, L.D., Duplessy, J.C., Blanc, P.L., 1987. Variations in mode of formation and temperature of oceanic deep waters over the past 125,000 years. Nature 327, 477–482. Lambeck, K., Chappell, J., 2001. Sealevel change during the last glacial cycle. Science 292, 679–686. Lea, D.W., Martin, P.A., Pak, D.K., Spero, H.J., 2002. Reconstructing a 350 ky history of sealevel using planktonic Mg/Ca and oxygen isotope records from a Cocos Ridge core. Quaternary Science Reviews 21, 283–293. Levitus, S., Boyer, T., 1994. World Ocean Atlas 1994. U.S. Department of Commerce, Washington D.C. Li, W.-X., Lundberg, J., Dickin, A.P., Ford, D.C., Schwarz, H.P., McNutt, R., Williams, D., 1989.High-precision mass-spectrometric uraniumseries dating of cave deposits and implications for paleoclimate studies. Nature 339, 534–536. Lundberg, J., Ford, D.C., 1994. Late Pleistocene sealevel change in the Bahamas from mass spectrometric U-series dating of submerged speleothem. Quaternary Science Reviews 13, 1–14. Martinson, D.G., Pisias, N.G., Hays, J.D., Imbrie, J., Moore Jr., T.C., Shackleton, N.J., 1987. Age dating and the orbital theory of the ice ages development of a high-resolution 0 to 300 000-year chrostratigraphy. Quaternary Research 27, 129. McManus, J., Oppo, D., Cullen, J., Healey, S., 2003. Marine isotope stage 11 (MIS 11) analog for Holocene and future climate? In: Earth’s climate and orbital eccentricity: the marine isotope stage 11 question. Geophysical Monographs 137, 69–85. Mo¨rner, N-A., 1971. Eustatic changes during the last 20,000 years and a method of separating the isostatic and eustatic factors in an uplifted area. Palaeogeography, Palaeoclimatology, Palaeoecology 9, 153–181. Murray-Wallace, C.V., 2002. Pleistocene coastal stratigraphy, sealevel highstands and neotectonism of the southern Australian passive continental

91

margin– a review. Journal of Quaternary Science 17(5–6), 469–489. Neumann, A.C., Hearty, P.J., 1996. Rapid sealevel changes at the close on the last interglacial (substage 5e) recorded in the Bahamian Island geology. Geology 24, 775–778. Pirazzoli, P.A., Radtke, U., Hantoro, W.S., Jouannic, C., Hoang, C.T., Causse, C., Borel Best, M., 1991. Quaternary raised coral-reef terraces on Sumba Island, Indonesia. Science 252, 1834–1836. Potter, E.-K., Lambeck, K., 2003. Reconciliation of sealevel observations in the Western North Atlantic during the last glacial cycle. Earth and Planetary Science Letters 217, 171–181. Rohling, E.J., Mayewski, P.A., Abu-Zied, R.H., Casford, J.S.L., Hayes, A., 2002. Holocene atmosphereocean interactions: records from Greenland and the Agaean Sea. Climate Dynamics 18, 587–593. Schellmann, G., Radtke, U., 2004. A revised morphoand chronostratigraphy of the late and middle Pleistocene coral reef terraces on Southern Barbados (West Indies). Earth-Science Reviews 64, 157–187. Scholz, D., Mangini, A., 2006. U-redistribution in fossil reef corals from Barbados, West Indies (this volume). Shackleton, N.J., 1987. Oxygen isotopes, ice volume and sealevel. Quaternary Science Reviews 6, 183–190. Shackleton, N., 2000. The 100,000 year ice-age cycle identified and found to lag temperature, carbon dioxide and orbital eccentricity, Science 289, 1897–1902. Shackleton, N.J., Pisias, N.G., 1985. Atmospheric carbon dioxide, orbital forcing, and climate. In: Sundquist, E.T., Broeker W.S. (Eds.), The Carbon Cycle and Atmospheric CO2: Natural Variations Archean to Present. Geophysical Monographs 32, 412–417. Shackleton, N.J., Opdyke N.D., 1973. Oxygen isotope and paleomagnetic stratigraphy of equatorial Pacific core V28–239, late Pliocene to latest Pleistocene. Geological Society of America Mem. 145, 449–464. Shackleton, N.J., Berger A., Peltier W.R., 1990. An alternative astronomical calibration on the lower Pleistocene time scale based on ODP site 677. Trans. Royal Society of Edinburgh, Earth Science 81, 2511–261. Siddall, M., Rohling, E.J., Almogi-Labin, A., Hemleben, Ch., Meischner, D., Schmelzer, I., Smeed, D.A., 2003. Sealevel fluctuations during the last glacial cycle. Nature 423, 853–858. Stearns, C.E., 1976. Estimates of position of sea level between 140,000 and 75,000 years ago. Quaternary Research 6, 445–449. Stein, M., Wasserburg, G.J., Aharon, P., Chen, J.H., Zhu, Z.R., Bloom, A., Chappell, J., 1993. TIMS U-series dating and stable isotopes of the last interglacial event in Papua New Guinea. Geochimica Cosmochimica Acta 57, 2541–2554.

92

M. Siddall, J. Chappell and E.-K. Potter

Stirling, C.H., Esat, T.M., McCulloch, M.T., Lambeck, K., 1995. High-precision U-series dating of corals from Western Australia and implications for the timing and duration of the Last Interglacial. Earth and Planetary Science Letters 135, 115–130. Stirling, C.H., Esat, T.M., Lambeck, K., McCulloch, M.T., 1998. Timing and duration of the last interglacial: evidence for a restricted interval of widespread coral reef growth. Earth and Planetary Science Letters 160, 745–762. Stirling, C.H., Esat, T.M., Lambeck, K., McCulloch, M.T., Blak,e S.G., Lee, D.-C., Halliday, A.N., 2001. Orbital forcing of the marine isotope stage 9 interglacial. Science 291, 290–293. Stocker, T.F., Wright, D.G., 1991. Rapid transitions of the ocean’s deep circulation induced by changes in the surface water fluxes. Nature 351, 729–732. Stocker, T.F., Johnsen, S.J., 2003. A minimum model for the bipolar seesaw. Paleoceanography 18, A-1087. Stocker, T.F., Clarke, G.K.C., Le Treut, H., Lindzen, R.S., Meleshko, V.P., Mugara, R.K., Palmer, T.N., Pierrehumbert, R.T., Sellers, P.J., Trenberth, K.E., Willebrand, J., 2001. Physical climate processes and feedbacks. In: Houghton, J.T., Ding, Y., Griggs, D.J., Noguer, M., van der Linden, P.J., Dai, X., Maskell, K., Johnson, C.A. (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, United Kingdom. Szabo, B.J., Ludwig, K.R., Muhs, D.R., Simmons, K.R., 1994. Thorium-230 ages of corals and duration of the last interglacial sealevel highstand on Oahu, Hawaii. Science 266(5182), 93–96. Thompson, W.G., Goldstein, S.L., 2005. Open-system coral ages reveal persistent suborbital sea level cycles. Science 308, 401–404. Toscano, M.A., Lundberg, J., 1999. Submerged Late Pleistocene reefs on the tectonically-stable SE Florida margin: high-precision geochronology, stratigraphy, resolution of Substage 5a sea level elevation, and orbital forcing, Quaternary Science Reviews 18(6), 753–767. Ve´zina, J., Jones, B., Ford, D., 1999. Sea level highstands over the last 500 000 years: evidence from the Ironshore Formation on Grand Cayman, British West Indies. Journal of Sedimentary Research 69(2), 317–327. Waelbroeck, C., Labeyrie, L., Michel, E., Duplessy, J.C., McManus, J.F., Lambeck, K., Balbon, E., Labracherie, M., 2002. Sealevel and deep water temperature changes derived from benthonic foraminifera isotopic records. Quaternary Science Reviews 21, 295–305. Wigley, T.M.L., Raper, S.C.B., 1987. Thermal expansion of sea water associated with global warming, Nature 397, 293–300.

8. Uranium-Series Dating of Peat from Central and Northern Europe Manfred Frechen1, Melanie Sierralta1, Deniz Oezen1 and Brigitte Urban2 1

Leibniz Institute for Applied Geosciences (GGA-Institut), Section Geochronology and Isotope Hydrology, Stilleweg 2, 30655 Hannover, Germany 2 University of Lu¨neburg, Herbert-Meyer-Str.7, 29556 Suderburg, Germany

ABSTRACT Interstadial and interglacial peat layers are widespread in the Northern Hemisphere and enable us to reconstruct the vegetation history. Uranium-series dating methods can provide a reliable and precise chronological frame for peat of the past 500 000 years. Uranium-series dating methods are based on the radioactive disequilibrium between 238 U and its radioactive daughter elements and the re-establishment of equilibrium. The activity ratio [230Th/234U] is a measure of the time elapsed since the formation of the peat. In this study, uranium-series dating of peat in Central and Northern Europe is reviewed including new thermal ionisation mass spectrometry (TIMS) 230Th/U dating results from recently investigated interglacial peat horizons in Central Europe. The suitability of the peat layers for dating depends on two essential assumptions: a closed-system behaviour excluding uranium migration after deposition and the contamination of peat by varying amounts of detrital thorium, which has to be corrected owing to its essential effect on calculating 230Th/U ages. The radiometric and TIMS dating results are in good agreement for certain fen peat layers correlating with MIS 3 and 5, as confirmed by independent age control through radiocarbon and luminescence dating methods. There are only a few approaches dating preEemian peat. Although the radiometric and TIMS dating results are less distinct, more reliable and precise absolute ages can be provided.

Keywords: uranium-series dating, peat, Pleistocene, palaeoclimate, Europe. 8.1 INTRODUCTION The reconstruction of past climate change allows an insight into what extent the environment has been altered and affected by climate forcing during the geological past. Terrestrial records such as loess/palaeosol sequences, fluvial terrace sequences, lake sediments, speleothems and peat bogs are distributed over all continents and can provide numerous important high-resolution archives of past climate and environment change. As anthropogenic activities have influenced the evolution of the present climate, it is of great interest to reconstruct the evolution of similar past interglacial periods to improve the prediction of climate change. In Central and Northern Europe, interstadial and interglacial peat layers are widespread and enable us to reconstruct the vegetation history. However, the reconstruction of climate and environmental change requires a reliable and precise chronological frame to determine the absolute timing of various climatic events during glacial and interglacial periods. Radiocarbon dating and uranium-series dating methods can provide reliable age results for peat deposits covering the time span of the past 60 000 years and the past 500 000 years, respectively. The direct dating of climatic and environmental events makes it possible to set up a more precise

94

Manfred Frechen et al.

chronological frame and therefore reconstruct climate conditions and changes during interstadial and interglacial periods of the past 500 000 years. Independent age control for aeolian and fluvial sediments sandwiching the interstadial and interglacial peat deposits can be provided by luminescence dating techniques (Frechen et al., 2003; Krbetschek and Degering, this volume) and hence improve the reliability of the chronological frame, at least for the last interglacial/glacial cycle. The uranium-series (230Th/U) dating method is based on the radioactive disequilibrium between 238U and its radioactive daughter elements and the re-establishment of equilibrium (Ivanovich and Harmon, 1992; Bourdon et al., 2003). Vogel and Kronfeld (1980) were the first who applied uranium-series dating to peat. Subsequently, van der Wijk et al. (1986), Heijnis (1992), Heijnis and van der Plicht (1992), Geyh and Techmer (1997), Geyh et al. (1997) and Rowe et al. (1997) reported variable degrees of success to date fen peat by applying alpha spectrometry to measure the isotope ratios. Over the recent years, uranium-series dating techniques have been significantly improved by replacing the radiometric alpha spectrometry through thermal ionisation mass spectrometry (TIMS). Through the development of the TIMS technique, the instrumentation and the methodology have been significantly enhanced, resulting in an essential reduction in sample size and analysis time, an improvement of the precision and sensitivity, as well as the extension of the dating range from 350 000 years to 500 000 years. Recently, multi-collectorinductively coupled plasma mass spectrometry (MC-ICPMS) has improved the analytical quality of uranium-series dating further (Goldstein and Stirling, 2003). The majority of TIMS dating applications of sediments from terrestrial records have been carried out on speleothem (Richards and Dorale, 2003), travertine (Mallick and Frank, 2002), peat, as well as on marine and lacustrine carbonates (Edwards et al.,

2003), yielding excellent age results up to about 500 000 years. The aim of this study was to review uranium-series dating of peat, including results from both the radiometric alpha spectrometry (230Th/234U) and TIMS (230Th/U). 230 Th/U dating results from recently investigated interglacial peat horizons in Central and Northern Europe range in age between 350 000 years and 40 000 years before present (BP). These interstadial and interglacial peat layers correlate with marine isotope stages (MIS) 9, 7, 5 and 3. The results including both radiometric 230Th/234U ages and new TIMS 230 Th/U ages are discussed to review the suitability of peat for uranium-series dating. 8.2 METHODOLOGY The natural radioactive isotope 238U decays through a complex series of intermediate nuclides with widely differing half-lives to stable 206Pb. In undisturbed sediments containing 238U for a geologically long period of time, radioactive equilibrium occurs. Essentially, the ratio of one isotope to another will be constant. Radioactive disequilibrium evolves following geochemical transitions, for example, weathering, until subsequently a dynamic secular equilibrium prevails. When the uranium decay is interrupted and some decay products are selectively removed, the uranium-series dating methods can be applied (Ivanovich and Harmon, 1992; for a comprehensive overview see Bourdon et al., 2003). 230Th/U dating is based on the assumption of the complete fractionation of uranium and thorium and furthermore assumes that no gain or loss of uranium and thorium occurred since the time of formation. 238 U decays via short-lived nuclides to 234 U, which decays with a half-life of 245 250 490 years to 230Th. Subsequently, 230 Th decays with a half-life of 75 690 230 years (Cheng et al., 2000) to its daughter element and via several short-lived nuclides to the stable isotope 206Pb. This decay is the

Uranium-Series Dating of Peat

only source of 230Th in nature. As long as the production of 230Th exceeds its radioactive decay, 230Th accumulates until the equilibrium is restored. The radioactive decay with time is expressed in terms of activity ratios of different isotopes, such as [230Th/234U]. If the decay chain is disturbed, the relative proportion of different isotopes will change, and the system will be in disequilibrium until equilibrium is restored through subsequent radioactive decay. The activity ratio [230Th/234U] is a measure of the time elapsed since the evolution of disequilibrium and so this activity ratio is a function of age (Equation (1), Fig. 8.1). The age can be directly determined, if the initial ratio [230Th/232Th] is zero. After about 500 000 years, approximately six half-lives, equilibrium is reached and the decay of 230Th balances its production. The time elapsed since the formation is given by Equation (1):

Th ¼ 238 U ð1 e230 t Þ 234 U 234 U 238 230 U þ 1 234 U 230 234

230

ð1 eð230 234 Þt Þ 10

2

ð1Þ

20 50 100

[234U]/[238U]

1.8

200

1.6

400 kyr

1.2

1

0

0.4

0.8

The ratios in brackets are the measured activity ratios, 230 and 234 represent decay constants, and t denotes the time elapsed since formation (Ivanovich and Harmon, 1992). Given measured [230Th/234U] and [234U/238U] ratios, Equation (1) cannot be solved directly, but for t an iterative procedure has to be applied. The ratios corrected for the initial Th can also be used to deduce the age from the graph (Fig. 8.1) Disequilibrium between the longer-lived isotopes can occur in natural systems like peat, speleothem, travertine, marine and lacustrine carbonates for a variety of reasons including the separation of different chemical elements during weathering or transport in water and deposition. Uranium-series dating of peat is based on the assumption of a complete fractionation of thorium and uranium during the formation of peat. After death of the plant and its burial in the peat bog, the peat layer has to behave geochemically as a closed system without postsedimentary mobilisation of uranium and thorium. In practice, there are many complications, and hence these dating assumptions are not necessarily fulfilled and have to be confirmed for each dating attempt. In the following, we will discuss in more detail the ‘closed-system behaviour’ and the ‘detrital correction’, which are the major sources of error for uranium-series dating. 8.2.1 Closed-system behaviour

300 1.4

95

1.2

[230Th]/[234U]

Fig. 8.1 Plot of activity ratios 234U/ 238U against 230 Th/ 234U in a closed system with no initial 230 Th. The horizontal curves show the development of the 234U/ 238U activity ratio with time. The dashed vertical lines are isochrons, lines of constant age.

Peat often gives evidence for an opensystem behaviour, which may be reflected by reversals of 230Th/U ages in the stratigraphic record, or the sample results themselves show evidence for open-system behaviour for uranium. Unfortunately, at present, there is no simple evidence during field or laboratory procedure for the selection of more suitable material for 230Th/U dating of peat. By weathering of host rock, uranium is leached from uranium-bearing minerals and dissolved in groundwater. During this process, U4þ is oxidised to U6þ in the ions. The crystal lattice and forms UO2þ 2

96

Manfred Frechen et al.

concentrations of uranium in groundwater may be up to 50 ng/g, in some cases even up to 2000 ng/g depending on the host rock (Ivanovich and Harmon, 1992). In contrast, thorium is almost insoluble in groundwater, but forms Th4þ ions, which tend to hydrolyse and polymerise. Thorium is strongly adsorbed onto clay minerals. Both uranium isotopes, 234U and 238U, have the same geochemical properties, but a natural process such as weathering will cause fractionation. The radioactive decay of 238U to 234U via 234Th and 234Pa is combined with a recoil of the 234U isotope. The recoil deforms the crystal lattice and allows 234U to move to positions closer to the surface. Therefore, 234U is more easily dissolved in water, resulting in 234 U/238U activity ratios up to 12 in groundwater (Ivanovich and Harmon, 1992). Seawater has a very constant global 234U/238U ratio of 1.14. Living plants and organic material such as peat take up uranium from the groundwater (Vogel and Kronfeld, 1980). Uranium is highly soluble in natural waters as uranyl ions ðUO2þ 2 Þ, or as uranyl carbonate complexes, and is transported by percolating groundwater to places where peat formation starts. Organic decomposition products like humic and fulvic acids have a large absorption capacity for uranium in the UO2þ 2 form (Szalay, 1958; Yliruokanen, 1980). Very stable uranyl organic complexes are formed by aromatic ring structures with hydroxyl and carboxyl groups (Szalay, 1958). Therefore, high uranium concentration in peat ranging from 1 to 100 ppm can be observed. The mobility of uranium in humic sediments or peat depends on the mobility of humic and fulvic acids. These acids are strongly bound to clay minerals implying low mobility of uranium in clay containing organic-rich material like peat (van der Wijk et al., 1986). Some encouraging results on dating fen peat bypassing the problem of postburial uranium uptake were reported by Heijnis et al. (1993), Geyh et al. (1997) and Rowe et al. (1997).

Burial by other sediments and compaction of peat can isolate organic-rich sediments from further uranium uptake or loss by percolating groundwater. Most of the uranium dissolved in groundwater is fixed and adsorbed in the upper and lower rim of peat layers about 10 cm thick, protecting the central part of the peat from further adsorption of uranium (Vogel and Kronfeld, 1980). It was found that the inner part of a 20–30 cm thick undisturbed peat layer might be considered as closed system (Heijnis and van der Plicht, 1992; Geyh et al., 1997). Therefore, the contaminated outer part of peat layers, 10 cm at the top and 10 cm at the bottom, as well as peat intercalated by thin sand layers or sand lenses have to be avoided for dating. A postsedimentary uptake of 234U results in a decreasing [230Th/234U] activity ratio and therefore in age underestimation. Loss of uranium owing to contact with oxygenated water, for example rain water, does change the isotopic composition in raised peat bogs. Thus, a reliable 230Th/U age is not possible for peat from raised peat bogs. Thorium is insoluble in groundwater, as concentrations of 230Th (IV) in groundwater are in the order of 8* 108 mg=g (Ivanovich and Harmon, 1992). Therefore, a postdepositional contamination is unlikely. In theory, the present-day 230Th can be solely attributed to the decay of its radioactive precursor and, providing that the peat remains a closed system, the amount of 230Th produced as 234U decays, will be a function of time and the initial uranium content of the sample. Each sample has to be investigated for opensystem behaviour. At present, no robust and reliable criteria are available to distinguish between a closed system and a quasi-closed system, allowing small amounts of uranium to pass through. Further investigations are essentially required. 8.2.2 Detrital correction At the time of formation, the concentration of 230Th is not necessarily zero. Peat may

Uranium-Series Dating of Peat

230

Th * ¼ 230 Th m 232 Th f0 e230 t ¼ 230 Th m f 232 Th

ð2Þ

where * denotes the radiogenic thorium, f0 is the initial 230Th/232Th activity ratio, is the decay constant of 230Th, f is the detritus correction factor, m means measured and t denotes time. At least three or preferably more coeval samples containing different detrital components must be analysed to construct isochrons (Fig. 8.2). The [234U/238U] and the [230Th/234U] activity ratios are obtained for the age calculation (see Equation (1)) from the slope of

20

Low detrital contamination 16

[230Th]/[232Th]

contain varying amounts of detrital thorium, which has an essential effect on calculating 230Th/U ages. Thorium is mostly adsorbed onto clay minerals, which are admixed by dust or by flowing water into the peat during its formation. The presence of detrital 230Th can be detected by measuring the amounts of the natural isotope 232Th in a sample. As 232Th has a long half-life ðt1=2 ¼ 1:39* 1010 aÞ, it is thought to be stable. The [230Th/232Th] activity ratio may be used as a basis for the correction of detrital contamination (Schwarcz and Latham, 1989). If initially detrital 230Th (and 232Th) was accumulated in the peat, the [230Th/234U] activity ratio is > 0 at t ¼ 0, and the calculated 230Th/U ages are too large, unless corrections are applied for the detrital 230Th. To correct for admixed detritus with a uniform [230Th/232Th] activity ratio at the time of formation, the isochron method is applied (Kaufman and Broecker, 1965; Osmond et al., 1970; Kaufman, 1971; Ku and Liang, 1984; Schwarcz and Latham, 1989; Luo and Ku, 1991) (see Fig. 8.2). This method is based on the assumption of a two-component mixing of radiogenic 230Th (230Th*) and detrital 230Th with a fixed activity ratio of [230Th/232Th], as described by Equation (2):

97

12

= [230Th*/234U] 8

High detrital sample 4

f = 230Th/ 232Th (detrital thorium) 0 0

4

8

12

16

20

24

[234U]/[232Th]

Fig. 8.2 Plot of activity ratios 230Th/ 232Th against 234 U/ 232Th illustrating the principle of the isochron approach for the detrital correction. The intercept of the line and the y-axis defines the correction factor for the detrital contamination (f-factor).

the best fit line of the two activity ratio diagrams [234U/232Th] versus [238U/232Th] and [230Th/232Th] versus [234U/232Th] activity ratios. An initial test to decide whether or not detrital correction is necessary can be applied by calculating 230Th/U ages with initial [230Th/232Th] activity ratios, f0 ¼ 0 and f0 ¼ 1. If the results are not in agreement with in the standard deviation or differ by more than 10%, then the [230Th/232Th] activity ratio has to be corrected applying the isochron method. An important assumption of the isochron method requires that the age span covered by the coeval samples does not exceed the confidence interval of the single 230Th/U age. To avoid methodologically enlarged errors, we used a modified procedure for the calculation of the isochron age, as described by Geyh (2001). The plot of the [230Th/232Th] activity ratio versus the [234U/232Th] activity ratio is used for determination of the detrital correction factor ( f ) and its standard deviation from the intersection of the isochron with the y-axis (Fig. 8.2). This f-factor is then used to

98

Manfred Frechen et al.

correct each coeval sample separately according to Kaufman and Broecker (1965) (Equation (2)). If the f-factor is > 10, then the 232Th activity and so the detrital 230Th activity is so small that a detrital correction does not alter the 230Th/U age. If f < 10, the difference between the corrected and uncorrected 230 Th/U ages can exceed up to some 100 kyr. The weighted mean of these corrected ages is the isochron-corrected 230 Th/U age. A Chi-square test is applied to verify whether the corrected dates belong to the same normal distribution. The Chisquare test should not be much larger than the number of samples used to calculate the isochron age. If the detrital contamination of the coeval samples differ only slightly, scatter plots appear, which do not enable a reliable determination of the slope of the isochron. If the difference is large enough, the slope increases with increasing age against 1, which correlates with the radioactive equilibrium. 8.3 EXPERIMENTAL DETAILS At the GGA-Institut, the chemistry for the extraction of uranium and thorium for TIMS 230 Th/U dating from the samples was adapted from the leachate/leachate technique (Schwarcz and Latham, 1989; Kaufman, 1993). To meet the isochron requirements, at least three separate coeval dry peat samples from the same depth of approximately 0.3–0.5 g each were combusted in an O2 flow at a temperature in the region of 800 C. The remaining ashes were treated with NaOH and dissolved in a concentrated HNO3/ HCl mixture. A 229Th spike and a 233 U–236U double spike were added to each sample and heated for about 18 h to achieve equilibrium between solution and spike. The leachate was separated from the residue by centrifuging before uranium and thorium were isolated from the leachate by co-precipitation with Fe(OH)3. The precipitate was re-dissolved in concentrated

HNO3. The final separation was achieved by conventional ion-exchange chromatography of uranium and thorium on a column of DOWEX 18 100–200 mesh. Thorium was washed from the column by HCl, while subsequently uranium was eluted from the column by HBr. The purified uranium and thorium fractions were loaded separately without any carrier on rhenium filaments. The isotopic ratios were measured by TIMS (Finnigan MAT262 RPQ), applying the double-filament technique. Conventional radiometric measurements of uranium and thorium isotopes were obtained through alpha spectrometry, the monitoring of alpha particles emitted during radioactive decay, until the mid-1990s (Hennig, 1979). Since 1997, TIMS has been employed to determine the isotope ratios and replaced the alpha spectrometry. As for AMS 14C dating, TIMS enables individual atoms to be counted directly. TIMS is more rapid and, as count rates are not restricted by the half-life of the isotopes, there is the potential for extending the age range of the method. Furthermore, TIMS offers a greater level of analytical precision.

8.4 APPLICATION In the following, we will present case studies applying both the radiometric 230 Th/234U and TIMS 230Th/U dating of peat deposits from Northern and Central Europe (Fig. 8.3), ranging in age from MIS 3 to MIS 9 and older. 8.4.1 Gossau, Switzerland (MIS 3) The section at Gossau is located in a gravel quarry near Gossau east of the Zu¨richsee (Kanton Zu¨rich) in Switzerland (Fig. 8.3) and was investigated in detail by Schlu¨chter et al. (1987). The lower part of the sequence consists of sand and gravel related to proximal delta deposits including delta foresets, delta topsets and flood loam. These sediments were most likely deposited during a

Uranium-Series Dating of Peat 0

100 mi

DK

100 km HH Bossel

NL

PL

Gröbern

GroßRohrheim

B

CZ

L

F

CH

Zell

Gossau

A

Fig. 8.3 Map showing the sections of interest in Central Europe. HH ¼ Hamburg-Isfeldstraße.

cold climate period correlating with MIS 4, as suggested by their pollen assemblage including Betula, Artemisia and Juniperus (Schlu¨chter et al., 1987). The middle part of the sequence consists of weathered gravel and overbank deposits intercalated by at least three compressed lignite horizons, which were correlated with the MIS 3 (Fig. 8.4). These interstadial deposits were covered by proximal fluvioglacial outwash gravel and lodgement till most likely deposited during the last glacial maximum (MIS 2). The lignite and the overbank deposits were investigated by radiocarbon, uranium-series and luminescence dating approaches (Geyh et al., 1997; Geyh and Schlu¨chter, 1998; Preusser et al., 2003). The delta sediments from below the lower lignite horizon gave IRSL age estimates ranging from 113 17 to 103 18 kyr. The lower lignite horizon yielded a radiometric 230 Th/234U isochron age of 49:1 3:7 kyr and AMS 14C ages ranging from 54:0 3:0 to 45:4 1:2 kyr. Samples from overbank deposits in between the lower and upper lignite horizons yielded IRSL age estimates

99

of 51:8 5:4 and 51:7 7:0 kyr. The middle lignite horizon gave radiometric 230Th/234U ages of 47:8 6:0 kyr (lower layer) and 37:6 2:3 kyr (upper layer). The AMS 14C ages range from 40:9 1:1 kyr (lower layer) to 33:0 2:5 kyr (upper layer). Overbank deposits from below the uppermost lignite horizon yielded IRSL age estimates of 30:1 3:2 and 29:0 3:9 kyr. The uppermost lignite gave a 230Th/234U age of 34:7 4 kyr and AMS 14C ages ranging from 29:5 1:2 to 28:3 0:4 kyr. Compared to the radiometrically determined 230Th/234U results, the radiocarbon ages are slightly underestimated. The IRSL age estimates are in good agreement with the AMS 14C and uranium-series ages. The multidisciplinary approach and the agreement of the dating results among the different methods confirm the suitability of lignite from the section at Gossau for uranium-series dating. It is most likely that the three humic-rich horizons correlate with interstadial periods during MIS 3 (Fig. 8.4). 8.4.2 Gro¨bern, Germany (MIS 5a) The Gro¨bern section is located in an opencast brown coal pit in the north-east of Bitterfeld in Saxony-Anhalt (Fig. 8.3). The basin of Gro¨bern was a sediment trap since the Late Saalian resulting in rather continuous lake sediments deposited during the Eemian and the Lower Weichselian, as recorded by palaeobotanical remains and geological estimates (Litt, 1990; Wansa and Wimmer, 1990; Eissmann and Litt, 1994). The complete vegetational succession of the Eemian is represented by an early birch zone, which is followed by a birch-pine-, pine-mixed-oak-forest-, a mixed-oak-foresthazel- and a hazel-yew-lime-zone, the latter characterising the climatic optimum of the last interglacial. The vegetational succession of the terminal part of the Eemian interglacial is characterised by a hornbeam-, hornbeam-fir- a pine-spruce-fir- and finally a pine-zone, marking the increasing boreal climatic character of the ending interglacial.

100

Manfred Frechen et al.

Lithology Gossau/Zell Zell

2

3

Gossau

2

Gossau

MIS

LUM age [kyr] 15 ± 2 13 ± 1 15 ± 3

230Th/U

age

[kyr]

34.7 ± 4

28.3 ± 0.4 29.3 ± 1.2

37.6 ± 2.3 47.8 ± 6.0

33.0 ± 2.5 40.9 ± 1.1

49.1 ± 3.7

45.4 ± 1.2 54.0 ± 3.0

30 ± 3

Zell

50 ± 2

4?

91 ± 2

122 ± 12 117 ± 15 124 ± 27

Zell

6

Zell

100 ± 27 100 ± 9 91 ± 9

5

AMS 14C age [kyr]

Silty sand Fine-grained sand Gravel superimposed by soil formation Fluvio-glacial gravel Till Lignite

Fig. 8.4 Idealised lithology and independent age control including AMS 14C ages (Geyh and Schlu¨chter, 1998) and luminescence age estimates (Preusser et al., 2001, 2003) for the Zell section and the Gossau section both located in Switzerland.

The early Weichselian stadials WF I (Herning) and WF III (Rederstall) are characterised by steppe-like conditions correlating with MIS 5d and MIS 5b. The early Weichselian interstadials WF II (Brørup) and WF IV (Odderade) were forest periods characterised by boreal taxa (Litt, 1990) and correlating with MIS 5c and MIS 5a, respectively (Fig. 8.5). Organic-rich sediments including fen peat, which correlate with the Brørup interstadial by means of pollen spectra, were deposited at the section at Gro¨bern. Independent age control is available by preliminary thermoluminescence (TL) age estimates (Krbetschek and Stolz, 1994). The preliminary TL age estimates range from 97 16 kyr to 119 15 kyr for the sediments, which correlate with the lower part

of the Eemian interglacial. A preliminary TL age estimate of 75 8 was determined for the sandy silt below the peat layer designated to correlate with the Odderade interstadial (Litt, 1990; Eissmann and Litt, 1994). Furthermore, the palynological results give evidence for a correlation with the Brørup and Odderade interstadials (Litt, 1990) suggesting a formation of the peat during MIS 5c and MIS 5a, respectively. For the first radiometric dating approach, samples were taken for alpha spectrometry from two horizons, a 50 cm thick peat layer at a depth of 5.8–7.0 m below surface and an organic-rich peat clay layer 70 cm thick underlying the peat layer. The peat and the peat clay are intercalated by thin layers of sand, probably causing an open-system

Uranium-Series Dating of Peat MIS [m] 0

5a

Lithology Gröbern

LUM age [kyr]

Odderade interstadial 75 ± 8 5 Brørup interstadial

5c

Modern soil Silt Silty mud

10

Silty sand Eemian interglacial

5e

15

Sand 97 ± 16 119 ± 15 104 ± 12

Gravel sand Calcareous mud Fine detritus mud Peat mud Till

Fig. 8.5 Lithology of the Upper Pleistocene sequence from the section at Gro¨bern (modified after Wansa and Wimmer, 1990) including preliminary TL age estimates (Krbetschek and Stolz, 1994). The uranium-series dating samples were taken from peat and peat clay, correlating with the Bro¨rup interstadial.

behaviour. The content of detrital material is relatively high. During a second field session, fen peat was sampled again from the same stratigraphic position to study the suitability of this material for TIMS 230Th/U dating. The radiometric and TIMS 230Th/U dating results are given in Table 8.1 and 8.2, as well as isochron plots and the Osmond– Ivanovich diagram (Fig. 8.6a, 8.6b) (Osmond and Ivanovich, 1992). The isochron plot indicates a large scatter of the data points. The activity ratios, as measured by alpha spectrometry, yielded an isochron for the peat, whereas activity ratios of material from the peat clay scatter largely. Although a detrital correction ðf0 ¼ 1Þ was applied, the Osmond–Ivanovich plot

101

([234U/238U] versus [230Th/238U]) indicates an open-system behaviour. The activity ratios of the peat scatter widely, while the activity ratios for the peat clay plot closely together. Detrital correction was not applied for the peat clay because it was impossible to fit an isochron by the data points (Fig. 8.6b; Table 8.2). The activity ratios, as measured by TIMS, gave only small differences for the three samples resulting in a close scatter plot in the Osmond–Ivanovich diagram (Fig. 8.6b). A reliable correction factor f could not be determined by the isochron plot owing to the intersection of the isochron with the negative part of the x-axis. However, a detrital correction is necessary for the samples from Gro¨bern, as evidenced by a visible large amount of detrital sediment and the measured [230Th/232Th] activity ratios smaller than 10. The uncorrected activity ratios were plotted in the Osmond–Ivanovich diagram. Owing to the small scatter of the thorium activity ratios, the three data points plot around an isochron of 200 kyr. However, the [234U/238U] activity ratios, as measured by TIMS, do not show large scattering (Table 8.2) and so do not confirm an open-system behaviour. Both the radiometric and the TIMS dating approach failed to determine 230Th/ U ages for the peat, as the ages could not be corrected for the detrital contamination (Fig. 8.6a, 8.6b). Based on preliminary TL age estimates and palynological investigations, it is likely that the sampled peat horizon correlates with the Brørup and Odderade interstadials designated to correlate with MIS 5c and 5a, respectively. 8.4.3 Zell, Luthern Valley, Switzerland (MIS 5c) The section at Zell is located in the Luthern Valley in central Switzerland (Fig. 8.3). The sequence consists of three gravel horizons (Fig. 8.4). The lower gravel horizon consists of homogenous gravel up to 44 m thick, which were deposited by a braided river

102

Table 8.1 Radiometric and TIMS dating results from Zell Samplea

Th (ppm)

Uh 1215 Uh 1216 Uh 1224 Uh 1225 Uh 1226 Uh 1227

9.75 9.85 13.9 10.8 12.7 8.78

1.35 1.23 0.99 1.05 1.02 1.13

TIMS 27 TIMS 28 TIMS 30 TIMS 60 TIMS 117

11.75 5.53 15.31 13.2 13.2

TIMS 648 TIMS 649 TIMS 650 TIMS 651 TIMS 652 TIMS 653

10.02 9.87 9.96 9.82 9.81 9.69

a

[234U/238U] b

[230Th/234U]

2

AR

1.0947 1.0878 1.0952 1.0724 1.0904 1.1054

0.0058 0.0055 0.0078 0.0045 0.0048 0.0065

0.80 3.48 0.90 1.47 1.23

1.1048 1.1225 1.1055 1.0958 1.0932

1.33 1.32 1.42 1.26 1.37 1.35

1.0806 1.0924 1.0822 1.0904 1.1015 1.1045

AR

b

[230Th/232Th]

2

AR

0.6464 0.6650 0.5652 0.6023 0.6178 0.5966

0.0077 0.0064 0.0077 0.0051 0.0058 0.0069

0.0017 0.0203 0.0028 0.0010 0.0030

0.5971 0.6631 0.6122 0.576 0.6032

0.0021 0.0022 0.0014 0.0028 0.0059 0.0027

0.6308 0.6284 0.6278 0.6235 0.6326 0.6363

b

Ageuncorrected þ

2

(ka)

14.8199 17.6431 26.4101 20.2507 25.5113 15.5336

0.2014 0.2572 0.4134 0.2752 0.3801 0.2458

95.5 116.1 89.0 98.6 102.5 96.6

1.6 2.1 2.0 1.4 1.7 1.9

1.6 2.1 1.9 1.4 1.7 1.9

0.0315 0.0101 0.0966 0.0018 0.0012

29.7211 3.6139 35.3635 17.2962 21.5031

2.1067 0.0312 7.5451 0.0638 0.0366

97.0 115.0 100.9 91.77 98.74

8 3.6 29 0.5 0.4

7 3.4 23 0.5 0.4

0.0028 0.0029 0.0010 0.0031 0.0027 0.0022

15.6192 15.5292 14.4948 16.0290 15.1197 15.3612

0.0862 0.1781 0.0358 0.0952 0.0446 0.0779

106.5 105.5 105.6 104.2 106.5 107.5

0.8 0.9 0.3 0.9 0.9 0.7

0.8 0.9 0.3 0.9 0.9 0.7

Samples with Uh notation were measured by alpha spectrometry, samples with prefix TIMS were measured by thermal ionisation mass spectrometry. AR denotes activity ratio (measured). c Corrected ages were calculated with f0 ¼ 1 (alpha spectrometry). b

Agecorrected

f-factor (ka) 1c 1c 1c 1c 1c 1c

þ

91 112.1 86.7 95.4 99.9 92.5

1.6 2.2 2 1.4 1.7 1.9

1.6 2.1 1.9 1.4 1.7 1.9

0.575 0.575 0.575 0.575 0.575

95.6 91.3 99.9 87.7 95.0

8.6 4.1 29.6 1.1 1.2

7.9 3.9 23.3 1.1 1.2

1.4064 1.4064 1.4064 1.4064 1.4064 1.4064

91.5 90.7 89.7 90.1 91.1 92.1

1.3 1.4 0.5 1.5 1.4 1.1

1.3 1.4 0.5 1.4 1.4 1.1

Manfred Frechen et al.

U (ppm)

Uranium-Series Dating of Peat (a) 8 Gröbern Alpha-clay

[230Th]/[232Th]

6

4

Alpha-peat

2

TIMS 0 0

4

8

12

16

[234U]/[232Th] (b)

Gröbern 10

2

20 50

[234U]/[238U]

100 200

1.6

300 400 kyr 1.2

0.8 0

0.4

0.8

1.2

1.6

[230Th]/[238U]

Fig. 8.6a Isochron plot showing the activity ratios 230Th/ 232Th against 234U/ 232Th of the peat and peat clay samples from Gro¨bern. Green dots and blue triangles show results from radiometric peat and peat clay measurements, respectively; red dots indicate the TIMS dating activity ratios. Fig. 8.6b Osmond–Ivanovich diagram for activity ratios from Gro¨bern. For symbol description see Fig. 8.6a. Two radiometric data points of the peat fall off the scale of the diagram (see Table 8.2).

system under cold climatic conditions (Ku¨ttel, 1989; Preusser et al., 2001). The middle gravel horizon consists of fine-grained sand intercalated by lignite horizons. The sediment was most likely accumulated in an oxbow during an interstadial climate. The interstadial deposits at the section at Zell

103

correlate with the Huttwil interstadial by means of pollen spectra (Ku¨ttel, 1989), the first interstadial following the last interglacial correlated with MIS 5e. A boreal forest type with Picea and Pinus is typical of the Huttwil interstadial, which is designated to correlate with MIS 5c. The upper gravel horizon consists of homogenous coarse-grained gravel deposited by a braided river system. Peat samples were taken for radiometric and TIMS 230Th/U dating from the upper lignite horizon. Radiometric measurements yielded an 230Th/234U age of 95 3 kyr (Geyh et al., 1997). The sandy gravel (‘Mittlere Zeller Schotter’) sandwiching the lignite gave luminescence age estimates ranging from 124 27 to 91 9 kyr for potassium-rich feldspars (Preusser et al., 2001). These IRSL age estimates show large uncertainties but are within their standard deviation in agreement with the radiometric dating results of the lignite. We have re-investigated the peat from the section at Zell in order to study the suitability of the material and the external reproducibility for TIMS 230Th/U dating of peat. Seventeen samples including six samples for alpha spectrometry and 11 samples for TIMS were taken from the central part of the lignite avoiding the outer 10 cm at the bottom and at the top of the lignite. The results of both uranium-series dating studies are presented in the isochron plot and the Osmond–Ivanovich diagram (Fig. 8.7a, 8.7b). A good agreement between the radiometric and TIMS results was found. In the isochron plot, the fit of the line differs only slightly in slope and intersection with the y-axis. A TIMS 230Th/U age was determined by the isochron method and yielded an age of 91 2 kyr (Table 8.1). As the plot of activity ratios in the Osmond– Ivanovich diagram shows a narrow cluster, we conclude that the lignite behaved like a closed system. Furthermore, there is no evidence for open-system behaviour from the uranium and thorium content of the layer (Table 8.1). The lignite from the section at Zell was suitable for uranium-series dating

104

Table 8.2 Radiometric and TIMS dating results from the peat sections presented Samplea Gro¨bern Uh 972 Uh 973 Uh 974

Peat

Clay

Peat

Hamburg-Isfeldstraße Uh 57 4.20–4.25 Uh 58 4.30–4.35 Uh 59 5.05–5.10 Uh 60 5.20–5.25 Uh 62 6.10–6.15 Uh 73 4.20–4.70 Uh 74 5.50–5.95 Uh 75 5.95–6.05 Uh 76 6.05–6.10 Uh 77 6.10–6.15 Uh 78 6.15–6.20

0.08 0.13 0.15 0.68 1.83 1.16 0.26 0.34 0.35 0.163 0.261 0.123 0.063 2.255 0.782 0.074 0.916 2.377 3.665 3.289

0.07 0.39 0.16 1.24 13.7 13.2 0.68 0.58 0.72 0.264 0.567 0.216 0.024 0.481 1.620 0.145 0.613 1.065 1.277 1.288

[234U/238U] AR

b

3.320 1.368 1.466

[230Th/234U] b

2

AR

0.210 0.072 0.061

4.743 3.613 3.738

2 0.462 0.181 0.177

[230Th/232Th] AR

b

12.311 1.433 4.037

Ageuncorrected

Agecorrected

f-factor

2

(ka)

þ

1.184 0.083 0.206

49.8 > 197:3 212.7

2.7 2.7 minimum age 26.5 20.9

(ka)

þ

1c

41.4

1c

188.5

27.8 22.2

c

74.8 14.7 69.1 – – –

110.9 74.8 5.1 5 7.7 7.1

35.8 73 132 79 117 42 99 64.6 99 187 87

0.977 0.981 1.021 1.1270 1.0519 0.9879

0.021 0.014 0.019 0.0033 0.0036 0.0042

1.310 1.444 6.403 1.3089 1.8572 1.4805

0.029 0.107 0.742 0.0038 0.0077 0.0088

1.627 4.513 12.445 1.1494 1.6235 1.2504

0.053 0.368 1.516 0.0060 0.0126 0.0126

180 41.9 78.1 209.6 216.5 204.2

19.7 4.2 7.6 3.9 6.0 6.8

16.1 4 7 3.8 5.7 6.4

1 1c 1c n.d. n.d. n.d.

1.566 1.288 1.160 1.279 1.033 1.147 1.307 1.130 1.053 0.974 1.071

0.097 0.044 0.073 0.096 0.029 0.049 0.161 0.088 0.018 0.068 0.017

0.531 0.782 0.868 0.672 0.687 0.727 0.814 0.561 0.658 0.838 0.608

0.08 0.111 0.059 0.065 0.087 0.067 0.074 0.043 0.049 0.093 0.048

1.560 1.409 1.738 2.288 10.115 1.225 1.652 2.878 4.707 7.121 5.064

0.396 0.304 0.149 0.363 9.518 0.056 0.082 0.146 0.553 0.785 0.38

77.8 151 200 114 125 135 163 87.4 115 201 100

18.2 57 58 23 36 29 51 12.1 17 155 14

15.5 38 35 18 27 23 32 10.4 14 54 12

1c 1c 1c 1c 1c 1c 1c 1c 1c 1c 1c

2.9

12.7 26 26 16 33 12 24 9.4 14 114 12

2.9

11.3 21 20 14 25 11 18 8.3 12 49 11

Manfred Frechen et al.

Uh 975 Uh 976 Uh 977 TIMS 632 TIMS 633 TIMS 634

Layer U Th Depth (m) (ppm) (ppm)

Bossel TIMS 426 TIMS 427 TIMS 428 TIMS 431 TIMS 433 TIMS 434 TIMS 430 TIMS 432 TIMS 435 TIMS 436

38 38 38 19 19 38 19 19 19 19

13.4 18.1 18.0 73.9 7.2 4.828 3.381 3.070 0.036 0.073 3.118 0.046 0.039 0.089 0.083

1.6 1.3 1.5 2.2 2.6

1.187 1.203 1.135 1.155 1.220

0.010 0.044 0.017 0.014 0.029

0.870 0.957 0.908 0.961 0.948

0.009 0.094 0.017 0.026 0.023

0.213 0.339 0.203 0.107 0.193 0.164 0.152 0.134 0.250 0.236

1.1584 1.1278 1.1671 0.9573 0.9820 1.1698 1.0783 1.0584 1.0358 1.0475

0.0019 0.0019 0.0019 0.0150 0.0608 0.0020 0.0104 0.0058 0.0076 0.0023

0.9839 0.9959 0.9842 1.1504 1.0347 0.9801 0.9891 0.9700 1.0073 0.9876

0.0020 0.0113 0.0041 0.0317 0.0644 0.0146 0.0176 0.0167 0.0264 0.0267

25.79 47.56 37.07 111.17 9.77 78.3865 33.9620 52.7383 1.1223 1.1722 66.3745 0.9775 0.8997 1.1309 1.1084

0.58 1.75 1.18 2.76 0.26

199 265 231 280 253

6 301 18 43 30

0.1722 0.4554 0.2494 0.0259 0.0175 1.1724 0.0148 0.0148 0.0287 0.0301

311 4.0 347 30 308 7.2 Age > 650 ka Age > 650 ka 301 24 363 78 331 48 543 5102 388 219

6 75 15 31 23

1c 1c 1c 1c 1c

195 263 229 279 244

3.8 24 6.7

0.674 0.674 0.674

298.0 309.5 290.0

0.674 n.d n.d n.d n.d

287.5 – – – –

20 45 33 173 70

5:9 298 18 43 28

74 15 31 22

3.5 3.4 20.4 17.3 6.0 5.7 n.d – n.d – 20.9 17.7

Uranium-Series Dating of Peat

Groß-Rohrheim Uh 288-1 19–21 Uh 288-2 19–21 Uh 288-3 19–21 Uh 316 21 Uh 317 18–19

The data concerning the sections at Groß-Rohrheim and Bossel were previously published by Schweiss (1988) and Geyh and Mu¨ller (2005), respectively. a Samples with Uh notation were measured by alpha spectrometry, samples with prefix TIMS were measured by thermal ionisation mass spectrometry. b AR denotes activity ratio (measured). c Corrected ages were calculated with f0 ¼ 1 (alpha). n:d: ¼ not determined, for explanations see text.

105

106 (a)

Manfred Frechen et al. 50

Zell Alpha 40

[230Th]/[232Th]

TIMS 30

20

10

0 0

20

40

60

80

[234U]/[232Th]

(b)

Zell 10

2

20 50

[234U]/[238U]

100 200

1.6

300 400 kyr 1.2

0.8 0

0.4

0.8

[

1.2

1.6

230Th]/[238U]

Fig. 8.7a Isochron plot showing the activity ratios 230 Th/ 232Th against 234U/ 232Th of the lignite samples from Zell. Blue dots show results from alpha spectrometry; red and black dots indicate results from TIMS measurements. Fig. 8.7b Osmond–Ivanovich diagram for radiometric and TIMS activity ratios of lignite from the Zell section. For symbol description see Fig. 8.7a.

applying both methods. Based on uraniumseries dating, the lignite most likely correlates with MIS 5c. 8.4.4 Hamburg–Isfeldstraße (MIS 5e) The section at Hamburg–Isfeldstraße (qee1) is located to the west of the city of Hamburg,

within a landscape dominated by Saalian moraines. Weichselian glacial deposits are exposed in the vicinity about 400 m away from the section of interest. The Saalian (Drenthian) ground moraine has a thickness of > 20 m in the study area. The section is located in a 6 m deep, almost round depression (70–80 m diameter), most likely formed by the melting of buried ice (Linke and Hallik, 2006). The lower part of the sediment sequence (Fig. 8.8) above a till horizon consists of 2.25 m thick organic-rich sediments including a complete pollen record correlating with the generalised pollen diagram of the Eemian interglacial in Middle and West Holstein, Germany. During the interglacial optimum, the area around the section at Hamburg–Isfeldstraße is nearly Taxus-free, which is the only difference to the Eemian pollen record, as defined by Menke and Tynni (1984). These interglacial deposits are covered by up to 4 m thick coversand and dune sand of Weichselian age. The original thickness of the organic-rich sediments was about 6 m and subsequently reduced by compaction to 2.25 m, as calculated by Linke and Hallik (2006). A detailed geological and palynological description was presented by the same authors. At the bottom of the organic-rich sediments, gyttja up to 0.32 m thick was formed at a depth between 6.35 and 6.03 m below surface, most likely owing to a higher water table, resulting in a water column of about 1 m during the early part of the Eemian. The upper part of the gyttja is very similar to fen peat and most likely forms a transition zone. Fen peat is exposed between 6.03 and 5.77 m below surface. Raised peat bog deposits with sphagnum imbricatum formed in a period with higher evaporation most likely during the climate optimum and cover the fen at a depth between 5.77 and 5.40 m below surface. At the end of the interglacial, swamp forest peat formed at a depth ranging from 5.40 to 4.65 m below surface most likely owing to a higher water table. This swamp forest peat is covered by raised peat

Uranium-Series Dating of Peat [m] 4.00

Lithology Sample Pollen 230Th/U age Hamburgzone [kyr] lsfeldstraße

(a)

107

8

Hamburg

+12.7

36.8 –11.3 VII

58

+12

73 4.50

6

+26

73 –21

[230Th]/[232Th]

57

42 –11 VI

Vb 5.00

2

+26

59

Va 60

4

132 –20

+16

79 –14

0 5.50

IVb

0

II I

6.50

Sand Sphagnum peat Swamp forest peat

8

(b)

10

Hamburg 10

2

20 50

Fen peat

100

Mud/gyttja Till

Fig. 8.8 Lithological sequence of the section at Hamburg-Isfeldstraße (modified after Linke and Hallik, in press) including sample positions, pollen zones and the 230Th/ 234U ages in kyr (¼ 1000 years).

[234U]/[238U]

III

+9.4 64.6 –8.3 +14 99 –12 +33 117 –25 +114 187 –49 +12 87–11

6

[234U]/[232Th]

99 –18 IVa

75 76 62+77 78

4

+24

74

6.00

2

200

1.6

300 400 kyr 1.2

0.8

bog deposits with sphagnum between 4.65 and 4.10 m below surface. A radiometric 230Th/234U weighted mean age of 103 5 kyr was determined on the basis of all 11 investigated samples taken through the whole profile (Geyh et al., 1997) (Fig. 8.9a, 8.9b). The peat sequence from Hamburg–Isfeldstraße consists of a combination of different types of peat, in fact a mixture of raised bog peat and fen peat including several transitions between ombrogenic and topogenic peat. More recent palynological investigations yielded a typical Eemian pollen spectrum, indicating an early interglacial deposition age of the fen peat (Linke and Hallik, 2006; H. Mu¨ller, unpublished data). Four samples from the lower part of the core (Uh 62, Uh 76, Uh 77, and Uh 78; Table 8.2) were taken from fen peat or from material of the transition zone between fen peat

0

0.4

0.8

1.2

1.6

[230Th]/[238U]

Fig. 8.9a Isochron plot showing the activity ratios 230 Th/ 232Th against 234U/ 232Th of the peat samples from Hamburg-Isfeldstraße. Fig. 8.9b Osmond– Ivanovich diagram for the activity ratios from Hamburg-Isfeldstraße.

and gyttja, as described by Linke and Hallik (2006) and were recalculated. The remaining samples were excluded from further calculations because they mainly derived from sphagnum peat, which is regarded to be material unsuitable for uranium-series dating, as described by Geyh and Techmer (1997). The aim of this study was to understand the potential underestimation of the 230Th/234U ages, as determined by Geyh et al. (1997) and later criticised by Linke and Hallik (2006).

108

Manfred Frechen et al.

All samples have high uranium and thorium content. A weighted mean age of 94 9 kyr was calculated for the corrected samples under study (Uh 62, Uh 76, Uh 77 and Uh 78). The Chi-square test gave 1.6 ðn ¼ 4Þ. Using the second approach following Geyh (2001), an isochron was fitted to data of Uh 62, Uh 76, Uh77 and Uh78. Corrected ages were calculated using the new determined correction f-factor (0.135) resulting in a weighted mean age of 104 10 (Chi-square ¼ 1:2; n ¼ 4). This application clearly demonstrates the need for a precise determination of the f-factor and shows the significant influence of different detritus corrections on the ages. The mean 230Th/234U age is still underestimated, if compared with the pollen record. An early Eemian deposition age would correlate with the time span ranging from about 130 to 125 kyr BP. According to the palynological results, the mean 230 Th/234U ages of 94 9 or 104 10 kyr are about 20–25% or 10% underestimated, respectively. Reasons for an age underestimation could be an unrecognised postdepositional uranium uptake by percolating groundwater. There is no distinct evidence for postdepositional migration of uranium from the Osmond–Ivanovich diagram (Fig. 8.9b). The underestimation could also be caused by an overestimated detrital correction factor ð f0 ¼ 1Þ. The fen peat from the section at Hamburg–Isfeldstraße is restricted suitable for uranium-series dating. TIMS 230Th/U measurements are required to re-study this age underestimation more precisely, as TIMS measurements offer the opportunity of smaller sample size than it is necessary for alpha spectrometry.

within the t7-Lower Terrace of the River Rhine, which is considered to be of Upper Pleistocene age (Schweiss, 1988). The gravel pit (Fig. 8.10) was explored by several drillings down to about 27.50 m below surface (Fig. 8.10). The sand and gravel is excavated down to about 22 m below surface by suction excavators. Mammal bones and wood remains are intercalated in fine-grained sediments between 11 and 15 m below surface. Wood pieces from Picea and/or Larix at 15 m depth gave a radiocarbon age of 32 500 620 BP (Hv 12531), indicating that this horizon correlates very likely with MIS 3. At a depth between 17 and 21 m below surface, sediment layers rich in wood fragments (pieces of oak with 30–60 cm diameter) and mammal remains are intercalated. The mammal remains include Hippopotamus amphibius, Sus scrofa and Cervus dama. Palynological results from the same sediments [m] Lithology 14C Age Groß-Rohrheim BP

The gravel pit KBC is located about 2 km in the south-west of Groß-Rohrheim in the Upper Rhine Graben, Germany (Fig. 8.3). The gravel pit is morphologically situated

Age

[kyr]

0

5

10

15

32,500 ± 620

+28

244 –22

195 ±5.9 +298 263 –74 20

>44,000

+18

229 –15 279

+43 –31

Fine-grained sand Sand Gravel

25

8.4.5 Gravel pit KBC near GroßRohrheim, Upper Rhine Graben (MIS 7)

230Th/U

Silt and peat Bones Wood

Fig. 8.10 Lithological sequence of the section at Groß-Rohrheim in the Upper Rhine Graben (modified after Schweiss, 1988) including 14C ages in years before present (BP) and 230Th/ 234U ages.

Uranium-Series Dating of Peat

8.4.6 Bossel (MIS 9 or older) Interglacial deposits were cored (GE00/1) at the section at Bossel, about 30 km west of Hamburg (Fig. 8.3). The coordinates are 29 139E and 30 209N. The section at Bossel was defined as one of the reference sites of

(a) 120

[230Th]/[232Th]

Groß-Rohrheim

80

40

0 0

80

40

120

[234U]/[232Th] (b)

Groß-Rohrheim 10

2

20 50 100

[234U]/[238U]

indicate a cooler period at the beginning or at the end of an interglacial period. The faunal and floral results indicate a deposition of these sediments during the last interglacial according to von Koenigswald and Beug (1988). A piece of black oak gave a radiocarbon age of > 44 000 BP (Hv 13031). Between 19 and 21 m below surface, peat lumps with diameters of about 15 cm were sampled and investigated by 230Th/234U dating. An isochron age of 214 8 kyr (Chisquare ¼ 12, n ¼ 5) was determined by Geyh et al. (1997). All samples have high uranium content but low thorium content. Reasons for a potential overestimation of the 230Th/234U ages were discussed in detail by Geyh and Hennig (in Schweiss, 1988). Furthermore, Geyh and Techmer (1997) described a loss of uranium in organic-rich sediments from the section at Groß Todtshorn in Lower Saxony, resulting in age overestimation. Loss of uranium from the peat resulting in an age overestimation by percolating groundwater is unlikely, as confirmed by the [234U/238U] activity ratios with a relatively high mean value of 1.2 (Fig. 8.11a, 8.11b; Table 8.2). Furthermore, the 230 Th/234U ages do not scatter in relation to their uranium content, varying by a factor of ten. The coherent results suggest that the true age of the peat should not be much older or younger than the apparent calculated 230Th/234U ages. The peat fragments from the section at Groß-Rohrheim are suitable for uranium-series dating. However, the exact position of the samples and the thickness of the peat layer as well as a complete pollen record are not available, and hence the chronological interpretation remains under discussion. This is part of an ongoing study.

109

200

1.6

300 400 kyr 1.2

0.8 0

0.4

0.8

1.2

1.6

[230Th]/[238U]

Fig. 8.11a Isochron plot showing the activity ratios 230 Th/ 232Th against 234U/ 232Th of the peat samples from Groß-Rohrheim. Fig. 8.11b Osmond–Ivanovich diagram for the activity ratios of the peat lumps from the section at Groß-Rohrheim.

the Holsteinian interglacial (Jerz and Linke, 1987). The sediment sequence (Fig. 8.12) consists of two organic-rich layers below and above marine Holsteinian deposits (Mu¨ller and Ho¨fle, 1994; Geyh and Mu¨ller, 2005). The lower part of the sequence consists of organic-rich, sandy-clayey sediments correlated with the transition period between the late Elsterian and the Holsteinian. In the lower part of the sequence, peat fragments are intercalated in sandy-clayey

110

Manfred Frechen et al.

[m]

Lithology Pollen 230Th/U age Bossel zone [kyr]

Saalian

16.50

17.70 17.90

XVI XV

18.60

XIII– XIVa Holsteinian

19.60

Corrected age not determined

20.10

VII?– XII

298 38.20 38.50

I–VI

310 290

Elsterian

288

+4 –3 +20 –17 ±6 +21 –18

Wadden sediments Silty sand Fine-grained sand Sand Gravel sand

40.00

Gravel

41.00

Calcerous mud Sandy peat

42.00

Peat

Fig. 8.12 Lithological sequence of the section at Bossel (modified after Mu¨ller and Ho¨fle, 1994) including pollen zones and 230Th/ U ages in kyr (¼ 1000 years). The upper peat layer was sampled between 19.10 and 19.30 m, the lower peat layer between 38.30 and 38.40 m.

sediments correlating with the onset of the Holsteinian interglacial. These deposits are covered by clayey to sandy silt and about 1.5 m thick calcareous gyttja. The gyttja is covered by 0.30 m thick fen peat. TIMS samples were taken from the fen peat at a depth between 38.30 and 38.40 m below surface. This peat is covered by approximately 16.2 m of compressed marine silty clay with plant remains. The upper part contains silty to sandy sediment layers including broken marine shells. A peat layer about 1 m thick is intercalated in the sediment between 19.6 and 18.6 m below surface. The lower part of the organic-rich sediments consists of fen peat and was sampled between 19.10 and 19.30 m below

surface. The peat is covered by reworked sandy ombrogenic peat. Radiometric 230Th/234U ages of peat from the section at Bossel yielded indefinite age results. A correlation with MIS 7 or older was not possible owing to the large standard deviation of the [230Th/232Th] and [234U/232Th] activity ratios (Geyh et al., 1997). TIMS measurements were carried out to test the suitability of this material for dating peat older than MIS 7. The [234U/238U] activity ratios of four samples taken from the fen peat at a depth of about 38 m below surface give evidence that the single TIMS 230Th/U ages are contemporaneous within their standard deviation (Fig. 8.13a–d). The analytical results are presented in Table 8.2. An isochron age of 296þ13 11 kyr was calculated with a detrital correction f0 of 0.674 (Fig. 8.13a). Six samples were taken from fen peat at a depth of 19 m below surface. The [234U/238U] activity ratios of two samples (TIMS 431 and 433) show large scattering, and hence these samples were excluded from further calculations (Table 8.2). The four samples used for further data processing gave a negative correlation for the detrital correction factor (Fig. 8.13c). Therefore, a detrital correction factor of f0 ¼ 0 was applied for calculating the single TIMS 230Th/U ages, resulting in an isochron age of 312 kyr. The Osmond–Ivanovich plot of detritus-corrected isotope ratios, [234U/238U] versus [230Th/238U] (Fig. 8.13b, 8.13d), shows that a closed-system behaviour is highly likely and that this dating assumption is fulfilled. The upper organic-rich horizon yielded a minimum age, so the fen peat correlates at least with MIS 9 or any older MIS. The lower humic-rich horizon correlates numerically with MIS 9, which is in agreement with the interpretation of Geyh and Mu¨ller (2005). It is statistically not acceptable to combine the results of the upper and lower humic-rich horizon owing to the large scattering of the [234U/238U] versus [230Th/238U] activity ratios. If we ignore our concern, a

Uranium-Series Dating of Peat (a)

111

(b)

Bossel (38 m)

Bossel (38 m)

80

10

2

20 50

[234U]/[238U]

[230Th]/[232Th]

100 60

40

200

1.6

300 400 kyr 1.2

20

0.8

0 0

20

40

60

0

80

0.4

[234U]/[232Th]

0.8

1.2

1.6

[230Th]/[238U] (d)

(c) Bossel (19 m)

10

2

1.2

Bossel (19 m) 20 50

[234U]/[238U]

[230Th]/[232Th]

100 0.8

300 400 kyr

0.4

1.2

0

0.8 0

0.4

0.8

1.2

200

1.6

0

0.4

0.8

1.2

1.6

[230Th]/[238U]

[234U]/[232Th]

Fig. 8.13a Isochron plot showing the activity ratios 230Th/ 232Th against 234U/ 232Th of the peat samples at a depth between 38.30 and 38.40 m from the Bossel section. Fig. 8.13b Osmond–Ivanovich diagram for the activity ratios of the peat at a depth between 38.30 and 38.40 m from the Bossel section. Fig. 8.13c Isochron plot showing the activity ratios 230Th/ 232Th against 234U/ 232Th of the peat samples at a depth between 19.10 and 19.30 m from the Bossel section. The intercept of the isochron and the y-axis is negative; therefore a detrital correction is not possible. Fig. 8.13d Osmond–Ivanovich diagram for the uncorrected activity ratios of the peat at a depth between 19.10 and 19.30 m from the Bossel section.

hypothetical weighted mean age of 297 31 kyr is determined, which also correlates with MIS 9. The lower fen peat from the section at Bossel seems to be suitable for TIMS 230Th/U dating. However, at present it persists under discussion whether the peat remained a closed system since deposition at about 300 kyr.

8.5 DISCUSSION A detailed discussion of the standard set of specification for the presentation of 230Th/U data in scientific publication is given by Ludwig (2003). Only the results from scientific publications are discussed in this study, as minimum data is provided to reproduce

112

Manfred Frechen et al.

the calculations. These are tables with isotope ratios, errors actually used for age and isochron calculations, standard deviation of isotope ratios or isochron methods and the necessary geological and palynological background information. Central and Northern Europe experienced peat formation during Middle and Upper Pleistocene interstadial and interglacial periods. Several uranium-series dating approaches have been carried out during the past 15 years to study the suitability of peat for uranium-series dating and to set up a more reliable chronological frame. Independent age control was provided for Eemian, Lower and Middle Weichselian peat layers by radiocarbon dating and luminescence dating methods. Palynological results and geological estimates were available for most of the peat layers studied. In Switzerland, three lignite horizons are intercalated in last glacial sediments at the section at Gossau. The lignites yielded radiometric 230Th/234U ages ranging from 49:1 3:7 to 34:7 4 kyr BP. These results are in excellent agreement with radiocarbon ages and luminescence age estimates (Geyh and Schlu¨chter, 1998; Preusser et al., 2003). The chronological study indicates that uranium-series dating of compressed peat ð¼ ligniteÞ correlating with MIS 3 is in principle possible. Heijnis (1992) studied the Chelford peat layer in Cheshire, England, which correlates palynologically with MIS 5a. Radiocarbon dating yielded an age > 63 000 BP, and luminescence dating of the overlying and underlying sediments gave age estimates of about 80 and 98 kyr, respectively. The corrected radiometric 230Th/234U age of 86þ26 21 kyr has a large uncertainty but is within the error in agreement with the palynological results and the independent age control. A sandy peat about 30 cm thick was studied from the section at Ta˚sjo¨ in Sweden by Heijnis (1992). Radiocarbon dating yielded a minimum age of 54 000 BP. The pollen spectra are not very specific, indicating a rather cold continental subarctic climate

correlating most likely with early parts of MIS 4. The radiometric corrected 230 Th/234U age gave 75 6 kyr and confirms the correlation with MIS 5/4 transition. At the section at Gro¨bern in SaxonyAnhalt, a very detailed Eemian and Lower Weichselian record is exposed in a brown coal pit. Fen peat is intercalated in the organic-rich sediments correlating with the Brørup interstadial (MIS 5c) by means of geological estimates, palynology and preliminary luminescence age estimates (Litt, 1990; Wansa and Wimmer, 1990; Eissmann and Litt, 1994; Krbetschek and Stolz, 1994). Radiometric 230Th/234U and TIMS 230Th/U dating showed that this fen peat is not suitable for uranium-series dating owing to open-system behaviour. The lignite from the section at Zell in Switzerland correlates with MIS 5c according to geological estimates and palynological investigations (Ku¨ttel, 1989). Radiometric measurements yielded an 230 Th/234U age of 95 3 kyr (Geyh et al., 1997), which is in excellent agreement with a TIMS 230Th/U age of 91 2 kyr determined by the isochron method (this study). The fluvial sediments sandwiching the lignite gave luminescence age estimates ranging from 124 27 to 91 9 kyr (Preusser et al., 2001). Although the luminescence age estimates show large uncertainties, they are within their error in agreement with the uranium-series ages confirming the correlation with MIS 5c. Heijnis and van der Plicht (1992) investigated peat from Allt Odhar located about 16 km south-east of Inverness in Scotland. The palynological results indicate an interstadial record (Walker, 1990). The radiocarbon ages gave minimum ages of > 51 000 BP. The peat yielded a corrected radiometric 230Th/234U age of 106þ11 10 kyr, and hence a correlation with MIS 5c is most likely. Heijnis (1992) studied also two sections from Finland to further test the suitability of uranium-series dating of peat. At the section at Tervola located in southern Finnish

Uranium-Series Dating of Peat

Lapland, an interstadial peat about 25 cm thick was investigated. A minimum radiocarbon age of > 48 000 BP and luminescence age estimates of 120 and 70 kyr for the underlying sediments and about 50 kyr for the overlying sediments were reported. A radiometric 230Th/234U age of 80–100 kyr was deduced from the isochron plot. A precise calculation was hampered by too large scattering in the isochron plot. According to Heijnis (1992), the peat layer from Tervola likely correlates with MIS 5c (or MIS 5a). A similar interpretation was found for a peat layer from the section at Oulainen located in the Gulf of Bothnia, Finland (Heijnis, 1992). Independent age control was provided by minimum radiocarbon ages of > 48 000 and > 63 000 BP, as well as by luminescence age estimates of 97 18 kyr and 150 30 kyr for the overlying and underlying sediments, respectively. A 230Th/234U age of 80–100 kyr was deduced from the slopes of the isochron plots, suggesting a correlation with MIS 5c (or MIS 5a). The pollen spectra and geological estimates of the sediment sequence from the section at Hamburg–Isfeldstraße indicate a correlation with MIS 5e. A mean 230Th/234U age of 103 5 kyr was calculated for fen peat and fen peat-like material from the lower part of the sequence in a previous study (Geyh et al., 1997). Since the detailed palynological investigation of Linke and Hallik (in press), it became necessary to re-calculate the 230Th/234U age from fen peat including material from the transition zone between fen peat and gyttja only. Depending on the different approaches for the determination of the detrital correction factor, the 230Th/234U ages range from 104 10 kyr to 94 9 kyr. These uraniumseries ages indicate age underestimation according to palynological and geological estimates. Reasons for the age underestimation could be open-system behaviour for uranium or an insufficient correction for detrital contamination. Systematic TIMS 230 Th/U dating is required to set up a more reliable chronological frame for the organic-

113

rich sequence from the section at HamburgIsfeldstraße. Heijnis et al. (1993) studied peat up to 30 cm thick deposited on top of beach sediments, at the Fenit section in Ireland. These sediments correlate with a period of high sealevel during the climate maximum of the last interglacial. Radiometric 230 Th/234U dating yielded a corrected age of 118þ9 8 kyr, which coincides with the termination of the last interglacial (MIS 5e), a result which is in good agreement with the palynological results and the geological estimates. In Greece, a sequence of thick peat layers intercalated by gyttja and lake marl from the Tenaghi Phillippon site was investigated by Heijnis and van der Plicht (1992) following the research of van der Wijk et al. (1986). A corrected radiometric 230 Th/234U age of 122þ15 14 kyr was calculated for the whole dataset. Deposition of peat during MIS 5e is likely to agree with palynological studies. To summarise, under certain circumstances peat is suitable for uranium-series dating of the time period covering MIS 5. Unfortunately, there is no independent age control available for organic-rich deposits including fen peat deposited prior to the last interglacial. Stratigraphic control is provided by geological estimates and pollen data only. There are only a few studies focussing on peat from Central and Northern Europe that correlate with the antepenultimate interglacial or older interglacials. At the section at Groß-Rohrheim located in the northern Upper Rhine Graben, peat fragments were studied by uranium-series dating. A radiometric 230Th/234U age of 214 8 kyr was determined (Geyh et al., 1997). Independent age control is not available. Furthermore, the geological estimates, as well as the floral and faunal remains, do not allow a precise chronostratigraphical interpretation. At the open-cast mine Scho¨ningen in eastern Lower Saxony, Germany, the Scho¨ningen interglacial and a second

114

Manfred Frechen et al.

interglacial, the Reinsdorf Interglacial described by Urban (1995, 1999), are intercalated between the Holsteinian deposits and the Drenthe till. Peaty deposits sampled from the Reinsdorf interglacial (Cycle II-1) were studied from the section at Scho¨ningen site 12. The chronostratigraphic position of the Reinsdorf interglacial is under discussion owing to the lack of clear botanical comparables in Northern Germany so far (Urban, this volume). Recently Mu¨ller (pers. comm., 2005) described two interglacials at the section at ‘Nachtigall’ located in Lower Saxony, which were found to be younger than the Holsteinian. The pollen record of the older interglacial from the section at Nachtigall is rich in Tilia and characterised by a late Abies phase immediately following the Holsteinian and the Early Saalian interstadials described in Scho¨ningen (Urban, 1999) and therefore has certain similarities with the Reinsdorf interglacial. The second subsequent interglacial is rich in Alnus and Pinus, but Abies is not present in the pollen spectra. These pollen spectra resemble those of the Scho¨ningen interglacial to a great extent (Urban, this volume). The uraniumseries dating approach gave preliminary uncorrected single 230Th/234U ages of 227 and 180 kyr (Heijnis, 1992) for peat of the Scho¨ningen interglacial. Postdepositional uptake of uranium would result in age underestimation. A necessary detrital correction would result in younger age estimates. The present dataset does not allow a more precise interpretation, as already outlined by Heijnis (1992). TIMS 230Th/U dating of peat taken from the organic-rich deposits of the Reinsdorf sequence comprising the Reinsdorf interglacial and two following interstadials about 10 m thick are currently in progress and might provide further tools for the correlation of interglacials younger than the Holsteinian interglacial. The section at Bossel was defined as one of the reference sites of the Holsteinian interglacial (Jerz and Linke, 1987). In a previous study, Geyh and Mu¨ller (2005)

determined a detritus-corrected mean age of 312 3 kyr for the lower peat layer and a detritus-corrected mean age of 327þ130 37 kyr for the upper peat layer. In this study, only TIMS data were considered for age calculations. The upper peat layer yielded a minimum age of 312 kyr. A detrital correction could not be applied owing to the negative intersection of the isochron with the y-axis. However, any detrital correction would result in younger 230 Th/U ages for this upper peat layer. The lower peat yielded an isochron age of 296þ13 11 kyr correlating numerically with MIS 9. In water-saturated sediments with moderate hydraulic gradients like those from the Bossel section, a relatively closedsystem behaviour is very likely, as evidenced by consistent uranium and thorium concentrations and isotope ratios in samples from the same unit. However, at present, it remains under discussion whether 230Th/U ages > 300 kyr are considered to be absolute ages or minimum ages. A minimal amount of uranium mobility cannot be excluded to occur after burial for such long time periods. A very detailed 230Th/234U dating approach was carried out on Middle Pleistocene interglacial peat deposits from the section at Tottenhill Quarry in Norfolk, England (Rowe et al., 1997). The pollen record of the peat layers from Tottenhill Quarry indicates palynological similarities with the Hoxnian stage (Walker, 1990). A sequence of consistent ages was determined through the peat profile resulting in a mean age of 317 14 kyr, which is consistent with deposition during MIS 9. In this study, Rowe et al. (1997) suggest that uranium uptake confined to a single early diagenetic event, which is thought to be coeval with peat formation. At the section at Tottenhill Quarry, the uranium uptake is interpreted as a result of diagenetic uranium enrichment of the peat by percolating groundwater rising from below. The coherence of the uranium-series ages over a wide range of uranium concentrations clearly suggests that uranium uptake was a rapid and

Uranium-Series Dating of Peat

unique event in the early period of peat formation. Rowe et al. (1997) demonstrated that the peat from Tottenhill Quarry has remained a closed system with respect to uranium mobilisation over a time period of 300 kyr and suggest that peat might be suitable for uranium-series dating. The TIMS 230Th/U dating studies of peat from the sections at Bossel and Tottenhill Quarry confirm the suitability of certain peat layers to set up a more reliable chronological frame and hence demonstrate the potential of dating interglacial deposits prior to MIS 5. However, more systematic and methodological studies are required. 8.6 CONCLUSION Certain interstadial and interglacial peat layers are suitable for absolute dating by uranium-series dating methods. The suitability of the peat layers depends on two essential assumptions: a closed-system behaviour excluding uranium migration after deposition and the contamination of peat by varying amounts of detrital thorium, which has to be corrected owing to its essential effect on calculating 230Th/U ages. Theoretically, the dating range of TIMS 230 Th/U ranges from about 1000 to 500 000 years. The best organic-rich sediment for 230 Th/U dating is fen peat and lignite. The radiometric and TIMS dating results are in good agreement with certain fen peat layers correlating with MIS 3 and 5, as confirmed by independent age control through radiocarbon and luminescence dating methods. There are only a few dating approaches about pre-Eemian peat that require a very careful interpretation. Independent age control is not available for the pre-Eemian age range at present. However, some encouraging refinements of the TIMS dating technique offer the exciting prospect of dating interstadial and interglacial peat deposits and other organic deposits that currently lie beyond the range of the radiocarbon technique. Further, systematic methodological

115

investigations are essential to test whether peat deposits formed prior to the last interglacial are suitable for uranium-series dating. ACKNOWLEDGEMENT At the GGA-Institut, the radiometric uranium-series dating facilities using alpha spectrometry was set up by Gerd J. Hennig in the 1980s and replaced by TIMS 230Th/U dating facilities in 1997 by Deniz Oezen, both under the supervision of the former head of the section Mebus A. Geyh. We appreciate the help of Thomas Litt during fieldwork at the section at Gro¨bern, Gerhard Linke and Astrid Techmer for providing unpublished material, Mebus A. Geyh for stimulating ideas, Gudrun Drewes and Sabine Mogwitz for the processing of the samples and Juliane Herrmann for doing the art work. Thanks to Detlev Degering and Augusto Mangini for valuable comments that improved the manuscript.

REFERENCES Bourdon, B., Henderson, G.M., Lundstrom, C.C., Turner, S.P., 2003. Introduction to U-series geochemistry. Reviews in Mineralogy and Geochemistry 52, 1–21. Cheng, H., Edwards, R.L., Hoff, J., Gallup, C.D., Richards, D.A., Asmeron, Y., 2000. The half-lives of uranium-234 and thorium-230. Chemical Geology 169, 17–33. Edwards, R.L., Gallup, C.D., Cheng, H., 2003. Uranium-series dating of marine and lacustrine carbonates. Reviews in Mineralogy and Geochemistry 52, 363–405. Eissmann, L., Litt, T., 1994. Klassische Quarta¨rfolge Mitteldeutschlands von der Elstereiszeit bis zum Holoza¨n unter besonderer Beru¨cksichtigung der Stratigraphie, Pala¨oo¨kologie und Vorgeschichte (Exkursion B1). Altenburger naturwissenschaftliche Forschungen 7, 250–356. Frechen, M., Oches, E.A., Kohfeld, K.E., 2003. Loess in Europe – mass accumulation rates during the Last Glacial Period. Quaternary Science Reviews 22, 1835–1857. Geyh, M.A., 2001. Reflections on the 230Th/U dating of dirty material. Geochronometria 20, 9–14.

116

Manfred Frechen et al.

Geyh, M.A., Techmer, A., 1997. 230Th/234U-Datierung der organogenen Sedimente der Bohrung Groß Todtshorn (Kr. Harburg; Niedersachsen). Schriftenreihe der Deutschen Geologischen Gesellschaft 4, 103–110. Geyh, M.A., Schlu¨chter, C., 1998. Zur Kalibration der 14 C-Zeitskala vor 22.000 Jahren v.h. GeoArchaeoRhein 2, 139–149. Geyh, M.A., Mu¨ller, H., 2005. Numerical 230Th/U dating and a palynological review of the Holsteinian/Hoxnian interglacial. Quaternary Science Reviews 24, 1861–1872. Geyh, M.A., Hennig, G., Oezen, D., 1997. U/ThDatierung interglazialer und interstadialer Niedermoortorfe und Lignite – Stand und Zukunft. Schriftenreihe der Deutschen Geologischen Gesellschaft 4, 187–199. Goldstein, S.J., Stirling, C.H., 2003. Techniques for measuring uranium-series nuclides: 1992–2002. Reviews in Mineralogy and Geochemistry 52, 23–57. Heijnis, H., 1992. Uranium/thorium dating of Late Pleistocene peat deposits in N.W. Europe. Ph.D. Thesis, Rijksuniversiteit Groningen, The Netherlands, 149 pp. Heijnis, H., van der Plicht, J., 1992. Uranium/thorium dating of Late Pleistocene peat deposits in NW Europe, uranium/thorium isotope systematics and open-system behaviour of peat layers. Chemical Geology 94, 161–171. Heijnis, H., Ruddock, J., Coxon, P., 1993. A uraniumthorium dated Late Eemian or Early Midlandian organic deposit from near Kilfenora between Spa and Fenit, Ireland. Journal of Quaternary Science 8, 31–43. Hennig, G.J., 1979. Beitra¨ge zur 230Th/234UAltersbestimmung von Ho¨hlensintern sowie ein Vergleich der erzielten Ergebnisse mit denen anderer Absolutdatierungsmethoden. Ph.D. Thesis, University of Cologne, Germany, 173 pp. Ivanovich, M., Harmon, R.S., 1992. Uranium-Series Disequilibrium. Clarendon Press, Oxford. Jerz, H., Linke, G., 1987. Arbeitsergebnisse der Subkommission Quarta¨rstratigraphie. Typusregion des Holstein-Interglazials. Eiszeitalter und Gegenwart 37, 145–148. Kaufman, A., 1971. U-series dating of Dead Sea basin carbonates. Geochimica et Cosmochimica Acta 35, 1269–1281. Kaufman, A., 1993. An evaluation of several methods for determining 230Th/U ages in impure carbonates. Geochimica et Cosmochimica Acta 57, 2303–2317. Kaufman, A., Broecker, W., 1965. Comparison of 14C and 230Th ages for carbonate minerals from lakes Lahontan and Bonneville. Journal of Geophysical Research 70, 4039–4054. von, W. Koenigswald, Beug, H.J., 1988. Schlussbetrachtungen. In: Koenigswald, W. von (Ed.), Zur

Pala¨oklimatologie des letzten Interglazials im Nordteil der Oberrheinebene. Pala¨oklimaforschung 4, 321–327. Krbetschek, M.R., Stolz, W., 1994. LumineszenzDatierungen an pleistoza¨nen Sedimenten aus Tagebauen des mitteldeutschen und Lausitzer Braunkohlereviers. Altenburger Naturwissenschaftliche Forschungen 7, 289–295. Krbetschek, M.R., Degering, D. (this volume). Dating of interglacial sediments by luminescence methods. In: Sirocko, F., Litt, T., Claussen, M. (Eds.), The Climate of Past Interglacials, Development in Paleoenvironmental Research, Elsevier. Ku, T.L., Liang, Z.C., 1984. The dating of impure carbonates with decay-series isotopes. Nuclear Instruments and Methods in Physics Research 223, 563–571. Ku¨ttel, M., 1989. Jungpleistoza¨n-Stratigraphie der Zentralschweiz. In: Rose, J., Schlu¨chter, C. (Eds.), Quaternary Type Sections: Imagination or Reality. Balkema, Rotterdam, 179–191. Litt, T., 1990. Pollenanalytische Untersuchungen zur Vegetations- und Klimaentwicklung wa¨hrend des Jungpleistoza¨ns in den Becken von Gro¨bern und Grabschu¨tz. Altenburger naturwissenschaftliche Forschungen 5, 92–105. Linke, G., Hallik, R. (2006). Das Eem-Vorkommen Hamburg-Isfeldstraße (Pleistoza¨n; Nordwestdeutschland). Geologie, Absolutdatierung, Pollenanalyse und Meeresspeegel-Bezug Verhandlungen haturwissenschaftlicher Verein Hamburg (NF) 42, 181–226. Ludwig, K.R., 2003. Mathematical–statistical treatment of data and errors for 230Th/U geochronology. Reviews in Mineralogy and Geochemistry 52, 631–656. Luo, S., Ku, T.L., 1991. U-series isochron dating: a generalized method employing total sample dissolution. Geochimica et Cosmochimica Acta 55, 555–564. Mallick, R., Frank, N., 2002. A new technique for precise uranium-series dating of travertine microsamples. Geochimica et Cosmochimica Acta 66, 4261–4272. Menke, B., Tynni, R., 1984. Das Eeminterglazial und das Weichselfru¨hglazial von Rederstall/Dithmarschen und ihre Bedeutung fu¨r die mitteleuropa¨ische Jungpleistoza¨n-Gliederung. Geologisches Jahrbuch, Reihe A, 76, 1–124. Mu¨ller, H., Ho¨fle, H.C., 1994. Das Holstein-Interglazialvorkommen bei Bossel westlich von Stade und Wanho¨den no¨rdlich Bremerhaven. Geologisches Jahrbuch, Reihe A, 134, 71–116. Osmond, J.K., Ivanovich, M., 1992. Uranium-series mobilization and surface hydrology. In: Ivanovich, M., Harmon, R.S. (Eds.), Uranium-Series Disequilibrium. Clarendon Press, Oxford, 259–289.

Uranium-Series Dating of Peat Osmond, J.K., May, J.P., Tanner, W.F., 1970. Age of the Cape Kennedy barrier-and-lagoon complex. Journal of Geophysical Research 75, 469–479. Preusser, F., Mu¨ller, B.U., Schlu¨chter, C., 2001. Luminescence dating of sediments from the Luthern Valley, Central Switzerland, and implications for the chronology of the last glacial cycle. Quaternary Research 55, 215–222. Preusser, F., Geyh, M.A., Schlu¨chter, C., 2003. Timing of Late Pleistocene climate change in lowland Switzerland. Quaternary Science Reviews 22, 1435–1445. Richards, D.A., Dorale, J.A., 2003. Uranium-series chronology and environmental applications of speleothems. Reviews in Mineralogy and Geochemistry 52, 407–460. Rowe, P.J., Richards, D.A., Atkinson, T.C., Bottrell, S.H., Cliff, R.A., 1997. Geochemistry and radiometric dating of a Middle Pleistocene peat. Geochimica et Cosmochimica Acta 61, 4201–4211. Schlu¨chter, C., Maisch, M., Suter, J., Fitze, P., Keller, W.A., Burga, C.A., Wynistorf, E., 1987. Das Schieferkohlen-Profil von Gossau (Kanton Zu¨rich) und seine stratigraphische Stellung innerhalb der letzten Eiszeit. Vierteljahresschrift der Naturforschenden Gesellschaft in Zu¨rich 132, 135–172. Schwarcz, H.P., Latham, A.G., 1989. Dirty calcites: 1. uranium series dating of contaminated calcite using leachates alone. Chemical Geology 80, 35–43. Schweiss, D., 1988. Jungpleistoza¨ne Sedimentation in der no¨rdlichen Oberrheinebene. In: Koenigswald, W. von (Ed.), Zur Pala¨oklimatologie des letzten Interglazials im Nordteil der Oberrheinebene. Pala¨oklimaforschung 4, 19–78. Szalay, A., 1958. The significance of humus in the geochemical enrichment of uranium. Proceedings

117

of the 2nd United Nations International Conference on the peaceful uses of Atomic Energy 2, 182–186. Urban, B., 1995. Palynological evidence of younger Middle Pleistocene Interglacials (Holsteinian, Reinsdorf, Scho¨ningen) in the Scho¨ningen open cast lignite mine (eastern Lower Saxony/Germany). Mededelingen Rijks Geologische Dienst 52, 175–186. Urban, B., 1999. Middle and Late Pleistocene biostratigraphy and paleoclimate of an open-pit coal mine Scho¨ningen, Germany. Chinese Science Bulletin 44, 30–37. Urban, B. (this volume). Interglacial pollen records from Scho¨ningen, North Germany. In: Sirocko, F., Litt, T., Claussen, M. (Eds.), The Climate of Past Interglacials, Elsevier. Vogel, J.C., Kronfeld, J., 1980. A new method for dating peat. South African Journal of Science 76, 557–558. Walker, M.L.C., 1990. The Allt Odhar interstadial site, Moy, Invernes-shire: results of pollen analyses. In: Auton, C.A., Firth, C.R., Merritt, J.W. (Eds.), Beauty to Nairn. Quaternary Research Association, field guide, 70–72. Wansa, S., Wimmer, R., 1990. Geologie des Jungpleistoza¨ns der Becken von Gro¨bern und Grabschu¨tz. Altenburger naturwissenschaftliche Forschungen 5, 49–91. van der, A. Wijk, El-Daoushy, F., Arends, A.R., Mook, W.G., 1986. Dating peat with U/Th disequilibrium: some geochemical considerations. Chemical Geology 59, 283–292. Yliruokanen, I., 1980. The occurrence of uranium in some Finnish peat bogs. Kemia Kemi 4, 213–217.

This page intentionally left blank

9. U-Redistribution in Fossil Reef Corals from Barbados, West Indies, and Sea-Level Reconstruction for MIS 6.5 Denis Scholz1, Augusto Mangini1 and Dieter Meischner2 1

Heidelberger Akademie der Wissenschaften, Im Neuenheimer Feld 229, 69120 Heidelberg, Germany 2 Am Weendelsgraben 6, 37077 Go¨ttingen, Germany

ABSTRACT Here, we present U-series data of a large number of fossil reef coral subsamples of the species Acropora palmata collected on Barbados. These corals grew during Marine Isotope Stage (MIS) 6.5 (175 000 years before present (BP)) and show substantial variations in their (234 U/238 U) and (230 Th/238 U) activity ratios and also U concentration. Detailed investigation of different subsamples ð> 4Þ from single coral specimens reveals that the assumptions of the presently available open-system models are not fulfilled. We show that some subsamples have gained U that was lost by nearby corals and term this diagenetic scenario U-redistribution. Small amounts of U-redistribution can only be detected by an analysis of a large number of samples as it is done in this study. Because U-redistribution results in significantly wrong coral ages, it may have consequences for the precise determination of the timing and duration of past interglacials. In this study, strictly reliable ages are selected by rejecting samples that suffered U-redistribution and application of the usual reliability criteria. These ages suggest that MIS 6.5 sea level was between 50 11 and 47 11 m relative to the present sea level from 176 100 2800 to 168 900 1400 years BP.

9.1 INTRODUCTION Fossil reef corals provide direct measurements of past sea-level fluctuations (Mesolella

et al., 1969; Fairbanks, 1989) and can be used as palaeoclimatic archives with seasonal resolution (Felis et al., 2004). An essential requirement for all applications is the determination of precise absolute coral ages. For corals older than 30 thousand years (kyr), the most frequently used method is 230Th/U dating, and the age errors resulting from the analytical precision of mass spectrometric measurements are between several per mil for Holocene and 1% for last interglacial corals (Edwards et al., 2003). Corals incorporate 234U and 238U exactly in the ratio as they occur in modern seawater (Edwards et al., 2003) which has a (234U/238U) activity ratio of 1:150 0:003, constant in both space and time (Delanghe et al., 2002). Several studies have shown that seawater (234U/238U) should not have varied more than 0:01– 0.02 compared to its modern value during the last several hundred kyr (Hamelin et al., 1991; Richter and Turekian, 1993; Robinson et al., 2004). Therefore, the initial (234U/238U) activity ratios of fossil reef corals should reproduce modern seawater (234U/238U) within error if the corals have behaved as closed systems. However, many fossil reef corals that have been exposed to meteoric waters for much of their postdepositional history display elevated initial 234 U values4 (i.e. 234 Uinit: values higher than modern-day seawater) 4 234 U ¼ f½ð234 U=238 UÞ=ð234 U=238 UÞeq 1g 103 . ð234 U=238 UÞeq is the atomic ratio at secular equilibrium and is equal to 238 =234 ¼ 5:489 105 , where 238 and 234 are the decay constants for 238U and 234U, respectively. 234 U is the measured value, the initial value is given by 234 Uinit: ¼ 234 U expð234 tÞ, where t is the age in years.

120

Denis Scholz, Augusto Mangini and Dieter Meischner

which cannot be explained by changes in seawater 234 U. The isotopic anomalies observed in fossil reef corals are assumed to result from postdepositional diagenetic alteration (i.e. postdepositional addition/loss/exchange of Th and/or U isotopes, Bar-Matthews et al., 1993; Fruijtier et al., 2000). Because the U–Th system seems to be more sensitive to diagenetic change than any other petrographic or geochemical parameter (Chen et al. 1991), it is not possible to preliminarily sort out disturbed corals on the basis of mineralogy and/or major or trace elementconcentration. Thus, the accuracy of 230Th/U ages of fossil reef corals is more limited due to diagenetic effects than analytical precision (Bard 1.20

100 kyr closedsystem isochron

a

g

gain and 234U gain b 238U loss or

b

c c

f

230 Th

gain or U loss

234U

loss 234 U and 238U gain e with δ234U < δ234Ucoral 230 f Th loss or U gain 234 U and 238U gain g with δ234U > δ234Ucoral d

1.10

e Seawater evolution curve

234U

230 Th

a

1.15

(234 U/238 U)

et al., 1992; Stirling et al., 1995). Only coral ages that fulfil several criteria are believed to be strictly reliable (Stirling et al., 1998). Many scenarios were proposed to explain changes in U-series isotopic composition with diagenetic alteration (Ku et al., 1990; Hamelin et al., 1991; Cheng et al., 1998; Fruijtier et al., 2000), but for a long time validation of one specific model by data has not been possible. Figure 9.1 illustrates the general effects of different diagenetic processes on a (234U/238U) versus (230Th/238U) plot. It must be emphasised that Fig. 9.1 is rather illustrative. The effects of all processes depend strongly on the timing (i.e. shortly after coral growth or later), the duration and the magnitude of the diagenetic

d

1.05

1.00 0.0

0.2

0.4

0.6

0.8

1.0

(230 Th/238 U)

Fig. 9.1 Illustration of the effects of different diagenetic processes on the isotopic composition of the coral. The straight curve represents the seawater evolution curve that describes the temporal development of the activity ratios measured in modern seawater under closed-system conditions. The intersection point of all arrows corresponds to a coral sample that behaved as a closed system for 100 kyr. Different diagenetic processes are indicated by different arrow colours. The dotted straight line is the 100 kyr closed-system isochron that indicates all possible combinations of (230Th/ 238U) and (234 U/ 238U) activity ratios, yielding a Th / U-age of 100 kyr. Diagenetic processes which shift the activity ratios to the right of this line produce apparent ‘older’ coral ages, the others produce apparent ‘younger’ coral ages. Most diagenetically altered fossil reef corals show elevated d234 U init: values (i.e. plot above the seawater evolution curve, Bard et al., 1991; Gallup et al., 1994; Fruijtier et al., 2000). This suggests that these corals were altered by one of the processes from group 1 or 3 (Table 9.1). In addition, several authors illustrated an empirical relationship between elevated d234 U values and higher Th/U ages (Stein et al., 1993; Zhu et al., 1993; Fruijtier et al., 2000) and also (230Th/238U) (Gallup et al., 1994; Stirling et al., 1998; Thompson et al., 2003; Villemant and Feuillet, 2003; Scholz et al., 2004). This is an indication for a process from group 1 (Table 9.1). While process (c) produces only slightly elevated d234 U values, the development of the activity ratios in scenario (b) strongly depends on the ratio of the added 234U and 230Th.

U-Redistribution in Fossil Reef Corals

event. Also, it must be considered that the diagenetic process may be either continuous or episodic. If two or more isotopes are involved (see processes (b), (e) and (g) in Fig. 9.1), the resulting change in the isotopic composition of the affected coral depends strongly on the ratio between the added/ lost isotopes. Finally, two or more of the described processes may proceed simultaneously or successively. Figure 9.1 allows dividing the processes into different groups, according to their effect on the isotopic composition of the coral. These groups are shown in Table 9.1. Recently, three studies suggesting correction approaches for diagenetically altered fossil reef corals were published (Thompson et al., 2003; Villemant and Feuillet, 2003; Scholz et al., 2004). The models of Thompson et al. (2003) and Villemant and Feuillet (2003) propose that elevated 234 Uinit: values are produced by continuous addition of particle-reactive 234Th and 230Th which is produced by decay of dissolved U and -recoil mobilisation of U daughters. 234Th is the initial decay product of 238U and has a 24.1-day half-life. Because 234Th rapidly decays to 234U, the processes proposed by Thompson et al. (2003) and Villemant and Feuillet (2003), respectively, result incontinuous addition of 234U and 230Th (scenario (b), Fig. 9.1 and Table 9.1). There are two differences between both models: firstly, the model of Villemant and Feuillet (2003) takes into account possible initial 230 Th excess and is an isochron approach that requires a suite of coeval samples, Table 9.1 Classification of the diagenetic processes illustrated in Fig. 9.1, according to their effect on coral age and 234 U Group

Effect on coral age

Effect on coral 234 U

1 2 3 4

Older Older Younger Younger

Higher Lower Higher Lower

Processes

b, c d a, g e, f

121

while the Thompson et al. (2003) model can be applied to a single coral (sub) sample; secondly, the model of Villemant and Feuillet (2003) includes the assumption that the addition ratio is related to the (234U/238U) activity ratio of the gaining coral. However, apart from these differences, both models rely on the same redistribution mechanisms and consequently calculate similar ‘open-system ages’. Both models can only be applied if the observed trend agrees with that proposed by the model within error. The model of Scholz et al. (2004) assumes different degrees of U addition and subsequent loss in different subsamples for single coral specimens. This scenario is a combination of processes (g) and (c) (Fig. 9.1) and results in a strong positive linear and correlation between 234 U 230 238 ( Th/ U). The model predicts that the open-system age or the isochron age can be calculated from the intersection point between the trend line (i.e. the isochron) and the seawater evolution curve. The main differences between the models of Thompson et al. (2003) and Villemant and Feuillet (2003), respectively, and the isochron approach of Scholz et al. (2004) are that the former suggest that (i) only Th is added to the corals and (ii) the slope of the trend line is model dependent, while Scholz et al. (2004) assume that (i) only U is mobilised (i.e. added and/or lost), and (ii) the isochron slope is obtained by a linear fit (i.e. depends on the specific diagenetic history of the analysed coral). It speaks in favour of the approach of Scholz et al. (2004) that the trends observed at different localities worldwide are not the same (Edwards et al., 2003). However, because the activity ratios measured in corals from a given terrace are not all co-linear within analytical error (Edwards et al., 2003) and the mechanisms of U addition and loss have remained unclear so far (Scholz et al., 2004), all three approaches have some weaknesses.

Denis Scholz, Augusto Mangini and Dieter Meischner

Limestone

O an ce

Foundation

Scotland–Clermont anticline

Cave Hill Fig. 9.1b

9.2.1 Study region and coral samples All samples presented in this study were collected at an outcrop below the University of the West Indies (UWI) which lies in the western part of Barbados (Fig. 9.2). The outcrop is a road cut section at the University Drive which is part of the Clermont Nose traverse (Mesolella et al., 1969). A detailed description of the limestone stratigraphy at the study region with corresponding figures is given by Speed and Cheng (2004). Fossil corals from this location were also investigated by U-series methods in other studies (Gallup et al., 2002; Thompson et al., 2003; Speed and Cheng, 2004). According to the notation of Gallup et al. (2002), the corals analysed in this study stem from the lower unit of the outcrop which is located between 30 and 35 m above sea level (asl). This unit contains corals that grew during MIS 6.5, a substage, within the glacial period MIS 6, which corresponds to the northern summer insolation peak around 175 kyr (Gallup et al., 2002). At present, the

Limestone es trough

St. Georg

Church arch Christ

5 km

9.2 MATERIAL AND METHODS

ic nt la At

Here, we present a detailed study of fossil corals collected on Barbados, West Indies, including U-series measurements on more than 50 subsamples. All samples grew during marine isotope stage (MIS) 6.5, a period corresponding to the Northern summer insolation peak at 175 kyr. Our results indicate that (i) many samples have gained additional U (Uadd) after deposition and (ii) that the source of the Uadd were coeval corals from the same reef. Under these specific conditions, none of the existing models can be applied to determine correct ages. We show here that the study of different subsamples ð> 4Þ from single corals is an additional criterion to detect altered corals and to identify reliable ages. Based on the new data, the magnitude, timing and duration of the MIS 6.5, sea-level peak are determined.

Caribean Ocean

122

13°00′N

59°30′W

Fig. 9.2 Barbados island. The red circle indicates the study region. All analysed samples stem from an outcrop at the University Drive which belongs to the Cave Hill area that is part of the Clermont Nose traverse. The figure was modified from Speed and Cheng (2004).

outcrop below the UWI is the only location worldwide where MIS 6.5 corals have been found. Figure 9.3 shows a photograph of the MIS 6.5 reef. Locations of coral sampling are indicated. The coral samples have been obtained in the form of drill cores using hollow drills ranging from 1.5 to 5 cm in diameter and a length of 25 cm (Fig. 9.4). All sampled corals are of the species Acropora palmata. One purpose of this study was to explore the diagenetic systems. To compare the isotopic variations within the reef on different scales, two locations (BB02-4 and BB02-5) were sampled (Fig. 9.3). At each location, four drill cores were taken within a distance of 20 to 30 cm to investigate the variations within different subsamples from the same reef, the same location and the same drill

U-Redistribution in Fossil Reef Corals

BB02-5(4 drillcores)

BB02-4(4 cores)

Fig. 9.3 Photograph of the MIS 6.5 reef deposit at the University Drive roadcut on Barbados. In the notation of Gallup et al. (2002), the MIS 6.5 reef corresponds to the lower unit (between 30 and 40 m a.s.l.). The white circles indicate the two locations where the coral samples were taken (i.e. BB02-4 and BB02-5). At each location, four drill cores were obtained.

core. Drill cores BB02-4-1 and -2 were obtained from one single coral specimen, as were BB02-4-3 and -4. 9.2.2 U-series measurements In contrast to other studies, we analysed not only the best preserved parts of the drill cores but also portions showing signs of diagenetic alteration because we wanted to

Fig. 9.4 Photograph of drill core BB02-5-1 (Acropora palmata coral). Each subsample is between 1 and 3 mm thick. Even surface areas (e.g. at the ends of pieces B and F) indicate where previous samples have been taken. Rough surface areas (e.g. at the other ends of pieces B and F) indicate natural sites of fracture or that the core broke during drilling.

123

investigate the effects of coral diagenesis on the U-series isotopic system. The aragonite content of some subsamples was measured by X-ray diffraction (XRD) analysis (Table 9.2). In addition, each drill core was examined under ultraviolet light to identify parts that are mineralogically altered. In contrast to the other subsamples that showed no special colour, those from drill core BB02-54 appeared whitish under ultraviolet light. Because the XRD analyses of the BB02-5-4 subsamples confirmed that substantial portions of this drill core consist of calcite (Table 9.2), this indicates that the white portions in the drill cores represent the calcite fraction. All other subsamples showed no white portions, indicating that these drill cores have negligible calcite content. Several subsamples were sawn from each drill core and analysed by U-series methods (Fig. 9.4). All measurements were performed using thermal ionisation mass spectrometry (TIMS). Details of sample preparation and analysis are described elsewhere (Scholz et al., 2004; Scholz, 2005). 9.3 RESULTS The U-series data of all analysed subsamples are listed in Table 9.2. The 238U concentrations of the samples collected at the two locations (i.e. BB02-4 and BB02-5, Fig. 9.3) show substantial differences. While the 238U contents of the samples from location BB024 lie between 3.07 and 3.38 ppm (mean value: 3:23 0:08 ppm, 1-error, n ¼ 19, Table 9.2), the samples from location BB025-1, -2 and -3 have higher 238U concentrations (3.38–4.07 ppm, mean value: 3:69 0:18 ppm, 1-error, n ¼ 33, Table 9.2). The U concentrations of the subsamples of drill core BB02-5-4 range from 1.52 to 3.17 ppm and are clearly lower than those of the other cores (Table 9.2). The mean U concentration observed in living Acropora palmata corals is 3:24 0:2 ppm (1-error, n ¼ 16, Cross and Cross, 1983). This indicates that the BB02-5-1, -2 and -3

124

Denis Scholz, Augusto Mangini and Dieter Meischner

Table 9.2 Samplea

Th (ppb)b

238

U (ppm)

234 U [‰]c

234 Uinit: (‰)d

(230Th/238U)

230

Th/U age (kyr)e

Aragonite content (%)

BB02-4-1 BB02-4-1-B BB02-4-1-C BB02-4-1-D BB02-4-1-B-li BB02-4-1-C-li BB02-4-1-C-re

0.19 0.58 0.14 0.42 0.25 0.16 0.24

3.23 3.21 3.38 3.22 3.2 3.3 3.23

101´ 1 101´ 3 96´ 2 99´ 3 98´ 1 96´ 1 101´ 1

164´ 2 165´ 4 157´ 3 161´ 4 160´ 2 158´ 2 164´ 2

0.889´ 0.003 0.893´ 0.006 0.893´ 0.005 0.887´ 0.004 0.891´ 0.003 0.893´ 0.004 0.891´ 0.004

171.6´ 1.6 173.2´ 3.2 175.6´ 2.4 171.6´ 2.3 174´ 1.4 175.9´ 1.8 172.7´ 1.7

– – – – – – –

BB02-4-2 BB02-4-2-B1 BB02-4-2-C1 BB02-4-2-D1

0.3 0.73 0.45 0.41

3.07 3.08 3.11 3.12

105´ 3 102´ 1 97´ 1 91´ 4

168´ 4 167´ 2 158´ 2 149´ 7

0.879´ 0.006 0.897´ 0.003 0.889´ 0.002 0.889´ 0.004

165.6´ 2.9 174.7´ 1.4 173.7´ 1.2 176.1´ 2.8

– – – >99

BB02-4-3 BB02-4-3-A1 BB02-4-3-A2 BB02-4-3-A3

0.29 0.04 0.32 0.32

3.24 3.3 3.31 3.29

96´ 2 98´ 2 96´ 1 94´ 1

155´ 2 156´ 3 154´ 1 152´ 2

0.878´ 0.003 0.869´ 0.008 0.88´ 0.002 0.881´ 0.002

168.9´ 1.5 164´ 3.4 169.8´ 0.9 170.9´ 1.2

– – 98 99

BB02-4-4 BB02-4-4-A BB02-4-4-C1 BB02-4-4-C2

1.61 0.48 0.55 0.81

3.27 3.2 3.32 3.25

102´ 2 101´ 2 100´ 1 97´ 4

164´ 2 165´ 4 162´ 2 156´ 6

0.88´ 0.003 0.896´ 0.004 0.884´ 0.002 0.878´ 0.008

167.3´ 1.4 174.8´ 2 169.6´ 1.1 168.5´ 3.8

– – – 93

BB02-5-1 BB02-5-1-D BB02-5-1-C-re BB02-5-1-E BB02-5-1-2 BB02-5-1-B BB02-5-1-F BB02-5-1-C-li BB02-5-1-3

0.15 0.32 0.1 0.16 0.17 0.31 0.37 0.15 0.21

4.07 3.99 3.99 3.97 3.53 3.64 3.89 3.62 3.65

99´ 1 97´ 2 97´ 1 95´ 1 102´ 3 92´ 1 94´ 1 98´ 1 92´ 2

149´ 2 146´ 3 149´ 2 147´ 2 159´ 4 143´ 2 148´ 2 153´ 2 153´ 4

0.818´ 0.003 0.825´ 0.004 0.833´ 0.003 0.837´ 0.002 0.857´ 0.006 0.845´ 0.004 0.856´ 0.002 0.853´ 0.004 0.896´ 0.008

143.9´ 1.1 147.2´ 1.6 150´ 1.1 152.2´ 1 157.5´ 2.7 156.6´ 1.6 159.9´ 1.1 157.7´ 1.6 178.9´ 4.4

– – – – – – – – –

BB02-5-2 BB02-5-2-3 BB02-5-2-E1 BB02-5-2-E2 BB02-5-2-2 BB02-5-2-C1 BB02-5-2-C2 BB02-5-2-C3 BB02-5-2-A1 BB02-5-2-A2 BB02-5-2-A3

–* 0.05 0.13 0.21 0.19 –* 0.06 0.02 –* –* 0.12

3.77 3.9 3.82 3.86 3.81 3.76 3.74 3.73 3.77 3.81 3.79

93´ 1 100´ 3 97´ 2 96´ 2 91´ 1 92´ 2 93´ 2 96´ 2 95´ 2 91´ 2 92´ 2

141´ 2 150´ 4 147´ 2 144´ 2 140´ 2 141´ 2 142´ 3 147´ 2 152´ 3 144´ 2 146´ 2

0.82´ 0.002 0.819´ 0.003 0.824´ 0.002 0.819´ 0.003 0.832´ 0.002 0.834´ 0.004 0.832´ 0.005 0.833´ 0.004 0.865´ 0.004 0.86´ 0.003 0.864´ 0.003

146.7´ 0.9 144.1´ 1.3 146.7´ 1 145.2´ 1.2 151.4´ 1 152.3´ 1.8 151.1´ 2.1 150.3´ 1.6 163.5´ 1.9 163´ 1.5 164.4´ 1.6

– – – – – – – – – – –

BB02-5-3 BB02-5-3-2 BB02-5-3-3 BB02-5-3-A BB02-5-3-B BB02-5-3-C BB02-5-3-D BB02-5-3-E BB02-5-3-F BB02-5-3-A-II

0.19 0.02 0.07 0.24 0.29 0.27 0.22 0.07 0.22 1.4

3.58 3.46 3.6 3.52 3.52 3.38 3.56 3.46 3.54 3.48

98´ 2 99´ 2 95´ 2 93´ 2 94´ 2 94´ 1 97´ 3 105´ 3 112´ 2 96´ 1

155´ 2 158´ 2 152´ 3 148´ 4 150´ 3 150´ 2 154´ 4 168´ 4 173´ 4 154´ 2

0.865´ 0.004 0.867´ 0.005 0.873´ 0.004 0.869´ 0.003 0.871´ 0.003 0.867´ 0.002 0.867´ 0.003 0.877´ 0.003 0.858´ 0.009 0.878´ 0.006

162.6´ 1.7 162.8´ 2 167.1´ 1.9 166.2´ 1.7 166.7´ 1.4 164.9´ 1 163.4´ 1.8 164.3´ 1.7 154.7´ 3.8 168.9´ 2.7

– – – – – – – – – 99

232

U-Redistribution in Fossil Reef Corals

125

Table 9.2 continued 232

238

Samplea

Th (ppb)b

U (ppm)

234 U [‰]c

234 Uinit: (‰)d

(230Th/238U)

230

Th/U age (kyr)e

Aragonite content (%)

BB02-5-3-C-II BB02-5-3-D-II BB02-5-3-E-II

1.44 1.16 0.56

3.5 3.59 3.52

96´ 1 99´ 1 96´ 1

153´ 2 157´ 2 155´ 2

0.874´ 0.002 0.871´ 0.002 0.876´ 0.002

167.3´ 1.1 164.7´ 1 167.6´ 1

98 98 97

BB02-5-4 BB02-5-4-B-W BB02-5-4-B-R BB02-5-4-B-M

0.01 0.01 0.03 0.02

1.87 1.52 3.17 2.17

86´ 2 86´ 2 94´ 2 95´ 3

165´ 3 186´ 3 151´ 3 160´ 6

0.976´ 0.003 1.022´ 0.004 0.876´ 0.002 0.911´ 0.013

231.9´ 3 275.2´ 4.7 168.8´ 1.4 185.2´ 7.4

– 10 30 60

a All samples are corals of the species Acropora palmata. Data from different drill cores are divided by a blank line. All uncertainties are quoted at 2-level. b For some subsamples (indicated by an asterisk), the resulting 232Th concentration was lower than the reproducible blank. c 234 U ¼ {[(234U/238U)/(234U/238U)eq] 1}*103. (234U/238U)eq is the atomic ratio at secular equilibrium and is equal to 238/234¼5.489* 10-5 where 238 and 234 are the decay constants for 238U and 234U, respectively. d 234 Uinit. ¼ 234U* exp(234T), where T is the age 230in years. 234 e 230 Th/U ages are calculated iteratively using: 238Th ¼ 1 expð230 TÞ þ 1000U 230 ð1 expðð230 234 ÞTÞÞ. T is the age in U act 234 230 years and 230 is the decay constant for Th. 238 ¼ 1:55125 1010 y1 ; 234 ¼ 2:8263 106 y1 ; 230 ¼ 9:1577 106 y1 as reported by Cheng et al. (2000). All activity ratios were corrected for contamination by detrital 238U, 234U and 230Th, assuming a silicate source with a 232Th/238U ratio of 3.8, and with 230Th, 234U and 238U in secular equilibrium. Because of the very low 232Th contents, the resulting differences between uncorrected and corrected ages are generally within the 2 error range of the ages. The uncertainty in the age includes the analytical contribution from 234 U and (230Th/238U).

subsamples have gained Uadd after deposition, and those from drill core BB02-5-4 have postdepositionally lost U. The 232Th contents of all subsamples are low (mean value: 0.31 ppb, Table 9.2). This indicates that no detrital Th (232Th as well as 230 Th) was incorporated during growth or postdepositionally added to these corals. The U-series ages display a rather large range from 143:9 1:1 to 275:2 4:7 kyr (Table 9.2). The ages from location BB02-4 range from 164 3:4 to 176:1 2:8 kyr (Table 9.2). The ages from locations BB025-1, -2 and -3 are generally lower, ranging from 143:9 1:1 to 178:9 4:4 kyr, and those from location BB02-5-4 are older, ranging from 168:8 1:4 to 275:2 4:7 kyr. This shows that the differences in the U concentration of the samples are also reflected in the 230Th/U-ages (Table 9.2). Aragonite concentrations were only measured for selected samples (Table 9.2). The subsamples from drill core BB02-5-4 have low aragonite concentrations ranging from 10 to 60%, indicating that substantial portions of the subsamples recrystallised from aragonite to calcite. Except for one sample (BB02-4-

4-C2), the aragonite content of all other analysed subsamples is greater than 97% (Table 9.2), suggesting no or minor recrystallisation. The U-series results of the analysed subsamples are shown on a 234 U 230 238 versus ( Th/ U) plot in Figures 9.5 and 9.6 which demonstrate that many samples plot on the seawater evolution curve. However, 22 samples plot significantly above the seawater evolution curve, indicating open-system behaviour. This is also manifested in their elevated 234 Uinit: values (Table 9.2). Interestingly, five subsamples exhibit 234 Uinit: value lower than modern seawater (Table 9.2). This was also reported for a few other corals from Barbados (Potter et al., 2004). In summary, the 234 Uinit: values range from 140 2 to 173 4‰.

9.4 DISCUSSION 9.4.1 234 U–(230Th/238U) correlations All drill cores were investigated extensively, and between four and thirteen subsamples within each drill core were analysed

126

Denis Scholz, Augusto Mangini and Dieter Meischner 120

(b)

BB02-4-1

120

115

115

110

110

105

105

δ 234U [‰]

δ 234U [‰]

(a)

100 95

100 95

90

90

85

85

80 0.80

0.85

0.90

0.95

1.00

80 0.80

1.05

BB02-4-2

0.85

(230Th/238U)

(c)

0.90

0.95

120

(d)

120 115

110

110

105

105

δ 234U [‰]

δ 234U [‰]

115

100 95

100 95

90

90

85

85

0.85

0.90

0.95

1.00

80 0.80

1.05

0.85

(230Th/238U) 120

(f)

BB02-4-1 & -4-2

110

110

105

105

100 95

95 90

85

85

0.95

(230Th/238U)

1.05

100

90

0.90

1.00

BB02-4-3 & -4-4

115

0.85

0.95

120

115

80 0.80

0.90

(230Th/238U)

δ 234U [‰]

δ 234U [‰]

(e)

1.05

BB02-4-4

BB02-4-3

80 0.80

1.00

(230Th/238U)

1.00

1.05

80 0.80

0.85

0.90

0.95

1.00

1.05

(230Th/238U)

Fig. 9.5 (A–F) d 234 U versus (230Th/ 238U) plots for the four drill cores from location BB02-4. Because drill cores BB02-4-1 and -2 as well as drill cores BB02-4-3 and -4 were obtained from the same coral specimen, the data from both drill cores are also combined in a single plot (Fig. 9.5E and 9.5F). Dotted lines around the seawater evolution curve represent the 5‰ uncertainty range. The red straight lines are EWLS regressions of the data that were obtained using ISOPLOT (Ludwig, 2003b). The black and blue straight lines are the addition lines proposed by the Thompson et al. (2003) model. The black addition lines were calculated assuming that the true age of the coral corresponds to the activity ratios at the intersection point of the isochron and the seawater evolution curve. For the blue lines, the mean value of the open-system ages (Thompson et al., 2003) was used as the true coral age. We note that not all plotted subsamples fulfil the applicability criteria of Thompson et al. (2003), but it is necessary to plot all subsamples to investigate whether the isochron dating method (Scholz et al., 2004) can be applied. Maybe the differences between the observed slopes and those predicted by the model partly arise from data points that do not meet the Thompson et al. (2003) criteria.

U-Redistribution in Fossil Reef Corals (b)

BB02-5-1

120

115

115

110

110

105

105

δ 234U [‰]

δ 234U [‰]

(a) 120

100 95

95 90

85

85

0.85

0.90

1.00

0.95

BB02-5-2

100

90

80 0.80

80 0.80

1.05

0.85

(230Th/238U) (d)

BB02-5-3

115

110

110

105

105

100 95

85

85

(

1.05

0.95

1.00

BB02-5-4

95 90

0.90

1.00

100

90

0.85

0.95

120

115

80 0.80

0.90

(230Th/238U)

δ 234U [‰]

δ 234U [‰]

(c) 120

127

80 0.80

1.05

0.85

230Th/238U)

0.90

(

0.95

1.00

1.05

230Th/238U)

(e) 120 115

δ 234U [‰]

110 105 100 95

Literature data

90 85 80 0.80

0.85

0.90

0.95

1.00

1.05

(230Th/238U)

Fig. 9.6 (A–E) d234 U versus (230Th/ 238U) plots for the drill cores from location BB02-5. Also shown are published coral data from the same outcrop (Gallup et al., 2002; Speed and Cheng, 2004, Fig. 9.6E). We note that not all plotted subsamples fulfil the applicability criteria of Thompson et al. (2003).

(Table 9.2). This should provide enough data points to detect linear correlations between 234 U and (230Th/238U) as predicted by the diagenetic models (Thompson et al., 2003; Villemant and Feuillet, 2003; Scholz et al., 2004).

Figures 9.5 and 9.6 show 234 U versus (230Th/238U) plots for each drill core. Because drill cores BB02-4-1 and -2 were obtained from a single coral, as were drill cores BB024-3 and -4, the data for these cores are also combined on single plots (Fig. 9.5E and 9.5F).

128

Denis Scholz, Augusto Mangini and Dieter Meischner

Also shown is a 234 U versus (230Th/238U) plot for published data from the MIS 6.5 reef below the UWI (Gallup et al., 2002; Speed and Cheng, 2004). The isochrons (red straight lines) were obtained by an error-weighted least-square (EWLS) regression using ISOPLOT (Ludwig, 2003b). ISOPLOT performs the EWLS algorithm of York (1969). This algorithm weights the data points according to their analytical x- and y-errors and also takes into account error correlations. To test if the isochron assumptions (Ludwig, 2003a, b) are fulfilled, the Probability of Fit was calculated for all isochrons (Table 9.3). While the data for some drill cores show medium-to-high correlations (BB02-4-1, BB02-4-3, BB02-5-4 and published data), those for other ones result in rather low correlation coefficients (Table 9.3). This shows that the samples from the MIS 6.5 reef on Barbados do not generally display a clear trend between 234 U and (230Th/238U), as suggested by the diagenetic models (Thompson et al., 2003; Villemant

and Feuillet, 2003; Scholz et al., 2004). This is also confirmed by the Probability of Fit values which are only for two corals higher than 15% (BB02-4-1 and BB02-4-3, Table 9.3). The calculated isochron slopes display a very large range of positive as well as negative values (Table 9.3). However, because the 95% confidence errors for the slopes are very large (ranging from 92 to 990%, Table 9.3), these differences are not significant. The main reasons for the large uncertainties are (i) data points scatter from the isochrons and (ii) small data point distances (Figs. 9.5 and 9.6). Thus, it is problematic to investigate the diagenetic processes that affected our corals on 234 U versus (230Th/238U) plots. The Thompson et al. (2003) model proposes addition line slopes of 335‰ for MIS 6.5 corals (326.7‰ for a true age of 150 kyr and 348.1‰ for 200 kyr). For illustration, these addition lines are also shown in Figs. 9.5 and 9.6 (black and blue straight lines, see figure captions for details). Six of the isochron slopes are

Table 9.3 Compilation of the fitting results for both the analysed drill cores and published data from the same outcrop (Gallup et al., 2002; Speed and Cheng, 2004) Drill core

Number of analysed subsamples

Correlation coefficient

Isochron slope (‰) 1493´ 1600 2625´ 26000 487´ 450 43´ 590 8207´ 57000 847´ 1200

Isochron intercept (‰) 1429´ 1500 2240´ 23000 524´ 390 63´ 530 7209´ 51000 649´ 1000

BB02-4-1 BB02-4-2 BB02-4-3 BB02-4-4 BB02-4-1 & -4-2 BB02-4-3 & -4-4

7 4 4 4 11 8

0.61 0.26 0.81 0.1 0.29 0.34

BB02-5-1 BB02-5-2 BB02-5-3 BB02-5-4

9 11 13 4

0.51 0.47 0.19 0.97

Literature data

12

0.68

0.44 0 0.16 0.05 0 0

97´ 120 39´ 110 2882´ 6500 64´ 80

177´ 100 125´ 90 2415´ 5600 151´ 77

0 0 0 0.02

130´ 150

0

260´ 170

Probability of Fit*

Isochron slopes and intercepts were obtained using ISOPLOT (Ludwig, 2003b). All quoted uncertainties are 95% confidence errors. * The Probability of Fit is the probability that, if the only reason for scatter from a straight line is the analytical errors assigned to the data points, the scatter of the data points will exceed the amount observed for the data (Ludwig, 2003b). In other words, the Probability of Fit tests if the measured data points and errors can be described by the assumed model (in this case if the data points plot on an isochron). If the Probability of Fit is higher than 15%, the model assumptions are considered justified (Ludwig, 2003b), and the 95% confidence errors of the isochron slope and intercept are calculated from the data point analytical errors, if not, they are calculated from the observed scatter and appropriately enlarged (see Ludwig, 2003b, for details).The low Probabilities of Fit indicate that only the data for drill cores BB02-4-1 and -3 fulfil the isochron assumptions, while the data for all other drill cores do not plot on an isochron within the assigned analytical data point errors. The large range in the isochron slopes suggests that samples from different drill cores were affected by different diagenetic processes.

U-Redistribution in Fossil Reef Corals

consistent with the value proposed by the Thompson et al. (2003) model within their 95% confidence errors (Table 9.3). However, any agreement (disagreement) arising from this comparison should not be interpreted as evidence for (against) the Thompson et al. (2003) model assumptions, because of the very large slope uncertainties associated with all isochrons (Table 9.3). Interestingly, the plot for published data from the same reef shows the best agreement with the addition line proposed by the Thompson et al. (2003) model (Fig. 9.6E, Table 9.3), suggesting the applicability of the model. It is important to note that these Useries data are not from one coral specimen, but from different corals from the MIS 6.5 reef. Therefore, these corals are not as strictly comparable as the drill core data because they are not necessarily coeval. This may be a reason for the observed scatter along the isochron (Fig. 9.6E). In addition, the resulting isochron slope error is rather large (65%, Table 9.3). This makes it generally difficult to prove/disprove the validity of the Thompson et al. (2003) model assumptions. Subsamples from drill cores BB02-5-1, -2 and -3 (Fig. 9.6A–C) have generally higher 238 U concentrations than those from location BB02-4. This indicates that the subsamples from these three drill cores gained Uadd after deposition. Because postdepositional Uadd uptake is one of the most important assumptions of the isochron approach (Scholz et al., 2004), these drill cores should be suitable for the application of isochron dating. However, to produce an isochron on a 234 U versus (230Th/238U) plot, the Uadd must have a different value than the coral (processes (e) and (g), Fig. 9.1). Many subsamples of drill cores BB02-5-1 and BB02-5-2 plot on the seawater evolution curve within error but show large variations in their (230Th/238U) activity ratios (Fig. 9.6A and 9.6B) and Useries ages (ranging from 140 to 180 kyr, Table 9.2). Because different subsamples from the same drill core (i.e. the same coral) must have the same age, the large (230Th/238U) variations are very strong

129

evidence that these subsamples are diagenetically altered. This must have been another process than suggested by Thompson et al. (2003) and Villemant and Feuillet (2003) because the isochron slopes for both drill cores are significantly different than proposed by these models (Table 9.3). Because of the large slope/intercept errors and the low Probability of Fit values (Table 9.3), isochron dating (Scholz et al., 2004) can also not be applied to these drill cores. The 234 U versus (230Th/238U) plot for drill core BB02-5-3 (Fig. 9.6C) reveals that only two of the 13 data points plot significantly above the seawater evolution curve. Because the Probability of Fit is 0 and the slope/intercept errors are very large (Table 9.3), it is not possible to decide unequivocally whether the assumptions of any of the diagenetic models are fulfilled. Therefore, a correction with all models should not be performed. The subsamples of drill core BB02-5-4 (Fig. 9.6D) show substantial concentrations of calcite and also significantly lower U concentrations (Table 9.2). This indicates postdepositional U loss due to recrystallisation from aragonite to calcite (Reeder et al., 2000). Because the models of Thompson et al. (2003) and Villemant and Feuillet (2003) do not account for U loss, they cannot be applied. In contrast, the data points plot on an isochron if U loss proceeds simultaneously in all subsamples (process (c), Fig. 9.1), and, in this case, isochron dating can be applied (Scholz et al., 2004). The correlation coefficient for this drill core is high (R ¼ 0:97, Table 9.3), but the Probability of Fit value is low (0.02, Table 9.3), and the slope/intercept errors are rather large (Table 9.3). Because the assumption of simultaneous U loss cannot be tested, isochron dating should not be applied to this drill core. In summary, the discussion above shows that one cannot unequivocally declare that the assumptions of one of the three diagenetic models (Thompson et al., 2003; Villemant and Feuillet, 2003; Scholz et al., 2004) are fulfilled.

130

Denis Scholz, Augusto Mangini and Dieter Meischner

Therefore, none of the models is applied to the MIS 6.5 corals. 9.4.2 Further investigation of the diagenetic processes As mentioned above, the 238U contents of the subsamples from location BB02-4 are in the usual range about 3 ppm (ranging from 3.07 to 3.38 ppm, Table 9.2), while those from location BB02-5 show substantial variations. It has been shown for Porites corals from Aqaba, Jordan, that postdepositional Uadd uptake produced substantially elevated 234 U values (Scholz et al., 2004). Even though the 234 U values of the Barbados corals are not elevated to the same extent, at least some of them plot significantly above the seawater evolution curve. This raises the question whether the elevated 234 U values of the Barbados corals were produced by postdepositional U uptake. Figure 9.7 shows the data of all subsamples except those of drill core BB02-5-4 on a plot of 234 U versus 238U content. It is evident that the 238U content of the BB02-4 samples (empty circles) is generally lower

than that of the subsamples from drill cores BB02-5-1, -2 and -3 (filled squares). Based on the usual U range for Acropora palmata corals, this suggests that the subsamples from drill cores BB02-5-1, -2 and -3 have postdepositionally gained Uadd. However, there is no trend between 234 U and 238U content detectable, suggesting that the Uadd had approximately the same 234 U value as the corals. Because a basic requirement for isochron dating is that the Uadd has a different (i.e. significantly higher or lower) 234 U value (Scholz et al., 2004), isochron dating cannot be applied. Figure 9.8 shows a plot of (230Th/238U) versus 238U concentration, which clearly illustrates that the subsamples from drill cores BB02-5-1, -2 and -3 gained Uadd. While there is no trend observable in the BB02-4 data (open circles), the BB02-5-1, -2 and -3 data (filled symbols) show a clear trend towards lower (230Th/238U) activity ratios with higher 238 U concentration. This confirms that the elevated 238U concentrations of the BB02-5 samples have been produced by postdepositional U addition because otherwise – if higher amounts of U had been incorporated during 0.92

115 0.90 110

(230Th/238U)

0.88

δ 234U [‰]

105

100

0.86

0.84

95 0.82 90 0.80 3.0

85 3.0

3.2

3.4 238U

3.6

3.8

4.0

3.2

3.4 238U

3.6

3.8

4.0

[ppm]

[ppm]

Fig. 9.7 Plot of d234 U versus 238U concentration for the subsamples from location BB02-4 (shown as empty circles) and drill cores BB02-5-1, -2 and -3 (shown as filled squares). The data of drill core BB025-4 are not shown because these samples have lower 238 U contents.

Fig. 9.8 Plot of (230Th/ 238U) versus 238U concentration for all analysed subsamples. Samples from location BB02-4 are shown as empty circles. Samples from location BB02-5 are shown separately for each drill core: BB02-5-1 as filled squares, BB02-5-2 as filled triangles and BB02-5-3 as filled circles. Data from drill core BB02-5-4 are not shown.

U-Redistribution in Fossil Reef Corals

coral growth – all subsamples would have different U concentration but the same (230Th/238U) activity ratio. Figures 9.7 and 9.8 show that (i) the subsamples from drill cores BB02-5-1, -2 and -3 gained substantial Uadd and (ii) the Uadd had approximately the same 234 U value as the gaining corals. The latter has been tested by a more detailed estimation of the 234 U value of the additional U component with a two-component mixing-model (Scholz, 2005). Furthermore, Scholz (2005) has shown that the data of drill cores BB02-5-1, -2 and -3 can be explained by U addition alone. This contrasts with the existing open-system models because these assume either subsequent U loss (Scholz et al., 2004) or addition of 230Th (Thompson et al., 2003; Villemant and Feuillet, 2003). Three of four subsamples of drill-core BB02-5-4 have a much lower U content than all other analysed samples (Table 9.2). This indicates that these samples have lost U after deposition, resulting in older 230Th/U ages. Because the aragonite contents of these samples range from 10 to 60% only (Table 9.2), U loss most likely occurred due to recrystallisation from aragonite to calcite (Reeder et al., 2000). Despite its low aragonite content (30%), subsample BB02-5-4-R has a ‘normal’ 238U content of 3.17 ppm (Table 9.2), suggesting that it has not been affected by U loss. The 230Th/U age of this subsample is 168:8 1:4 kyr (Table 9.2), broadly in agreement with the ages of the subsamples from location BB02-4 ð170 kyrÞ. Altogether, these observations show that the subsamples from drill cores BB02-5-1, -2 and -3 have postdepositionally gained U, while those from drill core BB02-5-4 show evidence for U loss. These data support the assumption of Scholz et al. (2004) that U can be postdepositionally mobilised in fossil reef corals. However, as discussed in the previous chapter, the isochron dating assumptions seem not to be fulfilled, and isochron dating cannot be applied. Because the Uadd had approximately the same 234 U

131

value as the gaining corals, we assume that some subsamples have taken up the U which was lost by others due to recrystallisation or dissolution processes. We term this diagenetic scenario U-redistribution. All corals within the MIS 6.5 reef have approximately the same age, and consequently, they all have similar 234 U values. Therefore, the mobilised U has the same 234 U value as the gaining corals. This explains why U addition does not produce elevated 234 U values (Fig. 9.7). Because U-redistribution results in data displacement to the left on a 234 U versus (230Th/238U) diagram (process (f) in Fig. 9.1), this scenario is also adequate to explain coral data which plot significantly below the seawater evolution curve. We conclude that the isotopic anomalies of the samples from location BB02-5 most likely result from U-redistribution within the reef. If one sample within the reef loses U, the released U is taken up by surrounding corals. It is evident from Figs. 9.7 and 9.8 that postdepositional U-redistribution 230 238 results in lower ( Th/ U) activity ratios, but not in elevated 234 U values. This in turn results in 230Th/U-ages that are significantly underestimated. Therefore, U-redistribution may have major consequences for U-series dating of fossil reef corals and, as a consequence, for the precise determination of the timing and duration of past interglacials. If only a small amount of Uadd was added to a fossil coral, the resulting alteration is difficult to detect by the existing reliability criteria (Stirling et al., 1998; Muhs et al., 2002). Because U-redistribution does not change coral 234 U, it cannot be detected by the 234 Uinit: criterion. Identification by comparison with the U concentration range obtained from living corals of the same coral species is also difficult because this range is normally rather large (i.e. 0:2 ppm for Acropora palmata corals, see above). For example, postdepositional gain of 0:2 ppm Uadd may not be detected by this criterion but would result in significant alteration of both (230Th/238U) and the

132

Denis Scholz, Augusto Mangini and Dieter Meischner

U-series age (strongly depending on the timing of addition). This shows that the U concentration criterion is not sufficient to detect all corals that suffered U-redistribution or uptake, and its application might result in acceptance of significantly wrong 230 Th/U ages of fossil reef corals. Here, we have demonstrated by investigation of a large number of (sub)samples that all subsamples from drill cores BB025-1, -2 and -3 have gained Uadd. If we had used the U concentration criterion (e.g. the 2-range obtained from living Acropora palmata corals, i.e. 3:23 0:4 ppm, see above), 15 of these subsamples would have been assumed to have normal U content (Table 9.2). Therefore, we suggest not to test for U-redistribution or uptake by application of a U-concentration range but, if possible, by analysis of a large number of (sub)samples (as in this study or in Potter et al., 2004) or combined Th/U and Pa/U dating (Edwards et al., 1997; Gallup et al., 2002; Cutler et al., 2003). These tests enable to detect even subtle postdepositional alteration of the U concentration of fossil corals. Thompson et al. (2003) and Thompson and Goldstein (2005) also use common reliability criteria (i.e. U concentrations within the range of modern corals of the same species, < 2 ppb 232Th, < 2% calcite) to sort out corals that were altered by diagenetic processes not described by their model. As explained above, these criteria are not sufficient to detect all samples that suffered U-redistribution. For example, Thompson et al. (2003) and Thompson and Goldstein (2005) accept all Acropora palmata corals with a 238U concentration between 2.64 and 3.84 ppm. Our study shows that all subsamples from drill cores BB02-5-1, -2 and -3 have gained significant Uadd after deposition, albeit many of these subsamples have 238U concentrations lower than 3.84 ppm. Therefore, the models of Thompson et al. (2003) and Villemant and Feuillet (2003) do not perform the appropriate corrections and calculate wrong open-system ages if they are

applied to these subsamples. Because the wide 238U concentration range used by Thompson et al. (2003) and Thompson and Goldstein (2005) is not appropriate to detect all corals which have suffered Uadd uptake, this also raises the question whether U gain, U loss or U-redistribution affected some of the samples that Thompson and Goldstein (2005) used for their sea-level reconstruction. As a rule, all open-system models (Thompson et al., 2003; Villemant and Feuillet, 2003; Scholz et al., 2004) should only be applied to corals which unequivocally have not been affected by other processes than those assumed by these models. 9.5 MIS 6.5 – AN EXTRAORDINARY CLIMATE PERIOD MIS 6.5 is a period of distinctive climate signature which is characterised by an unconventional combination of insolation, ice sheet extent and greenhouse gas concentration (Tzedakis et al., 2003). While northern summer insolation (Berger and Loutre, 1991) and atmospheric 18 O values were at interglacial levels, there was no corresponding increase in atmospheric CO2 (Petit et al., 1999). Based on one strictly reliable coral age from Gallup et al. (2002), sea level was approximately 40 m relative to the present sea level (r.p.s.l.). Several continental records indicate more humid and slightly warmer conditions during MIS 6.5 than during the main part of the glacial period MIS 6 but much colder and drier conditions than during an interglacial (Ayalon et al., 2002; Bard et al., 2002b; Plagnes et al., 2002; Tzedakis et al., 2003). Another outstanding climatic feature of MIS 6.5 is the extreme minimum in the Dole effect (i.e. the natural enrichment of 18 O in the atmosphere with respect to its value in the ocean, Malaize et al., 1999). This minimum has been explained by a decoupling between high and low latitudes (Me´lie`res et al., 1997; Malaize et al., 1999), the former dominated by the presence of ice

U-Redistribution in Fossil Reef Corals

(explaining the observed cold MIS 6.5 climate), the latter directly responding to the insolation increase (Tzedakis et al., 2003). A direct measurement of the magnitude, timing and duration of the MIS 6.5 sea-level peak from coral data should help to understand the complex climatic signature of this extraordinary climate period. 9.6 TIMING, MAGNITUDE AND DURATION OF THE MIS 6.5 SEA-LEVEL PEAK In this chapter, we estimate the sea level during MIS 6.5 based on strictly reliable coral ages and compare the results with other sea-level reconstructions. Strictly reliable ages are selected by the following criteria: (i) no evidence for U uptake, U loss or U-redistribution (on the basis of the results from the previous chapters), (ii) 232Th concentrations less than 2 ppb, (iii) 234 Uinit: ¼ 149 4‰, and (iv) 98% aragonite. Because all subsamples from location BB02-5 were affected by U-redistribution, the following discussion only refers to the subsamples from location BB02-4. All subsamples from location BB02-4 have 232Th concentrations lower than 2 ppb (Table 9.2). Only five of the subsamples from this location have 234 Uinit: values within the range of 149 4‰: BB02-4-2D1, BB02-4-3, BB02-4-3-A2, BB02-4-3-A3 and BB02-4-4-C2 (Table 9.2). The last subsample has only 93% aragonite (Table 9. 2) and is therefore not considered as reliable. Overall, four subsamples with strictly

133

reliable ages ranging from 176 3 to 169 1 kyr remain (Table 9.4). Because the original dataset consisted of 56 subsamples, this highlights the complexity of determining strictly reliable U-series ages for fossil reef corals. Interestingly, the ages of subsamples BB02-4-2-D1 (176:1 2:8 kyr, Table 9.4) and BB02-4-3 (168:9 1:4 kyr, Table 9. 4) are substantially different, though the two drill cores were collected at a distance of 30 cm. However, because the two subsamples do not belong to a single coral specimen, and considering the rather large age error of subsample BB02-4-2-D1, its age is considered as strictly reliable. To estimate the palaeo sea level from the coral data, an uplift correction must be subtracted from the present elevation of the corals (i.e. þ36 m r.p.s.l., Table 9.4). This correction is usually calculated from the coral age and the local uplift rate (Mesolella et al., 1969; Cabioch and Ayliffe, 2001; Gallup et al., 2002). However, the uplift rate for our study region below the UWI is controversial. Taylor and Mann (1991) estimated an uplift rate of 0.44 m/kyr based on the height of the Last Interglacial First High Cliff, and this value was subsequently used by Gallup et al. (2002) for sea-level reconstruction. Recently, Thompson and Goldstein (2005) calculated a value of 0:45 0:01 m=kyr for this site. In contrast, Speed and Cheng (2004) determined an uplift rate of 0:53 0:04 m=kyr based on the present-day elevation of the last interglacial shoreline angle. Finally, Schellmann and Radtke (2004) conclude that the reef terraces below the UWI form warped anticlines, and the

Table 9.4 Sample BB02-4-2-D1 BB02-4-3 BB02-4-3-A2 BB02-4-3-A3

Elevation (m r.p.s.l.) 36´ 3 36´ 3 36´ 3 36´ 3

Age (kyr) 176.1´ 2.8 168.9´ 1.4 169.8´ 0.9 170.9´ 1.2

Palaeo sea level (m r.p.s.l.) 50.3´ 11.1 46.8´ 10.6 47.2´ 10.6 47.7´ 10.7

The uplift rate is estimated as 0:49 0:06 m=kyr (details are given in the text). Palaeo sea-level errors were calculated by linear error propagation. Palaeo sea levels (P) were calculated from P ¼ E U t, where E is the present elevation in metres relative to the present sea level, U is the uplift rate in metres per kyr and t is the coral age in kyr.

134

Denis Scholz, Augusto Mangini and Dieter Meischner

Sea level (m r.p.s.l.)

uplift rates varied spatially and, most probably, temporally and should not be used for sea-level reconstructions. Nevertheless, we calculate MIS 6.5 sea level for this location on Barbados because it is unique, and the coral data from this locality represent the only direct measurement of MIS 6.5 sea level presently available. To account for the uncertainty at this site, we estimate palaeo sea levels (Table 9.4 and Fig. 9.9) using the average of the quoted values (i.e. 0.49 m/ kyr) and a rather large uncertainty of 0.06 m/kyr. We then compare our results with published coral data of Gallup et al. (2002), Speed and Cheng (2004) and Thompson and Goldstein (2005). The strictly reliable coral data from this study (Table 9.4) suggest that MIS 6.5 sea level was between 50 11 and 47 11 m r.p.s.l. from 176 3 to 169 1 kyr (Table 9.4, Fig. 9.9). Timing and duration of the sea-level peak are in good agreement with the corresponding peak in Northern summer insolation that has its maximum value at 174 kyr and the subsequent minimum at 160 kyr (Berger and Loutre, 1991). Even if the lower boundary for the MIS 6.5 sea-level peak is based on only one coral age (BB02-4-2-D1,

Table 9.4) and therefore not very accurately defined, this suggests that insolation changes have been the primary trigger for ice sheet melting before/during MIS 6.5, in accordance with the Milankovitch theory of climate change (Milankovitch, 1941). Figure 9.9 shows a comparison of our sealevel estimation with other reconstructions for MIS 6.5. It is evident that these reconstructions display large differences ranging from values around 20 (Shackleton, 2000) to 70 m r.p.s.l. (Waelbroeck et al., 2002). Gallup et al. (2002) estimated a sea level of 38 5 m r.p.s.l. at 168 1:3 kyr (Fig. 9.9) based on combined Th/U and Pa/U dating of corals from the same outcrop below the UWI. The difference in the magnitude of their sea-level estimate results from the different uplift rate (0.44 m/kyr, see above) they used. However, both estimates agree within error. Recently, Thompson and Goldstein (2005) published a sea-level curve which covers MIS 6.5 based on open-system coral ages (Thompson et al., 2003, filled yellow circles in Fig. 9.9). Their results indicate a MIS 6.5 duration from 172:9 1:8 to 163:5 2:1 kyr, and a magnitude between 38 2 and 46 2 m r.p.s.l. (Fig. 9.9). All MIS 6.5

–10

–10

–20

–20

–30

–30

–40

–40

–50

–50

–60

–60

–70

–70

–80

–80

–90

–90

–100

Coral data (this study) Gallup et al. (2002) Thompson and Goldstein (2005) Specmap (Imbrie et al., 1984) Shackleton (2000) Waelbroeck et al. (2002) Lea et al. (2002) Siddall et al. (2003) Bard et al. (2002)

–100 155

160

165

170

175

180

185

Age (kyr)

Fig. 9.9 Comparison of the calculated MIS 6.5 sea levels with other reconstructions (see figure legend for the respective references). Coral based data are shown as filled squares [black: this study (Table 9.4), red: Gallup et al. (2002)]. The yellow circles are open-system coral data from Thompson and Goldstein (2005). Straight curves are reconstructions based on stable isotopes. Connected filled squares represent data from a submerged speleothem.

U-Redistribution in Fossil Reef Corals

samples they used in their study are published data from the outcrop below the UWI (Gallup et al., 2002; Speed and Cheng, 2004). The resulting differences in palaeo sea-level magnitude arise from the different uplift rate they used. Because of the large uplift rate uncertainty we used for our reconstruction, both MIS 6.5 sea-level estimates agree within error. It is evident from Fig. 9.9 that our reconstruction and that of Thompson and Goldstein (2005) suggest a similar timing for the onset of the MIS 6.5 sea-level peak but a substantially different duration. According to Thompson and Goldstein (2005), the onset was at 172:9 1:8 kyr which is in agreement with our estimate within error (i.e. 176:1 2:9 kyr, Table 9.4). In contrast, our reconstruction suggests a MIS 6.5 duration until 168:9 1:4 kyr (Table 9.4, Fig. 9.9), while they calculated a duration until 163:5 2:1 kyr (Fig. 9.9). We note that a large number of our subsamples also yield ages between 160 and 165 kyr (especially those from drill cores BB02-5-2 and -3, Table 9.2). However, we have demonstrated that all these subsamples have suffered U-redistribution and, therefore, cannot be considered strictly reliable. In addition, we have shown that the Thompson et al. (2003) model cannot be used to correct our coral data. Because the corals used by Thompson and Goldstein (2005) are from the same outcrop, this raises the question whether some of their samples might have suffered U-redistribution or uptake, as some of our subsamples have. Except for two samples, all their MIS 6.5 corals have U concentrations within the 1-range of living Acropora palmata corals, suggesting that these corals have not gained U after deposition. However, as mentioned above, subtle but significant changes in U content can only be detected by analysis of a large number of samples. The connected filled squares (shown in magenta, Fig. 9.9) are based on submerged speleothem data from Argentarola cave, Italy (Bard et al., 2002a). This speleothem

135

was sampled at 18:5 m r.p.s.l. and shows continuous growth between 189:7 1:5 and 145:2 1:1 kyr, indicating that sealevel was not higher than 18:5 m r.p.s.l. during that time, in agreement with our coral data. All other curves shown in Fig. 9.9 are based on stable oxygen isotopes. The SPECMAP curve (Imbrie et al., 1984) was linearly converted to sealevel as a first-order approximation using two fixed points: 120 m r.p.s.l. for the last glacial maximum and þ6 m r.p.s.l. for the last interglacial highstand. This results in maximum values of 65 m r.p.s.l. during MIS 6.5 (Fig. 9.9). The difference to the values around 50 m r.p.s.l. estimated from coral records suggests that the SPECMAP 18 O signal might be affected by deep-sea temperature changes. Thus, it is essential to untangle the deep-sea temperature component from the ice volume component to reliably reconstruct palaeo sealevels from stable isotope signals. The large variations between the different reconstructions (Fig. 9.9) highlight the difficulty of such corrections. The new coral record agrees best with the reconstructions of Lea et al. (2002) and Siddall et al. (2003) which are assumed to be accurate within 24 m and 12 m, respectively (Fig. 9.9). While the reconstruction of Waelbroeck et al. (2002), which is associated with an uncertainty of at least 13 m, results in lower values between 75 and 60 m r.p.s.l. (Fig. 9.9), the estimate of Shackleton (2000) yields much higher sealevels between 20 and 30 m r.p.s.l. (Fig. 9.9). Because the timescales of all stable isotope-based records were derived by orbital tuning, either by correlating to SPECMAP (Lea et al., 2002; Waelbroeck et al., 2002; Siddall et al., 2003) or by more sophisticated methods (Shackleton, 2000), the coral data from this study and the reconstruction from Thompson and Goldstein (2005) represent the only absolutely dated sea-level record for MIS 6.5. Here, the strictly reliable coral ages were very carefully selected, especially with respect to postdepositional U-redistribution. Therefore, these data

136

Denis Scholz, Augusto Mangini and Dieter Meischner

probably represent the most accurate estimate of timing and duration of MIS 6.5 sealevel from corals at present. Figure 9.10 shows the sea-level curve for the last 180 kyr based on strictly reliable coral data (modified from Cutler et al., 2003). The curve is composed of data from various studies (see caption of Fig. 9.10 for details). The MIS 6.5 coral data from this study are included. For the period between the last interglacial highstand and MIS 6.5, no coral data are available. Therefore, it is not possible to give a sea-level estimate for this period, which is indicated by the question mark (Fig. 9.10). Because fossil coral data are the only direct measurement of past sea-level fluctuations and are precisely, absolutely dated by U-series methods, this sea-level curve (Fig. 9.10) represents one of the most accurate records in magnitude as well as in timing for the last 180 kyr which is presently available.

20

Sea level (m r.p.s.l.)

0 –20 –40 –60 –80

?

–100 –120 –140 0

20

40

60

80

100

120

140

160

180

Age (kyr)

Fig. 9.10 Coral-based sea-level curve for the last 180 kyr. The figure has been modified from Cutler et al. (2003) and is composed of strictly reliable coral data from the following references: Bard et al., 1990a, b, 1996; Chen et al., 1991; Edwards et al., 1993, 1997; Ludwig et al., 1996; Toscano and Lundberg, 1999; Gallup et al., 2002; Cutler et al., 2003. The red symbols represent the MIS 6.5 sea-level reconstruction from this study (Fig. 9.9). The question mark between the last interglacial highstand and the MIS 6.5 peak indicates that no coral data from this period are available.

9.7 CONCLUSIONS A large number of Acropora palmata subsamples from an outcrop below the UWI, Barbados, were investigated by U-series methods. Small degrees of U uptake and loss cannot be identified with the common reliability criteria. To identify U-redistribution, we analysed a large number of subsamples of each coral specimen and used the variations in U concentration and activity ratios as a further reliability criterion. All subsamples from location BB02-5 were affected by U-redistribution, and the presently available diagenetic models (Thompson et al., 2003; Villemant and Feuillet, 2003; Scholz et al., 2004) cannot be applied to these coral data. The subsamples from location BB02-4 show less evidence for Uadd uptake/loss. After the application of further reliability criteria, only four strictly reliable ages from an initial set of 56 subsamples remained, highlighting the difficulty of finding fossil coral samples suitable for U-series dating. MIS 6.5 sealevel was estimated on the basis of the strictly reliable coral ages and ranges from 50 11 to 47 11 m r.p.s.l. between 176:1 2:8 and 168:9 1:4 kyr. The timing and duration of the MIS 6.5 highstand are in agreement with the Milankovitch theory. The uncertainty in the uplift rate at the study site below the UWI was accounted for by assuming a rather large error (0:49 0:06 m=kyr). Because of the additional criterion used here for coral data screening, our results deliver probably the most accurate reconstruction in timing and duration for MIS 6.5 from fossil corals presently available. Our reconstruction agrees best with the reconstructions of Lea et al. (2002) and Siddall et al. (2003).

ACKNOWLEDGEMENTS We thank T. Felis and M. Zuther for XRD analysis. Thorough reviews by two anonymous reviewers helped substantially to

U-Redistribution in Fossil Reef Corals

improve the manuscript. We greatly appreciated the comments of W. Thompson which were very helpful. This work was funded by BMBF grant 01 LD 0041 (DEKLIM).

REFERENCES Ayalon, A., Bar-Matthews, M., Kaufman, A., 2002. Climatic conditions during marine oxygen isotope stage 6 in the eastern Mediterranean region from the isotopic composition of speleothems of Soreq Cave, Israel. Geology 30, 303–306. Bar-Matthews, M., Wasserburg, G. J., Chen, J. H., 1993. Diagenesis of fossil coral skeletons: Correlation between trace elements, textures, and 234 U/238U. Geochimica et Cosmochimica Acta 57, 257–276. Bard, E., Hamelin, B., Fairbanks, R. G., 1990a. U–Th ages obtained by mass spectrometry in corals from Barbados: sea level during the past 130,000 years. Nature 346, 456–458. Bard, E., Hamelin, B., Fairbanks, R. G., Zindler, A., 1990b. Calibration of the 14C timescale over the past 30 000 years using mass spectrometric U–Th ages from Barbados corals. Nature 345, 405–410. Bard, E., Fairbanks, R. G., Hamelin, B., Zindler, A., Hoang, C. T., 1991. Uranium-234 anomalies in corals older than 150 000 years. Geochimica et Cosmochimica Acta 55, 2385–2390. Bard, E., Fairbanks, R. G., Hamelin, B., 1992. How accurate are the U–Th ages obtained by mass spectrometry on coral terraces. In: Kukla, G., Went, E. (Eds.), Start of a Glacial. Springer-Verlag, Berlin, pp. 15–21. Bard, E., Hamelin, B., Arnold, M., Montaggioni, L., Cabioch, G., Faure, G., Rougerie, F., 1996. Deglacial sealevel record from Tahiti corals and the timing of global meltwater discharge. Nature 382, 241–244. Bard, E., Antonioli, F., Silenzi, S., 2002a. Sea level during the penultimate interglacial period based on a submerged stalagmite from Argentarola Cave (Italy). Earth and Planetary Science Letters 196, 135–146. Bard, E., Delaygue, G., Rostek, F., Antonioli, F., Silenzi, S., Schrag, D. P., 2002b. Hydrological conditions over the western Mediterranean basin during the deposition of the cold Sapropel 6 (ca. 175 kyr BP). Earth and Planetary Science Letters 202, 481–494. Berger, A., Loutre, M.-F., 1991. Insolation values for the climate of the last 10 million years. Quarternary Science Reviews 10, 297–317.

137

Cabioch, G., Ayliffe, L. K., 2001. Raised coral terraces at Malakula, Vanatu, Southwest Pacific, indicate high sea level during Marine Isotope Stage 3. Quaternary Research 56, 357–365. Chen, J. H., Curran, H. A., White, B., Wasserburg, G. J., 1991. Precise chronology of the last interglacial period: 234U–230Th data from fossil coral reefs in the Bahamas. Geological Society of America Bulletin 103, 82–97. Cheng, H., Edwards, R. L., Murrell, M. T., Benjamin, T.M., 1998. Uranium-thorium-protactinium dating systematics. Geochimica et Cosmochimica Acta 62, 3437–3452. Cheng, H., Edwards, R. L., Hoff, J., Gallup, C.D., Richards, D.A., Asmerom, Y., 2000. The half-lives of uranium-234 and thorium-230. Chemical Geology 169, 17–33. Cross, T. S., Cross, B. W., 1983. U, Sr and Mg in Holocene and Pleistocene corals A. palmata and M. annularis. Journal of Sedimentary Petrology 53, 587–594. Cutler, K. B., Edwards, R. L., Taylor, F. W., Cheng, H., Adkins, J., Gallup, C. D., Cutler, P. M., Burr, G. S., Bloom, A. L., 2003. Rapid sea level fall and deepocean temperature change since the last interglacial period. Earth and Planetary Science Letters 206, 253–271. Delanghe, D., Bard, E., Hamelin, B., 2002. New TIMS constraints on the uranium-238 and uranium-234 in seawaters from the main ocean basins and the Mediterranean Sea. Marine Chemistry 80, 79–93. Edwards, R. L., Beck, J. W., Burr, G.S., Donahue, D. J., Druffel, E.R.M., Taylor, F.W., 1993. A large drop in atmospheric 14C/12C and reduced melting during the Younger Dryas, documented with 230Th ages of corals. Science 260, 962–968. Edwards, R. L., Cheng, H., Murrell, M. T., Goldstein, S. J., 1997. Protactinium-231 dating of carbonates by thermal ionization mass spectrometry: Implications for Quaternary climate change. Science 276, 782–786. Edwards, R. L., Gallup, C. D., Cheng, H., 2003. Uranium-series dating of marine and lacustrine carbonates. In: Bourdon, B., Henderson, G. M., Lundstrom, C. C., Turner, S. P. (Eds.), Uraniumseries Geochemistry. Mineralogical Society of America, pp. 656. Fairbanks, R. G., 1989. A 17,000-year glacio-eustatic sea level record: influence of glacial melting rates on the Younger Dryas event and deep-ocean circulation. Nature 342, 637–642. 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, 164–168.

138

Denis Scholz, Augusto Mangini and Dieter Meischner

Fruijtier, C., Elliott, T., Schlager, W., 2000. Mass-spectrometric 234U–230Th ages from the Key Largo Formation, Florida Keys, United States: Constraints on diagenetic age disturbance. Geological Society of America Bulletin 112 (2), 267–277. Gallup, C. D., Edwards, R. L., Johnson, R. G., 1994. The timing of high sea levels over the past 200 000 years. Science 263, 796–800. Gallup, C. D., Cheng, H., Taylor, F. W., Edwards, R.L., 2002. Direct determination of the timing of sea level change during Termination II. Science 295, 310–313. Hamelin, B., Bard, E., Zindler, A., Fairbanks, R.G., 1991. 234U/238U mass spectrometry of corals: How accurate is the U-Th age of the last interglacial period? Earth and Planetary Science Letters 106, 169–180. Imbrie, J., Hays, J. D., McIntyre, A., Mix, A. C., Morley, J. J., Pisias, N. G., Prell, W.L., Shackleton, N. J., 1984. The orbital theory of Pleistocene climate: Support from a revised chronology of the marine 18 O record. In: Berger, A., Imbrie, J., Hays, J., Kukla, G.J., Saltzman, E. (Eds.), Milankovitch and Climate. D. Reidel, Boston, pp. 269–305. Ku, T.-L., Ivanovich, M., Luo, S., 1990. U-series dating of Last interglacial high sea stands: Barbados revisited. Quaternary Research 33, 129–147. Lea, D. W., Martin, P. A., Pak, D. K., Spero, H. J., 2002. Reconstructing a 350 ky history of sealevel using planctonic Mg/Ca and oxygen isotope records from a Cocos Ridge core. Quaternary Science Reviews 21, 283–293. Ludwig, K. R., 2003a. Mathematical-statistical treatment of data and errors for 230Th/U geochronology. In: Bourdon, B., Henderson, G.M., Lundstrom, C. C., Turner, S.P. (Eds.), Uranium-series Geochemistry. Mineralogical Society of America, pp. 656. Ludwig, K. R., 2003b. User’s Manual for Isoplot 3.00. A geochronological toolkit for Microsoft Excel, Berkeley Geochronology Center Special Publication No. 4, 70 pp. Ludwig, K. R., Muhs, D.R., Simmons, K. R., Halley, R. B., Shinn, E. A., 1996. Sealevel records at 80 ka from tectonically stable platforms: Florida and Bermuda. Geology 24, 211–214. Malaize, B., Paillard, D., Jouzel, J., Raynaud, D., 1999. The Dole effect over the last two glacial–interglacial cycles. Journal of Geophysical Research 104, 14199–14208. Me´lie`res, M.-A., Rossignol-Strick, M., Malaize´, B., 1997. Relation between low latitude insolation and 18 O change of atmospheric oxygen for the last 200 kyrs, as revealed by Mediterranean sapropels. Geophysical Research Letters 24, 1235–1238. Mesolella, K. J., Matthews, R. K., Broecker, W. S., Thurber, D. L., 1969. The astronomical theory of

climatic change: Barbados Data. Journal of Geology 77, 250–274. Milankovitch, M. M., 1941. Canon of Insolation and the Ice Age Problem (English Translations, Washington, D. C.). Ko¨niglich Serbische Akademie, Beograd. Muhs, D. R., Simmons, K. R., Steinke, B., 2002. Timing and warmth of the Last Interglacial period: new U-series evidence from Hawaii and Bermuda and a new fossil compilation for North America. Quaternary Science Reviews 21, 1355–1383. Petit, J. R., Jouzel, J., Raynaud, D., Barkov, N.I., Barnola, J.-M., Basile, I., Bender, M. L., Chappellaz, J., Davis, M., Delaygue, G., Delmotte, M., Kotlyakov, V.M., Legrand, M., Lipenkov, V. Y., Lorius, C., Pepin, L., Ritz, C., Saltzman, E., Stievenard, M., 1999. Climate and atmospheric history of the past 420,000 years from the Vostok ice core, Antarctica. Nature 399, 429–436. Plagnes, V., Causse, C., Genty, D., Paterne, M., Blamart, D., 2002. A discontinuous climatic record from 187 to 74 ka from a speleothem of the Clamouse Cave (south of France). Earth and Planetary Science Letters 201, 87–103. Potter, E.-K., Esat, T. M., Schellmann, G., Radtke, U., Lambeck, K., McCulloch, M.T., 2004. Suborbitalperiod sea level oscillations during marine isotope substages 5a and 5c. Earth and Planetary Science Letters 225, 191–204. Reeder, R. J., Nugent, M., Lamble, G. M., Tait, C. D., Morris, D. E., 2000. Uranyl incorporation into calcite and aragonite: XAFS and luminescence studies. Environmental Science & Technology 34, 638–644. Richter, F.M., Turekian, K.K., 1993. Simple models for the geochemical response of the ocean to climatic and tectonic forcing. Earth and Planetary Science Letters 119, 121–131. Robinson, L. F., Henderson, G. M., Hall, L., Matthews, I., 2004. Climatic control of riverine and seawater uranium-isotope ratios. Science 305, 851–854. Schellmann, G., Radtke, U., 2004. A revised morphoand chronostratigraphy of the Late and Middle Pleistocene coral reef terraces on Southern Barbados (West Indies). Earth-Science Reviews 64, 157–187. Scholz, D., 2005. U-series dating of diagenetically altered fossil reef corals and the application for sealevel reconstruction. PhD Thesis, University of Heidelberg, 172 pp. Scholz, D., Mangini, A., Felis, T., 2004. U-series dating of diagenetically altered fossil reef corals. Earth and Planetary Science Letters 218, 163–178. Shackleton, N. J ., 2000. The 100 000-year ice-age cycle identified and found to lag temperature, carbon dioxide, and orbital eccentricity. Science 289, 1897–1902. Siddall, M., Rohling, E. J., Almogi-Labin, A., Hemleben, C., Meischner, D., Schmelzer, I., Smeed, D. A.,

U-Redistribution in Fossil Reef Corals 2003. Sea level fluctuations during the last glacial cycle. Nature 423, 853–858. Speed, R. C., Cheng, H., 2004. Evolution of marine terraces and sealevel in the last interglacial, Cave Hill, Barbados. Geological Society of America Bulletin 115, 219–232. Stein, M., Wasserburg, G. J., Aharon, P., Chen, J. H., Zhu, Z. R., Bloom, A., Chappell, J., 1993. TIMS Useries dating and stable isotopes of the last interglacial event in Papua New Guinea. Geochimica et Cosmochimica Acta 57, 2541–2554. Stirling, C. H., Esat, T.M., McCulloch, M. T., Lambeck, K., 1995. High-precision U-series dating of corals from Western Australia and implications for the timing and duration of the Last Interglacial. Earth and Planetary Science Letters 135, 115–130. Stirling, C. H., Esat, T. M., Lambeck, K., McCulloch, M.T., 1998. Timing and duration of the Last Interglacial: evidence for a restricted interval of widespread coral reef growth. Earth and Planetary Science Letters 160, 745–762. Taylor, F. W., Mann, P., 1991. Late Quaternary folding of coral reef terraces, Barbados. Geology 19, 103–106. Thompson, W. G., Spiegelmann, M.W., Goldstein, S. L., Speed, R. C., 2003. An open-system model for U-series age determinations of fossil corals. Earth and Planetary Science Letters 210, 365–381. Thompson, W. G., Goldstein, S.J., 2005. Open-system coral ages reveal persistent suborbital sea level cycles. Science 308, 401–404.

139

Toscano, M. A., Lundberg, J., 1999. Submerged Late Pleistocene reefs on the tectonically-stable S. E. Florida margin: high-precision geochronology, stratigraphy, resolution of substage 5a sea level elevation, and orbital forcing. Quaternary Science Reviews 18, 753–767. Tzedakis, C., McManus, J. F., Hooghiemstra, H., Oppo, D.W., Wijmstra, T.A., 2003. Comparison of changes in vegetation in northeast Greece with records of climate variability on orbital and suborbital frequencies over the last 450 000 years. Earth and Planetary Science Letters 212, 197–212. Villemant, B., Feuillet, N., 2003. Dating open systems by the 238U-234U-230Th method: application to Quaternary reef terraces. Earth and Planetary Science Letters 210, 105–118. Waelbroeck, C., Labeyrie, L., Michel, E., Duplessy, J. C., McManus, J. F., Lambeck, K., Balbon, E., Labracherie, M., 2002. Sealevel and deep water temperature changes derived from benthic foraminifera isotopic records. Quaternary Science Reviews 21, 295–305. York, D., 1969. Least squares fitting of a straight line with correlated errors. Earth and Planetary Science Letters 5, 320–324. Zhu, Z. R., Wyrwoll, K.-H., Collins, L. B., Chen, J. H., Wasserburg, G. J., Eisenhauer, A., 1993. High-precision U-series dating of Last Interglacial events by mass spectrometry: Houtman Abrolhos Islands, Western Australia. Earth and Planetary Science Letters 118, 281–293.

This page intentionally left blank

10. Holocene and Eemian Varve Types of Eifel Maar Lake Sediments Bert Rein, Knut Ja¨ger, Yvonne Kocot, Kirsten Grimm and Frank Sirocko Institute for Geosciences, Johannes Gutenberg-University, Becherweg 21, 55099 Mainz, Germany

ABSTRACT Varves of the Holocene and of the last interglacial were investigated in two sediment sequences from Eifel maar lakes. The modern maar with Schalkenmehrener Maar Lake and the dry maar lake West Hoher List have the same size, are two kilometres apart at the same altitude, but the Eemian lake was much deeper. The sediments of both lakes are dominated by autochthonous sediments, mainly from diatom-dominated algae. Differences in the palaeoproductivity and in calcite precipitation are probably not climatically controlled but due to lake basin morphometry and the carbonate reservoir in the catchment areas. The occurrence of dry periods with aeolian dust deposition during the last interglacial is the major difference of the last interglacial to the Holocene sedimentation. The discrimination of natural sedimentation processes versus human-controlled clastic input during the last 3000 years is however poorly understood and complicates the interpretation of the clastic sediments during the time of the Little Ice Age and their comparison with the sedimentation during the Late Eemian Aridity Pulse (LEAP). 10.1 INTRODUCTION The varves in the sediments of the Eifel maar lakes were first studied by Negendank and co-workers in several maar lakes since 1984. They described the seasonal layers of varves from the lakes of Meerfelder Maar (Zolitschka, 1986, 1988, 1990; Poth, 1993; Poth and Negendank, 1993; Brauer et al., 2001),

Holzmaar (Zolitschka, 1990; Brauer, 1994; Diel, 1995; Rein, 1996; Rein and Negendank, 1997; Zolitschka et al., 2000; Brauer et al., 2001; Baier et al., 2004; Bru¨chmann and Negendank, 2004; Kienel et al., 2005), Weinfelder Maar (Brauer, 1988; Brauer and Negendank, 1993) and finally Schalkenmehrener Maar (SMM) (Heinz, 1991; Heinz et al., 1993). The varve thicknesses were partly measured microscopically as one potential palaeoenvironmental proxy. In greater detail, the temporal and spatial variability of the varve thickness and of the seasonal sublayers were measured and described together with the downcore variation of the components in the seasonal sublayers of the varves in follow-up studies (Rein, 1996; Rein and Negendank, 1997). During the last years, studies on Holocene sediments of the Eifel maar lakes concentrated on the Holzmaar and on selected time windows for which diatom assemblages were explored with seasonal resolution (Baier et al., 2004; Bru¨chmann and Negendank, 2004; Kienel et al., 2005). Apart from the acquisition of palaeoenvironmental information from the varves, the development of high-resolution continuous varvechronologies was a primary goal of Negendank and co-workers (references above). Since 2000, Sirocko and co-workers drilled the sediments of the dry maars to reconstruct the history of palaeoclimate, volcanism and environments in the Eifel/ central Europe during the last glacial cycles. The Eifel laminated sediment archive (ELSA) at Mainz comprises 1700 m of sediment cores from 30 dry maar lakes (Schaber and Sirocko, 2005), which cover at the moment the last 140 000 years (Sirocko et al.,

142

Bert Rein et al.

2005, Seelos and Sirocko, 2005, this volume) and the Holsteinian (Diehl and Sirocko, this volume). Work on a record from MIS 7 is under preparation. Dating of the dry maar cores was accomplished by tuning, radiocarbon and OSL dates, and a floating varve chronology exists for the last interglacial in the dry maar West Hoher List (HL). In 2005, several long cores reaching down into the last glacial period were retrieved from the SMM Lake to add the Holocene to the ELSA chronology (SM2-4). Renewed varve counting is not intended. Instead, the age model for the Holocene ELSA cores is derived from correlation with the varve counted continuous composite cores SMM1ab and SMM2ab, which were drilled and studied by the Negendank group (Rein and Negendank, 1997). We will, however, compare the sedimentology of the Holocene varves with those from the last interglacial to identify similarities/differences in the course of the two interglacials (Fig. 10.1a, 10.1b). The internal structure of sediment deposition during one annual cycle (i.e. one varve) reflects the seasonal change in the catchment area of a lake basin and is the result of several, partly competitive processes. 10.2 THE HOLOCENE SMM LAKE The SMM has the shape of three overlapping circular structures (Fig. 10.1b). The polygenetic character of this phreatomagmatic structure has been recognized by Frechen (1962) and Bu¨chel and Krawczyk (1986). The maar has a diameter of 1000 m in N–S and 1400 m in E–W direction (Bu¨chel and Krawczyk, 1986). The crater rims reach up to 500 m above sealevel (a.s.l.), whereas the lake level is at 420.5 m a.s.l. (Table 10.1). The maar lake covers the western part of the maar, is 550 m by 600 m wide and has a small overflow in the south. The catchment area, excluding the influence of remote aeolian sources, is restricted to the maar structure. The modern lake covers a fifth of the

maar, while an ancient larger lake also occupied the filled-in eastern maar section. This part of the maar is today covered with a peat bog (Straka, 1975). The lake has steep slopes of about 30 that reach water depths of more than 19 m within < 100 150 m from the lake margin (Fig. 10.1c). The lake bottom is flat, and maximum water depth is 21.5 m. Limnological research in the maar lake has been summarized by Scharf and Bjo¨rk (1992). The growth of diatoms is limited by the availability of dissolved siliceous acid and soluble reactive phosphorus (Schettler et al., 1999). The climate of the Eifel area shows strong oceanic influences. Around the lakes West HL and SMM, the current mean annual temperature is 7.5 C, the annual rainfall is 760 mm with maxima from June to August and November to February (Station Mehren, DWD in van Haaren, 1988). 10.2.1 Sediments In a first drilling campaign in 1989 (Fig. 10.1c), 11 core sequences, 8 to 11 m long, were recovered from SMM Lake (Heinz, 1991; Rein, 1991; Heinz et al., 1993; Rein and Negendank, 1993). Late glacial and Holocene varves were described and counted in one of the cores from the central part of the lake (Heinz, 1991). Two important time markers, already known from other maar lakes, could be identified; the tephra of the Laachersee eruption (LST) is an important time marker between northern Italy and southern Scandinavia, the tephra of the nearby Ulmener Maar eruption (UMT) is a regional isochrone (Fig. 10.2). In Lake Holzmaar, the LST is radiocarbon and varvechronologically dated to 12 900 cal. years BP, and the UMT has an age of 11 000 cal. years BP (Zolitschka et al., 2000). In 1992, two pairs of 11-m long cores, each of the pairs forming one overlapping continuous sediment sequence (SMM1ab and SMM2ab), were drilled using the Usinger piston corer in the deep eastern central part of the lake (Rein and Negendank, 1997). In this part of the lake Holocene sedimentation rate is highest as was known from the

Holocene and Eemian Varve Types 2555

2560

143

2565

2570

(a) 5565

5565 Ulmener Maar DAUN Gemündener Maar

Weinfelder Maar 5560

5560 Schalkenmehrener Maar dry maar west of Hoher List Pulvermaar 5555

5555

Holzmaar Ueßbach

Meerfelder Maar Sam

Kle

bac met

5550

A

lf

h

s er

2555

Lie

ll Ky ine

5550

2565

2560

21.5

SM2 SM4

SMM 1a,b

5

2

km

SM3 20.5

10 5 19.515

(c)

21.

(b)

2570

0

SMM 2a,b

100 m

Fig. 10.1 Topographic (LVA RLP) and bathymetric data of the study area. (a) Digital elevation model (DEM) of the Westeifel volcanic field around the maar lakes, (b) DEM of the Dauner Maar Group and the dry maar West Hoher List, (c) bathymetric map of the lake in the Schalkenmehrener Maar with core locations (letters with Roman number – drillings in 1989, SMM1ab and SMM2ab (1992) and ELSA cores SM2-4 (2005).

sediment analysis of the previous coring campaign. From 11 to 8.4 m of sediment depth (s.d.) (Fig. 10.2), the cores consist of greyish silt/clay laminites that were deposited during the last glacial maximum (Heinz et al., 1993). The late glacial and Holocene sediments comprise laminated diatomaceous gyttja interrupted during the Younger Dryas period (7.1–6.8 m s.d.) only when silt/clay laminites, turbidites and deposits from dust storms were again deposited. The

LST is found at 7.25 m s.d., shortly before the clastic sedimentation of the Younger Dryas commences, whereas the UMT (2.5 mm thick) follows these sediments within the Preboreal sediments at 6.7 m s.d. A refined varve analysis of 8300 years of sedimentation above the UMT in these cores is described in this paper. The same sequence of clastic and organic sediments was found in the recently drilled ELSA cores (SM2-4) recovered in September 2005 (Figs. 10.1c

144

Bert Rein et al.

Table 10.1 Comparison of Holocene and Eemian sites and sediments

Regional settings Water depth Lake diameter Crater diameter (catchment area) Altitude lake plane Maximum altitude in catchment area

Holocene

Eemian

max 30 m 550 600 m 1400 1000 m 420 m a.s.l. 500 m a.s.l.

65 m 500 m 1000 800 m 420 m a.s.l. 530 m a.s.l.

Floating varve chronology fixed by radiocarbon dating 0:1 > 1 mm Variation determined by the thickness of diatom layers

Floating varve chronology

Spring/early summer Stephanodiscus sp. Fragilaria sp. Summer Cyclotella sp., Nitschia sp.

Spring/early summer Stephanodiscus sp. Fragilaria sp. Summer Cyclotella radiosa, Cyclotella sp.

Very limited occurrence

Abundant, forms thick layers, often half and more of the varve thickness Rare Common

Distance between both archives is 2 km Dating Varve thickness

Most important diatoms

Autigenic minerals Calcite

Vivianite Pyrite Clay, silt

Common Common Rare

and 10.2). The LST is a 6-cm thick sandy layer, whereas the UMTcan only be reliably detected in thin sections and is not shown in Fig. 10.2. 10.2.2 Dating A method was developed on thin-section profiles from Lake Holzmaar for the detailed correlation of varved sediment sequences within the lake (Rein, 1996; Rein and Negendank, 1997). This method was adapted to study the varved sequences of SMM Lake (Rein and Negendank, 1997). Specific varves with a very characteristic internal seasonal structure or unusual components as well as event layers are used for the correlation of thin-section profiles from different sites within the lake. The micro-reference layers (Rein, 1996) are used as a grid to compare the number and

0:1 > 1 mm Thickest varves when lithic input is high

Rare Though abundant during some ‘‘dry’’ periods

composition of varves between these layers (in SMM: 377 micro-reference layers between 11 000 and 2700 cal. BP). Between the reference layers, a further varve-by-varve correlation was conducted. The varve-by-varve correlation allowed to recognize varves that are missing compared to the other sequence. Compared to a synthetic master varve chronology that integrates the maximum of varves between each of the reference layers, 99% of the varves could be found in the individual cores. This is comparable to values found in Lake Holzmaar where three thin-section profiles were compared by this method (Rein, 1996; Rein and Negendank, 1997). For the study presented here, 8300 varves above the UMT were studied. All ages rely on the age of the UMT (11 000 cal. BP, Zolitschka et al., 2000), which is the starting

Holocene and Eemian Varve Types Holocene interglacial (b)

Eemian interglacial (a) 111

HL2

112

1

113

2 Late last interglacial

117

4 5 6

LEAP

121 122

slump

123 124 125 126

slump

Eemian interglacial (in sensu stricto)

120

Saalian 127

7 8

2

3 Diatom layer 5

4

Depth (m)

118 119

0

1

Holocene interglacial

116

SM4 Core loss

SM2

SMM

3

Age (kyr BP)

115

Age (kyr BP)

0

Weichselian

114

145

9

10 11

UMT

12

Younger Dryas LST BøllingAllerød

13 14

5

empty

Weichselian

15

7

Silt/clay laminites

Diatomaceous gyttja

Turbidites

Plant rests

UMT - Ulmener Maar Tephra

6

LST - Lake Laach Tephra

Fig. 10.2 Lithological profiles of lake sediments from (a) the dry maar West Hoher List and from (b) the lake in the Schalkenmehrener Maar.

point for the floating varve chronology of SMM Lake (Fig. 10.3a). The varve analysis ends with a 12-mm thick turbidite (Fig. 10.3a) that mainly contains a graded bedding of lithics in the silt and clay fraction. Other turbidites with prevailing lithic matter and comparable thickness follow (Fig. 10.2). These turbidites, as well as contemporaneous turbidites in Lake Holzmaar (Rein, 1996), were interpreted as testifying to increased human activity in the lake catchment (Zolitschka, 1990). Thus, because of possibly strong human impact, younger sediments are not considered for comparison with varves of the last interglacial. Only two turbidites, 3 and 1 mm thick, could be found in the entire sequence between the UMT and the 12-mm thick turbidite at 2700 cal. BP (Fig. 10.3a). They look different

and consist of resuspended lake sediments. At their bottom plant remains, large diatoms and silt prevail followed by a graded bedding of diatoms and lithics in the silt and clay fraction. The appearance of varves in the temporal vicinity of these turbidites is unchanged, whereas varves after the 12-mm thick turbidite contain more lithic components. 10.2.3 Varves The laminated diatomaceous gyttja of SMM Lake is made up of varves that consist of two or more seasonal sublaminae, essentially, a sequence of diatom assemblages which form during consecutive diatom blooms within a year (Fig. 10.4a). One or two diatom layers form the basal part of the varves with small centric diatoms in the lower layer (mainly

146

Bert Rein et al. EEMIAN 1000

100

(b)

10

(a)

10000

HOLOCENE

38

54.3

26

Turbidite 3 mm

113

33

35 37

34 36

3

39

40

111 4.0 Turbidite 12 mm Turbidite 1 mm

32

4

31

115 25

5.0

29

30

5

28 27

kyr BP

20 22 24 26

19 21 23 25

18 17 16 15

23 24, 22 119

21 19

14

18

11 13

12

9

9 10

10 121 6.0 56.0

7

8

20

Depth (m)

7

117 Depth (m)

cal kyr BP

6

55.0

ELSA core HL2

SMM

6

SLUMP

57.4

5

9

4

123

18

10 57.6

3

SLUMP 125

58.4 10

1

2

10

6.7

11

0.01

1.0 0.1 Varve thickness (mm)

10

14 127 0.01

16 15 59.0 0.1 1.0 Varve thickness (mm)

10

Fig. 10.3 Varve thickness in SMM2ab and HL2 lake sediments and occurrence of microlithozones (numbered bars).

Stephanodiscus sp.) followed by small pennate diatoms in the upper layer (mainly Fragilaria sp.) of the basal part. Frequently, only the layer dominated by Fragilaria sp. occurs. On these nearly pure diatom layers follow larger centric (mainly Cyclotella spp.) and pennate diatoms (mainly Nitschia sp.), mostly mixed with aquatic, rarely terrestrial plant remains and very fine organic matter. The SMM Lake sediments were last studied in 1997; thus before more detailed seasonal diatom species analysis in varves of Eifel maar lakes were carried out (Baier et al., 2004; Bru¨chmann and Negendank, 2004; Kienel et al., 2005). The most pronounced difference of SMM Lake to other modern lakes is in the clastic sedimentation, in particular the complete absence of distinct clay and silt layers in

SMM Lake, except a few traces below 5.1 m s.d. (7400–6600 cal. BP). The lakes Meerfelder Maar (Zolitschka, 1990) and Holzmaar (Zolitschka, 1990; Rein, 1996) show such distinct clay layers, but these two lakes are fed by small local creeks. The lithic supply in SMM Lake only increases with the late Holocene deposition of turbidites after 2.7 kyr. Calcite could not be found in the Holocene sediments of SMM1ab and 2ab, but small amounts of calcite (< 0:5% of the sediment) occasionally occur in large thin sections of other cores from the centre of the lake (AII, SM2 and SM4). In these other cores, the occurrence of small amounts of calcite is limited to the periods between 11 400–9900 cal. BP and 9000–8000 cal. BP,

Varve = year

(b)

0.1–0.3 mm Spring Late summer + to winter summer

Varve = year

(a)

0.1–1(2) mm Late Spring summer + to winter summer

Holocene and Eemian Varve Types

HOLOCENE

EEMIAN

0.3–1.0 mm Late summer Spring + to winter summer

Clay

Varve = year

where its occurrence is limited to the period of the early Holocene thermal optimum (Zolitschka, 1990; Rein, 1996). 10.2.4 Microlithozones

Very fine organic matter

(c)

147

Larger diatoms, pelagial + littoral Small diatoms, pelagial Plant rests Vivianite Pyrite Calcite Silt

Fig. 10.4 Idealized varves for (a) SMM and (b, c) HL2.

although no single calcite layer could be found in these time windows. Calcite is more abundant and partly layer forming between 9900 and 9000 cal. BP ( corresponds to the time of microlithozone 3 in Fig. 10.5). The missing calcite in the thin sections of SMM1ab and SMM2ab is due to the late subsampling of these cores when calcite had already dissolved and was replaced by the growth of gypsum crystals on the surface of the split cores (Schettler et al., 1999). Occasional layer-forming calcite occurrence is also known from thin sections of the Bølling/Allerød sediments (AII/III, Fig. 10.1c; Heinz, 1991). So far, it is not clear whether calcite precipitation played a larger role than is documented in the sedimentary record of SMM Lake. The spatially and temporally spotty occurrence of calcite might be secondary and emphasized due to unfavourable preservation conditions for calcite. Similarly, calcite is not an important Holocene sediment compound in Lake Holzmaar

The occurrence and shape of compositionally different seasonal laminae make the succession of components within a varve and thus the varve composition highly variable over time. However, within narrow sediment sections, varve composition is almost the same. Sections in which the varves look almost the same with only minor annual variations are called microlithozones (labelled bars in Fig. 10.3a, Rein, 1996). Between two subsequent microlithozones there may exist both minor and major differences (Figs. 10.5 and 10.6) reflecting smaller or more effective changes in the lake or its catchment area (see ‘The succession of microlithozones’ below). The insignificant flux of allochthonous sediments into SMM Lake is the reason why only 40 microlithozones could be distinguished in the sediments between 11 000 cal. BP (UMT) and 2700 cal. BP. In the Lake Holzmaar sediments, 119 microlithozones were distinguished over the same period (Rein, 1996). SMM Lake varves mainly consist of diatoms. Years (varves) with diatom minima coincide with higher amounts of plant remains and very fine organic matter (e.g. MLZ of varve types 30d, 32c in Fig. 10.6). In these varves, the layers with either small or large diatoms frequently are missing, or at least one of these seasonal layers is very thin. The larger diatoms often are missing after very thick basal diatom layers (e.g. Fig. 10.5: varve types 7 and 8) or the layer with the larger diatoms mainly consists of resuspended large pennate diatoms from the littoral zone. During some periods, diatoms are not layer forming (e.g. varve types 30d, 32c in Fig. 10.6). In this case, their place is taken by very fine organic matter, and chrysophycean cysts appear more prominent.

148

Bert Rein et al.

6a

12

18b

25b

5

11b

18a

25a

17b

24b

11a

4b

10b

4a

10a

17a

24a

3b

9

16

23

3a

8

15b

22

2

7b

15a

21

1b

7a

14

20

1a

6b

13

19

Transition between 1 and 2

Fig. 10.5 Varve composition of microlithozones 1–25 in Schalkenmehrener Maar Lake (for legend see Fig. 10.6).

10.2.5 Varve thickness The thickness of each varve (Fig. 10.3a) was microscopically measured in the thinsection sequence of SMM2. In SMM1, varve thickness was only measured for selected

sections. Over a distance of 100 m, the varve thicknesses of SMM1 and SMM2 show a very good match, as earlier observed in Lake Holzmaar (Rein, 1996; Rein and Negendank, 1997). The variation of annual sediment accumulation is considerable,

Holocene and Eemian Varve Types

30c

33b

38a

30b

33a

37b

149

Small diatoms 30a

32c

37a

Large diatoms Chrysophy. cysts

29

32b

36b

Very fine organic matter Plant rests Clay

28b

32a

36a

40

28a

31c

35b

39c

28

31b

35a

39b

27

31a

34b

39a

26

30d

34a

38b

Fig. 10.6 Varve composition of microlithozones 26–40 in Schalkenmehrener Maar Lake.

despite the deposition of more or less autochthonous sediments only. Varve thickness varies between < 0:1 mm and larger than 2 mm. However, varve thickness does not change randomly but is organized in periods of thin and thicker varves as earlier

observed in Lake Holzmaar (Rein, 1996; Rein and Negendank, 1997). Long-term changes can be seen with superimposed short- and mid-term variability. A relation between varve thickness and varve shape is obvious. Thick layers with small diatoms

150

Bert Rein et al.

occur in the thickest varves (e.g. varve type 16 in Fig. 10.5) and are lacking in the thinnest varves (varve type 11b in Fig. 10.5). The amount of deposited diatoms controls the varve thickness, with only a few exceptions during years when diatom blooms did not occur in the pelagial and plant remains with attached diatoms were transported into the deep lake basin. Allochthonous (terrestrial) lithic and organic matter is rare and like newly formed minerals (calcite, pyrite and vivianite) have only little effect on varve thickness (Rein and Negendank, 1997). 10.2.6 The succession of microlithozones (MLZ)1–8 The succession of microlithozones in Figs. 10.5 and 10.6 is a compositional signal of changes in the lake (nutrient conditions, temperature, stability of lake slopes, lake level, expanses with flat water and water plants) and in the lake surroundings (erosivity of precipitation, erodibility of soils, slope stability, vegetational cover, deposition of aeolian dust and tephra material and weathering). The succession of microlithozones represents an effective (not genetic) classification of the sediments. The seasonal change of algae production or allochthonous matter entering the lake is documented in the sedimentary archive, but the processes which caused these variations are ambiguous. The palaeoclimatic interpretation still requires the process by which nutrients were supplied. Nutrients can be supplied with surface run-off, by groundwater discharge or by lake circulation that brings nutrients from the lake bottom into the euphotic level. Since SMM Lake has no inflow and the lithic matter in the sediments is rare before 2700 cal. BP, superficial discharge into the lake can only be of subordinate importance. For SMM Lake, Schettler et al. (1999) found that millennial changes of nutrient supply and of diatom growth in the lake are limited by the amount of groundwater discharge into the lake. In the following, the very simplified interpretation is

restricted to the MLZ 1 through MLZ 8 (11 000–8500 cal. BP). During the first 600 years (MLZ 1 and 2) of the record, varves are thin, and spring diatom blooms are of minor (MLZ 1b) or no (MLZ 1a) importance. The varves comprise mainly the summer diatom blooms and the deposition of aquatic plant remains. After 10 400 cal: BP (MLZ 3), varves become much thicker because the spring diatom blooms gain importance. At the same time, the input of aquatic plant remains is reduced. After 9700 cal. BP, the contribution of spring diatom blooms is fading and plant remains are first slightly (MLZ 4) then strongly increasing (MLZ 5–7a). At the same time, increased amounts of clay (MLZ 5) indicate allochthonous sediment sources or a destabilized littoral zone, possibly reworking of sediment at a lowered lake level. When varve types 7a and 8 are deposited, the varve thickness distinctly increases with the re-appearance of major spring diatom blooms ( 8700 cal: BP). Generally, during periods with major spring blooms, the amount of plant remains decreases and it increases when spring diatom blooms are absent. Since diatom blooms are nutrient limited in SMM Lake, this pattern can be explained with changed precipitation/evapotranspiration. Increased groundwater discharge from the catchment supplies nutrients (thick diatom blooms) and water to keep the lake level high. During drier periods, nutrients support only minor diatom blooms and lake level drops. When the former lake bottom emerges in the littoral zone, sediments are eroded and macrorests drifted into the lake’s centre, explaining the higher amounts of large pennate diatoms and the geochemical and mineralogical pattern (Schettler et al., 1999). This is a simplified view that reflects the varve thickness, varve composition and geochemical data on multicentennial and millennial timescales. Probably, variability on shorter timescales is introduced by variables as the thermal stratification and vertical mixing of lake water

Holocene and Eemian Varve Types

according to winter temperatures or the intensity of winds during periods of unstratified lakes. Bru¨chmann and Negendank (2004) interpreted a very detailed database of geochemical data and detailed seasonal diatom assemblages (10 100–9100 cal. BP) in the adjacent Lake Holzmaar in terms of these parameters. 10.3 THE EEMIAN DRY MAAR LAKE WEST HL The dry maar lake West HL is 1000 m 800 m in diameter and is only 2 km west-southwest of SMM Lake (Fig. 10.1). The palaeolake level was at about 420 m a.s.l., thus, at the same altitude as the modern SMM Lake. The crater rims reach up to 530 m a.s.l. (Table 10.1). Both maars are not only close to each other but also comparable in their morphometrical characteristics and their topographical position. However, water depth distinguishes both lakes. The Eemian maar lake was much deeper, with about 65 m of water depth (Table 10.1). A total of six cores have been retrieved from West HL by the ELSA project in 2002– 2004 (Schaber and Sirocko, 2005). We focus here on the 104 m long core HL2, which contains lake sediments between 10 and 65 m. Pollen analysis and OSL dates reveal that sediments between 54.5 and 59 m s.d. were deposited during the last interglacial. Two sections with blockslides (allochthonous sandy) and slumped sediments interrupt the varved sequence between 58.5 and 57.6 m and between 57.45 and 56 m s.d. (Fig. 10.2) (Sirocko et al., 2005). 10.3.1 Varves The last interglacial varves in the sediments of the core HL2 consist of two or more seasonal sublaminae, with diatoms, calcite, as well as lithics and detrital organic matter as most frequent components (Fig. 10.3b, 10.3c). As in the SMM Lake varves described above, most varves reflect a

151

sequence of diatom assemblages that form during consecutive diatom blooms of a year. One, seldom two, diatom layers form the basal part of the varves. The basal layers contain much more frequent small centric diatoms (Stephanodiscus sp.) than small pennate forms (Fragilaria spp.) The succession of centric diatom blooms followed by blooms of pennate forms, which is frequently observed in SMM Lake and Lake Holzmaar, is not common. The basal diatom layers frequently are more or less pure algae layers; larger centric (Cyclotella radiosa and Cyclotella sp.) and pennate diatoms (e.g. Navicula spp, Eunotia sp., Cocconeis sp. and Pinnularia sp.) are mixed with plant remains and lithic matter in the following seasonal layer. The abundance of lithic particles of clay and silt size increases in the top of the varves. In some rare cases, deposition of the clay fraction continues into the basal part of the next varve. The importance of calcite precipitation with the formation of frequently massive calcite layers mainly distinguishes the HL2 varves from the Holocene varves of the SMM Lake (and Holzmaar Lake). Calcite occurs mainly in distinct layers, but can be also diffusely distributed within the diatom layer. Calcite is found in different seasonal levels of the deposition of a year. Massive calcite layers formed above the basal diatom layers, between the basal diatom layers and the seasonal sublaminae with larger diatoms or in the layer with the larger diatoms. In the detrital layers at the top of the varves, calcite only occurs as intraclasts. 10.3.2 Microlithozones Carbonaceous varves were deposited in HL during warm and temperate periods and are the most common varve type (Figs. 10.2, 10.3, 10.7 and 10.8). Generally, the transition between microlithozones is a gradual change of varve type during the warm periods of the last interglacial when varve composition also is very similar over long periods of time. However, at first sight, it is

152

Bert Rein et al.

26

19

25

18

24

16

23

15

22

14

21

1 mm

20

1 mm

10

1 mm Very fine organic matter

Plant rests

Clay

Diatoms

Silt

Calcite

Pennate diatoms

1 mm

Fig. 10.7 Varve composition of varve types 10–20 in HL2 (for legend see Fig. 10.8).

remarkable that varve types normally change after turbidites. This might be an indication that varves were possibly eroded by the turbidites extinguishing the otherwise observed gradual change between varve types. The transition into and within the periods with thicker varves at 126 kyr ago and at 118 kyr ago is more rapid, sometimes within years. Generally, these transitions are characterized by a drastic change of the lithic components in the varves which also changed the varve thickness (Fig. 10.3b, varve types (VT) 14–16 in

Fig. 10.8 Varve composition of varve types 21–26 in HL2.

Fig. 10.7, VT 21–24 in Fig. 10.8), a variability not observed in the Holocene sections. The high amount of silt in the varves (VT 14, 15, 16, 21, 22, 24 and 26 in Fig. 10.7 and Fig. 10.8), deposited from aeolian dust (Sirocko et al., 2005; Seelos and Sirocko, this volume), is not observed in the early and mid-Holocene varves of SMM Lake (see above). 10.3.3 Varve thickness The thickness of the interglacial varves (Fig. 10.3b) was not individually measured (as for the Holocene sediments Rein, 1996; Rein

Holocene and Eemian Varve Types

and Negendank, 1997), but determined as a mean varve thickness from the number of varves in 25-mm sediment sections. Varves are between 0.1 and 1 mm thick in the interglacial sediments of HL (Fig. 10.3b). Thick varves always contain lithic detritus. Carbonaceous varves (VT 19–20 in Fig. 10.7 and type 25 in Fig. 10.8) are typically very thin. These thin varves are most common throughout the last interglacial (see also supplementary material to Sirocko et al., 2005). Major changes of varve thickness occurred at the beginning and the end of the interglacial, most pronounced at 118 kyr ago during the arid and cold interval that Sirocko et al. (2005) called LEAP (LEAP) (Fig. 10.3b). Since varve thickness determination differs from the method used in the Holocene sediments, it should be emphasized that short-term variability of varve thickness on the scale of decades is at the moment still suppressed by the method used for the last interglacial. We do not have detailed geochemical and faunal data yet; thus, the palaeoclimatic interpretation of the microlithozones in the sediments of palaeolake HL remain mostly descriptive, protocolling the sedimentation and changes of sedimentation in the lake. However, doubtlessly apparent from the microlithozones is the existence of dry periods (Figs. 10.7 and 10.8 VT 14–16, 23, 24, 26) with increased aeolian dust transport around the palaeolake. 10.4 COMPARISON OF SEDIMENTATION DURING THE EEMIAN AND HOLOCENE INTERGLACIALS The seasonal composition of lacustrine varves was documented in two neighbouring maar structures of comparable size but from different interglacial periods. The Holocene varves were deposited in SMM Lake, the last interglacial varves in the dry maar West HL. The sediments of both lakes are dominated by autochthonous sediments, basically, the

153

remains of diatoms (Table 10.1) and aquatic plant remains. Whereas lithic components are nearly absent in SMM Lake, very fine layers of clay frequently occurred in the dry maar lake West HL where silt layers are more frequent in several horizons (VT 14–16, 23, 24 and 26). In both lakes, varve thickness varies between < 0:1 and 1 mm, with higher variability in SMM Lake. The centennial to millennial scale variability of varve thickness in the Holocene Lake is related to variations in the amount of groundwater discharge which is the major source of external nutrient supply (Schettler et al., 1999). Varve thickness larger than 0.3 mm occurs in the palaeolake only at the beginning ( 126 kyr ago) and at the end (112 kyr ago) of the last interglacial. A spike of high varve thickness during the interglacial is at 118 kyr ago (LEAP, 470 years duration, Sirocko et al., 2005). Varve thickness in these sections is not caused by thick diatom layers as in SMM Lake, but it is caused by greater abundance of clastic aeolian silt (Seelos and Sirocko, 2005) (Table 10.1). Thus, on average, diatom primary production and nutrient availability was higher in SMM Lake. This might be an indication that nutrient reflux (phosphorous) from the organic-rich sediments in SMM Lake was of greater importance than in the much deeper palaeolake West HL (Table 10.1). Apart from this, the extensive calcite precipitation in the palaeolake could have been a very effective mechanism that removed nutrients (phosphorous) from the water column (Fo¨rstner and Patchineelam, 1976). Occasional calcite precipitation in SMM Lake is apparently biogenetically induced since it follows thick diatom layers during summer. In contrast, regular annual calcite precipitation in rather thin varves associated with thin diatom layers characterizes the interglacial dry maar lake West HL. Even if the Eemian period might have been slightly warmer than the Holocene in middle Europe (see Ku¨hl and Litt, this volume), the deeper palaeolake HL probably compensated

154

Bert Rein et al.

this climatic signal compared to the rather shallow and thus more rapidly warming Holocene SMM Lake. Artificial and natural wells in the HL maar show calcite crusts, which indicate that the Devonian rocks in this area bear carbonaceous sediments, which is not unusual in Eifel Devonian strata. Thus, the frequent occurrence of calcite layers in the interglacial varves cannot be attributed to climate-controlled processes, but are most probably simply caused by local geology and palaeo water depths. An obvious climatic signal that distinguishes the sediments of both lakes is the frequent occurrence of aeolian dust storms at 126; 118 and 112 kyr ago (Fig. 10.3b). Dust storms were first abundant immediately before the algae-dominated sedimentation of the last interglacial began ( 126 kyr ago, duration less than 200 years). The second period with increased dust storm activity is documented at 118 kyr ago (duration 470 years). The third period at 112 kyr ago (duration 100 years) ends the organic-dominated sedimentation of the last interglacial. Comparable dry periods as during the last interglacial are not documented in SMM Lake for the early and middle Holocene period but only for the Younger Dryas period immediately before the Holocene period commences. In all Eifel maar lakes, the Younger Dryas period is documented as a lithic-dominated period within the diatom-dominated late glacial and Holocene organic sediments (Zolitschka et al., 1992; Heinz et al., 1993; Poth and Negendank, 1993). It is much more complicated to detect if dry events occurred during the late Holocene. The late Holocene sediments of SMM Lake contain large amounts of lithic matter. Turbidites and dust layers are abundant in part of the iron age sediments and in the deposits of the Roman and medieval period but also during the Little Ice Age (290– 630 cal. BP, Rein, 1991; Heinz et al., 1993) as also observed in Holzmaar Lake (Zolitschka, 1998, 2002). However, variations of the lithic flux are obviously related to known periods

of human settlement activity and cultivation in the region around the maar lakes (Negendank and Zolitschka, 1993). The extent of human occupation into the high regions of the Eifel around the maar lakes occurred during periods of favourably mild climate (e.g. van Haaren, 1988). Increased settlement and land use activities are observed during warmer periods (Roman Empire, Medieval Period), and settlement activity decreased during periods of assumed unfavourable climate (early medieval migration of the peoples, Little Ice Age). However, from chronicle data (van Haaren, 1988), it seems that during the climatically unfavourable period of the Little Ice Age, the maar craters were local centres for the retreat of part of the population from the surrounding plateaus. The population at close vicinity to the lakes slightly increased, whereas medieval villages on the surrounding plateaus were deserted. Thus, the lithic maximum in the sediments of the Little Ice Age (290–630 cal. BP, Heinz et al., 1993) which is also observed in Lake Holzmaar (Zolitschka, 2002) needs no climatological explanation but could be explained by anthropogenically induced soil erosion only. So it leaves a major challenge to examine the Little Ice Age sediments in the very detail to discriminate evidence for natural and human forcing. So far, no evidence exists for extreme dry periods within the Holocene as observed in the deposits of the last interglacial palaeolake.

REFERENCES Baier, J., Lu¨cke, A., Negendank, J.F.W., Schleser, G.-H., Zolitschka, B., 2004. Diatom and geochemical evidence of mid- to late Holocene climatic changes at Lake Holzmaar, West-Eifel (Germany). Quaternary International 113, 81–96. Brauer, A., 1988. Versuch einer Erfassung alter Seespiegelsta¨nde an ausgesuchten Eifelmaaren und mikrostratigraphische Untersuchungen an Sedimenten des Weinfelder Maares, Diploma Thesis, University Trier, 117 pp.

Holocene and Eemian Varve Types Brauer, A., 1994. Weichselzeitliche Sedimente des Holzmaares – Warvenchronologie des Hochglazials und Nachweis von Klimaschwankungen. Documenta Naturae 85, 210 pp. Brauer, A., Negendank, J.F.W., 1993. Palaeoenvironmental reconstruction of the late- and postglacial sedimentary record of Lake Weinfelder Maar. In: Negendank, J.F.W., Zolitschka, B. (Eds.), Palaeolimnology of European maar lakes. Lecture Notes in Earth Sciences 49, 223–235. Brauer, A., Litt, T., Negendank, J.F.W., Zolitschka, B., 2001. Lateglacial varve chronology and biostratigraphy of lakes Holzmaar and Meerfelder Maar, Germany. Boreas 30, 83–88. Bu¨chel, G., 1984. Die Maare im Vulkanfeld der Westeifel, ihr geophysikalischer Nachweis, ihr Alter und ihre Beziehung zur Tektonik der Erdkruste. Ph.D. Thesis, Universita¨t Mainz, unpublished, 385 pp. Bu¨chel, G., Krawczyk, E., 1986. Zur Genese der Dauner Maare im Vulkanfeld der Westeifel. Mainzer Geowissenschaftliche Mitteilungen 15, 219–238. Bru¨chmann, C., Negendank, J.F.W., 2004. Indication of climatically induced natural eutrophication during the early Holocene period, based on annually laminated sediment from Lake Holzmaar, Germany. Quaternary International 123–125, 117–134. Diehl, M., and Sirocko, F., 2006. New Holsteinian pollen record from the dry maar at Do¨ttingen (Eifel). In: Sirocko, F., Litt, T., Claussen, M. (Eds.), The climate of past interglacials, developments in paleo- and environmental research. Elsevier, Amsterdam (this volume). Diel, H.J., 1995. Mikrostratigraphische Abgrenzung der Ju¨ngeren Dryas mit Hilfe von drei warvierten Sedimentprofilen aus dem Holzmaar. Diploma Thesis, University Trier, 113 pp. Fo¨rstner, U., Patchineelam, S.R., 1976. Bindung und Mobilisation von Schwermetallen in fluviatilen Sedimenten. Chemiker-Zeitung 100, 49–57. Frechen, J., 1962. Fu¨hrer zu vulkanisch-petrologischen Exkursionen im Siebengebirge am Rhein, Laacher Vulkangebiet und Maargebiete der Westeifel. Schweizerbart, Berlin, 151 pp. van Haaren, C., 1988. Eifelmaare. Landschaftso¨kologisch-historische Betrachtung und Naturschutzplanung. Pollichia 16, 548 pp. Heinz, T., 1991. Pala¨olimnologische und spektralanalytische Untersuchungen an jahreszeitlich geschichteten Sedimenten des Schalkenmehrener Maares/ West. Diploma Thesis, Universita¨t Trier, 113 pp. Heinz, T., Rein, B., Negendank, J.F.W., 1993. Sediments and basin analysis of Schalkenmehrener maar lake. In: Negendank, J.F.W., Zolitschka, B. (Eds.), Palaeolimnology of European maar lakes. Lecture Notes in Earth Sciences 49, 149–161.

155

Kienel, U., Schwab, M., Schettler, G., 2005. Distinguishing climatic from direct anthropogenic influences during the past 400 years in varved sediments from Lake Holzmaar (Eifel, Germany). Journal of Palaeolimnology 33(3), 327–347. Ku¨hl, N., Litt, T., 2006. Quantitative time series reconstructions of Holsteinian and Eemian temperatures using botanical data. In: Sirocko, F., Litt, T., Claussen, M. (Eds.), The climate of past interglacials, developments in paleo- and environmental research. Elsevier, Amsterdam (this volume). Negendank, J.F.W., Zolitschka, B., 1993. Maars and maar lakes of the Westeifel volcanic field. In: Negendank, J.F.W., Zolitschka, B. (Eds.), Palaeolimnology of European maar lakes. Lecture Notes in Earth Sciences, Springer Verlag, 61–80. Poth, D., 1993. Mikrostratigraphische und spektralanalytische Untersuchungen an jahreszeitlich geschichteten spa¨tglazialen und holoza¨nen Sedimenten des Meerfelder Maares (Westeifel). Diploma Thesis, University Trier, 107 pp. Poth, D., Negendank, J.F.W., 1993. Palaeoclimate reconstruction at the Pleistocene/Holocene transition – a varve dated microstratigraphic record from Lake Meerfelder Maar. In: Negendank, J.F.W., Zolitschka, B. (Eds.), Palaeolimnology of European maar lakes. Lecture Notes in Earth Sciences, Springer Verlag, 209–222. Rein, B., 1991. Versuch einer Rekonstruktion des Pala¨oenvironments anhand hoch-zeitauflo¨sender geochemischer und sedimentologischer Untersuchungen an spa¨t- und postglazialen Sedimenten des Schalkenmehrener Maarsees (Westeifel/Bundesrepublik Deutschland). Diploma Thesis, University Trier, 109 pp. Rein, B., 1996. Die Warvenchronologie des Holzmaares, Vergleichende Untersuchung an drei Sedimentprofilen. Doctoral Thesis, University Potsdam, 128 pp. Rein, B., Negendank, J.F.W., 1993. Organic carbon contents from Schalkenmehrener Maar Lake: A palaeoclimate indicator. In: Negendank, J.F.W., Zolitschka, B. (Eds.), Palaeolimnology of European maar lakes. Lecture Notes in Earth Sciences, Springer Verlag, 163–171. Rein, B., Negendank, J.F.W., 1997. Die Warvenchronologien des Holzmaar und Schalkenmehrener Maar, NE 154/24–1–3: DFG-Abschlußbericht, unpublished, 86 pp. Schaber, K., Sirocko, F., 2005. Lithologie und Stratigraphie der spa¨tpleistoza¨nen Trockenmaare der Eifel. Mainzer Geowiss. Mitt. 33, 295–340. Scharf, B.W., Bjo¨rk, S. (Eds.), 1992. Limnology of Eifel maar lakes. Archiv fu¨r Hydrobiologie – Advances in Limnology 38, 348 pp. Scharf, B.W., Menn, U., 1992. Hydrology and morphometry. In: Scharf, B.W., Bjo¨rk, S. (Eds.), Limnology of

156

Bert Rein et al.

Eifel maar lakes. Archiv fu¨r Hydrobiologie – Advances in Limnology 38, 43–62. Schettler, G., Rein, B., Negendank, J.F.W., 1999. Geochemical evidence for Holocene palaeodischarge variations in lacustrine records from the Westeifel volcanic field, Germany: Schalkenmehrener Maar and Meerfelder Maar. The Holocene 9, 381–400. Seelos, K., Sirocko, F., 2005. RADIUS – Rapid Particle Analysis of digital images by ultra-high-resolution scanning of thin sections. Sedimentology 52, 669–681. Seelos, K., Sirocko, F., 2006. Abrupt cooling events at the very end of the last interglacial. In: Sirocko, F., Litt, T., Claussen, M. (Eds.), The climate of past interglacials, developments in paleo- and environmental research. Elsevier, Amsterdam (this volume). Sirocko, F., Seelos, K., Schaber, K., Rein, B., Dreher, F., Diehl, M., Lehne, R., Ja¨ger, K., Krebetschek, M., Degering, D., 2005. A late Eemian aridity pulse in central Europe during the last glacial inception. Nature 436, 833–836. Straka, H., 1975. Die spa¨tquarta¨re Vegetationsgeschichte der Vulkaneifel. Beitra¨ge zur Landespflege in Rheinland-Pfalz Beiheft 3, Oppenheim, 163 pp. Zolitschka, B., 1986. Warvenchronologie des Meerfelder Maares – Licht- und elektronenmikroskopische

Untersuchungen spa¨tglazialer und holoza¨ner Seesedimente. Diploma Thesis, University Trier, 119 pp. Zolitschka, B., 1988, Spa¨tquarta¨re Sedimentationsgeschichte des Meerfelder Maares (Westeifel). Mikrostratigraphie jahreszeitlich geschichteter Sedimente. Eiszeitalter und Gegenwart 38, 87–93. Zolitschka, B., 1990. Spa¨tquarta¨re jahreszeitlich geschichtete Seesedimente ausgewa¨hlter Eifelmaare. Documenta Naturae 60, 226 pp. Zolitschka, B., 1998. A 14,000 year sediment yield record from western Germany based on annually-laminated lake sediments. Geomorphology 22, 1–17. Zolitschka, B., 2002. Late Quaternary sediment yield variations – natural versus human forcing. Zeitschrift fu¨r Geomorphologie N. F., Suppl.-Bd. 128, 1–15. Zolitschka, B., Haverkamp, B., Negendank, J.F.W., 1992. Younger Dryas oscillation – varve dated microstratigraphic, palynological and palaeomagnetic records from Lake Holzmaar, Germany. In: Bard, E., Broecker, W. S. (Eds.), The last deglaciation: Absolute and radiocarbon chronologies, NATO ASI series I 2, 81–101. Zolitschka, B., Brauer, A., Negendank, J.F.W., Stockhausen, H., Lang, A., 2000. Annually dated late Weichselian continental palaeoclimate record from the Eifel, Germany. Geology 28, 783–786.

11. Dating of Interglacial Sediments by Luminescence Methods D. Degering and M.R. Krbetschek Saxon Academy of Sciences in Leipzig, Quaternary Geochronology Section at the TU Bergakademie Freiberg, Leipziger Str. 23, D-09596 Freiberg/Sa., Germany

ABSTRACT Luminescence dating methods like optically stimulated luminescence (OSL) or radiofluorescence (RF) offer possibilities for age determination on a large variety of interglacial sediments. Nevertheless, sediments deposited in warm stages show a number of peculiarities which have to be considered for reliable results. Especially, the effects of radioactive disequilibria and partial bleaching must be discussed carefully. Age determination by infrared-RF on sediments is possible up to MIS 9. The approaches for dating interglacial sediments are illustrated exemplarily on deposits from Central and Eastern Europe. 11.1 INTRODUCTION Palaeoclimatic investigations call for a correlation of global and regional archives of climate changes. Stratigraphic information alone is insufficient to give an unambiguous timescale. Among other methods, luminescence dating has a great potential to date typical sediments from glacial/interglacial cycles. The present paper briefly reviews the application of conventional luminescence methods (TL, OSL and IRSL) for dating sediments and introduces the recently developed radiofluorescence (RF) method. The correctness of luminescence age data depends on typical properties of interglacial sediments (e.g. lake sediments) like transportation processes, changes in water level and chemical conditions. Particularly, the problem of radioactive disequilibria in the uranium decay chain, caused by

geochemical processes, has to be taken into account. In the following, special attention will be given to analytical procedures, modelling of disequilibria and potential sources of error in the age calculation. Typical examples of dating interglacial sediment sequences from Central and Eastern Europe by luminescence methods will also be presented. 11.2 BASIC PRINCIPLES OF LUMINESCENCE DATING Luminescence dating of sediments has become an important tool in Quaternary geochronology during the last decades. The broad interest of geosciences in such techniques is especially related to the wide range of sediments which can be dated – either regarding their origin or age. The general principle of luminescencebased age determination is the solid-state dosimetry of ionising radiation, i.e. each mineral grain to be dated is considered as a tiny, natural dosimeter. Crystal mineral lattice defects (e.g. due to elemental substitution) of minerals may act as charge traps and recombination centres and thus as a base for the different luminescence processes. During their geological history, minerals are permanently exposed to a complex field of external radiation including the different types of radiation released by the radioactive decay of natural radionuclei (mainly 238U, 235U, 232Th, their instable daughter nuclei and 40K) and, to a minor extent, cosmic radiation. Owing to the impact of natural ionising radiation, the number of trapped charges increases continuously and the luminescence capability

158

D. Degering and M.R. Krbetschek

grows with time. It will reach saturation only if all traps are filled, and no charge loss (signal fading) occurs within the period to be dated. Traps may also be re-emptied by a number of different processes such as light exposure or heating. Both primary mineral formation and secondary emptying of traps are datable events where the ‘luminescence clock’ is set to zero. It should be mentioned that incomplete emptying is a partial reset event that may – analogous to other dating techniques in geosciences – bias the significance of the determined luminescence age. The concentration of trapped charge carriers is a measure of the absorbed radiation energy per mass, the dose (D). It can be experimentally determined by comparing the ‘natural’ luminescence signal of an untreated sample with the luminescence induced by known radiation doses, applied by an artificial, calibrated irradiation source. The experiment can be done in two ways. The natural luminescence is either increased stepwise by additional irradiations (additive [A] dose method) or laboratory irradiation is applied to regenerate the luminescence to its initial (‘natural’) level after a complete depletion of the natural luminescence signal (regenerative [R] dose method). In both cases, a calibration curve is obtained, which enables assigning the natural luminescence signal to a known radiation dose and thus to determine quantitatively the natural dose absorbed by the sample. The determined radiation energy dose is called equivalent dose (or palaeodose) DE, and its unit is the Gray ð1 Gy ¼ 1 J kg1 Þ. The luminescence build-up is quantita_ (the tively linked to the energy dose rate D absorbed dose per unit time, given e.g. in mGy a1 ) that affects the mineral. It can be calculated from radionuclide concentrations in the sediment and mineral grains plus cosmic contribution. Since the half-life of the major primordial radioisotopes are in the order of 109 a, their decay causes an almost constant irradiation of the sediment grains over the entire dating time range of some

100 kyr (provided that the radioactive equilibrium in the decay series is undisturbed). The dose actually deposited in an individual grain depends on a number of factors such as sediment moisture, grain size and internal radioactivity (e.g. 40K in potassium feldspars). Radioactive disequilibria (particularly common in the decay chain of 238U) caused by geochemical processes lead furthermore to time-dependent dose rates. Once all factors are reliably determined, the age (t) of a sample is simply calculated from the measured equivalent energy dose DE according to: _ t ¼ DE = D Luminescence-based age determination usually utilises quartz and feldspar (especially K-feldspar) grains. Stimulations by heat (thermoluminescence – TL) or by visible and infrared light (optically stimulated luminescence – OSL and infrared-stimulated luminescence – IRSL) are applied to determine the naturally absorbed dose. OSL signals of quartz and feldspar are much more light sensitive than most of their TL components, resulting in a much faster signal reset prior to deposition. To minimise the problems of insufficient optical bleaching, OSL dating replaced the TL techniques to a large extent in the 1980s. More detailed information about TL and OSL dating is given in Aitken (1985, 1998), Wintle (1997) and Boetter-Jensen et al. (2003). More recently, RF – i.e. the prompt luminescence during exposure to ionising radiation – was applied for sediment dating. Potassium feldspars (microcline and orthoclase) show a prominent near-infrared (865 nm) radiofluorescence emission (IRRF), which gives a measure of the amount of charge (electrons) that can be captured additionally to the already trapped electrons. IR-RF dating is generally carried out using a single-aliquot or single-grain regenerative (SAR or SGR) dose-determination protocol. Compared to other luminescence

Dating of Interglacial Sediments

11.3 SPECIFICS OF DATING INTERGLACIAL DEPOSITS Dating of Quaternary sediments is limited by two main factors: the saturation of the luminescence versus dose function and the long-time stability of the charge concentration in the traps. The first depends on the number of traps and on their probability to be filled by ionising radiation. The saturation level of the luminescence signal is reached in dependence of the total dose rate, i.e. the intensity of the ionising radiation in a sediment will govern the maximum determinable age. In materials with higher concentration of radioisotopes (high dose rate), the traps will be more quickly filled than in such with low concentration. The mineral itself may also contain a remarkable contribution to the overall dose rate (e.g. 40K in potassium feldspar). Therefore, OSL methods are restricted to ages up to about 130 kyr (MIS 5e) (see also Boetter-Jensen et al., 2003). In the case of quartz, stable OSL signals reach saturation mainly depending on the external dose rate. In the case of feldspar (IRSL), the age limit is

set by reasons of signal instability. From its signal versus dose characteristics, IR-RF dating is the only method to date sediments of older interglacials, up to 300–350 kyr (MIS 9). Independent age control, which is currently available up to MIS 7, shows a good agreement with IR-RF data (Fig. 11.1). It should be noted that the IR-RF signal is (although slightly) less optically sensitive compared to OSL/IRSL (Trautmann, 1999). Lower optical sensitivity may result in insufficient bleaching, that is, in an incomplete emptying of electronic traps. As a result, the luminescence signal is only partially reset and the apparent age becomes too old. This effect has to be considered mainly in sediments showing non-aeolian transport like interglacial limnic sediments. Since no general method exists for doubtless detection of partial bleaching, it should not be excluded from the discussion of the dating results. Single-grain methods possess in principle the capability to detect partial bleaching by investigating the palaeodose distribution for a large number of grains. For this, the dosimetric conditions on the microscale must be simulated to separate their contribution from the total dose distribution. Although loess/palaeosol sequences, fluvial terraces and coastal sediments provide information on warm stage conditions in continental sedimentary archives, a detailed reconstruction of past Quaternary climate is 300

Age by independent method (kyr)

methods, the IR-RF dose characteristics and consequently the obtained dose values can be obtained with much higher precision. In addition to the advantages of SAR or SGR dose protocols in IR-RF dating, the artificial calibrated radiation dose is applied continuously while measuring the luminescence. Therefore, the palaeodose can be determined on the basis of several tens to a few hundreds IR-RF versus radiation dose points (usually < 10 in TL or OSL/IRSL dating), forming a mathematically well-defined stretched exponential dose–response curve. The final information (i.e. the palaeodose corresponding to the ‘natural’ signal) is basically the same as for other luminescence methods. The RF-based dating technique has been described in detail in Trautmann et al. (2000), Erfurt et al. (2003) and Erfurt and Krbetschek (2003).

159

250

200 Erfurt (2003): IR-OSL U/Th 14 C

150

100

China /Qilian Shan (Hetzel et al., 2004): 10Be/ 21Ne

50

This work: OSL

0 0

50

100

150

200

250

300

IR-RF age (kyr)

Fig. 11.1 Comparison of age data from IR-RF measurements with independent dating methods.

160

D. Degering and M.R. Krbetschek

often only possible in limnic sediment basins. Such deposits are well suited for the application of palynological and geochemical analyses with high sample density, necessary to describe the climate quantitatively and continuously over time. Varved sediments might even allow up to one-year resolution. Long records may also contain sediments of more than one interglacial stage. Interglacial sedimentation usually starts with the deposition of sands and silts with no or low organic content. They represent cool conditions at the end of the cold stage. Sedimentation during the interglacial is dominated by organic rich deposits like muds, mostly rich in carbonates, and peats. Cooling at the end of a warm stage generally leads again to inorganic sedimentation, often interrupted by organic rich layers deposited under interstadial climate conditions. This primary sedimentation is strongly influenced by water chemistry, sediment load and by the hydrological conditions (precipitation, groundwater flow and surface water inflow). Regarding luminescence dating, also other sediment components may play an important role. An appreciable aeolian (optically well bleached) component within primary limnic sediments can be assumed since the sedimentation rate in such basins is often relatively low. On the other hand, such continuous sequences can be disturbed by small landslides, even in compact layers, which disturb the stratigraphy and have no defined zero-point or represent an older stage of sedimentation because optical bleaching could not or only partially take place. An open problem remains the occurrence of permafrost during the glacials. The volume increase of water at freezing amounts to about 9%. This may lead to a loosening of the sediment and thus to a decrease of the external dose rate. The second effect is the formation of ice lenses in fine-grained material which is connected with a water flow directly at the grain surfaces. This can lead to the transport of radionuclides and to leaching of the concerned grain surfaces. The processes, although well

known for permafrost soils, are at present not considered in the common dose rateestimation procedures. 11.4 RADIOACTIVE DISEQUILIBRIA The special conditions of sedimentation described above require the consideration of exchange processes at solid/liquid interfaces. This involves the interactions of the aqueous phase with mineral grain surfaces (weathering and sorption) as well as with organic material (complexation and sorption) and also the solubility. Since these processes are very specific, sedimentary systems often show to a large extent differentiation between elements. The natural radioactive decay series of 238,235 U and 232Th are typical examples of such multi-element systems. Since the 238U and 232Th series determine besides 40K the external dose rate in luminescence-dating procedures, such differentiation processes influence strongly the age calculations. Table 11.1 summarises some important members of these two decay series together with their chemical and decay properties. In the case of an undisturbed geological object, the activities within the decay series tend to reach the ‘radioactive equilibrium’, i.e. the state of equal activities for all participating nuclides. Differentiation in at least temporarily open systems and migration processes lead to a loss of this balance, to radioactive disequilibria. The time evolution of those disequilibria depends on nuclide and system parameters: . In closed systems the excess or depletion of a daughter nuclide (with a shorter halflife than its mother’s) disappears according to their decay constant, i.e. after seven half-lives the difference is reduced to < 1% of its original value. . In open systems, a stationary state is reached when the decay rate (activity) of the excess/depletion is identical with the in/output rate of the respective nuclide.

Dating of Interglacial Sediments

161

Table 11.1 Characteristics of members of the 238U and 232Th decay series Decay series

Nuclide

Half-life (Firestone and Ekstro¨m)

Important chemical properties

238

238

4:468109 yr 2:455105 yr

U series

Th Ra 222 Rn 210 Pb

7:538104 yr 1:6103 yr 3.8235 days 22.3 yr

Actinide, heavy metal, forms organic complexes, valence states U(IV) (low solubility) and U(VI) (soluble as uranyl ion ðUO2 Þ2þ ) Actinide, tetravalent, low solubility Earth alkali metal, soluble Inert gas Heavy metal, forms organic complexes

232

1:4051010 yr 5.75 yr 1.9116 yr

cf. 230Th cf. 226Ra cf. 230Th

234

U U

230 226

232

Th series

Th Ra 228 Th 228

Therefore, radioactive disequilibria in closed systems are only detectable and important for dose rate calculations if the half-life of the daughter is of the order of magnitude of the sample age, e.g. for interglacial sediments MIS 5e larger than about 10 kyr. As a consequence, it is only necessary to consider 234U/230Th disequilibria in this case. In open systems, parameters responsible for the migration behaviour like solubility and reactivity are essential. In a sediment layer especially radium and, under oxidising conditions, uranium show a tendency to move. But also processes like continuous outgassing of inert 222Rn or input of the radon successor 210 Pb into the uppermost sediments can lead to distortions of the equilibrium. In general, it is necessary to study carefully the geochemical situation at sampling sites showing radioactive disequilibria. In some cases, it requires testing of several possible model approaches. The radioactivity analysis of the sampled interglacial sediment should also include the determination of most of the long-living natural radionuclides. Low-level, high-resolution gammaspectrometry will often fulfil most of these demands. Alphaspectrometric measurements of the uranium and/or thorium nuclides are time consuming and should be used only in cases where gammaspectrometric results are below the detection limits.

11.5 EXPERIMENTAL Samples were taken from outcrops by driving opaque cylinders into the sediment surface or from cores as ‘D’-shaped slices under laboratory save light. In both cases, the outermost, light exposed layer was removed. In our laboratory, the material was wet sieved, organics and carbonates were removed. Quartz and potassium feldspars were extracted from the 100200 mm grain fraction by a combination of density separation (2.62–2.67 g/cm3 for quartz, 2.53–2.58 g/cm3 for K-feldspar) and flotation. A layer of about 20 mm was etched away at the grain surfaces to suppress the influence of the external alpha dose and to remove surface contaminations. A polymineralic fine-grain fraction was gained by sieving in an ultrasonic bath down to the 510 mm range. All work was performed under dimmed red light. The moisture content of the sediments was determined by drying the material at 60 C for about 12 h. To estimate the saturation moisture, samples were elutriated, then centrifuged and the overlying liquid was decanted. For the OSL measurements, aliquots were prepared by fixing some milligrams of the material by silicone oil onto aluminium sample disks. Monolayers of K-feldspar coarse grains, fixed on an adhesive tape in special sample holders, were used for RF measurements.

162

D. Degering and M.R. Krbetschek

The optically stimulated luminescence signals were recorded by a Risø automated luminescence reader DA 15 equipped with a 90 Sr/90Y beta source for the irradiations. Dating on quartz coarse grains was carried out according to the SAR protocol by Murray and Wintle (2000). The aliquots were stimulated with high-power blue LEDs at 470 nm. Separation of the quartz luminescence in the near UV from stimulation light was achieved by using a Hoya U 340 filter (peak wavelength 340 nm, FWHM 70 nm). All measurements were made at 125 C. K-feldspars were investigated by applying a multiple aliquot additive dose (MAA) protocol and a SAR protocol according to Preusser (2003), respectively. The grains were stimulated in both cases by LEDs in the near-infrared at 870 nm, the blue emission of the sample at 410 nm was extracted from the whole spectrum by an interference filter and the measuring temperature was set to 30 C. The polymineralic fine-grain fraction was dated by the MAA protocol under the same measuring conditions like for K-feldspar coarse grains. Anomalous fading was tested on aliquots stored for at least 12 months. We used an equipment developed in our laboratory as described by Erfurt et al. (2003) for recording the near-IR-RF emission of coarse-grain K-feldspars. The samples were in close contact to underlying 137Cs sources. A Hamamatsu GaAs:Cs photomultiplier (PMT)-type R943-02 in combination with an interference filter measured the fluorescence emission at 865 nm. The IR-SAR protocol proposed by Erfurt and Krbetschek (2003) was applied for the reconstruction of the signal-dose dependency. According to Du¨tsch and Krbetschek (1997), the internal potassium concentration of feldspar grains can be analysed by studying the peak position of the red radiophosphorescence at about 720 nm. Afterglow spectra were measured by a CCD camera-based spectrometer (Rieser et al., 1994), if necessary. In our investigations on interglacial sediments, we used mainly gammaspectrometry

to analyse the radionuclide contents and to detect possible radioactive disequilibria. For the analysis, two low-level gammaspectrometry systems equipped with a 36% p-type and a 38% n-type HPGe detector were used, respectively. The sample material was filled into gas-proof measuring containers of cylindrical shape or so-called Marinelli beakers which cover the cylindrical detector on each side. Measuring times up to 48 h were necessary in most cases. The effect of self-attenuation was taken into account for the activity calculation. Generally, the following long-living nuclides could be determined (in some cases by using gamma-emitting short-living daughter nuclides): 238U, (230Th), 226Ra, 210Pb from 238 U series, 228Ra, 228Th from 232Th series and 40K. To investigate the time evolution of radionuclide disequilibria, a mathematical model was developed which allowed considering three characteristic cases: (1) Closed system without any mass transfer to the environment; the present activity ratios are directly coupled to their initial values during the formation of the object, (2) Open system with supply/discharge of one or more radionuclides from the decay series; the amount Q of the activity exchange rate is constant in time, (3) Open system with supply/discharge of one or more radionuclides from the decay series; characterised by parameter P as the ratio of the activity with mass transfer and the activity without mass transfer (i.e. the exchange is proportional to the number of nuclei). According to the model, each activity term can be written as a superposition of exponential decays: Ai ðtÞ ¼

i X Qj j¼1

j

j

P ðtt0 Þ

þ bij e

j

Dating of Interglacial Sediments i with bij ¼ i =Pi bði1Þj 8 j < i; j =Pj i iP 1 P Q bii ¼ Ai ðt0 Þ jj bij

j¼1

Ai t t0 i Qi, Pi

j¼1

– activity of nuclide number i in the decay series – time – moment of the activity analysis – decay constant of nuclide i – parameters for open systems as described above

In the framework of this model, the resulting external dose rate becomes time dependent. Therefore, the sample age is no more the simple ratio of palaeodose to dose rate and must be calculated by an iterative method (Degering et al., 2006).

11.6 AGE DETERMINATION ON INTERGLACIAL SEDIMENTS – CASE STUDIES Within the joint project ‘EEM – Climate Change at the very end of a warm stage’, part of the German climate research program DEKLIM, several groups investigated a number of interglacial deposits mainly from Middle and Eastern Europe. Age data obtained on these sediments are summarised in Appendix A. We present in this chapter some typical examples to illustrate the peculiarities of luminescence dating on interglacial sediments. A special attention was turned on the application of IR-RF for dating sediments from MIS 5e and older. 11.6.1 The Eemian sequence at Klinge The limnic sediment sequence at the section Klinge is situated in eastern Germany and is correlated with the Eemian interglacial. A detailed description of this palaeoclimatic record is given in Knetsch et al. (2005). The interglacial is represented by a large organic rich layer stacked on a postglacial clay layer. Only the sediment from the bottom of the

163

sequence, just above the clay layer, yielded sufficient coarse grain material. All other samples were predominantly fine grained. Owing to the potential for comparing results of different, independent luminescence methods, our work was focused on the two samples from lowest layer, Kli1 and Kli8, respectively. A schematic presentation of the sequence together with the results of the radionuclide analyses and the sample positions is given in Fig. 11.2. Compared to the upper part of the sequence, the sandy layer, including samples Kli1 and Kli8, showed obvious radioactive disequilibria in the 238U series (238U excess over 226Ra) and furthermore higher activity ratios of 238U, 226Ra and 40K, respectively, to 232Th. To understand the disequilibria in the 238U series, two models were tested: (i) Additional supply of 238,234U via the aqueous path into the sediment during its deposition. The present 226Ra content stands for the concentration of its mother nuclide 230Th (below detection limit of gammaspectrometry in this case). After its formation, the system was closed and the 234U/230Th imbalance vanishes with time. Maximum ages can therefore be determined by assuming a zero initial enrichment of 230 Th. They were found to be within the confidence limits not conflicting with the assumption of an Eemian sequence. A point of criticism for this model is the missing distinct 238U excess in the mud and peat layers since organic matter is known for its high uranium affinity. (ii) Lowered groundwater table during the operation of an adjacent open-cast pit (‘Jaenschwalde’). In reducing environments like peat, uranium is often found in the insoluble U(IV) state. If the organic matter is no more water saturated, the chemical conditions may change considerably. For instance, the

164

D. Degering and M.R. Krbetschek Specific activity (Bq/kg) 5

0

200

400

600 40

0

5

Ratio 10 15

20

5

40

K/ 232Th

K

Kli7 4

Peat

Organic-rich mud Sandy silt Clay

Depth scale (m)

Silty clay

4

4

3

3

Kli6 Kli5

238

2

U series: 238 U 226 Ra

238

2

U/ 232Th 226 Ra/ 232Th

2 Kli4

1

232

Th series: Total

0

1 0

0 0 10 20 30 40 50 60 70 0.0 Specific activity (Bq/kg)

0.5

1.0 1.5 Ratio

2.0

Kli2 Kli1 Kli8

2.5

Fig. 11.2 Overview of the stratigraphic sequence of the Eemian deposits at Klinge together with the results of the radionuclide analyses and the position of the samples taken. The right part shows the depth dependence of several ratios between natural radionuclides. A dotted line indicates the presumed 238U/ 232Th in the detritus.

start of oxidation results in the formation of the soluble uranyl ion ðUðVIÞO2 Þ2þ . Also radium may be released in consequence of this process. The 238U- and 226Ra- accumulation in Kli1 and Kli8 reflects a recent event. For almost the complete age, the 238U series was in equilibrium and the activities remained at a constant, not elevated level. Its value was reconstructed from the 232Th activity, assuming that the 238U/232Th ratio of the above sediment layers represents that of the mineral component, the detritus level. Nevertheless, a clear U/Ra loss was not found in the overlying layers. The difference in the age results using the different approaches amounted to about 2–3 kyr and was therefore definitely within the limits of measuring uncertainties. All the results given below are based on assumption (ii). Palaeodose determinations on Kli1 and Kli8 were carried out by quartz-SAR, Kfeldspar-SAR and IR-RF. It was found that the scattering of the K-feldspar and the IRRF results was surprisingly high. The potassium content of some aliquots as analysed

by the red radiophosphorescence varied between 5 and 13%, in contrast to the structurally given value of 12.5% for K-feldspar. Thus, an average K-concentration of 8:5 2:0% was set for all age calculations. All samples showed moisture content well below saturation level, as determined by laboratory experiments. Probably the groundwater was lowered rather late (due to open-cast mining) so water saturation was assumed for age calculation. Table 11.2 gives a summary of the parameters used for dating and the individual results for different measurements. The age results using quartz-SAR with their large uncertainties emphasise the influence of the nearly saturated luminescence signal. On the other hand, feldspar results vary due to variations in internal potassium content. Overall, the results confirm the classification of the investigated layer as sediments deposited at the start of the Eemian. 11.6.2 The limnic sediment sequence at Cheremoshnik/Russia The limnic sequence Cheremoshnik 3 is situated in central Russia near Rostov (NE

Dating of Interglacial Sediments

165

Table 11.2 Summary of parameters used for luminescence dating on outcrop Klinge Sample

Moisture (%)

Kli1

49

Cosmic dose rate (mGy=a) 110´ 11

Kli8

51

110´ 11

Luminescence method Quartz-SAR IR-RF on K-feldspar IR-RF on K-feldspar

Palaeodose (Gy) 128´ 34 231.7´ 9.3 203´ 13

Age (kyr) 104´ 28 137´ 15 120´ 15

Quartz-SAR IR-RF on K-feldspar IR-RF on K-feldspar K-feldspar-SAR

174´ 38 277´ 22 228´ 13 224´ 18

134´ 30 158´ 21 130´ 15 127´ 16

The model of recent U/Ra migration was applied for age calculation.

of Moscow). Interglacial sediments above a Saalian till were palynologically classified as Eemian deposits (Grichuk, 1982). Owing to appearing radioactive disequilibria and the stratigraphic conditions, it gives a further example for special features appearing in dating interglacial deposits. A schematic overview of the stratigraphy and the radionuclide profile is shown in Fig. 11.3. From palynological and lithological indications, it was concluded that only the region below the clay layer at 2.9 m depth is undisturbed. The upper part is presumably disturbed by a landslide (E.Y. Novenko, personal communication) so reliable external dose calculations are not possible. Therefore, just dating results of sample CHE 3-1 are shown here.

Radionuclide concentrations at this site were distinctly elevated in the 238U series, especially in the organic mud. This allowed the measurement of 230Th by gammaspectrometry, which enabled better constraining the evolution of the 238U-series disequilibria. In the lower part of the section, the 226Ra activity exceeded that of 238U, whereas in the organic mud it falls below that of 230Th. If the clay layers CHE 3-1A and CHE 3-5 represent the detritus value of the 238 U/232Th ratio, then uranium accumulation appears in all organic parts but to a different extent. A special enhancement is found in the detrital organic mud (containing sample CHE 3-4). The mineral content in this part of the section is very small, as

0

0 232

Th series: 236U series: 238 U Total 230 Th 226 Ra

Silt

Clay Coarse detrital organic mud Silt (clayey) Clayey sand

Depth scale (m)

Detrital organic mud

U/ 232Th 226 Ra/ 232Th

1

1

2

2

2

3

3

3

1

Clay

238

CHE 3-5

CHE 3-4

4 0

100 200 300 0 Specific activity (Bq/kg)

4 5 10 15 20 25 30 Ratio

CHE 3-3 CHE 3-2 CHE 3-1 CHE 3-1A

Fig. 11.3 Overview of the stratigraphic sequence of the Eemian deposit at Cheremoshnik 3 together with the results of the radionuclide analyses and the position of the samples taken. CHE 3-1A was used for radionuclide analysis only. The right part shows the depth dependence of several ratios between natural radionuclides. A dotted line indicates the presumed 238U/ 232Th in the detritus.

166

D. Degering and M.R. Krbetschek

indicated by the low 232Th content. Under the assumption of a closed system, an estimation of the age by the 238,234U/230Th disequilibrium in CHE 3-4 gives 106 27 kyr. The difference between 226Ra and 230Th found for this sample cannot be explained within a closed-system model. Using an open-system model, a migration of radium in the sediment column would be allowed. With that, radium was sorbed and accumulated in the permeable silt (CHE 3-1) just above the damming clay. Such a process could only start after the outcrop was formed by a valley incision, i.e. probably within the Holocene. This time period is fortunately too short for a larger modification of the mean dose rate in the past. For calculations of the external dose rate in CHE 3-1, it was presumed that the detritus had the same 238U/232Th ratio as for CHE 3-1A and CHE 3-5 and that uranium was additionally accumulated in the organic phase leading to an initial 234U/230Th imbalance. The start of the radium accumulation was assumed to be at 5 kyr before present. The influence of the different radioactivity contents in CHE 3-1 and CHE 3-1A for the calculation of the gamma dose rate was taken into account as a further step. The applied procedure is described in Aitken (1985). By considering the mentioned procedure in dose rate determination and further parameters listed in Table 11.3, an age of 132 13 kyr was calculated and confirmed the idea of an early Eemian sequence in the lower part of Cheremoshnik 3. 11.6.3 The Kieselgur (diatomite) deposit at Munster The Kieselgur deposits at Munster/Breloh located in the Lu¨neburg Heath of Lower

Saxony (Germany) were formed in lakes within glaciofluvial sands. On the basis of the pollen profiles they are described as a Holsteinian interglacial sequence (Mu¨ller, 1974). A recently drilled core was investigated by the University of Mainz (Diehl et al., 2004). The lower part of the core contains well-laminated sediments. In its upper part, the Kieselgur is mixed with mineral and organic components. It ends abruptly and is overlain by a sandy sediment which contains a little admixed Kieselgur (cf. Fig. 11.4). Material for luminescence dating was taken in the vicinity of the Kieselgur–sand transition and was analysed by IR-RF. The low radioisotope content of the diatomite and the sand resulted in a low external annual dose and in a relatively low total palaeodose. As a consequence, the natural RF signal was far from the saturation level of the regenerated signal-to-dose function (Fig. 11.4) and thus, the uncertainty in the palaeodose determination was low. Parameters for the age calculation are listed in Table 11.4 and 11.5 together with the dating results. The groundwater table was approximately at 5 m below surface (F. Sirocko, personal communication), so the saturation moisture, as determined in laboratory experiments, was used for the calculations. At a first glance, the saturation moisture of sample MU 3 (diatomite) is surprisingly high (212%). It can be easily explained, however, by the definition of the moisture content as the ratio of water mass to dry mass and by the very low bulk density of Kieselgur. Within its uncertainty range, the age of the Kieselgur sample MU 3 covers perfectly the oxygen isotope stage 9 (e.g. Jouzel and EPICA Community Members, 2004).

Table 11.3 Summary of parameters used for luminescence dating on outcrop Cheremoshnik 3 Sample

Moisture (%)

Cosmic dose rate (mGy=a)

Luminescence method

Palaeodose (Gy)

Age (kyr)

CHE 3-1

23

128 13

IR-RF on K-feldspar

399 36

132 13

The model of 234U/230Th disequilibrium plus a recent 226Ra migration was applied for age calculation.

Kieselgur with a few sandy layers (varved)

RF intensity (arbitrary units)

MU 2 5

10

15

2400 2200 2000 1800 1600 1400

nat. signal

1200 1000 800 600 400 200 0 0

200

400

600

800

1000

Dose (Gy)

MU 3

20

25

Sand

167

9

0

Depth scale (m)

Sand and gravel with clay layers (Kieselgur admixed)

Dating of Interglacial Sediments

10

Fig. 11.4 Left: Overview of the stratigraphic sequence in core MU (Munster) (Diehl et al., 2004). Middle: Photograph of the core section at 9–10 m depth. The regions used for luminescence dating are marked. Right: Signal-to-dose dependency of an aliquot from MU 3 as obtained by a regeneration procedure. The intensity of the ‘natural’ signal is also shown.

Table 11.4 Results of the radionuclide analysis on core MU (Munster) Sample

Specific activity of 238U decay series (Bq/kg)

Specific activity of 232Th decay series (Bq/kg)

Specific activity of 40K (Bq/kg)

MU 2 MU 3

7:39 0:34 15:87 0:55

6:72 0:43 14:40 0:74

190:8 3:9 92:4 3:0

No radioactive disequilibrium was found in the decay series.

Table 11.5 Summary of parameters used for luminescence dating on core MU (Munster) Sample

Moisture (%)

Cosmic dose rate (mGy=a)

Luminescence method IR-RF on K-feldspar IR-RF on K-feldspar

MU 2

20

70 7

MU 3

212

66 7

The Holsteinian has already been correlated with MIS 9 by several authors (see Frechen et al., 2005 and the references cited therein). Sample MU 2 is located just above the Kieselgur section in a pollen free part of the core. In the lower region, the pollen sequence is parallel to those recorded by Mu¨ller (1974) but truncated at 9.5 m depth. The upper part is obviously erosively removed. Therefore,

Palaeodose (Gy) 624´ 19

Age (kyr) 433 24

329.3´ 8.3

334 21

the sandy layer is likely deposited in connection with a Saalian ice advance, whose maximum extension reached to the south of Munster (Liedtke, 1981). To understand the astonishing ‘age reversal’, one should keep in mind that the bleaching conditions for the overlaying sand layers are much more unfavourable than for the limnic Kieselgur material. The mass ratio of

168

D. Degering and M.R. Krbetschek

the grain size fractions >200 mm, 63200 mm and <63 mm, respectively is about 1:1:2 in the Kieselgur sample MU 3. Considering a portion of fine-grained diatomite, this ratio is an indication of an only minor alluvial contribution which would result in the prevalence of one grain size fraction. A major part of the grains in MU 3 is therefore of aeolian origin, and complete bleaching can be presumed. Furthermore, the assumption of insufficient bleaching would result in a correlation of the interglacial sediments with at least MIS 7 which is not supported by the palynological results. If the Kieselgur deposit was formed during an older interglacial (e.g. MIS 11), there must have occurred a signal loss of about one-fourth during 400 kyr. For this effect, no indication is found so far in the age range with independent age control (< 230 kyr, cf. Fig. 11.1). It is assumed that the age of sample MU 3 gives the correct value, and the higher age of the sand layer is the result of partial bleaching. The Holsteinian Kieselgur deposit at Munster/Breloh is thus correlated with MIS 9.

11.7 SUMMARY Luminescence methods are appropriate for dating sediments from past interglacials. The age range and geochemical properties of such deposits make special demands to the used procedures and the accurate age calculation, respectively. The IR-RF of K-feldspar is an especially valuable tool for dating sediments from MIS 5e and older. Special efforts have been made to include the influence of radioactive disequilibria in the determination of the external dose. This requires a careful analysis of the geological conditions at the sampling place. Age determinations on interglacial sediments are illustrated by some case studies which involve the comparison between different methods, the consideration of radioactive disequilibria and the effect of insufficient bleaching.

REFERENCES Aitken, M.J., 1985. Thermoluminescence Dating. Academic Press, London. Aitken, M.J., 1998. An Introduction to Optical Dating – The Dating of Quaternary Sediments by the Use of Photon-Stimulated Luminescence. Oxford University Press, Oxford. Boetter-Jensen, L., McKeever, S.W.S., Wintle, A.G., 2003. Optically Stimulated Luminescence Dosimetry. Elsevier, Amsterdam. Degering, D., Kulig, G., Krbetschek, M.R., 2006. ADELE – a novel software for age determination based on luminescence and electron spin resonance (submitted to Radiation Measurements). Diehl, M., Sirocko, F., Degering, D., Krbetschek, M., 2004. German pre-Eemian interglacial records of Munster Breloh, Bonstorf, Bilshausen, Niedersachsen, and Do¨ttingen, Eifel – Lecture presented at: Third workshop of the DEKLIM-EEM Project ‘‘Climate change at the very end of a warm stage’’, Potsdam, Germany, 8 March–12 March, 2004. Du¨tsch, C., Krbetschek, M.R., 1997. New methods for a better internal 40K dose rate determination. Radiation Measurements 27, 377–381. Erfurt, G., 2003. Radiolumineszenzspektroskopie und -dosimetrie an Feldspa¨ten und synthetischen Luminophoren fu¨r die geochronometrische Anwendung. PhD thesis, Technical University Bergakademie Freiberg. Erfurt, G., Krbetschek, M.R., 2003. A single-aliquot regenerative-dose dating protocol applied to the infrared radiofluorescence (IR-RF) of coarse-grain K feldspar. Ancient TL 21, 21–28. Erfurt, G., Krbetschek, M.K., Bortolot, V.J., Preusser, F., 2003. A fully automated multi-spectral radioluminescence reading system for geochronometry and dosimetry. Nuclear Instruments and Methods in Physics Research (B) 207, 487–499. Firestone, R.B., Ekstro¨m, L.P. LBNL Isotopes Project – LUNDS Universitet, WWW Table of Radioactive Isotopes, http:/ie.lbl.gov/toi/. Frechen, M., Sierralta, M., Oezen, D., Urban, B., 2005. Uranium-series dating of peat from Central and Northern Europe. In: Sirocko, F., Litt, T., Claussen, M., (eds.), The Climate of Past Interglacials, Developments in Paleoenvironmental Research (this volume). Grichuk, V.P., 1982. Vegetation of Europe during Late Pleistocene. In: Gerasimov, I.P., Velichko, A.A. (eds.), Paleogeography of Europe During the Last One Hundred Thousand Years. Nauka, Moscow, 79–85. Hetzel, R., Tao, M., Niedermann, S., Strecker, M.R., Ivy-Ochs, S., Kubik, P.W., Gao, B., 2004. Implications of the fault scaling law for the growth of topography: mountain ranges in the broken foreland of north-east Tibet. Terra Nova 16, 157–162.

Dating of Interglacial Sediments Jouzel, J., EPICA Community Members, 2004. EPICA Dome C Ice Cores Deuterium Data. IGBP PAGES, World Data Center for Paleoclimatology, Data Contribution Series # 2004-038. NOAA/NGDC Paleoclimatology Program, Boulder CO, USA. Knetsch, S., Bo¨ttger, T., Junge, F.W., Morgenstern, P., 2005. Element- und isotopengeochemische Untersuchungen am eemzeitlichen limnischen Profil Klinge, Niederlausitz. Natur und Landschaft in der Niederlausitz (in press). Liedtke, H., 1981. Die nordischen Vereisungen in Mitteleuropa. Zentralausschuß fu¨r deutsche Landeskunde, Selbstverlag, Trier. Mu¨ller, H., 1974. Pollenanalytische Untersuchungen und Jahresschichtenza¨hlungen an der holsteinzeitlichen Kieselgur von Munster-Breloh. Geologisches Jahrbuch A 21, 107–140. Murray, A.S., Wintle, A.G., 2000. Luminescence dating of quartz using an improved single aliquot

169

regenerative-dose protocol. Radiation Measurements 32, 57–73. Preusser, F., 2003. IRSL dating of K-rich feldspars using the SAR protocol: Comparison with independent age control. Ancient TL 21, 17–23. Rieser, U., Krbetschek, M.R., Stolz, W., 1994. CCDcamera based high sensitivity spectrometer. Radiation Measurements 23, 523–528. Trautmann, T., 1999. Radiolumineszenzuntersuchungen an Feldspat. PhD thesis, Technical University Bergakademie Freiberg. Trautmann, T., Krbetschek, M.R., Dietrich, A., Stolz, W., 2000. The basic principle of radioluminescence dating and a first model approach. Radiation Measurements 32, 487–492. Wintle, A.G., 1997. Luminescence dating: laboratory procedures and protocols. Radiation Measurements 27, 769–817.

170

Appendix A Summary of dating results and data used for age calculation Sample

In-situ water content (%)

Saturation water content (%)

Applied Cosmic Specific activity in sediment (Bq/kg) water dose rate 232 40 content (mGy=a) 238U Th K (%)

15´ 2

49´ 5

49´ 5

110´ 11

Kli4 Kli7 Kli8

49´ 5 7´ 1 14´ 2

89´ 9 56´ 6 51´ 5

89´ 9 56´ 6 51´ 5

131´ 13 34.89´ 0.61 42.8´ 1.3 669.6´ 9.2 192´ 19 35.52´ 0.78 46.2´ 1.5 699´ 11 ´ ´ ´ 110 11 33.3 2.3 21.94 0.97 379.3´ 7.1

Kli9

19´ 2

47´ 2

47´ 2

110´ 11 24.44´ 0.95 16.92´ 0.79 327.1´ 1.4

Site Pljos, Russia Pl 1/02-1 Pl 1/02-2

14´ 2 64´ 6

30´ 3 107´ 11

19´ 2 75´ 8

Pl 1/02-3

32´ 3

56´ 6

39´ 4

109´ 11 25.69´ 0.63

30.9´ 1.4

490.5´ 2.3

Pl 1/02-4

23´ 2

51´ 5

35´ 4

158´ 16 19.48´ 0.54

20.6´ 1.1

396.8´ 2.7

Pl 1/02-5

18´ 2

60´ 6

25´ 3

202´ 20 30.92´ 0.74

35.5´ 1.6

410.9´ 2.1

Pl 1/02-6

65´ 7

131´ 13

79´ 8

125´ 13 21.39´ 0.64

23.4´ 1.2

396.0´ 3.1

79´ 8 86´ 9

27.8´ 2.4

18.73´ 0.87 355.9´ 6.8

15.22´ 0.42 16.23´ 0.82 402.0´ 1.0 38.7´ 0.9 25.0´ 1.2 462.7´ 3.0

Palaeodose Age (kyr) (Gy)

IR-RF IR-RF Q-SAR

231.7´ 9.3 203´ 13 128´ 34

FG-MAA FG-MAA IR-RF IR-RF FS-SAR Q-SAR IR-RF

271´ 49 307´ 33 277´ 22 228´ 13 224´ 18 174´ 38 205´ 14

Remarks

137´ 15 Radioactive disequilibrium, 120´ 15 K-content in 104´ 28 K-feldspars only 8.5´ 2.0%, layered geometry 108´ 22 93´ 13 158´ 21 Radioactive disequilibrium, 130´ 15 K-content in 127´ 16 K-feldspars only 134´ 30 8.5´ 2.0% 124´ 15 Radioactive disequilibrium, K-content in K-feldspars only 8.5´ 2.0%, layered geometry

IR-RF 342´ 16 157.3´ 9.2 IR-RF 323´ 109 157´ 54 FS-MAA 322´ 41 156´ 22 ´ IR-RF 316 27 124´ 12 FG-MAA 402´ 138 154´ 54 Q-SAR 106´ 45 57´ 24 IR-RF 216´ 59 101´ 28 IR-RF 146´ 19 68.2´ 9.3 FS-SAR 153´ 12 71.4´ 6.5 FS-MAA 200´ 44 93´ 21 Q-SAR 156´ 61 100´ 39 FG-MAA 195´ 36 63´ 13 Q-SAR 139.6´ 9.4 65.4´ 6.5 IR-RF 212´ 34 119´ 20 ´ FS-MAA 200 59 113´ 34

D. Degering and M.R. Krbetschek

Site Klinge, Germany Kli1

Dating method*

Site Cheremoshnik 3, Russia CHE 3-1 17´ 2

CHE 3-2

40´ 4

108´ 11 189´ 19

Site Cheremoshnik 5, Russia CHE 5-1 65´ 7

23´ 2

128´ 13

85.1´ 4.3

27.5´ 1.4

444.1´ 3.1

IR-RF

399´ 36

122´ 12 138´ 14

59.7´ 2.9

19.1´ 1.0

309.9´ 2.5

FG-MAA

215´ 34

470.7´ 3.4

IR-RF FS-MAA FG-MAA IR-RF

543´ 32 425´ 90 336´ 69 377´ 20

558.8´ 7.9 411.9´ 5.9 550.7´ 8.8 425.9´ 6.8

IR-RF IR-RF IR-RF IR-RF

473´ 22 439´ 18 699´ 34 732´ 44

156.0´ 9.1 185.7´ 9.6 246´ 15 302´ 21

677´ 34 680´ 34 538´ 33 602.0´ 8.8 236.9´ 1.9

FG-MAA 441´ 49 FG-MAA 517´ 62 FG-MAA 166.8´ 9.5 IR-RF 307´ 29 IR-RF 184´ 29

98´ 15 117´ 18 44.5´ 5.1 87.5´ 9.4 107´ 20

131´ 13

79´ 8

134´ 13

41.1´ 1.0

25.6´ 1.3

7.21´ 0.33

7.11´ 0.44 280.3´ 4.8

15´ 2

50´ 5

30´ 3

143´ 14

Site Gaxun Nur, China I 70-3 I 70-19 I 70-26 I 70-35

20´ 2 11´ 2 21´ 2 21´ 2

54´ 5 37´ 4 34´ 3 38´ 4

20´ 2 11´ 2 21´ 2 21´ 2

50´ 5 19´ 2 12´ 1 7´ 1

Eifel dry maars, Germany U2-104,5 m 18´ 2 U2-122,5 m 18´ 2 WD 42.05 m 19´ 2 HL 2–4 (42.15–42.30 m) 16´ 2 HL 2–6 (53.31–53.78 m) 56´ 6

50´ 5 53´ 5 51´ 5 69´ 7 85´ 9

25´ 3 25´ 3 33´ 3 16´ 2 56´ 6

1.0´ 0.1 48.9´ 2.5 1.0´ 0.1 46.7´ 2.3 9´ 1 45.6´ 1.9 0.9´ 0.1 36.41´ 0.69 0.6´ 0.1 16.41´ 0.50

Core Munster MU 2 (9.00–9.25 m) MU 3 (9.65–9.95 m)

15´ 2 20´ 2 ´ 138 14 212´ 21

24.32´ 0.58 36.0´ 1.3 13.86´ 0.38 17.67´ 0.70 22.14´ 0.72 31.7´ 1.4 18.77´ 0.59 26.2´ 1.1 51.1´ 2.6 50.3´ 2.5 49.9´ 1.6 55.8´ 1.7 21.3´ 1.1

256´ 21 200´ 44 154´ 35 217´ 16 Unclear radioactive disquilibria in adjacent organic layers, layered geometry

20´ 2 212´ 21

70´ 7 66´ 7

7.39´ 0.34 6.72´ 0.43 190.8´ 3.9 15.87´ 0.55 14.40´ 0.74 92.4´ 3.0

IR-RF IR-RF

624´ 19 329.3´ 8.3

433´ 24 334´ 21

Core Bonstorf BT 2 (10.30–10.50 m) BT 3 (11.25–11.45 m)

21´ 2 52´ 5

25´ 3 62´ 6

25´ 3 62´ 6

62´ 6 57´ 6

3.30´ 0.16 2.91´ 0.19 131.5´ 2.1 10.62´ 0.41 11.56´ 0.57 197.6´ 4.0

IR-RF IR-RF

341´ 15 350´ 12

297´ 19 265´ 17

Core Bilshausen Bi 1-1 (13.38–13.48 m)

25´ 3

38´ 4

38´ 3

47´ 5

43.28´ 0.87

IR-RF

1235´ 56

303´ 19

50.0´ 1.8

1011´ 13

Dating of Interglacial Sediments

CHE 5-2

132´ 13 Radioactive disequilibria in sample and adjacent layer, layered geometry 140´ 27 Radioactive disequilibria in sample and adjacent layer, layered geometry

*

171

Explanation of abbreviations: FS – K-FeldSpar coarse grains ðmainly 90160 mmÞ. FG – polymineralic fine grains ð510 mmÞ. Q – quartz coarse grains (mainly 90160 mm). MAA – multi aliquot additive dose protocol. SAR – single aliquot regenerative dose protocol. IR-RF – InfraRed RadioFluorescence on K-feldspar coarse grains (mainly 90160 mm), SAR protocol.

This page intentionally left blank

12. Neanderthal Presence and Behaviour in Central and Northwestern Europe During MIS 5e Stefan Wenzel Forschungsbereich Altsteinzeit des Ro¨misch-Germanischen Zentralmuseums Mainz, Schloß Monrepos, D-56567 Neuwied, Germany

ABSTRACT The dense interglacial forests of Central and Northwestern Europe are considered to have been less productive environments and therefore less favourable habitats for humans than the grasslands of glacial Europe. However, there are several sites which attest to the presence of hominids from the beginning of MIS 5e to its climatic optimum (Quercetum mixtum-Corylus phase). So far, there is no evidence of hominids during the Carpinus phase, and only a few archaeological sites are known from younger biozones of the Eemian. Britain seems to have been totally unpopulated due to its island position. The Early Neanderthals inhabiting Europe hunted big game and presumably cached goods and practised symbolic behaviour. Artefacts made of material found more than 60 km away from the outcrops are indicative of mobility and social networks no different to those of Neanderthals living under more continental climatic conditions. 12.1 THE EEMIAN INTERGLACIAL The last interglacial sensu stricto or Eemian (Fig. 12.1) (MIS 5e) covers approximately the time between 128 000 and 115 000 years before present (BP). Two shorter and less intensive warm phases separated by cool and dry phases followed the Eemian and preceded the first maximum of the last cold stage (Reille et al., 2000; Martrat et al., 2004). The climate of the last interglacial was warmer than the Holocene. This is indicated by sea levels 3–8 m higher than today

(Bruggemann et al., 2004; Siddall and Chappell, this volume), along with the presence of animals and plants in Central Europe whose ranges in the Holocene do not extend as far north. Examples of such plants are box tree (Buxus sempervirens) and honeysuckle Lonicera arborea (Schweigert, 1991). Exotic inhabitants of Central Europe in the last interglacial period are hippopotamus (Hippopotamus amphibius), water buffalo (Bubalus murrensis) and fallow deer (Dama dama) (von Koenigswald, 2002). Indicators of warm climatic conditions familiar to us are roe deer (Capreolus capreolus) and wild boar (Sus scrofa), while straight-tusked elephant (Palaeoloxodon antiquus) and Merck’s rhinoceros (Stephanorhinus kirchbergensis) became extinct at the end of MIS 5 (Rink et al., 1996). Even if Europe north of the high mountain chains was settled by hominids for the first time approximately 600 000 years ago in a phase of similarly warm climate (MIS 15a), the interglacials were in each case only comparatively short episodes in which the hominids had to adapt to changes from an open to a wooded landscape. The present paper aims to describe the archaeological record of the last interglacial and to examine the ways in which such adaptations to new environments could have occurred. 12.2 CONDITIONS FOR THE PRESERVATION OF SITES FROM THE LAST INTERGLACIAL During the last interglacial period, processes of weathering and soil formation

174

Stefan Wenzel

11

20

12

19 17 1316 18 14 15

2

1

8

21 10 9

26

27 30 28 29

22 24

23

5 4 3 6 25 Travertine deposits

6

Caves Lake basins

7

River deposits Raised beaches

Fig. 12.1 Map of last interglacial sensu stricto (MIS 5e) sites with evidence of hominid presence. Travertine deposits: 2 Caours; 10 Stuttgart-Untertu¨rkheim; 11 Stuttgart-Bad Cannstatt, Seelberg; 12 Steinmu¨hle near Veltheim; 13 Burgtonna; 14 Weimar, Parktravertin; 15 Taubach; 26 Bojnice III; 27 Ga´novce; 28 Hoˆrka-Ondrej (Kaminska´ et al., 2000); 29 Hranovnica-Hincava (?); 30 Beharovce-Sobotisko (?); caves: 3 St Marcel, level u (Moncel et al., 2004); 4 Payre, level D (Bouteaux, 2003); 5 Moula-Guercy, levels XV-XIV (Defleur et al., 2001); 21 Ku˚lna Cave, level 11; 22 Vindija, complex K; 23 Krapina, levels 2–4; 24 Veternica Cave, level j; 25 Crvena Stijna, levels XXIX-XVIII; lake basins: 11 Lehringen (Houben, 2003); 16 Neumark-Nord 1 (?); 17 Rabutz; 18 Grabschu¨tz; 19 Gro¨bern; 20 Pho¨ben (Wenzel, 2002); River deposits: 9 Wallertheim; 8 Saccopastore; Raised beaches: 1 Port Racine; 7 Balzi Rossi (near Mentone), zona dell’ex-Casino (modified from Wenzel 1998, Fig. 1).

took place under forest cover. Nevertheless, erosion was a significant geological phenomenon of this time period, resulting in the loss of upper portions of the soil formation. Owing to this factor, and the comparatively short duration of MIS 5e, sites evincing the presence of hominids in the last interglacial period are rare (Wenzel, 1998; Speleers, 2000; Tuffreau, 2001). There are four groups of sites where the deposits were only minimally affected by erosion:

. Lake basins Lake basins were formed by glacial exaration of troughs by the Saalian continental ice sheet [e.g. Gro¨bern (Wansa and Wimmer, 1990)], by glaciological and cryogenic deformation (diapiric upwelling of brown coal) [e.g. NeumarkNord (Eissmann, 2002; Mania et al., 2004)] or by subrosion of salt and gypsum [e.g. Ko¨nigsaue, Eemian lacustrine basin with Middle Palaeolithic finds in

Neanderthal Presence and Behaviour During MIS 5e

Early Weichselian layers (Mania and Toepfer, 1973)]. Several of these lake basins lie in an area which was only reached by ice in the Drenthe phase of the Saalian glaciation, but not during the Warthe phase. Discussions concerning the dating of those basins overlying the Drenthe basal moraine are closely linked to the heavily disputed question of the existence of an interglacial period between the Drenthe and Warthe ice advances of the Saalian cold stage (Eissmann, 1990, 2002; Lippstreu and Strackebrandt, 2003; and contra: Mania, 2001; Nowel, 2003). Interglacial basin deposits stratified between the basal moraines of the Drenthe and Warthe phases have not been found so far in Central Europe, thus providing a substantial argument against the existence of an Intra-Saalian Interglacial. Furthermore, in some lake basins where Eemian sediments fill out gulleys formerly sealed by dead ice, it is not clear how the dead ice could have survived an interglacial period between the Drenthe and Warthe glaciations (Eissmann, 1990). Fallow deer remains recovered in the lake basin of Neumark-Nord 1 are somewhat larger than those known from the Eemian (Pfeiffer, 1999). However, this possible biostratigraphic indication for a preEemian age is too isolated to assign the lake basin deposits at this site to an otherwise not compellingly proven ‘Intra-Saalian Interglacial’. Further clarification may come from a stratigraphical relation between the interglacial horizons of Neumark-Nord 1 and of the recently discovered basin Neumark-Nord 2 (Mania et al., 2004). . River and beach deposits River and beach deposits are only preserved at sites where they were protected by thick overlying layers (Cliquet, 1992; Adler et al., 2003). . Caves Interglacial layers were occasionally protected by thick layers of frost debris, as at Krapina (Malez, 1978). The attribution of

175

cave sediments to the Eemian is often problematic, due to bioturbation and other processes of redeposition, or the presence of intrusive elements, such as animals indicative of cold climatic conditions in layer 11 at Ku˚lna cave (Valoch, 1988). . Travertine deposits Travertine owes its formation to mineral waters rich in calcium carbonate, which often flow out of the ground at the edge of floodplains (Adam and Berckhemer, 1983). This rock normally contains many cavities, so that it is not subject to frost shattering. According to plant fossils and faunal remains, travertine developed only during interglacial periods. However, there is evidence for the formation of travertine in Central Europe during MIS 5c. The travertine deposit of StuttgartUntertu¨rkheim consists of the mainly solid ‘lower travertine’ and the travertine sand of the ‘upper travertine’ (Fig. 12.2A). These deposits are separated by a disconformity and, in the western part of the site, by an earthy ‘intermediate layer’ containing the remains of steppe rodents such as great jerboa (Allactaga jaculus) and steppe lemming (Lagurus lagurus) (Adam and Berckhemer, 1983). According to Soergel (1940), the intermediate layer consists of two parts. The steppe rodent horizon is a 10–15 cm thick, dark humic layer which contained fine quartz sand, pieces of Keuper marl and small quantities of clay (12% clastic material). According to Koban (1993), it is a limestone rendzina. A layer of grey-violet sediment consisting of 55% clastic material is preserved occasionally above and below the steppe rodent horizon. According to terrace stratigraphy and biostratigraphic indications, the lower travertine can be assigned to MIS 5e, while the steppe rodent layer is assigned to MIS 5d and the upper travertine to MIS 5c (Reiff, 1994; Wenzel, 1994; Braun et al., 1999; Geyh et al., 1999). An alternative chronological model by Frank et al. (2000) places the major

176

Stefan Wenzel

(a)

(b)

Surface ca 233,00 m NN

SW

Lagurus lagurus

W

NE

E

152

I

151

1 150

ut

2

149 3a

srl 148 It

147

4

146 Quarry floor

ca 226, 15 m NN

1m

3

5

1m 145 0

10

20

30 m

Fig. 12.2 A. Stuttgart-Untertu¨rkheim. Profil of the west corner of the Biedermann quarry with tree trunk cavety 1. l: loess or loam, ut: upper travertine, srl: steppe rodent layer, lt: lower travertine. B. The Tata travertine. 1 teterata type travertine; 2 calcareous silt accumulated in teterata basin; 3 eolian sand/loess; 3a Palaeolithic site; 4 terrace gravel and sand; 5 triasic limestone. (a: modified after Adam and Berckhemer, 1983; b: modified after Pe´csi et al., 1988, Fig. 4; Lagurus lagurus after Heinrich, 1990).

part of the lower travertine and all of the upper travertine in MIS 5c. This model is based on 230Th/U mass spectrometry datings and assumes a major disconformity within the lower travertine. However, there was no evidence of a hiatus in the lower travertine during earlier studies of the deposits, when the sections were in a better condition of preservation than in recent years. Nor is it consistent with the occurrence of thermophilous plants even in the upper parts of the lower travertine (Schweigert, 1991). A discontinuance of travertine formation during MIS 5d is also evident in the travertine deposits at Tata (H). At this site, sandy loess was deposited during MIS 5d, and the layer is also characterized by the remains of steppe rodents (Fig. 12.2B) (AllsworthJones, 1986; Pe´csi et al., 1988). A faunal assemblage recovered in a horizontal fissure in the upper part of the travertine sequence of Burgtonna South was found to include extinct porcupine Hystrix cf. vinogradovi, indicative of still warm but open conditions (Maul, 2002; Meyrick and Maul, 2002). This faunal assemblage, as well as terrestrial molluscs from the relatively warm and dry phases 6 or 8 of the molluscan succession from Burgtonna Quarry I (Mania, 1978; Meyrick, 2002), may correlate with the

recently described late Eemian aridity pulse in Central Europe (Sirocko et al., 2005). 12.3 THE ECOLOGICAL BASIS OF HOMINID SETTLEMENT AND THE CHRONOLOGICAL POSITION OF ARCHAEOLOGICAL SITES WITHIN THE LAST INTERGLACIAL During the last interglacial, with its moderate climate and heavy precipitation, hominid occupation was possible as far north as Finland (Schulz et al., 2002). Even so, the Eemian phase did not offer Early Neanderthals in Central Europe particularly good environmental conditions (Roebroeks et al., 1992; Wenzel, 1998), since there is less density and variety of game in forests than in open landscapes. Furthermore, those species of animals which could be targeted during large-scale herd hunting are more or less absent in forests. Even in the ‘shade forest phase’ of MIS 5e, the forests of Central Europe were more open in comparison with those found during the Holocene, prior to human interference. A major factor was that the European Beech (Fagus sylvatica) – the characteristic shade tree in the Holocene – attained only local importance in the Eemian. Saplings of

Neanderthal Presence and Behaviour During MIS 5e

the shade plants of the Eemian, European Hornbeam (Carpinus betulus) and Spruce (Picea sp.) require more light than saplings of the European Beech. Open forests at the beginning and end of the last interglacial period are also testified by the presence of animals preferring open conditions; for example the horse, whose remains were recovered in lake deposits at Lehringen and Rabutz. It is possible that foraging by large animals, such as rhinoceroses and elephants, constantly produced clearings in the forest (May, 1993). The chronological distribution of sites evincing the presence of humans possibly reflects a deterioration of environmental conditions within MIS 5e (Fig. 12.3). More than 80% of the datable deposits originate from the first third of this time period. This cannot be explained by the fact ‘that most of the Eemian sedimentary basins were filled up around the middle part of the interglacial’ (Roebroeks and Speleers,

–2 110

–8 δ18O [‰ PDB] –4

2002), as most of these basins contain not only the complete Eemian sequence but also deposits from the interstadials of the Early Weichselian. The rarity of hominid settlement traces dating to the last two-thirds of MIS 5e on the continent appears to be reflected by a complete absence of settlement in Britain, possibly related to its island position (White and Schreve, 2000; Ashton, 2002). While dating by pollen analysis is possible for sediments in lake basins, the age determination of find layers in some travertine deposits is based on the assumption that the snail Helicigona banatica was only present in Central Europe for a short span of time in MIS 5e, during the Corylus-Quercetum mixtum or the CorylusTilia phases (Mania, 1973, 110). This corresponds with the findings at Ga´novce (SK), where the molluscan fauna C contains H. banatica and correlates with the second Palaeolithic find layer, assigned to a

110

–10

177

7

7

6b

6b

120

120

3

6a

EEM

5

5

130

4b 4a 1-3

l.h.s. b.s. TAU

?

2 1 Gá

H-O

L

Gr

R

N-N 1

4b 4a 1-3

H.b.

130

6a

Fig. 12.3 Chronological position of archaeological sites dating to MIS 5e in relation to the vegetation phases of the last interglacial as well as to the oxygen isotope fluctuations in the lake basin of Gro¨bern (Saxony-Anhalt). Vegetaion phases: 1 Betula zone, 2 Pinus-Betula zone, 3 Pinus-Quercetum mixtum zone, 4a Quercetum mixtum-Corylus zone, 4b Corylus-Taxus-Tilia zone, 5 Carpinus zone, 6a Carpinus-Abies zone, 6b PinusPicea-Abies zone, 7 Pinus zone. H.b.: presumed occurrence of Helicigona banatica. Sites: TAU: Taubach, b.s. ‘‘bone sand’’, l.h.s. ‘lower humic sand’; Ga´: Ga´novce; H-O: Hoˆrka-Ondrej; L: Lehringen; Gr: Gro¨bern; R: Rabutz, NN Neumark-Nord 1 (modified from Boettger et al., 2000 and Wenzel 1998, Fig. 105 by O. Jo¨ris/ S. Wenzel).

178

Stefan Wenzel

phase with mixed-oak forest by pollen analysis and preserved leaf imprints (Vlcˇek, 1969). These data compare well with the results of pollen analyses, which indicate that during the last interglacial period the highest mean annual temperatures were already attained in the Quercetum mixtum– Corylus phase and prevailed up to the Corylus-Taxus-Tilia phase (Caspers et al., 1995). For individual find layers the following statements are possible:

– At Neumark-Nord 1 (D), archaeological and palaeontological finds have been recovered on the upper and lower shore terraces of a shallow lake, which correspond to different phases of regression. If the interglacial sediments belong to MIS 5e, the shore terraces can be attributed to the Quercetum mixtum–Corylus zone or Corylus-Taxus-Tilia zone of this interglacial and pre-dating the Carpinus zone (Litt, 1994).

– At Taubach (D), a small accumulation of artefacts in ‘the lower humic sand’ is associated with the occurrence of H. banatica. The main archaeological horizon, investigated around 1900, was in loose travertine (‘bone sand’) at the base of the travertine deposit (Steiner, 1977). – Ga´novce (SK), level 1: Pine-birchhorizon; Ga´novce, level 2: phase with mixed-oak forest, presence of H. banatica (Vlcˇek, 1969); Ga´novce, level 3: ‘time of the mixed forest with dominance of conifers with spruce, fir und pine’ (Ba´nesz, 1990). The travertine sequence closes with a horizon which is characterized by pine, birch and willows associated with faunal elements of the glacial steppe (Svoboda, 2002). – Layer 12 in site area C1 at Hoˆrka-Ondrej (SK) underlies layer 10, which contains a Helicigona banatica-fauna (Kaminska´ et al., 2000). – The ‘elephant stratum’ at Lehringen (D) belongs to the Tilia-Ulmus-Corylus zone (Thieme and Veil, 1985) or Corylus-TaxusTilia zone (Litt, 1990). – The straight-tusked elephant remains at Gro¨bern (D) date into the final phase of the Corylus-Taxus-Tilia zone (Weber and Litt, 1991). – According to the analysis of macrofossils, the palaeolithic horizon at Rabutz (D) was probably deposited before the Carpinus zone (Toepfer, 1958). K. Erd (1990) considers a correlation with the Quercetum mixtum zone and the Corylus-Taxus-Tilia zone conceivable.

There are several archaeological sites in the northwestern part of Europe which belong to MIS 5d and testify to the hominid occupation of open environments after the Eemian sensu stricto, such as Wallertheim C (¼ Bachablagerung B1 of the old excavations) (Gaudzinski, 1995; Adler et al., 2003); Saint-Sauveur (Antoine et al., 1995) and the sites of Tata and Stuttgart-Untertu¨rkheim (steppe rodent layer) mentioned above. 12.4 THE EARLY NEANDERTHALS The hominids of the last interglacial were Early Neanderthals, and their remains have been discovered at the following sites: – Krapina (Croatia): 874 bone fragments and teeth of Early Neanderthals were recovered at the sandstone abri of Krapina, some 55 km north of Zagreb. These finds can be assigned mainly to the cultural layers 3 and 4 and accumulated in a complex of sediments deriving from dilapidated sandstone which overlays the basal fluvial deposits (GorjanovicKramberger, 1906; Radovcˇic´ et al., 1988). The fluvial deposits and the lower cultural layers 1–4 belong to MIS 5e, according to their altitude above the valley floor and an associated interglacial fauna (Patou-Mathis, 1997). Stratified above layer 4 are thick sand deposits (3þ metres thick) with large frost-shattered blocks, clearly indicating a change in the climate.

Neanderthal Presence and Behaviour During MIS 5e

Cut marks on some of the Neanderthal bones are discussed to have been inflicted in association with secondary burials (Russell, 1987b; Wolpoff, 1999). In addition, some bones bear pathological features (cranium Krapina 4; ulna fragment Krapina 180), which suggest that handicapped and wounded individuals were supported by the group (GorjanovicKramberger, 1906; Trinkaus, 1983; Klein, 1999). Defects on the teeth of the Early Neanderthals from Krapina are indicative of nutritional deficits during childhood (Pettitt, 2000). – Baume Moula-Guercy near Soyons (Arde`che, F): a sequence of cultural horizons has been excavated in the Baume (¼ cave) Moula Guercy since 1992. According to faunal analyses, levels XIV and XV belong to MIS 5e (Defleur et al., 2001; Moncel, 2003). In layer XV, 78 hominid fossils were found which can be assigned to a minimum of six individuals. Numerous bones exhibit cut marks. Many of the bones are broken; only bones of the hands and feet are intact (Defleur et al., 1999). – Saccopastore (Latium, I): In 1929 and 1935, two fossil hominid skulls were found in a gravel pit in a lower terrace of the river Aniene, about 2.5 km before the river drains into the Tiber (Condemi, 1992). According to terrace stratigraphy, and interglacial faunal remains from the underlying layers, the finds date to MIS 5e. The almost intact skull Saccopastore 1 belonged to an adult woman. Only parts of the face of the Saccopastore 2 skull are preserved. It is probably from a man younger than 35 years of age. – Ga´novce (SK): An almost complete travertine endocast (revealing the size and shape of the brain) was found in the travertine dome of Ga´novce in 1926. Fragments of the parietal bone, the left temporal bone and a part of the squama occipitalis were adhering to the cast. Between 1955 and 1966, the casts of a radius and a fibula were also recovered (Vlcˇek, 1991).

179

– Taubach (near Weimar, D): A left lower first molar of a 12- to 14-year-old child (quarry Sonnrein, main find layer) and a left lower first deciduous molar of an individual about nine years old (quarry Mehlhorn, lower travertine) were found in the travertine at Taubach (BehmBlancke, 1959/60; Vlcˇek, 1991). The teeth have the morphology similar to that of other Neanderthal teeth. The milk molar from the Melhorn quarry is of Pleistocene age according to ESR dating (Schu¨ler, 2003). The skulls from Saccopastore display characteristics not present in Classic Neanderthals, such as a smaller endocranial capacity, a shorter back of the skull (only preserved on Saccopastore 1) and the position of the meatus acusticus externus, which is not at the same height as the zygomatic bone (Condemi, 1992; Wolpoff, 1999). Characteristics of Classic Neanderthals, such as the forward protrusion of the facial region (prognathism of the alveoles), are only moderately pronounced. In contrast to the skulls of Classic Neanderthals, maxillary notches (Vandermeersch, 1990, Fig. 28) are present on both of the specimens from Saccopastore. But the notches are not as strongly pronounced as those on the skulls of modern humans (Condemi, 1992). Comparisons of features of the two specimens from Saccopastore as well as comparisons of the Saccopastore finds with the skulls from Krapina show that although some characteristics, for instance the form of the orbita, were quite variable, many other characteristics were common to all Early Neanderthals (Condemi, 1992, 145; Wolpoff, 1999, 654). Neanderthal postcranial bones display a pronounced musculature when compared with those of modern humans (Stringer and Gamble, 1993; Henke and Rothe, 1999). Neanderthals were under more physical stress in their early youth than early modern humans (Gibbons, 1996; Pettitt, 2000). With a body height of 1.55–1.65 m

180

Stefan Wenzel

the European Neanderthal was on the average somewhat shorter than modern humans today. Early Neanderthals possessed some cranial characteristics which are distinct from those of their successors the Classic Neanderthals, but are similar to those of their predecessors as well as recent human populations (Henke and Rothe, 1994). The Early Neanderthals were part of a development to a form of human which seems to have vanished about 38 000 years ago, presumably without leaving us with much of their DNA (Stringer and Gamble, 1993; Henke and Rothe, 1999; Caramelli et al., 2003; Schmitz, 2005).

12.5 TOOLS AND WEAPONS Artefacts made of flint and of other materials which tend to break with conchoidal fracture dominate the archaeological record of the last interglacial. The production of levallois flakes was common. The form of these flakes was predetermined by the form of the core and by the arrangement of previous flake scars on the core surface (Boe¨da, 1994; Klein, 1999). The flakes were mainly produced in series. This technique did not lead to as large flakes as when striking off only one endproduct. However, the cores were used very effectively (Fig. 12.4, 6). With

1

2

m 1

3 4 5 0

5 cm

7

0

6

Fig. 12.4 Artefacts from the last interglacial. 1 levallois flake, 2 scraper, 3 point, 4 biface tool, 5 retoucher, 6 levallois core, 7 lance (reconstruction showing the position of removed branches). 1.2.6 from Rabutz, 3 from Stuttgart-Untertu¨rkheim (lower cultural layer of the lower travertine), 4–5 Ku˚lna Cave; 7 Lehringen. 1.2.6 baltic flint, 3 Muschelkalk chert; 4 porcellanite, 5 bone, 7 yew wood (Toepfer, 1958; Valoch, 1988; Thieme and Veil, 1985).

Neanderthal Presence and Behaviour During MIS 5e

another method of flake production, one changed the striking and flaking surface of the cores depending upon need. Here one obtained less standardized flakes, but one could exploit even quite reduced cores (Moncel and Neruda, 2000). Sharp-edged flakes were suitable for many activities, such as butchering game, without further modification. But the edges of the flakes could be made less steep and more stable by striking them with a retoucher made of bone (Fig. 12.4, 5). By doing this, the outline of a tool was also formed, and scrapers (Fig. 12.4, 2), points (Fig. 12.4, 3) and denticulated pieces were produced using this technique (BehmBlancke, 1959/60; Kaminska´ et al., 2000; Schu¨ler, 2000). In some cases, the base of the scrapers or points were thinned by flaking (Fig. 12.4, 3): these tools were probably easier to mount in shafts. A flake with adhesions of a substratum from oak bark was discovered at NeumarkNord 1 (Mania, 2004), perhaps indicating the use of this material in hafting the flake to a handle. Bifacially worked tools in the form of small handaxes are known from the Ku˚lna cave in the Moravian karst (Fig. 12.4, 4), and a bifacially worked scraper was found in the travertine of Burgtonna near Gotha (Toepfer, 1978). The modified tools from the last interglacial period are tool forms found during the Middle Palaeolithic, but bifacial tools appear much scarcer than in some phases before and after the last interglacial (Jo¨ris, 2003). The only wooden find dating to the last interglacial is a lance fashioned from yew from Lehringen (Fig. 12.4, 7), which was found between the ribs of a straight-tusked elephant (Thieme and Veil, 1985). The production of this weapon must have been time consuming, since the surface of the lance had been completely modified and several branches removed. The point of the lance is slightly asymmetrical so that the point itself is made of the harder, outer wood rather than the softer, inner pith. To what extent

181

the point of the lance was hardened by fire is unclear. The base of the lance is rounded; it may have served as a multifunction tool, perhaps a digging stick.

12.6 SUBSISTENCE 12.6.1 Gathering plant food Unambiguous evidence for the collection and preparation of vegetable foods during the last interglacial is not available. However, we may speculate that hominids living in MIS 5e might have found compensation for the unfavourable economic conditions caused by the dense forest in part by eating larger portions of plant food. Meat is rather a luxury commodity for most modern hunter-gatherers, except for some groups of Eskimo (Lee, 1977). Plants and small land animals are often available, require less effort and are better predictible in their procurement than hunting larger animals. Food gathering is the basis of modern hunter-gatherer subsistence; this soure of food is general far from being exhausted (Lee, 1977). Only from the early Upper Palaeolithic onwards local overharvesting of certain species seems to have occurred (Stiner and Munro, 2002). The best evidence to date of Neanderthals exploiting plant food is seen at Rabutz, where charred shells of hazelnut (Corylus avellana) were possibly burnt in an attempt to make the nuts tastier and to prevent decay by roasting (Toepfer, 1958). Considering that interglacial vegetation offered an abundance of edible and medicinal plants (Mai, 1983), there should be more evidence for their use; however their absence might simply be a factor of preservation in the archaeological record. 12.6.2 Use of aquatic resources There is much evidence that the hominids of the Eemian phase utilized beaches, where

182

Stefan Wenzel

they may have collected seafood. Artefactbearing beach deposits dating to MIS 5e are known from Saint Germain the Vaux/Port Racine (Somme, F) (Cliquet, 1992); Balzi Rossi (Liguria, I) (Vicino, 1976); Elea (Pelepones, GR) (Kowalczyk et al., 1992); Ras Aamer, Wadi Haula and El Atrun (Libya) (McBurney and Hey, 1955; Wenzel, 1998); as well as the Red Sea coast of Eritrea (Bruggemann et al., 2004). Human footprints dating to MIS 5e have even been revealed on the beach of a lagoon near the southern tip of Africa (Roberts and Berger, 1997). It is unclear to what extent freshwater fish were exploited by Neanderthals. The few fish remains from Taubach could also have been caught by carnivores. However, there is evidence for fishing activities from layer VIII at Grotte Vaufrey (Dordogne, France), dating to the penultimate interglacial (Le Gall, 1989). 12.6.3 Hunting The most impressive example of Neanderthal hunting during the last interglacial comes from the site of Lehringen near Verden an der Aller, some 35 km southeast of Bremen (Adam, 1951; Thieme and Veil, 1985). In 1948, an elephant butchering site was discovered there in lake deposits. A lance of yew wood, approximately 2.38 m in length, was found between the ribs of a 45year-old straight-tusked elephant, providing unequivocal evidence of the hunting of the animal. Twenty-seven flint flakes bearing traces of meat-cutting activities were distributed around the head of the elephant. It is less clear whether the straight-tusked elephant found at the Gro¨bern site (NE Bitterfeld; Saxony-Anhalt) had been hunted. Found in the riparian zone of a former lake, the elephant was old and had suffered from inflammation of the bone (Weber and Litt, 1991). However, Neanderthals seem to have loosened some of its bones and moved the left tusk (Fig. 12.5). Some artefacts and fragments of bones of other animals were found together with the remains of the elephant.

In the travertine deposit of Taubach near Weimar, the majority of the remains of larger mammals belong to the extinct rhinoceros Stephanorhinus kirchbergensis and to brown bears (Ursus arctos). Wolfgang Soergel (1922) had already noted a high proportion of remains of young rhinoceroses at the site and, using age statistics, tried to prove that the rhinoceroses had been killed by humans. New investigations determined numerous cut marks on the bones of both rhinoceroses and bears at Taubach (Bratlund, 2000). At least 76 rhinoceroses, comprising 51 young and subadult animals, as well as 52 bears were present. The mortality profile of the rhinoceroces deviates from the classical profile of a naturally dying community, since too few old animals are represented. Instead, the profile resembles that of populations which died during a mass mortality, which in this context suggests hunting by humans. The mortality profile of the bears displays a dominance of adult, but not very old animals, also suggestive of humanhunting activities (Bratlund, 2000). Cut marks on the bones of bison, beaver and red deer indicate that these animals were also exploited by hominids. Merck’s rhinoceros was possibly also hunted at Krapina. Patou-Mathis (1997) counted 42 individuals of this taxon from all the find layers. She gives minimum numbers of individuals for each layer, but mingles the animals from all layers in tables and discussion when dealing with the age structure. Only the lowermost layers up to cultural layer 4 belong to MIS 5e. Nineteen Merck’s rhinoceros were young animals, less than six years of age. Remains of 14 bison and 10 aurochs have been identified, as well as 14 beavers, while many other animals are represented by only a few finds. Bones of cave bears were predominantly found in the postinterglacial layers, and probably represent animals which died during hibernation. With a minimum of five individuals, fallow deer dominates the fauna from horizon A at Wallertheim in the Rheinhessen region.

Neanderthal Presence and Behaviour During MIS 5e

183

1 2 3 1m

Fig. 12.5 Palaeoloxodon antiquus skeleton found in situ at the site of Gro¨bern together with lithic artefacts (Weber and Litt 1991, Fig. 5). Disjuncted anatomical connections of the left (1) and of the right (2) side of the body as well of the skull (3) (Weber and Litt, 1991).

Remains of at least four large bovids (aurochs or bison), and one each horse, wolf and beaver were also recovered (Conard and Prindiville, 2000). Owing to the strong decomposition of the bones found in floodplain sediments, cut marks are preserved on only one bone fragment. But the bones of fallow deer and bovids seem to have been broken intentionally by hominids. It is not clear whether horses were hunted at Wallertheim A. Traces of gnawing on horse bones indicate that carnivores played a role in the accumulation of the equid remains. A total of 382 bones and more than 6700 stone artefacts were recovered from Wallertheim A during investigations of a site 176 m2 in size. The finds were found more or less in

situ, enabling the structure of the camp to be examined. A possible hearth may be indicated by a small concentration of burnt bones. At Neumark-Nord 1, the nearly complete skeleton of a juvenile male aurochs was found on the lower shore terrace (Mania, 2004; Mania et al., 2004). A few stone artefacts lay in the proximity of the carcass, but traces of butchering are absent. Some of the aurochs bones display carnivore gnawing. Partial skeletons of Merck’s rhinoceros and of straight-tusked elephant were also discovered in association with few artefacts. Numerous complete skeletons of fallow deer were found within the former lake at Neumark-Nord 1, representing animals that

184

Stefan Wenzel

seem to have died naturally at the site (Mania, 1990; Gaudzinski, 2004). At the southern margin of the lake basin, a considerable number of bones and of artefacts have been found in distant position to the former shore, originating from campsites on the elevated plain outside the basin (Mania, 1990). The remains of a ‘background-fauna’ which perished naturally, or which were killed by carnivores, are far more numerous at some sites than remains of animals showing traces of hominid activities. This applies particularly to the lake deposits of the last interglacial, in which complete skeletons of deer and wild cattle have been repeatedly discovered. Some of these remains do not even exhibit carnivore gnawing. So far, cutmarked bones have been determined at only at a few sites, for example at Lehringen (bear, beaver) and Rabutz (Merck’s rhinoceros) (Wenzel, 1998). With such a large number of well-investigated lake basins dating to the Eemian (Eissmann and Litt, 1994; Bjo¨rck et al., 2000), offering some of the best conditions of preservation known to archaeology, the low overall number of archaeological sites that have been found so far is surprising. This might reflect the low population density during the last interglacial. In Taubach, the rudiments of a particular kind of Neanderthal behaviour can be recognized which is more pronounced on some sites from the time after MIS 5e – specialization on the hunt on certain taxa of big game (Gaudzinski, 1995; 2004). As in the case of bovids at some post-Eemian sites, it is unlikely that all the rhinoceroses from Taubach were captured in a single hunting event. The hunting sites were repeatedly visited, and certain species were singled out for capture. This contradicts the opportunistic food procurement strategy occasionally attributed to the Neanderthals. The absence of a concept for the hunting of large animals would have to be recognized by evidence of hunting mainly small mammals and scavenging

from animal carcasses. Even today some hunter-gatherer groups occasionally utilize the carcasses of dead animals as an additional source of protein (Foley, 1992). However, large quantities of animal carcasses are found only during droughts in Africa and only during the winter in Europe (Stiner, 1994). Thus, they can only have played an occasional role as a source of food. Evidence of the hunting of small game is missing.

12.6.4 Creation of supplies? The ability of the Neanderthals to plan on a long-term basis and store supplies has often been questioned (Stringer and Gamble, 1993). However, one example of the possible storage of goods is found in the travertine site of Stuttgart-Untertu¨rkheim, where the cavities of decayed tree trunks seem to have functioned as depots for materials (Fig. 12.6) (Adam and Berckhemer, 1983; Wenzel, 1998). These cavities were covered by travertine and contained animal bones and considerable numbers of large pebbles, some of which are chipped. The absence of river gravels or smaller rubble in the cavities, and the fact that the travertine is almost free of pebbles, indicates that these stones were deliberately placed here. However, the location near a travertine spring seems not to be suitable for storing food, which may be preferently done on dry ground. More recently, comparable features dating to the time of the Classic Neanderthals were observed at La Quina (Dordogne, F). Forty-two large, plateshaped pebbles were found in a pit approximately 70–80 cm in diameter and 40 cm in depth (Debe´nath, 1992). This structure was interpreted by the excavator as a food cache. Whether this interpretation is correct or not, the evidence from Untertu¨rkheim and La Quina suggests that Neanderthals practised storing food or raw materials on a small scale.

Neanderthal Presence and Behaviour During MIS 5e

0

185

10 cm

30 cm

Cavity

110 cm

S. 3. 30 cm

Bone

Sinter

Yellowish brown loam

70 cm Brown fluvial loam (Letten/Aualehm)

1

2

Fig. 12.6 Stuttgart-Untertu¨rkheim. 1 Cavity of a decayed tree trunk (S3) containing pebbles presumably deposited by Early Neanderthals (sketch by F. Berckhemer, modified); 2 worked pebble from tree trunk cavity S3.

12.6.5 Use of the landscape The occurrence of exogenous raw materials at last interglacial sites provides information about the extent of the area Early Neanderthals might have in their movements over the course of the year. Local raw materials from sources 10–15 km away from the respective sites could have been collected and brought to the place of residence as part of the daily activities, as has often been observed in modern hunter-gatherer groups (Floss, 1994). There is however evidence that those larger distances were covered. For example, level 11 at Ku˚lna cave in the Moravian karst contained several tools made of a soft, coloured porcellanite from a source 60 km away (Valoch, 1988). In addition, at Burgtonna in Thuringia, a scraper of quartzite from an outcrop on the far side of the Thuringian forest was found in the travertine (Toepfer, 1978). Although there is stronger evidence at some sites of longdistance raw material transfer in postEemian contexts (Adler et al., 2003, 69), the occurrence of ‘exotic´ materials in the Middle Palaeolithic is rather rare and mostly limited to modified tools and resharpening waste. The mobility of Early Neanderthals as reflected by long distance transfer of artefacts does not differ substantially from that observed in the Middle Palaeolithic under different climatic conditions.

12.7 RELIGIOUS CONCEPTIONS The bones of the Early Neanderthals from Krapina display traces of manipulations indicative of mortuary behaviour by these humans. Traces of cut marks, fractures and patches of burning on the bones had already been observed during excavation and interpreted as evidence of cannibalism (Gorjanovic-Kramberger, 1906). Later analyses showed that the bones had been broken by rock fall or movement of sediment, and that there were no clear traces of a deliberate smashing of the bones to obtain marrow (Russell, 1987a). In addition, the few burned bones seem to have been affected accidentally. The anatomical location, sharpness and frequency of cut marks on the Neanderthal remains from Krapina (Fig. 12.7) differ from those usually produced during the butchering of animals where defleshing rather than disarticulation is of interest and one tries to avoid scratching the bones so as not to dull the implements (Russell, 1987b). The sharpness of many Krapina cut marks indicates that they were made directly on the bone and not through flesh. The pattern of the Krapina striations matches skeletal collections of North American Indian tribes, who practised disarticulation after defleshing (and after partial decomposition) of corpses prior to transfer of the bones for burial (Russell, 1987b). Russell’s arguments

186

Stefan Wenzel

Fig. 12.7 Krapina. Cut-marked humerus fragments of Early Neanderthals (Ullrich, 1989, Fig. 3).

for secondary burials in Krapina seem to be plausible, but the possibility of cannibalism cannot be totally excluded (Wolpoff 1999, 658). The preparation of human bones is also known from historical contexts in Europe and the bones of prominent crusaders, such as the emperor Barbarossa, have been defleshed in order to bring them back to Europe (Peter-Ro¨cher, 1994, 30). Characteristic of the Krapina assemblage is a dominance of the upper molars of adult Neanderthals; adult lower molars are less well represented (Turk and Dirjec, 1991). This discrepancy could derive from bringing skulls without lower jaws into the abri, yet still representative of the complete corpse. It has been discussed whether cannibalism was practised during the last interglacial at the Baume Moula-Guercy near Soyons (Arde`che, F) (Defleur et al., 1999; Pettitt, 2002). The bones of six individuals were recovered in layer XV, close to two hearths. The bones are fractured and bear traces of cut marks. The location of the cut marks and the pattern of the fractures are very similar to traces produced on the bones of deer during dismemberment and marrow procurement. In the period postdating the last interglacial, there is more evidence of Neanderthals dismembering their dead and removing bones from graves. Cut marks are present on the cranium of the eponymous Neanderthal from the small Feldhofer Grotte in the Neandertal Valley (Schmitz and Thissen, 2000). Cut marks have also been observed on the bones of Classic

Neanderthals from Marillac (near Angouleˆme, Charente) and Combe Grenal (Dordogne). Evidence of the removal of bones from graves has been observed at Kebara Cave (Mount Carmel, Israel), where the skull had been lifted away from the decayed cadaver of a Neanderthal (Bar-Yosef and Vandermeersch, 1993). This action must have been carried with great care, since only an upper molar and the hyoid bone remained behind with the lower jaw, and none of the other skeletal elements had been disturbed. A comparable situation was found in Regourdou Cave (near Lascaux, Dordogne) (Bosinski, 1993). The possible secondary burials of Krapina in the context of the presumed ‘handling and circulation of skeletal parts’ (Larsson, 2001) and the numerous Neanderthal burials of post-Eemian times (Defleur, 1993; Riel-Salvatore and Clark, 2001; Pettitt, 2002) point to the ability of Neanderthals to reflect on life and death (Wolpoff, 1999). Further evidence of activities which go beyond those necessary to ensure survival, such as the application of colouring materials and the collection of curiosities (e.g. fossils and minerals) (Leroi-Gourhan, 1984; Scha¨fer, 1996), is known from the period both before and after MIS 5e. It is possible that numerous shed antlers of red deer from the travertine of Taubach (Kahlke, 1977) and from the underlying stratum of the travertine of Caours (Somme, F) (Breuil and Barral, 1955; Somme´, 1989), as well as many other accumulations of shed antlers dating to the Middle Palaeolithic, can be placed into a similar, nonutilitarian context, while

Neanderthal Presence and Behaviour During MIS 5e

an accumulation caused by carnivores (or the red deer when shedding the antlers) cannot be totally excluded (Conard, 1992; Street, 2002; Wenzel, 1998). The Sami of Finland, as well as different peoples of Siberia, and some tribes of North American Indians (zu Wied, 2001, 232) gathered shed antlers for religious reasons well into historical times. But in general, evidence indicating that Neanderthals were capable of symbolic thinking is rare and vague. Only modern humans, who settled Europe about 38 000 BP, possessed a prominent and large cultural repertoire, which can be regularly and clearly identified in archaeological contexts. 12.8 CONCLUSION Although some scholars have argued that the interglacial forested environments were less favourable for human habitation than the open landscapes of previous and subsequent periods, many aspects of the archaeological record of the Eemian in fact point to a continuity of Neanderthal ways of life independent of the environment. In regard to their physical characteristics, tools, subsistence and mobility, the Neanderthals living in the forests of the last interglacial do not seem to have differed dramatically from their direct predecessors and successors which were inhabiting an open landscape. The picture changes for the last twothirds of the last interglacial, as evinced by the scarcity of archaeological sites dating to this time period, perhaps indicating a major deterioration in environmental conditions. Were Early Neanderthals really absent from northern Europe in the later part of the last interglacial and was their absence the only response of hominids to the prevailing ecological conditions? In order to come to a better understanding of the history of hominid settlement within the Eemian, it would be necessary to determine the chronological position of more archaeological sites within this interglacial.

187

It is clear from the archaeological evidence that the Early Neanderthals of the Eemian were able to hunt big game, but we still know little about the role of plant food in their diet. One may ask if plant food was more important during shifts and alterations in vegetation zones. This is possible, since studies of modern huntergatherers show that gathering of plant food increases in importance accordingly with habitation in lower latitudes (Lee, 1977, 42 and tab. 9–10). Additional archaeobotanic evidence from MIS 5e can begin to answer these questions. It might also complement the seasonal information from the analysis of the fauna and thus lead to a better insight into Neanderthal mobility patterns (BarYosef, 2004, 336). Other productive avenues of research on this topic involve observations on microwear patterns on hominid teeth from MIS 21 onwards, up to now missing for Early Neanderthals from the last interglacial of northern Europe; these seem to indicate that there is as well a correlation between the portion of abrasive plant food and climate as well as a long-term trend of dietary habits towards a bigger component of meat (Pe´rez-Pe´rez et al., 2003; Jo¨ris 2005). The spatial analysis of last interglacial campsites may not only help to evaluate taphonomic processes and to isolate individual occupations, but can provide a valuable source of information about the social organization and behavioural patterns of Early Neanderthals. Thus far, Wallertheim A which represents ‘a single short-term occupation within a landscape regularly and repetedly traversed by hominids’ (Adler et al., 2003, 53) is the only site from MIS 5e in northern Europe where spatial analyses have been conducted in detail. It may be possible to find a relationship between site types as defined by activities and data on seasonality in order to describe the settlement systems within ecological phases of MIS 5e. In summary, there is much evidence that supports our impression of a continuity of the behaviour of hominids in the Eemian,

188

Stefan Wenzel

yet our knowledge remains incomplete. Future research will undoubtedly contribute to a fuller understanding of human reactions on ecological changes. ACKNOWLEDGEMENTS The author thanks L. Niven and E. Turner for helpful critic, various informations and for improving the English text, O. Jo¨ris for many discussions, as well as F. Sirocko, J. van der Meer and two anonymous reviewers for their suggestions. REFERENCES Adam, K.D., 1951. Der Waldelefant von Lehringen, eine Jagdbeute des diluvialen Menschen. Quarta¨r 5, 79–92. Adam, K.D., Berckhemer, F., 1983. Der Urmensch und seine Umwelt im Eiszeitalter auf Untertu¨rkheimer Markung. Ein Beitrag zur Urgeschichte des Neckartales. Bu¨rgerverein Untertu¨rkheim, Stuttgart, 1–88. Adler, D., Prindiville, T.J., Conard, N.J., 2003. Patterns of spatial organization and land use during the Eemian Interglacial in the Rhineland: new data from Wallertheim, Germany. Eurasian Prehistory 1 (2), 25–78. Allsworth-Jones, P., 1986. The Szeletian and the transition from Middle to Upper Palaeolithic in Central Europe. Clarendon Press, Oxford, XVIII, 412 pp., XII. Antoine, P., Munout, A.-V., Kolfschoten, T. van, Limonin, N., 1995. Une occupation du pale´olithique moyen en contexte fluviatile dans la se´quence de la tre`s basse terasse de la Somme a` Saint-Sauveur (Somme). Bulletin de la Socie´te´ Pre´historique Franc¸aise 92, 201–212. Ashton, N., 2002. Absence of humans in Britain during the last interglacial (oxygen isotope stage 5e). In: Tuffreau, A., Roebroeks, W. (Dir.), Le Dernier Interglaciaire et les occupations humaines du Pale´olithique moyen. Colloque Crouy-Saint-Andre´, septembre 1997. Publications du CERP 8. 93–103. Ba´nesz, L., 1990. Mittelpala¨olithische kleinfo¨rmige Industrie aus den Travertinfundstellen der Zips. Slovenska´ Archaeolo´gia 38, 45–88. Bar-Yosef, O., 2004. Eat what is there: hunting and gathering in the world of Neanderthals and their neighbours. International Journal of Osteoarchaeology 14, 333–342.

Bar-Yosef, O., Vandermeersch, B., 1993. Modern humans in the Levant. Scientific American, April 1993, 64–70. Behm-Blancke, G., 1959/60. Altsteinzeitliche Rastpla¨tze im Travertingebiet von Taubach, Weimar, Ehringsdorf. Alt-Thu¨ringen 4. VI, 246 pp. Bjo¨rck, S., Noe-Nygaard, N., Wolin, J., HoumarkNielsen, M., Hansen, H.J., Snowball, I., 2000. Eemian lake development, hydrology and climate: a multi-stratigraphic study of the Hollerup site in Denmark. Quaternary Science Reviews 19, 509–536. Boe¨da, E., 1994. Le concept levallois: variabilite´ des me´thodes. Monographie du Centre de Recherches Arche´ologiques 9, 280 pp. Boettger, T., Junge, F.W., Litt, Th., 2000. Stable climatic conditions in central Germany during the last interglacial. Journal of Quaternary Science 15 (5), 469–473. Bosinski, G., 1993. Der Neandertaler und seine Zeit. Archa¨ologie im Ruhrgebiet 1 (1991), 25–48. Bouteaux, A., 2003. E´tude arche´ozoologique du site de Payre en Ardeche. In: Patou-Mathis, M., Bocherens, H., (Eds.), Le roˆle de l’environnement dans les comportements des chasseurs-cueilleurs pre´historique. BAR International Series 1005, 99–110. Bratlund, B., 2000. Taubach revisited. Jahrbuch des Ro¨misch-Germanischen Zentralmuseums Mainz 46, 61–174. Braun, M., Frank, N., Goppelsro¨der, A., Kind, C.-J., Mangini, A., Mu¨ller, K.P., Niederho¨fer, H.-J., Wagner, G.R.A., Ziegler, R., 1999. Deckerstraße – Eine mittelpala¨olithische Travertin-Fundstelle in Stuttgart – Bad Cannstatt. Fundberichte aus Baden-Wu¨rttemberg 22, 13–44. Breuil, H., Barral, L., 1955. BMIS de Cervide´s et autres os travaille´s sommairement au Pale´olithique ancien du Vieux Monde et au Mouste´rien des Grottes de Grimaldi et de l’Observatoire de Monaco. Bulletin du Muse´e d’Anthropologie Pre´historique de Monaco 2, 3–11. Bruggemann, J.H., Buffler, R.T., Guillaume, M.M.M., Walter, R.C., Cosel, R. von, Ghebretensae, B.N., Berhe, S.M., 2004. Stratigraphy, palaeoenvironments and model for the deposition of the Abdur Reef Limestone: context for an important archaeological site from the last interglacial on the Red Sea coast of Eritrea. Palaeogeography, Palaeoclimatology, palaeoecology 203, 179–206. Caramelli, D., Lalueza-Fox, C., Vernesi, C., Lari, M., Casoli, A., Mallegni, F., Chiarelli, B., Dupanloup, I., Bertranpetit, J., Barbujani, G., Bertorelle, G., 2003. Evidence for a genetic discontinuity between Neanderthals and 24,000-year-old anatomical modern Europeans. Proceedings of the National Academy of Sciences, U. S. A. 100, 6593–6597.

Neanderthal Presence and Behaviour During MIS 5e Caspers, G., Jordan, H., Merkt, J., Meyer, K.-D., Mu¨ller, H., Streif, H., 1995. Niedersachsen. In: Benda, L. (Ed.), Das Quarta¨r Deutschlands. Gebru¨der Borntraeger, Berlin, Stuttgart, pp. 23–58. Cliquet, D., 1992. Le gisement Pale´olithique moyen de Saint-Germain-des-Vaux/Port Racine (Manche) dans son cadre re´gional. E´tudes et Recherches Arche´ologiques de l’Universite´ de Lie`ge 63, 644 pp. Conard, N.J., 1992. To¨nchesberg and its position in the paleolithic prehistory of Northern Europe. Monographien des Ro¨misch-Germanisches Zentralmuseums Mainz 20. XI, 176 pp. Conard, N.J., Prindiville, T.J., 2000. Middle Palaeolithic hunting economies in the Rhineland. International Journal of Osteoarchaeology 10, 286–309. Condemi, S., 1992. Les hommes fossiles de Saccopastore et leurs relations phyloge´ne´tiques. Cahiers de Pale´oanthropologie. XIV, 174 pp. Debe´nath, A., 1992. De pierre et d’os In: Debe´nath, A., Tournepiche, J.F. (Eds.), Ne´andertal en PoitouCharentes. Association Re´gionale des Conservateurs des Musee´es de Poitou-Charentes, Angouleˆme, 100–119. Defleur, A., 1993. Les se´pultures mouste´riennes. CNRS E´ditions, Paris, 325 pp. Defleur, A., White, T., Valensi, P., Slimak, L., Cre´gutBonnoure, E`., 1999. Neanderthal Cannibalism at Moula-Guercy, Arde`che, France. Science 286, 128–131. Defleur, A., Cre´gut-Bonnoure, E`., Desclaux, E., Thinon, M., 2001. Pre´sentation pale´o-environnementale du remplissage de la Baume Moula-Guercy a` Soyons (Arde`che): implications pale´oclimatiques et chronologiques. L’Anthropologie 105, 369–408. Eißmann, L., 1990. Das mitteleuropa¨ische Umfeld der Eemvorkommen des Saale-Elbe-Gebietes und Schlußfolgerungen zur Stratigraphie des ju¨ngeren Quarta¨rs. In: Eißmann, L. (Ed.), Die Eemwarmzeit und die fru¨he Weichseleiszeit im Saale-ElbeGebiet: Geologie, Pala¨ontologie, Palo¨kologie. Ein Beitrag zum ju¨ngeren Quarta¨r in Mitteleuropa. Altenburger naturwissenschaftliche Forschungen 5, 11–48. Eissmann, L., 2002. Quaternary geology of eastern Germany (Saxony, Saxony-Anhalt, South Brandenburg, Thuringia), type area of the Elsterian and Saalian Stages in Europe. Quaternary Science Reviews 21, 1275–1346. Eißmann, L., Litt, Th. (Eds.), 1994. Das Quarta¨r Mitteldeutschlands. Ein Leitfaden und Exkursions¨ bersicht u¨ber das Pra¨quarta¨r fu¨hrer. Mit einer U des Saale-Elbe-Gebietes. Altenburger Naturwissenschaftliche Forschungen 7, 458 pp. Erd, K., 1990. Pollenstratigraphie des interglazialen Beckentons von Rabutz su¨do¨stlich Halle/

189

Saale. In: Eißmann, L. (Ed.), Die Eemwarmzeit und die fru¨he Weichseleiszeit im SaaleElbe-Gebiet: Geologie, Pala¨ontologie, Palo¨kologie. Ein Beitrag zum ju¨ngeren Quarta¨r in Miteleuropa. Altenburger naturwissenschaftliche Forschungen 5, 141–147. Floss, H., 1994. Rohmaterialversorgung im Pala¨olithikum des Mittelrheingebietes. Monographien des Ro¨misch-Germanischen Zentralmuseums Mainz 21. XIII, 407 pp. Foley, R., 1992. Studying human evolution by analogy. In: Jones, S., Martin, R., Pilbeam, D. (Eds.), The Cambridge Encyclopedia of Human Evolution. Cambridge University Press, Cambridge, 335–430. Frank, N., Braun, M., Hambach, U., Mangini, A., Wagner, G., 2000. Warm period growth of travertine during the Last Interglacial in Southern Germany. Quaternary Research 54, 38–48. Gaudzinski, S., 1995. Wallertheim Revisited: a Re-analysis of the fauna from the Middle Palaeolithic site of Wallertheim (Rheinhessen/Germany). Journal of Archaeological Science 22, 51–66. Gaudzinski, S., 2004. A matter of high resolution? The Eemian Interglacial (MIS 5e) in Northcentral Europe and Middle Palaeolithic subsistence. International Journal of Osteoarchaeology 14, 201–211. Geyh, M., Reiff, W., Frank, N., 1999. Grenzen der radiometrischen 230Th/U-Altersbestimmung der Sauerwasserkalkvorkommen (Travertine) in Stuttgart. Zeitschrift der deutschen geologischen Gesellschaft 150, 703–733. Gibbons, A., 1996. Did Neandertals lose an evolutionary ‘arms’ race? Science 272, 1586–1587. Gorjanovic-Kramberger, D., 1906. Der diluviale Mensch von Krapina in Kroatien. Ein Beitrag zur Pala¨oanthropologie. Studien u¨ber die Entwicklungsmechanik des Primatenskelettes mit besonderer Beru¨cksichtigung der Anthropologie und Descendenzlehre. C.W. Kreidel, Wiesbaden. I-XI, pp. 59–277 Heinrich, W.D., 1990. Nachweis von Lagurus lagurus (Pallas, 1773) fu¨r das Pleistoza¨n von NeumarkNord, Kreis Merseburg. In: Mania, D., Thomae, M., Litt, T., Weber, T., Neumark-Gro¨bern. Beitra¨ge zur Jagd des mittelpala¨olithischen Menschen. Vero¨ffentlichungen des Landesmuseums fu¨r Vorgeschichte in Halle 43, pp. 167–175. Henke, W., Rothe, H., 1994. Pala¨oanthropologie. Springer-Verlag, Berlin. XIII, 699 pp. Henke, W., Rothe, H., 1999. Stammesgeschichte des Menschen. Eine Einfu¨hrung. Springer-Verlag, Berlin. X, 347 pp. Houben, C., 2003. Die Wirbeltierfauna aus dem Interglazial von Lehringen (Niedersachsen, Deutschland). Eiszeitalter und Gegenwart 52, 25–39.

190

Stefan Wenzel

Jo¨ris, O., 2003. Zur chronostratigraphischen Stellung der spa¨tmittelpala¨olithischen Keilmessergruppen. Der Versuch einer kulturgeographischen Abgrenzung einer mittelpala¨olithischen Formengruppe. Bericht der Ro¨misch-Germanischen Kommission 84, 49–153. Jo¨ris, O., 2005. Aus einer anderen Welt – Europa zur Zeit des Neandertalers. In: Conard, N.J., Ko¨lbl, St., Schu¨rle, W. (Eds.), Vom Neandertaler zum modernen Menschen. Alb und Donau, Kunst und Kultur 46, pp. 47–70, 200–202. Kahlke, H.-D., 1977. Die Cervidenreste aus den Travertinen von Taubach. Quarta¨rpala¨ontologie 2, 209–223. Kaminska´, L’., Ford, D.C., Hajnalova´, E., Hora´cˇek, I., Kovanda, J., Lozˇek, V., Mlı´kovsky´, J., Smolı´kova´, L., 2000. Hoˆrka-Ondrej. Research of a Middle Palaeolithic travertine locality. Archaeologica Slovacia Monographiae Fontes 17, 202 pp. Klein, R.G., 1999. The human career. Human biological and cultural origins. Second edition. The University of Chicago Press, Chicago. XVII, 810 pp. Koban, Chr. G., 1993. Faziesanalyse und Genese der quarta¨ren Sauerwasserkalke von Stuttgart, BadenWu¨rttemberg. Profil 5, 47–118. von, W. Koenigswald 2002. Lebendige Eiszeit. Klima und Tierwelt im Wandel. Theiss, Stuttgart, 190 pp. Kowalczyk, G., Winter, K.-P., Steinich, G., Reisch, L., 1992. Jungpleistoza¨ne Strandterrassen in Su¨dostLakonien (Pelepones, Griechenland). Schriftenreihe fu¨r Geowissenschaften 1, 72 pp. Larsson, L., 2001. Comment. In: Riel-Salvatore, J., Clark, G.A., Grave Markers. Middle and Early Upper Palaeolithic burials and the use of Chronotypology in contemporary Paleolithic research. Current Anthropology 42, 449–479. Lee, R.B., 1977. What hunters do for a living, or, how to make out on scarce resources. In: Lee, R.B., Devore, I. (Eds.), Man the hunter. Aldine Publishing Company, Chicago, 30–48. Le Gall, O., 1989. Analyse palethnologique de l’ichtofaune de la Grotte Vaufrey. In: Rigaud, J.-Ph. (Ed.), La Grotte Vaufrey, Pale´oenvironments, chronologie et activites humains. Me´moirs de la Socie´te´ Pre´historique Franc¸aise 19, pp. 565–568. Leroi-Gourhan, A., 1984. Hand und Wort. Die Evolu¨ bersetzt tion von Technik, Sprache und Kunst. U von M. Bischoff. 3. Aufl. Suhrkamp, Frankfurt am Main, 531 pp. Lippstreu, L., Strackebrandt, W., 2003. Ja¨nschwalde und die Gliederung des Saale-Komplexes – ein Komentar zum Beitrag von Werner Nowel. Eiszeitalter und Gegenwart 52, 47–83. Litt, Th., 1990. Pollenanalytische Untersuchungen zur Vegetations- und Klimaentwicklung wa¨hrend des Jungpleistoza¨ns in den Becken von Gro¨bern und Grabschu¨tz In: Eißmann, L. (Ed.), Die

Eemwarmzeit und die fru¨he Weichselkaltzeit im Saale-Elbe-Gebiet: Geologie, Pala¨ontologie, Palo¨kologie. Altenburger Naturwissenschaftliche Forschungen 5, 92–105. Litt, Th., 1994. Zur stratigraphischen Einstufung des Interglazials von Neumark-Nord aufgrund neuer pollenanalytischer Befunde. In: Eißmann, L., Litt, Th. (Eds.), Das Quarta¨r Mitteldeutschlands. Ein ¨ berLeitfaden und Exkursionsfu¨hrer. Mit einer U sicht u¨ber das Pra¨quarta¨r des Saale-Elbe-Gebietes. Altenburger Naturwissenschaftliche Forschungen 7, pp. 328–333. Malez, M., 1978. Stratigrafiski, paleofaunski i paleolitski odnosi krapinskog nalazisˇta. (Stratigraphische, pala¨ofaunistische und pala¨olithische Verha¨ltnisse des Fundortes Krapina’’. In: Malez, M. (Hrsg.), Krapinski pracovjek i evolucija hominida. Zbornik predavanja odrzanih na znanstvenom skupu ‘‘Krapinski pracovjek i evolucija hominida’’ u Krapini, dne 17. rujna 1976. Jugoslavenska Akademija Znaosti i Umjetnost, Zagreb, pp. 61–102, supplement 1. Mania, D., 1973. Pala¨oo¨klogie, Faunenentwicklung und Stratigraphie des Eiszeitalters im mittleren Elbe-Saale-Gebiet auf Grund von Molluskengesellschaften. Geologie 21, Beiheft 78/79, Berlin, 175 pp. Mania, D., 1978. Die Molluskenfauna aus den Travertinen von Burgtonna in Thu¨ringen. Quarta¨rpala¨ontologie 3, 69–85. ¨ kologie und mittelMania, D., 1990. Stratigraphie, O pala¨olithische Jagdbefunde des Interglazials von Neumark-Nord (Geiseltal). In: Mania, D., Thomae, M., Litt, T., Weber, T., Neumark-Gro¨bern. Beitra¨ge zur Jagd des mittelpala¨olithischen Menschen. Vero¨ffentlichungen des Landesmuseums fu¨r Vorgeschichte in Halle 43, pp. 9–130. Mania, D., 2001. Die Deckschichtenfolge von Lengefeld-Bad Ko¨sen im mittleren Saaletal – ein Typusprofil fu¨r die Quarta¨rstratigraphie. Praehistorica Thuringica 6/7, 103–131. Mania, D., 2004. In den Jagdgru¨nden des Menschen vor 200 000 Jahren im Geiseltal. In: Meller, H. (Ed.), Pala¨olithikum und Mesolithikum. Kataloge zur Dauerausstellung im Landesmuseum fu¨r Vorgeschichte Halle 1. Landesamt fu¨r Denkmalpflege und Archa¨ologie Sachsen- Anhalt – Landesmuseum fu¨r Vorgeschichte, Halle (Saale), 122–149. Mania, D., Toepfer, V., 1973. Ko¨nigsaue. Gliederung, ¨ kologie und mittelpala¨olithische Funde der letzO ten Eiszeit. Vero¨ffentlichungen des Landesmuseums fu¨r Vorgeschichte in Halle 26, 164 pp. Mania, D., Bru¨hl, E., Laurat, Th., 2004. Zum Stand der Grabungen im Tagebau Neumark-Nord. Stra¨ kologie und Archa¨ologie des Mittelpatigraphie, O la¨olithikums im spa¨ten Mittel- und fru¨hen Jungpleistoza¨n. Fu¨hrer zum Feldkolloquium

Neanderthal Presence and Behaviour During MIS 5e Frankleben und Gosek, 2. und 3. August 2004. Landesamt fu¨r Denkmalpflege und Archa¨ologie Sachsen- Anhalt, Halle, 52 pp. Mai, D.H., 1983. Die fossile Pflanzenwelt des interglazialen Travertins von Bilzingsleben (Kreis Artern, Thu¨ringen). In: Mai, D.H., Mania, D., No¨tzold, T., Toepfer, V., Vlcek, E., Heinrich, W.-D., Bilzingsleben II. Homo erectus – seine Kultur und seine Umwelt. Vero¨ffentlichungen des Landesmuseums fu¨r Vorgeschichte in Halle 36, pp. 45–129. Martrat, B., Grimalt, J.O., Lopez-Martinez, C., Cacho, I., Sierro, F.J., Abel Flores, J., Zahn, R., Canals, M., Curtis, J.H., Hodell, D.A., 2004. Abrupt temperature changes in the western Mediterranean over the past 250 000 years. Science 306, 1762–1765. Maul, L., 2002. The Quaternary small mammal faunas of Thuringia (Central Germany). In: Meyrick, R.A., Schreve, D.C. (Eds.), The Quaternary of Central Germany. Field guide. Quaternary Research Association, London. pp. 79–95, 195–230. May, Th., 1993. Beeinflußten Großsa¨uger die Waldvegetation der pleistoza¨nen Warmzeiten Mitteleuropas? Natur und Museum 123 (6), 157–170. McBurney, C.B.M., Hey, R.W., 1955. Prehistory and Pleistocene Geology in Cyrenaican Libya. Occasional Publications of the Cambridge University Museum of Archaeology and Ethnology 4. XII, 314 pp. Meyrick, R.A., 2002. The Quaternary molluscan faunas of Thuringia (Central Germany). In: Meyrick, R.A., Schreve, D.C. (Eds.), The Quaternary of Central Germany. Field guide. Quaternary Research Association, London, 31–49, 195–230. Meyrick, R.A., Maul, L.C., 2002. Stratigraphy and biostratigraphy of the Eemian deposits of Burgtonna. In: Meyrick, R.A., Schreve, D.C. (Eds.), The Quaternary of Central Germany. Field guide. Quaternary Research Association, London, 145–161, 195–230. Moncel, M.-H., 2003. L’exploitation de l’espace et la mobilite´ des groupes humains au travers des assemblages lithiques a` la fin du Ple´istoce`ne moyen et au de´but du Ple´istoce`ne supe´rieur. La moyenne vale´e du Rhoˆne entre Droˆme et Arde`che, France. BAR International Series 1184, 179 pp. Moncel, M.-H., Neruda, P., 2000. The Ku˚lna level 11: some observations on the debitage rules and aims. The originality of a Middle Palaeolithic microlithic assemblage (Ku˚lna Cave, Czech Republic). Anthropologie 38, 219–247. Moncel, M.-H., Daujeard, C., Cre´gut-Bonnoure, E`., Fernandez, Ph., Faure, M., Gue´rin, Cl., 2004. L’occupation de la grotte de Saint-Marcel (Arde`che, France) au Pale´olithique moyen: strate´gie d’exploitation de l’environnement et type

191

d’occupation de la grotte. L’exemple des couches i, j et j’. Bulletin de la Socie´te´ Pre´historique Franc¸aise 101, 257–304. Nowel, W., 2003. Zur Korrelation der Glazialfolgen im Saale-Komplex Nord- und Mitteldeutschlands am Beispiel des Tagebaus Ja¨nschwalde in Brandenburg. Eiszeitalter und Gegenwart 52, 47–83. Patou-Mathis, M., 1997. Analyses taphonomique et palethnographique du materiel osseux de Krapina (Croatie): Nouvelles donne´es sur la faune et les restes humains. Pre´histoire Europe´enne 10, 63–90. Pe´csi, M., Scheuer, Gy., Schweitzer, F., 1988. Neogene and Quaternary geomorphical surfaces and litho-stratigraphical units in the Transdanubian Mountains. In: Pe´csi, M., Starkel, L. (Eds.), Paleogeography of the Carpartian regions. Theory – Methodology – Practice 47, pp. 11–41. Pe´rez-Pe´rez, A., Espurz, V., Bermu´dez de Castro, J.M., de Lumley, M.A., Turbo´n, D., 2003. Non-occlusal dental microwear variability in a sample of Middle and Late Pleistocene human populations from Europe and the Near East. Journal of Human Evolution 44, 497–513. Peter-Ro¨cher, H., 1994. Kannibalismus in der pra¨historischen Forschung. Studien zu einer paradigmatischen Deutung und ihren Grundlagen. Universita¨tsforschungen zur pra¨historischen Archa¨ologie 20. IX, 264 pp. Pettitt, P.B., 2000. Neanderthal lifecycles: developmental and social phases in the lives of the last archaics. World Archaeology 31, 351–366. Pettitt, P.B., 2002. The Neanderthal dead: exploring mortuary variability in Middle Palaeolithic Eurasia. Before Farming 2002/1, 1–19. Pfeiffer, Th., 1999: Sexualdimorphismus, Ontogenie und innerartliche Variabilita¨t der pleistoza¨nen Cervidenpopulationen von Dama Dama geiseliana Pfeiffer 1998 und Cervus elaphus L. (Cervidae, Mammalia) aus Neumark-Nord (Sachsen Anhalt, Deutschland). Berliner geowissenschaftliche Abhandlungen E30, 207–313. Radovcˇic´, J., Smith, F.H., Trinkaus, E., Wolpoff, M.H., 1988. The Krapina hominids: an illustrated catalog of skeletal collection. Mladost and Croatian Natural History Museum, Zagreb, 118 pp. Reiff, W., 1994. Die Abfolge der quarta¨ren Travertine im Stuttgarter Raum – ihre stratigraphische Zuordnung und o¨kologische Auswertung. Ethnographisch-Archa¨ologische Zeitschrift 35, 41–52. Reille, M., de Beaulieu, J.-L., Svobodova, H., Andrieu-Ponel, V., Coeury, C., 2000. Pollen analytical biostratigraphy of the last five climatic cycles from a long continental sequence from the Velay region (Massif Central, France). Journal of Quaternary Science 15, 665–685. Riel-Salvatore, J., Clark, G.A., 2001. Grave Markers. Middle and Early Upper Palaeolithic burials and

192

Stefan Wenzel

the use of Chronotypology in contemporary Paleolithic research. Current Anthropology 42, 449–479. Rink, W.J., Schwarcz, H.P., Valoch, K., Seitl, L., Stringer, C.B., 1996: ESR Dating of Micoquian Industry and Neanderthal Remains at Ku˚lna Cave, Czech Republic. Journal of Archaeological Science 23, 889–901. Roberts, D., Berger, L.R., 1997. Last Interglacial (c. 117 kyr) human footprints from South Africa. South African Journal of Science 93, 349–350. Roebroeks, W., Speleers, B., 2002. Last Interglacial (Eemian) human occupation of the North European plain and adjactent areas. In: Tuffreau, A., Roebroeks, W. (Dir.), Le Dernier Interglaciaire et les occupations humaines du Pale´olithique moyen. Colloque Crouy-Saint-Andre´, septembre 1997. Publications du CERP 8. 31–39. Roebroeks, W., Conard, N.J., Kolfschoten, T. van, 1992. Dense forests, cold steppes and the Palaeolithic Settlement of Northern Europe. Current Anthropology 33, 551–586. Russell, M.D., 1987a. Bone Breakage in the Krapina Hominid Collection. American Journal of Physical Anthropology 72, 373–379. Russell, M.D., 1987b. Mortuary Practices at the Krapina Neandertal Site. American Journal of Physical Anthropology 72, 381–397. Scha¨fer, J., 1996. Die Wertscha¨tzung außergewo¨hnlicher Gegensta¨nde (non-utilitarian objects) im Alt-und Mittelpala¨olithikum. EthnographischArcha¨ologische Zeitschrift 36, 173–190. Schmitz, R.W., 2005. Neue Funde aus dem Neandertal. In: Conard, N.J., Ko¨lbl, St., Schu¨rle, W. (Eds.), Vom Neandertaler zum modernen Menschen. Alb und Donau, Kunst und Kultur 46. pp. 153–168, 206–207. Schmitz, R.W., Thissen, J., 2000. Neandertal. Die Geschichte geht weiter. Spektrum Akademischer Verlag, Heidelberg and Berlin, XXI, 327 pp. Schu¨ler, T., 2000. Mittelpala¨olithische Artefakte aus dem Travertinsteinbruch von Burgtonna, Lkr. Gotha. Ausgrabungen und Funde im Freistaat Thu¨ringen 4 (1999), 1–6. Schu¨ler, T., 2003. Zersto¨rungsfreie ESR-Untersuchungen zur stratigraphischen Zuordnung von pleistoza¨nen Hominiden-Za¨hnen aus Altsammlungsbesta¨nden. In: Burdukiewicz, J.M., Fiedler, L., Heinrich, W.-D., Justus, A., Bru¨hl, E. (Eds.), Erkenntnisja¨ger. Kultur und Umwelt des fru¨hen Menschen. Vero¨ffentlichungen des Landesamtes fu¨r Archa¨ologie Sachsen-Anhalt 57/II. Halle (Saale) 2003, 537–540. Schulz, H.-P., Eriksson, B., Hirva, H., Hutha, P., Jungner, H., Purhonen, P., Ukkonen, P., Rankama, T., 2002. Excavations at Susiluola Cave. Suomen Museo, 109, 5–45. Schweigert, G., 1991. Die Flora der Eem-interglazialen Travertine von Stuttgart-Untertu¨rkheim

(Baden-Wu¨rttemberg). Stuttgarter Beitra¨ge zur Naturkunde, Serie B, 178. Staatliches Museum fu¨r Naturkunde Stuttgart, Stuttgart, 43 pp. Siddall, M., Chappell, J. (this volume). Sealevel change during past interglacials. Sirocko, F., Seelos, K., Schaber, K., Rein, B., Dreher, F., Diehl, M., Lehne, R., Ja¨ger, K., Krbetschek, M., Degering, D., 2005. A late Eemian aridity pulse in central Europe during the last glacial inception. Nature 436, 833–836. Soergel, W., 1922. Die Jagd der Vorzeit. Fischer, Jena, IV, 149 pp. Soergel, W., 1940. Zur biologischen Beurteilung diluvialer Sa¨ugetierfaunen. Sitzungsberichte der Heidelberger Akademie der Wissenschaften, Mathematisch-naturwissenschaftliche Klasse, Jahrgang 1940, 4. Abhandlung. Weis, Heidelberg, 39 pp. Somme´, J., 1989. Tuf de Caours (Somme). In: Le´colle, F. (Ed.), Les tufs et travertins quaternaires des bassins de la Seine et de la Somme et du littoral cauchMIS. Essai d’inventaire. Centre de ge´omorphologie de Caen, Bulletin 37 (not paginated site no. 31). Speleers, B., 2000. The relevance of the Eemian for the study of the Palaeolithic occupation of Europe. Geologie en Mijnbouw 79, 283–291. Steiner, W., 1977. Das geologische Profil des Travertin-Komplexes von Taubach bei Weimar. Quarta¨rpala¨ontologie 2, 83–118. Stiner, M.C., 1994. Honor among thieves. A zooarchaeological study of Neandertal ecology. Princeton University Press, Princeton. XXII, 447 pp. Stiner, M.C., Munro, N.D., 2002. Approaches to prehistoric diet breath, demography, and prey ranking systems in time and space. Journal of Archaeological Method and Theory 9, 181–214. Street, M., 2002. Plaidter Hummerich. An early Weichselian Middle Palaeolithic site in the Central Rhineland, Germany. Monographien des Ro¨mischGermanischen Zentralmuseums Mainz 45. VII, 160 pp. Stringer, Chr., Gamble, C., 1993. In search of the Neanderthals. Solving the puzzle of human origins. Thames & Hudson, London, 247 pp. Svoboda, J., 2002. The Last Interglacial in Southeast Central Europe: a short note. In: Tuffreau, A., Roebroeks, W. (Dir.), Le Dernier Interglaciaire et les occupations humaines du Pale´olithique moyen. Colloque Crouy-Saint-Andre´, septembre 1997. Publications du CERP 8. Lille, 15–19. Thieme, H., Veil, St., 1985. Neue Untersuchungen zum eemzeitlichen Elefanten-Jagdplatz Lehringen, Ldkr.Verden. Die Kunde N.F. 36, 11–58. Toepfer, V., 1958. Steingera¨te und Palo¨kologie der mittelpala¨olithischen Fundstelle Rabutz bei Halle (Saale). Jahresschrift fu¨r mitteldeutsche Vorgeschichte 41/42, 140–177.

Neanderthal Presence and Behaviour During MIS 5e Toepfer, V., 1978. Die pala¨olithischen Funde im Travertin von Burgtonna in Thu¨ringen. Quarta¨rpala¨ontologie 3, 175–178. Trinkaus, E., 1983. The Shanidar Neandertals. Academic Press, New York, London, XXIV, 502 pp. Tuffreau, A., 2001. Contextes et modalite´s des occupations humaines au Pale´olithique moyen dans la France septentrionale. In: Conard, N.J. (Ed.), Settlement dynamics of the Middle Palaeolithic and Middle Stone Age. Tu¨bingen Publications in Prehistory, pp. 293–314. Turk, I., Dirjec, J., 1991. Krapinski kanibalizem, kult lobanj in prokopi. Primerjalna tafonomska analiza fosilnih ostankov vrste Homo sapiens neanderthalensis iz Krapine (Hrvsˇka). (Zusammenfassung: Der Krapiner Kannibalismus, Scha¨delkult und Bestattungen.Vergleichende tafonomische Analyse der Fossilreste der Art Homo sapiens neanderthalensis aus Krapina (Kroatien). Procˇilo o raziskovanju paleolita, neolita i eneolita v Sloveniji 19, 131–144. Ullrich, H., 1989. Kannibalismus im Pala¨olithikum. In: Schlette, F., Kaufmann, D. (Eds.), Religion und Kult in ur- und fru¨hgeschichtlicher Zeit. Historiker-Gesellschaft der DDR, 13. Tagung der Fachgruppe Ur- und Fru¨hgeschichte vom 4. bis 6. November 1985 in Halle (Saale). Akademie-Verlag, Berlin, 51–71. Valoch, K., 1988. Die Erforschung der Ku˚lna-Ho¨hle 1961–76. Mit Beitra¨gen von J. Jelinek, W.G. Mook, R. Musil, E. Opravil, L. Seitl, L. Smolı´kova, H. Svobodova´. Anthropos 24 (N.S. 16). Moravske´ muzeum – Anthropos Institut, Brno, 318 pp. Vandermeersch, B., 1990. Les ne´andertaliens et les premiers hommes modernes. In: 5 millions d’annees. L’aventure humaine. Exposition Bruxelles 1990. Bruxelles. pp. 68–86. Vicino, G., 1976: Site du Casino. In: Lumley, H. de, Barral, L. (Eds.), Sites pale´olithiques de la re´gion de Nice et grottes de Grimaldi. Union

193

Internationale des Sciences Pre´historiques et Protohistoriques, IXe Congre`s, Nice, 13–18 septembre 1976. Livret-guide de l’excursion B1. Nice, 136–148. Vlcˇek, E., 1969. Neandertaler der Tschechoslowakei. Academia, Prag, 276 pp., 57 pl. Vlcˇek, E., 1991. L’Homme fossile en Europe centrale. L’Anthropologie 95, 409–472. Wansa, St., Wimmer, R., 1990. Geologie des Jungpleistoza¨ns der Becken von Gro¨bern und Grabschu¨tz. Altenburger Naturwissenschaftliche Forschungen 5, 49–91. Weber, Th., Litt, Th., 1991. Der Waldelefantenfund von Gro¨bern, Kr. Gra¨fenhainichen. Jagdbefund oder Nekrophagie? Archa¨ologisches Korrespondenzblatt 21, 17–32. Wenzel, St., 1994. Die pala¨olithischen Funde aus dem letztinterglazialen Travertin von StuttgartUntertu¨rkheim. Ethnographisch Archa¨ologische Zeitschrift 35, 59–66. Wenzel, St., 1998. Die Funde aus dem Travertin von Stuttgart-Untertu¨rkheim und die Archa¨ologie der letzten Warmzeit in Mitteleuropa. Universita¨tsforschungen zur pra¨historischen Archa¨ologie 52, 272 pp. Wenzel, St., 2002. Leben im Wald – die Archa¨ologie der letzten Warmzeit vor 125000 Jahren. Mitteilungen der Gesellschaft fu¨r Urgeschichte Blaubeuren 11, 35–63. White, M.J., Schreve, D.C., 2000. Island Britain – Peninsular Britain: Palaeogeography, colonisation, and the Lower Palaeolithic settlement of the British Isles. Proceedings of the Prehistoric Society 66, 1–28. Wied, Maximilian Prinz zu, 2001. Reise in das innere Nord-America in den Jahren 1832–1834. Illustriert von Karl Bodmer. Taschen, Cologne, 263 pp. Wolpoff, M.H., 1999. Paleoanthropology. McGrawHill, Boston, LVIII, 878 pp.

This page intentionally left blank

Section 3 Climate and Vegetation in Europe During MIS 5 (ed. Maria Fernanda Sa´nchez Gon˜i)

This page intentionally left blank

13. Introduction to Climate and Vegetation in Europe During MIS5 Marı´a Fernanda Sa´nchez Gon˜i Ecole Pratique des Hautes Etudes, DGO, UMR-CNRS 5805, EPOC, Universite´ Bordeaux 1, Avenue des Faculte´s, 33405 Talence

The paradox of the last interglacial consists in the fact that while being the best documented among similar warm periods, its study has created a body of knowledge that reveals more and more complexity and multiplies rather than answers questions on its impact on different Earth reservoirs. Perhaps, the principal reason for this complexity is historical. The main climatic trends that characterised this period were discovered over more than one century in parallel with the development of the various approaches and methods that have made palaeoclimatology the composite discipline that we know now. The first evidence for the warm period that marine geologists call marine isotopic stage (MIS) 5 was found 150 years ago in European sediments. It was called Eemian by Harting in 1852 when he described a stratigraphic unit from the Amsterdam and Amersfoort areas characterised by clays and sands rich in Mediterranean and Lusitanian warm species of mollusc and diatom fossils (Bosch et al., 2000). Subsequent pollen investigations conducted in Denmark between the two World Wars characterised the Eemian as a period of widespread Quercus, Corylus and Carpinus woodland (Jessen and Milthers, 1928). In the 1960s, Zagwijn (1961) showed a similar expansion of the deciduous forest in Germany. Meanwhile, on the other side of the Atlantic, Emiliani published the first generalised temperature curve for the surface water of the central Caribbean over the last 425 000 years, thus paving the way for marine isotopic stratigraphy. Emiliani’s curve (Emiliani, 1955, 1966) identified seventeen isotopic stages

alternating between glacial and interglacial periods. MIS 5 was the last interglacial, occurring at around 100 000 years ago. Shackleton (1969) refined this stratigraphy and proposed, based on the correlation between marine and terrestrial records, that only the first 18 O minimum values of MIS 5, named MIS 5e, had to be considered the equivalent of the Eemian as identified on land. This was confirmed fifteen years later by Turon (1984) who established the first direct though low-resolution correlation between pollen and benthic isotopic data in a core retrieved off the north-western Iberian margin. Both Turon’s results and the counting of annually laminated layers in the sequence of Bispingen, northern Germany (Mu¨ller, 1974) gave 11 000-year duration for the Eemian, between 126 and 115 kyr. We know now that MIS 5, often referred to as the last interglacial sensu lato (LIGA Members, 1991), was a period of minimum ice volume that expanded from c. 130 to 75 kyr. The first and warmest interval, MIS 5e characterised by sealevels 0 to 6 m higher than those of the present day (Stirling et al., 1998; Schellmann and Radtke, 2004), was followed by two cold episodes in the ocean (MIS 5d and MIS 5b) alternating with two warm ones (MIS 5c and MIS 5a). In western Europe, the Eemian warmest period led to two temperate intervals interpreted as interglacials (palynological definition): St. Germain I and St. Germain II separated by two very cold phases: Me´lisey I and Me´lisey II (Woillard, 1978). The St. Germain I forest period was interrupted by the Montaigu cold suborbital event (Reille et al., 1992).

Marı´a Fernanda Sa´nchez Gon˜i

198

Recent direct correlation between highresolution isotopic and pollen data shows that MIS 5 encompasses the Zeifen interstadial, the short stadial event before the Eemian and the four cold/warm cycles, Me´lisey I/Saint-Germain Ia, Montaigu/ Saint-Germain Ic, Me´lisey II/Saint-Germain II and Stadial I/Ognon I interstadial, following it (Sa´nchez Gon˜i et al., 1999, 2005) (Fig. 13.1). MIS 5e has long been considered the best analogue for the present interglacial

MD95-2042 (SW Iberian margin, 37°48′N; 10°10′W)

O

ic

n

Be

a

Pl

4

GIS19

Stadial II Ognon I Stadial I St-Germain II Mélisey II

C19

GIS20

C20

80

5a

82.9

GIS21

C21

5b

GS 22

90

δ

on

t nk

GIS22

C22

GS 23

St-Germain lc

5c

100

GIS23 GIS24 110

C23

Montaigu event St-Germain la Mélisey I

GS 24 C24 GS 25

5d

C25

GIS25 C26

116.1±0.9

GS 26 120

Eemian

5e Stadial Zeifen Interstadial

128±1 129±0.8 132±2

130

Stadial –40

‰

–30

5 ‰

4

2

3 ‰

1

0 10 °C

20

0 20 40 0 20 %

0 20

LANTERNE GLACIAL

70

δ

70

80

LAST INTERGLACIAL COMPLEX

δ

i th

bs ru sh

90

100

110

120

LINEXERT GLACIAL

Age

c

O

ss tre

& s ee ts us n tr lan l tu a t p E & rn be ane ser te EA fe C ra us err -de s e n A i t i te p C p di m I e r m o a e Is ER Te C M Se id um H

) (s s a st ts cy rm ys e T no oc d i y n S d h di -S m ac d ar .p ol C U N W

O

k 37

18

18

18

NorthGRIP

Kyr BP

and has for this reason attracted the interest of many Quaternary geologists, palaeoclimatologists and modellers. The paper by Kukla and Matthews (1972) remains famous in which he predicts an imminent ice age after observing that our interglacial has already lasted for 11 000 years, i.e. the same time span of MIS 5e. We know now that from an astronomical point of view, the best analogue for the present interglacial is MIS 11, due to the weaker character of its insolation changes.

130

0 20 40 0 0 %

0 20 40 0 20 40 0 20

Fig. 13.1 Comparison between the 18 O record from NorthGRIP ice core and the multiproxy record from the MIS 5 interval in marine core MD95-2042. From left to right: the benthic isotopic curve indicating the marine isotopic stratigraphy, the planktic isotopic curve reflecting sea-surface hydrological changes, mainly in temperature and salinity (Shackleton et al., 2002), the alkenone-derived sea-surface temperatures (Pailler and Bard, 2002), the percentage curve of the polar foraminifera Neogloboquadrina pachyderma (s), the percentage curves of cold and warm dinocyst assemblages (Eynaud et al., 2000), the terrestrial stratigraphy and the curves of the main pollen ecological groups: Temperate and humid tree curve mainly includes Abies, Betula, Corylus, deciduous Quercus, Carpinus betulus, Fraxinus excelsior-type, Acer, Ulmus, Alnus, Hedera and Salix; Mediterranean tree and shrub curve includes evergreen Quercus, Olea, Phillyrea, Pistacia and Cistus; semi-desert plant curve includes Artemisia, Chenopodiaceae and Ephedra dystachia and E. fragilis-types; humid plant curve refers to Ericaceae. The NorthGRIP age scale is based on the GRIP ss09sea or GRIP 2001 chronology (Johnsen et al., 2001). The chronology of most of the MIS 5 interval in core MD95-2042 is derived from Shackleton et al. (2003); however, the chronology of the interval between 81.9 and 69 kyr is based on Shackleton et al. (2004). C26 to C19 refer to marine cold events identified by Chapman and Shackleton (1999) in the North Atlantic Ocean. GS and GIS refer to Greenland stadials and interstadials, respectively.

Introduction to Climate and Vegetation

During the last decades, hypotheses on the timing, duration, warmth amplitude, degree of climatic stability of MIS 5e and mechanisms leading to glacial inception have been repeatedly challenged (Dansgaard et al., 1993; Turner, 2002; Tzedakis, 2003). 13.1 TIMING AND DURATION Although it is widely accepted that climate variations during the MIS 5 were largely determined by insolation, other factors and feedback mechanisms certainly played a crucial role in determining the timing and duration of the penultimate deglaciation, the full interglacial and the last glacial inception. The onset of the penultimate deglaciation predicted by the theory of the Ice Ages of Milankovitch at around 128 kyr has been challenged by Winograd et al. (1992) who sees an early warming before deglaciation in the 18 O of the Devils Hole speleotheme record. This observation, consistent with uranium series-dated corals indicating a sea level close to modern levels as early as 132 to 135 kyr, is not easily reconcilable with the insolation forcing of the penultimate deglaciation. However, as Herbert et al. (2002) point out, Devils Hole 18 O record, detecting a local 18 O signature of temperature and precipitations over the northern Pacific region, cannot be considered a reliable indicator of ice volume changes. A lively debate also concerns the full interglacial and its end. In two recent works, Kukla and collaborators (Kukla et al., 1997, 2002b; Kukla, 2000) propose, based on an indirect correlation between astronomically tuned benthic isotopic curves from several North Atlantic deep-sea cores and the Grand Pile pollen record from northeastern France, that the Eemian in northeastern France is not the terrestrial equivalent of MIS 5e but encompasses the entire MIS 5e and a part of MIS 5d. This implies that the Eemian in this region lasted 22 000 years,

199

between 130 and 108 kyr. Recent studies (Shackleton et al., 2002, 2003; Sa´nchez Gon˜i et al., 2005), based on an astronomically independent age model and the direct correlation of marine, ice and terrestrial climatic records from cores MD95-2042 and MD99-2331 retrieved off western Iberian margin, reach conclusions that challenge Kukla et al.’s interpretation. These authors show that (1) the Eemian in Iberia started 6000 years later than the onset of MIS 5e (132 kyr), and (2) it ended 6000 years after the end of MIS 5e (115 kyr). In other words, according to this study, the Iberian Eemian lasted 16 000 years (from 126 to 110 kyr), that is 6000 years less than suggested by Kukla et al. for northeastern France. Considering the northernmost location of La Grande Pile compared to the Iberian records, the hypothesis of an onset of the Eemian in the Vosges region at 130 kyr is difficult to reconcile with the observation, based on direct correlation of ice and terrestrial proxies, that typical Eemian forest formations only appear in the South at 126 kyr. A discrepancy between the two studies also exists for the end of the Eemian, which is situated at 108 kyr in northeastern France and at 110 kyr in the South. The reason for this difference is that Kukla et al. (1997: Fig. 2) correlate the increase of IRD concentration and that of N. pachyderma (s) percentages that mark the C24 cold event in the North Atlantic (108 kyr) (GS25 in the ice core terminology, NorthGRIP Members, 2004) with the end of the Eemian conifer forest, while the direct correlation between pollen spectra and marine proxies (Shackleton et al., 2003; Sa´nchez Gon˜i et al., 2005) indicates that this event is contemporaneous with the open vegetation of the Me´lisey 1. The study conducted on the Iberian margin also indicates that a time lag exists between ice volume decay/growth and forest development in Europe during the last interglacial sensu stricto: (1) the minimum in ice volume is reached at 128 kyr, 2000 years

200

Marı´a Fernanda Sa´nchez Gon˜i

before the onset of the Eemian forest (126 kyr), and (2) the substantial accumulation of ice in high northern latitudes (MIS 5e/MIS 5d transition) occurs at 115 kyr, i.e. 5000 years before the demise of the temperate and conifer forests in Iberia and France, respectively. The 11 000-year duration of the Eemian in northern European latitudes above 50 N (Turner, 2002) suggests that tundra vegetation expanded during the first 5000 years of MIS 5d when forest still occupied southern latitudes. This occurred synchronously with the C26 (115 kyr) and C25 (112 kyr) cold events in the North Atlantic surface temperatures (Chapman and Shackleton, 1999). Both marine cold events may correlate with the gradual GS26 cooling in Greenland (Fig. 13.1). The observed time lag between ice volume and vegetation shifts in Europe is changing our view on the response of the different Earth’s reservoirs to fluctuations in insolation. It is known that the weak decrease in summer insolation at 65 N predicted by Milankovitch at 126–115 kyr could not solely trigger glacial inception. Simulations using the atmosphere–ocean global circulation model have highlighted the role of changes in North Atlantic sea-surface temperature gradient between low and high latitudes as the major triggering factor for the last glacial inception (Khodri et al., 2001). Using the earth model of intermediate complexity (EMIC) MoBidiC, Crucifix and Loutre (2002) and Loutre et al. (this volume) attribute this role to the extension of the tundra that would double the albedo at 122–120 kyr. Pollen data are in agreement with this prediction as they identify a replacement of the temperate forest by Conifers at 52 N as early as 120 kyr. This suggests a southward displacement of the boreal vegetation belt of around 10 by comparison to its location (60 N–70 N) at the beginning of the Eemian (Sa´nchez Gon˜i et al., 2005). Kageyama et al. (2004, this volume) recently confirmed, using another EMIC model, the major role of vegetation changes as a triggering factor for ice growth, sea-surface

temperature decrease acting only as an enhancer of the glaciation process. 13.2 PACING AND AMPLITUDE OF THE EEMIAN CLIMATIC VARIABILITY The Eemian interglacial has been traditionally seen as a period characterised by a rather uniform warm climate, as observed for the Holocene and predicted by astronomical parameters. This view was contradicted by the identification of highamplitude climatic changes in the atmosphere of Greenland detected by the GRIP isotopic record (Dansgaard et al., 1993) and by magnetic susceptibility measurements of the Lac de Bouchet in the French Massif Central (Thouveny et al., 1994). Such cold spells would imply the existence of suborbital forcing episodes not predicted by Milankovich theory. These coolings, however, are neither recorded by cold water indicators in the central North Atlantic (McManus et al., 1994) nor by high-resolution terrestrial and marine records from western Iberian margin deep-sea cores (Sa´nchez Gon˜i et al., 1999, 2005), nor in European pollen and diatom archives (Cheddadi et al., 1998; Frogley et al., 1999; Rioual et al., 2001). These last sequences unambiguously indicate that the Eemian climatic variability was similar to that of the Holocene and involved no return to glacial conditions. The high amplitude of changes in the GRIP isotopic signal is now explained as an artefact created by ice flow (Chappellaz et al., 1997; Landais et al., 2003). The first major cold episode associated with substantial iceberg discharges is labelled C24 and coincides (110–107 kyr), based on the direct correlation between marine proxies and pollen, with the Me´lisey I steppic period on land (Shackleton et al., 2002, 2003; Sa´nchez Gon˜i et al., 2005). A weak though substantial cooling in the middle of the Eemian (ca. 121 kyr) is nevertheless recorded in marine pollen and sea-surface temperature records from the Iberian

Introduction to Climate and Vegetation

margin (Sa´nchez Gon˜i et al., 2005). It can be correlated with the so-called Eemian cooling previously detected in the North Atlantic (Cortijo et al., 1994), which indicates the southward displacement of the Polar Front, when ice volume was still at the minimum. Such a cold spell has been related to a reduction of the upper North Atlantic deep-water ventilation (Maslin et al., 1998). Several weak cooling episodes (1–2 C lower than today) within the Eemian time interval, one of which seems to coincide with the 121-kyr cooling, have been detected in a high Alpine stalagmite and interpreted as the result of changes in solar activity (Holzka¨mper et al., 2004). The new NGRIP isotopic record extending back to 123 kyr indicates warmer temperatures in Greenland between 123 and 120 kyr than during the Holocene and a unique gradual cooling from 120 to 115 kyr (NorthGRIP Members, 2004). The time of this cooling seems to coincide with the first southward displacement of vegetation belts in western Europe and of the North Atlantic current (Mu¨ller and Kukla, 2004) in response to summer insolation decrease at higher northern latitudes. Also, this cooling would coincide with the late Eemian aridity pulse in Central Europe recorded in the Eifel region and associated with dust storms and bushfire (Sirocko et al., 2005; Seelos and Sirocko, this volume). The moderate increase in insolation that characterises the following period, between 115 and 110 kyr, sees a slight increase in temperate trees in southwestern Iberia (Sa´nchez Gon˜i et al., 1999) and a contemporaneous warming in Greenland, called GIS25 in NGRIP. This warming, however, does not seem to influence the ongoing ice growth, which probably determined the end of the Eemian in northwestern Europe. The albedo increase at that time may have overridden the effect of increasing summer insolation. In sum, a consensus exists at present (but see Rousseau et al., this volume) on the relatively stable nature of the Eemian climate but also on the presence of short and lowamplitude cooling phases during this period.

201

Was the last interglacial optimum warmer than the Holocene? Oxygen isotope and micropalaeontological analyses of 52 cores across the world ocean conducted in the framework of the CLIMAP project concluded that the warmth of the last interglacial ocean was not significantly different from the present one (CLIMAP, 1984). In contrast, a more recent project (LIGA Members, 1991) seems to reveal, based on a mapped compilation of data from the Northern Hemisphere, that the last interglacial was on average warmer than the present in both terrestrial and marine environments. This conclusion is consistent with the 18 O signal from NorthGRIP (NorthGRIP Members, 2004) and pollen-derived quantitative climatic reconstructions from northern Germany (Ku¨hl and Litt, 2003) but in disagreement with other marine and terrestrial reconstructions (Maslin et al., 1998; Frogley et al., 1999; Sa´nchez Gon˜i et al., 2000, 2005). Clearly, many questions concerning the frequency, amplitude and abruptness of the climatic and environmental changes that have characterised the last interglacial at different locations remain open. This chapter of the book brings together contributions mainly focusing on the impact of MIS 5 and in particular the MIS 5e on Eurasia and the North Atlantic. The mechanisms that have driven climatic changes during the period at hand are the focus of Sections 1 and 5. The timing of these events is thoroughly addressed in Section 2. Comparison of the last interglacial with preceding similar warming intervals is the main topic of Section 4. Seelos and Sirocko’s paper takes us to the heart of the debate that constitutes one of the major focuses of this book. They document in detail by means of high-resolution pollen and grain-size data from the Eifel region a number of phases of widespread dust dispersal within MIS 5 and suggest that these phases are coeval with the cold events in Greenland and in North Atlantic sea-surface temperatures. In particular, they detect a major occurrence of dust in

202

Marı´a Fernanda Sa´nchez Gon˜i

northern Germany that tentatively correlates with the last glacial inception centred at around 118 kyr. In their challenging paper, Rousseau and collaborators seem to identify large amplitude coolings – an average of 10 C in winter temperature – within the Eemian in northeastern France. They reach this conclusion by applying the BIOME 4 vegetation model to the new pollen data from La Grande Pile and constraining the results with 13 C measurements. This conclusion is, however, contradicted by the paper of Ku¨hl and Litt, which presents pollen reconstructions of January and July temperatures at two northern German Eemian sites in addition to the same site of La Grande Pile. Using a method based on probability density functions ( pdf method), they show that the Eemian started with a relatively fast warming and ended with a pronounced cooling, while during its course it was marked by uninterrupted interglacial conditions. Moving to the northeast of Europe, Velichko et al. (this volume) suggest, based on pollen sequences distributed in the area between 50 to 55 N and 10 to 40 E, that after the Eemian climatic optimum, environment and climate changes became more pronounced eastward. Boettger et al. (this volume) uses pollen and geochemical tracers to address a controversial issue, namely the weak-amplitude warming event at the end of the Eemian as recorded in Central and Eastern Europe, and its possible connection to the slight warming GIS25 in the NorthGRIP record. They conclude that this warming was a global phenomenon affecting the whole northern hemisphere and compare it with the current global warming. Of course, the main difference between the two events lies in the fact that the present warming is associated with an increase in human-derived CO2 concentrations, while the Eemian was characterised, as clearly indicated by the Antarctic record, by a gradual decline in CO2 values. In contrast with the conclusions reached by Boettger et al., Mu¨ller and Sa´nchez Gon˜i compare the two longest pollen records from southern

Germany with northern and southern European vegetation archives and suggest steep vegetation gradients between northern and southern Germany during the inception of the last glacial involving the albedo increase and the southward migration of the North Atlantic current. A detailed study of the Gulf Stream fluctuations during MIS 5e and MIS 5d by Vautravers and co-workers (this volume), based on the study of two ODP cores located on the western end of the subtropical gyre, detects high-frequency cold events at subtropical latitudes for the time span between 134 and 108 kyr and convincingly shows that these latitudes were also dramatically affected by shifts in conditions associated with the penultimate deglaciation. The following period is characterised by a sequence of climatic events with the warmest, peaking at 125 kyr, interpreted as the maximum influence of the Gulf Stream. Superimposed onto the declining temperature from 122 to 108 kyr, a number of weak cooling events in winter, most of them likely associated with ice rafting, are also detected. Wu¨nnemann et al.’s (this volume) contribution documents, by analysing geochemical precipitates and sediment grain size of a core from the centre of the Gaxun Nur Basin (Gobi desert, NW China), the close relationship between environmental conditions in distant desert regions of NW China, global ice volume and North Atlantic and Greenland regional climates. During MIS 5e, they detect a large and slightly saline lake, which filled the entire basin as a result of summer monsoon precipitation increase associated with a strong northward shift beyond the modern limit. They date the end of this monsoon enhancement at around 118 kyr, when geochemical and sedimentological proxies indicate an intensification of aeolian transport in NW China. Their noteworthy conclusions are similar to those reached by Seelos and Sirocko for Central Europe. Finally, the last contribution of this chapter moves us to the southern Pacific

Introduction to Climate and Vegetation

ocean. The paper by Rein et al. (this volume) focuses on the abrupt, likely orbitally driven changes of El Nin˜o activity during MIS 5e and MIS 5d. Using proxies for sea-surface temperatures and lithic flux estimates from the continent, they document a sharp drop of El Nin˜o activity at around 122 kyr. In conclusion, the complexity, in both time and space, of the nonlinear climatic signal in response to insolation changes during the penultimate deglaciation, last interglacial and last glacial inception is clearly shown by the collection of papers presented in this chapter. If problems of nomenclature and stratigraphy, reflecting the history of the discipline, are on the way to being solved, others, due to uncertainties in the chronology of a number of MIS 5 records, require further investigation. Our scientific community should concentrate its efforts on accurately correlating the available records. It is only through this approach that we will be able to document the climatic variability of MIS 5 in an integrated way, link the processes reflected in different reservoirs and propose reliable scenarios for the mechanisms underlying the climatic variability of the last interglacial. ACKNOWLEDGEMENTS To Sigfus Johnsen for his contribution and the stimulating discussion during the writing of this work. This is contribution n 1612 REFERENCES Bosch, J. H. A., Cleveringa, P., Meijer, T., 2000. The Eemian stage in the Netherlands: history, character and new research. Geologie en Mijnbouw/ Journal of Geosciences 79, 135–145. Boettger, T., Junge, F. W., Knetsch, S., Novenko, E. Yu., Borisova, O. K., Velichko, A. A. (this volume). Indications of short-term climate warming at the very end of the Eemian in terrestrial records of Central and Eastern Europe. In: Sirocko, Claussen, M. F., Litt, T. & Sa´nchez Gon˜i, M. F. (eds.) The climate of past integlacials.

203

Chapman, M. R., Shackleton, N. J., 1999. Global ice-volume fluctuations, North Atlantic ice-rafted events, and deep-ocean circulation changes between 130 and 70 ka. Geology 27, 795–798. Chappellaz, J., Brook, E., Blunier, T., Malaize, B., 1997. CH4 and 18 O of O2 records from Antarctic and Greenland ice: a clue for stratigraphic disturbance in the bottom part of the GRIP and GISP2 ice-cores. Journal of Geophysical Research 102, 26547–26557. Cheddadi, R., Mamakowa, K., Guiot, J., de Beaulieu, J.-L., Reille, M., Andrieu, V., Granoszewski, W., Peyron, O., 1998. Was the climate of the Eemian stable? A quantitative climate reconstruction from seven European pollen records. Palaeogeography, Palaeoclimatology, Palaeoecology 143, 73–85. CLIMAP., 1984. The last interglacial ocean. Quaternary Research 21, 123–224. Cortijo, E., Duplessy, J.-C., Labeyrie, L., Leclaire, H., Duprat, J., van Weering, T. C. E., 1994. Eemian cooling in the Norwegian Sea and North Atlantic Ocean preceding continental ice-sheet growth. Nature 372, 446–449. Crucifix, M., Loutre, M. F., 2002. Transient simulations over the last interglacial period (126–115 kyr BP): feedback and forcing analysis. Climate Dynamics 19, 417–433. Dansgaard, W., Johnsen, S. J., Clausen, H. B., Dahl-Jensen, D., Gundestrup, N. S., Hammer, C. U., Hvidberg, C. S., Steffensen, J. P., Sveinbjo¨rnsdottir, A. E., Jouzel, J., Bond, G., 1993. Evidence for general instability of past climate from a 250-kyr ice-core record. Nature 364, 218–220. Emiliani, C., 1955. Pleistocene temperature variations in the Mediterranean. Quaternaria 2, 87–98. Emiliani, C., 1966. Isotopic paleotemperatures. Science 154, 851–857. Eynaud, F., Turon, J.-L., Sa´nchez Gon˜i, M. F., Gendreau, S., 2000. Dinoflagellate cyst evidence of Heinrich-like events off Portugal during Marine Isotopic Stage 5. Marine Micropaleontology 40, 9–21. Frogley, M. R., Tzedakis, P. C., Heaton, T. H. E., 1999. Climate variability in northwest Greece during the last interglacial. Science 285, 1886–1889. Harting, P., 1852. De bodem onder Amsterdam onderzocht en beschreven. Verhandelingen 1e klas Koninklijk Nederlands Instituut van Wetenschappen, 2e Reeks 8, 282–290. Herbert, T. D., Schuffert, J. D., Andreasen, D., Heusser, L., Lyle, M., Mix, A., Ravelo, A. C., Stott, L. D., Herguera, J. C., 2002. Response to ‘‘The California Current, Devils Hole, and Pleistocene climate’’. Science 296, doi: 10.1126/science.296.5565.7a. Holzka¨mper, S., Mangini, A., Spo¨tl, C., Mudelsee, M., 2004. Timing and progression of the Last Interglacial derived from a high alpine stalagmite.

204

Marı´a Fernanda Sa´nchez Gon˜i

Geophysical Research Letters 31, doi: 10.1029/ 2003GL019112. Jessen, K., Milthers, V., 1928. Stratigraphical and palaeontological studies of interglacial freshwater deposits in Jutland and northwest Germany. Danmarks Geologiske Underso¨gelse, II 48, 1–380. Johnsen, S. J., Dahl-Jensen, D., Gundestrup, N., Steffensen, J. P., Clausen, H. B., Miller, H., Masson-Delmotte, V., Sveinbjo¨rnsdottir, A. E., White, J., 2001. Oxygen isotope and palaeotemperature records from six Greenland ice-core stations: Camp Century, Dye-3, GRIP, GISP2, Renland and NorthGRIP. Journal of Quaternary Science, 16, 299–307. Kageyama, M., Charbit, S., Ritz, C., Khodri, M., Ramstein, G., 2004. Quantifying ice-sheet feedbacks during the last glacial inception. Geophysical Research Letters 31, 1–4. Kageyama, M., Charbit, S., Ritz, C., Khodri, M., Ramstein, G. (this volume). Mechanisms leading to the last glacial inception over North America: results from the CLIMBER-GREMLINS atmosphere-ocean-vegetation-northern hemisphere icesheet model. In : Sirocko, Claussen, M.F., Litt, T. & Sa´nchez Gon˜i, M.F. (eds.) The climate of past integlacials. Khodri, M., Leclainche, Y., Ramstein, G., Braconnot, P., Marti, O., Cortijo, E., 2001. Simulating the amplification of orbital forcing by ocean feedbacks in the last glaciation. Nature 410, 570–574. Ku¨hl, N., Litt, T., 2003. Quantitative time series reconstruction of Eemian temperature at three European sites using pollen data. Vegetation History and Archaeobotany 13, 205–214. Ku¨hl, N., Litt, T. (this volume). Quantitative time series reconstructions of Holsteinian and Eemian temperatures using botanical data. In : Sirocko, Claussen, M. F., Litt, T. & Sa´nchez Gon˜i, M. F. (eds.) The climate of past integlacials. Kukla, G. J., 2000. The last interglacial. Science 287, 987–988. Kukla, G., Matthews, R. K., 1972. When will the present interglacial end? Science 178, 190–191. Kukla, G., McManus, J. F., Rousseau, D.-D., Chuine, I., 1997. How long and how stable was the last interglacial? Quaternary Science Reviews 16, 605–612. Kukla, G. J., de Beaulieu, J.-L., Svobodova, H., Andrieu-Ponel, V., Thouveny, N., Stockhausen, H., 2002. Tentative correlation of pollen records of the last interglacial at Grande Pile and Ribains with marine isotope stages. Quaternary Research 58, 32–35. Landais, A., Chappellaz, J., Delmotte, M., Jouzel, J., Blunier, T., Bourg, C., Caillon, C., Cherrier, S., Malaize´, B., Masson-Delmotte, V., Raynaud, D., Schwander, J., Steffensen, J. P., 2003. A tentative

reconstruction of the last interglacial and glacial inception in Greenland based on new gas measurements in the GRIP ice core. Journal of Geophysical Research 108, n D18, 4563, doi: 10.1029/ 2002JD003147. LIGA Members, 1991. Report of 1st discussion group: the last interglacial in high latitudes of the northern hemisphere: terrestrial and marine evidence. Quaternary International 10–12, 9–28. Loutre, M. F., Berger, A., Crucifix M., Desprat, S., Sa´nchez Gon˜i, M. F. (this volume). Interglacials as simulated by MoBidiC and the LLN 2-D NH climate models. In : Sirocko, Claussen, M. F., Litt, T. & Sa´nchez Gon˜i, M. F. (eds.) The climate of past integlacials. Maslin, M., Sarnthein, M., Knaack, J. J., Grootes, P., Tzedakis, C. (1998). Intra-interglacial cold events: an Eemian-Holocene comparison. In: Cramp, A., MacLeod, C. J., Lee, S. V. & Jones, E. J. W. (eds.) Geological evolution of oceans basins: Results from the ocean drilling program. Geological Society, Special Publications, London, 91–99. McManus, J. F., Bond, G. C., Broecker, W. S., Johnsen, S., Labeyrie, L., Higgins, S., 1994. High-resolution climate records from the North Atlantic during the last interglacial. Nature 371, 326–329. Mu¨ller, H., 1974. Pollenanalytische Untersuchungen und Jahresschichtenza¨hlungen an der eemzeitlichen Kieselgur von Bispingen/Luhe. Geologisches Jahrbuch A 21, 149–169. Mu¨ller, U. C., Kukla, G. J., 2004. North Atlantic Current and European environments during the declining stage of the last interglacial. Geology 32, 1009–1012. Mu¨ller, U. C., Sa´nchez Gon˜i, M. F. (this volume). Vegetation dynamics in southern Germany during marine isotope stage 5 ( 130 to 70 kyr ago). In : Sirocko, Claussen, M. F., Litt, T. & Sa´nchez Gon˜i, M. F. (eds.) The climate of past integlacials. NorthGRIP Members, 2004. High resolution climate record of the Northern Hemisphere reaching into the last glacial interglacial period. Nature 431, 147–151. Pailler, D. Bard, E., 2002. High frequency palaeoceanographic changes during the past 140,000 yr recorded by the organic matter in sediments of the Iberian margin. Palaeogeography, Palaeoclimatology, Palaeoecology 181, 431–452. Reille, M., Guiot, J., de Beaulieu, J.-L., 1992. The Montaigu event: an abrupt climatic change during the early Wu¨rm in Europe. In: Went, G. J. K. E. (ed.) Start of a glacial. NATO ASI Series. SpringerVerlag, Berlin Heidelberg, 85–95. Rein, B., Sirocko, F., Lu¨ckge, A., Reinhardt, L., Wolf, A., Dullo, W. C. (this volume). Abrupt change of El Nin˜o activity off Peru during stage MIS 5e-5d. In: Sirocko, Claussen, M. F., Litt, T. &

Introduction to Climate and Vegetation Sa´nchez Gon˜i, M.F. (eds.) The climate of past integlacials. Rioual, P., Andrieu-Ponel, V., Rietti-Shati, M., Battarbee, R. W., de Beaulieu, J.-L., Cheddaddi, R., Reille, M., Svobodova, H., Shemesh, A., 2001. High-resolution record of climate stability in France during the last interglacial period. Nature 413, 293–296. Rousseau, D.-D., Hatte´, C., Duzer, D., Schevin, P., Kukla, G., Guiot, J. (this volume). Estimates of temperature and precipitation variations during the Eemian interglacial: New data from the Grande Pile record (GP XXI). In: Sirocko, Claussen, M. F., Litt, T. & Sa´nchez Gon˜i, M. F. (eds.) The climate of past integlacials. Sa´nchez Gon˜i, M. F., Eynaud, F., Turon, J.-L., Shackleton, N. J., 1999. High resolution palynological record off the Iberian margin: direct land-sea correlation for the last interglacial complex. Earth and Planetary Science Letters 171, 123–137. Sa´nchez Gon˜i, M. F., Eynaud, F., Turon, J.-L., Shackleton, N. J., Cayre, O., 2000. Direct land-sea correlation for the Eemian and its comparison with the Holocene: a high resolution palynological record off the Iberian margin. Geologie en Mijnbouw/ Netherlands. Journal of Geosciences 79, 345–354. Sa´nchez Gon˜i, M. F., Loutre, M. F., M., C., Peyron, O., Santos, L., Duprat, J., Malaize´, B., Turon, J.-L., Peypouquet, J.-P., 2005. Increasing vegetation and climate gradient in Western Europe over the last glacial inception (122–110 ka): data-model comparison. Earth and Planetary Science Letters 231, 111–130. Schellmann, G., Radtke, U., 2004. A revised morphoand chronostratigraphy of the late and middle Pleistocene coral reef terraces on Southern Barbados (West Indies). Earth Science Reviews 64, 157–187. Seelos, K., Sirocko, F. (this volume). Abrupt cooling events at the very end of the last interglacial. In: Sirocko, Claussen, M. F., Litt, T. & Sa´nchez Gon˜i, M. F. (eds.) The climate of past integlacials. Shackleton, N. J., 1969. The last interglacial in the marine and terrestrial records. Proceedings of the Royal Society B 174, 135–154. Shackleton, N. J., Chapman, M., Sa´nchez-Gon˜i, M. F., Pailler, D., Lancelot, Y., 2002. The classic marine isotope substage 5e. Quaternary Research 58, 14–16. Shackleton, N. J., Sa´nchez Gon˜i, M. F., Pailler, D., Lancelot, Y., 2003. Marine isotope substage 5e and the Eemian interglacial. Global and Planetary Change 757, 1–5. Shackleton, N. J., Fairbanks, R. G., Chiu, T., Parrenin, F., 2004. Absolute calibration of the Greenland time scale: implications for Antarctic time scales and for 14 C. Quaternary Science Reviews 23, 1513–1523.

205

Sirocko, F., Seelos, K., Schaber, K., Rein, B., Dreher, F., Dielh, M., Lehne, R., Ja¨ger, K., Krbetschek, M., Degering, D., 2005. A late Eemian aridity pulse in central Europe during the last glacial inception. Nature 436, 833–836. Stirling, C. H., Esat, T. M., Lambeck, K., McCulloch, M. T., 1998. Timing and duration of the last interglacial: evidence for a restricted interval of widespread coral reef growth. Earth and Planetary Science Letters 160, 745–762. Thouveny, N., de Beaulieu, J. L., Bonifay, E., Creer, K. M., Guiot, J., Icole, M., Johnsen, S., Jouzel, J., Reille, M., Williams, T., Williamson, D., 1994. Climate variations in Europe over the past 140 kyr deduced from rock magnetism. Nature 371, 503–506. Turner, C., 2002. Problems of the duration of the Eemian interglacial in Europe North of the Alps. Quaternary Research 58, 45–48. Turon, J.-L., 1984. Direct land/sea correlations in the last interglacial complex. Nature 309, 673–676. Tzedakis, P. C., 2003. Timing and duration of last interglacial conditions in Europe: a chronicle of a changing chronology. Quaternary Science Reviews 22, 763–768. Vautravers, M. J., Bianchi, G., Shackleton, N. J. (this volume). Subtropical NW Atlantic surface water variability during the last interglacial. In: Sirocko, Claussen, M. F., Litt, T. & Sa´nchez Gon˜i, M. F. (eds.) The climate of past integlacials. Velichko, A. A., Novenko, E. Yu., Zelikson, E. M., Boettger, T., Junge, F. W. (this volume). Comparative analysis of vegetation and climate changes during the Eemian Interglacial in Central and Eastern Europe. In: Sirocko, Claussen, M. F., Litt, T. & Sa´nchez Gon˜i, M. F. (eds.) The climate of past integlacials. Winograd, I. ., Coplen, T. B., Landwehr, J. M., Riggs, A.C., Ludwig, K. R., Szabo, B. J., Kolesar, P. T., Revesz, K. M., 1992. Continuous 500,000-year climate record from vein calcite in Devils-Hole, Nevada. Science 258, 255–260. Woillard, G. M., 1978. Grande Pile Peat Bog: a continuous Pollen Record for the last 140.000 years. Quaternary Research 9, 1–21. Wu¨nnemann, B., Hartmann, K., Altmann, N., Hambach, U., Pachur, H. J. (this volume). Interglacial and glacial fingerprints from lake deposits in the Gobi Desert, NW China. In : Sirocko, Claussen, M. F., Litt, T. & Sa´nchez Gon˜i, M. F. (eds.) The climate of past integlacials. Zagwijn, W. H., 1961. Vegetation, climate and radiocarbon datings in the Late Pleistocene of the Netherlands. Part I: Eemian and Early Weichselian. Mededelingen Geologische Stichting, Nieuwe Serie 14: 15–45.

This page intentionally left blank

14. Abrupt Cooling Events at the Very End of the Last Interglacial Klemens Seelos and Frank Sirocko Institute for Geosciences, Johannes Gutenberg-University, Becherweg 21, 55099 Mainz, Germany

ABSTRACT A comparison of a last interglacial annually laminated and varve counted maar lake record from the Eifel/Germany, with a laminated lake sediment record from Northern Germany, shows that high-resolution cores can be correlated across central Europe by dust/loess content, if the resolution of grain-size data is on the order of decades/centuries. Phases of widespread dust dispersal are the same as the cold events in the Greenland ice and North Atlantic seasurface temperature patterns. The first occurrence of dust in Northern Germany and in the Eifel is during the late Eemian aridity pulse (LEAP, Sirocko et al., 2005) which is called C26 in ocean records (McManus, this volume). This cold and arid event occurred exactly at the time of the last glacial inception at 118 kyr. Vegetation change in Northern Germany and the Eifel is out of phase after the LEAP. A taiga/ tundra vegetation characterizes Northern Germany between the LEAP and C24, whereas at the same time a Carpinusdominated temperate forest spread in the Eifel region, comparable to the Carpinusdominated forests in France (Sa´nchez Gon˜i et al., 2005). A drastic cooling, associated with widespread aridity, came with the C24 cold event, when the vegetation of central Europe changed to a tundra or shrub tundra.

14.1 INTRODUCTION In this paper, we present pollen and dust records from three German lake sediment

cores to quantify the response of the central European vegetation and climate to the changes in the hydrography and seasurface temperatures (SST) of the North Atlantic (Fig. 14.1). Two cores are located in the Eifel region (West Germany), and one core was drilled in SchleswigHolstein/Northern Germany (Fig. 14.2). The stratigraphy of the two Eifel cores is based on 14C and OSL datings (Schaber and Sirocko, 2005) and tuning of the stadial/interstadial succession of the last 140 kyr to the NorthGRIP record (Sirocko et al., 2005). The correlation of the Eifel and North German cores in this paper is, however, based on the eolian dust content of the sediments. Pollen data are completely neglected for this synchronization. We use the dust correlation, because the response of vegetation change to a cooling can be time transgressive, strongly affected by local processes like access to water and regional thresholds in the winter/summer temperatures. Deflation of dust, however, responds today and in the past to seasonal dryness almost immediately over vast areas (e.g. Sirocko and Sarnthein, 1989), and windtransported quartz grains leave welldetectable traces in lake sediments (Seelos and Sirocko, 2005). 14.1.1 NorthGRIP The 18 O record of the NorthGRIP ice core (Fig. 14.3) starts at 122 kyr and documents the temperature evolution from the last interglacial into the last glacial (NorthGRIP Members, 2004). The ice core record shows

208

Klemens Seelos and Frank Sirocko N

a) winter

Denmark Baltic sea

a No r h Se t

RS1

Elb

e

ELSA ine Rh

Germany

HL2, Ei 2 b) summer

France

<2°C

6°–10°C

16°–20°C

2°–6°C

10°–16°C

>20°C

Land

Fig. 14.1 Equal area projection of North Atlantic sea-surface temperatures derived from satellite-borne advanced very high resolution Radiometer: (a) during summer, ( b) during winter of the year 2002, together with location of the Eifel and ELSA cores.

a continuously decreasing temperature trend from 120 to 116 kyr, the latter part of which is called the Stadial S26, or C25 (following the terminology of McManus et al., 1994). Temperature then increases slightly, but very abrupt at 115 kyr, which is the beginning of the DO25 event from 115 to 111 kyr, when the downward temperature trend continues until 111 kyr into the first severe cooling event, named Stadial 25 by the NorthGRIP Members, which is the equivalent of the Cool Event 24 in earlier papers (Johnsen et al., 1992; Grootes et al., 1993). All cold events from today back to the C23 were associated with abundant aeolian dust microparticles (Ruth et al., 2003).

Fig. 14.2 The locations of the West German Eifel cores HL2 (dry maar west of Hoher List), Ei2 (Eigelbach dry maar) and the North German core RS1 (rim depression near Rederstall, Schleswig Holstein).

Respective data for C24 and the late Eemian have not been presented yet. The NorthGRIP dust is derived from the Chinese loess desert and transported by west winds over North America to Greenland. Accordingly, we can expect that the concentrations of dust in all stadial ice core sections document phases of strong continental aridity at times of low air temperature all over the northern hemisphere. 14.1.2 Ice-rafted debris in the North Atlantic All stadials of the last glacial cycle were associated with high amounts of ice-rafted debris (IRD) in the North Atlantic, which apparently stems from surging continental glaciers (McManus et al., 1994; Chapman and Shackleton, 1999). SST were much cooler during the stadials (van Kreveld et al., 2000) all over the North Atlantic down to the Portugese margin (Sa´nchez Gon˜i et al., 1999), which might be caused by the melting

Abrupt Cooling Events at the Very End of the Last Interglacial NorthGRIP – δ18O (NorthGRIP Members, 2004)

209

NorthGRIP – Microparticles (Ruth et al., 2003)

30

30

40

40

50

50

60

60

70

70

80

80

90

90

100

100

110

120

Age (kyr BP)

Age (kyr BP)

ELSA-HL2 Loess Detection

C23 DO24 C24 DO25 C25 LEAP

200 600 1000 [thousand counts / ml]

110

120

(< cold | warm >) 130

130 –46 –42 –38 –34

0

max

Fig. 14.3 Comparison of the 18 O record of the NorthGRIP ice core (NorthGRIP Members, 2004), the NorthGRIP microparticle record (Ruth et al., 2003) and the RADIUS loess detection record of the HL2 core sequence 132–63 kyr BP (sample interval of 500 m).

ice, but also by a reduction of deep-water formation in the North (Rahmstorf, 2002) and thus reduction of the advection of warm subtropical waters into the high latitudes via the Gulf stream and North Atlantic drift (Khodri et al., 2001). 14.1.3 Central European climate Pollen are the most established proxy data of the continental climate change during the last interglacial. A distinct succession of trees was observed from the Netherlands (Zagwijn, 1996), Denmark (Bjo¨rck et al.,

2000), Northern Germany (Menke and Tynni, 1984), West Germany (Mu¨ller, 1974), East Germany (Eissmann, 1990), Poland to South Germany (Mu¨ller et al., 2003) and France (Woillard, 1978). This abundant evidence lead to several synthesis papers on the climate of the Eemian (e.g. Frenzel et al., 1992; Guiot et al., 1993; Aalbersberg and Litt, 1998; Cheddadi et al., 1998; Mu¨ller and Kukla, 2004; Ku¨hl and Litt, this volume). Pollen profiles further to the south still show an ‘Eemian pattern’, but in particular the early and late part of the interglacial appear different to the north European

210

Klemens Seelos and Frank Sirocko

conditions (e.g. de Beaulieu and Reille, 1992; Tzedakis et al., 2001). See also a compilation of numerous last interglacial time series by Mu¨ller and Kukla (2004). Sa´nchez-Gon˜i et al. (2005) first observed a time-transgressive development between the North German response to the late Eemian cooling, compared to a lagged response in central and southern France. The patterns of the tree succession in northern Europe are often generalized, see an example in Aalbersberg and Litt (1998), but already Menke and Tynni (1984) noticed a number of local leads/lags in the first/last occurrence of pollen taxa, when they carefully compared about 20 sites with a succession of trees during the Eemian. These authors first presented the record from Rederstall, which we recorded in 2004 to obtain a parallel pollen and dust record from Northern Germany. We use this record to compare the sediment structures with our ELSA cores from the Eifel. Rederstall, like all other North German sites, shows an interglacial forest with seven pollen zones, beginning with a pioneer vegetation of Betula and Pinus, and leading through a mixed oak forest phase to a final Pinus phase which is then followed by the expansion of grasses. See a typical Eemian pollen profile in the paper by Ku¨hl and Litt (this volume). The duration of the Eemian was first extrapolated from an annual layer counting of about 3500 years at the site of Bispingen in northern Germany (Mu¨ller, 1974, personal communication 2005), an estimation corroborated by Hahne et al. (1994). The duration of the entire temperate forest phase in North German was apparently near to 10 000 years. Dating of the Eemian was done for more than twenty years by tuning the temperate forest phase to the 65N July insolation maximum at 126 kyr (Berger, 1978; Berger, this volume). Absolute dating of Eemian lake sediments was done by luminescence methods, however often with error bars of more than 10 000 years, see Degering and Krbetschek (this volume). Precise dating of

the last interglacial was recently presented by Holzka¨mper et al. (2004, 2005), who applied U-series dating to stalagmites from the Spannagel cave, which is located in the Alps at a height of 2500 m, i.e. on the modern snow line. Stalagmites can only grow at this height, if climatic conditions are similar to those of today (see also Spo¨tl et al., this volume). The Spannagel record dates the very first interglacial conditions at 135 kyr, followed by a climatic deterioration and the beginning of the Eemian at 128 kyr (Holzka¨mper et al., 2004, 2005). Stalagmites terminated to grow at 118 kyr; thus, they represent an interglacial of 10 000 years, however, with at least one phase with no stalagmite growth. The duration of the interglacial at Spannagel is similar to Mu¨ller’s varve counts. The major difference at Spannagel is, however, the early growth phase from 135 to 132 kyr, which was never observed in the pollen records of Northern German, but is documented in pollen records from Southern Germany (Zeifen) and marine 18 O records from the North Sea (Seidenkrantz et al., 1996) and of Portugal (Sa´nchez Gon˜i et al., 2005). 14.2 LITHOLOGY AND STRATIGRAPHY OF SEDIMENT CORES Twenty-two long sediment cores have been drilled by the ELSA project (Eifel Laminated Sediment Archive) between 2001 and 2005 in 16 Eifel dry maar lakes. The cores are up to 155 m long and document the entire time from the Holocene to 140 kyr. Four cores reach the last interglacial, but only core HL2 (dry maar west of Hoher List) shows annually laminated sediments in the last interglacial section (Schaber and Sirocko, 2005; Sirocko et al., 2005; Rein et al., this volume). Varves have been counted from 10-cm long thin sections of 30 mm thickness. The same thin sections are also used for grainsize analysis (see below). The stratigraphy

Abrupt Cooling Events at the Very End of the Last Interglacial

14.2.1 Core ELSA-HL2 The ELSA-HL2 core (dry maar west of Hoher List) is characterized by three long event layers (Fig. 14.4). The section 57.4–56.0 m shows two slumps which are separated by a laminated sequence of about 1000 years. The second event is a block slide, which starts at a depth of 54.4 m. The third event layer is a strongly disturbed sediment sequence from 51.05 to 50.4 m. The sequence shows typical seismites, which must have been caused by strong earthquakes. The remaining sections of the core HL2 comprise undisturbed, completely fine laminated sediments (see figures of varves in the paper by Number of particles (< 63 µm)

56

Number of particles (< 63 µm)

Slump

52

C24

57

Block slide

Slump

53

HL2 core sequence 60–50 m Number of particles (<63 µm)

58

54

50

Seismit

C25

55

51

59

52 0 200

600 [abs.]

56 1000

Depth (m)

C24

Depth (m)

LEAP

Depth [m]

of these cores is based on 66 AMS 14C dates (Schaber and Sirocko, 2005) and eight luminescence dates (Degering and Krbetschek, this volume). The final stratigraphy, however, comes from tuning the greyscale curves to the succession of DO events (Sirocko et al., 2005). Additional stratigraphic information comes from the occurrence of the phonolitic Du¨mpelmaar tephra, which was Ar/Ar dated (van den Bogaard et al., 1989) to 116 000 þ 16 000 yr BP and is now tuned to 105 kyr BP very shortly before the beginning of DO23 (Sirocko et al., 2005). The fine tuning of the stratigraphy is based on the greyscale record, but includes also significant lithological changes of grain size, organic carbon content, lamination thickness and occurrence of light diatom/calcite layers. The similarity between ELSA greyscale and NorthGRIP 18 O is very high during the last termination, the early MIS3, MIS4 and down to the C24 event. Similarity is, however, low for the Holocene and last interglacial, because the ELSA interglacial laminae consist of both dark organic carbon-rich layers, but also of light calcite and diatom layers. The stratigraphy of the last interglacial is not done by tuning, but strictly by varve counting. The entire varve counted section reaches from 106 kyr BP down to 132 kyr BP. The connection of the floating ELSA varve chronology to the NorthGRIP record is done at the beginning of C24, which is most clearly visible in both records (see Fig. 2 of Sirocko, et al. 2005). The varve counted stratigraphy is continuous from C23 down to 122 kyr BP, where we find a massive slump. All varve counts below the slump are then floating again. The link of these early Eemian floating varve chronologies to an absolute timescale will be done by the analysis of stalagmite growth periods in the Alpine Spannagel cave (Holzka¨mper and Sirocko, in preparation). This cave is at the height of the modern snow line, and stalagmites can only grow if temperature conditions were similar to those of today.

211

0 200

600 [abs.]

60 1000

0 200

600 [abs.]

1000

Fig. 14.4 Core photos and the ‘number of quartz particles <63 mm (100 m sample interval)’ record of the HL2 sequence 60–50 m (dry maar west of Hoher List, Eifel).

212

Klemens Seelos and Frank Sirocko

Rein et al., this volume). Sedimentation rates during the last interglacial are very low ð80300 mm=yearÞ. The grain-size analysis shows a mean value around 30 mm, and the number of quartz particles is seldom higher than 100 per 100 mm sample segment (Fig. 14.4). 14.2.2 Core ELSA-Ei2 The depth interval from 38 to 48 m in ELSA core Ei2 (dry maar of Eigelbach, Fig. 14.5) is characterized by similar sediment structures as HL2. Again, we observe several sections of increased quartz content. We observed slumps at the respective depth

38

Number of particles (<63 µm)

42

Number of particles (<63 µm)

46

Number of particles (<63 µm)

LEAP

LEAP

43

44

40

Slump

48 0 200

600

1000

[abs.]

Ei2 core sequence 48–38 m

45

Depth (m)

41

Depth (m)

Depth (m)

C24

47

cool event C24 110.5 - 108.5 kyrs BP

39

C25

42

46 0 200

600 [abs.]

1000

0 200

600 [abs.]

1000

Fig. 14.5 Core photos and the ‘number of quartz particles <63 m (100 m sample interval)’ record of the Ei2 sequence 48–38 m (Eigelbach dry maar, Eifel).

intervals, but the seismites of core HL2 are not represented. Apparently, core Ei2 is from a location with a slope large enough to cause slumping during strong earthquakes. There are, however, also fundamental differences between the particle analysis results of the two ELSA cores. In opposition to the HL2 core, the mean size of quartz particles in Ei2 is larger than 40 mm. In addition, sedimentation of organics is significant higher. This is most likely caused by the size of the maar and the lithology of the surrounding area. 14.2.3 Core RS1 from the site of Rederstall, Northern Germany The drilling site of Rederstall is located in the rim depression of the large Oldensworther salt wall in the western part of Schleswig Holstein (Fig. 14.2). The last interglacial sediments were known to be at this location at 35–44.5 m depth (Fig. 14.6), because this site was palynological analysed by Menke and Tynni (1984) already twenty years ago. The diatom assemblage reflects brackish water conditions during the early Eemian; thus salt water from the North Sea penetrated into the rim depression during the Eemian sea-level highstand. The decrease of sealevel after 118 kyr (Lambeck and Chappell, 2001) must have affected this site strongly, which was apparently very near to the shore during the Eemian, but propably several kilometres away at 110 kyr. The nonexistence of turbidites and block slides in this record points to a relatively shallow basin with gently inclined slopes. The sediments of the whole core sequence are undisturbed. Some sections show lamination, but these sediments are not characterized by seasonal varves (Menke and Tynni, 1984). Like in the Eifel region, however, mean particle sizes are around 40 mm, and we detect dust intervals with a coarsening and better sorting of the quartz fraction (Fig. 14.6). Sedimentation rates are

Abrupt Cooling Events at the Very End of the Last Interglacial

35

Number of particles (<63 µm)

39

Number of particles (<63 µm)

43

Number of particles (<63 µm)

C24

40

44 Depth (m)

36

0

200

600 [abs.]

1000

RS1 core sequence 44.5–35 m 37

41

C25

42

Depth (m)

Depth (m)

38

39

43 1000

0 200

600 [abs.]

LEAP 0 200

600 [abs.]

1000

Fig. 14.6 Core photos and the ‘number of quartz particles < 63 m (100 m sample interval)’ record of the RS1 sequence 44.5–35 m (Rederstall, Schleswig-Holstein).

similar to the respective values of the maar lake records.

14.3 METHODS 14.3.1 Grain-size analysis of eolian dust The particle analysis method RADIUS (rapid particle-size analysis of digital images by ultra-high-resolution scanning of thin sections; Seelos and Sirocko, 2005) allows the detection of different coloured particles in terms of size and shape on digital images of thin sections. Sediments like

213

dust layers, organics, turbidites and tephra are identified by a pattern-recognition module included in RADIUS, which is a combination of different statistical routines (i.e. sorting parameters, modes and particle-size gradients) (Seelos and Sirocko, 2005). Clastic sediment structures are characterized by specific particle-size distributions, which are associated with transport and deposition processes (McTainsh et al., 1997; Stuut et al., 2002). The best-known examples are turbidites, which are normally graded. Mafic pyroclastic sediments are characterized by a large number of dark particles beneath bright schist and quartz fragments under polarized light. Another sediment type that must be distinguished from loess is a quartz-rich deposit from phreatomagmatic maar explosions that have fragmented the autochthonous Devonian rocks to small pieces of silt grain size. This type of sediment is observed in the HL2 core only in the section immediately overlaying the tuff from the initial maar explosion. These sediment structures are visually similar to loess (Clapperton, 1993; Iriondo, 1997, 1999), but they are separated by the RADIUS loess-detection algorithm, because these volcanogenic airborne sediments show a distinct grading and large amounts of debris from the authigenic rocks. The mean size of these grains, however, is indeed similar to loess. Typical dust/loess sequences are represented by mean grain sizes around 40 mm and high packaging densities of 600–1000 quartz particles per 100 mm sample segment (Table 14.1). The dust sediments are well sorted and show no grading at all (Seelos and Sirocko, 2005). To document the nature of the three cores used in this study, we show core photos and the ‘number of particles > 63 mm’ record against depth in Fig. 14.4–14.6. 14.3.2 Pollen Pollen preparation followed the techniques of Berglund and Ralska-Jasiewiczowa (1986) respectively Faegri and Iverson

214

Klemens Seelos and Frank Sirocko

Table 14.1 Numerical setting of the RADIUS loess detection algorithms (Seelos and Sirocko, 2005) Loess sequences do not show significant grain-size gradients The mean grain sizes are equal over the whole loess sequence The maximum grain sizes are equal over the whole loess sequence The percentage of dark particles (organics, ashes) is very low The distribution width of grain-size classes is low for loesses The values of the sorting parameters are small for loess sequences The distribution of loess sequences is always monomodal – the modes are small The symmetry of loess distribution histograms is high – values for the skewness are small

(1989), modified after T. Litt (University of Bonn, Germany). About 1 cm3 of sediment was treated with potassium hydroxide solution (KOH), hydrochloric acid (HCl) and hydrofluoric acid (HF). Acetoloysis, using acetic acid (C2H4O2) and a mixture of acetic anhydride (C4H6O3) and sulphuric acid (H2SO4), was optional. The samples were archived with liquid, anhydrous glycerol (C3H8O3). Centrifugation was done at 3000–3500 r.p.m. for 5 minutes. The samples were sieved through a 200-mm sifter and later filtrated through a 10-mm filter. Lycopodium-spore tablets (Department of Geology, Quaternary Sciences, So¨lvegatan 12, SE-223 62 LUND, Sweden) were added for calibration of absolute pollen content. Pollen counting was done under five-hundred-fold magnification. The pollen spectra are based on the sum of 100% of all terrestrial pollen (TP), which were divided into arboreal pollen (AP) and nonarboreal terrestrial pollen (NATP). The former contain all trees including hazel, ivy, holly and mistletoe. Cyperaceae was treated as terrestrial pollen. The pollen for HL2 and Ei2 (Figs. 14.8, 14.9) were counted by Frank Dreher (Mainz) and were published in the paper by Sirocko et al. (2005). The pollen data for Rederstall are taken from the original publication by Menke and Tynni (1984). Our new core was drilled 10 m away from the original coring site of Menke. Both Rederstall cores show an event lamination and five lithological

0 þ=0:05 40 þ=15 mm 100 þ=15 mm < 5% < 3% 7–9 classes 1.2–2.2 4063 mm

5 5 1 1 2 1 2 1

0 þ=0:1 0=0:05

1 2

marker layers that we used to correlate the old and new core. In addition, we counted several pollen samples of our core to precisely correlate the older Rederstall data with the new grain-size records. There is a depth offset of 30 cm between the core of Menke and Tynni (1984) and the RS1. 14.4 SYNTHESIS OF DUST AND POLLEN RECORDS The RADIUS loess detection record shows grain-size patterns, which are very similar in all three cores (Fig. 14.7). The last interglacial between 132 and 118 kyr BP shows no loess activity, but several short intervals of dust deposition. These short events could have a similar origin as the LEAP, but this has not been studied in detail yet, because of the open question on the stratigraphy of the middle and early parts of the Eemian (see above). The first pulse of high dust rates was observed in HL2 at 118 kyr during an event that Sirocko et al. (2005) called LEAP (late Eemian aridity pulse). This event is dated to 118 kyr BP by varve counting of the HL2 core sequence and represents a first cooling/aridification exactly at the time of the last glacial inception. The varve counted age matches the age of first drop of global sea level as reconstructed from marine terraces (Lambeck and Chappell, 2001), it also matches the time of last stalagmite growth

Abrupt Cooling Events at the Very End of the Last Interglacial

Quartz (20–63 µm) (% of total)

δ18O – NorthGRIP (NorthGRIP Members, 2004) Loess content (%) Loess content (%) Quartz (RADIUS (20–63 µm) (RADIUS detection) (% of total) detection)

West of Hoher List

Eigelbach

Loess content (%) Quartz (RADIUS (20–63 µm) detection) (% of total)

215

Rederstall

100 102

Dümpelmaar Tephra

104

C23 106 108

C24

110

Age (varve kyr BP)

112 114 116

C25

118

LEAP

120 122 124 126 128 130 132 25

75

25 50 75 100

25

75

25 50 75 100

25

75

25 50 75 100

Fig. 14.7 ‘RADIUS dust/loess detection’ and ‘Quartz <63 m (% of total)’ of the three core sequences Ei2, HL2 and RS1 against time. The Du¨mpelmaar Tephra (116 þ= 16 kyr BP, van den Bogaard et al, 1989) is detectable in the HL2 and the Ei2 core.

in the Spannagel cave (Holzka¨mper et al., 2004, 2005). The LEAP was interpreted by Sirocko et al. (2005) to be caused by southward displacement of the North Atlantic drift, i.e. a severe cooling of the North Atlantic temperatures during the initial growth of ice. If the signal is indeed controlled by the ocean hydrography, we can expect to observe a synchronous signal all over Central and Northern Europe. Accordingly, we use this first occurrence of post-Eemian dust/loess to synchronize the cores. Starting from the first dust occurrence, all other loess phases line up nicely in all three cores and

we attributed them to the C events C26, C25, C24 and C23 (Fig. 14.7 ), because they almost exactly match the time series of the NorthGRIP stadials (NorthGRIP) and North Atlantic cold events (McManus et al., 1994; Chapman and Shackleton, 1999). Figures 14.8–14.10 show the pollen record of the cores HL2, Ei2 and RS1 for the last interglacial plotted on the HL2 chronology. Both Ei2 and RS1 were tuned at the marker points shown in Table 14.2. In the following, we will compare the climate evolution during the last interglacial in the Eifel and in Northern Germany.

216

Klemens Seelos and Frank Sirocko Core HL2 (Hohe List, Eifel) 5

80

70

40

20 10

5

5

40

10 5

50

5

C25

lm

ry s/oa lus k Al /ha nu ze Ti l lia s/al Ca /lim de rp e r inu tre s/h e or nb ea m Er A ica ce bies ae / /(h fir ea Po th ac ) ea e/ gr as Cy se pe s ra ce a/ se dg es

cu er

Co

ce

us /e

Qu

Ul m

ru sp a/ ce Pi

Pi nu s/p

h irc tu

la

/b

illo

ine

LEAP

Be

/w lix Sa

10

C24

w

Age (kyr BP)

100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140

Fig. 14.8 Pollen content in core HL2 (counted by Frank Dreher, Mainz), plotted versus varve counted age.

14.4.1 C26, LEAP The first indication of late Eemian climate change is the first occurrence of dust at 118 kyr, which lasted 450 years. During this period, we have distinguished 52 individual dust layers (Sirocko et al., 2005). The sediment section of this event in core HL2 shows two distinct charcoal layers, which indicate bushfires and strong aridity. Sirocko et al. (2005) called this interval of cold and arid conditions the LEAP. The mean sedimentation rate during the LEAP is about 650 mm=yr in core HL2, thus six times higher than before and after the event. The pollen content reveals a dominance of grass

vegetation at the expense of the tree taxa ( Fig. 14.8 ). The organic fraction disappeared at the beginning of these varved sediments within a few years, reflecting the extremely abrupt and severe nature of this event. Tree pollen are not completely absent, but we do not know how much of these pollen grains were reworked when lake level could have lowered in response to the aridity. Lowering, however, could not be very strong, because we do not observe any large slumping during this interval. The lake level of a maar represents either the groundwater table or is determined by an overflow outlet; most probable, the lake level has not changed at all during the

Abrupt Cooling Events at the Very End of the Last Interglacial

217

Core Ei2 (Eigelbach, Eifel) 5

50

70

50

20 10

5

5

30

5 5

40

5

C25

s lus /oak / Al ha z n Ti us/ el lia ald /lim e Ca e r rp tre inu e s/h or nb ea Er m ica Ab ce ies ae /(h /fir ea Po th ac ) ea e/ gr as Cy se pe s ra ce a/ se dg es

lm

ry

cu

Co

er Qu

Ul

m

us

/e

ce ru sp Pi ce a/

h irc tu

la

/b

illo

Pi nu s/p ine

LEAP

Be

/w lix Sa

5

C24

w

Age (kyr BP)

100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140

Fig. 14.9 Pollen content in core Ei2 (counted by Frank Dreher, Mainz), plotted versus varve counted age.

LEAP, indicating that the aridity was only seasonal, most probably in the late summer. Thirty per cent of grass pollen however indicates a vegetation near to a shrub tundra. It is likely that the high elevations of the Eifel (400–500 m above modern sea level) were almost deforested during this event, but the trees survived in refuge areas nearby, like the local creek valleys and in particular the very warm Rhine and Mosel gorge (only 100 m above modern sea level), where water supply is plentiful and summer temperatures today are several degrees higher than the temperatures on the high

Eifel, at a distance approximately 20 km from the Eifel maar lakes. The LEAP (C26) is also clearly identified in the dust records of core Ei2 and the North German Rederstall core RS1 (Fig. 14.7). The pollen sample resolution of core HL2 is sufficient to show the event nicely, the pollen sample spacing in core Ei2 is too coarse. The lithology of core Ei2 (Fig. 14.9) shows similar sediment structures as the HL2 sequence during this time (Fig. 14.7), but the sedimentation rates are higher and reach 3700 mm=yr, indicating deflation areas near by the Eigelbach site. A total of 48 dust layers are

218

Klemens Seelos and Frank Sirocko Core RS1 (Rederstall, Northern Germany) 30

90

50

5

30

50

20

10

30

tre e

or

Er

ica

s/h

Ca

rp

inu

lia Ti

nb ea Ab m ce i ae es/f /(h ir Po ea ac th ea ) e/ gr Cy a ss pe es ra ce a/ se dg es

e

r lde /lim

el az

nu Al

lus /h ry Co

s/o cu

s/a

ak

lm /e

er

us

Qu

ru

ce m Ul

sp Pi ce a/

s/p nu Pi

la

/b

illo

irc

h

ine

LEAP

tu

/w

40 10

C25

Be

lix Sa

10 5

C24

w

Age [kyr BP]

5 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 128 130 131 132 133 134 135 136 137 138 139 140

Fig. 14.10 Pollen content for Rederstall: taken from the original publication by Menke and Tynni (1984). The depth offset between the core of Menke and Tynni and the RS1 is 30 cm. The data of the new core were shifted by 30 cm to match the old pollen counts. This aligns the major boundaries in the lithologies but allows slight offsets.

counted in Ei2 during the LEAP, which is only four less than in the HL2 section (see above). The LEAP started in the north German core from Rederstall with a thick dust layer. The counting of single dust events in the Rederstall core is impossible, because the sequence is not laminated and individual events are not detectable, only a general enrichment of eolian particles. The LEAP is not an event in Rederstall but represents the end of the Eemian. In particular, Carpinus and Abies disappear and do not return again. Plotting the original pollen counts by Menke and Tynni (1984) on our corrected timescale from Fig. 14.7, we observe a spread

of grasses in a forest dominated by Pinus and Betula, representing most probably a taiga vegetation (Fig. 14.10). The site of Gro¨bern in eastern Germany shows a tundra vegetation at that time (Ku¨hl and Litt, this volume), which is different to Rederstall, but even today differences of several degrees characterize even today the winter temperatures between these two locations in northwest and southeastern Germany. Whether the traces of Ulmus and Quercus pollen at Rederstall represent real growth of these taxa during the event, or whether they are just reworked, cannot be decided yet. Alnus, however, clearly continued to grow during the event and at least for another 8000 years.

Abrupt Cooling Events at the Very End of the Last Interglacial

219

Table 14.2 Marker points for the tuning of cores Ei2 and RS1 Depth

Age

Source

Marker points for tuning

Ei2 34.49 34.75 35.05 38.20 38.73 41.15 41.50 42.50 44.95 46.25 47.27 51.10 53.50 57.50 61.00 61.76 62.50

104712 105050 105200 106100 108650 111000 115380 115500 118100 118600 122000 125795 126928 130457 132070 132116 132188

HL2 NorthGrip NorthGrip NorthGrip NorthGrip NorthGrip HL2 NorthGrip HL2 HL2 HL2 HL2 HL2 HL2 HL2 HL2 HL2

Tephra DMT Loess peak Loess peak End of DO24 (loess marker) Start of DO24 (loess marker) End of DO25 (loess marker) Carpinus peak Start of DO25 (loess marker) Loess peak Loess peak Top of a slide/tephra, limy Betula peak Corylus decrease Ulmus decrease NATP pollen decrease Tephra layers Start of warm phase – decrease of loess

RS1 36.0 36.4 37.4 39.1 41.0 43.0

110500 112000 116000 118000 122000 126000

HL2 HL2 HL2 HL2 HL2 HL2

Start of C24 – loess increase Loess peak Loess peak Start of the LEAP – loess increase Abies increase Start of Eemian (sensu stricto) – loess decrease

The Eifel grasslands of the LEAP were apparently reforested with a pioneer vegetation of Betula and Pinus during an interval of 30 years at the end of the LEAP (Fig. 14.8). All other trees reappear immediately after this transition phase, which allows us to infer that the refuge areas of the temperate tree taxa during the LEAP were not very far away (see above). Marine records from the North Atlantic show a cold event immediately at the beginning of the global sea-level drop at 118 kyr (Chapman and Shackleton, 1999; Sa´nchez Gon˜i et al., 1999; Lehman et al., 2002). This cold event is characterized by an SST reduction of 2–3 C, but it is not associated with IRD at the NEAP 18 K site in the centre of the North Atlantic. Accordingly, it shows not up in the IRD synthesis record of McManus (this volume), but it is obvious in the

temperature record of Chapman and Shackleton (1999) who named the event C26. As this event occurred at the very beginning of early Weichselian ice-sheet growth, it cannot be associated with icesheet surges, which explains why the IRD is missing. Mu¨ller and Kukla (2004) and Sa´nchez Gon˜i et al. (2005) were the first who compared this event with European pollen profiles. Sirocko et al. (2005) called this event LEAP, instead of C26, because this would indicate a causal mechanism like for the later C events, which are clearly associated with surging continental glaciers. The mechanisms that caused the LEAP are not well understood yet. All the above authors attributed it to a southward displacement of the North Atlantic drift. It is, however, also visible in low latitudes, where Lehman et al. (2002) observed it clearly in the sediments on

220

Klemens Seelos and Frank Sirocko

the Bermuda Rise. The cause for the displacement is thus still an open question. Modelling experiments by Khodri et al. (2001), Berger and Loutre (2002), Calov et al. (2005), Kubatzki (this volume), Kaspar et al. (this volume) highlight the role of insolation change to be the principal cause of the last glacial inception. Khodri

55.0

et al. (2001) observed a change in the SST fields of the North Atlantic in response to the insolation change. The evidence of a combined aridity and temperature change in central Europe during the LEAP would indeed perfectly match this scenario. The response of the continental climate to the LEAP is, however, regionally different.

Thin-section images

Event detection Loess Corg Tephra

2 C26/LEAP Core HL2 (Eifel) 121.8–116.2 kyr BP

55.1

55.2

55.3

40 mm

LEAP 55.4 2 55.4

55.5

1

1 55.6

2000 µm

55.7

20 mm

55.8

55.9

2000 µm

56.0

Fig. 14.11 C26/LEAP: Core photo of the HL2 core sequence 56–55 m (121.8–116.2 kyrs BP); RADIUS event detection for dust/loess, Corg and tephra in 500 m sample resolution; polarized thin-section images.

Abrupt Cooling Events at the Very End of the Last Interglacial Event detection

Loess 46.0

Corg

221

Thin-section images

Tephra

C26/LEAP

2

Core Ei2 (Eifel) 121.6–118.2 kyr BP

46.1

LEAP

46.2

20 mm

2 46.3

46.4

2000 µm

46.5

1

46.6

46.7

20 mm

46.8

1

46.9

2000 µm

47.0

Fig. 14.12 C26/ LEAP: Core photo of the Ei2 core sequence 47–46 m (121.6–118.2 kyr BP); RADIUS event detection for dust/loess, Corg and tephra in 500 m sample resolution; polarized thin-section images.

Based on the correlation of the dust and loess events (Fig. 14.7), we see the following picture emerging: dust from local deflation of grassland areas (however no continent wide loess yet) occurred both in Northern Germany and in the Eifel during the LEAP. In comparison with the Eifel, where thermophilous trees continued to

grow after the LEAP, the temperate trees disappear in Northern Germany with the LEAP and do not return. 14.4.2 C25 The first IRD event, C25, is called Stadial 26 by the NorthGRIP members. This slight cold

222

Klemens Seelos and Frank Sirocko Event detection

Loess 39.0

Thin-section images

Corg 2 C26/LEAP LEAP

39.1

2

Core Ei2 (Northern Germany) 119.9–117.9 kyr BP

39.2

20 mm

39.3

1

39.4

2000 µm

39.5

1

39.6

39.7

20 mm

39.8

39.9

2000 µm

40.0

Fig. 14.13 C26 / LEAP: Core photo of the RS1 core sequence 40–39 m (119.9–117.9 kyr BP); RADIUS event detection for dust / loess and Corg in 500 m sample resolution; polarized thin-section images.

event at 116 kyr is associated with the first traces of IRD in the North Atlantic (McManus et al., 1994; McManus, this volume), reflecting that the continental glaciers have reached a size since the inception at 118 kyr, which was sufficient to make a first surge.

Pinus became more and more widespread during the C25 event in Northern Germany (Fig. 14.10), but at the same time Carpinus dominated the vegetation in the Eifel again ( Figs. 14.8, 14.9). The distance between Northern Germany and the Eifel is 500 km, which

Abrupt Cooling Events at the Very End of the Last Interglacial

could well explain the different climatic conditions. Loess detection in HL2 and Ei2 during the cool event C25 is significant (Fig. 14.7), but quite low. Loess detection for the RE1 core during the same period is explicitly higher at 80–90% loess per 500 mm sample. This pattern, i.e. a Pinus-dominated taiga forest in Northern Germany, and a Carpinus-dominated temperate forest in the Eifel, is typical of C25 and continued during the early part of DO25. Apparently, the Eifel was at this time climatically much closer connected to France than to north, east or south of Germany.

14.4.3 DO25 The vegetation during DO25 is very similar to C25 and the gradient in the vegetation zones remained during the early part of this warm phase. The later part of DO25 is, however, characterized by a continuous decline of Carpinus and all other temperate forest taxa in the Eifel, which began at a time, when Betula dominated in Northern Germany over Pinus. It appears only during the late DO25 that the amplitude for dust/ loess in the HL2 and the RS1 core reach clearly the level of loess. The grain-size data indicate a well correlatable (Fig. 14.7) dust phase immediately before the severe C24 cooling.

14.4.4 C24 Sedimentation pattern in HL2 change abruptly at a depth of 52.9 m; the loess detection increases within several millimetres of sediment up to 70–80%. The content of organic components decreases during the same time and reaches values near zero. This transition occurred within a few years. The silty sediments of the section 51.25–52.9 m are laminated, and every single loess storm is detectable, however not documented here in detail. The second ELSA core Ei2 shows similar sediment

223

structures (Fig. 14.14). At a depth of 41.1 m, the first single loess storms are detected. In analogy to the HL2 core, the number of loess events increases during the cool event C24. On the whole, the dust/loess layers of the Ei2 core are coarser than the corresponding HL2 sediments. The RS1 record differs to some extent: the loess contents are also high and very stable (around 90%) during the C24, but the transition from D25 to C24 is smoother. The Rederstall core sequence RS1 shows some differences compared to the Eifel records. The organic-rich sediments at the bottom of the core (35.5–36.5 m) are not laminated. The percentage of clastic components in the warm D25 sediments is higher than in the corresponding Eifel core sequences. The grass pollen strongly increase both in the Eifel and in Northern Germany strongly with the onset of C24 to 40–50% of all pollen; only the pollen of Betula and Pinus are abundant and we have to infer a shrub tundra vegetation for this time. The DO25 climate/vegetation gradient across central Europe apparently disappeared, and all of northern Europe was covered by a tundra at a time when the North Atlantic cooled strongly, and icebergs reached even the north of the Iberian Peninsula (Sa´nchez Gon˜i et al., 2005). The details of this pattern were reconstructed from three cores on the basis of a correlation of aeolian dust layers. The inferences on the length of the last interglacial corroborates both the points of Turner (2002), who opted for a short interglacial, but it also corroborated the point of Kukla et al. (2002), who had envisioned the pattern of a longer duration than 10 000 years first (Kukla et al., 1997; Kukla, 2000). We will stop our evaluation of the climate evolution during the last glacial inception at this point. The Eifel maar records cover the entire last glacial cycle (see Fig. 14.2 in Sirocko et al., 2005). We are currently doing the grain-size and pollen analysis for the entire section, but this will be presented later.

224

Klemens Seelos and Frank Sirocko Thin-section images

Event detection Loess

Corg

Tephra

54.0

C24

2

Core HL2 (Eifel) 116.2–111.6 kyr BP

54.1

13 mm

54.2

C24

54.3

54.4 2000 µm

2

54.5

1

54.6

1 54.7

20 mm

54.8

54.9

2000 µm

55.0

Fig. 14.14 C24: Core photo of the HL2 core sequence 55–54 m (116.2–111.6 kyr BP); RADIUS event detection for dust/ loess, Corg and tephra in 500 m sample resolution; polarized thin-section images.

14.5 CONCLUSIONS (1) Sediment cores across central Europe can be correlated by dust/loess content, if the resolution of the grainsize data is on the order of centuries.

The phases of dust dispersal are the same as the cold events in the Greenland ice and North Atlantic SST patterns. (2) The first major occurrence of dust is during the LEAP (C26), i.e. the period

Abrupt Cooling Events at the Very End of the Last Interglacial Event detection Loess 40.5

Corg

225

Thin-section images

Tephra

2

C24 Core Ei2 (Eifel) 113.7–110.3 kyr BP

40.6

40.7

30 mm

C24 40.8

40.9

1

41.0 2000 µm

2

40 mm

1

41.1

41.2

41.3

41.4

2000 µm

41.5

Fig. 14.15 C24: Core photo of the Ei2 core sequence 41.5–40.5 m (113.7–110.3 kyr BP); RADIUS event detection for dust/ loess, Corg and tephra in 500 m sample resolution; polarized thin-section images.

of the last glacial inception at 118 kyr. These dust layers are, however, fine grained and have some clay content, thus, not comparable to the real loesses, which have no clay content,

and became important with the C24 cold event. Ocean records demonstrate that the LEAP is clearly associated with changes in the low- and highlatitude Atlantic SST, but without any

226

Klemens Seelos and Frank Sirocko Event detection Loess 35.5

Thin-section images

Corg 2

C24 Core RS1 (Northern Germany) 112.4–109.2 kyrs

35.6

35.7

C24 30 mm

35.8

35.9

1

36.0 2000 µm

2

36.1

1 40 mm

36.2

36.3

36.4

36.5

2000 µm

Fig. 14.16 C24: Core photo of the RS1 core sequence 36.5–35.5 m (112.4–109.2 kyr BP); RADIUS event detection for dust/ loess, Corg and tephra in 500 m sample resolution; polarized thin-section images.

IRD deposition, because the glaciers did not exist at this time, but just started to grow. (3) The afforestation in the Eifel after the LEAP came from nearby refuge areas,

most probably the warm Rhine and Mosel valleys. (4) The vegetation change in North Germany and the Eifel is out of phase after the LEAP. A taiga/shrub tundra/

Abrupt Cooling Events at the Very End of the Last Interglacial

tundra vegetation characterizes Northern Germany between the LEAP and C24, whereas at the same time a Carpinus-dominated temperate forest spread in the Eifel region, but already deteriorated before the C24 cool event. (5) A drastic cooling, associated with widespread aridity, came with the C24 cold event, when the vegetation all over central Europe changed to a tundra/shrub tundra. (6) The forcing of the arid/cold event LEAP is most likely related to a threshold in insolation, which affects the hydrography of the North Atlantic. Other scenarios are possible and could invoke primary changes in the intensity of the solar irradiation (similar to the forcing of the Little Ice Age) or instabilities of a Greenland ice sheet which was smaller and instable during the last interglacial. ACKNOWLEDGEMENTS We thank the Johannes Gutenberg-University of Mainz, and the German Ministry for Research and Education for the very generous funding of this study over the last five years.

REFERENCES Aalbersberg, G., and Litt, T., 1998. Multi-proxi climate reconstructions for the Eemian and Early Weichselian. Journal of Quaternary Science 13, 367–390. Berger, A. L., 1978. Long term variations of caloric solar radiation resulting from the earth’s orbital elements. Quaternary Research 9, 139–167. Berger, A., and Loutre, M. F., 2002. An exceptionally long interglacial ahead? Science 297, 1287–1288. Berglund, B. E., and Ralska-Jasiewiczowa. 1986. In: Berklund, B. E. (Ed.) Handbook of Holocene Palaeoecology and Palaeohydrology, Wiley, Chichester, 455–484. Bjo¨rck, S., Noe-Nygaard, N., Wolin, J., HoumarkNielsen, M., Hansen, H. J., and Snowball, I., 2000. Eemian lake development, hydrology and climate: a multi-stratigraphic study of the

227

Hollerup site in Denmark. Quaternary Science Reviews 19, 509–536. Calov, R., Ganopolski, A., Claussen, M., Petukhov, V., and Greve, R., 2005. Transient simulation of the last glacial inception. Part I: Glacial, inception as a bifurcation in the climate system. Climate Dynamics, 24 (6), 545–561, doi: 10.1007/s00382-005-0007-6. Chapman, M. R., and Shackleton, N. J., 1999. Global ice-volume fluctuations, North Atlantic ice-rafting events, and deep-ocean circulation changes between 130 and 70 kyr. Geology 27, 795–798. Cheddadi, R., Mamakowa, K., Guiot, J., de Beaulieu, J. L., Reille, M., Andrieu, V., Granoszewski, W., and Peyron, O., 1998. Was the climate the Eemian stable? A quantitative climate reconstruction from seven European pollen records. Palaeogeography, Palaeoclimatology, Palaeoecology. 143; 1–3, 73–85. Clapperton, C., 1993. Quaternary Geology and Geomorphology of South America, Amsterdam, Elsevier. de Beaulieu, J. L., and Reille, M., 1992. Long Pleistocene pollen sequences from the Velay Plateau (Massif Central, France). Vegetation History Archaeobotany 1, 233–242. Eissmann, L., 1990. Das mitteleuropa¨ische Umfeld der Eemvorkommen des Saale-Elbe-Gebietes und Schlußfolgerungen zur Stratigraphie des ju¨ngeren Quarta¨rs. Altenburger naturwissenschaftliche Forschung 5, 11–48. Faegri, K., and Iversen, J., 1989. Textbook of Pollen Analysis, Wiley, New York. Frenzel, B., Pecsi, M., and Velichko, A. A., 1992. Atlas of Paleoclimates and Paleoenvironments of the Northern Hemisphere. Grootes, P. M., Stuiver, M., White, J. W. C., Johnsen, S., and Jouzel, J., 1993. Comparison of oxygen isotope records from the GISP2 and GRIP Greenland ice cores. Nature 366, 552–554. Guiot, J., de Beaulieu, J. L., Cheddadi, R., David, F., Ponel, P., and Reille, M., 1993. The climate in Western Europe during the last glacial/interglacial cycle derived from pollen and insect remains. Palaeogeography, Palaeoclimatology, Palaeoecology 103, 73–93. Hahne, J., Kemle, S., Merkt, J., and Meyer, K. D., 1994. Eem-, Weichsel- und Saalezeitliche Ablagerungen in der Bohrung ‘‘Quakenbru¨ck GE2’’. Geologische Jahrbuch A 134, 9–69. Holzka¨mper, S., Mangini, A., Spo¨tl, C., Mudelsee, M., 2004. Timing and progression of the Last Interglacial derived from a high alpine stalagmite. Geophysical Research Letters 31, L07201, doi:10.1029/ 2003GL019112. Holzka¨mper, S., Spo¨tl, C., and Mangini, A., 2005. High-precision constraints on timing of Alpine warm periods during the middle to late Pleistocene using speleothem growth periods. Earth and Planetary Science Letters 236 (3–4), 751–764.

228

Klemens Seelos and Frank Sirocko

Iriondo, M., 1997. Models of Aeolian silt deposition in the Upper Quaternary of South America. Journal for South American Earth Sciences 10 (1), 71–79. Iriondo, M., 1999. The origin of the silt particles in the loess question. Quaternary International 62, 3–9. Johnsen, S. J., Clausen, H. B., Dansgaard, W., Fuhrer, K., Gundestrup, N., Hammer, C. U., Iversen, P., Jouzel, J., Staufer, B., and Steffensen, J. P., 1992. Irregular glacial interstadials recorded in a new Greenland ice core. Nature 359, 311–313. Khodri, M., Leclainche, Y., Ramstein, G., Braconnot, P., Marti, O., and Cortijo, E., 2001. Simulating the amplification of orbital forcing by ocean feedbacks in the last glaciation. Nature 410, 570–574. Kukla, G. J., 2000. The last interglacial. Science 287, 987–988. Kukla, G., McManus, J. F., Rousseau, D. D., and Chuine, I., 1997. How long and how stable was the last interglacial? Quaternary Science Reviews 16, 605–612. Kukla, G. J., Bender, M. L., de Beaulieu, J.-L., Bond, G., Broecker, W. S., Cleveringa, P., Gavin, J. E., Herbert, T. D., Imbrie, J., Jouzel, J., Keigwin, L. D., Knudsen, K.-L., McManus, J. F., Merkt, J., Muhs, D. R., and Mu¨ller, H., 2002. Last interglacial climates. Quaternary Research 58, 2–13. Lambeck, K., and Chappell, J., 2001. Sealevel change through the last glacial cycle. Science 292, 679–685. Lehman, S. J., Sachs, J. P., Rotwell, A. M., Keigwin, L. D., and Boyle, E. A., 2002. Relation of subtropical Atlantic temperature, high-latitude ice rafting, deep water formation, and European climate 130 000–60 000 years ago. Quaternary Science Reviews 21, 1917–1924. McManus, J. F., Bond, G. C., Broecker, W. S., Johnsen, S., Labeyrie, L., and Higgins, S., 1994. High resolution climate records from the North Atlantic during the last interglacial. Nature 371, 326. McTainsh, G. H., Nickling, W. G., Lynch, A. W., 1997. Dust deposition and particle-size in Mali. Catena 29, 307–322. Menke, B., and Tynni, R., 1984. Das Eeminterglazial und das Weichselfru¨hglazial von Rederstall/ Dithmarschen und ihre Bedeutung fu¨r die mitteleuropa¨ische Jungpleistoza¨n-Gliederung. Geologisches Jahrbuch A 76, 3–120. Mu¨ller, H., 1974. Pollenanalytische Untersuchungen und Jahresschichtenza¨hlungen an der eemzeitlichen Kieselgur von Bispingen/Luhe. Geologisches Jahrbuch A 21, 149–169. Mu¨ller, U. C., Pross, J., and Bibus, E., 2003. Vegetation response to rapid climate change in central Europe during the past 140 000 yr based on evidence from the Fu¨hramoos pollen record. Quaternary Research 59, 235–245. Mu¨ller, U. C., and Kukla, G. J., 2004. North Atlantic Current and European environments during the

declining stage of the last interglacial. Geology 32, 1009–1012. NorthGRIP Members, 2004. High resolution record of Northern Hemisphere climate extending into the last interglacial period. Nature 43, 147–151. Rahmstorf, S., 2002. Ocean circulation and climate during the past 120 000 years. Nature 419, 207–214. Ruth, U., Wagenbach, D., Steffensen, J. P., and Bigler, M., 2003. Continuous record of microparticle concentration and size distribution in the central Greenland NGRIP ice core during the last glacial period. Journal of Geophysical Research 108 (D3): ACL 1–1. Sa´nchez Gon˜i, M. F., Eynaud, F., Turon, J.-L., and Shackleton, N. J., 1999. High resolution palynological record off the Iberian margin: direct land-sea correlation for the last interglacial complex. Earth and Planetary Science Letters 171, 123–137. Sa´nchez Gon˜i, M. F., Loutre, M. F., Crucifix M., Peyron, O., Santos, L., Duprat, J., Malaize´, B., Turon, J.-L., and Peypouquet, J.-P., 2005. Increasing vegetation and climate gradient in Western Europe over the last glacial inception (122–110 kyr) : datamodel comparison. Earth and Planetary Science Letters 231, 111–130. Schaber, K., and Sirocko, F., 2005. Lithology and stratigraphy of the ELSA cores: Eifel laminated sediment archive. Mainzer Geowissenschaftliche Mitteilungen 33, 295–340. Seelos, K., and Sirocko, F., 2005. RADIUS – Rapid particle analysis of digital images by ultra-high resolution scanning of thin sections. Sedimentology 52, 669–681. Seidenkrantz, M.-S., Bornmalm, L., Johnsen, S. J., Knudsen, K. L., Kuijpers, A., Lauritzen, S.-E., Leroy, S. A. G., Mergeal, I., Schweger, C., and Van Vliet-Lanoe¨, B., 1996. Two-step deglaciation at the oxygen isotope stage 6/5E transition: The ZeifenKattegat climate oscillation. Quaternary Science Review 15, 63–75. Sirocko, F., and Sarnthein, M., 1989. Wind-borne deposits in the Northwestern Indian Ocean: record of Holocene sediments versus modern satellite data. In: Leinen, M., and Sarnthein M. (Eds.) Paleoclimatology and Paleometeorology: Modern and Past Patterns of Glacial Atmospheric Transport. NATO ASI Series, C, Math and Phys. Sciences. Kluwer Academic Publishers, Dordrecht, Boston, London, 401–433. Sirocko, F., Seelos, K., Schaber, K., Rein, B., Dreher, F., Diehl, M., Lehne, R., Ja¨ger, K., Krbetschek, M., and Degering, D., 2005. A late Eemian aridity pulse in central Europe during the last glacial inception. Nature 436, 753–888. Stuut, J. B. W., Prins, M. A., Schneider, R. S., Weltje, G., Jansen, J. H., and Postma, G., 2002. A 300 kyr record of aridity and wind strength in

Abrupt Cooling Events at the Very End of the Last Interglacial south western Africa: evidence from grain-size distributions of sediments on Walvis Ridge, SE Atlantic. Marine Geology 180: 221–233. Turner, C., 2002. Problems of the Duration of the Eemian Interglacial in Europe North of the Alps. Quaternary Research 58, 45–48. Tzedakis, P. C., Andrieu, V., de Beaulieu, J.-L., Birks, H. J. B., Crowhurst, S., Follieri, M., Hooghiemstra, H., Magri, D., Reille, M., Sadori, L., Shackleton, N. J., and Wijmstra, T. A., 2001. Establishing a terrestrial chronological framework as a basis for biostratigraphical comparisons. Quaternary Science Reviews 20, 1583–1592. van den Bogaard, P., Hall, C. M., Schmincke, H.-U., and York, D., 1989. Precise single-grain 40Ar/39Ar

229

dating of a cold to warm climate transition in Central Europe. Nature 342, 523–525. van Kreveld, S., Sarnthein, M., Erlenkeuser, H., Grootes, P., Jung, S., Nadeau, M. J., Plaumann, U., and Voelker, A., 2000. Potential links between surging ice sheets, circulation changes, and the Daansgaard-Oeschger cycles in the Irminger Sea, 60–18 kyr. Paleoceanography 15, 425–442. Woillard, G. M., 1978. Grande pile peat bog: A continuous pollen record for the last 140 000 years. Quaternary Research 9, 1–21. Zagwijn, W. H., 1996. An analysis of Eemian climate in western and central Europe. Quaternary Science Reviews 15, 451–469.

This page intentionally left blank

15. Estimates of Temperature and Precipitation Variations During the Eemian Interglacial: New Data From the Grande Pile Record (GP XXI) Denis-Didier Rousseau1,2,3, Christine Hatte´4, Danielle Duzer1, Patrick Schevin1, George Kukla2 and Joel Guiot5 1

Universite´ Montpellier II, Institut des Sciences de l’Evolution (UMR CNRS-UM2 5554) case 61, pl. E. Bataillon, 34095 Montpellier Cedex 5, France 2 Lamont-Doherty Earth Observatory of Columbia University, Palisades, NY 10964, USA 3 Universita¨t Bayreuth, Lehrstuhl fu¨r Geomorphologie, Universita¨tsstr. 30, 95447 Bayreuth, Germany 4 Laboratoire des Sciences du Climat et de l’Environnement, UMR 1572 CEA/CNRS, Domaine du CNRS, Avenue de la Terrasse, 91198 Gif sur Yvette Cedex, France 5 CEREGE, Europoˆle Me´diterrane´en de l’Arbois, BP 80-13545 Aix-en-Provence, Cedex 04, France

ABSTRACT New data were obtained from previously unanalysed Grande Pile core samples (GP XXI) corresponding to the penultimate glacial up to the St. Germain 1 interstadial. Parallel sampling for pollen and carbon isotopes was performed. The biostratigraphy is based on pollen grains as proposed by Woillard (1978) for Grande Pile X. The age scale is from Kukla et al. (1997) except for the start of Eemian, which follows the timing defined by Shackleton et al. (2002). The pollen data were first processed to determine their biome scores. The BIOME4 vegetation model was then run in the inverse mode using the determined biome scores and measured 13 C values as constraints to reconstruct the climate parameters. These results are compared with those previously published and show that the Eemian sensu stricto and more generally the penultimate interglacial period was not stable or uniform contrary to previous terrestrial reconstructions but is in agreement with variations observed in the North Atlantic Ocean. An earlier comparison of the bottom part of the Grande Pile record (E. France) (Woillard, 1978), covering the Linexert

glaciation through the Ognon stadials, with the marine record V29-191 from the North Atlantic Ocean (McManus et al., 1994) covering MIS 6 to 4, provoked a wide debate in the palaeoclimatology community (Kukla et al., 1997). First, assuming a constant sedimentation rate in both records, the comparison between the variation in continental arboreal and temperate deciduous trees with the IRD occurrences and Neogloboquadrina pachyderma sinistral counts showed that all events characterizing the sea-surface conditions in the North Atlantic Ocean are recorded in the terrestrial record. Indeed, all the observed changes in the vegetation, mostly declines in arboreal pollen, are related to events of strong iceberg discharges, named C events in the marine record, and high values of N. pachyderma sinistral (Fig. 15.1). Second, in the earlier comparison (Kukla et al., 1997), the observed variation in benthic forams 18 O showed that the terrestrial interglacial, named Eemian sensu lato after Woillard (1978), was not only associated with marine isotope substage (MIS) 5e as previously assumed (Shackleton, 1969; Mangerud et al., 1979), but also included part of MIS 5d. This observation was

Denis-Didier Rousseau et al. LA GRANDE PILE δ18O (‰)

Depth N. pachyderma s. (%) (cm) 0

20

40

60

80

100

3

AP + decid. Trees (%)

3.5

4 100

80

60

40

20

VEG. 0

Pollen Depth zones (cm)

TR ST

V29-191

OPT WT CT TG

232

620

1140

W21

5a

W22

800

II

II OGN I

I

4

C21

2 ME II

C22

5

C23

Me 3

W23

3

W24

C24

900

1200

1300

1400

1500

ME I

5d

WE

1600

6

5

1000

St. GER II

700

ST I

5e

EEMIAN

W20

4

St. GER II

C18 C19 C20

1700

4

1800 1070 0

10

20

30

40

50

6

1 Linexert Stadial

1830

IRD (%)

Fig. 15.1 Comparison between marine (McManus et al., 1994) and terrestrial (Woillard, 1978) proxies during marine isotope stage (MIS) 5, showing coeval variations in the North Atlantic sea-surface and in the terrestrial vegetation. The benthic 18 O indicates that the Eemian interval includes MIS 5e and part of MIS 5d (from Kukla et al., (1997) modified). W20–24, warm sea-surface water intervals; C18-24, cold sea-surface water intervals; 4, 5a, 5d, 5e, 6 marine isotope stages; WE: Woillard event; AP, arboreal pollen; decid, deciduous; VEG, vegetation zones (Woillard, 1978); OPT, forest of climatic optimum; WT, warm temperate forest; CT, cold temperate forest; TG, taiga; TR, treeless shrubland; ST, steppe; ST I St. Germain I interstadial, OGN, Ognon stadial; ME I, II, Melisey I and II stadials; Me, Montaigu event.

important as the strong change to open steppe vegetation identified in the Grande Pile record as Melisey 1 is not contemporaneous with the continental ice-sheet inception, but is instead related to a strong discharge of icebergs in the North Atlantic Ocean (McManus et al., 1994). This Heinrich-like event implies the presence of ice caps prior to this calving and release of icebergs. Lastly, the duration of the Eemian interglacial was questioned, considering whether or not it was restricted only to the time of deciduous forest. Further studies indicated that the duration of the Eemian varies according to the geographical position of the analysed sites (Kukla, 2000; Kukla et al., 2002a, 2002b, 2002c; McManus et al., 2002; Turner, 2002a, 2002b; Tzedakis et al., 2002; van Kolfschoten and Gibbard, 2000).

Since then, other studies have supported the main conclusions of the paper by Kukla et al. (1997) especially that the last interglacial on the continent is not the equivalent of MIS 5e. The best evidence comes from the combined 18 O and pollen analyses from a core off Portugal, MD95-2042 (Sa´nchez Gon˜i et al., 1999; Shackleton et al., 2002). This record, located south of the iceberg limit defined by Ruddiman and McIntyre (1981), and south of the main IRD belt, indicates that the Eemian in southern Europe includes part of MIS 5e and part of MIS 5d (Sa´nchez Gon˜i et al., 1999; Shackleton et al., 2002). A recent synthesis of Eemian records in Europe and 18 O and IRD from North Atlantic cores supports the interpretation by Kukla et al. (1997) on the continental– marine correlation (Mu¨ller and Kukla,

Estimates of Temperature and Precipitation Variations

2004), showing that the vegetation variation has a strong geographical gradient, which is related to variations in the North Atlantic circulation affecting the NADW formation zone. Climatic estimates proposed for this period are rare and based only on methods referring to modern conditions (analogs) calibrated on modern pCO2 values (Guiot et al., 1989; Guiot, 1990; Cheddadi et al., 1998; Klotz et al., 2003; Mu¨ller et al., 2003; Sa´nchez Gon˜i et al., 2005). Recently, a new method was developed to reconstruct climatic parameters by inverse modelling of biomes and 13 C on loess organic matter (Hatte´ and Guiot, 2005). We applied this method to new data from the Grande Pile locality, GPXXI. This was performed on the levels showing C/N ratio higher than 15 to ensure an atmospheric origin of the organic matter and not a mixture between superior plants and algae (qualified levels are above 1810 cm, younger than 127.7 kyr). Furthermore, a preliminary study combining inverse modelling and the possible 13 C ranges for each expected biomes was performed in the study conditions (altitude, latitude and longitude of La Grande Pile) and under different CO2 concentrations (Hatte´ et al., 2006). This study allowed characterizing a bacterial degradation mostly inducing a 1‰ depletion of the original isotopic composition of the plants along the whole record. We therefore applied the inverse modelling procedure to the C/N qualified levels and taking into account a 1‰ shift of the measured 13 C. This new core was retrieved at the same time and same location as the other cores studied by Woillard (1973, 1978, 1979a) and is preserved in the LDEO core repository. The Grand Pile site (47 449N, 6 309E, 330 m a.s.l) is located in a bog in the southern Vosges Mountains in France. A preliminary study of the bulk density of the base of the sequence permitted the identification of different stratigraphical units, especially the mineral intervals corresponding to the

233

Linexert glaciation (sandy clay) and the Melisey 1 stadial (silty clay), the time interval we focus on here, the Eemian, being represented by a gyttja. The thickness of the units fits with what has been already described in other previously studied cores from the Grande Pile locality (Woillard, 1973, 1978, 1979a, 1979b, 1980; Woillard and Frenzel, 1991; de Beaulieu and Reille, 1992). The new sequence has been sampled every 10 cm for pollen and 13 C measurements (Fig. 15.2) using this preliminary stratigraphy. The pollen preparation followed the classical protocol, and the bulk 13 C was measured in Montpellier using an elementary analyser coupled with a Micromass optima mass spectrometer. All the used methods are described in Hatte´ et al. (2006) and Rousseau et al. (2006). The pollen analysis of 45 samples results in the identification of the classical vegetation succession defined by Woillard (1978) for the interval including the Linexert glaciation to the St. Germain I interstadial. This allows a comparison of our results with the previous analyses of Woillard. Note that the new core yields a complete record of the Linexert glaciation to St. Germain 1 interval. We then applied the classical Grande Pile pollen biostratigraphy and use the age model defined in Kukla et al. (1997) for the top of the studied interval and Shackleton et al. (2002) for the refined chronology of the base of the Eemian (Fig. 15.3). The measurement of the bulk 13 C indicates values varying between about 32 and 22‰, with the lightest measurements yielded by the lowest samples in the core. Conversely, the upper part of the sequence shows values varying around a mean value of 29‰. The biomisation of the pollen data, following the protocol defined by Prentice et al. (1996), yields scores for 12 different possible biomes among 28 potential ones. The inverse modelling of the biomes and of the 13 C allows the palaeoclimate reconstruction to be constrained with this method (Hatte´ et al., 2006; Rousseau et al., 2006). It also takes into account variations in the

234

Denis-Didier Rousseau et al.

La Grande Pile XXI Stratigraphy

–33 –30 –27 –24 –21

1450

Mo

1500

E1 Me Melisey1 E7

1550

E6

1600

δ13C (‰)

St Germain I

Artemisia

Gramineae

Pinus

Picea

Abies

Taxus

Alnus

Corylus

Carpinus

Main pollen taxa identified (%)

Quercus

Ulmus

Betula

Salix

Cupressaceae

Depth (cm)

Eemian

E5

1650 1700

E4

1750

E3 E2 E1 Linexert

1800 1850 0 10 0 5 0 DDR2005

25

0 10

0

25

50

0

25

50

0

25 10

0

25

50

0

25

0

25

0

25

50

0

25

0 5

analysts D. Duzer & P. Schevin

Fig. 15.2 Biostratigraphy of Grande Pile XXI. Plot of the percentage of the main pollen taxa showing the classical vegetation succession occurring during the penultimate interglacial. 13 C measures from parallel samples of the pollen ones. Biostratigraphy nomenclature according to Woillard (1978) (from Rousseau et al. (2006) modified). Age (kyr) 100

105

110

115

120

125

130

1400

Depth (cm)

1500

1600

1700

1800

1900

Fig. 15.3 Age model used for La Grande Pile core XXI, with solid circles representing control ages. Ages interpolated between control points (from Rousseau et al. (2006) modified).

global CO2 concentration and the isotopic composition of atmospheric CO2. The mean annual temperature, the mean temperatures of the coldest and warmest months and the annual precipitation are then reconstructed. The temperature and precipitation estimates show three successive warm intervals prior to a strong cooling at the end of the penultimate interglacial. The first Eemian warming at about 125 kyr at La Grande Pile is marked in all temperature parameters (about 12 C for the mean annual temperature) and is associated with a strong increase in annual precipitation (Fig. 15.4). This warming appears to have been strong in both summer and winter as expressed by the warmest and coldest months, respectively (anomalies of about 19 and 20 C) (Fig. 15.4). A cooling trend first occurs at about 124 kyr, just after the interglacial optimum, as is indicated by all the temperature parameters (124.5–123.8 kyr for mtwa and Tann, and 123.8–123 kyr for mtco), while the mean annual precipitation signal does not show

Estimates of Temperature and Precipitation Variations

Climatic estimates Age (kyr)

–10

Mean annual temperature (°C) –5

0

5

Vegetation

Annual precipitation (mm yr–1)

10

15

20

235

Anomalles In mean temperature Coldest month (°C) –30 –25 –20 –15 –10 –5

0

Arboreal pollen (%)

Warmest month (°C) 5

10

0

20

40

60

80

100 100

100

Me 105

105

ME 1 110

WE

110

115

115

120

120

125

125

130

130 0

200 400 600 800 1000 1200 1400

–15 –10 –5

0

5

10

15

20

Deciduous AP Total AP

Fig. 15.4 Climate estimates from the last interglacial at Grande Pile XXI compared to vegetation indices corresponding to the arboreal (AP) and deciduous arboreal pollen (DAP) percentages. The modern values at Grande Pile are Tann: 9.5 C, Pann: 1080 mm, mtco: 0:5 C, mtwa: 18.5 C. The fat lines show the most probable values, whereas the 95% confidence level intervals are presented with hair lines. The two open circles in mtwa reconstructions correspond to estimates, which remain questionable. Me, Montaigu event; ME1, Melisey I stadial; WE, Woillard event (from Rousseau et al. (2006) modified).

any particular trend. The temperature estimates indicate a warming, between 120 and 118 kyr, mainly expressed in the coldest month mean temperatures, although the anomaly values remain negative and follow a small cold interval between 122 and 121 kyr. The third Eemian warm interval at Grande Pile is between 117 and 115 kyr and is mostly expressed in the annual and coldest month mean temperatures, but only slightly in the warmest month mean temperature. The precipitation again does not show any particular trend. A decrease in mean annual, coldest month and warmest month temperatures associated with decreasing annual precipitation occurred at about 115 kyr (Fig. 15.4). This corresponds to a decrease in the deciduous arboreal vegetation (Fig. 15.1). Estimates of mean annual and coldest month temperature indicate colder conditions from the Woillard event (WE) (Woillard,

1979b) until the Melisey I stadial and are correlated with cold marine events C25 and C24 (McManus et al., 1994; Kukla et al., 1997). The change in the vegetation at Melisey I had been interpreted previously as resulting from the IRD released in the North Atlantic, thus following the first glacial inception. Indeed, the Melisey I stadial at about 107 kyr is marked by low temperature values and is characterized by a strong decrease in annual precipitation of about 500 mm yr 1 as can be seen from our reconstruction, even if due to the sampling resolution this corresponds to only one point. Precipitation remained high between 110 and 107 kyr, while the mean annual and the coldest month temperatures were already very low. The decrease in precipitation lags the cooling by 3000 yr. The discussion of the reconstructed climatic patterns based on the new Grande Pile core requires comparing them with

236

Denis-Didier Rousseau et al.

other evidence, especially from the North Atlantic Ocean. The first warming observed at the base of the Eemian could correspond to the northward migration of the polar front within the Greenland Seas described at 126–125 kyr. The temperature decrease reconstructed at 124 kyr fits with the first cooling described in the Nordic Seas (Fronval et al., 1998), which corresponds to reduced influx in this region of Atlantic water and dominance of Arctic waters. Indeed, records from the Norwegian Sea and Northeast Atlantic Ocean (Cortijo et al., 1994; Maslin et al., 1995; Fronval and Jansen, 1996; Fronval and Jansen, 1997) indicate a sharp decrease in both temperature and salinity. This change in the thermohaline circulation in the Norwegian Sea may have been important enough to be related to our cooling trend. The cold interval estimated on land at about 123–121 kyr could correspond to the increase in IRD deposition recorded at about 122 kyr in the northern Greenland Sea (Fronval et al., 1998). The third warming matches the second warm interval during which the Arctic front reached its most westerly location in the Iceland Sea region. The cold decrease at about 115 kyr could be related, considering our time resolution, to the C26 cold event corresponding to a surface cooling in the North Atlantic (Shackleton et al., 2002). Finally, the WE at about 110 kyr may correspond to the IRD event C25 in the Nordic Seas interpreted as an advance of the Scandinavian, Svalbard-Barents and Greenland ice sheets surrounding the Nordic Seas. This event was also identified in the North Atlantic as the first large-scale cooling of the surface water after MIS 5e (Chapman and Shackleton, 1999; McManus et al., 2002). The strong seasonality recorded between 110 and 107 kyr appears to be in agreement with the results of the 2D earth model of intermediate complexity (EMIC) MoBidiC performed to simulate the climate between 126 and 100 kyr (Sa´nchez Gon˜i et al., 2005). Indeed, the model indicates no continental ice in the Northern Hemisphere

between 124 and 119 kyr. Ice sheets then grow until 109 kyr, with partial melting afterwards until 104 kyr. Our results indicate a decreasing trend in all temperature parameters from about 117 kyr until the lowest values estimated at 110 kyr, with a severe decrease in precipitation occurring at about 107 kyr. One of the main results of our investigation is that the climate of the last interglacial was not uniformly warm. Our results are in agreement with other continental western European records (Guiot et al., 1993; Thouveny et al., 1994; Klotz et al., 2004, Muller et al., 2005) and marine records (McManus et al., 1994; Fronval and Jansen, 1997; Fronval et al., 1998; Chapman and Shackleton, 1999; Shackleton et al., 2002), and could support the hypothesis of a modified thermohaline circulation during the Eemian affecting the climate on the western side of Europe (Broecker, 1998). On the other hand, the characterization of the several coolings disagree with the interpretation from the same locality (Field et al., 1994; Cheddadi et al., 1998; Rioual et al., 2001) or off the Iberian peninsula (Sa´nchez Gon˜i et al., 2005) which show a roughly regular warm interval at the same time. ACKNOWLEDGEMENT We express our thanks to Frank Sirocko for the invitation to the final meeting of the DEKLIM program and his invitation to contribute to this book, P.C. Tzedakis and S. Klotz for valuable comments and corrections. This study was funded by a CNRSNSF exchange program and completed during the stay of the lead author in the University of Bayreuth thanks to a von Humboldt research award. This is ISEM contribution 2005-087, LDEO contribution 6971 and LSCE contribution 1769. Sample material used in this project provided by the Lamont-Doherty Earth Observatory Deep-Sea Sample Repository. Support for the collection and curating facilities of the

Estimates of Temperature and Precipitation Variations

core collection is provided by the National Science Foundation through Grant OCE0002380 and the Office of Naval Research through Grant N00014-02-1-0073.

REFERENCES Broecker, W. S., 1998. The end of the present interglacial: How and when? Quaternary Science Reviews 17, 689–694. Chapman, M. R., Shackleton, N. J., 1999. Global icevolume fluctuations, North Atlantic ice-rafting events, and deep-ocean circulation changes between 130 and 70 ka. Geology 27, 795–798. Cheddadi, R., Mamakowa, K., Guiot, J., de Beaulieu, J. L., Reille, M., Andrieu, V., Granoszewski, W., Peyron, O., 1998. Was the climate of the Eemian stable? A quantitative climate reconstruction from seven European pollen records. Palaeogeography, Palaeoclimatology, Palaeoecology 143, 73–85. Cortijo, E., Duplessy, J. C., Labeyrie, L., Leclaire, H., Duprat, J., van Weering, T. C. E., 1994. Eemian cooling in the Norwegian Sea and North Atlantic Ocean preceding continental ice-sheet growth. Nature 372, 446–449. de Beaulieu, J. L., Reille, M., 1992. The last climatic cycle at La Grande Pile (Vosges, France) a new pollen profile. Quaternary Science Reviews 11, 431–438. Field, M. H., Huntley, B., Mu¨ller, H., 1994. Eemian climate fluctuations observed in a European pollen record. Nature 371, 779–783. Fronval, T., Jansen, E., 1996. Rapid changes in ocean circulation and heat flux in the Nordic seas during the last interglacial period. Nature 383, 806–810. Fronval, T., Jansen, E., 1997. Eemian and early Weichselian (140–60 ka) paleoceanography and paleoclimate in the Nordic seas with comparisons to Holocene conditions. Paleoceanography 12, 443–462. Fronval, T., Jansen, E., Haflidason, H., Sejrup, H. P., 1998. Variability in surface and deep water conditions in the Nordic Seas during the Last Interglacial period. Quaternary Science Reviews 17, 963–985. Guiot, J., 1990. Methodology of the last climatic cycle reconstruction from pollen data. Palaeogeography, Palaeoclimatology, Palaeoecology 80, 49–69. Guiot, J., Pons, A., de Beaulieu, J. L., Reille, M., 1989. A 140 000 year climatic reconstruction from two European pollen records. Nature 338, 309–313. Guiot, J., de Beaulieu, J. L., Cheddadi, R., David, F., Ponel, P., Reille, M., 1993. The climate in Western Europe during the last glacial/interglacial cycle

237

derived from pollen and insect remains. Palaeogeography, Palaeoclimatology, Palaeoecology 103, 73–93. Hatte´, C., Guiot, J., 2005. Paleoprecipitation reconstruction by inverse modelling using the isotopic signal of loess organic matter: Application to the Nussloch loess sequence (Rhine Valley, Germany). Climate Dynamics 25, 315–327. Hatte´, C., Rousseau, D. D., Guiot, J., Kukla, G., 2006. Shift in original isotopic signature of plants due to degradation: Contribution of the modeling and paleoclimatic implications. Global Biogeochemical Cycles (submitted for publication). Klotz, S., Guiot, J., Mosbrugger, V., 2003. Continental European Eemian and early Wu¨rmian climate evolution: comparing signals using different quantitative reconstruction approaches based on pollen. Global and Planetary Change 36, 277–294. Klotz, S., Mu¨ller, U. G., Mosbrugger, V., de Beaulieu, J. L., Reille, M., 2004. Eemian to early Wu¨rmian climate dynamics: History and pattern of changes in central Europe. Palaeogeography, Palaeoclimatology, Palaeoecology 211, 107–126. Kukla, G. J., 2000. The last interglacial. Science 287, 987–988. Kukla, G., McManus, J. F., Rousseau, D. D., Chuine, I., 1997. How long and how stable was the last interglacial? Quaternary Science Reviews 16, 605–612. Kukla, G. J., Bender, M. L., de Beaulieu, J. L., Bond, G., Broecker, W. S., Cleveringa, P., Gavin, J. E., Herbert, D., Imbrie, J., Jouzel, J., Keigwin, L. D., Knudsen, K. L., McManus, J. F., Merkt, J., Muhs, D., Mu¨ller, H., Poore, R. Z., Porter, S. C., Seret, G., Shackleton, N. J., Turner, C., Tzedakis, P. C., Winograd, I. J., 2002a. Last interglacial climates. Quaternary Research 58, 2–13. Kukla, G. J., Clement, A. C., Cane, M. A., Gavin, J. E., Zebiak, S. E., 2002b. Last interglacial and early glacial ENSO. Quaternary Research 58, 27–31. Kukla, G. J., de Beaulieu, J. L., Svobodova, H., Andrieu-Ponel, V., Thouveny, N., Stockhausen, H., 2002c. Tentative correlation of pollen records of the last interglacial at Grande pile and Ribains with marine isotope stages. Quaternary Research 58, 32–35. Mangerud, J., Sonstegaardt, E., Sejrup, H. P., 1979. Correlation of the Eemian (interglacial) stage and the deep-sea oxygen-isotope stratigraphy. Nature 277, 189–192. Maslin, M. A., Shackleton, N. J., Pflaumann, U., 1995. Surface water temperature, salinity, and density changes in the northeast Atlantic during the last 45 000 years: Heinrich events, deep water formation, and climatic rebounds. Paleoceanography 10, 527–544.

238

Denis-Didier Rousseau et al.

McManus, J. F., Bond, G. C., Broecker, W. S., Johnsen, S., Labeyrie, L., Higgins, S., 1994. High-resolution climate records from the North Atlantic during the last interglacial. Nature 371, 326–329. McManus, J. F., Oppo, D. W., Keigwin, L. D., Cullen, J. L., Bond, G. C., 2002. Last interglacial and prolonged interglacial warmth in the North Atlantic. Quaternary Research 58, 17–21. Mu¨ller, U. C., Kukla, G. J., 2004. North Atlantic current and European environments during the declining stage of the last interglacial. Geology 32, 1009–1012. Mu¨ller, U. C., Pross, J., Bibus, E., 2003. Vegetation response to rapid climate change in Central Europe during the past 140 000 yr based on evidence from the Fu¨ramoos pollen record. Quaternary Research 59, 235–245. Mu¨ller, U. G., Klotz, S., Geyh, M. A., Pross, J., Bond, G. C., 2005. Cyclic climate fluctuations during the last interglacial in central Europe. Geology 33, 449–452. Prentice, I. C., Guiot, J., Jolly, D., Cheddadi, R., 1996. Reconstructing biomes from palaeoecological data: a general method and its application to European pollen data at 0 and 6 ka. Climate Dynamics 12, 185–194. Rioual, P., Andrieu-Ponel, V., Rietti-Shati, M., Battarbee, R. W., de Beaulieu, J. L., Cheddadi, R., Reille, M., Svobodova, H., Shemesh, A., 2001. High-resolution record of climate stability in France during the last interglacial period. Nature 413, 293–296. Rousseau, D. D., Hatte´, C., Guiot, J., Duzer, D., Schevin, P., Kukla, G., 2006. Reconstruction of the Grande Pile Eemian using inverse modeling of biomes and 13 C. Quaternary Science Reviews (in press). Ruddiman, W. F., McIntyre, A., 1981. The North Atlantic Ocean during the last deglaciation. Palaeogeography, Palaeoclimatology, Palaeoecology 35, 145–214. Sa´nchez Gon˜i, M. F., Eynaud, F., Turon, J. L., Shackleton, N. J., 1999. High resolution palynological record off the Iberian margin: direct landsea correlation for the Last interglacial complex. Earth and Planetary Science Letters 171, 123–137. Sa´nchez Gon˜i, M. F., Loutre, M. F., Crucifix, M., Peyron, O., Santos, L., Duprat, J., Malaize, B., Turon, J. L., Peypouquet, J. P., 2005. Increasing vegetation and climate gradient in Western Europe over the Last Glacial Inception (122–110 ka): data-

model comparison. Earth and Planetary Science Letters 231, 111–130. Shackleton, N. J., 1969. The last interglacial in the marine and terrestrial records. Proceedings of the Royal Society of London 174, 135–154. Shackleton, N. J., Chapman, M., Sa´nchez-Goni, M. F., Pailler, D., Lancelot, Y., 2002. The classic marine isotope substage 5e. Quaternary Research 58, 14–16. Thouveny, N., de Beaulieu, J. L., Bonifay, E., Creer, K. M., Guiot, J., Icole, M., Johnsen, S., Jouzel, J., Reille, M., Williams, T., Williamson, D., 1994. Climate variations in Europe over the past 140 kyr deduced from rock magnetism. Nature 371, 503–506. Turner, C., 2002a. Formal status and vegetational development of the Eemian interglacial in Northwestern and Southern Europe. Quaternary Research 58, 41–44. Turner, C., 2002b. Problems of the duration of the Eemian interglacial in Europe North of the Alps. Quaternary Research 58, 45–48. Tzedakis, P. C., Frogley, M. R., Heaton, T. H. E., 2002. Duration of last interglacial conditions in Northwestern Greece. Quaternary Research 58, 53–55. van Kolfschoten, T., Gibbard, P. L., 2000. The Eemianlocal sequences, global perspectives: introduction. Geologie en Mijnbouw 79, 129–133. Woillard, G., 1973. Mise en e´vidence de l’Ee´mien sur le Plateau de Haute-Saoˆne. Comptes Rendus de l’Acade´mie des Sciences de Paris 276, 939–942. Woillard, G., 1978. Grande Pile Peat Bog: A Continuous Pollen Record for the Last 140 000 Years. Quaternary Research 9, 1–21. Woillard, G., 1979a. The last interglacial–glacial cycle at Grande Pile in northeastern France. Bulletin de la Socie´te´ belge de Ge´ologie 88, 51–69. Woillard, G. M., 1979b. Abrupt end of the last interglacial s.s. in north-east France. Nature 281, 558–562. Woillard, G., 1980. The pollen record of Grande Pile (N.E. France) and the climatic chronology through the last interglacial–glacial cycle. In: J. Chaline (Ed.), Proble`mes de stratigraphie quaternaire en France et dans les pays limitrophes, pp. 95–103. CRDP Acade´mie de Dijon, Dijon. Woillard, G., Frenzel, B., 1991. Forest changes at the end of the Last Interglacial. In: B. Frenzel (Ed.), Klimageschlichtliche Probleme der letzten 130 000 Jahre. Pala¨oklimaforschung 1, 37–50.

16. Quantitative Time-Series Reconstructions of Holsteinian and Eemian Temperatures Using Botanical Data Norbert Ku¨hl and Thomas Litt Institute for Palaeontology, University of Bonn, Nussallee 8, 53115 Bonn, Germany

ABSTRACT Botanical fossils have successfully been used for quantitative climate reconstructions. Recent developments emphasize the need for statistical approaches which are robust to methodological problems such as the lack of modern analogues and which can quantify uncertainties. Therefore, a method based on probability density functions ( pdf method) was chosen to reconstruct January and July mean temperatures throughout the Holsteinian interglacial stage at two Central European sites. The reconstructions were compared with those of the Eemian interglacial stage for two sites located very close to the Holsteinian sites. The reconstructions quantify similarities and differences between the Eemian and the Holsteinian temperature development. Both interglacials start with a relatively fast warming and end with a distinct cooling. During their course, they show uninterrupted interglacial conditions. However, the Holsteinian seems to be less stable than the Eemian with some intra-interglacial coolings. The magnitude of the main cooling in the mid-Holsteinian is reconstructed as approximately 5 C for January temperature. No great change is reconstructed for July temperature during this episode. The temperature course within the two interglacial stages differs. Reconstructed Holsteinian January and July temperatures were higher in the later part of the interglacial with highest reconstructed most probable values of 2 C and almost 20 C, respectively. The lowest reconstructed temperatures were 2 C for January and 17.5 C for July. In contrast, the Eemian had

its temperature optimum during its early phase. For the Eemian, the trend is more pronounced in January than July temperature with a continuous decrease of 3 C before the beginning of the early Weichselian. For the Holsteinian, a decrease of 10–15 C in average January and 3 C in July temperature is reconstructed at the very end of the interglacial, which resembles in magnitude to the temperature decrease at the end of the Eemian. 16.1 INTRODUCTION Reliable reconstructions of Quaternary climate development are crucial for understanding past glacial–interglacial cycles and for validating general circulation models which may be used to simulate future climate scenarios. Long-term climate change during the Pleistocene can be assessed by climate reconstructions that translate information of proxy data into climatic information. Botanical data are well established as proxy data, because plants strongly depend on climate, of which temperature is a decisive factor for the survival and prospering. The reliability of the reconstructions, however, depends on the quality of the proxy data as well as on the choice of the appropriate transfer functions which convert the information of the proxy data, e.g. presence of a specific taxon or species community, into values of climatic parameters, e.g. temperature. The Eemian and the Holsteinian show a vegetational development which is characteristic of the respective interglacial

240

Norbert Ku¨hl and Thomas Litt

stage. Since the successions differ between the interglacials, sequences of different age can be distinguished biostratigraphically. Records of the same age, however, show a remarkably similar vegetational succession over large areas of Europe. This allows the establishment of pollen assemblage zones of a specific interglacial that can be correlated. The vegetation succession within interglacials may be related to climatic change, but may also be due to very different reasons, e.g. soil maturity. Insolation change is commonly considered as a main factor for the existence of glacial–interglacial cycles. Data-model comparisons are useful to test this hypothesis (e.g., Kaspar et al., 2005) and can lead to an advanced understanding of the climate system. In this regard, the last interglacial, the Eemian, is of particular use, because it is the best investigated interglacial before the onset of the Holocene, and records exist of the complete interglacial, including its end. Because the insolation curve shows overlay of eccentricity, obliquity and precession cycles, which leads to different insolation conditions during different interglacials (Berger, 1978), a comparison of the climate evolution of different interglacials is important for understanding past climate variability. We applied a method based on probability density functions ( pdfs) to pollen records of Holsteinian and Eemian age to explore similarities and differences between these interglacials with respect to temperature. 16.2 DEFINITION OF HOLSTEINIAN AND EEMIAN The Holsteinian and Eemian interglacials are defined on sedimentary records in North Central Europe. Today, they can be recognized and correlated by their typical vegetation succession based on palynostratigraphy. More recently, radiometric dating techniques assist in the correct geochronological attribution.

The term Holsteinian originates from Geikie (1894), who described interglacial marine sediments as ‘Holsteinian beds’. Grahle (1936) used the term Holsteinian for interglacial marine deposits prior to the last interglacial, the Eemian (so-called marine interglacial I after Gottsche, 1898). Hallik (1960) for the first time defined palynostratigraphically Holsteinian sediments in terms of Grahle and correlated them with limnic-terrestrial interglacial records. Type sections of the Holsteinian are HamburgDockenhuden (marine deposits) and Bossel, west of Hamburg (limnic deposits) (see Jerz and Linke, 1987). The vegetation succession of the Holsteinian warm period was later described by several authors which assigned regionally differing pollen assemblage zones (Erd, 1970; Mu¨ller, 1974b). The INQUA Subcommission of European Quarternary Stratigraphy defined the lower boundary of the Holsteinian as the transition from subarctic (still late Elsterian) to boreal conditions, and the upper boundary as the transition from boreal to subarctic (Saalian) conditions (Jerz and Linke, 1987). The duration of the Holsteinian is estimated to be about 15–16 000 years, based on varve counts of Mu¨ller (1974b) at Munster-Breloh and Meyer (1974) at Hetendorf. The records of these sites are included in the reconstructions below. New Th/U datings based on peat deposits from the type section Bossel indicate an age of about 310–330 kyr BP (Geyh and Mu¨ller, 2005), which would correspond to marine isotope stage (MIS) 9. The term Eemian originates from Harting (1874), who called certain sediments from a warm period after a rivulet near Amersfoort (Netherlands). The term was at first used for marine sediments of the last interglacial in Denmark, Northern Germany and the Netherlands (Madsen et al., 1908). Later it was expanded to isochronous terrestrial sediments (Jessen and Milthers, 1928). At that time, pollen analysis already played an important role for the biostratigraphic

Quantitative Time-Series Reconstructions

classification and correlation of this interglacial. The type section (Amersfoort) is palynostratigraphically defined by Zagwijn (1961). The section Amsterdam Terminal has been suggested as a new parastratotype, because it is more complete (van Leeuwen et al., 2000). The U/Th age of the uppermost part of these interglacial deposits (transition from pollen assemblage zone E6 to E7, according to the zones after Menke and Tynni, 1984) is 118:2 6:3 kyr BP (van Leeuwen et al., 2000). At present, pollen assemblage zones of the Eemian in Central Europe are defined slightly different by different authors, according to regional differences (Ku¨hl and Litt, 2003). Nevertheless, the main characteristics of the vegetational succession are unambiguously visible in most parts of Europe (Menke and Tynni, 1984). The Eemian has been correlated with MIS 5e (Shackleton, 1969); however, recent direct correlation of marine and pollen stratigraphies shows a certain offset between the Eemian and MIS 5e (Shackleton et al., 2003; Sanchez-Gon˜i et al., 2005). 16.3. CLIMATE RECONSTRUCTIONS BASED ON PALAEOBOTANICAL DATA Until recently, terrestrial palaeoenvironmental reconstructions have relied overwhelmingly upon interpretations of pollen-stratigraphical records (Berglund et al., 1996), the major climatic events being defined on the basis of inferred vegetational changes (Gibbard and van Kolfschoten, 2004). These were considered to reflect contemporaneous climatic conditions and to be in equilibrium with them. Reconstructions can be attempted simultaneously for a range of different climatic parameters. Multivariate reconstructions are preferable that include the parameters that are most relevant for the proxy indicator. For plants, temperature is the main factor, with precipitation playing a major role as well. Rather than annual average

241

temperatures, the temperature course throughout the year, which determines the continentality, is important for plants. Consequently, January and July temperatures as representing the coldest and warmest month are often chosen as parameters which are suitable to be reconstructed. A classical study which established the use of palaeobotanical data for the generation of quantitative palaeoclimatic inferences was that of Iversen (1944). His indicator species method employed observations of the modern geographical range limits of particular plant taxa and their correspondence to modern climatic parameters (e.g. mean July and/or mean January temperature). This approach has been widely adopted and was extended by additional taxa for which the modern climatic range limits could be established (e.g. Frenzel, 1967, 1991; Kolstrup and Wijmstra, 1977; Grichuk et al., 1984; Zagwijn, 1994, 1996; Litt et al., 1996; Aalbersberg and Litt, 1998; Hoffmann et al., 1998; Isarin and Bohncke, 1999). However, the indicator species approach has its limitations, because the inferences are based on the presence/absence of a selected number of the total assemblage of plants represented in the fossil sample. An alternative approach is the use of multivariate analysis of the relative abundances of all of the taxa represented in a fossil assemblage. Where pollen data are used for this purpose, for example, climatic inferences are derived from studies of the numerical relationship between modern pollen rain or pollen accumulation data and climatic gradients (such as the ‘transfer function’ approach – e.g., Guiot, 1990). The transfer function approach is based on the assumptions that (a) climate is the primary cause of quantitative variations in the fossil pollen record and (b) there is a linear relationship between climatic variables and pollen flora composition (Webb, 1980). It also generally assumes equilibrium between biota and prevailing climate, which is not necessarily the case. From the palaeoecological point of view, these assumptions are difficult to test,

242

Norbert Ku¨hl and Thomas Litt

and may be questioned, because many plants, and in particular trees, migrated relatively slow across continents and were influenced by complex processes of competition and succession (Birks and Gordon, 1985). Hemispheric climate changes, on the other hand, appear to have been very abrupt (NGRIP Members, 2004). Nevertheless, investigations on annually laminated sediments in central Europe have shown that the vegetation reacts immediately after climate fluctuations (Litt et al., 2001). A major problem with this approach is the limited availability of modern analogues which resemble fossil pollen assemblages, in particular with respect to European palynological data. The vegetation of virtually all of Europe has been altered so radically by human activities since the Neolithic period, and perhaps over a longer time span, that it is extremely difficult to find convincing modern assemblage analogues for much of the fossil palaeobotanical record. Despite these difficulties, a number of researchers have employed this approach in an attempt to reconstruct the late Quaternary palaeoclimate history of Europe, including Huntley and Prentice (1988), Guiot et al. (1989), Guiot (1990), Pons et al. (1992), Field et al. (1994), Cheddadi et al. (1998) and Rioual et al. (2001). More recently, attempts have been made to apply the mutual climatic range method developed by beetle specialists (see below) to the analysis of fossil plant macrofossil and pollen records (Sinka and Atkinson, 1999; Pross et al., 2000; Klotz et al., 2004). This method is affected by its strong dependency on climatic thresholds (or climatic ‘spheres’) that determine the climatic range. No account is taken of the probability distributions within the climatic limits. Moreover, the limit of the distribution area and therefore climatic threshold values of a taxon are difficult to determine. For example, as early as 1902, the contemporary distribution limit of Corylus (hazel) served Andersson (1902) for reconstructing Holocene temperatures. Samuelsson (1915),

however, determined a summer temperature of 1.9 C (Andersson: 2.4 C) higher than that of today as sufficient to explain the northernmost distribution limit of Corylus during the Holocene. Another problem in determining thresholds is the definition and dynamic of the distribution limit. Nonfruiting and nonflowering populations are to be neglected when establishing today’s climate–taxon relationships, if pollen and macro remains are used as proxy data. In favourable years, however, those stands may be able to produce pollen and/or fruits, leading to a different determination of their average climatic thresholds depending on the chosen year. In addition, other factors also affect the occurrence of a taxon near its distribution limit, for example snow cover, early or late frost, soil and competition which hamper the determination of temperature threshold values. Attempts are therefore being made to refine this approach by, for example, the use of a plant–climate transfer function model which quantifies the probability density of climate in regard to the occurrence of the diagnostic taxa (Ku¨hl et al., 2002). The area of occurrence of each plant taxon can be defined in ‘climate-space’, in much the same way as has been developed for the analysis of fossil beetle assemblages (Atkinson et al., 1987). However, in this case, the probability for a plant to occur under given climatic conditions can be defined. The resulting nonparametric probability density functions ( pdfs) can be expressed as parametric Gaussian functions (Fig. 16.1). Because more than one climate parameter usually determines a plant’s area of occurrence, the method has been extended by estimating a two-dimensional pdf, for example combining both January and July mean temperatures, normally the warmest and coldest months in NW Europe. In contrast to the indicator species and mutual climate range approaches, this method expresses climate dependencies of the

Quantitative Time-Series Reconstructions

243

Mean temperature July in °C

llex aquifolium 40

% January

35

40 20

30

–10 % July

25

0

10

20

30 °C

0

10

20

30 °C

40 20 –10

20 15 10 5 0 –35

–30

–25

–20

–15

–10

–5

0

5

10

15

5

10

15

Mean temperature January in °C

Mean temperature July in °C

Carpinus betulus 40

% January

35

40 20

30

–10 % July

25

0

10

20

30 °C

0

10

20

30 °C

40 20 –10

20 15 10 5 0 –35

–30

–25

–20

–15

–10

–5

0

Mean temperature January in °C

Fig. 16.1 Presence/absence data and probability density functions (pdfs) of Ilex aquifolium and Carpinus betulus. Gridded climatological data were used (New et al., 1999) with a spatial resolution of 0:5 0:5 . Each circle represents the temperature value of one 0:5 0:5 grid cell and includes the information if Ilex aquifolium and Carpinus betulus, respectively, are present (green) or absent (red) in a grid cell. Upper left: The one-dimensional pdfs (histogram and normal distribution) for January and July mean temperatures which were calculated using the same data points as for the two-dimensional plot. The two-dimensional pdf is visualized in the scatter plot by an ellipse, indicating 1:65sð¼ 90% probabilityÞ of the pdf, and its centre indicates the maximum of the pdf.

different plant taxa as a pdf with mean and variance values, which are interpreted as the most probable climate and uncertainty ranges, respectively (Fig. 16.1).

Potentially, the method can use all taxa of the fossil record for the reconstruction. However, a selection procedure is advantagous to exclude the use of taxa with very

244

Norbert Ku¨hl and Thomas Litt

similar pdfs, because the use of numerous similar pdfs leads to a narrowing of the uncertainty range which may not be justified by the data and may therefore introduce a bias (Ku¨hl et al., 2002). The technique chooses the most significant taxa for the reconstruction, because it is based on the Mahalanobis distance which not only compares means, but also the standard deviation for evaluating the similarity of pdfs (Ku¨hl et al., 2002). The procedure is applied to every sample independently. Every taxon that is selected influences the joint pdf, but the taxa with the smallest climate tolerance influence the reconstruction result most. As a consequence, genera which often have a broad climate range if they include many species either are not involved in or have a minor effect on the reconstruction result. 16.4 PALAEOBOTANICAL AND CHRONOLOGICAL DATA Pollen serves as palaeobotanical proxy data that indicate the composition of the palaeoflora. They are preserved in sediments, especially where anoxic conditions have prevented the fossils from corrosion. Airborne pollen is very abundant in the atmosphere, given that anemophilous plants are growing in the area. Therefore, pollen may be found in high quantities in suitable sediments. We use presence-absence-data, i.e. we interpret the fossil record in regard to which plants might have grown in the area during the time of pollen deposition and sedimentation. The possibility of input of allochtonous pollen must carefully be evaluated, because pollen may be transported long distances or redeposited. The pollen records used for the reconstructions provide pollen data in good temporal resolution. Two Holsteinian sites were chosen (Fig. 16.2). One is Gro¨bern-Schmerz (Eißmann et al., 1995), which spans the complete interglacial (Fig. 16.3). We compiled the second Holsteinian record from Hetendorf (Meyer, 1974) and Munster-Breloh

60°

Bispingen Hetendorf/ MunsterBreloh

50°

Gröbern Gröbern-Schmerz

La Grande Pile

40°

350°

0°

10°

20°

30°

Fig. 16.2 Location of Eemian and (in italic) Holsteinian sites.

(KS 416/71, Mu¨ller, 1974b), in the following referred to as Hetendorf/Munster-Breloh (Fig. 16.2). The Hetendorf record covers the first seven pollen zones after Meyer (1974) and Mu¨ller (1974b), while the record of MunsterBreloh starts within zone VI and provides a record at high temporal resolution to zone XVI after Mu¨ller (1974b). The very first pollen zone I at Hetendorf contains a large amount of tertiary pollen which indicates redeposition. This zone is therefore not included in the temperature reconstruction. Varves in the Hetendorf record have been counted from zone III to the early part of VI and at Munster-Breloh from zone VII to early XII. The duration of zone II, the larger part of zones VI and XII, and zones XIII to XVI could not be detemined by varve counting and had to be estimated. Mu¨ller (1974b) and Meyer (1974) provide an estimation of these parts of the profile by extrapolating the results of the varve counts of the other zones for which they considered the varve counts of the adjoined zones and the sediment composition. Three Eemian sites were chosen to reconstruct the temperature of the last interglacial (Fig. 16.2). The first site, Gro¨bern, not only provides a Holsteinian record (Fig. 16.3), but also an excellent pollen record for the

Quantitative Time-Series Reconstructions

245

900

Bu x Fa us gu Pt s e C roc el a H tis rya e Vi der sc a Ile um Li x g Vi ustr bu um Vi rn ti u Hs m er bs

Ac e Ta r xu s Fr ax T i in u l C ia s ar pi nu s Ab ie s

U lm Q us ue rc Pi us ce C a or yl u Al s nu s

us Pi n

H ip Ju pop n h Sa ipe ae l ru Be ix s tu la

Gröbern-Schmerz (Holsteinian)

Zone 7

1000 1100

6

1200 1300 1400

Depth in cm

1500 1600

5

1700 1800 1900 2000

4

2100 2200 3 2300 2400

2

2500 2600

1 20 40 60

20 40 60

20

20

20

20

20

20

20

Per cent

Fig. 16.3 Simplified pollen diagram of the Holsteinian site Gro¨bern-Schmerz (Analysis: Ko¨hler). After Eißmann et al. (1995), revised and redrawn.

Eemian (Litt, 1994) (Fig. 16.4). The second site is Bispingen (Mu¨ller, 1974a) which is located close to the Holsteinian sites Hetendorf and Munster-Breloh. This allows the comparison of Holsteinian with Eemian records that originate from the same area. The third Eemian record is La Grande Pile (de Beaulieu and Reille, 1992). For Bispingen, varve counting could be performed for a large part of the Eemian. No profile is as yet available which could be completely varve counted. Bispingen provides so far the best available record of the duration of the specific pollen zones, and the terrestrial Eemian as a whole, in northern central Europe. However, these counts only reach into the Carpinus phase of the Eemian (pollen zone 5 after Menke and Tynni, 1984), and the duration of the upper part of the interglacial (parts of pollen zone 5 and zones 6 and 7) was estimated by Mu¨ller (1974a) and is therefore less certain. Gro¨bern

and La Grand Pile were correlated with Bispingen biostratigraphically. The beginning of the interglacial as recorded by the immigration of interglacial vegetation can be regarded as relatively isochronic (e.g. Tzedakis, 2003). The beginning of the different pollen zones as defined by the subsequent domination of different taxa may have been somewhat asynchronous, but is likely not to have differed considerably within central Europe. The Bispingen record allows an estimation of the duration in which certain climatic variations took place. This is particularly valuable since sedimentation rates differ through time and a graph based on sample depths does not provide reliable temporal information. For example, in the Bispingen record the sedimentation rate of 10.5 cm in pollen zone 2 is considerably higher than that of the varve counted pollen zone 1 (7 cm/century) or the partially counted pollen zone 5 (Table 16.1).

Norbert Ku¨hl and Thomas Litt

246

H ed Vi era sc Li um g Ile ust x rum H er bs

ax Ab inu ie s s

Fr

Al n Ac us Ta er x Ti us lia C ar pi nu s

Pi ce a C or yl us

Pi nu s

U lm Q us ue rc us

700

H ip Ju pop n h Sa ipe ae li ru rha m Be x s no tu id la es

Gröbern (Eemian)

Zone 7 6b

750

6a 800 5

Depth in cm

850

4b

900 4a 950 3

1000

2

1050

1 1100 S 20 40 1150

20 40 60 80

20 40

20 40 60

Per cent

20 40 60

20 40

20 40

%

Fig. 16.4 Simplified pollen diagram of the Eemian site Gro¨bern (after Litt, 1994). Table 16.1 Sedimentation rates of the varve counted pollen assemblage zones of the Eemian record from Bispingen Zones after Mu¨ller (1974a)

Zones after Menke and Tynni (1984)

Thickness of zones in cm (Mu¨ller, 1974a)

Varve counted years (Mu¨ller, 1974a)

Sedimentation rate cm/century

IVa þ b IIIc IIIb IIIa IIb IIa I

5 4b

240 96 59 36 47 21 7

4000** 1100* 700 450 450 200 100

6.0 8.7 8.4 8.0 10.4 10.5 7.0

*

4a 3 2 1

partly counted; estimated.

**

It has to be noted, though, that the timescale and the duration of the Eemian interglacial in different areas is subject to debate (Kukla et al., 1997; Turner, 2002; Tzedakis, 2003). For example, Kukla et al. (2002) postulated a duration of about 20 000 years based on the Grand Pile record. To provide

the results independently from a timescale which might be subject to change and to enable comparison to the original depth data, our reconstructions are presented on a depth scale (Figs. 16.5 and 16.6). To provide a supposedly more realistic view of timing and duration of climatic changes,

Quantitative Time-Series Reconstructions

graphs using the tentative timescale are presented as well (Figs. 16.7 and 16.8). New data may shed more light on the duration of the Eemian in the future. Nevertheless, at least the comparison between the Bispingen and Gro¨bern record based on the synchronization of pollen assemblage zones appears to be unambiguous. 16.5 RECONSTRUCTIONS The reconstruction and comparison of Holsteinian and Eemian January and July mean temperatures is visualized in two different ways. One is presenting the reconstructions in a graph, with colours referring to the reconstructed probability densities (Figs. 16.5 and 16.6). These graphs plot the reconstructed values versus sediment depth. The other types of graphs (Figs. 16.7 and 16.8) plot the reconstructed values on a timescale as derived from varve counting and correlation between sites (see data section). These figures indicate the reconstructed most probable temperature and their uncertainty ranges. This range corresponds to the standard deviation multiplied by 1.65. This factor provides a 90% probability within the range and includes most – though not all – of the temperature values where the plants found in the fossil sample occur at present. An occurrence of the plant under climatic conditions outside the range is improbable, but it has to be noted that this uncertainty range is not equivalent to minimum and maximum thresholds. In the text, however, the usual notation of is given. The quantification of uncertainty is a very important component of any reconstruction based on (botanical) proxy data, even though this is often omitted in publications. While reconstructions aim at obtaining results as precise as possible, for plants it is essential to be able to survive and prosper in different climates. A plant species with a narrow climatic tolerance can survive well only as long as climate does not change. In

247

order to survive in the long term, the ability to deal with changing climate is an advantage for plants and therefore, having a broad climatic tolerance is an advantage. This tolerance must not be neglected. It is therefore mandatory for any reconstruction based on proxy data to quantify the uncertainty of the reconstructions and include it in figures. 16.5.1 Holsteinian The vegetation development of the interglacial started with birch dominating in the initial phase, followed by pine. Subsequently, thermophilous species started to occur in the area. The taxa that were present during the initial phase (zones I–III after Mu¨ller, 1974b) are relatively insensitive to climate and have a broad climate range. This means that reconstructions are associated with a broad uncertainty. The reconstructed most probable January/July temperature and their standard deviation for this pre-Holsteinian/early Holsteinian phase is around 6 4:2=15 2:4 C (Gro¨bern-Schmerz) and 1413:5=163:4 C (Hetendorf). As soon as thermophilous taxa occur, substantially higher temperatures are reconstructed (Fig. 16.5). These resemble today’s temperatures, but show some variation within the interglacial. In addition, the uncertainty range is considerably smaller. A temperature trend towards higher temperatures is observable in the course of the Holsteinian. In particular, July temperatures show an increase from early Holsteinian values from around 17.5 C to about 18.5 C, with highest most probable values of 19.7 C. Plotting the reconstructions on a timescale reveals that the warmer interval spans about half of the entire warm phase (Fig. 16.7). Reconstructed most probable winter temperatures generally resemble the July trend of increasing values with increasing duration of the interglacial. This is in particular observable at Hetendorf/Munster-Breloh and to some degree also at Gro¨bern-Schmerz. A significant difference to the reconstructions for July, however, is the observation of

Norbert Ku¨hl and Thomas Litt

248 Gröbern-Schmerz

0.4

pd

0.3

30

0.2 0.1

10 0

0.0

TJuly (°C)

20

0.4

–10

0.3 0.2

0

0.1

–10 –20

0.0

TJanuary (°C)

10

–30 1500 II

1000 XI a+b

VI+VII+VIII IX+X

XII

500 c

0 XIV

XIII

III+IV+V

III II IV

V

VI

VII

VIII

IX

b

XII

c

XIV XIII

0.2

20

0.1

10 0

0.4

–10

0.3

10

0.2

0

0.1

–10 –20

0.0

TJanuary (°C)

pd

0.3

a

0.0

TJuly (°C)

30

X XI

0.4

Hetendorf/Munster-Breloh

–30 2000

1500

1000

500

0

Relative depth in cm

Fig. 16.5 Reconstructed probability densities (pd) of January and July mean temperatures during the Holsteinian interglacial at Gro¨bern-Schmerz and Hetendorf/Munster-Breloh. The dashed white line indicates the mean m. The x-axis is a relative depth scale. The pollen diagrams were correlated using pollen floristic characteristics. The Roman Numerals indicate pollen assemblage zones of Hetendorf/Munster-Breloh after Mu¨ller (1974b). See Fig. 16.7 for the plot on a timescale.

relatively high mean January temperatures during the early Holsteinian (pollen zone VII after Mu¨ller, 1974b). This appears more pronounced at Gro¨bern-Schmerz, but is also observable at Hetendorf/Munster-Breloh (Figs. 16.5 and 16.7). The latter record shows high values in January temperature only in a few samples in the early Holsteinian. Possibly, the moderate number of pollen grains

counted per sample in the Hetendorf/Munster-Breloh record accounts for the lack of thermophilous rare taxa in more samples. During the interglacial, no extreme fluctuations are reconstructed neither for the Gro¨bern-Schmerz nor for the Hetendorf/ Munster-Breloh record. This is somewhat surprising, because distinct events are seen in the pollen diagram of Munster-Breloh and of Gro¨bern-Schmerz. In the record of Munster-Breloh, two pollen zones (VIII and XI after Mu¨ller, 1974b) show a pronounced decrease in pollen percentages of some thermophilous taxa. In zone VIII, Taxus changes significantly, with percentages decreasing from over 5 to 1% and not recovering for centuries. However, Taxus seems not to be affected by the second event. In contrast, Carpinus is heavily affected by both events. Before the first event, it was not a significant component of the forests, with pollen percentages of 2%. The second event led to a decrease of Carpinus as one of the important forest components. After this event, Carpinus was almost absent at Munster-Breloh. Other taxa such as Quercus and Corylus also show a major decrease in their pollen representation at Munster-Breloh, in particular during the first event. It has to be noted, however, that this zone VIII event is only represented by one or two pollen spectra in the Munster-Breloh pollen record. The temporal resolution of the Gro¨bern-Schmerz record is not high enough to detect the event itself. Nevertheless, the long-lasting decrease of Taxus from 12 to 2% and the first expansion of Carpinus allow to locate this event at the transition from pollen zone 3 to 4 after Erd (1970) (Fig. 16.3). For the zone VIII event, the lowest winter temperatures of the Holsteinian are reconstructed with January temperatures of 7:1 3:2 C for Hetendorf/MunsterBreloh. For Gro¨bern-Schmerz, too, a decrease is reconstructed, but only as low as 3:8 2:2 C for January, which may be explained by the abovementioned lower temporal resolution compared with MunsterBreloh. Reconstructions of July temperature

Quantitative Time-Series Reconstructions

0.0 0.1 0.2 0.3 0.4

TJuly (°C)

20 10 0

0.0 0.1 0.2 0.3 0.4

–10

TJanuary (°C)

10 0 –10 –20 –30 1100

1200

E1 E2 E3

1000

E4a

E4b

900

E5

E6

E7

TJuly (°C)

30 20 10 0 –10

TJanuary (°C)

10 0 –10 –20 –30 2600

2500

E1 E3 E4a E2

2400

2300

E4b

2200

E5

2100

E6

2000

1900

0.0 0.1 0.2 0.3 0.4

0.0 0.1 0.2 0.3 0.4

Bispingen

E7

0.0 0.1 0.2 0.3 0.4

La Grande Pile 30 20 10 0 –10

0.0 0.1 0.2 0.3 0.4

The temperature reconstructions with the pdf method yield the following results for the climate development of the Eemian. At the beginning of the Eemian, a steep increase both in January and in July temperatures is reconstructed at all three sites (Fig. 16.6). This temperature increase obviously took place during a relatively short period (Fig. 16.8). Mu¨ller (1974a) quantifies its duration to several hundred years by varve counting. After reaching the temperature optimum at a relatively early stage, the reconstructed temperature variability indicates comparatively stable conditions during the entire Eemian. Some minor variations in average temperature might have taken

pd

30

TJuly (°C)

16.5.2. Eemian

Gröbern

10

TJanuary (°C)

do not show a significant decrease during this period for any of the two records (Fig. 16.7). The zone XI event, characterized by a sharp degline of the Carpinus curve, can be located at Gro¨bern at the very beginning of pollen zone 5 after Erd (1970) (Fig. 16.3). Recontructions of this zone also show a temperature decrease, both in January as well as in July temperature. However, the magnitude does not exceed the variation that is visible in the first part of the Holsteinian (Fig. 16.9). At the transition of pollen zone XIIb to XIIc after Mu¨ller (1974b), an additional event is visible in the Munster-Breloh profile which is expressed by changes in the pollen representation of many tree species. This event is expressed less distinct in Gro¨bern-Schmerz, where the total tree pollen concentration is not affected. Nevertheless, composition of trees also changed at Gro¨bern-Schmerz in pollen zone 5 after Erd (1970), which presumably can be correlated with the zone XIIb/c event at Munster-Breloh. In particular, Betula pollen reaches values of 30% (Fig. 16.3). In the rest of the profile, it has low percentages which may be attributed to long-distance transport. The continuous presence of thermophilous taxa keeps the reconstruction from indicating strong oscillations during this period.

249

0 –10 –20 –30 1800

1750

1700

1650

1600

1550

Relative depth in cm

Fig. 16.6 Reconstructed probability densities (pd) of January and July mean temperatures during the Eemian interglacial at three European sites (redrawn after Ku¨hl and Litt, 2003). The dashed white line indicates the mean m. The x-axis is a depth scale. The pollen diagrams were correlated using pollen floristic characteristics. Pollen assemblage zones E1–E7 after Menke and Tynni (1984). See Fig. 16.8 for the plot on a timescale.

250

Norbert Ku¨hl and Thomas Litt

place. However, the probability density distributions (Fig. 16.6) indicate that one should be cautious about overinterpreting minor fluctuations that are visible in the reconstructions. The reconstruction results are confirmed by stable oxygen isotope investigations from one of the sites discussed here, Gro¨bern, which show uninterrupted interglacial conditions (Boettger et al., 2000). Changes in climate may have been the reason for shifts in vegetation that are expressed in the typical succession patterns of the interglacial (Fig. 16.4), but we may exclude a major temperature shift as the main reason. Rather, for example, precipitation change and soil development might be taken into account as a possible reason for the intra-Eemian vegetational succession. It is an important result that strong oscillations are not apparent within the Eemian (Ku¨hl and Litt, 2003). Extreme events as previously postulated by other reconstructions were most likely due to methodological limitations. As discussed by Ku¨hl and Litt (2003), Carpinus and also Picea may provide an explanation for reconstructions of mean January temperature as low as 20 C based on the modern analogue technique (Field et al., 1994). Today, pollen spectra with high values of Carpinus can hardly be obtained in the distribution area of the species. Human alteration of the vegetation and the presence of Fagus which has been dominating in central Europe during the late Holocene, but was absent during the Eemian, make surface samples difficult to find that are analogous to pollen spectra of Eemian age. Nevertheless, a long-term cooling trend in January temperature is reconstructed for the Eemian. For Bispingen, the difference between the early and late part of the Eemian accounts to 3 C. The plants that lead to the reconstruction of this decrease are mainly conifers (Picea and Abies). In particular, Picea seems to be a special case. The modern January temperature range of this conifer is between 19 C and 0 C. The

contemporaneous presence of more thermophilous taxa during the Eemian, however, indicates relative warm temperatures, partly above 0 C. Potentially, Picea may have occurred in areas with January temperatures above 0 C. While this can currently only be assumed, it emphasizes the importance of using reconstruction methods that are robust to such uncertainties. The transition from the end of the Eemian into the first stadial appears far more drastic than the temperature decrease within the Eemian. In particular, winter temperatures are affected and show a change to 8 C. The change in July temperatures was less dramatic. It has to be considered, however, that the span between lowest and highest temperature is much narrower for the July than for the January mean temperature. Hence, optimal temperature varies less for July than for January between species. The reconstructed decrease in July temperature alone is not sufficient to explain the end-Eemian collapse of forests as recorded in the pollen records. Most likely, this is a combined effect of July and January temperatures. For example, at present, areas with average January temperatures below 10 C may still be inhabited by broad-leaved trees, if July average temperatures are high enough ð 18 CÞ. Vice versa, oceanic areas such as Great Britain with relatively high winter temperatures (mean January temperatures somewhat above 0 C) bear broad-leaved trees, although July temperatures are below 15 C. Areas with July temperatures below 12 C in average are dominated by Atlantic dwarf-shrub heaths in northern Great Britain.

16.6 COMPARISON OF INTERGLACIALS The records allow the comparison of two different interglacials at the same location or in the same geographical area. The Holsteinian and Eemian record of Gro¨bern are situated

Quantitative Time-Series Reconstructions

very close to each other, and Munster-Breloh and Bispingen lie at a distance of 15 km. Reconstructions for the Holsteinian and the Eemian show certain similarities between the interglacials. The increase in temperature at the beginning and the decrease at the end of the interglacials had a similar magnitude and did not take a long time (in the order of centuries, Figs. 16.7 and 16.8). January and July temperatures within the interglacials were comparable to present climate conditions within a few degrees

251

variation. Although variation and climatic trend within the interglacials seem to have differed, our reconstruction indicates that major climatic setbacks in a magnitude greater than several degrees in mean July and more than 5 C in mean January temperature are unlikely. Nevertheless, differences between the interglacials are apparent as well. First, the duration of the interglacials was different, the Holsteinian obviously having lasted a few millennia longer than the Eemian.

Gröbern-Schmerz II

III+IV+V

VI+VII+VIII

IX+X

XI

XII

XIII

XIV

25

TJuly (°C)

20 15 10 5

TJanuary (°C)

0 –5 –10 –15 –20 2000

0

4000

6000

8000

10 000

12 000

14 000

16 000

Duration in years Hetendorf / Munster-Breloh II III IV

V

VI

VII

VIII

IX

X

XI

XII

XIII

XIV

TJuly (°C)

25 20 15 10 5

TJanuary (°C)

0 –5 –10 –15 –20 0

2000

4000

6000

8000

10 000

12 000

14 000

16 000

Duration in years

Fig. 16.7 Reconstruction of January (blue) and July (red) temperatures during the Holsteinian interglacial at Gro¨bern-Schmerz and Hetendorf/Munster-Breloh. In contrast to Fig. 16.5, the reconstructed most probable temperature and 1:65s uncertainty range is shown, and the x-axis values are based on the age estimation in Meyer (1974) and Mu¨ller (1974b). The roman numerals indicate pollen zones after Mu¨ller (1974b).

Norbert Ku¨hl and Thomas Litt

252 Gröbern E1 E2 E3 E4a E4b

E5

E6

E7

8000

10 000

E6

E7

8000

10 000

TJuly (°C)

25 20 15 10 5

TJanuary (°C)

0 –5 –10 –15 –20 0

2000

4000

6000

Duration in years Bispingen E1 E2 E3 E4a E4b

E5

TJuly (°C)

25 20 15 10 5

TJanuary (°C)

0 –5 –10 –15 –20 0

2000

4000

6000

Duration in years

Fig. 16.8 Reconstruction of January (blue) and July (red) temperatures during the Eemian interglacial at Bispingen and Gro¨bern (after Ku¨hl and Litt, 2003). In contrast to Fig. 16.6, reconstructed most probable temperature and 1:65s uncertainty range is shown, and the x-axis values are based on the age estimation in Mu¨ller (1974a). Pollen assemblage zones E1–E7 after Menke and Tynni (1984).

Second, the maximum reconstructed temperatures for the Holsteinian are higher than the highest values for the Eemian. Third, the reconstruction of Holsteinian temperatures shows for Hetendorf/Munster-Breloh as well as for Gro¨bern a trend different from the Eemian. Temperature decreased within the Eemian but increased within the Holsteinian. The first half of the Holsteinian is characterized by temperatures somewhat lower than today. In the second half, the reconstructed mean temperatures are higher, in particular the July temperature. In contrast, the Eemian temperatures show a decrease

which is more pronounced for January than July temperatures. In addition, the reconstructions of Holsteinian temperatures suggest that climate varied more during the Holsteinian than during the Eemian. Pollen diagrams show that the Holsteinian vegetation changed considerably in some events. However, the events are not visible in all pollen diagrams. Two reasons may account for this. The events had a relatively short duration, and therefore can only be seen in records with a good temporal resolution. In addition, it seems the vegetation was not completely destroyed and could recover relatively fast. If plants would have had to migrate from distant refugia, this would have resulted in a considerable lag in vegetation recovery. In conclusion, the reconstructions point to a strong warming and cooling at the beginning and end of the Eemian and Holsteinian, respectively, but also reveal a unique climate history during the course of the Holsteinan compared with the Eemian.

REFERENCES Aalbersberg, G., Litt, T., 1998. Multiproxy climate reconstructions for the Eemian and Early Weichselian. Journal of Quaternary Science 13, 367–390. Andersson, G., 1902. Hasseln i Sverige fordom och nu. Sveriges Geologiska Undersoekning C 3, Stockholm, 168 pp. Atkinson, T.C., Briffa, K.R., Coope, G.R., 1987. Seasonal temperatures in Britain during the past 22,000 years, reconstructed using beetle remains. Nature 325, 587–592. Berger, A.L., 1978. Long term variations of daily insolation and Quaternary climate changes. Journal of Atmospheric Sciences 35, 2362–2367. Berglund, B., Birks, H.J.B., Ralska-Jasiewiczowa, M., Wright, H.E., 1996. Palaeoecological Events During the Last 15000 Years. Wiley, Chichester, 764 pp. Birks, H.J.B., Gordon, A.D., 1985. Numerical Methods in Quaternary Pollen Analysis. Academic Press, London, 317 pp. Boettger, T., Junge, F.W., Litt, T., 2000. Stable climatic conditions in central Germany during the last interglacial. Journal of Quaternary Science 15, 469–473. Cheddadi, R., Mamakowa, K., Guiot, J., de Beaulieu, J.L., Reille, M., Andrieu, V., Granoszewski, W.,

Quantitative Time-Series Reconstructions Peyron, O., 1998. Was the climate of the Eemian stable? A quantitative climate reconstruction from seven European pollen records. Palaeogeography, Palaeoclimatology, Palaeoecology 143, 73–85. de Beaulieu, J.L., Reille, M., 1992. Long Pleistocene pollen sequences from the Velay Plateau (Massif Central, France). Vegetation History and Archaeobotany 1, 233–242. Eißmann, L., Litt, T., Wansa, S., 1995. Elsterian and Saalian deposits in their type area in central Germany. In: Ehlers, J., Kozarski, S., Gibbard, P.L. (Eds.), Glacial Deposits in North-East Europe. Balkema, Rotterdam, pp. 439–464. Erd, K., 1970. Pollenanalytical classification of the Middle Pleistocene in the German Democratic Republic. Palaeogeography, Palaeoclimatology, Palaeoecology 8, 129–145. Field, M.H., Huntley, B., Mu¨ller, H., 1994. Eemian climate fluctuations observed in a European pollen record. Nature 371, 779–783. Frenzel, B., 1967. Die Klimaschwankungen des Eiszeitalters. Vieweg, Braunschweig, 291 pp. (in German). Frenzel, B., 1991. Das Klima des letzten Interglazials in Europa. In: Frenzel, B. (Ed.), Klimageschichtliche Probleme der letzten 130.000 Jahre. Fischer, Stuttgart, pp. 51–78. (in German). Geikie, J., 1894. The Great Ice Age, and Its Relation to the Antiquity of Man. Stanfort, London, 850 pp. Geyh, M.A., Mu¨ller, H., 2005. Numerical 230 Th/U dating and a palynological review of the Holsteinian/Hoxnian Interglacial. Quaternary Science Reviews 24, 1861–1872. Gibbard, P., van Kolfschoten, T., 2004. The Pleistocene and Holocene Epochs. In: Gradstein, F.M., Ogg, J.G., Smith, A.G. (Eds.), A Geologic Time Scale 2004. Cambridge University Press, pp. 441–452. Gottsche, C., 1898. Die Endmora¨nen und das marine Diluvium Schleswig-Holstein´s, im Auftrage der Geographischen Gesellschaft in Hamburg untersucht. Mittheilungen der Geographischen Gesellschaft in Hamburg XIV Theil II: Das marine Diluvium, 1–74. (in German). Grahle, H.-O., 1936. Die Ablagerungen der HolsteinSee (Mar. Interglaz. I), ihre Verbreitung und Schichtenfolge in Schleswig-Holstein. Abhandlungen der Preußischen Geologischen Landesanstalt 172, 1–110. (in German). Grichuk, V.P., Gurtovaya, Y.Y., Zelikson, E.M., Borisova, O.K., 1984. Methods and results of Late Pleistocene paleoclimatic reconstructions. In: Velichko, A.A. (Ed.), Late Quaternary Environments of the Soviet Union. Longman, London, pp. 251–260. Guiot, J., 1990. Methodology of the last climatic cycle reconstruction in France from pollen data. Palaeogeography, Palaeoclimatology, Palaeoecology 80, 49–69.

253

Guiot, J., Pons, A., de Beaulieu, J.L., Reille, M., 1989. A 140,000-year climatic reconstruction from two European pollen records. Nature 338, 309–313. Hallik, R., 1960. Die Vegetationsentwicklung der Holstein-Warmzeit in Nordwestdeutschland und die Altersstellung der Kieselgurlager der su¨dlichen Lu¨neburger Heide. Zeitschrift der Deutschen Geologischen Gesellschaft 112, 326–333. (in German). Harting, P., 1874. De bodem van het Eemdal. Verslagen en Verhandelingen Koninklijke Academie van Wetenschappen II Deel VIII, 282–290. Hoffmann, M.H., Litt, T., Ja¨ger, E.J., 1998. Ecology and climate of the early Weichselian flora from Gro¨bern (Germany). Review of Palaeobotany and Palynology 102, 259–276. Huntley, B., Prentice, J.C., 1988. July temperatures in Europe from pollen data, 6000 years before present. Science 241, 687–690. Isarin, R.F.B., Bohncke, S., 1999. Mean July temperatures during the Younger Dryas in Northwestern and Central Europe as inferred from climate indicator species. Quaternary Research 51, 158–173. Iversen, J., 1944. Viscum, Hedera and Ilex as climate indicators. Geologiska Fo¨reningens Fo¨rhandlingar 66, 463–483. Jerz, H., Linke, G., 1987. Arbeitsergebnisse der Subkommission fu¨r Quarta¨rstratigraphie. Typusregion des Holstein-Interglazials. Eiszeitalter und Gegenwart 37, 145–148. Jessen, K., Milthers, V., 1928. Stratigraphical and paleontological studies of interglacial fresh-water deposits in Jutland and Northwest Germany. Danmarks Geologiske Undersoegelse Raekke II 48, 1–379. Kaspar, F., Ku¨hl, N., Cubasch, U., Litt, T., 2005. A model-data-comparison of European temperatures in the Eemian interglacial. Geophysical Research Letters 32, L11703, doi:10.1029/2005GL022456. Klotz, S., Mu¨ller, U., Mosbrugger, V., de Beaulieu, J.L., Reille, M., 2004. Eemian to early Wu¨rmian climate dynamics: history and pattern of changes in Central Europe. Palaeogeography, Palaeoclimatology, Palaeoecology 211, 107–126. Kolstrup, E., Wijmstra, T.A., 1977. A palynological investigation of the Moershooft, Hengelo, and Denekamp Interstadials in the Netherlands. Geologie en Mijnbouw 56, 85–102. Ku¨hl, N., Litt, T., 2003. Quantitative time series reconstruction of Eemian temperature at three European sites using pollen data. Vegetation History and Archaeobotany 12, 205–214. Ku¨hl, N., Gebhardt, C., Litt, T., Hense, A., 2002. Probability density functions as botanical– climatological transfer functions for climate reconstruction. Quaternary Research 58, 381–392. Kukla, G., McManus, J.F., Rousseau, D.D., Chuine, I., 1997. How long and how stable was the last interglacial? Quaternary Science Reviews 16, 605–612.

254

Norbert Ku¨hl and Thomas Litt

Kukla, G.J., de Beaulieu, J.L., Svobodova, H., AndrieuPonel, V., Thouveny, N., Stockhausen, H., 2002. Tentative correlation of pollen records of the last interglacial at Grand Pile and Ribains with marine isotope stages. Quaternary Research 58, 32–35. Litt, T., 1994. Pala¨oo¨kologie, Pala¨obotanik und Stratigraphie des Jungquarta¨rs im nordmitteleuropa¨ischen Tiefland. Dissertationes Botanicae 227. Cramer, Berlin Stuttgart, 185 pp. Litt, T., Junge, F.W., Bo¨ttger, T., 1996. Climate during the Eemian in north-central Europe – a critical review of the palaeobotanical and stable isotope data from central Germany. Vegetation History and Archaeobotany 5, 247–256. Litt, T., Brauer, A., Goslar, T., Merkt, J., Balaga, K., Mu¨ller, H., Ralska-Jasiewiczowa, M., Stebich, M., Negendank, J.F.W., 2001. Correlation and synchronisation of late glacial continental sequences in northern central Europe based on annually laminated lacustrine sediments. Quaternary Science Reviews 20, 1233–1249. Madsen, V., Nordmann, V., Hartz, N., 1908. EemZonerne. Studier over Cyprinaleret og andre Eem-Aflejringer i Danmark, Nord-Tyskland og Holland. Danmarks Geologiske Undersogelse Raekke II 17, 1–302. Menke, B., Tynni, R., 1984. Das Eeminterglazial und das Weichselfru¨hglazial von Rederstall/Dithmarschen und ihre Bedeutung fu¨r die mitteleuropa¨ische Jungpleistoza¨n-Gliederung. Geologisches Jahrbuch A 76, 3–120. Meyer, K.J., 1974. Pollenanalytische Untersuchungen und Jahresschichtenza¨hlungen an der holsteinzeitlichen Kieselgur von Hetendorf. Geologisches Jahrbuch A 21, 87–105. Mu¨ller, H., 1974a. Pollenanalytische Untersuchungen und Jahresschichtenza¨hlung an der eem-zeitlichen Kieselgur von Bispingen/Luhe. Geologisches Jahrbuch A 21, 149–169. Mu¨ller, H., 1974b. Pollenanalytische Untersuchungen und Jahresschichtenza¨hlungen an der holsteinzeitlichen Kieselgur von Munster-Breloh. Geologisches Jahrbuch A 21, 107–140. New, M.G., Hulme, M., Jones, P.D., 1999. Representing 20th century space-time climate variability. Journal of Climate 12, 829–856. NGRIP Members, 2004. High-resolution record of Northern Hemisphere climate extending into the last interglacial period. Nature 431, 147–151. Pons, A., Guiot, J., de Beaulieu, J.L., Reille, M., 1992. Recent contributions to the climatology of the last glacial–interglacial cycle based on French pollen sequences. Quaternary Science Reviews 11, 439–448. Pross, J., Klotz, S., Mosbrugger, V., 2000. Reconstructing palaeotemperatures for the Early and Middle Pleistocene using the mutual climatic range

method based on plant fossils. Quaternary Science Reviews 19, 1785–1799. Rioual, P., Andrieu-Ponel, V., Rietti-Shati, M., Battarbee, R.W., de Beaulieu, J.L., Cheddadi, R., Reille, M., Svobodova, H., Shemesh, A., 2001. Highresolution record of climate stability in France during the last interglacial period. Nature 413, 293–296. ¨ ber den Ru¨ckgang der HaselSamuelsson, G., 1915. U grenze und anderer pflanzengeographischer Grenzlinien in Skandinavien. Bulletin of Geology 13, 93–114. Sanchez-Gon˜i, M.F., Loutre, M.F., Crucifix, M., Peyron, O., Santos, L., Duprax, J., Malaize, B., Turon, J.-L., Peypouquet, J.-P., 2005. Increasing vegetation and climate gradient in Western Europe over the Last Glacial Inception (122-110 ka), datamodel comparison. Earth and Planetary Science Letters 231, 111–130. Shackleton, N.J., 1969. The last interglacial in the marine and terrestrial records. Proceedings of the Royal Society of London 174, 135–154. Shackleton, N.J., Sa´nchez-Gon˜i, M.F., Pailler, D., Lancelot, Y., 2003. Marine Isotope Substage 5e and the Eemian Interglacial. Global and Planetary Change 36, 151–155. Sinka, K.J., Atkinson, T.C., 1999. A mutual climatic range method for reconstructing palaeoclimate from plant remains. Journal of the Geological Society 156, 381–396. Turner, C., 2002. Problems of the duration of the Eemian interglacial in Europe North of the Alps. Quaternary Research 58, 45–48. Tzedakis, P.C., 2003. Timing and duration of last interglacial conditions in Europe: a chronicle of a changing chronology. Quaternary Science Reviews 22, 763–768. van Leeuwen, R.J.W., Beets, D.J., Bosch, J.H.A., Burger, A.W., Cleveringa, P., van Harten, D., Herngreen, G.F.W., Kruk, R.W., Langereis, C.G., M., T., Pouwer, R., de Wolf, H., 2000. Stratigraphy and integrated facies analysis of the Saalian and Eemian sediments in the Amsterdam-Terminal borehole, the Netherlands. Geologie en Mijnbouw 79, 161–196. Webb III, T., 1980. The reconstruction of climatic sequences from botanical data. Journal of Interdisciplinary History 10, 749–772. Zagwijn, W.H., 1961. Vegetation, climate and radiocarbon datings in the Late Pleistocene of the Netherlands, Part I: Eemian and Early Weichselian. Mededelingen van de Geologische Stichting Nieuwe Series 14, 15–45. Zagwijn, W.H., 1994. Reconstruction of climate change during the Holocene in western and central Europe based on pollen records of indicator species. Vegetation History and Archaeobotany 3, 65–88. Zagwijn, W.H., 1996. An analysis of Eemian climate in western and central Europe. Quaternary Science Reviews 15, 451–469.

17. Comparative Analysis of Vegetation and Climate Changes During the Eemian Interglacial in Central and Eastern Europe A.A. Velichko1, E.Y. Novenko1, E.M. Zelikson1, T. Boettger2 and F.W. Junge3 1

Institute of Geography RAS, Department of the Evolutionary Geography, Staromonetny 29, 119017 Moscow, Russia 2 UFZ Centre for Environmental Research Leipzig-Halle, Dept. of Isotope Hydrology, Th.-Lieser-Str. 4, D-06120 Halle, Germany 3 Saxon Academy of Sciences at Leipzig, Dynamics of Contaminants in Catchment Areas working group, Karl-Tauchnitz-Straße 1, D-04107 Leipzig, Germany

ABSTRACT The spatial–temporal landscape dynamics through the Eemian interglacial (including preceding and succeeding transitional phases) have been examined along a latitudinal transect (50–55 N). Three Eemian pollen diagrams are presented. As follows from comparison of the data from Central and Eastern Europe, changes of environment and climate became more contrasting from west to east. At the same time, the main phases in the evolution of vegetation appear to be similar throughout the latitudinal belt. The interglacial optimum was characterised by an essential similarity of vegetation all over the region investigated. Plant communities of the cooler intervals (at the beginning and closer to the end of the interglacial) differed noticeably from west to east. Significant contrasts in environmental and climatic fluctuations mark the Saalian/Eemian boundary (transition from MIS 6 to MIS 5e). Vegetation dynamics at this boundary resemble those detected at the transition from Weichselian to Holocene (Allero¨d and Younger Dryas). 17.1 INTRODUCTION Palaeogeographical reconstructions of warm intervals of the late Pleistocene increasingly attract attention of researchers because of their importance for a better understanding of modern environmental

processes and their anticipated changes. Successive phases of vegetation evolution inferred from pollen data have been traced from west to east along a latitudinal transect (50–55 N; Fig. 17.1). Palaeogeographical reconstructions are based on three key sections: Klinge (Fig. 17.2) in Central Europe and Cheremoshnik (Fig. 17.3) and Il’inskoye (Fig. 17.4) on the East European Plain (Velichko et al., 2005) and supplemented by published data on sites in Central and Eastern Europe showing evolution of vegetation through the last interglacial. Vegetation maps for four time slices of the Eemian have been compiled. The data obtained characterise the landscape evolution during the late Saalian (Dniepr) termination, the Eemian interglacial and the early phases of the Weichsel (Valdai) glacial stage. Analysis of the vegetation succession during the Saalian late glacial revealed some similarities between successions at the end of this glaciation and at the Weichselian late glacial/Holocene transition (Allero¨d and Younger Dryas). No less interesting are the climatic oscillations and vegetation changes at the Eemian–Weichselian transition. This observed structural similarity between the last interglacial and the Holocene provides us with an insight into climatic fluctuations that may be expected during the later phases of the Holocene. As the Eemian interval is beyond the range of radiocarbon dating, and only a

256

A.A. Velichko et al. 10°W

70° N

Artic

10°E

0°

20°

30°

Circle

40°

White Sea

Norwegian Sea

23

60°N

2 17 19

North

Baltic Sea

Sea 24

12 13 11

5 1 7 6

a

b

10 9

4

8

16

18

14

schemes can be compared with each other. Table 17.1 summarises the correlation between biozones from west to east within the latitudinal belt. We assume that the boundaries between biozones were not contemporaneous especially at the beginning and at the end of the interglacial; however existing data do not allow us to discuss this volume.

3

20

17.2 RECONSTRUCTION OF VEGETATION

21

15 22

Black Sea

Fig. 17.1 Location of Eemian sites (a) Sections, those pollen diagrams are presented in the paper: 1. Klinge (Velichko et al., 2005); 2. Cheremoshnik (Velichko et al., 2005); 3. Il’inskoye (Velichko et al., 2005); (b) Published data: 4. Gro¨bern (Litt, 1994); 5. Kittlitz (Erd, 1973); 6. Grabschu¨tz (Litt, 1994); 7. Scho¨nfeld (Erd, 1991); 8. Imbramowice (Mamakowa, 1989); 9. Kerkwitz (Erd, 1960); 10. Zbytki (Kuszell, 1997); 11. Rogaczewo (Kuszell, 1997); 12. Nakło (Norys´kiewicz, 1978); 13. Głowczyn (Niklewski, 1968); 14. Otapy (Bitner, 1956); 15. Horoszki (Granoszewski, 2003); 16. Yonensis-Maximonis (Kondratene, 1996); 17. Pushkari (Grichuk, 1961); 18. Murava (Tzapenko and Mahnach, 1959); 19. Nizhnyaya Boyarshchina (Grichuk, 1982); 20. Mikulino (Grichuk, 1961); 21. Tarasovo (Tzapenko and Mahnach, 1959); 22. Kolodiev (Gurtovaya, 1983); 23. Domanovo (Gei et al., 2000); 24. Rederstall (Menke and Tynni, 1984).

few sites have been dated by the uranium– thorium disequilibrium method, it seems reasonable that individual sections should be correlated on a biostratigraphic basis. There are a number of regional schemes of biostratigraphical zonation (Erd, 1973; Grichuk, 1982; Menke and Tynni, 1984; Mamakova, 1989, Table 17.1). As a great number of Eemian sites show the typical succession of trees, various zonation

17.2.1 The stage of late glacial vegetation Despite scarcity of palynological data for the end of Saalian (Dniepr) glacial epoch (the lowermost part of the Klinge and Cheremoshnik profiles, Fig. 17.2, 17.3), we can nevertheless suppose that at this time Central and Eastern Europe were occupied by a complex vegetation, which included open birch and pine forests, cold-tolerant shrub communities (with Betula nana, B. humilis), meadows and bogs. The transition to interglacial conditions encouraged the spread of forest in Central and Eastern Europe. The role of shrub communities in the landscape increased significantly. Elements of the periglacial vegetation and communities with heliophytes (Helianthemum, Hippophae rhamnoides) probably survived in favourable habitats. Noteworthy is the appearance of aquatic plants, known to be for their rather thermophilous. The late glacial plant succession suggests rather unstable climatic conditions. Two substages of vegetation development can be identified. The composition of pollen assemblages obtained for the lower part of the Klinge profile (Velichko et al., 2005) and of previously published sections in east Germany and Poland (e.g. Kittlitz: Erd, 1973; Gro¨bern: Litt, 1994; Imbramowice: Mamakowa, 1989; Zbytki: Kuszell, 1997) show that pine forest dominated in

0

Biostratigraphic zone by Menke and Tynni,1984

LPAZ

Total pollen concentration

257

AP/NAP/Sporas

Sphagnum

Polypodiaceae

Hedera Viscum Salix Alnus: Alnaster-type Betula humilis Betula nana Hippophae rhamnoides Artemisia Chenopodiaceae Cyperaceae Ericales Helianthemum Poaceae

Corylus

Alnus

Ilex Taxus

Carpinus

Quercus Ulmus Tilia cordata

Betula sect. Albae

Pinus sylvestris

Abies Picea Pinus cembra

Comparative Analysis of Vegetation and Climate Changes

20 40 60

7 E7

80 100 120 140 160 180

6

200

E6

220

5b

240 260

E5

5a

280 300

4 E4b

320 340

3 E4a

360 380

Gap in sedimentation

400 420

2 E1 1b 1a

440 460 480

lateglacial

500

S

520 540 10 20 10

20 40 60 20 40 60 80

20 10 10 20 40 60

10

20 40 20 40 60

10 10 10 10

10 1010 10

10

20 40 10

20 40 60 80 100 2000 2 4000

2

10 grains per cm .

1

2

3

4

5

6

7

8

9

Fig. 17.2 Pollen diagram of the Klinge section. Pollen sum ¼ AP þ NAP þ Spores. Clear curves represent 10 exaggeration of base curves, ‘þ’ – redeposited taxa, ‘‘.’’ – presence of taxa under 2%. Lithology; 1 – silt; 2 – laminated silt; 3 – peat with clay; 4 – peat; 5 – mud; 6 – silty mud; 7 – sandy silt; 8 – silty clay; 9 – silt with clay.

the earlier part of this period. Then, birch forest and herb communities (with Artemisia, Chenopodiaceae and Poaceae) expanded over the area. In late glacial pollen assemblages in sites, located in eastern Poland (e.g. Otapy: Bitner, 1956; Horoszki: Granoszewski, 2003) and Byelorussia (Murava, Tarasovo: Tzapenko and Mahnach, 1959; Savchenko and Pavlovskaya, 1999) at similar latitudes, spruce plays a noticeable role alongside pine. Farther to the east (Mikulino: Grichuk, 1961), a Picea maximum is typical of pollen diagrams from

this time interval all over the East European Plain (the so-called ‘lower spruce maximum’). Thus, spruce woodlands were widespread in the eastern part of the territory under consideration, whereas pine forests occupied the western region. Climatic and vegetation dynamics at the glacial/interglacial transition show greater contrast in the eastern part of the latitudinal belt. Two phases of vegetation can clearly be determined in the pollen diagram from the Cheremoshnik section in the Upper Volga

Depth, cm

1 Pinus sylvestris

20 40 60 10 20 40

20 40 60

2

20 40 60

20 10

Tilia platyphyllos

20 40

20

20 40 60

20

20 20

Osmunda cinnamomea

4 M3

3 M2

2 1 M1

140

160

180

200

220

240

260

280

300

380 Biostratigraphic zone by V.P. Grichuk,1961

6

LPAZ

AP/NAP/Spores

10

Sphagnum

20 40 60 80 10

Polypodiaceae

10 10 10

Poaceae

LPAZ Biostratigraphics zone by V.P. Grichuk, 1961

Algae

AP/NAP

Pteredium aqilinum Selaginella selaginoides Sphagnum Nymphaea Nuphar Typha latifolia

Spores

Ranunculaceae

20 40 10

Polypodiaceae

Cyperaceae Dipsacaceae Ephedra Helianthemum Polygonaceae Ranunculaceae Rosaceae Thalictrum Osmunda Botrychium lunaria Lycopodium annotinum Lycopodium selago

NAP

Cyperaceae Ephedra Ericales

10

Poaceae

Viburnum Salix Alnus: Alnaster type Betula humilis Betula nana Juniperus Humulus lupulus Artemisia Chenopodiaceae

Corylus

Alnus

Ulmus Tilia cordata Tilia platyphyllos Carpinus Fraxinus

Quercus

Pinus sibirica Larix Betula sect. Albae Acer

Pinus sylvestris

Abies Picea

AP

Viscum album Viburnum Salix Alnus: Alnaster type Betula humilis Betula nana Artemisia

20

Corylus avellana

10 10

Alnus

10

Fraxinus exelsior

20 40

Carpinus betulus

20 40

Ulmus

2 10 10

Tilia cordata

1 20 10

Quercus

20 40 60

Acer

20

Betula sect. Albae

20 40 60

Larix

20 40

Pinus sect. Cembra

150 160 170 180 190 200 210 220 230 240 250 260 270 280 290 300 310 320 330 340 350 360 370 380

Picea

Lithology

Depth, cm

258 A.A. Velichko et al.

7 M7

Gap in sedimentation

5a M4

20 40 60 80100 10

lateglacial

3

Fig. 17.3 Pollen diagram of the Cheremoshnik section. Pollen sum ¼ AP þ NAP, ‘‘.’’ – presence of taxa under 2%. Lithology: 1 – peat; 2 – gyttja; 3 – till.

6 M8

5 M7

4 M6

3 M5

320

2

M4

340

360

1 M2-3

400

20 20 20 40 60 80100

3

Fig. 17.4 Pollen diagram of the Il’inskoye section. Pollen sum ¼ AP þ NAP þ Spores, ‘‘.’’ – presence of taxa under 2%. Lithology: 1 – loam; 2 – peat; 3 – gyttja.

Comparative Analysis of Vegetation and Climate Changes

259

Table 17.1 Eemian biostratigraphic correlation B. Menke and R. Tynni (1984)

K. Erd (1973)

K. Mamakowa (1989)

V.P. Grichuk (1961, 1982)

E7 Pinus E6 Pinus-Picea-Abies

E9 Pinus-Betula E8 Pinus-Picea-AlnusAbies E7 Carpinus-Alnus-Picea (Abies) E6 Carpinus-Taxus-TiliaAlnus (Picea) E5b Corylus-Taxus-TiliaAlnus E5a Corylus-Taxus-QuercusAlnus E4 Corylus-Quercus-AlnusPinus E3 Pinus-Quercus-BetulaUlmus E2 Pinus-Betula -Ulmus

E7 Pinus E6 Pinus-PiceaAbies E5 CarpinusPicea-Alnus

M8 Pinus-Picea-Betula M7 Picea

E4 CorylusQuercus-Tilia

M5 Tilia-Quercus-UlmusCorylus M4 Quercus-UlmusCorylus

E5 Carpinus-Picea

E4b Corylus-TaxusTilia E4a Quercetum mixtum-Corylus

E3 Pinus-Quercetum mixtum E2 Pinus-Betula E1 Betula Saalian late-glacial

E1 Betula-Pinus

region (Fig. 17.3). Spruce forest with shrubs and elements of periglacial-steppe vegetation occurred in the earliest phase. Then birch forest occupied the territory. The role of bushes (Betula nana, B. humilis, Alnus (Alnaster) virides ssp. fruticosus) and steppe-like communities (with Poaceae, Ephedra, Artemisia, Chenopodiaceae) increase significantly. 17.2.2 The Eemian Interglacial 17.2.2.1 The stage of birch and pine–birch forests A rapid afforestation of the area reflects a dramatic change in the regional vegetation caused by the warming at the beginning of the Eemian. The vegetation changes appear to be similar all over the latitudinal belt. Dense forests spread over the region under consideration (first pine and later birch). Oak, elm, hazel and alder gradually penetrated the forest communities during this period, but in eastern regions thermophilous plants appeared later – by the end of that time. After that forest communities with broad-leaved trees dominated the plant cover.

E3 QuercusFraxinus-Ulmus E2 Pinus-Betula Ulmus E1 Pinus-Betula

M6 Carpinus

M3 Pinus-Betula-(QuercusUlmus- Corylus) M2 Betula M1 Picea

17.2.2.2 The stage of coniferous – broadleaved forests (Fig. 17.5a) Pine-oak forests with broad-leaved trees (Ulmus, Tilia, Acer and Fraxinus) and hazel communities expanded widely in Central Europe. Broad-leaved forests (dominated by Quercus and Ulmus) and hazel communities spread over the East European Plain. The appearance of Hedera helix and Viscum album indicates warming and increasing humidity. Hedera pollen was recorded in Eemian deposits in Lithuania (Kondratene, 1996) and the Precarpathians (Gurtovaya, 1983). Viscum album penetrated the central region of the East European Plain (e.g. Il’inskoye section Fig. 17.4, Velichko et al., 2005). 17.2.2.3 The stage of mixed broad-leaved forests (Fig. 17.5b) Forests formed by lime (Tilia cordata and Tilia platyphyllos), oak and hornbeam occupied the area. Important components were Ulmus, Fraxinus and Acer. Hazel continued to play a cospicuous role in the prevailing communities. The interglacial optimum was

260

A.A. Velichko et al. 10°E

(a)

30°

20°

60°N

(c)

10°E

30°

20°

60°N

North Sea

North Sea

Baltic Sea

50°N

Baltic Sea

50°N

1

2

5

4

3

1

2

Pollen zone E5

Pollen zone E4a 30°

20°

10°E

(b) 60°N

4

3

10°E

(d)

30°

20°

60°N

North Sea

North Sea

Baltic Sea

50°N

Baltic Sea

50°N

1

2

3

4

Pollen zone E4b

1

2

3

4

Pollen zone E6

Fig. 17.5 Reconstruction of the vegetation along the latitudinal profile for various time slices in the Eemian (Biostratigraphical units by Menke and Tynni, 1984). (a) The stage of coniferous – broad-leaved forests (pollen zone E4a). 1 – pine-oak forest with broad-leaved trees (Ulmus, Tilia, Acer and Fraxinus); 2 – broadleaved forest (dominated by Quercus and Ulmus); 3 – the easternmost location of Viscum album pollen; 4 – the easternmost location of Hedera helix pollen; 5 – location of Eemian sites. (b) The stage of mixed broad-leaved forests (pollen zone E4b). 1 – broad-leaved forest formed by lime and yew with admixture of oak and hornbeam; 2 – broad-leaved forest formed by lime and hornbeam with admixture of oak and swamped forest of alder (Alnus glutinosa); 3 – broad-leaved forest formed by lime, oak and elm; 4 – the easternmost location of Ilex aquifolium pollen. (c) The stage of hornbeam forests (pollen zone E5). 1 – hornbeam forest with lime, yew and spruce; 2 – hornbeam forest with lime and oak and swamp forest of alder (Alnus glutinosa); 3 – Hornbeam forest with lime and oak; 4 – mixed broad-leaved forest with spruce. (d) The stage of spruce and pine forests with hornbeam (pollen zone E6). 1 – spruce and pine forest with hornbeam, oak and fir; 2 – spruce and pine forest with hornbeam, fir and alder; 3 – spruce and pine forest with admixture of broad-leaved trees; 4 – spruce and pine-birch forest.

marked by essentially similar vegetation over the latitudinal transect. Nevertheless, differentiation between Central and East Europe took place. In western regions, yew (Taxus baccata) was one of the significant taxa in forest communities. Exceptionally high pollen values of Taxus baccata at a number of sections in Central Europe

suggest that in the interglacial optimum it might form communities under the canopy of other trees, similar to yew forest in southern England nowadays. In this case, some researchers distinguish a Taxus pollen zone (Jung et al., 1972) or a CorylusTaxus-Tilia one (Menke and Tynni, 1984) in biostratigraphic schemes for the Eemian.

Comparative Analysis of Vegetation and Climate Changes

The important components of the vegetation in Poland were communities of Alnus incana on wet soils and swamp forests of A. glutinosa. During the period under consideration, thermophilous species such as Taxus baccata and Ilex aquifolium advanced far to the east compared to their modern range. Pollen and macrofossils of holly (a species restricted to West Europe at present) have been found in eastern Germany (Erd, 1960, 1973; Litt, 1994) and Poland (Mamakowa, 1989; Kuszell, 1997). There is evidence for Ilex aquifolium having occurred as far as the Precarpathians (Kolodiev section; Gurtovaya, 1983). Broadleaved trees, noted for their rather high requirement of humidity – Tilia platyphyllos and Carpinus betulus, appeared in forest communities of the eastern region considerably later than in the west. Therefore, the Carpinus peak in pollen diagrams in eastern regions (see Il’inskoye site, Fig. 17.4) is short. It should be noted that these species are absent in the modern flora of the eastern part of the investigated territory. 17.2.2.4 The stage of hornbeam forests (Fig. 17.5c) Hornbeam became dominant in zonal forest formations during the second half of the interglacial, while at its end spruce and fir (only in the west) were also present in the forest communities. Broad-leaved trees (Quercus, Tilia, Ulmus, Fraxinus, as well as Ilex and Taxus) persisted in the forest, but were not abundant. 17.2.2.5 The stage of spruce and pine forests with hornbeam (Fig. 17.5d) This stage is characterised by a reduction of forest area. Alongside dense forest, open woodland appeared. The role of broadleaved trees fell substantially. Pollen diagrams reflect provincial differentiation of plant cover during this stage. Pollen spectra of sediments in the western region are

261

distinguished by the presence of Abies and relatively high percentages of thermophilous trees; pollen assemblages from eastern sections lack these characteristics. The conspicuous maximum of spruce pollen is typical of the end of the Eemian in the pollen sequences all over the East European Plain (the ‘upper spruce maximum’). 17.2.2.6 The stage of open pine forests The final stage of the interglacial was marked by pine and birch forests with admixture of spruce and fir (in the west) and some steppe-like areas (with Ephedra and Artemisia), shrub communities, wet meadows (with Selaginella selaginoides and Lycopodium pungens) and wetlands. The role of cold-tolerant shrubs (dwarf birch and green alder) in the plant cover increased with the onset of colder climate. On the whole, forests became more open, than they were at the beginning of this stage, as indicated by the decrease in tree pollen content and by higher frequencies of herbaceous pollen in the assemblages (Artemisia, Poaceae and Chenopodiaceae). 17.2.3 Vegetation at the beginning of the early Weichselian (Valdai) glacial stage Unlike the late glacial of the Saalian glacial stage, which was distinguished by noticeable changes in the plant cover, vegetation evolution at the boundary between the interglacial and glacial epochs in Central Europe proceeded gradually. Pollen sequences where the final phases of the interglacial are present, e.g., Rederstall (Menke and Tynni, 1984), Kittlitz (Erd, 1973) and Klinge, show that the end of the Eemian cycle was characterised by gradual changes of vegetation. In the Gro¨bern profile, a brief warming phase at the end of the interglacial is reflected by stable isotope data (Boettger et al., 2000). While woodlands persisted in western regions, open periglacial vegetation spread over eastern ones.

262

A.A. Velichko et al.

Similar features of vegetation dynamics are recognised in the eastern part of the transect. Pollen records from the Mikulino section (stratotype for the last interglacial) enabled Grichuk (1961) to distinguish the first Early Valdai interstadial (Upper Volga interstadial) despite some gaps in sedimentation at the transition. In pollen assemblages of this interstadial, Betula pollen reaches its maximum, Picea and Pinus are also abundant. The continuous pollen sequence from the second half of the Eemian/Mikulino interglacial to the onset of the Early Valdai glacial stage has been found in the Pushkari section in the Vitebsk region (Grichuk, 1961) and in Domanovo section in the Vologda area (Gei et al., 2000). The last zone of the Mikulino interglacial is in the pollen diagrams followed by a phase marked by a significant peak of Betula sect. Nanae pollen against the background of decreasing tree pollen percentages. Still higher in the diagrams the Early Valdai warming phase occurs. The lower boundary of the interstadial is identified by relatively high values of pollen of pine and tree birch. Composition of the pollen assemblages of the next phase suggests a transition to cooling.

Weichsel the climate became colder and more continental (Klinge: TI ¼ 10 , TVII ¼ þ16 ). Comparison of the reconstructed climatic parameters with those of today indicates a larger positive deviation of winter than of summer temperatures. As for winter temperatures, their positive deviations were larger in the east (Il’inskoye section) than in the west (Klinge) – 10 C and 2 C respectively. Summer (July) temperature deviations were no more than 1 C both in the east and in the west. At the Eemian (Mikulino interglacial) optimum, the latitudinal gradient of temperatures was considerably reduced. Continentality of climate in the east of the continent was much less as a result of (a) 20 Klinge t° C

II’inskoye

16

12

Present value: Klinge TVII = 18° C II’inskoye TVII = 19° C

4

17.3 CLIMATIC RECONSTRUCTIONS Klinge

Climatic characteristics (mean temperatures of the warmest and coldest months) for the Eemian were determined using the floristic method of palaeoclimatic reconstruction – the method of climagrams that was developed in the Laboratory of Evolutionary Geography IG RAS (Grichuk, 1985). Curves of mean temperatures of the coldest and warmest months – January and July (Fig. 17.6) – show that, obviously, the heat supply rose gradually from the beginning of the interglacial (Klinge: TI ¼ 2 , TVII ¼ þ18 , Il’inskoye: TI ¼ 6 , TVII ¼ þ18 ) to its middle part, reaching maximal values (Klinge: TI ¼ þ2 , TVII ¼ þ19 , Il’inskoye: TI ¼ 0 , TVII ¼ þ19 ). Then the temperature (curve) decreased, and with the onset of the

0 t° C

II’inskoye

–4

–8 Present value: Klinge TI = 0° C II’inskoye TI = –10° C –12 E2

E3+ E4b E5 E6 E7 E4a Biostratigtaphical subdivisions by Menke & Tynni (1984).

Fig. 17.6 Temperature reconstruction of the Klinge and Il’inskoye sections. (a) Reconstruction of the mean temperature of the warmest month. (b) Reconstruction of the mean temperature of the coldest month.

Comparative Analysis of Vegetation and Climate Changes

evident penetration of oceanic influences farther eastwards. This accounts for the fact that summer temperatures differed insignificantly from modern ones all over the transect, while winter temperatures did not drop below 0 C anywhere. The interglacial optimum was marked by increasing precipitation. In the eastern region, the deviation of precipitation was not more the 100 mm while that in the western part of transect reached 300 mm (Velichko et al., 1991). The presented climatic reconstructions are in high agreement with the findings of other studies of the Eemian (Velichko et al., 1991; Zagwijn, 1996; Aalbersberg and Litt, 1998). 17.4 CONCLUSIONS Comparison of palynological materials from Central and Eastern Europe reveals that changes of environment and climate became more contrasting from west to east. At the same time, the main phases in the evolution of vegetation appear to be similar throughout the latitudinal belt under consideration. The interglacial optimum was marked by an essential similarity of vegetation all over the region investigated. Nevertheless, a floristic provincial differentiation is detectable. Mixed broad-leaved forests in Central Europe included species that require a certain oceanicity of climate (Ilex aquifolium, Hedera helix, Taxus baccata, etc). The participation of these plants decreases eastward. Of those species, only Tilia platyphyllos and Viscum album are found in the Eemian pollen assemblages in the eastern part of the transect. On the other hand, plant communities of the cooler intervals (at the beginning and closer to the end of the interglacial) differed noticeably from west to east, primarily in the proportion of broad-leaved species in zonal vegetation formations. Significant contrasts in environmental and climatic fluctuations mark the Saalian/ Eemian boundary (transition from MIS 6 to MIS 5e). Vegetation dynamics at this boundary resemble those detected at the transition

263

from Weichselian to Holocene (Allero¨d and Younger Dryas).

REFERENCES Aalbersberg, G. & Litt, T., 1998. Multiproxy climate reconstructions for the Eemian and early Weichselian. Journal of Quaternary Science 13, 367–390. Bitner, K., 1956. Flora interglacjalna w Otapach. Biuletyn Instytutu Geologicznego 100, 61–142. Boettger, T., Junge, F. W. & Litt, Th., 2000. Stable isotope conditions in central Germany during the last interglacial. Journal of Quaternary Science 15, 469–473. Erd, K., 1960. Das Eem-Interglazial von KerkwitzAtterwasch bei Guben. Wissenschaftliche Zeitschrift der pa¨dagogischen Hochschule Potsdam, Mathematisch-naturwissenschaftliche Reihe 6, 1–2, 107–118. Erd, K., 1973. Pollenanalytische Untersuchungen im Pleistoza¨n der DDR. Abhandlungen des Zentralen Geologischen Institutes Berlin 18, 1–7. Erd, K., 1991. Vegetationsentwicklung und Pollenanalysen im Eeminterglazial und Weichsel-Fru¨hglazial von Scho¨nfeld, Kreis Calau. Natur und Landschaft in der Niederlausitz 1, 71–81. Gei, V.P., Pleshivtzeva, E.S. & Auslinder, V.G., 2000. Stratigraphy. In: Zarrina, T.P. and Shik, S.M. (Eds.), Problems of stratigraphy of Quaternary deposits and marginal till formation on Vologda region (North-West of Russia), Proceedings of the International Symposium GEOS, Moscow, 31–61. Granoszewski, W., 2003. Late Pleistocene vegetation history and climate at Horoszki Duz´e, Eastern Poland: a palaeobotanical study. Acta Paleobotanica. Supplementtum 4. 3–95. Grichuk, V.P., 1961. Fossil floras as a paleontological basis of Quaternary stratigraphy. In: Markov K.K. (Ed.), Relief and stratigraphy of Quaternary deposits on the north-west of Russian Plain. USSR Academy of Sciences Press, Moscow, 25–71. Grichuk, V.P., 1982. Vegetation of Europe during Late Pleistocene. In: Gerasimov, I.P. and Velichko, A.A. (Eds.), Paleogeography of Europe during the last one hundred thousand years. Nauka, Moscow, 79–85. Grichuk, V.P., 1985. Reconstruktsia skaliarnykh klimaticheskikh pokasatelei po floristicheskim materialam i otsenka ee tochnosti. In: Gerasimov, I.P. and Velichko, A.A. (Eds.), Metody reconstruktsii paleoklimatov. Nauka, Moscow, 20–28.

264

A.A. Velichko et al.

Gurtovaya, E.E., 1983. Flora and vegetation in the east of Middle Europe during Mikulino Interglacial. USSR Academy of Sciences, Izvestiya, Seria Geografija 4, 78–86. Jung, W., Beug, H.J. & Dehm, R., 1972. Das Riss/ Wu¨rm-Interglazial von Zeifen, LandkreisLaufen a.d. Salzach. Bayerische Akademie der Wissenschaften, mathematisch-naturwissenschaftliche Klasse, Abhandlungen, Neue Folge 151, 1–131. Kondratene, O., 1996. Stratigraphy and paleogeography of Quaternary in Lithuania. Academia-Press, Vilnius, 214 pp. Kuszell, T., 1997. Palinostratygrafia osadow interglacjału eemskiego i wczesnego vistulianu w południowej Wielkopolsce i na Dolnym S´lasku. Wydawnictwo Uniwersytetu Wrocławskiego, Wrocław, 69 pp. Litt, T., 1994. Pala¨oo¨kologie, Pala¨obotanik und Stratigraphie des Jungquarta¨rs im nordmitteleuropa¨ischen Tiefland (unter besonderer Beru¨cksichtigung des Elbe-Saale-Gebietes). Dissertationes Botanicae 227, 1–185. Mamakowa, K., 1989. Late Middle Polish Glaciation, Eemian and Early Vistulian vegetation at Imbramowice near Wroclaw and the pollen stratigraphy of this part of the Pleistocene in Poland. Acta Paleobotanica 29, 1, 11–179. Menke, B. & Tynni, R., 1984. Das Eeminterglazial und das Weichselfru¨hglazial von Rederstall/

Dithmarschen und ihre Bedeutung fu¨r die mitteleuropa¨ische Jungpleistoza¨n-Gliederung. Geologisches Jahrbuch A 76, 3–120. Niklewski, R., 1968. Interglacjal Eemski w Nakle nad Notecia. Acta Paleobotanica 19, 67–112. Norys´kiewicz, K., 1978. Interglacjał Eemski w Gło´wczynie koło Wyszogrodu. Monographiae Botanicae 27, 125–191. Savchenko, I.E. & Pavlovskaya, I.E. 1999. Muravian (Eemian) and Early Poozerian deposits at Azarichi (Eastern Belarus). Acta Paleobotanica. Supplementtum 2. 523–527. Tzapenko, M.M. & Mahnach, H.A., 1959. Quaternary sediments of Byelorussia. Isdatelstwo Academii Nauk BSSR, Minsk, 225 pp. Velichko, A.A., Borisova, O.K., Gurtovaya, E.E. & Zelikson, E.M., 1991. Climatic rhythm of the Last Interglacial in Northern Eurasia. Quaternary International 10–12, 191–213. Velichko, A.A., Novenko, E.Y., Pisareva, V.V., Zelikson, E.M., Boettger, T. & Junge, F.W., 2005. Vegetation and climate changes during the Eemian Interglacial in Central and East Europe: comparative analysis of pollen data. Boreas 34, 207–219. Zagwijn, W.H., 1996. An analysis of Eemian climate in Western and Central Europe. Quaternary Science Reviews 15, 451–469.

18. Indications of Short-Term Climate Warming at the Very End of the Eemian in Terrestrial Records of Central and Eastern Europe T. Boettger1, F.W. Junge2, S. Knetsch1,2, E.Y. Novenko3, O.K. Borisova3, K.V. Kremenetski3,4 and A.A. Velichko3 1

Department of Isotope Hydrology, UFZ Centre for Environmental Research Leipzig-Halle, Theodor-Lieser-Str. 4, D-06120 Halle, Germany 2 Research group ‘‘Pollutants dynamics in catchment areas’’, Saxon Academy of Sciences at Leipzig, Karl-Tauchnitz-Str. 1, D-04107 Leipzig, Germany 3 Laboratory of Evolutionary Geography, Institute of Geography, Russian Academy of Sciences, Staromonetny 29, 109017 Moscow, Russia 4 Department of Geography, University of California, 1255 Bunche Hall, Los Angeles, CA 90095 1524, USA

ABSTRACT Geochemical and palynological studies of lacustrine sediments from the standard Eemian–Early Weichselian profiles Gro¨bern, Neumark-Nord and Klinge (Germany, Central Europe) document at least two warming events during the transition from the Eemian to the Early Weichselian. The first pronounced warming phase takes place towards the very end of the Eemian Interglacial during pollen assemblage zone E7, just before the actual transition into the Weichselian Glacial period. Its amplitude is not on the scale of the Eemian climatic optimum, but is comparable with the conditions found in the first Early Weichselian Interstadial (Bro¨rup). An additional event of climatic amelioration was detected within the period of the first Weichselian Stadial (Herning). In the high-resolution Eemian–Early Weichselian limnic sequence from Ples in the Upper Volga region (Russia, Eastern Europe), we also found indications of climate warming events at the very end of the Eemian during pollen assemblage zone E7 and within the Herning Stadial recorded both in palynological and in geochemical records. Furthermore, the 18 O

results of the new Greenland ice core presented by the North Greenland Ice Core Project (NGRIP) members record ‘a hitherto unrecognised warm period initiated by an abrupt climate warming about 115 000 years ago (towards the end of the Last Interglacial), before glacial conditions were fully developed’. In this paper, we discuss possible correlations between our terrestrial results in Central and Eastern Europe and their possible connection to the NGRIP record. It appears that both in Central and Eastern Europe and in Greenland, warming phases towards the end of the Last Interglacial preceded the final transition to glacial conditions. Thus, natural warming episodes during the end of the Last Interglacial appear to be a global phenomenon for the Northern Hemisphere. 18.1 INTRODUCTION The Quaternary large-scale climate is marked by changes between glacial and interglacial intervals. Periods of global transitions from warm to cold or from cold to warm conditions are unstable and characterised by mid- and short-term climate variations.

266

T. Boettger et al.

For analysis of large-scale transitions from cold to warm, the well-studied change from Late Weichselian to Holocene is of particular importance. This rapid termination, only a few millennia long period comprised a number of Late Glacial phases: Meiendorf-Oldest Dryas-BøllingOlder Dryas-Allerød-Younger Dryas (e.g. Firbas, 1949; Iversen, 1954; Hughen et al., 1996). The climatic change from Late Weichselian to Holocene is well known through numerous multiproxy studies of Central European profiles (Eicher and Siegenthaler, 1976; Eicher et al., 1981, 1991; Boettger et al., 1998, 2004; Scharf et al., 2005). The transition from Saalian Glacial to Last Interglacial (Eemian) shows similar climatic phases as the transition from Weichselian Glacial to Holocene (Seidenkrantz and Knudsen, 1994; Novenko et al., 2005). On the other hand, there are only a few well-studied continental profiles, which document the change from warm (Interglacial) to cold (Early Glacial) stages with its mid- and short-term climate stages. For the transition from Eemian to Early Weichselian in Central Europe, the standard profile Gro¨bern in central Germany (Wansa and Wimmer, 1990; Eissmann, 1990, 1994) is well known. This sequence allows highresolution study of both the full Eemian Interglacial and the Early Weichselian with its stadial–interstadial changes (HerningBro¨rup-Schalkholz-Odderade). Furthermore, the results of the Gro¨bern profile document a pronounced warming phase towards the very end of the Eemian, before the actual transition to glacial conditions (Boettger et al., 2000). This can be classified as a first sign for increasing climatic instability within the transition from interglacial to glacial. Also, the results from a new Greenland ice core presented by the North Greenland Ice Core Project members (NGRIP members, 2004) ‘reveals a hitherto unrecognised warm period initiated by an abrupt climate warming about 115 000 years ago (towards the

end of the Last Interglacial), before glacial conditions were fully developed’. The situation at the very end of the Eemian can be compared with the currently observable global warming and is therefore very important. On the other hand, various simulations using climate models and strong linear correlations with the insolation signal lead to the assumption that the current interglacial climate could persist much longer, even up to 50 000 years (Loutre and Berger, 2003; Berger and Loutre, 2005). The main objective of this study was to contribute to a better understanding of climatic variability in the final part of Eemian in its spatial and temporal distribution by means of selected European continental profiles. It is under discussion whether the observed few short warming phases during the change from interglacial to glacial are only of a local nature or whether they could be interpreted as regional or global events.

18.2 RESULTS AND DISCUSSION 18.2.1. Limnic sequences in Central and Eastern Europe Because of its large number of open-cast mines, central Germany is one of the best explored areas in Europe when it comes to Cenozoic sediments. Many Eemian and Holsteinian limnic profiles on top of Saalian or Elsterian tills are known here. Interglacial profiles discussed in this paper (Fig. 18.1) – Gro¨bern (Wansa and Wimmer, 1990), Neumark-Nord (Eissmann, 2002) from the Saale-Elbe district, Klinge (Striegler, 1986) from Niederlausitz (Lower Lusatia) and Ples (Grichuk and Grichuk, 1959) from the region Upper Volga in Russia – have developed on top of the youngest moraines of the Saalian glaciation. Their stratigraphic position clearly indicates a Last Interglacial age. This is the basis for a spatial comparison of climate dynamics during the transition from Eemian to Early Weichselian attempted in this study.

Indications of Short-Term Climate Warming

267

1

5

2

4

3

Reference sites ( ): 1 NorthGRIP 2 Gröbern

Investigated sites ( ): 3 Neumark-Nord 4 Klinge 5 Ples

Fig. 18.1 Location of investigated limnic profiles in Central and Eastern Europe and of the reference site NGRIP ice core.

The applied methods have previously been discussed in Boettger and Junge (1994) and Boettger et al. (1998) – stable isotopes, and in Velichko et al. (2005) – palynology. 18.2.1.1 Central Europe (Gro¨bern, Neumark-Nord, Klinge) . Gro¨bern The Eemian–Early Weichselian limnic profile Gro¨bern is exposed in an open-cast lignite mine (Wansa and Wimmer, 1990). Numerous geochemical (Boettger et al., 2000), palynological (Litt, 1990, 1994; Litt et al., 1996) and faunal (Walkling and Coope, 1996) studies make it one of the best-studied Last Interglacial sequences in Germany. The bowl-shaped sequence of lake sediments starts on top of the Saalian till. The low part is represented by an unlaminated silt of Late Saalian age. Concordant above it, the up to 4-m thick limnic Eemian sequence follows. This can be subdivided into seven pollen assemblage zones (PAZ) and is

lithologically characterised by closely alternating finely laminated carbonates and coarse or fine detrital muds. This is followed by approximately 7 m of Early Weichselian sediments: coarse clay muds within the stadials and organogenic sediments (calcareous muds, coarse and fine detrital muds and peat) in the interstadials. This sequence includes the Herning stadial as the first cold period after the Eemian, the first forested interstadial Bro¨rup, followed by the sediments of the Rederstall stadial and the Odderade interstadial and partly of the Schalkholz stadial (Wansa and Wimmer, 1990). The comprehensive results of complete high-resolution isotope geochemical studies of the Gro¨bern sequence are presented in Boettger et al. (2000). For Fig. 18.2, we selected the 18 O curve of carbonate as the most representative proxy for variation of temperature. It should be noted that in both the Eemian and the Early Weichselian stages of profile Gro¨bern, a close correlation was observed between the curves recording the

268

T. Boettger et al.

Fig. 18.2 Correlation of data from Central European profile Gro¨bern, the East European profile Ples and NGRIP ice core (PAZ*: after Menke and Tynni, 1984; IR-OSL data: after Degering and Krbetschek, this volume).

carbonate level, its isotopic values ð13 C, 18 OÞ and the content and 13 C value of total organic carbon in the sediment. Within PAZ E7 (Fig. 18.2), at the very end of Eemian and just before the complete change to glacial conditions to the Early Weichselian, a brief phase of warming is also indicated by geochemical indicators in several samples. This is reflected in the simultaneous increase of contents and stable isotope values of both lake carbonates and organic sediments. This warming followed a continuous temperature decrease from the middle of PAZ E6b, which had set in after the climatic optimum (PAZ E4b/5). Its amplitude is not on the scale of the Eemian climatic optimum, but is instead comparable to the conditions found in the first Early Weichselian Interstadial. Furthermore, within the first Weichselian stadial (Herning, PAZ WFI), we observe evident signs of additional climatic amelioration. This warming event can be recognised both in a renewed increase in limnic carbonate production and in increased 18 O values (Fig. 18.2) of carbonate (Boettger et al., 2000).

There is no palynological evidence for these climatic fluctuations. This may be due to the slower or insufficient reaction of the vegetation to short-term warming phases – in contrast to the immediate reaction to changes in temperature of isotopic fractionation during the course of carbonate precipitation, a purely chemical process. On the other hand, a shorter duration of the vegetation period, in spite of the warming, could be the cause for insufficient plant growth. . Neumark-Nord The Eemian profile Neumark-Nord (Eissmann, 2002) is exposed in the Geiseltal open-cast lignite mine not far from the town Merseburg (Saxony-Anhalt). In order to improve the profile resolution, five subprofiles were sampled over a distance of about 200 m at the middle of the basin in areas where each subprofile is the most extensive. Thus, the comprehensive thickness of the studied interglacial sequence is about 10 m. Above the Saalian till this limnic sequence is characterised in its lower part by coarse clay with very few molluscs (PAZ

Polypodiaceae

Poaceae

Artemisia Chenopodiaceae Cyperaceae

Ulmus

269

Picea Abies Quercus

60

Carpinus

40

Corylus

20

Betula

Corg, wt. % 0

0

Alnus

Pinus

Indications of Short-Term Climate Warming

AP/NAP/Spores

0 N8

E7

1

1

2

2

3

3

4

4

5

5

6

6

E4

7

7

E2

8

N7

N6b

E6

E5

Depth, m

N6a

N5

N4 N3 N2

C org

8

N org E1 Saalian

PAZ*

Age IR-OSL (kyr) 132 ± 12

9

144 ± 14

10 0

20 40 60

80 100 0

CaCO3, wt. %

N1

9

1 2 Norg, wt. %

3

10

2040 60 80

20 20 40 60

20 40 60

20 40 60 80

20

20 20 20

20 20

20 40 60 80 Local PAZ

Fig. 18.3 Neumark-Nord. New IR-OSL data, carbonate and organic matter content (C, N) and palynological findings (PAZ*: after Menke and Tynni, 1984).

E1 to E2; local PAZ N1 to N2; Fig. 18.3), organic and calcareous silts in the middle of profile (PAZ E4 to lower part of E6; local PAZ N3 to N6a) and a rearrangement zone (upper part of E6 to beginning of E7; local PAZ N6b to middle of N7) in the upper section. The profile is completed by a compact layer of non-redeposited calcareous muds in the upper half of PAZ E7 (upper part of local PAZ N7, N8). According to the geochemical data, the carbonate content has initially been derived from clastic carbonate of marine origin (Triassic formation) and consequently was replaced by autochthonous limnic carbonate during the transition from the end of the Saalian glacial to interglacial conditions. The carbon and nitrogen contents correlate significantly throughout the whole profile ðrlinear ¼ 0:99Þ and mirror the rate of formation of submerged organic matter in the lake. The two areas of high organic productivity (‘liver muds’) during the climatic optimum (PAZ E4 to E6; local PAZ N4 to N6a) can be observed by a distinct increase in both elements. Stable C:N atomic ratios

of about 10 –16 throughout the entire period indicate constant subaquareous bioproductivity. The end of the interglacial (uppermost part of the PAZ E7; local PAZ N8) is characterised by an increasing carbonate level, which is supported by a rapid increase in molluscs. Additionally, organic matter content also increases (Fig. 18.3). The last reflects the beginning of a siltingup process similar to that observed in PAZs E6 and E4, in the latter with formation of ‘liver muds’ as the final stage. The strong spread of molluscs joint together with a renewed increase in organic matter production towards the very end of the Eemian can be associated with the brief warming phase postulated for PAZ E7 in the Gro¨bern profile (Boettger et al., 2000). The palynological findings of the assembled Neumark-Nord profile are shown in Fig. 18.3 (Boettger et al., 2003). The pollen analysis of this profile revealed a tree sequence typical of the Eemian: birch pine - elm - oak - hazel - hornbeam - pine. The geological position above the Saale till

270

T. Boettger et al.

and the palynological data indicate that the Neumark-Nord profile should definitely be classified as of Eemian interglacial age. Profile segments cover the transition from the Late Saalian to the Eemian (Late Saalian, local PAZ N1) and the whole interglacial period up to the spread of pine and birch at the end of the Eemian (local PAZ N7 and N8). Some peculiarities are found in the profile. First, the sharp change in pollen spectra at the transition between local pollen zones N2 and N3 may suggest a hiatus. The local pollen zone N3 is the zone of Corylus and Quercetum mixtum and corresponds to the PAZ E4 zone. The rearrangement zone in the upper section of the profile extended up to the beginning of the local PAZ N7 and is clearly documented by lithology but is not visible in the pollen data. PAZ E7 can be subdivided into two parts (local PAZ N7 and N8). PAZ N7 is dominated by Pinus pollen. At the end of PAZ E7 pine and spruce-pine forest with minor amount of oak, ash and elm were predominant in the plant cover. In zone N8, the percentages of Betula increase, and at some levels the amount of birch pollen significantly exceeds pine. An increase in Betula percentages is accompanied by an increase in nonarboreal pollen (Artemisia and Chenopodiaceae). Comparable findings are described also in earlier investigations of the Neumark-Nord profile (Litt, 1994). We observe pronounced fluctuations in arboreal vegetation during local PAZ N8. Pollen data indicate at least two stages of degradation of coniferous forest vegetation, while birch woodlands became widespread. Together with geochemical data, these can be interpreted as an indicator of increased climatic instability during the very end of the Eemian. . Klinge The interglacial Klinge profile was exposed in the former open-cast lignite mine Ja¨nschwalde near Cottbus. It is one of the best-investigated Eemian sequences

in Lower Lusatia (Striegler, 1986; Cepek et al., 1994). This lacustrine sequence is developed in a dead ice basin above Saalian till and has in its interglacial part a thickness of about 4.5 m. The investigated profile includes sediments from the Late Saalian and the complete Eemian Interglacial. The sequence was palynologically investigated in 1986 by Seifert-Eulen (unpublished; Fig. 18.4). Because of better past outcrops at that time, those results are better suitable to characterisation of especially the very end of Eemian. Reinterpretation of these older results together with the findings of new 2001 sampling (Strahl, 2004; Velichko et al., 2005) is in progress now (Novenko et al., submitted). As follows from the complete palynological data set of the Klinge profile, the last stage of the Eemian Interglacial was characterised by noticeable changes in plant cover. The beginning of PAZ E7 was marked by pine and birch forests with an admixture of spruce. In the middle part of the interval, the role of forest communities in the plant cover decreased, while birch woodland with dry meadow became widespread. At the end of this stage, pine forest reappeared. These features of vegetation point to a cooling which had been replaced by a short relatively warm interval before the transition to Early Glacial conditions (Novenko et al., submitted). 18.2.1.2 Eastern Europe . Ples The lacustrine Ples profile is situated in the Upper Volga region in Russia and comprises the period from the Late Saalian, the complete Eemian Interglacial and the Early Weichselian including its first interstadial Bro¨rup (Grichuk and Grichuk, 1959). This very important Eastern European Late Pleistocene profile was sampled and investigated in detail (Borisova et al., submitted) in 2002 within the scope of the German–Russian scientific cooperation

Indications of Short-Term Climate Warming

271

Fig. 18.4 Simplified pollen diagram of Klinge section (analyses: M. Seifert-Eulen).

(UFZ Centre for Environmental Research Leipzig-Halle, Saxon Academy of Sciences at Leipzig, Institute of Geography RAS Moscow, Fig. 18.5). The lithological composition of this carbonate-free profile is mainly determined by clastic sediments (silt and clay) in the cool periods and organic-rich muds and peat in the warm stages. The whole profile is about 8.5 m thick and exposed in a deep gully close to the Volga River. The results of our investigations show that as in Central Europe, two warm phases occur in Eastern Europe at the interglacial– glacial transition. These events are indicated both by the geochemical and by the palynological data (Fig. 18.2). The first warming event has a smaller magnitude in comparison with Central Europe and occurs at the

very end of the Eemian. Within the local PAZ Pl 6 (which corresponds to PAZ E7 by Menke and Tynni, 1984), an increase in the organic carbon content and decrease of 13 C values occur, which is confirmed at least by six samples. This can be attributed to an increase in lake productivity accompanied by a rise of 12C-enriched carbon in the lake water due to intensification of such processes as organic matter decay in the water column and on the lake bottom (Wolfe et al., 1999) during warming events in transition to glacial conditions. Changes in the composition of pollen spectra, such as a considerable increase in herb percentages and concentrations accompanied by a reduction in tree and shrub content and total pollen concentration, suggest a degradation of forest

272

T. Boettger et al.

Fig. 18.5 The participants of the German–Russian expedition in 2002 in front of the Late Pleistocene sequence Ples near Volga River. From left to right: Front row: Olga Borisova, Tatjana Boettger, Elena Novenko, Andrei A. Velichko; Middle row: Frank W. Junge, Stefan Knetsch, Anna Kosmakova, Konstantin Kremenetski; Back row: Detlef Degering, Yuri Kononov, Taras Samborski.

communities during the transition from interglacial to glacial conditions. The fall of arboreal pollen content occurs at the beginning of this zone, after which tree and shrub percentages increase again but mainly by coniferous trees (Picea, Pinus sylvestris and Pinus sibirica) and birch. Obviously, the role of forest communities in a complex plant cover increased, which indicates a slight warming in the second half of the PAZ E7. The first new radiofluorescence data (Degering and Krbetschek, this volume) from Ples profile marked for this interval point at an age of approximately 124 12 kyr. The second pronounced warm phase appears in the local PAZ Pl 8 (or PAZ WFI following Menke and Tynni, 1984) within the first Weichselian Herning Stadial.

The radiofluorescence data (Degering and Krbetschek, this volume) for this event show an age of approximately 119 20 kyr. As indicated by the organic carbon content, the submersed primary production in the lake increases, which can be seen in the Corg content, while the 13 C values decrease again (Fig. 18.2). In this zone, pollen spectra are characterised by a noticeable increase of tree birch pollen (up to 70–85%). Other trees and shrubs are poorly represented with the exception of Betula nana. Pollen of herbaceous plants typical of birch communities (Polygonum bistorta, Sanguisorba officinalis, Thalictrum and Valeriana) pregnantly occur in this interval. The presence of Nuphar pollen in PAZ Pl 8, registered here for the first time since the onset of the glacial epoch,

Indications of Short-Term Climate Warming

indicates some warming. Therefore, we can conclude that birch open forest with dense undergrowth of dwarf birch replaced the tundra vegetation. Apparently, the climate amelioration, which gave an impulse to the development of birch woodlands only, was of minor magnitude and duration. Betula alba, being a typical ‘pioneer’ tree, responded to it quickly, while other more demanding tree species did not spread over the area. The spread of Betula can also indicate cold-derived open environments. But the combination of all geochemical and palynological findings point rather to an increase in temperature. 18.3 CONCLUSIONS At the end of the Eemian, pollen and geochemical data of all investigated profiles at the West–East European transect show a gradual cooling during the interglacial– glacial transition. This can be correlated with the continuous temperature fall at approximately 120 000 years BP in the NGRIP ice core from central Greenland (Fig. 18.2). Within this global climatic transition, we have found in three Central European profiles (Gro¨bern, NeumarkNord and Klinge) indications for short warming phases towards the very end of the Last Interglacial. Furthermore, in the East European limnic profile at Ples (Upper Volga region), we found minor warming oscillations during the very end of the Eemian (PAZ E7). Moreover, within the first Weichselian Stadial (Herning), we found clear evidence of a warming event in all investigated profiles in Central as well as in Eastern Europe covering this period. Independent of chronological relations to the warming Dansgaard-Oeschger event (DO) 25 in the NGRIP ice core, the correlation of the detected climatic fluctuations in continental limnic sequences in Central and Eastern Europe is well established. The first fluctuation occurred definitely before the

273

beginning of the first Early Weichselian Interstadial, the second within it. But there are some problems with the correlation of DO 25 in NGRIP with terrestrial records. Johnsen et al. (2005) for example reports that the warming event DO 25 lasted some 4 kyr. The duration of the observed warming events in terrestrial records presented here is unclear. In our opinion, the assumed duration of DO 25 seems to be too long for correlation with the first warming event at the very end of the Eemian (PAZ E7). From this point of view, comparison to the second warming event during the first Weichselian Herning stadial (PAZ WFI) seems to be plausible as well. Therefore, the precise correlation between the investigated terrestrial limnic profiles and the results of NGRIP ice core remains under discussion. Nevertheless, it appears that in Central and Eastern Europe and in Greenland (NGRIP Members, 2004), a warming phase towards the end of the Last Interglacial preceded the final transition to glacial conditions. Thus, this warming phase appears to be a global phenomenon for the Northern Hemisphere. ACKNOWLEDGEMENTS This work was supported financially by the subproject ‘EEM’ of the DEKLIM program (grant number 01LD0041) of the German Ministry of Education and Research (BMBF). Our thanks go to Ms. M. Seifert-Eulen (LfUG Freiberg, Germany) for helpful discussions and friendly permission for using of her yet unpublished palynological data from the Klinge section. We thank Ms. U. Helmstedt (Halle) and Ms. I. Flu¨gel (Leipzig) for carrying out stable isotope analysis.

REFERENCES Berger, A., Loutre, M.F., 2005. Is our interglacial going to be exceptionally long? Book of Abstracts

274

T. Boettger et al.

of DEKLIM/PAGES conference in Mainz, Germany, March 7th–10th, 2005, pp. 82–83. Boettger, T., Junge, F.W. 1994. Stabile Isotope als pala¨oklimatische und pala¨oo¨kologische Indikatoren. In: Eissmann, L., Litt, Th. (Eds.), The Quaternary in Central Germany. An Overview and Excursion Guide Book with a Review on the Pre-Quaternary in the Saale-Elbe Area. Altenburger Naturwissenschaftliche Forschungen 7, pp. 283–289. Boettger, T., Hiller, A., Junge, F.W., Litt, Th., Mania, D., Scheele, N., 1998. Late Glacial stable isotope record, radiocarbon stratigraphy, pollen and mollusc analyses from Geiseltal area, central Germany. Boreas 27, 88–100. Boettger, T., Junge, F.W., Litt, Th., 2000. Stable climatic conditions in central Germany during the last interglacial. Journal of Quaternary Science 15, 469–473. Boettger, T., Junge, F.W., Kremenetski, K., 2003. Climatic conditions in the Eemian Interglacial recorded by stable isotopes from a Neumark-Nord limnic sediment sequence (central Germany). Book of Abstracts XVI. INQUA Congress (Reno, Nevada, USA), pp. 184. Boettger, T., Hiller, A., Stottmeister, L., Junge, F.W., 2004. First isotope studies on the Late Weichselian part of the limnic type sequence from the former Lake Aschersleben (Saxony-Anhalt, Germany). Studia Quaternaria 21, 207–211. Borisova, O.K., Novenko, E. Yu, Velichko, A.A., Kremenetski, K.V., Junge, F.W., Boettger, T., submitted. Vegetation and climate changes during the Eemian and Early Weichselian in the upper Volga region, Russia. Quaternary Science Reviews. Cepek, A.G., Hellwig, D., Nowel, W., 1994. Zur Gliederung des Saale-Komplexes im Niederlausitzer Braunkohlenrevier. Brandenburgische Geowissenschaftliche Beitra¨ge 1, 43–83. Degering, D., Krbetschek, M.R., 2006. Dating of interglacial sediments by luminescence methods (This volume). Eicher, U., Siegenthaler, U., 1976. Palynological and oxygen isotope investigations on Late-Glacial sediment cores from Swiss lakes. Boreas 5, 109–117. Eicher, U., Siegenthaler, U., Wegmu¨ller, S., 1981. Pollen and oxygen isotope analyses on Late and PostGlacial sediments of the Tourbie`re de Chirens (Dauphone´, France). Quaternary Research 15, 160–170. Eicher, U., Oeschger, H., Siegenthaler, U., 1991. Pollenanalyse und Isotopenmessungen an Seekreiden. In: Frenzel, B. (Ed.), Klimatische Probleme der letzten 130 000 Jahre. Fischer Verlag, Stuttgart, pp. 127–138. Eissmann, L., 1990. Das mitteleuropa¨ische Umfeld der Eemvorkommen des Saale-Elbe-Gebietes und

Schlußfolgerungen zur Stratigraphie des ju¨ngeren Quarta¨rs. In: Eissmann, L., Litt, Th. (Eds.), Die Eemwarmzeit und die fru¨he Weichselzeit im Saale-Elbe-Gebiet: Geologie, Pala¨ontologie, Pala¨oo¨kologie. Altenburger Naturwissenschaftliche Forschungen 5, pp. 11–48. Eissmann, L., 1994. Klassische Quarta¨rfolge Mitteldeutschlands von der Elsterzeit bis zum Holoza¨n unter besonderer Beru¨cksichtigung der Stratigraphie, Pala¨oo¨kologie und Vorgeschichte. In: Eissmann, L., Litt, Th. (Eds.), The Quaternary in Central Germany. An Overview and Excursion Guide Book with a Review on the Pre-Quaternary in the Saale-Elbe Area. Altenburger Naturwissenschaftliche Forschungen 7, pp. 250–337. Eissmann, L., 2002. Quaternary geology of eastern Germany (Saxony, Saxony-Anhalt, South Brandenburg, Thuringia), type area of the Elsterian and Saalian Stages in Europe. Quaternary Science Reviews 21, 1275–1346. Firbas, F., 1949. Spa¨t- und nacheiszeitliche Waldgeschichte Mitteleuropas no¨rdlich der Alpen. I: Allgemeine Waldgeschichte. Gustav Fischer Verlag, Jena, 480 pp. Grichuk, V.P., Grichuk, M.P., 1959. Paleolake deposits in the Ples region. In: Markov, K.K., Popov, A.I. (Eds.), Lednikovyi Period na Territorii Evropeiskoy Chasti SSSR i Sibiri. Izdatel’stvo Moskovskogo Universiteta, Moscow, pp. 39–63. Hughen, K.A., Overpeck, J.T., Peterson, L.S., Trumbore, S., 1996. Rapid climate changes in the tropical Atlantic region during the last deglaciation. Nature 380, 51–54. Iversen, J., 1954. The Late-glacial flora of Denmark and its relation to climate and soil. Danmark Geologisk Undersuchunks II 80, 87–119. Johnsen, S.J., Steffensen, J.P., Dahl-Jensen, D., Landais, A., Chappellaz, J., 2005. In and out of a glacial. Book of Abstracts of DEKLIM/PAGES conference in Mainz, Germany, March 7th–10th, 2005, pp. 41–42. Litt, Th., 1990. Pollenanalytische Untersuchungen zur Vegetations- und Klimaentwicklung wa¨hrend des Jungpleistoza¨ns in den Becken von Gro¨bern und Grabschu¨tz. In: Eissmann, L. (Ed.), Die Eemwarmzeit und die fru¨he Weichselkaltzeit im SaaleElbe-Gebiet: Geologie, Pala¨ontologie, Pala¨oo¨kologie. Altenburger Naturwissenschaftliche Forschungen 5, pp. 92–105. Litt, Th., 1994. Pala¨oo¨kologie, Pala¨obotanik und Stratigraphie des Jungquarta¨rs im nordmitteleuropa¨ischen Tiefland. Dissertationes Botanicae 227, pp. 1–185. Litt, Th., Junge, F.W., Boettger, T., 1996. Climate during the Eemian in north-central Europe – a critical review of the palaeobotanical and stable isotope

Indications of Short-Term Climate Warming data from central Germany. Vegetation History and Archaeobotany 5, 247–256. Loutre, M.F., Berger, A., 2003. Marine Isotope Stage 11 as an analogue for the present interglacial. Global and Planetary Change 36, 209–217. Menke, B., Tynni, R., 1984. Das Eeminterglazial und das Weichselfru¨hglazial von Rederstall/ Dithmarschen und ihre Bedeutung fu¨r die mitteleuropa¨ische Jungpleistoza¨n-Gliederung . Geologisches Jahrbuch Reihe A, 76, 3–120. NGRIP Members, 2004. High-resolution record of Northern Hemisphere climate extending into the last interglacial period. Nature 431, 147–151. Novenko, E.Y., Velichko, A.A., Suganova, I.S., Junge F.W., Boettger T., 2005. Dynamics of vegetation at the Late Pleistocene Glacial–Interglacial transition (new data from the center of the East European Plain). In: Winter, H., Balabanis, P. (Eds.), Proceedings of the Workshop ‘‘Reconstruction of Quaternary palaeoclimate and palaeoenviroments and their abrupt changes’’. Polish Geological Institute Special Papers 16, pp. 77–82. Novenko, E.Y., Seifert-Eulen, M., Boettger, T., Junge, F. W., submitted. Eemian and Early Weichselian vegetation and climate history in Central Europe: a case study from Klinge section. Review of Palaeobotany and Palynology. Scharf, B.W., Bittmann, F., Boettger, T., 2005. Freshwater ostracods (Crustacea) from the Lateglacial site at Miesenheim, Germany, and temperature reconstruction during the Meiendorf Interstadial. Palaeogeography, Palaeoclimatology, Palaeoecology 225, 203–215. Seidenkrantz, M.S., Knudsen, K.L., 1994. Marine high resolution records of the Last Interglacial in

275

Northwest Europe: a review. Ge´ographie physique Quaternaire 48, 157–168. Strahl, J., 2004. Das Geotop Klinge – pollenanalytische Untersuchungen an den saalespa¨tglazialen bis weichselfru¨hglazialen Ablagerungen der ehemaligen Dominalgrube von Klinge, Tagebau Ja¨nschwalde. Brandenburger Geowissenschaftliche Beitra¨ge 11, 111–121. Striegler, U., 1986. Zum Eem-Interglazial von Klinge. In: Cepek, A.G. (Ed.), Kurzreferate und Exkursionsfu¨hrer, ‘‘25 Jahre AK Quarta¨rgeologie’’ der GGW vom 10. bis 13. Juli 1986, Berlin, pp. 39–40. Velichko, A.A., Novenko, E.Y., Pisareva, V.V., Zelikson, E.M., Boettger, T., Junge, F.W., 2005. Vegetation and climate changes during Eemian in Central and East Europe: comparative analysis of pollen data. Boreas 34, 207–219. Walkling, A.P., Coope, R., 1996. Climatic reconstruction from the Eemian/Early Weichselian transition in Central Europe based on the coleopteran record from Gro¨bern, Germany. Boreas 25, 145–159. Wansa, S., Wimmer, R., 1990. Geologie des Jungpleistoza¨ns der Becken von Gro¨bern und Grabschu¨tz. In: Eissmann, L., Litt, Th. (Eds.), Die Eemwarmzeit und die fru¨he Weichselzeit im Saale-Elbe-Gebiet: Geologie, Pala¨ontologie, Pala¨oo¨kologie. Altenburger Naturwissenschaftliche Forschungen 5, pp. 49–92. Wolfe, B.B., Edwards, T.W.D., Aravena, R., 1999. Changes in carbon and nitrogen cycling during tree-line retreat recorded in the isotopic content of lacustrine organic matter, western Taimyr Peninsula, Russia. The Holocene 9, 215–222.

This page intentionally left blank

19. Vegetation Dynamics in Southern Germany During Marine Isotope Stage 5 (130 to 70 kyr Ago) Ulrich C. Mu¨ller1 and Maria F. Sa´nchez Gon˜i2 1

Institute of Geology and Palaeontology, Johann Wolfgang Goethe University Frankfurt, Senckenberganlage 32–34, 60054 Frankfurt, Germany 2 Ecole Pratique des Hautes Etudes, De´partement de Ge´ologie et Oce´anographie, UMR 5805 EPOC – OASU, Universite´ Bordeaux 1, Avenue des Faculte´s, 33405 Talence, France

ABSTRACT This contribution reviews the vegetation dynamics in southern Germany from the onset of the Eemian interglacial to the end of the early Weichselian glacial, an interval nearly equivalent to marine isotope stage (MIS) 5. The two longest pollen records in this region, which reflect the vegetation change of MIS 5, are correlated and plotted using a common tentative timescale. The comparison of these records and other pollen records from northern and southern Europe allows to evaluate shifting vegetation and climate gradients. The spread of Eemian forests occurred 6 kyr after the onset of MIS 5e. Eemian thermophilous deciduous forests prevailed for the rest of MIS 5e and were replaced by coniferdominated forests close to the MIS 5e/5d transition. Presumably, the coniferdominated forests persisted in southern Germany well within MIS 5d, whereas they declined 5 kyr earlier in northern Germany. This suggests steep vegetation gradients between northern and southern Germany at the inception of the last glacial. The early Weichselian in southern Germany was characterised by three interstadials (Bro¨rup, Odderade, and Du¨rnten) with conifer-dominated forests separated by stadials with tundra-steppe biomes. The Bro¨rup interstadial correlates with some degree of diachroneity with MIS 5c, whereas both the Odderade and Du¨rnten interstadials correlate with MIS 5a. The end of the Du¨rnten interstadial shows the final demise of early

Weichselian woodlands and the spread of steppe biomes with the onset of MIS 4. 19.1 INTRODUCTION Climate has been changing persistently through Earth history, especially during the past two million years of the Quaternary (e.g. Mix et al., 1995; Zachos et al., 2001; EPICA Community Members, 2004). Therefore, it is reasonable to assume that climate change will continue in the future, irrespective of whether human has a significant influence on global climate change or not (Ruddiman, 2003, 2005; Kukla and Gavin, 2005). In this context, it appears mandatory to gain information about the potential effects of inevitable future climate variability – especially of abrupt climate change. The investigation of past interglacials can yield valuable information on climate change and related environmental response during periods that resemble the present Holocene interglacial. In contrast to the Holocene, past interglacials, having run their full course and demise, can also teach us about environmental responses to human-independent climate change at the end of a warm stage (e.g. Kukla et al., 2002; Mu¨ller and Kukla, 2004; Sa´nchez Gon˜i et al., 2005; Sirocko et al., 2005). Reconstructions of past changes can be performed using various methods among which pollen analysis has been proven to be one of the most useful tools (e.g. Mangerud et al., 1979; Mangerud, 1989; Sa´nchez Gon˜i et al., 2000; Mu¨ller et al.,

Ulrich C. Mu¨ller and Maria F. Sa´nchez Gon˜i

278

2005). Pollen analysis is well suited to examine the impact of rapid climate fluctuations on terrestrial ecosystems since the response of vegetation to climate change is pronounced and fast as it has been shown by analysis of annually laminated sediments (e.g. Allen et al., 1999; Tinner and Lotter, 2001). Precise absolute dating of interglacial records, however, is still a notorious problem. Therefore, most interglacial pollen records are presented on a depth scale rather than a tentative timescale. Furthermore, the procedure to calculate percentages shown in pollen records is not standardised, therefore hampering the comparability of different records. The decision about which taxa are excluded from the calculation sum depends on what the different authors consider to be components of local vegetation signals. Finally, the graphical presentation of complex pollen data sets in a diagram varies between authors. As a consequence of missing chronologies, different calculation procedures and varying presentation forms, pollen records are often difficult to read and even more difficult to compare with each other. Here we review the vegetation dynamics in southern Germany from the onset of the

Eemian interglacial to the end of the early Weichselian warm stages, a time interval that closely correlates with MIS 5 as it has been shown by combined analysis of pollen and marine proxies in cores off western Iberia (Sa´nchez Gon˜i et al., 1999, 2000, 2005; Shackleton et al., 2003). The two pollen records in southern Germany which cover the entire time interval of MIS 5, Samerberg (Gru¨ger, 1989) in the east and Fu¨ramoos (Mu¨ller et al., 2003) in the west (Fig. 19.1), are presented on a common timescale that has been adopted from the records MD95-2042 and MD992231 off western Iberia (Shackleton et al., 2003; Sa´nchez Gon˜i et al., 2005). Although the transfer of the Iberian chronology to southern Germany can provide only a tentative timescale for the latter region, it allows a direct comparison of the pollen signal between the two German records. Furthermore, both records are presented graphical in the same way, and the calculation procedure of pollen percentages is the same for both records. The comparison of the two records in southern Germany and with other pollen records from northern and southern Europe allows to evaluate shifting vegetation and climate gradients during the substages of MIS 5.

N Bispingen

Gröbern

MD99-2331

Füramoos Samerberg

MD95-2042 100

Fig. 19.1 Location of some sites mentioned in the text.

500 km

Vegetation Dynamics in Southern Germany During MIS 5

19.2 AGE MODEL The construction of the age model is based on two steps: First, major climatic events documented in both pollen records from southern Germany are correlated with each other (Table 19.1, left site). This step assumes synchronicity of major climate events within southern Germany. For example, the prominent warming allowing the rapid spread of forests at the beginning of the Eemian interglacial must have occurred at the same time in the east and west of southern Germany. Further, the coldest point associated with the peak abundance of nonarboreal pollen within the Melisey 1 stadial is assumed to be simultaneous in both the southern German records. A list of the seven correlation points used is given in the left column of Table 19.1. In a second step, certain ages are assigned to the correlation points (Table 19.1, right site). The age data are adopted from the chronology of the pollen record of core MD95-2042 off Portugal, which reflects the same major climatic events during stage 5 as listed in Table 19.1 (Shackleton et al., 2003; Sa´nchez Gon˜i et al., 2005). This transfer of a chronology over 2200 km provides a tentative timescale for the two south German records. Subsequently, all age data, if not stated otherwise, are based on linear interpolation between these tuning points.

279

19.3 VEGETATION DYNAMICS DURING MIS 5 19.3.1 MIS 5e During the Saale late glacial prior to the beginning of the Eemian interglacial (assigned to 126 kyr), both southern German pollen records (Figs. 19.2 and 19.3) reflect an open, steppe-like vegetation. The Samerberg record in the east shows higher percentages of the steppe taxon Artemisia than the Fu¨ramoos record in the west. This suggests dryer conditions to the east during the Saale late glacial. In marine cores off western Iberia, combined pollen and marine proxy analysis have shown that the palynostratigraphic late glacial already belongs to the early part of MIS 5e (Sa´nchez Gon˜i et al., 1999, 2005; Shackleton et al., 2002). Most likely, the spread of Eemian tree species replacing late glacial steppe biomes occurred slightly later in Germany than in Iberia because trees had to migrate from refugia in the southern parts of Europe. Therefore, we conclude that the palynostratigraphic late glacial in southern Germany occurred well within MIS 5e. The reforestation at the beginning of the Eemian interglacial, which took place 6 kyr after the onset of MIS 5e (Shackleton et al., 2002; Sa´nchez Gon˜i et al., 2005) was associated with a rapid spread of the

Table 19.1 Tie points and assigned ages used for the establishment of the age model for the pollen records Fu¨ramoos (Fig. 19.2) and Samerberg (Fig. 19.3) Step 1

Step 2

Correlation points between Fu¨ramoos (Mu¨ller et al., 2003) and Samerberg (Gru¨ger, 1989)

Age adopted from core MD95-2042 (Sa´nchez Gon˜i et al., 2005)

(7) End Du¨rnten/Ognon interstadial (6) End of Odderade/St. Germain 2 interstadial (5) End of Melisey 2 stadial (4) End of Bro¨rup/St. Germain 1 interstadial (3) Peak abundance of nonarboreal within Montaigu event (2) Peak abundance of nonarboreal within Melisey 1 stadial (1) Reforestation at beginning of Eemian interglacial

73 kyr 75.5 kyr 81.5 kyr 87 kyr 103 kyr 107 kyr 126 kyr

See text for further explanation.

Ulrich C. Mu¨ller and Maria F. Sa´nchez Gon˜i

280

40 5

50 15

70 10

30

20

50

Plenigla

20

S3 Dü

(7) (6)

Odder IS

MIS 5a

MIS 5b

80

5

(5)

MeS2

70

MIS 4

50

(4) Brörup IS

MIS 5c

90

100

MIS 5e

120

MeS1

(2)

Eemian Intergl

110

MIS 5d

(3)

ea Pic

ies Ab

s nu

xu

s

rpi Ca

Ta

us ryl Co

us xin

Fra

erc

us

us Qu

Ulm

a tul Be

MIS 6

kyr BP

130

P Ar inus bo rea Gr He l am rbs in Ar eae tem isia Ju Sali nip x eru s

SLG

(1)

Fig. 19.2 Vegetation changes during marine isotope stage (MIS) 5 as reflected in the Fu¨ramoos record, southwest Germany (Fig. 19.1). The timescale to left has been adopted from core MD95-2042 off western Iberia (Shackleton et al., 2003; Sa´nchez Gon˜i et al., 2005). Dashed lines, arrows and numbers in brackets indicate position of tie points (c.f. Table 19.1). SLG ¼ Saale late glacial, Eemian Intergl ¼ Eemian interglacial, MeS1 ¼ Melisey 1 stadial, Bro¨rup IS ¼ Bro¨rup interstadial, MeS2 ¼ Melisey 2 stadial, Odder IS ¼ Odderade interstadial, S3 ¼ third stadial, Du¨ ¼ Du¨rnten interstadial, Plenigla ¼ Pleniglacial. Pollen data from Mu¨ller et al. (2003).

pioneer trees Betula and Pinus in southern Germany. During the dominance of pioneer woodlands centred at around 125 kyr BP, both records show the immigration of the oak-mixed elements Ulmus, Quercus and Fraxinus, but also of Corylus and Picea. Subsequently, both sites exhibit the spread of Ulmus, Quercus and Fraxinus. The difference between both sites, however, is, that the oakmixed forest spread was associated with a major spread of Picea in the east and Corylus in the west (Figs. 19.2 and 19.3). This points to a more continental climate and/or the stronger influence of the Alps in the east (Mu¨ller, 2000). The dominance of oak-mixed forests at ca. 123 kyr BP indicates that the Eemian maximum of mean summer temperatures was

during its early part. According to pollenbased quantitative climate reconstructions, mean summer temperatures were around 2 C higher than today at Samerberg and Fu¨ramoos (Klotz et al., 2003). After this optimum, the spread of Taxus documented in both records (with higher percentages in the east) points to a slight cooling of summer temperatures. Subsequent to the Taxus peak, Abies and Carpinus spread simultaneously at the western site (Fig. 19.2). In the east, however, Picea and Abies had become an important constituent of the forest before Carpinus spread (Fig. 19.3). Thus, Carpinus had to compete with Picea and Abies when spreading. Therefore, this taxon remained less important in the east of southern Germany (Gru¨ger, 1995). When

Vegetation Dynamics in Southern Germany During MIS 5 40 10

30

40 10

40

80

(7) (6)

S3 Dü

MIS 4

20 10

Odder IS

MIS 5a

MIS 5b

80

50

(5)

MeS2

70

10

Plenigla

50

281

(4) Brörup IS

MIS 5c

90

100

MIS 5e

120

MeS1

(2)

Eemian Intergl

110

MIS 5d

(3)

ea Pic

s

ies Ab

nu

xu

s

rpi Ca

Ta

us ryl

us xin

Co

us erc

Fra

Qu

tul a Ulm us

Be

bo

Ar

Pin

MIS 6

Kyr BP

130

us rea He l Gr am rbs in Ar eae tem isi Sa a lix Ju nip eru s

SLG

(1)

Fig. 19.3 Vegetation changes during marine isotope stage (MIS) 5 as reflected in the Samerberg record, southeast Germany (Fig. 19.1). The timescale to left has been adopted from core MD95-2042 off western Iberia (Shackleton et al., 2003; Sa´nchez Gon˜i et al., 2005). Dashed lines, arrows and numbers in brackets indicate position of tie points (c.f. Table 19.1). SLG ¼ Saale late glacial, Eemian Intergl ¼ Eemian interglacial, MeS1 ¼ Melisey 1 stadial, Bro¨rup IS ¼ Bro¨rup interstadial, MeS2 ¼ Melisey 2 stadial, Odder IS ¼ Odderade interstadial, S3 ¼ third stadial, Du¨ ¼ Du¨rnten interstadial, Plenigla ¼ Pleniglacial. Pollen data from Gru¨ger (1989).

considering other Eemian pollen records in the northern and western alpine foreland, e.g. Mondsee (Drescher-Schneider, 2000), Zeifen (Jung et al., 1972), Eurach (Beug, 1979), Wurzach (Gru¨ger and Schreiner, 1993), Jammertal (Mu¨ller, 2000; Mu¨ller et al., 2005), Gondiswil (Wegmu¨ller, 1992), Grande Pile (Woillard, 1978; Beaulieu and Reille, 1992) and Les Echets (Beaulieu and Reille 1984), it appears that Carpinus migrated from the west to the east. Picea and Abies, in contrast, migrated from the east to the west (Gru¨ger, 1995). Pollen-based climate reconstructions suggest that the Eemian optimum of mean winter temperatures (slightly above 0 C) occurred at both sites during the Abies phase (Klotz et al., 2003)

centred at around 117 kyr. Subsequent to this optimum, the decline of thermophilous tree taxa associated with the spread of less cold-sensitive conifers, indicates that both winter and summer temperatures declined. 19.3.2 MIS 5d The post-temperate phase of the Eemian in southern Germany, lasting from around 115 to 110 kyr BP, shows minor percentages of thermophilous trees and the dominance of the more cold-resistant coniferous Picea and Pinus. Pollen-based climate reconstructions suggest that mean winter and summer temperatures had decreased by around 7 and 2 C, respectively, during the

282

Ulrich C. Mu¨ller and Maria F. Sa´nchez Gon˜i

post-temperate phase (Klotz et al., 2003). The stronger decline in mean winter temperatures might be associated with a southward displacement of the North Atlantic Current at around 115 kyr (Mu¨ller and Kukla, 2004). Finally, the replacement of Picea and Pinus by grasses and herbal heliophytes such as Artemisia reflects the opening of late Eemian woodlands, which marks the end of the biostratigraphic Eemian in southern Germany at ca. 110 kyr (Figs. 19.2 and 19.3). Combined pollen and oxygen isotope analysis from cores off western Iberia show that the last 6 kyr of Eemian woodlands occurred already during MIS 5d, (Shackleton et al., 2003; Sa´nchez Gon˜i et al., 2005). We suppose that the post-temperate phase with conifer woodlands lasting 5 kyr in southern Germany also belongs already to MIS 5d. In the cores off western Iberia, the peak abundance of nonarboreal pollen during the Melisey 1 stadial is coeval to the sea-surface temperature (SST) cold event C24 at 107 kyr during late MIS 5d (Sa´nchez Gon˜i et al., 1999, 2000, 2005). It has been proposed that the peak abundance of nonarboreal pollen during the Melisey 1 occurred nearly coeval across Europe (e.g. Kukla et al., 1997; Forsstro¨m, 2001; Mu¨ller and Kukla, 2004). Therefore, we tied this peak of nonarboreal taxa to 107 kyr (Table 19.1). In contrast, the opening of forests at the beginning of the Melisey 1 stadial occurred earlier in more northern areas than in the south of Europe. Furthermore, the reforestation at the end of the stadial should be later in regions north of southern Europe. Therefore, the existence of open vegetation biomes may have lasted longer in the north than in the south of Europe. In southern Germany, climate conditions associated with this first Weichselian stadial affected an open tundrasteppe biome which lasted for presumably 4 to 5 kyr (Figs. 19.2 and 19.3). 19.3.3 MIS 5c The beginning of the Bro¨rup (¼ St. Germain 1) interstadial is recorded in core MD95-2042

off western Iberia during the latest MIS 5d when oxygen isotopes of planktonic foraminifers indicate already an SST rise (Shackleton et al., 2002). Owing to plant migration lags the reforestation with the onset of the Bro¨rup interstadial occurred rather slightly later in southern Germany than in southern Europe. The reforestation in southern Germany was marked by a rapid spread of the pioneer tree taxa Juniperus, Betula and Pinus at both sites (Figs. 19.2 and 19.3). Subsequent to the spread of Picea, coniferous woodlands consisting mainly of Picea and Pinus dominated southern Germany at 104 kyr. Some few pollen grains from thermophilous deciduous trees Ulmus, Quercus, Fraxinus, Corylus and Carpinus are also found. However, it is not clear whether these grains are a result of long-distance transport or whether single stands of these thermophilous trees were present close to the sites in southern Germany. A short and abrupt cooling in the lower third of the Bro¨rup interstadial, termed Montaigu event (Woillard, 1978), has been found to match the North Atlantic SST cold event C23 at ca. 103 kyr (Shackleton et al., 2003; Sa´nchez Gon˜i et al., 2005). As the Montaigu event occurred probably synchronously across Europe (e.g. Kukla et al., 1997; Forsstro¨m, 2001; Mu¨ller et al. 2003), we tied the peak abundance of nonarboreal taxa during the Montaigu event to 103 kyr (Table 19.1). In southern Germany, the Montaigu event is documented by a minor opening of conifer forests and a decrease of Picea to the benefit of Pinus (Figs. 19.2 and 19.3). The subsequent recovery of forests is indicated by a re-spread of Picea in both pollen records. Thermophilous deciduous tree taxa also migrated into the alpine foreland, reaching in total a maximum of 16% at Fu¨ramoos and 2% at Samerberg around 3 kyr after the Montaigu event. This suggests that the climatic optimum of the Bro¨rup interstadial occurred after the Montaigu event. A comparison with the Eemian pollen record of Les Echets (Beaulieu and Reille, 1984) in the French alpine foreland shows that the climate

Vegetation Dynamics in Southern Germany During MIS 5

optimum of the Bro¨rup interstadial alternatively may have occurred before the Montaigu event. The higher percentages of thermophilous deciduous trees at Fu¨ramoos than at Samerberg indicate that these taxa migrated from west to east during the Bro¨rup interstadial. In the later part of the Bro¨rup interstadial, decreased Picea and increased Pinus percentages suggest a moderate cooling phase which is centred at 91 kyr (Figs. 19.2 and 19.3). This cooling phase may correspond to North Atlantic SST cold event C22 (Kukla et al., 1997; Forsstro¨m, 2001; Mu¨ller et al., 2003). A late Bro¨rup climate warming is suggested by a rise of Picea percentages at both sites before the final decline of Picea and Pinus documents the end of the interstadial. 19.3.4 MIS 5b Further up core, the lithology of the German cores indicates significant changes in sedimentation rates. Therefore, the top of each subsequent palynostratigraphic unit had to be tied to the Iberian chronology (Table 19.1). This renders estimates of differences in the duration of units between both sites impossible. The Melisey 2 stadial has been found to correlate with some degree of diachroneity with MIS 5b centred at 85 kyr in the marine cores off western Iberia (Sa´nchez Gon˜i et al., 1999, 2005). In southern Germany, Melisey 2 was characterised by a substantial deforestation accompanied by a spread of Betula species. High percentages of Artemisia and Gramineae consistently found at both sites suggest the presence of steppe biomes. In comparison with the Melisey 1 stadial, Melisey 2 shows higher percentages of Artemisia which suggests that the later stadial experienced a more continental climate. 19.3.5 MIS 5a The reforestation which marks the beginning of the Odderade (¼ St. Germain 2) interstadial is recorded in the Iberian cores

283

at 81:5 kyr close to the MIS 5b/5a transition (Sa´nchez Gon˜i et al., 1999, 2005). Most likely, the reforestation occurred slightly later in Germany than in Iberia because tree taxa had to migrate from refugia in southern Europe. The reforestation started with a spread of Pinus in southern Germany. Subsequently, Picea spread and thermophilous deciduous trees such as Ulmus, Quercus, Fraxinus, Corylus and Carpinus migrated into southern Germany early within the Odderade interstadial. In total, pollen grains from thermophilous deciduous trees reach a maximum of 20% at the western site Fu¨ramoos and 5% at the eastern site Samerberg. This indicates that the climate optimum occurred in a very early phase of the Odderade which is in agreement with 18 O-based SST estimates from the cores off western Iberia. The early recurrence of thermophilous deciduous trees suggests that their refugia must have persisted in areas north or at least west of the alpine foreland during the preceding Melisey 2 stadial (Mu¨ller et al., 2003). After the climate optimum, a co-dominance of Picea and Pinus is recorded at Fu¨ramoos, whereas Picea prevailed at Samerberg. This suggests the presence of coniferous forest biomes in southern Germany during the early part of MIS 5a. Subsequently, the Odderade interstadial shows a decline of Picea which points to a gradual cooling. The end of the Odderade is marked by a final peak of Pinus at Samerberg and a subsequent rapid decline of Picea and Pinus in both records. The stadial subsequent to the Odderade, termed ‘Stadial 1’ in the Iberian cores, has been found within MIS 5a during a 1 kyr lasting period when 18 O values from planktonic foraminifers show an interim SST setback (Sa´nchez Gon˜i et al., 1999, 2000). Therefore, the end of the Odderade interstadial should not be considered an equivalent to the end of MIS 5 as it has been proposed before (Behre, 1989). In southern Germany, this stadial (labelled S3 ¼ stadial 3 in Figs. 19.2 and 19.3) subsequent to the Odderade was associated

284

Ulrich C. Mu¨ller and Maria F. Sa´nchez Gon˜i

with a major deforestation. High percentages of Gramineae and Artemisia suggest the presence of steppe biomes. The third interstadial above the Eemian, referred to as ‘Ognon’ in the Iberian cores, has been found in the uppermost part of MIS 5a during a final phase with light 18 O values from planktonic foraminifers (Sa´nchez Gon˜i et al., 1999, 2000, 2005). In the northern alpine foreland, the third interstadial above the Eemian has been termed ‘Du¨rnten’ (Welten, 1982). Since the Du¨rnten and Ognon interstadials are recorded in the same palynostratigraphic position, it is concluded that the Du¨rnten interstadial belongs to MIS 5a as well (Figs. 19.2 and 19.3). The pollen records in southern Germany show that the Du¨rnten interstadial was characterised by a spread of Pinus and Picea in the west and Picea in the east. The increased percentages of these conifers in combination with the finding of Selaginella selaginoides suggest that the timberline was located close to the sites in southern Germany. The end of the Du¨rnten interstadial is marked by the final demise of early Weichselian woodlands and the spread of steppe biomes in southern Germany with the onset of MIS 4. 19.4 VEGETATION GRADIENTS DURING THE DECLINE OF THE LAST INTERGLACIAL The duration and timing of the last interglacial has been a matter of intensive debate (e.g. Kukla et al., 1997; Forsstro¨m, 2001; Turner, 2002; Muhs, 2002; Shackleton et al., 2003; Tzedakis, 2003; Mu¨ller and Kukla, 2004). This issue holds implications for the potential duration of the present Holocene interglacial which has lasted already 11.5 kyr. Pollen analysis from annually laminated lake sediments in northern Germany implies that Eemian woodlands lasted for around 11 kyr (Mu¨ller, 1974), respectively 9.5 kyr (Hahne et al., 1994), and 10 to 11 kyr (Caspers et al., 2002). The interval

comprising Eemian woodlands has been proposed to span the period from ca. 126 to 115.5 kyr (Ku¨hl and Litt, 2003). This duration of Eemian woodlands in northern Germany, however, is in contrast to recently presented results from southern Europe. Based on pollen analysis and varve counting from Lago Grande di Monticchio in southern Italy, Eemian woodlands persisted from ca. 127 to ca. 109 kyr (Allen et al., 2005), thus having a duration of ca. 18 kyr. Similar values for the timing and duration of Eemian woodlands in southern Europe have been proposed before by Kukla et al. (1997), Kukla (2000), Tzedakis et al. (2002) and corroborated by Shackleton et al. (2002, 2003). Apparently, Eemian woodlands persisted around 7 kyr longer in southern Europe than in northern Germany. How do we explain this contrast? It is generally assumed that forests spread rather rapid across Europe at the beginning of the Eemian interglacial, similar to the rapid reforestation at the onset of the Holocene (e.g. Kukla et al. 2002; Ku¨hl and Litt, 2003, Mu¨ller et al., 2003). However, most workers believe that there was a several thousand years lasting phase during the declining stage of the last interglacial when Eemian forests persisted in southern Europe, whereas open tundra-like vegetation prevailed in northern Europe (e.g. Kukla et al., 1997, 2002; Shackleton et al. 2002, 2003; Tzedakis et al., 2002). The feasibility of such steep vegetation gradients has been doubted (Turner, 2002). To explain the steep gradients, a reconstruction has been provided, which suggests that the steepening of vegetation gradients in Europe was associated with a southward displacement of the North Atlantic Current to the southwest of the Scotland–Iceland Ridge at the transition from MIS 5e to MIS 5d (Mu¨ller and Kukla, 2004). This would result in a substantial cooling in the Nordic Seas, thus also of northwest Europe. A stronger North Atlantic circulation, limited to southwest of the Scotland–Iceland Ridge, may have helped to maintain the steep climate and

Vegetation Dynamics in Southern Germany During MIS 5

vegetation gradients by preferentially supplying heat to southern Europe at the expense of northern Europe (McManus, written communication, 2004). Tzedakis (2003) suggested that the decline of tree populations in northern Europe at ca. 115 kyr may have been caused by orbital changes that reduced insolation in the growing season. Such orbital changes, however, appear too slow to explain the rather abrupt decline of forests in northern Europe. Most likely, the continuous decrease in high-latitude summer insolation producing the southward displacement of the vegetation belts from 122 to 120 kyr, associated albedo feedbacks, and expanding ice sheet (Sa´nchez Gon˜i et al., 2005) in concert with the southward displacement of the North Atlantic current (Mu¨ller and Kukla, 2004) forced the system to cross a critical threshold leading to the northern European forest decline. The two south German pollen records display the opening of woodlands to have occurred in this region at 110 kyr (Figs. 19.2 and 19.3). This age estimate should be considered as tentative as it is based on linear interpolation between two tie points, i.e. (1) the reforestation at the beginning of the Eemian at 126 kyr and (2) the peak abundance of nonarboreal during the Melisey 1 stadial at 107 kyr (Table 19.1). However, when the same tie points are applied to Eemian pollen records from northern Germany such as Gro¨bern (Litt, 1994), linear interpolation places the opening of late Eemian woodlands at 115 kyr, which yields a duration of 11 kyr for the Eemian in northern Germany (Fig. 19.4). This result based on simple linear interpolation matches the abovementioned 11 kyr duration of the Eemian in northern Germany based on varve counting (cf. Mu¨ller, 1974; Ku¨hl and Litt, 2003). The agreement between linear interpolation and varve counting for northern Germany may suggest that linear interpolation applied to southern Germany may yield likely results as well. According to linear interpolation, Eemian woodlands lasted for

285

Tie point (2)

Tie point (1) Eemian 100% 80

Northern Germany Gröbern

60 40 20

Eemian 100% 80

Southern Germany Füramoos

60 40 20 105

110

MIS 5d Weichselian

115

120

125

MIS 5e Eemian

130 kyr BP

MIS 6 Saalian

Fig. 19.4 Proposed correlation of last interglacial pollen records from northern and southern Germany. Linear interpolation between the beginning of the Eemian at 126 kyr (tie point 1) and the peak abundance of nonarboreal taxa during the first Weichselian stadial at 107 kyr (tie point 2) suggests that Eemian woodlands declined 5 kyr earlier in northern than in southern Germany. Black curve ¼ total tree taxa, red curve ¼ total thermophilous taxa, green curve ¼ Pinus taxa. Pollen data from Gro¨bern are adopted from Litt (1994) and pollen data from Fu¨ramoos are adopted from Mu¨ller et al. (2003).

16 kyr in southern Germany, which is 5 kyr longer than in northern Germany (Fig. 19.4). A closer inspection of Fig. 19.4 shows that the spread of Pinus (green curve) at around 115 kyr was associated with a minor ð 5%Þ decrease of total tree taxa (black curve) in southern Germany, however, with a substantial ð 30%Þ decrease of total tree taxa in northern Germany. This suggests that Pinus-rich tundra-steppe biomes existed in northern Germany, whereas Pinusdominated forests persisted in southern Germany during an around 5 kyr long-lasting phase in the declining stage of the last interglacial. As shown above, we suppose that this Pinus-dominated phase belongs already to MIS 5d, a period when ice sheets expanded

286

Ulrich C. Mu¨ller and Maria F. Sa´nchez Gon˜i

over the Scandinavian mountains (Mangerud, 2004). Accordingly, the reconstruction results suggest a significant steepening of climate and vegetation gradients between north and southern Germany at the inception of the last glacial. ACKNOWLEDGEMENTS We thank E. Gru¨ger for providing the raw count data of the Samerberg pollen record. J. Mangerud and J. Pross are thanked for fruitful discussions. Financial support through DFG grant PR 651/3-1 is gratefully acknowledged.

REFERENCES Allen, J.R.M., Brandt, U., Brauer, A., Hubberten, H.-W., Huntley, B., Keller, J., Kraml, M., Mackensen, A., Mingram, J., Negendank, J.F.W., Nowaczyk, N.R., Oberha¨nsli, H., Watts, W.A., Wulf, S., Zolitschka, B., 1999. Rapid environmental changes in southern Europe during the Last Glacial period. Nature 400, 740–743. Allen, J.R.M., Huntley, B., The Monticchio Working Group, 2005. The Eemian vegetation and climate record from Lago Grande di Monticchio, southern Italy. Poster presentation, DEKLIM-PAGES Conference, 7.–10.03.2005, Mainz, Germany. Beaulieu, J.-L. de, Reille, M., 1984. A long Upper Pleistocene pollen record from Les Echets, near Lyon, France. Boreas 13, 111–132. Beaulieu, J.-L. de, Reille, M., 1992. The last climatic cycle at La Grande Pile (Vosges, France): a new pollen profile. Quaternary Science Reviews 11, 431–438. Behre, K.-E., 1989. Biostratigraphy of the last glacial period in Europe. Quaternary Science Reviews 8, 25–44. Beug, H.-J., 1979. Vegetationsgeschichtlich-pollenanalytische Untersuchungen am Riss/Wu¨rmInterglazial von Eurach am Starnberger See/Obb. Geologica Bavarica 80, 91–106. Caspers, G., Merkt, J., Mu¨ller H., Freund, H., 2002. The Eemian interglaciation in northwestern Germany. Quaternary Research 58, 49–52. Drescher-Schneider, R., 2000. The Riss-Wu¨rm interglacial from West to East in the Alps: an overview of the vegetational succession and climatic development. Geologie en Mijnbouw 79, 233–239.

EPICA Community Members, 2004. Eight glacial cycles from an Antarctic ice core. Nature 429, 623–628. Forsstro¨m, L., 2001. Duration of interglacials: a controversial question. Quaternary Science Reviews 20, 1577–1586. Gru¨ger, E., 1989. Palynostratigraphy of the last interglacial/glacial cycle in Germany. Quaternary International 3/4, 69–79. Gru¨ger, E., 1995. Correlation of Middle-European Late-Pleistocene pollen sequences of the Pfefferbichl and Zeifen types. Mededelingen Rijks Geologische Dienst 52, 97–104. Gru¨ger, E., Schreiner, A., 1993. Riss/Wu¨rm- und wu¨rmzeitliche Ablagerungen im Wurzacher Becken (Rheingletschergebiet). Neues Jahrbuch Geologisch Pala¨ontologische Abhandlungen 189, 81–117. Hahne, J., Kemle, S., Merkt, J., Meyer, K.-D., 1994. Eem-, weichsel- und saalezeitliche Ablagerungen der Bohrung ‘‘Quakenbru¨ck GE 2’’. Geologisches Jahrbuch A 134, 9–69. Jung, W., Beug, H.-J., Dehm, R., 1972. Das Riss/ Wu¨rm-Interglazial von Zeifen, Landkreis Laufen a.d. Salzach. Bayrische Akademie der Wissenschaften, Abhandlungen Mathematischnaturwissenschaftliche Klasse, Neue Folge 151, Mu¨nchen (131 pp). Klotz, S., Guiot, J., Mosbrugger, V., 2003. Continental European Eemian and early Wu¨rmian climate evolution: Comparing signals using different quantitative reconstruction approaches based on pollen. Global and Planetary Change 36, 277–294. Ku¨hl, N., Litt, T., 2003. Quantitative time series reconstruction of Eemian temperature at three European sites using pollen data. Vegetation History and Archaeobotany 12, 205–214. Kukla, G.J., 2000. The last interglacial. Science 287, 987–988. Kukla, G.J., Gavin, J., 2005. Did glacials start with global warming?. Quaternary Science Reviews 24, 1547–1557. Kukla, G.J., McManus, J.F., Rousseau, D.D., Chuine, I., 1997. How long and how stable was the last interglacial? Quaternary Science Reviews 16, 605–612. Kukla, G.J., Bender, M.L., Beaulieu, J.-L. de, Bond, G., Broecker, W.S., Cleveringa, P., Gavin, J.E., Herbert, T.D., Imbrie, J., Jouzel, J., Keigwin, L.D., Knudsen, K.-L., McManus, J.F., Merkt, J., Muhs, D.R., Mu¨ller, H., Poore, R.Z., Porter, S.C., Seret, G., Shackleton, N.J., Turner, C., Tzedakis, P.C., and Winograd, I.J., 2002. Last interglacial climates. Quaternary Research 58, 2–13. Litt, T., 1994. Pala¨oo¨kologie, Pala¨obotanik und Stratigraphie des Jungquarta¨rs im nordmitteleuropa¨ischen Tiefland unter besonderer Beru¨cksichtigung des Elbe-Saale-Gebietes. Dissertationes Botanicae 227, Borntra¨ger, Stuttgart (185 pp).

Vegetation Dynamics in Southern Germany During MIS 5 Mangerud, J., 1989. Correlation of the Eemian and the Weichselian with deep sea oxygen isotope stratigraphy. Quaternary International 3/4, 1–4. Mangerud, J., 2004. Ice sheet limits on Norway and the Norwegian continental shelf. In Ehlers, J., Gibbard, P., (eds.). Quaternary Glaciations – Extent and Chronology. Vol. 1. Europe, Elsevier, Amsterdam, 271–294. Mangerud, J., Sønstegaard, E., Sejrup, H.-P., 1979. Correlation of the Eemian (interglacial) Stage and the deep-sea oxygen-isotope stratigraphy. Nature 277, 189–192. Mix, A.C., Pisias, N.G., Rugh, W., Wilson, J., Morey, A., Hagelberg, T.K., 1995. Benthic foraminifer stable isotope record from Site 849 (0–5 Ma); local and global climate changes. Proceedings of the Ocean Drilling Program, Scientific Results 138, 371–412. Muhs, D.R., 2002. Evidence for the timing and duration of the last Interglacial period from highprecision uranium-series ages of corals on tectonically stable coastlines. Quaternary Research 58, 36–40. Mu¨ller, H., 1974. Pollenanalytische Untersuchungen und Jahresschichtenza¨hlungen an der eemzeitlichen Kieselgur von Bispingen/Luhe. Geologisches Jahrbuch A 21, 148–169. Mu¨ller, U.C., 2000. A late Pleistocene pollen sequence from the Jammertal, SW Germany with particular reference to location and altitude as factors determining Eemian forest composition. Vegetation History and Archaeobotany 9, 125–131. Mu¨ller, U.C., Kukla, G.J., 2004. North Atlantic current and European environments during the declining stage of the last interglacial. Geology 32, 1009–1012. Mu¨ller, U.C., Pross, J., Bibus, E., 2003. Vegetation response to rapid climate change in Central Europe during the past 140 000 yr based on evidence from the Fu¨ramoos pollen record. Quaternary Research 59, 235–245. Mu¨ller, U.C., Klotz, S., Geyh, M.A., Pross, J., Bond G.C., 2005. Cyclic climate changes during the Eemian Interglacial in Central Europe. Geology 33, 449–452. Ruddiman, W. F., 2003. The Anthropogenic Greenhouse era began thousands of years ago. Climate Change 61, 261–293. Ruddiman, W. F., 2005. Cold climate during the closest Stage 11 analog to recent Millennia. Quaternary Science Reviews 24, 1111–1121. Sa´nchez Gon˜i, M.F., Eynaud, F., Turon, J.L., Shackleton, N.J., 1999. High resolution palynological record off the Iberian margin: direct land–sea correlation for the last interglacial complex. Earth and Planetary Science Letters 171, 123–137.

287

Sa´nchez Gon˜i, M.F., Eynaud, F., Turon, J.-L., Shackleton, N.J., Cayre, O., 2000. Direct land–sea correlation for the Eemian and its comparison with the Holocene: a high resolution palynological record off the Iberian margin, Netherlands. Journal Geoscience 79, 345–354. Sa´nchez Gon˜i, M.F., Loutre, M.F., Crucifix, M., Peyron, O., Santos, L., Duprat, J., Malaize´, B., Turon, J.-L., Peypouquet, J.-P., 2005. Increasing vegetation and climate gradient in Western Europe over the Last Glacial Inception (122–110 ka): datamodel comparison. Earth and Planetary Science Letters 231, 111–130. Shackleton, N.J., Chapman, M., Sa´nchez-Gon˜i, M.F., Pailler, D., Lancelot, Y., 2002. The classic marine isotope substage 5e. Quaternary Research 58, 14–16. Shackleton, N.J., Sa´nchez Gon˜i, M.F., Pailler, D., Lancelot, Y., 2003. Marine isotope substage 5e and the Eemian Interglacial. Global and Planetary Change 36, 151–155. Sirocko, F., Seelos, K., Schaber, K., Rein, B., Dreher, F., Diehl, M., Lehne, R., Ja¨ger, K., Krbetschek, M., Degering, D., 2005. A late Eemian aridity pulse in central Europe during the last glacial inception. Nature 436, 833–836. Tinner, W., Lotter, A.F., 2001. Central European vegetation response to abrupt climate change at 8.2 ka. Geology 29, 551–554. Turner, C., 2002. Problems of the duration of the Eemian Interglacial in Europe north of the Alps. Quaternary Research 58, 45–48. Tzedakis, P.C., 2003. Timing and duration of last interglacial conditions in Europe: A chronicle of a changing chronology. Quaternary Science Reviews 22, 763–768. Tzedakis, P.C., Frogley, M.R., Heaton, T.H.E., 2002. Duration of last interglacial conditions in northwestern Greece. Quaternary Research 58, 53–55. Wegmu¨ller, S., 1992. Vegetationsgeschichtliche und stratigraphische Untersuchungen an Schieferkohlen des no¨rdlichen Alpenvorlandes. Denkschriften der Schweizerischen Akademie der Naturwissenschaften 102, Basel, Boston, Berlin (82 pp). Welten, M., 1982. Pollenanalytische Untersuchungen im ju¨ngeren Quarta¨r des no¨rdlichen Alpenvorlandes der Schweiz. Beitra¨ge zur geologischen Karte der Schweiz N.F. 156, Bern (174 pp). Woillard, G.M., 1978. Grande Pile peat bog: A continuous pollen record for the last 140,000 years. Quaternary Research 9, 1–21. Zachos, J., Pagini, M., Sloan, L., Thomas, E., Billups, K., 2001. Trends, rhythms, and aberrations in global climate 65 Ma to present. Science 292, 686–693.

This page intentionally left blank

20. Subtropical NW Atlantic Surface Water Variability During the Last Interglacial M.J. Vautravers1,2, G. Bianchil3 and N.J. Sackleton1* 1

Godwin Laboratory for Palaeoclimate Research, Department of Earth Sciences, Downing Street, University of Cambridge, Cambridge, CB2 3EQ, UK 2 now at, British Antarctic Survey, High Cross, Madingley Road, Cambridge, CB3 0ET 3 now at, School of Earth, Ocean and Planetary Sciences, Park Place, Cardiff University, Cardiff CF10 3YE *Deceased 24th January 2006

ABSTRACT A core from the western end of the subtropical gyre (ODP Leg 172 Site 1060, 31 45N, 74 24W, 3400 m water depth) in close proximity with the Gulf Stream was studied at high resolution to clarify the surface hydrology changes involved during the last interglacial, their relations and impact over deep circulation as well as continental climate as recorded by pollen. The occurrence of high-frequency cold events at subtropical latitudes is also examined. For the time span from 134 to 108 kyr, eight distinct periods can be recognised on the basis of planktonic foraminifera assemblages. These faunal changes are related mainly to changes in the summer stratification tracing the Gulf Stream current influence and late winter mixing controlling the formation of the subtropical mode water. In addition, our results indicate that during the deglaciation, the sea level rise first during 4 kyr, while sea-surface temperatures stay cold. Only at 128 kyr do the two vary in phase sharply and rapidly; after that date the Gulf Stream increases its influence and reaches a maximum on the core site after 125 kyr. Therefore, the onset of the warm current seems to predate by 2 kyr the onset of the Eemian on land as recorded off Portugal. During the interval 122–116, the sea-surface temperatures are maximum in winter but start to decrease in summer. Superimposed on this cooling trend, we found several

high-frequency cooling events better marked in winter (around 2 C). For most of these, cooling peaks of lithic concentration ð90150 mmÞ probably representing periodic ice rafting to the core site were found. We found that the repeated pacing of these events is about 1400 years. Keywords: sea-surface temperature, last interglacial, climate variability, Gulf Stream, planktonic foraminifera, stable isotopes 20.1 INTRODUCTION For the last glacial period, ice cores from Greenland have provided evidence that dramatic changes have occurred during ice age involving as much as 10 C temperature changes within few decades. It is admitted that these changes have happened simultaneously with sea-surface temperature (SST) fluctuations in North Atlantic area. Moreover, these high-frequency SST fluctuations and changes in the North Atlantic deep flow have been shown to be linked to variability of northward surface heat transfer (Broecker et al., 1992, Keigwin and Jones, 1994; Labeyrie et al., 1995; Maslin and Shackleton, 1995; Chapman and Shackleton, 1998). For interglacial periods, results on mid-latitude North Atlantic cores reveal abrupt events expressing a millennial scale climate cycle operating independently of the glacial/ interglacial alternations (Bond et al., 1997). During the Holocene, subtropical cores

290

M.J. Vautravers, G. Bianchil and N.J. Sackleton*

from the Sargasso Sea record SSTs 1 C cooler during the Little Ice Age and 1 C warmer SSTs during the Medieval Warm interval than at present (Keigwin, 1996). For the previous interglacial stage [marine isotopic stage 5e (MIS 5e)] various climatic proxies from marine, ice and terrestrial records are indicative of either stability (Woillard 1978, Keigwin et al., 1994; McManus et al., 1994; Adkins et al., 1997; Chappellaz et al., 1997; Bauch et al., 2000) or instability (Dansgaard et al., 1993; Guiot et al., 1993; Cortijo et al., 1994; Field et al., 1994; Thouveny et al., 1994; Seidenkrantz et al., 1995; Fronval and Jansen, 1996; Maslin et al., 1996, 1998; An and Porter, 1997; Seidenkrantz and Knudsen, 1997; Frogley et al., 1999; Oppo et al., 2001). Recently, Muller and Kukla (2004), building on previous palynological studies from a range of sites in Europe, have linked the southward displacement of the North Atlantic Current to that of the tree line in southern Europe at the end of the last interglacial interval. The Gulf Stream is the precursor of North Atlantic Current in the western subtropical

90°

85°

80°

Atlantic in proximity of North America. Since site ODP 1060 has been shown to experience fluctuations that can easily be correlated with the climate recorded in the Greenland ice as demonstrated for MIS3 (Vautravers et al., 2004), it is of importance to document how this warm current fluctuates during the previous interglacial. We have carried out a detailed multiproxy study, based on planktonic foraminifer counts, lithics counts and stable isotopes at a very sensitive location for the present climate system. ODP site 1060 (31 459N, 74 249W, 3400 m) is located below the opposing flow of the Gulf Stream, and the Deep Western Boundary Current (DWBC) (Fig. 20.1). The surface current is thought to be the most important cross-latitudinal carrier of heat and salt and is a key element of present-day global thermohaline circulation. These two are the main shallow and deep branches parts of the global conveyor belt in the North Atlantic (Gordon, 1986). Therefore, this core site provides the opportunity of addressing the question of possible subtropical high-frequency variability on

75°

70°

65°

60°

38°

38°

36°

36°

34°

34°

32°

32°

Site ODP 1060

30°

30°

28°

28°

26°

26°

24°

24°

90°

85°

80°

75°

70°

65°

60°

Fig. 20.1 General bathymetric map of the studied area with location of ODP Leg 172 site 1060 and the broad path of the Gulf Stream current.

Subtropical NW Atlantic Surface Water Variability

the path of the Gulf Stream and its possible consequences for thermohaline circulation during the previous interglacial. 20.2 ENVIRONMENTAL BACKGROUND 20.2.1 Location of the site The Blake Bahamas Outer Ridge (BBOR) is a sedimentary drift extending from the eastern continental margin of North America to the abyssal depths. The ridge is thought to be the result of cross-influences of the Gulf Stream and DWBC, coupled with a significant supply of sediment from the Blake Plateau and the continental margin (McCave and Tucholke, 1986). Because of its high sedimentation rates, the BBOR is a remarkable archive of proxy climate informations. 20.2.2 Surface hydrology The Gulf Stream is one of the best-observed surface currents in the world. From the Florida Straits, the Florida current follows the continental coast to Cape Hatteras, where it turns east. At the Florida Straits, the mean northward transport is close to 30 Sv (Schmitz and McCartney, 1993); approximately 45% of the water is from the South Atlantic. This transport is seasonal, with a peak in late summer. Downstream from the Straits of Florida, the current is increased to 100 Sv by re-circulating gyres. Passing Cape Hatteras, as the flow turns east over deep waters it is producing many meanders, and the flow exhibits two peaks, one in late winter and one in summer. The meanders develop into cold rings, which are transmitted southwards within the warm waters. Conversely, warm rings are produced and transmitted northwards. The meridian position of the Gulf Stream also exhibits a seasonal variability, which is linked to its water transport. Increasing flow enhances the northward transport of vorticity, which is compensated by a

291

south-eastward shift of the current (Marchese, 1999). The monthly temperatures and salinity distribution at the core site (between 0 and 1000 m water depth; Levitus, 1982) indicate that during summer the Gulf Stream has a maximal influence on the core site, with salinity and temperatures at about 36‰ and over 25 C respectively. During the course of the year, these parameters also show the occurrence of late winter mixing above 350 m water depth, which leads to the production of the salty subtropical mode water also called 18 C waters (McCartney, 1982).

20.3 MATERIALS AND METHODS 20.3.1 Material Site 1060 was collected during ODP Leg 172 in order to monitor the variability of the northward heat transport from a subtropical area (Keigwin et al., 1998). The samples presented here are taken from the interval 34.50 to 39.22 m below sea floor (mbsf) in hole 1060 C core 5H, where the presence of last interglacial was recognised using the onboard colour reflectance and magnetic susceptibility records [ODP Leg 172 initial report (Keigwin et al., 1998)]. There is no core break within the studied samples. The sampling resolution is 2 cm except at the base of the interval, where four samples spaced at 50 cm extended the record. The dry weight of the samples is between 8 and 14 g. Prior to analyses, bulk samples were disaggregated in distilled water, and wetsieved over a 63-mm mesh sieve. The coarse size fraction ð> 63 mmÞ was dry sieved over a 150-mm sieve and split until an average number of planktonic foraminifera, about 450 to 600 specimens, was obtained for the micropalaeontological study (Figs. 20.2 and 20.3). We have used the taxonomy of Kennett and Srinivasan (1983). The size fraction under 63 mm obtained after settling of the sediment and drying of the excess water was retained for other studies. Percentage fragments testifying to the quality of

292

M.J. Vautravers, G. Bianchil and N.J. Sackleton*

5d

5e

6

50

G. ruber white 0

20

G. ruber pink 0

G. ruber total 0

G. sacculifer 0

Globiberinoides total

50

30 70

40

G. inflata 0

G. bulloides 0

G. falconensis 0

G. glutinata

20 10

30

N. pachyderma

(right)

0

N. pachyderma

(left)

0 1

G. truncatulinoides (left)

0

30

20

G. truncatulinoides % left 0 100 among G. truncatulinoides total G. menardii P. obliquiloculata

0

0

10

10

0 108

110

112

114

116

118

120

122

124

126

128

130

132

134

Age (kyr BP)

Fig. 20.2 Relative abundances of dominant and subordinate planktonic foraminiferal species.

samples in terms of foraminifera preservation available for micropalaeontology were reported in Bianchi et al. (2001a). 20.3.2 Faunal counts and SST estimates Planktonic foraminiferal census counts were obtained from the > 150-mm size fraction in order to reconstruct SST. During the counting procedure, the subsample was displayed on a regularly gridded tray under the light microscope. Only whole planktonic foraminifera ð> 1=2 testÞ were considered in the faunal determination. Broken tests were counted as fragments (resulting from carbonate dissolution). The percentage of planktonic foraminifera fragment is low (less than 25%) which indicates that the dissolution at

this site was not an important factor during the last interglacial. On the contrary, higher values are found during MIS 5d and late MIS 6 (Bianchi et al., 2001a). As a consequence of low carbonate dissolution, the relative faunal abundances have only minor chances to have been affected by selective dissolution, which could have produced relict assemblages leading to misleading variable SST reconstructions during the last interglacial. The percentages of species were calculated as proportion among whole planktonic tests; the down core distribution of the main species is given in Fig. 20.2. SST estimates are based on a micropalaeontological modern analogue technique from planktonic foraminiferal assemblages (Fig. 20.4). Since the NW subtropical

Subtropical NW Atlantic Surface Water Variability

20.3.3 Terrigeous input

1.0 G. ruber

1

0.0 0.0 2

G. inflata –1.0 0.0

3 G. truncatulinoides (l) –1.0 0.5 G. bulloides –0.5 0.5

N. pachyderma (r)

–0.5 0.5

–0.5 0.5

293

4

5

6 Dissolution Summer stratification

7

–0.5 108 110 112 114 116 118 120 122 124 126 128 130 132 134

Age (kyr BP)

Fig. 20.3 Down core evolution of the seven faunal factor loading in order of importance from top to bottom: (1) G. ruber (w), (2) G. inflata, (3) G. truncatulinoides (l), (4) G. bulloides, (5) N. pachyderma (r), (6) dissolution, (7) summer stratification.

Atlantic has a rather poor coverage in core top samples available for faunal transfer functions, we chose to use the SIMMAX method (Pflaumann et al., 1996), because this last method use core top samples from both the north and the south Atlantic. The SIMMAX transfer function (Pflaumann et al., 1996) is configured to determine the five best analogues among a database of 947 core top samples from the South and North Atlantic (Pflauman et al., 2003) with a dissimilarity threshold of 0.9. The use of surface sample from the south Atlantic is desirable and advantageous for this study because it provides a greater choice of analogue samples corresponding to closest hydrographic conditions. The error bar associated with the Simmax method is about 1 C.

Site 1060 is located far from any direct terrigenous input sources. Only very fine terrigenous particles arrive at the site. These are transported by the DWBC from the Northeast Bermuda Rise, where the overlying deep water is the most turbid in the basin (Biscaye and Eittreim, 1977) as a result of the advection of clays and silts by the deep Gulf Stream return flow (Laine and Hollister, 1981). The ultimate source of this terrigenous sediment is probably eastern Canada. Detrital grains larger than 150 mm are exceedingly rare and can only be transported by icebergs at this subtropical latitude. Only one quartz grain of this size was found in the 111 samples examined for the micropalaeontological study. Since the earlier works of Ruddiman (1977) and Fillon et al. (1981) in the north Atlantic smaller grains, provided they are larger than 63 mm, have been unambiguously linked to this mode of transport (McCave et al., 1995; Bond et al., 1997). We have examined our samples in the 90- to 150-mm fraction size. Lithic particles were counted among other biogenic components (Fig. 20.4). Owing to their small size and low abundance, they were counted in a single group. Although identification was nonsystematic, they consisted mainly of clean quartz, clear volcanic glass, some quartz with hematite coating and rare dark minerals. Micas and pyrite grains were not taken into account, because micas can too easily be displaced under strong deep-current activity, while pyrite can be formed in the sediment through the diagenesis of organic material.

20.4 THE TEMPORAL FRAMEWORK Figure 20.4 shows the benthic 18 O sequence versus time. We defined the so-called benthic 18 O ‘plateau’, indicating a high sea-level stand in the core between 36.22 and 35.60 mbsf. The average 18 O value for ‘the plateau’ is 3.22‰ based on 34

294

M.J. Vautravers, G. Bianchil and N.J. Sackleton* 112

114

116

118

120

122

124

126

128

130

132

134

3.3

n lithic (90–150 µm)/g

25

3.8

20 4.3 15

16

14

4.8

BENTHIC δ18O

2.8

10 5.3

5 15

0 27.5

13

Summer

9

11 12

7

1

5

6

8

2 4 3

10

26.5 25.5 24.5

SST (°C)

23.5 22.5 21.5

Winter

20.5 19.5 18.5 17.5 16.5 15.5 108

110

112

114

116

118

120

122

124

126

128

130

132

134

Age (ka)

Fig. 20.4 Reconstruction of surface conditions between 134 and 108 kyr at site 1060: From bottom to top. Simmax sea-surface temperatures for winter, summer, lithic concentration in the 90- to 150-m fraction size, benthic 18O from Bianchi et al. (2001a).

measurements. More rigorously, there is no real ‘plateau’, and the benthic 18 O shows an increase of 0.4‰ between 3 and 3.4‰ during the interval 128–116 kyr; nevertheless to be consistent with other authors we keep the denomination of ‘plateau’. We keep in mind that this 0.4‰ increase in the 18O ratio most probably corresponds to two degree cooling of the water masse the depth of the core. We follow Shackleton et al. (2003) and attempt to constrain at least partially the benthic 18 O record of 1060 in terms of radiometric age determinations as well as through the SPECMAP (Imbrie et al., 1984). The dates for the beginning and end of the plateau are derived from coral dating, giving 128 =þ 1 kyr and 116:1 þ= 0:9 kyr (Stirling et al., 1998). At the base of our record (39.22 mbsf), we found the heaviest values (5‰), which we identified as the event 6.2, corresponding to an age of 135 kyr in the chronology of Martinson et al.

(1987). Within the termination, we use a control point at 36.59 mbsf for the transition between marine isotope stages 6 and 5 at 130 kyr with an isotopic value of 4‰ corresponding to a middle slope value between the glacial value and the beginning of the interglacial. Finally, we used the heaviest benthic 18 O values as the peak of substage MIS 5d, with an SPECMAP-based age of 110.79 kyr. The resulting temporal frame is established on the basis of linear interpolation between the five, 18 O control points. From this age model, the sedimentation rate decreases across the 6/5 transition from an excess of 50 cm/kyr during MIS 6 to a low 5 cm/kyr during the 5e plateau. Such dramatic fluctuations in sediment deposition rate are common on the Blake-Bahamas Outer Ridge and occur from at least 2 500 to 4800 m water (Bianchi et al., 2001b). Nonetheless, there is not always a clear and consistent pattern across this depth range

Subtropical NW Atlantic Surface Water Variability

because of shifts in both the hydrologic properties of the DWBC and terrigenous supply through time which may have contrasting effect at sites separated by only a few hundred metre depth (Keigwin et al., 1998). 20.5 RESULTS AND DISCUSSION 20.5.1 Reconstructing sea-surface conditions through faunal variability 20.5.1.1 Dominant species Relative abundances of dominant and subordinate species of planktonic foraminiferal are plotted from top to bottom of Fig. 20.2. At present, G. ruber (white) variety is a subtropical shallow water dwelling species found year round in the Sargasso Sea (Deusser and Ross, 1989). The lowest representations of this species are found in our record at 130 and 111.5 kyr with 10%. A maximum value of 40% occurred at 121.8 kyr. During the first part of MIS 5e (128–123 kyr), this species shows an increasing trend towards the end of the interval. The pink variety of G. ruber is the shallow water dweller species and prefers warm stratified summer surface waters (Deusser and Ross, 1989). Present relative abundances of the two varieties, deduced from core top samples show that the pink variety prefers warmer water. At site 1060, its abundance peaks during the first half of the 18 O plateau reaching 12% and decreases thereafter to reach minimum values after 112.5 kyr. The abundance of this species is interpreted as a direct indicator of the Gulf Stream influence over the site. Altogether, the two varieties of G. ruber make up 35 to 45% of the total fauna between 128 and 116 kyr. G. ruber is tolerant of extreme water salinity (high or low). But in particular, the pink variety is found in the Caribbean Sea associated with lenses of low-salinity surface water; therefore the pink variety would develop at the beginning of stage 5 under the influence of the lower salinity water of the Gulf Stream during summer in

295

comparison with the salinity of the surface gyre water. On the basis of this species (both varieties), the first part of stage 5e would certainly appear as the warmest. As for the previous species, G. sacculifer is a surfacewater dwelling species, living within the 50 m of water depth (Hemleben et al., 1989). It also bears dinoflagelate symbionts and develops preferentially in summer under oligotrophic conditions. This species is commonly used to trace tropical water masses. In the Atlantic, G. sacculifer develops preferentially over G. ruber when the salinity is between 34.5 and 36‰ (Be and Hutson, 1977). Its abundances within the studied interval of site 1060 vary from 0 to 20%. Peak values over 12% are observed between 124.5 and 119 kyr. During the first part of the plateau from 128 to 124.5 kyr G. sacculifer shows low percentages. If one were to interpret the values of G. sacculifer only in terms of temperature, it would appear that the warmest period was the later part of the ‘plateau’. On the same basis, the 130- to 128-kyr interval would indicate a slight increase of surface temperatures compared to the previous period. It is more likely, however, that changes in surface salinity also account for the changes in the relative abundances of the two main globigerinoides species during stage 5. The entire globigerinoides genus represents up to a maximum of 60% of the total fauna. The highest values of this genus clearly and unambiguously delimit the onset of the interval with the warmest summer conditions from 128 kyr under the influence of an active Gulf Stream system. However, the end of the interglacial around 116 kyr is not so clearly delimitated by this faunal group. We should note the difference between the beginning of this interval well marked by the sudden increase of the genus, whereas its end (at 116 kyr) occurs without any abrupt change. The values at 116 kyr are close to the middle values between the maximum and the lowest values. More generally, we notice a continuous decrease of the genus throughout the interval from 122 to 110 kyr. This shows (i) that the sea-surface

296

M.J. Vautravers, G. Bianchil and N.J. Sackleton*

conditions have already deteriorated during the so-called plateau; and (ii) that the end of the plateau is not characterised by any abrupt change of the surface conditions but that on the contrary the surface conditions were changing gradually. 20.5.1.2 Productivity-related species We have examined the fluctuations of other colder species like G. inflata, which is a transitional species (Be and Hutson, 1977) living in the winter–spring mixed layers. This species flourishes in high phosphate waters, where it feeds on diatoms. In the studied interval, this species fluctuates between 30 and 2% of the assemblage; the highest values are found before 129 kyr and after 112.5 kyr delimiting the coldest intervals. The lowest values are found around 122 kyr, probably indicating the time of the warmest winter season. Also considered as a cold species, G. bulloides is a shallow dwelling species with a preference for waters sustaining high primary productivity, from the subpolar to the subtropical domain (Deusser et al., 1981; Thunell and Renoylds, 1984). In the Sargasso Sea, it lives year round but is at its lowest abundance during summer (Deusser and Ross, 1989). In the record, G. bulloides reaches a maximum of 10%, values over 6% are found between 116 and 112 kyr; this maybe either linked to colder summer or higher primary productivity. Two periods of minimal abundance are found, between 126 and 125 kyr and between 112 and 109 kyr. The last of this interval may be due to selective carbonate dissolution, which also increases the relative representation of G. inflata. G. falconensis is often considered to be the warm water variant of G. bulloides and bears symbionts, whereas G. bulloides does not. G. falconenis is most abundant between 120 and 111 kyr, reaching 7% with significant fluctuations around 115 kyr. Clearly, the two species G. bulloides and G. falconensis evidence progressive cooling starting from 120 in the plateau; at first the G. falconensis

dominates later G. bulloides dominates. G. falconensis is absent in the two relative cold interval centred at 129 kyr and 109 kyr. N. pachyderma right is a subpolar species, with its highest abundances (over 10%) restricted to the transition 6/5, and abundances of about after 114 kyr with 5%. N. pachyderma left polar species was present only during the first cold period (131–129 kyr). Among recent planktonic foraminifera, G. glutinata is the most widely distributed species in the modern ocean. It is considered to be a cosmopolitan subtropical species, which tolerates a wide range of salinities and temperatures feeding on diatoms; this species develops within area of eddies where there is good nutrient mixing within the subsurface as this is the case in the Carribean Sea (Schmuker and Schieble, 2002). It makes up over 10% of the assemblage between 128 and 123 kyr and after 117 kyr, possibly indicating the time of maximal occurrence of eddies on the core site. 20.5.1.3 Deep dweller species Species of the genus Globorotalitae are all deep-water dweller species. Among them, G. truncatulinoides, and in particular the left coiling variety of this species, characterises deep winter mixing as evidenced by isotopic analysis of adult specimens (Deusser and Ross, 1989). In the north Atlantic surface, sample G. truncatulinoides (left) is found between about 20 N and 35 N. The right coiling variety is more restricted to the upper water column, indicating shallower winter mixing, and is also more abundant north of 35 N (Ottens, 1992). The Gulf Stream also delimits the western boundary between a left coiling province (Sargasso Sea) and a dextral coiling province (Caribbean and equatorial area) (Be and Tolderlund, 1971). The influence of differential winter mixing also applied for these two provinces. On the basis of the percentage of left coiling specimens among the total G. truncatulinoides population, we may discriminate the influence of the two provinces

Subtropical NW Atlantic Surface Water Variability

of origin on the core site at anytime. During the studied time interval, the most contrasted period is the end of MIS 6 (132– 130 kyr) characterised by very deep winter mixing in the Sargasso Sea. On the contrary, the interval 130–126 kyr shows an extremely reduced mixing. This latter pattern could be explained by an increasing intensity of the Gulf Stream water coming out of the Gulf of Mexico and the Caribbean Sea through Florida Strait and expanding southwards over the core site at the end of the first period. Again increased winter mixing characterised the 124- to 116-kyr time slice. During the 116- to 114-kyr interval, the winter mixing seems to have been less efficient, and the dominant dextral form attests of the Caribbean Sea influence during this interval. From the temporal evolution of the relative abundance of the left and right variety of G. truncatulinoides, we recognised three intervals of significant reduced winter mixing during the interval 130–126 kyr, at about 120 kyr and again between 115 and 113.8 kyr. We should notice that strong winter cooling leading to enhanced winter mixing in the subtropical gyre is consistent with positive phase of the meteorological anomaly recognised by (Dickson et al., 1996). On this basis, the second part of the last interglacial would be associated with the most intense meridional surface circulation. 20.5.1.4 Minor species Several other minor species are significant for the climatic interpretation. Among them, G. menardii (including the related G. tumida) is present in percentages exceeding 1% between 129.4 and 113.3 kyr. Moreover, during the termination, G. menardii appeared at the core site at 129.5 kyr, preceding the beginning of the ‘plateau’ by 1.5 kyr. Taken at face value, this early reappearance of G. menardii could indicate (i) not only an early warming on the core site (ii) but also more likely points to the water originating from the transfer of south hemisphere Indian Ocean surface and subsurface water

297

through an increase of both the Aghulas Current and the Gulf Stream to predate the time of full Interglacial conditions in the North Atlantic as recorded on the BBOR by large shift of the benthic 18 O towards the lightest values. This associated early warming although rather moderated is confirmed by the Simmax temperatures as shown on Fig. 20.4. But during this particular interval, G. menardii co-exists with the two cold species N. pachyderma right and N. pachyderma left, this last species representing 1% of the fauna at 129.5 kyr. The only explanation for the co-existence of these species is that the coldest episode for the studied interval also corresponds to a period when evidence of warming exists. This could happen at a seasonal and/or interannual timescale or even larger scale or/and traces the existence of cold rings over the sites. Whatever the scale, it seems that the sporadic southward penetration of coldest North Atlantic influence (Labrador current) co-exists with the opposite warm water influence emerging from the Caribbean Sea through the Gulf Stream current. For example, among the warm water species, P. obliquiloculata and G. sacculifer show the same pattern as G. menardii during this cold interval reinforcing our previous interpretation of the record. It should be noted that aside from being indicators of the Gulf Stream activity, these three species are also dissolution resistant; therefore a selective dissolution event cannot be completely ruled out at that time, but neither the low percentage of fragments nor the benthic 13 C (Bianchi et al., 2001a, 2001b) can support this hypothesis. It is more likely that during this initial warming interval, the surface water remained cold as evidenced by the presence of N. pachyderma left and characterised by an important stratification as evidenced by the absence of the deep dweller species G. truncatulinoides. In these conditions, tropical species could have developed during rare but warm summer months; alternatively they could have been transported at subsurface depths from a southern area through the operation of a

298

M.J. Vautravers, G. Bianchil and N.J. Sackleton*

warm current, while the uppermost surface layer was cool and under subpolar surface water influence. Such hydrological pattern would imply the presence of cold and fresh water overlying the warm and salty water of the subtropical gyre. 20.5.2 Faunal factors A principal component analysis was run on the species counts and resulted in seven factors, accounting for 98% of the total variance and confirms some of the trends seen in the individual species Fig. 20.3. The first factor accounts for 45.8% of the total variance and content of the tropical species. The dominant species is the tropical species G. ruber (white), and the subordinate species are G. ruber (pink), G. sacculifer and G. glutinata. This factor shows maximum loading from 128 to 113 kyr. The second (transitional) factor accounts for 34.8% of the fauna and is dominated by G. inflata. The other species associated with this factor are N. dutertrei and N. pachyderma right. This transitional factor shows trends opposite to the first factor. The third factor accounts for 11.8% of the total variance and groups G. truncatulinoides left, G. glutinata and G. hirsuta. This factor likely indicates the intensity of the winter mixing, since G. truncatulinoides (left) and G. hirsuta are both deepdwelling species. This factor increases from 130 to 110 kyr. The fourth factor accounts for 2.2% of the total variance. It regroups G. glutinata and G. bulloides with N. dutertrei, G. falconensis and G. truncatulinoides (right). This factor contains subpolar and opportunistic species as well as high productivity species and intermediate winter mixing species. This factor represents mixed influences of the subpolar surface water and subtropical surface water as well as being link orientated towards high productivity environment. It may be an indicator of the intensity of the meandering as well shallowto-intermediate water mixing on the core site. This factor is well represented between 130–124 kyr and 116–111.5 kyr. It could

indicate the development of mature eddies on the core site during the period of maximal flow of the Gulf Stream current. However, it would also point to a second peak of activity of the warm current after the end of the plateau and during the glacial inception (116–111.5 kyr). The fifth factor accounts for only 1.3% of the variance and is clearly dominated by the subpolar species N. pachyderma (right). This factor is restricted to the 130- to 128.5-kyr period. The sixth factor groups together dissolution-resistant species (such as P. obliculocuata and G. tumida) and evolved in a manner similar to the percentage of planktonic fragments. The seventh factor representing only 1.5% of the total variance could be linked to summer stratification and is led by the species G. ruber (pink). This factor is important before 123 kyr. 20.5.3 Sea-surface temperatures During the period of interest, the summer temperatures (Fig. 20.4) range between 21.5 and 27.5 C. The warmest temperature, obtained at 125 kyr, corresponds exactly to the present maximal temperatures observed in July at the core site. The lowest summer temperature estimate is 21.5 C, which correspond to the present winter temperatures at this site and is observed at 129.5 kyr. During the winter season, temperatures range between 15.7 and 25.5 C at 122.5 kyr. The Levitus Atlas indicated that at present the winter temperatures are on average about 22 C for the five winter months (December to April). The SST reconstruction would imply that the winter mixed layer at 122.5 kyr was warmer by almost 4 C than at present; this situation could be related to an extremely significant reduction of winter westerly winds over the subtropical latitudes, leading to a reduced mixing depth within the subtropical gyre. However, the contribution of the deep dweller to the total fauna is not reduced during this interval and contradicts this hypothesis. It is more likely that the warm surface water

Subtropical NW Atlantic Surface Water Variability

characterising the Gulf Stream area in summer was influential on the core site for a longer time interval each year. This finding would need to be confirmed by other proxy data obtained from the area. We notice a significant cooling of 8 C in winter and 6 C in summer between the warmest part of MIS 5e and the coldest part of MIS 5d. The decrease in temperatures started at 123 kyr. The general cooling slope is marked by two cold events at 117.5 and 115 kyr, after which the temperatures increase slightly but do not recover to their previous levels. During this warmer period, the winter–summer gradient seems to be increased. The SST records also clearly display a series of cooling events centred at 127, 123, 118 and 115.5 kyr; the amplitude of these coolings is of about 1 to 3 C. After 120 kyr, the cooling events are more pronounced during the winter season than during summer. The last significant warming, in between two cold events, occurred during the interval 115.7–112 kyr and is more pronounced in summer. After this interval, a general cooling of 4 C in winter and 2 C in summer occurred. 20.5.4 Lithic content Concentrations of lithic grains (number/ gram of dry sediment) were calculated from census counts within the 90- to 150-mm fraction size. During the time span from 132 to 111 kyr, we recognised 15 lithic peaks. Five reached values of over 10 grains per gram, at 130, 121, 114.5 and at 112 kyr. These peaks are numbered from 1 at 133 to 16 at kyr. The event at 133 kyr is not discussed due to the low resolution of the sampling. Nonetheless, we can notice the similar amplitude when compared to the other events. This is surprising since this event is concomitant with massive release of iceberg. All these events correspond to surface water-cooling events described in the previous section on SSTs. One of these icerafting events, at 121 kyr, is more pronounced than the other; this could support the idea of a major

299

cooling event in the middle of the Eemian (Cortijo et al., 1994, An and Porter, 1997, Maslin et al., 1998). During this 21-kyr interval (132–111 kyr), we can identify 15 analogous events, which give an average event rate of one in every 1400 years, and support the idea of pervasive millennial scale variability during MIS 5e as well as during the Holocene (Bond et al., 1997). 20.6 DISCUSSION As a result of this detailed micropalaeontological study of MIS 5e in core 1060. We propose to distinguish eight main hydrographic situations between 132 and 108 kyr within the subtropical domain presently influenced by the Gulf Stream. (1) The first interval 132–130 kyr is characterised by significant deep winter mixing as shown by the coiling ratio of G. truncatulinoides (left). The percentage of G. ruber pink variety summer proxy for stratification stays very low, indicating a very weak influence of the Gulf Stream on the core site. During this interval, the sea level starts rising. (2) The second period, earliest MIS 5e (130– 128 kyr), is characterised by a weak winter mixing, but also by weak summer stratification. During this period, temperatures reached their lowest reconstructed values. Polar and subpolar cold water species are represented over the core site. About the same time, several warm water species, such as G. menardii, G. sacculifer and P. obliquiloculata, returned to this area. This seems to indicate a restart of the Gulf Stream activity in the area, but masked by a cold surface subpolar layer. In terms of the SSTs, this cold episode occurring during termination II could be compared with the cold Younger Dryas of termination I. Extending the comparison between the two terminations, the early warm water species reappearance during the cold

300

M.J. Vautravers, G. Bianchil and N.J. Sackleton*

event (restart of efficient Gulf Stream activity at 130 kyr) provides some support for the surprisingly warm Younger Dryas reconstructed by SST alkenone methods for a subtropical North Atlantic core (Rhu¨lemann et al., 1999). (3) The period of time spanning 128 to 125 kyr shows both shallow winter mixing and strong summer stratification associated with the warmest temperatures only reached at the end of the interval. We interpret this period as the interval of maximal Gulf Stream activity, also corresponding to the local Northern hemisphere summer insolation maximum. (4) The fourth period, from 125 to 123 kyr, saw a decrease of summer stratification and an increase of winter mixing. This pattern can be interpreted as showing a decrease in the intensity of the Gulf Stream, accompanied by an increasing influence of the water of the subtropical gyre over the core site. (5) The 123- to 116-kyr interval is mainly characterised by strong winter mixing, whereas the summer stratification records its minimum over the core site at the end of this period. At the same time, the thermohaline circulation reached its maximum (Bianchi et al., 2001a, 2001b). We observed that although the interval is still within the benthic plateau, SSTs start to decrease from as early as about 122 kyr in summer. This development evidences that the sustained warm conditions are associated with the winter season and not the summer and contrast to previous evidence of warm summer subtropics, while the northern latitudes cooled (Cortijo et al., 1999). Indeed during this interval, winter temperatures are the warmest. At present, at interdecadal timescales, the North Atlantic Oscillation model links the occurrence of strong winter mixing in the Sargasso Sea and the production of a strong subtropical mode mass to the production of

a strong deep cold water mass in the Greenland Sea (Dickson et al., 1996). By analogy, we interpreted the second part of the benthic plateau as a kind of prolonged NAO maximum, leading to the most efficient thermohaline circulation as recorded at 3400 depth (Bianchi et al., 2001a, 2001b). Therefore, it appears that during the previous interglacial, the subtropical North Atlantic conditions are more powerful upon the THC though the efficiency of the winter mixing leading to the production of a strong subtropical mode water mass when it increases rather than by the extent of the summer SST warming. (6) During the interval 116–114 kyr, after the end of the benthic plateau, we found evidence of increased Gulf Stream activity (increase of G. truncatulinoides right), but this time without the cooccurrence of G. ruber pink, which is the mark for summer stratification. During this period, a major ice volume growth occurs. The operation of such heat transfer may have maintained warmth within the middle North Atlantic for a substantial part of stage 5d, explaining why the warm Eemian climate was continued, while the extent of the northern ice sheet almost attained its maximum during that period. (7) During the 114–111 kyr period we see that the summer SST stay warm. (8) On the contrary, after 111 kyr, cooler conditions occurred during both seasons. Superimposed on this general hydrological pattern in eight periods, we found that as far south as 30 N, even during the ice volume minimum, it is possible to find evidence of 15 ice-rafted events, all related to SST coolings. Of the order of 1 to 2 C, these coolings are in the range of those implied using isotopic method, for the Holocene in the same area (Keigwin, 1996). The periodicity of these events follows a 1400-year cycle. This study clearly demonstrates the importance of presenting full

Subtropical NW Atlantic Surface Water Variability

planktonic assemblage results as well as using these data for SST reconstructions in order to understand past climatic changes especially during interglacial intervals when we expect the magnitude of departing events from the average stage to be rather small. REFERENCES Adkins, J.F., Boyle, E.A., Keigwin, L., Cortijo, E., 1997. Variability of the North Atlantic thermohaline circulation during the last interglacial period. Nature 390, 154–157. An, Z., Porter, S.C., 1997. Millennial-scale climatic oscillations during the last interglaciation in central China. Geology 25, 603–606. Bauch, H.A., Erlenkauser, H., Jung, S.J.A. Thiede, J., 2000. Surface and deep water changes in the subpolar North Atlantic during Termination II and the last interglaciation. Paleoceanography 15, 1, 76–84. Be, A.W.H., Tolderlund, D.S., 1971. Distribution and ecology of living planktonic foraminifera in surface waters of the Atlantic and Indian Oceans. In: Funnell, B.M. and Riedel, W.R. (Eds.), Micropaleontology of Oceans, University Press, London, pp. 105–149. Be, A.W.H., Hutson, W.H., 1977. Ecology of planktonic foraminifera and biogeographic patterns of life and fossil assemblages in the Indian Ocean. Micropaleontology 23, 369–414. Bianchi, G.G., Vautravers, M., Shackleton, N.J., 2001a. Deep flow variability under apparently stable North Atlantic Deep Water production during the last interglacial of the sub-tropical NW Atlantic. Paleoceanography 16, 306–316. Bianchi, G.G., Oppo, D.W., McManus, J.F., Vautravers, M., Shackleton, N.J., 2001b. Detailed deep circulation of the sub-tropical NW Atlantic during MIS 5 (135–64) ka. Abstract Book ICP VII, Sapporo, 83–84. Biscaye P.E., Eittreim S.L., 1977. Suspended particule loads and transports in the nepheloı¨ode layer of the Abyssal Atlantic Ocean. Marine Geology 23, 155–172. Bond, G., Showers, W., Cheseby, M., Lotti, R., Almasi, P., deMenocal, P., Priore, P., Cullen, H., Hajdas, I., Bonani, G., 1997. A pervasive millennial-scale cycle in North Atlantic Holocene and glacial climates. Science 278, 1257–1265. Broecker, W., Bond, G., Klas, M., Clark, E., McManus, J., 1992. Origin of the Northern Atlantic’s Heinrich events. Climate Dynamics 6, 265–273.

301

Chapman M., Shackleton N.J., 1998. Millenial-scale fluctuations in North Atlantic heat flux during the last 150 000 years. Earth and Planetary Science Letters 159, 57–70. Chappellaz, J.A., Brook, E., Blunier, T., Malaize, B., 1997. CH4 and 18O of O2 records from Antarctic and Greenland ice: a clue for stratigraphic disturbance in the bottom part of the GRIP and GISP2 ice-cores. Journal of Geophysical Research 102, 26547–26557. Cortijo, E., Duplessy, J.C., Labeyrie, L., Leclaire, H., Duprat, J., Van Weering, T.C.E., 1994. Eemian cooling in the Norvegian Sea and North Atlantic ocean preceding continental ice-sheet growth. Nature 372, 446–449. Cortijo, E., Lehman, S., Keigwin, L., Chapman, M., Paillard, D., Labeyrie, L., 1999. Changes in meridional temperature and salinity gradients in the North Atlantic Ocean (30 –72 N) during the last interglacial period. Paleoceanography 14, 23–33. Dansgaard, W., Johnsen, S.J., Clausen, H.B., DahlJensen, D., Gundestrup, C.U., Hammer, C.S., Hivdberg, J.P., Steffensen, A.E., Sveinbo¨rnsdottir, A.E., Jouzel, J., Bond, G., 1993. Evidence for general instability of past climate from a 250-kyr icecore record. Nature 364, 218–220. Deusser, W.G., Ross, E.H., 1989. Seasonnally abundant planktonic foraminifera of the Sargasso sea succession, deep-water fluxes, isotopic compositions, and palaeoceanographic implication. Journal of Foraminiferal Research 19, 268–293. Deusser, W.G., Ross, E.H., Hemleben, C., Spindler, M., 1981. Seasonal changes in species composition, numbers, mass, size and isotopic composition of planktonic foraminifera settling into the deep Sargasso Sea. Palaeogeography, Palaeoclimatology, Palaeoecology 33, 103–127. Dickson, R., Lazier, J., Meinke, J., Rhines, P., Swift, J., 1996. Long term coordinated changes in the convective activity of the North Atlantic. Progress in Oceanography 38, 241–295. Duplessy, J.C., Shackleton, N.J., Matthews, R.K., Prell, W.L., Ruddiman, W.F., Caralp, M., Hendy, C., 1984. 13C record of benthic foraminifera in the last interglacial ocean: implications for the carbon cycle and the global deep water circulation. Quaternary Research 21, 225–243. Field, M.H., Huntley, B., Mu¨ller, H., 1994. Eemian climate fluctuations observed in a European pollen record. Nature 371, 779–783. Frogley, M.R., Tzedakis, P.C., Heaton, T.H.E., 1999. Climate variability in Northwest Greece during the last interglacial. Science 285, 1886–1889. Fronval, T., Jansen, E., 1996. Rapid changes in ocean circulation and heat flux in the Nordic seas during the last interglacial period. Nature 383, 806–810.

302

M.J. Vautravers, G. Bianchil and N.J. Sackleton*

Gordon, A.L., 1986. Interocean exchange of thermocline water. Journal of Geophysical Research 91, 5037–5046. Guiot, J., Baulieu, J.L., Cheddadi, R., David, F., Ponnel, P., Reille, M., 1993. The climate in western Europe during the late glacial/interglacial cycle derive from pollen and insects remains. Paleogeography, Paleoclimatology, Palaeoecology 103. Hemleben, C., Spindler, M., Anderson, O.R., 1989. Modern Planktonic Foraminifera, Springer, 363 pp. Imbrie, J., Hays J.D., Martinson, D.G., McIntyre, A., Mix, A.C., Morley, J.J., Pisias, N.G., Prell, W.L., Shackleton, N.J., 1984. The orbital theory of Pleistocene climate: support from a revised chronology of the marine 18 O record. In: Berger et al. (Eds.), Milankovitch and Climate, Part 1, D. Riedel, Hingham, MA, pp. 269–305. Keigwin, L.D., 1996. The Little Ice Age and Medieval Period in the Sargasso Sea. Science 274, 1504–1507. Keigwin, L.D., Jones, G.A., 1994. Western North Atlantic evidence for millennial-scale changes in ocean circulation and climate. Journal of Geophysical Research 99, 12397–12410. Keigwin, L.D., Curry, W.B., Lehman, S.J., Johnsen, S., 1994. The role of the deep ocean in the North Atlantic climate change between 70 and 130 kyr ago. Nature 371, 323–325. Keigwin, L.D., Rio, D., Acton, G.D., 1998. Proceedings of the Ocean Drilling Program, Initial reports Introduction, Shipboard Scientific party 172, 7–12. Labeyrie, L., Vidal, L., Cortijo, E., Paterne, M., Arnold, M., Duplessy, J.C., Vautravers, M., Labracherie, M., Duprat, J., Turon, J.L., Grousset, F., Van Weering, T., 1995. Surface and deep hydrology of the northern Atlantic Ocean during the last 150000 years. Philosophical Transaction of the Royal Society of London 348, 255–264. Laine E.P., Hollister C.D., 1981. Geological effects of the Gulf Stream system on the Northern Bermuda Rise. Marine Geology 39, 277–310. Levitus, S., 1982. Climatological Atlas of the World Ocean. National Oceanic and Atmospheric Administration 173. Marchese, P.J., 1999. Variability in the Gulf Stream recirculation gyre. Journal of Geophysical Research 104, 29549–29560. Martinson, D.G., Pisias, N.G., Hays, J.D., Imbrie, J., Moore, T.C., Shackleton, N.J., 1987. Age dating and the orbital theory of the ice ages: development of a high-resolution 0–300 000 year chronostratigraphy. Quaternary Research 27, 1–30. Maslin, M.A., Shackleton, N.J., 1995. Surface water temperature, salinity, and density changes in the Northeast Atlantic during the last 45 000 years: Heinrich events, deep water formation, and climatic rebounds. Paleoceanography 10, 527–544.

Maslin, M., Sarnthein, M., Knaack, J.J., 1996. Subtropical eastern Atlantic climate during the Eemian. Naturwissenschaften 83, 122–126. Maslin, M., Sarnthein, M., Knaack, J.-J., Grootes, P., Tzedakis, C., 1998. Intra-interglacial cold events: an Eemian-Holocene comparison. In: Cramp, A. (Eds.), Geological Evolution of Ocean Basins: Results from the Ocean Drilling Program Special Publications, Geological Society, London, pp. 91–99. McManus, J.F., Bond, G.C., Broecker, W.S., Johnsen, S., Labeyrie, L., Higgins, S., 1994. High-resolution climate records from the North Atlantic during the last interglacial. Nature 371, 326–329. McCartney, M.S., 1982. The subtropical recirculation of Mode Waters. Journal of Marine Research 40, 427–480. McCave, I.N., Tucholke, B.E., 1986. Deep current controled sedimentation in the western North Atlantic. In: The Geology of North America. The Western North Atlantic Region edited by pp. McCave, I.N., Manighetti, B., Robinson, S.G., 1995. Sortable silt and fine sediment size/composition slicing: parameteres for palaeocurrent speed and palaeoceanography. Paleoceanography 10, 593–610. Muller, U.C., Kukla, G.J. 2004. North Atlantic Current and European environments during the declining stage of the last interglacial. Geology 32, n 12, 1009–1012, doi:101130/G20901. Oppo, D.W., Keigwin, L.D., McManus, J.F., Cullen, J.L., 2001. Persistent suborbital climate variability in marine isotope stage 5 and Termination II. Paleoceanography 16, 3, 280–292. Ottens, J.J., 1992. Planktic Foraminifera as Indicators of Ocean Environments in the Northeast Atlantic, PhD Thesis. Pflaumann, U., Duprat, J., Pujol, C., Labeyrie, L.D., 1996. SIMMAX: A modern analog technique to deduce Atlantic sea surface temperatures from planktonic foraminifera in deep sea sediments. Paleoceanography 111, 15–35. Ru¨hlemann, C., Mulitza, S., Mu¨ller, P., Wefer, G., Zahn, R., 1999. Warming of the tropical Atlantic Ocean and slowdown of thermohaline circulation during the last deglaciation. Nature 402, 511–514. Sa´nchez Gon˜i, M.F., Eynaud, F., Turon, J.L., Shackleton, N.J., 1999. High resolution palynological record off the Iberian margin: direct land-sea correlation for the Last Interglacial complex. Earth and Planetary Science Letters 171, 123–137. Schmitz, W.J., McCartney, M.S., 1993. On the North Atlantic circulation. Reviews of Geophysics 31, 29–49. Schmuker, B., Schiebel R., 2002. Planktic foraminifera and hydrography of the Eastern and Northern Carribean Sea. Marine Micropaleontology 46, 387–403.

Subtropical NW Atlantic Surface Water Variability Seidenkrantz, M.-S., Knudsen, K.L., 1997. Eemian climatic and hydrographical instability on a marine shelf in Northern Denmark. Quaternary Research 47, 218–234. Seidenkrantz, M.-S., Kristensen, P., Knudsen, K.L., 1995. Marine evidence for climatic instability during the last interglacial in shelf records from Northwest Europe. Journal of Quaternary Science 10, 77–82. Stirling, C.H., Esat, T.M., Lambeck, K., McCulloch, M.T., 1998. Timing and duration of the Last Interglacial: evidence for a restricted interval of widespread coral reef growth. Earth and Planetary Science Letters 160, 745–762. Shackleton, N.J., Sa´nchez-Gon˜i, M.-F., Pailler, D., Lancelot, Y., 2003. Marine Isotope Substage 5e and the Eemian Interglacial. Gobal and Planetary Change 757, 1–5.

303

Thouveny, N., de Beaulieu, J.-L., Bonifay, E., Cre´er, K.M., Guiot, J., Icole, M., Johnsen, S., Jouzel, J., Reille, M., Williams, T., Williamson, D., 1994. Climate variations in Europe over the past 140kyr deduced from rock magnetism. Nature 371, 503–506. Thunell, R.C., Reynolds, L.A., 1984. Sedimentation of planktonic foraminifera: Seasonal changes in species flux. Micropaleontology 30, 243–262. Vautravers, M.J., Shackleton, N.J., Lopez-Martinez, C., Grimalt, J.O., 2004. Gulf Stream variability during marine isotope stage 3. Paleoceanography V19, doi: 10.1029. Woillard, G.M., 1978. Grande Pile Peat Bog: A continuous Pollen Record for the last 140 000 years. Quaternary Research 9, 1–21.

This page intentionally left blank

21. Abrupt Change of El Nin˜o Activity off Peru During Stage MIS 5e-d Rein, Bert1, Frank Sirocko1, Andreas Lu¨ckge2, Lutz Reinhardt2, Anja Wolf2,3 and Wolf-Christian Dullo3 1

Johannes Gutenberg Universita¨t Mainz, Becherweg 21, 55099 Mainz, Germany Bundesanstalt fu¨r Geowissenschaften und Rohstoffe (BGR), Stilleweg 2, 30655 Hannover, Germany 3 IFM-Geomar, Wischhofstr. 1-3, 24148 Kiel, Germany 2

ABSTRACT High-resolution proxy data for El Nin˜o variability during the last glacial cycle were derived from a laminated marine sediment core from a region whose climatology and oceanography is strongly affected by ENSO variability. The proxies used are the seasurface temperature and the lithic flux from the continent onto the continental shelf that is largely controlled by the river flood discharge after strong El Nin˜o rainfall in northern and northern central Peru. The focus of this paper is on an abrupt, possibly orbitally driven change of El Nin˜o activity within marine isotope stage 5e. A similar sharp decline of El Nin˜o activity is also observed during the middle of the Holocene interglacial 8 kyr ago. Thus, the major last interglacial and the major Holocene periods of El Nin˜o weakness as documented in the sediments commenced during times with very similar seasonal insolation maximum. Whereas El Nin˜o activity strengthened during the late Holocene with more favourable seasonal heating, El Nin˜o activity did not recover during the very early glacial period with even more favourable insolation conditions but changed boundary conditions. Strong El Nin˜o activity during marine isotope stage 5d was postulated after experiments with the Zebiak and Cane ENSO model and so considered as a major source of moisture excess for the build-up of North American ice sheets. The data from Peru reveal that modelled strong El Nin˜o activity

did not occur in reality and was thus not able to supply excessive moisture to the mid- and higher latitudes for the growth of ice sheets. 21.1 INTRODUCTION The El Nin˜o Southern Oscillation is the major source of global interannual climate variability (Philander, 1990). However, little is known about the long-term variability of this phenomenon. The behaviour of this system under different boundary conditions (e.g. global climate) is also unknown. This is due to the sparse availability of long proxy data records from the tropical ENSO region, where proxy variability can be attributed with some probability to ENSO variability. Some records exist for the late and middle Holocene period (Wells, 1990; Diaz and Markgraf, 1992; Sandweiss et al., 1996; Keefer et al., 1998; Fontugne et al., 1999; Rodbell et al., 1999; Corre`ge et al., 2000; Diaz and Markgraf, 2000; Sandweiss et al., 2001; Tudhope et al., 2001; Andrus et al., 2002; Moy et al., 2002; Nu´n˜ez et al., 2002; Cobb et al., 2003; Gagan et al., 2004; McGregor and Gagan, 2004). However, early Holocene data (Markgraf and Diaz, 2000; Moy et al., 2002) or data for the transition from the last glacial maximum (LGM) into the current interglacial (Koutavas et al., 2002; Koutavas and Lynch-Stieglitz, 2003; Skilbeck et al., 2004) are very sparse. Decadal to multidecadal coral time series provide

306

Bert Rein et al.

(occurrence of rainstorms) and oceanography (sea-surface temperature and primary production) of the study area is strongly affected by ENSO variability. A detailed view on the El Nin˜o activity during the last 20 000 years is given by Rein et al. (2005). The present paper shows the proxy data for El Nin˜o activity in Peru during the last interglacial and during the transition into the very early glacial period when ice sheets in the higher latitudes commenced growth and sea-surface temperatures in the tropical Pacific cooled (Lea et al., 2000; Lea, 2004). It compares El Nin˜o activity with similar seasonality of maximum heat flux but different boundary conditions during the late Holocene and 110 kyr ago. 21.2 STUDY SITE AND MATERIAL The study site where core SO147-106KL was taken (80 km off Lima/Peru; 12 039S, 77 39.89W, 184 m water depth, Fig. 21.1) is 11°S

P

ODP 680B

Huacho

PERU c cifi Pa an e Oc

E

10

0

50

00

4SL 12°S

R

Lima

106KL

12

50 km

0 10 0

0

U

20

13°S

15

106KL 10

50 0

00

1 2090 0

160 170 180

information on ENSO variability in the western equatorial Pacific during MIS 5e-d. One coral is from North Sulawesi (Hughen et al., 1999), six are from northern Papua New Guinea (Tudhope et al., 2001). Thus, our view of how ENSO might have behaved in the past relies heavily on model experiments (Clement et al.,1999; OttoBliesner et al., 2003). The Zebiak and Cane (ZC) model (Zebiak and Cane, 1987; Clement et al., 1999) has now been used for nearly two decades to understand and predict ENSO activity (Kukla et al., 2002). The active domain of the model is restricted to the tropical Pacific between 124 E–80 W and 29 N–29 S, and boundary conditions outside the domain are invariable and linearized around a current average climatology (Clement and Cane, 1999). The omission of processes from outside the equatorial Pacific and the simplified assumption of current boundary conditions (e.g. sea-surface temperature, large-scale atmospheric and oceanic circulation and ice volume) may impair the model’s ability to fully simulate palaeoclimatic variation. It is best applied to times when the global climate (boundary condition) was similar to that of today (Kukla et al., 2002). In agreement with proxy data, the model experiments revealed an El Nin˜o weakness during the middle of the Holocene period due to the seasonality of the equatorial equinoctial insolation (Clement et al., 1999). Kukla et al. (2002) discussed a theoretical maximum of El Nin˜o activity in the ZC model (due to seasonality) around 110 kyr and its implications as a source of moisture for the build-up of northern ice-sheets. However, despite the model results, it is not clear that such an El Nin˜o activity maximum really occurred. We present El Nin˜o proxy data that span the time since the last interglacial, 130 kyr ago. The data are derived from laminated marine sediment cores from the highdeposition continental shelf area off Peru (Rein and Sirocko, 2002/2003; Rein, 2003; Rein et al., 2004). The current climate

0

Pisco 78°W

77°W

14°S 76°W

Fig. 21.1 Study area with bathymetry, topography (GTOPO 30) and core locations.

Abrupt Change of El Nin˜o Activity off PERU

located in a small basin on the edge of the continental shelf. The palaeo-water depth at the site and corrected for sediment thickness was about 75 m during the LGM sea-level lowstand (Fairbanks, 1992), and between 210 and 130 m during marine isotope stage 5 (MIS 5) according to the sea-level reconstructions of Shackleton (2000) and Lambeck and Chappell (2001). The basin, the mud lens off Callao, is sheltered by a landward ridge (Fig. 21.1) against erosion by turbidites and coarse-grained mass flows. A 20-m long sediment sequence (106KL) was recovered from this basin during cruise Sonne-147 (Dullo et al., 2000). Oceanography and climate along the coast of Peru are fundamentally linked to ENSO variability. Figure 21.2 shows the position of 106KL below the warm surface water anomaly during the 1997/98 El Nin˜o event. Between the El Nin˜o events, seaward Ekman transport of surface waters by trade winds causes extensive upwelling of cool and nutrient-rich deep water. These waters make the Peruvian continental shelf one of the most bioproductive marine systems. Organic matter decay causes strong oxygen minimum conditions between 50 and 650 m water depths, favouring the preservation of laminated diatomaceous and diatom-bearing oozes (Kudraß, 2000). During El Nin˜o events, upwelling of nutrient-rich water (and therefore 20°N 10°N 0°

–

10°S 20°S NOAA

–5

106KL 150°W

120°W

90°W

–4 –3 –2 –1 0 1 2 3 4 Temperature anomaly (degree Celsius) El Niño January 1998

5

Fig. 21.2 Eastern tropical Pacific sea-surface temperature anomalies during the 1997/98 El Nin˜o (http://www.cdc.noaa.gov/map/clim/sst_olr/el_ Nin˜o_ anim.shtml) and position of site 106KL.

307

bioproduction) is subdued by the deepening of the thermocline caused by the arrival of a downwelling Kelvin wave (Carranza, 1891). Although wind-induced upwellings still occur during an El Nin˜o episode, the depth of the thermocline implies that upwelling is fed with nutrient-depleted surface waters (Arntz and Fahrbach, 1991). Upwelling does not reach below the thermocline into the nutrientrich deeper water levels. Apart from the reduction of marine primary bioproduction in the pelagic level, extensive rainfall reaches into some regions of otherwise (hyper-) arid coastal deserts (Philander, 1990). Precipitation run-off erodes finegrained lithics from soils. These are flushed via rivers into the sea where they are dispersed over hundreds of kilometres along the continental shelf by the Peru Current (0–30 m of water depth) and the southward flowing countercurrent at depth below 30 m (of water depth) (Scheidegger and Krissek, 1982). The surface current reverses its flow direction towards the south with the arrival of the Kelvin wave during El Nin˜o events (Arntz and Fahrbach, 1991), resulting in the southward transport of suspended sediments by both the surface current and the undercurrent during El Nin˜o situations. Since ocean currents transport the fine-grained lithics over large distances, sedimentary archives on the outer continental shelf integrate the discharge of riverine lithic suspension from a great number of river catchments along the Peruvian coast. Sediment remobilization by the undercurrent apparently occurs at water depths between 250 and 400 m, where extensive fields of mudwaves were mapped during cruise Sonne-147 along the continental slope of Peru between 10.5 S and 13 S (Reinhardt et al., 2002). No mudwaves were observed at any water depth 30 km around site 106KL. The sedimentation rate in the mud lens off Callao is at least twice that of all other (more exposed) locations on the outer continental shelf

308

Bert Rein et al.

between 9 S and 14 S. This testifies to this basin being an extraordinary sediment trap. Currently, strong El Nin˜o rainfall occurs especially in the northern part of Peru, whereas strong El Nin˜o rainfall was less regular south of 11 S during the last century (Wells, 1990; Ortlieb, 2000). In the southern part of the catchment area of 106KL, in central Peru between 11 S and 14 S, weak monthly precipitation maxima (2–6 mm) occur during June to September and are not El Nin˜o related. However, these small amounts of rainfall are not known to be capable of generating floods strong enough to disperse large amounts of fine-grained suspended sediments over wide areas of the continental shelf. The contribution of aeolian dust (compared to riverine sediments) can be neglected in the catchment area of site 106KL on the Peruvian continental shelf area (Scheidegger and Krissek, 1982). Therefore, we use lithics in the shelf sediments as a proxy for strong flood inducing El Nin˜o rainfall activity on the continent.

21.3 METHODS 21.3.1 Chlorins (discrete samples at 20-cm intervals) Chlorins were determined in 20-cm intervals according to methods described in Boto and Bunt (1978), Welschmeyer (1994) and Chen et al. (2000) in the laboratories of IFM-Geomar Kiel. Eight millilitres of acetone (90%) was added to 10 mg of homogenized sample powder in two parallel extractions. The samples were ground for 3 minutes in an agate ball mill and centrifuged for 15 minutes. From the resulting clear solution, 1.5 ml was taken (consequently a total of 3 ml for each sample). The samples were exposed to blue light, and red-light emission was measured with a Turner Fluorometer TD-700. The amount of emitted red light is proportional to the concentration of pigments. The standard

deviation of this method was better than 4.7%. 21.3.2 Alkenones (discrete samples at 10-cm intervals) Palaeo-SST estimations are derived from alkenones (1-cm samples at 10-cm intervals) applying the temperature calibration equation of Prahl et al. (1988). Alkenone analyses and temperature calibration was carried out at the BGR Hannover. To extract the alkenones, a Dionex Accelerated Solvent Extractor 200 was used. Between 1 and 2 g of fine-grained sediment per sample was extracted at 80 C and 1200 psi pressure in three cycles using dichloromethane/methanol (95:5; v:v) as the eluent. The extracts were dried under a stream of nitrogen and saponified with 0.5 ml 1-propanolic KOH (5%) for 24 h at 20 C followed by removal of the KOH and coeluting polar compounds by a solidphase extraction using silica gel columns. These purified extracts were analysed by GC with an HP-6890 instrument equipped with an HP PTV Inlet on a DB-1 capillary column (30 m 0:25 mm i.d.; film thickness 0:25 mm) coupled to a flame ionization detector. The samples were injected splitless in dichloromethane using a cool injection program with solvent venting. The carrier gas was hydrogen at a flow of 0.9 ml/min. A temperature programme of 2 min, isothermal at 56 C, 56–150 C at 24 C/min, 150–320 C at 4.7 C/min and 10 min isothermal was used, giving a good separation of all major compounds. The alkenones were identified by retention times. Quantification was performed relative to external calibration with n-C36 alkane (Doose-Rolinski et al., 2001). Precision and accuracy of the alkenone method were checked by repeated measurements ðn ¼ 11Þ of control material from a surface sample of the Peru upwelling area. This material was repeatedly measured between the analyses of the other samples. The reproducibility of the SST estimations

Abrupt Change of El Nin˜o Activity off PERU

of this surface sample was better than 0:2 C ðp ¼ 0:1Þ. 21.3.3 Oxygen isotopes (discrete samples at 10 cm intervals) One centimetre thick sediment slices were sampled at 10-cm intervals and analysed at the IFM-Geomar Kiel (Wolf, 2003). If available in the sample, about 28 tests of the benthic foraminifera Bolivina seminuda were picked from the 125- to 250- mm sediment fraction of each sample. Oxygen isotope measurements were made using a FINNIGAN 252 mass spectrometer. The ratios of 18 O/16O are reported as relative deviation ( notation) with reference to the PDB (PeeDeeBelemnite) standard (National Bureau of Standards NBS 19). 21.3.4 Colour logging data – chlorin and lithics (continuous logging at 2-mm intervals) Pigment and lithic data were derived from photospectrometric logging with a GretagSpectrolino (GretagMcBeth, Switzerland) in the spectral laboratory at Mainz University. With this device, calibrated reflectance spectra (between 380 nm and 730 nm wavelength) were measured on the split sediment cores at consecutive 2-mm intervals. The reflectance spectra display several photosynthesic pigment (chlorin and carotenoid)-related absorption bands. An absorption band minimum at 660–670 nm wavelength is caused by the absorption by early diagenetic chlorophyll derivates (chlorins – mostly phaeophorbid-a (Rein, 2003). The depth (intensity) of the absorption band is a function of the pigment concentration in the sediment. In the same way, the ratio of reflectance between 570 (R570) and 630 (R630) nm wavelength monitors a characteristic change of the reflectance spectrum continuum. A shift of maximum reflectance is correlated with the lithic concentration in the sediment (Rein and Sirocko, 2002/2003; Rein, 2003; Rein et al.,

309

2004). Thirty replicate analyses on 20 samples (600 analyses) reveal a mean standard error of 2 ‰ for the chlorin absorption band, and 1:5 ‰ for the ratio was used as a measure for the lithic concentration. The absorption intensity in the visible part of the electromagnetic spectrum is not a function of the volume (weight) of an absorbing component but of its surface area in a sample. Thus, this method already detects small concentrations of fine-grained components (pigments and lithics), but the signal intensity increases nonlinearly with increasing concentration. A nonlinear increase of signal intensity with increasing concentration, especially of fine-grained components, is frequently described for mixtures of solids (e.g. Pieters and Englert, 1993). Therefore, the absorption intensity data derived from the reflectance spectra had to be transformed to approximated concentrations. First, absorption intensity (A) was normalized (Anorm) to a range between 0 and 1. The zero point for the pigment data could be defined by the theoretical continuum over the pigment absorption band (Rein and Sirocko, 2002/2003). For the lithic data, the two layers with the lowest ratio values (R570/R630), from which only nonweighable traces of lithics could be extracted, represent the zero point. This concentration was set to zero. Transfer functions were applied to these normalized intensities. The transfer function (pigment concentration Anorm 2 ) for the in situ measured pigment data was found by matching the in situ measured chlorin absorption intensity to the concentration of conventionally measured chlorin data after extraction (Rein et al., 2005). The transfer function to approximate the lithic concentrations (lithic concentration Anorm 1:2 ) from the absorption intensity is based on a regression between the normalized spectral signal and the known lithic concentration in a series (Rein et al., 2005). The result of these transformations are relative concentrations of a component,

310

Bert Rein et al.

with a concentration of zero for samples in which the component is not or almost not (lithics) present. A concentration of 1 represents the highest concentration of a component occurring in the core. The highest (not normalized) lithic concentration found in a 2-mm thick sample was 92 weight %. The multiplication of the relative concentrations of a component with the sediment accumulation per time unit results in relative flux/accumulation rates expressed as percentage flux of the maximum flux in the record. The partly stepwise appearance of the flux rate curves depends on the temporal resolution with which accumulation rate changes can be resolved (density of dating points). Short-term extreme fluxes or individual events may be suppressed by this method.

21.4 RESULTS 21.4.1 Dating Dating of the sediments of the last 19 600 years in piston core 106KL is described in Rein (2003), Rein et al. (2004) and Rein et al. (2005). A prominent reflector (10.75–10.81 m of sediment depth in 106KL, top dated to 19 600 cal BP) was identified in the sediment echolot soundings that can be traced in the shelf sediments along the Peruvian coast (Reinhardt et al., 2002). This prominent reflector was dated to the last sea-level lowstand in ODP core 680B (Suess and von Huene, 1988) where the stratigraphy relies on oxygen isotopes of benthic foraminifera. The long ODP core also shows that a comparable strong reflector is found during the preceding sea-level lowstand during MIS 6 (Wefer et al., 1990). The latter was also identified in sediment echograms at Site 106KL, but was not penetrated. Thus, the oldest sediments in 106KL are younger than the sea-level lowstand during MIS 6. The dating of the lower part of the core had to use a rather unconventional method since preservation conditions for

foraminifers are less favourable in the organic-rich sediments. The scarce availability of foraminifera for isotope analysis imposed major problems concerning the temporal resolution and noise in the oxygen isotope data and its value as the standard stratigraphic mean off Peru. For these reasons, the oxygen isotope data of 106KL was only used to test the stratigraphy which was derived from an alternative method. This method relies on the observation by Wefer et al. (1990) that organic matter contents (%Corg) in the ODP cores from the Peruvian continental shelf resemble variations of oxygen isotopes determined on the same cores. Thus, they used organic carbon contents to enhance the resolution of their age model, in sediment section for which oxygen isotope data were not available. Rein and Sirocko (2002/2003) showed that organic carbon contents in the Peruvian shelf sediments can be estimated from one fraction of the organic matter, the derivates of photosynthetic chlorophyll pigments (chlorins). Organic carbon contents are highly correlated ðr2 ¼ 0:96Þ to the chlorin contents of the sediment. The high correlation of photosynthetic pigments with the organic carbon contents is only explicable if the organic carbon of the sediments is to a large degree supplied by the algae biomass. The value of chlorins as a palaeo-productivity proxy has been demonstrated by Harris et al. (1996). The limited use of chlorins in the past is due to the rather time-consuming data acquisition compared to organic carbon measurements. A very fast method to estimate chlorin concentrations was developed on the continental shelf sediments off Peru (Rein and Sirocko, 2002/2003; Rein et al., 2005). This method was meanwhile also successfully applied to marine and lacustrine sediments from other regions. At site 106KL, relative maxima of photosynthetic pigments (Fig. 21.3c – chlorins, not shown – carotenoids) occur during isotope substages 5e, 5c, 5a, 3 and 1, whereas low

Abrupt Change of El Nin˜o Activity off PERU

Hole 893A / California

5.51 5.50

3.0

(c)

3.5 100

Tuning

4.0

4.5 10

1

26

106KL Alkenone SST (°C)

(d)

45AMS 14 C (Rein et al., 2004, 2005)

Two 14C 41 kyr conv. BP

106KL – Chlorine concentration (% of maximum)

δ18O (PDB) (Kennett, 1995)

2.5

(a)

6.0

5.40

5.30

5.20

Martinson et al. (1987) MIS5a 5b 5c 5d MIS5e

4.23 5.0 5.10

2

MIS4 3.30 4.0 4.22

(b)

MIS3

MIS2

311

error bar (reproducibility)

24

22

20

1.0

(e)

1.5

106KL

2.5

δ 18O (PDB)

2.0

3.0

0

10

20

30

40

50

60

70

80

90

100

110

120

130

Age (kyr BP)

Fig. 21.3 Dating of core 106 KL. (a) Marine isotope stages (MIS) (Martinson et al., 1987) with isotope dates and (b) oxygen isotope data from ODP 893A (Kennett, 1995) used for tuning with (c) 106KL pigment record. (grey) Radiocarbon dating of the last 20 kyr supported by Cs and Pb isotope data in the topmost sediments (Rein et al., 2004, 2005). (d) Alkenone-derived 106 KL SST and (e) 106 KL benthic foraminifera oxygen isotope data.

concentrations are observed in 106KL during the substages 5d, 5b, 4 and during the last glacial maximum. The highest chlorin concentrations are found in the sediments of

the Holocene and the last interglacial (MIS 5e). In the eastern Pacific off Peru, nutrients are supplied by upwelling of subantarctic

312

Bert Rein et al.

water. The process that produced the precessional variability of marine primary production is probably related to chemical (nutrient) changes and reduced surface water flux to depth in the southern Ocean (Francois et al., 1997; Sigman and Boyle, 2000). This had consequences for the chemical properties of the upwelled waters in the eastern tropical Pacific and so on the primary production (Loubere, 2000, 2002). It might be assumed that sea-level change and the thereby induced migrations of the upwelling cell (zone of maximum primary production) or changed export of nutrients from the continent into the sea should also affect long-term variability of the bioproductivity proxy. However, minimum productivity (pigments) during MIS 5b, and despite high sea level (Shackleton, 2000; Lambeck and Chappell, 2001), does not support this case. That pigment concentrations increase prior to the lithic flux between 18 and 17 kyr (Rein et al., 2005) during a period of very low sea level and dry climate in Peru also speaks against this. The dating approach of the present study is the tuning of the chlorin concentrations in core 106KL to a high-resolution oxygen isotope stratigraphy. The existing ODP 18 O stratigraphies from the Peruvian continental shelf are rather coarse (Wefer et al., 1990). The Santa Barbara 18 O stratigraphy of Kennett (1995) reveals the highest temporal resolution of all available marine oxygen isotope stratigraphies for the eastern Pacific over the last 140 kyr. Therefore, the Santa Barbara stratigraphy was used to transfer the SPECMAP isotope dates of Martinson et al. (1987) on the 106KL chlorin record. These correlations are supported during MIS 3 by AMS 14C radiocarbon dating at 41 000 years BP (Fig. 21.3, Table 21.1). The positive correlation which was assumed between chlorins and SPECMAP and which was the basis for the tuning is confirmed by 18 O data of core 106KL. These oxygen isotope data were dated with the age model relying on the above-described

Table 21.1 Age model for 106 KL (dating of post-LGM section of the record is described in (Rein et al., 2004, 2005) Isotope substage dates

Sediment depth Age model (cm rounded to (kyr) 0.5 cm)

3.30 4.00 4.22 4.23 5.00 5.10 5.20 5.30 5.40 Minimum 5d 5.51 5.50 5.53 6.00 AMS14C 41680 þ 460= 430

1101.5 1108 1135.5 1152.5 1243.5 1279.5 1343 1389.5 1470 1509 1647 1739 1809 1983 1086.5

50:20 3:8 59:00 5:5 64:00 6:3 68:80 4:2 73:90 2:6 79:25 3:6 90:95 6:8 99:40 3:4 110:80 6:3 115:00 6:3 122:56 2:4 123:82 2:6 125:19 2:9 130:00 3 41.0

Sediment depths in 106 KL and ages and errors (kiloyears (kyr) ago) of isotope substages (Martinson et al., 1987) to which photosynthesis pigment data were tuned (see Fig. 21.3).

tuning of the chlorin concentration data to the SPECMAP. Although the relative dating errors between SPECMAP-derived stratigraphies is at best about 1–2 kyr, the absolute dating error can be several thousand years during MIS 5e (2.5–3 kyr, Martinson et al., 1987) and MIS 5d (6.3 kyr, Martinson et al., 1987) as shown in Table 21.1. This has to be considered with all ages mentioned in the following when the sediment proxy data is compared to astronomical data (insolation/ theoretical heat flux variability) on an absolute timescale.

21.4.2 Sediment accumulation and lithics The mean sample resolution (years per 2 mm of sediment) of proxy data which was derived from reflectance spectra logging is shown in Fig. 21.4. It reveals that the temporal resolution of proxy data differs

Abrupt Change of El Nin˜o Activity off PERU

increased lithic contents but represent sections with increased diatom contents.

0.1

Mean sampling resolution (years/2-mm sample)

313

1

10

21.4.3 Sea-surface temperature (SST)

100

1000 0

20

40

60

80

100

120

Age (kyr)

Fig. 21.4 Mean temporal resolution of photospectrometrically derived proxy data.

largely along the core. Mean sediment accumulation rates were smallest during MIS 2 to MIS 4 (3.4 cm/kyr) and highest during the late and early Holocene ð 100 cm=kyrÞ. Sedimentation rates in these parts of the Holocene are higher than in any other core section. Mean Holocene or late glacial sedimentation rates (both 5060 cm=kyr) are only reached during the earlier part of the last interglacial and during Termination II ð 45 cm=kyrÞ. Since mean sedimentation rate is averaged over longer periods in the pre-LGM part of 106KL record (coarser grid of ages), shorter periods with significantly higher deposition might not be resolved. During the middle of MIS 5e ( 122:5 2:5 kyr ago due to the age model), mean sediment accumulation dropped by an order of magnitude until 115 6:3 kyr ago and dropped again by another order of magnitude during stages 4 through 2. The drop of sedimentation rate during the middle of MIS 5e occurs with a rapid change of sediment characteristics within 20 cm of sediment. The early part of MIS 5e and the sediments deposited during Termination II are characterized by yellowish diatomaceous sediments with higher contents of lithics (right column of core photographs in Fig. 21.5). After the second SST maximum (marked with encircled ‘3’ in Fig. 21.5), sediment becomes darker, and yellowish layers are not related to

The modern sea-surface temperature off Peru shows seasonal and interannual variability. Current monthly means in the study area are between 15 and 22 C, with the cooler temperatures occurring during May to October/November and higher temperatures from January to April (monthly seasurface temperature maps are available at the website of the `lnstituto del Mar del Peru; http://200.60.133.147/uprsig/sst_prov.html). The annual mean is about 18 C. The cool Peru-Chile current and especially the upwelling of cool water are responsible for the cold sea-surface temperature that is observed along the South American west coast. The surface water temperature is normally 5–7 C cooler along the coast than 1000 km off the coast at 12 S (SST maps provided by the Instituto del Mar del Peru). The alkenone record from which our SST estimations are derived ends in the 1960s. The estimated temperature for the uppermost sample is 21:3 0:2 C (Fig. 21.3d) (minimum error due to reproducibility). This is the coolest reconstructed SST over the last 500 years (range 21.3–21.9 C; Rein et al., 2005) and also a rather cool temperature compared to most of the Holocene and late glacial samples. However, it is in the upper range of modern SST variation. This is not surprising because the coccolithophoridae which synthesize the analysed alkenones occur preferentially during the months with warmer SST and less nutrient availability. Coccolithophoridae reach their maximum abundance during the El Nin˜o warm water anomalies when surface water is nutrient depleted. Thus, the alkenone thermometer is here preferentially a recorder of warm water conditions. SSTs were 2–3 C warmer during the last interglacial period than during the

314

Bert Rein et al. Age (kyr BP)

100

100 Sea level (m below present)

2. SST max 1

kyr BP

0

110

120

130

(a)

–50

(Lambeck and Chappell, 2001) 1

109

5

115

100

0.1

10

0.01

(c) 1

4 117.5

1. SST max

26

2 1 (d)

SST (°C) (alkenones)

127

119.5

3

Lithics – flux rate (% of maximum)

125

(b) Sediment accumulation (mm/yr)

SST min

0.1

25

4

3

24

5 23

22 100 122.5

2

110

120

130

Age (kyr BP) 130

High sediment accumulation Low

x

Lithological boundaries

Fig. 21.5 Proxy data 100–130 kyr ago. Left: Core photographs with ages (black numbers) and lithological boundaries (numbered dots). Right: (a) Sea level (Lambeck and Chappell, 2001) and 106 KL proxy data, (b) sediment accumulation rate, (c) flux rate of lithics, (d) reconstructed SST with error bars.

Holocene (Fig. 21.3d). The same is observed in other eastern Pacific records: SST was 2 C warmer in the equatorial Pacific (Lea, 2004) and 2–3 C warmer off California (Herbert et al., 2001). The cooling of the sea surface off Peru already began several thousand years before the end of MIS 5e, according to our

age model within the middle of MIS 5e. SST cooling already levelled off during the inception of the last glacial between 118 and 116 kyr ago. Temperature variations are less than 1 C between 40 and 120 kyr, only the MIS 4 signal appears as a more prominent temperature decline of 1–2 C (Fig. 21.3d).

Abrupt Change of El Nin˜o Activity off PERU

After 40 kyr BP, SSTs decline by 4 C to the LGM minimum temperatures of 18.6 C. This is about 2–3 C below the temperature level of the last centuries and in accordance with SST changes observed in the eastern and western tropical Pacific (Lea et al., 2000; Rosenthal et al., 2003; Lea, 2004). 21.5 DISCUSSION Drastic changes of the flux rates occurred during MIS 5e and the very last Late glacial. Figure 21.5 shows the lithic accumulation together with the estimated SST and the sea-level reconstruction of Lambeck and Chappell (2001). There are two important points to consider when looking at the timing of flux changes: (1) The dating of the core that, independent of the relative dating method used, always bears a possible error in absolute age of several thousand years for sediments that old (Table 21.1). This is important for the comparison of the El Nin˜o proxy data with insolation data on an absolute time scale. (2) Strong sedimentation rate changes should ideally coincide with prominent lithological boundaries. If the latter is not the case, then sedimentation rate change can also occur somewhere between the dating points used. Prominent lithological changes occurred during stage 5e and the early glacial (MIS 5d), according to our age model at 124, 122.5, 117.5 and 109 kyr ago (Fig. 21.5). The lithological boundaries at 124, 122.5 and 109 kyr coincide with breakpoints in the sediment flux, whereas the change of sediment flux at 115 kyr is without any change of sediment characteristic at that point of the core (Fig. 21.5). A prominent lithological change occurred already at 117.5 kyr ago. So, this might be the real breakpoint where flux rate should change to a lower level. The flux of lithics increased during Termination II and during the early stage 5e. It culminated sharply between 124 and 122.5 kyr ago during the MIS 5e SST maximum, which is contemporaneous with the

315

SST maxima observed in the equatorial Pacific (Lea et al., 2000; Lea, 2004). The lithic flux before 122.5 kyr ago looks very similar to that during the last late glacial and early Holocene, including the sharp drop after the 5e and Holocene maxima (Fig. 21.6). The abruptness of this drop (the Holocene drop and the sharp increase about 3000 years later are reproduced in other cores from different sedimentary basins on the Peru continental shelf; Rein, unpublished data) implies that some kind of threshold was passed. As opposed to the late Holocene when the lithic flux recovered, it apparently did not recover to stage 5e strength during stage 5d, although mean seasonal heat flux conditions were favourable (Fig. 21.6, Table 21.2). A lowered sealevel cannot be the only reason for the low lithic flux (around 110 kyr ago) which is at a similar or even lower level than during the LGM (Fig. 21.6) when the sealevel was significantly lower than during MIS 5d (Fig. 21.6). Keeping the uncertainties regarding the absolute ages of stage 5 ages in mind (Table 21.2), the most prominent flux decrease (by a factor of 10, similarly observed during the middle of the Holocene) occurred at about 122.5 kyr, without any indication of erosion in the core and several thousand years before sea level began to drop. Thus, no hiatus or changed distribution pathways for lithics between river mouths and site 106KL can be made responsible for the drop of lithic flux. This drop was contemporaneous with the end of the MIS 5e SST maximum when SST first dropped and then gradually cooled towards the early glacial SST level. This level is reached between 118 and 116 kyr ago, similar to the Galapagos area (Lea 2004). The lithic flux continued to decline until 109 kyr ago. Assuming current-like conditions, and assuming that most of the lithics reach the continental shelf area off central Peru with El Nin˜o floods, the intensity of El Nin˜o flood events dropped sharply within stage 5e, several thousand years before its end. At that time (on an absolute timescale), a minimum of the seasonality index is approached

316

Bert Rein et al. Age (kyr BP) 0

5

10

Age (kyr BP) 15

20

100

110

120

130 0

(Fairbanks, 1992)

–100

–380

–410

–440

(Lambeck and Chappell, 2001)

–350

–50

–100

–380

–410 100

?

–440

60

10 30

0 1

Lithic flux (% of maximum)

Equatorial Mar–Sep insolation (W/m2) (Berger, 1978)

90

EPICA ? dD ‰ (EPICA community members, 2004)

–50

Sea level (m) below modern

Sea level (m) below modern

0

–30

–60

0.1 0

5

10

15

20

Age (kyr)

100

110

120

130

Age (kyr)

Fig. 21.6 Comparison of the time windows 0–20 kyr and 100–130 kyr ago. (Top) Sea level reconstructions (Fairbanks, 1992; Lambeck and Chappell, 2001) and EPICA (red) deuterium data (EPICA Community Members, 2004). (Bottom) Lithic flux rate and (black) and equatorial March minus September insolation (orange) (Berger, 1978).

and perihelion occurs during boreal late summer. A prominent change of El Nin˜o activity was also observed during the middle of the Holocene (Rein et al., 2005) when the seasonality of heat flux due to insolation was similar (Fig. 21.6). Clement et al. (1999) supplied a mechanism for the middle Holocene El Nin˜o weakness derived from experiments with the

Zebiak and Cane ENSO model. They make a minimum of the seasonality of equatorial insolation responsible. The seasonality is calculated from the difference between the March and September equinoctial insolation. This index is used in the ZC ENSO model as an important forcing element. When the seasonality index is low, the Earth reaches its perihelion position during

Abrupt Change of El Nin˜o Activity off PERU

317

Table 21.2 Comparison of El Nin˜o activity off Peru during the Holocene and MIS 5e Holocene

Marine isotope substage 5e

Dating accuracy (estimated)

Small dating uncertainties: maximum 0:3 kyr

Sedimentation rate (SR)

Much higher than during the glacial period between 100 and 20 kyr ago

Large uncertainties of absolute ages: 2:53 kyr (Martinson et al., 1987/SPECMAP error) Relative to SPECMAP: 12 kyr (absolute age uncertainties important for comparison with insolation) Much higher than during the glacial period between 100 and 20 kyr ago

Lithic flux from the continent onto the continental shelf

Highest SR of the last 130 kyr. Very low SR during the middle of the Holocene period

First half: SR almost as high as during the Holocene period Second Half : SR only 10–20% that of the first half, but distinctly higher than during the glacial part of the record

El Nin˜o activity

High during the early and late Holocene (but with decadal and centennial periods of weakness) Low between 8 (8.9) and 5.6 cal kyr. Abrupt drop of El Nin˜o activity during the middle of the Holocene Low activity coincides with maximum external heat flux during summer/late summer (unfavourable seasonal heating)

High activity during the first half, Low activity during the second half.

Abrupt increase of El Nin˜o activity during the late Holocene with more favourable season of maximum heat flux

September. The maximum heat flux during the boreal late summer and early autumn strengthens the atmospheric convection centre in the western equatorial Pacific, and thus the trade winds. It also shifts the convection centre northwards compared to average conditions. All these, on average, make conditions less favourable for the development of strong El Nin˜o activity. On the contrary, when the seasonality index is high, the Earth reaches its aphelion position during September. The reduced heat flux during the boreal late summer and early autumn weakens the atmospheric convection centre in the western equatorial Pacific and thus the trade winds. Generally speaking, these are favourable conditions for the development of frequent strong El Nin˜o events. Based on the modelled maximum of El Nin˜o frequency, Kukla et al. (2002) proposed

Abrupt drop of El Nin˜o activity during the MIS 5e Low activity coincides with maximum external heat flux during summer/ late summer (unfavourable seasonal heating) El Nin˜o activity does not reinforce after 5e with more favourable season of maximum heat flux. ! changed boundary conditions

that stronger El Nin˜o activity at the end of the last interglacial could have been involved in the build-up of northern ice sheets by supplying surplus moisture along a steepened gradient towards higher latitudes. However, contrary to expectations after studies with the ZC ENSO model, the aphelion situation during September around 110 kyr ago apparently had little effect on long-term El Nin˜o activity in Peru. From the data presented here, it does not look as if El Nin˜o activity was especially strong during the early glacial period and played an outstanding role as a source of moisture for the growing of the North American ice sheet. From this data, El Nin˜o activity was almost as weak as and partly weaker than during the ENSO weakness within the middle of the Holocene. Tudhope et al. (2001) provided decadal to multidecadal coral data

318

Bert Rein et al.

sets with ENSO variability from the western Pacific within several time windows (modern, 2–3 kyr, 6:5 kyr, 38–42 kyr, 85 kyr; 112 kyr, 118–128 kyr, 130 kyr ago) over the last glacial–interglacial cycle. With the exception of the 6.5-kyr old corals, the corals from 112 kyr ago show the lowest ENSO variability of all records, thus matching inferences from above. The existing proxy data do not promote strong El Nin˜o activity during stage 5d as implied by the ZC model. Suppose that seasonality of maximum heat flux is a driver for long-term ENSO variability during the interglacials, then mean climatology (boundary conditions in the ZC ENSO model) within the tropics and/or in the extra-tropics must have been distinctly different from the present and Holocene conditions (ZC model is optimized for present conditions) during stage 5d. The most effective of the boundary conditions might be the SST in the equatorial Pacific because it directly influences the transfer of moisture into the atmosphere. SST reconstructions from the equatorial Pacific reveal 12 C cooler than current temperatures in the western (site ODP 806B; Lea et al., 2000) and eastern equatorial Pacific (site TR163-19; Lea, 2004). So, boundary conditions were certainly different from those used for the ZC model runs. Distinct changes of the mean climatology are also documented outside the tropics. The ice cores from Greenland (NorthGRIP Members, 2004) and from Antarctica (EPICA Community Members, 2004) document a distinctly cooler 5d climate than during stage 5e and during the Holocene. Possible teleconnections from Antarctica into the eastern Pacific are easily found via the Chile–Peru current flowing northwards along the South American coast and via upwelling of water masses from intermediate levels that also formed in the sub-Antarctic region (Sirocko, 2003). Thus, the cooling of surface water off Peru and reduced rainfall in Peru may also reflect Antarctic influences beginning well within the last interglacial (Fig. 21.6).

21.6 CONCLUSIONS Proxy data for ENSO variability in and off Peru are derived from a long, laminated marine core that spans the time of the glacial–interglacial cycle since 130 kyr. The data comprise decadal to centennial resolving alkenone-derived SST estimations and subannual to decadal resolving proxy data for the flux of lithics from the continent onto the continental shelf area off Peru. Current interannual SST variability is linked to ENSO variability, and the amount of lithic supply is controlled by the discharge of river floods after strong El Nin˜o rainfall events in northern and northern central Peru. During the last interglacial, we observed a sharp drop of El Nin˜o activity, similar to that during the Holocene when perihelion occurred during the late summer. However, during MIS 5d, the strength of El Nin˜o activity did not recover with more favourable insolation conditions as was observed during the late Holocene. The strong El Nin˜o activity that was indicated by the ZC ENSO model according to orbital forcing did not occur in Peru at the beginning of the last glacial. The transition into a glacial world apparently changed critical boundary conditions, which are linearized around a current mean climatology in the ZC ENSO model.

ACKNOWLEDGEMENTS This work was supported by the German ‘Bundesministerium fu¨r Bildung und Forschung’ (cruise and core material) and the Johannes-Gutenberg University Mainz providing the position for B.R.

REFERENCES Andrus, C.F.T., Crowe, D.E., Sandweiss, D.H., Reitz, E.J., Romanek, C.S., 2002. Otolith 18 O record of mid-Holocene sea surface temperature in Peru. Science 295, 1508–1511.

Abrupt Change of El Nin˜o Activity off PERU Arntz, W., Fahrbach, E., 1991. El Nin˜o – Klimaexperiment der Natur. Basel, Birkha¨user Verlag, 263 p. Berger, A., 1978, Long-term variations of daily insolation and Quaternary climatic changes. Journal of Atmospheric Sciences 35, 2362–2367. Boto K.G., Bunt J.S., 1978. Selective excitation fluorometry for determination of chlorophylls and pheophytins. Analytical Chemistry 50, 392–395. Carranza, L., 1891. Contracorriente maritime observada en Payta y Pacasmayo. Boletı´n Sociedad Geografica Lima 1, 344–345. Chen R.F., Jiang Y., Zhao M., 2000. Solid-phase fluorescence determination of chlorin in marine sediments. Organic Geochemistry 31, 1755–1763. Clement, A.C., Cane, M.A., 1999. A role for the tropical Pacific coupled ocean-atmosphere system on Milankowitch and millenial timescales. Part I: A modeling study of tropical Pacific variability. In: Clark, P.U., Webb, R.S., Keigwin, L.D., (Eds.). Mechanisms of global climate change at millennial time scales, Washington, DC, American Geophysical Union, 363–371. Clement, A.C., Seager, R., Cane, M.A., 1999. Orbital controls on the El Nin˜o/Southern Oscillation and the tropical climate. Paleoceanography 14, 441–456. Cobb, K.M., Charles, C.D., Cheng, H., Edwards, R.L., 2003. El Nin˜o/Southern Oscillation and tropical Pacific climate during the last millennium. Nature 424, 271–276. Corre`ge, T., Delcroix, T., Re´cy, J., Beck, W., Cabioch, G., Le Cornec, F., 2000. Evidence for stronger El Nin˜oSouthern Oscillation (ENSO) events in a mid-Holocene massive coral. Paleoceanography 15, 465–470. Diaz, H.F., Markgraf, V., 1992. El Nin˜o, Cambridge, Cambridge University Press, p. 476. Doose-Rolinski, H., Rogalla, U., Scheeder, G., Lu¨ckge, A., von Rad, U., 2001. High-resolution temperature and evaporation changes during the late Holocene in the northeastern Arabian Sea. Paleoceanography 16, 358–367. Dullo, W.C., Rein, B., Wolf, A., Biebow, N., Schaber, K., Sirocko, F., 2000. Core descriptions and reflectance spectra, In: Kudrass, H. (Ed.). Cruise report Sonne 147, Peru Upwelling. BGR Hannover, Hannover, Report 0120607, 102–119. EPICA Community Members, 2004. Eight glacial cycles from an Antarctic ice core. Nature 429, 623–628, doi: 10.1038/nature02599. Fairbanks, R., 1992. Barbados Sea level and Th/U 14C Calibration. IGBP PAGES/World Data Center – A for Paleoclimatology Data Contribution Series # 92-020. NOAA/NGDC Paleoclimatology Program, Boulder, CO, USA. Fontugne, M., Usselmann, P., Lavalle´e, D., Julien, M., Hatte´, C., 1999. El Nin˜o variability in the coastal desert of Southern Peru during the mid-Holocene. Quaternary Research 52, 171–179.

319

Gagan, M.K., Hendy, E.J., Haberle, S.G., Hantoro, W.S., 2004. Post-glacial evolution of the IndoPacific Warm Pool and the El Nin˜o-Southern Oscillation. Quaternary International 118–119, 127–143. Harris, P.G., Zhao, M., Rosell-Mele´, A., Tiedemann, R., Sarnthein, M., Maxwell, J.R., 1996, Chlorin accumulation rate as a proxy for Quaternary marine primary production. Nature 383, 63–65. Herbert, T.D., Schuffert, J.D., Andreasen, D., Heusser, L., Lyle, M., Mix, A., Ravelo, A.C., Stott, L.D., Herguera, J.C., 2001. Collapse of the California Current during glacial maxima linked to climate change on land. Science 293, 71–76. Hughen, K.A., Schrag, D.P., Jacobsen, S.B., 1999. El Nin˜o during the last interglacial period recorded by a fossil coral from Indonesia, Geophysical Research Letters 26, 3129–3132. Keefer, D.K., de France, S.D., Moseley, M.E., Richardson III, J.B., Satterlee, D.R., Day-Lewis, A., 1998. Early maritime economy and El Nin˜o events at Quebrada Tacahuay, Peru. Science 281, 1833–1835. Kennett, J.P., 1995. Latest Quaternary benthic oxygen and carbon isotope stratigraphy: Hole 893A, Santa Barbara Basin, California. In: Kennett, J.P., Baldauf, J.G., Lyle, M. (Eds.). Proceedings of the Ocean Drilling Program, Scientific Results, 3–18. Koutavas, A., Lynch-Stieglitz, J., 2003. Glacial– interglacial dynamics of the eastern equatorial Pacific cold tongue-Intertropical Convergence Zone system reconstructed from oxygen isotope records. Paleoceanography 18, 1089–1104. Koutavas, A., Lynch-Stieglitz, J., Marchitto, T.M., Sachs, J.P., 2002. El Nin˜o-like pattern in ice age tropical Pacific sea surface temperature. Science 297, 226–230. Kudraß, H., 2000. Cruise report Sonne 147, Peru Upwelling, BGR Hannover, Hannover, Report 0120607, 177 pp. Kukla, G.J., Clement, A.C., Cane, M.A., Gavin, J.E., Zebiak, S.E., 2002. Last interglacial and early glacial ENSO. Quaternary Research 58, 27–31. Lambeck, K., Chappell, J., 2001. Sea level change through the last glacial cycle. Science 292, 679–685. Lea, D.W., 2004. The 100 000-yr cycle in tropical SST, greenhouse forcing, and climate sensitivity: Journal of Climate 17, 2170–2179. Lea, D.W., Pak, D.K., Spero, H.J., 2000. Climate impact of late Quaternary equatorial Pacific sea surface temperature variations. Science 289, 1719–1724. Loubere, P., 2000. Marine control of biological production in the eastern equatorial Pacific Ocean. Nature 406, 497–500. Loubere, P., 2002. Remote vs. local control of changes in eastern equatorial pacific bioproductivity from the last glacial maximum to the present. Global and Planetary Change 35, 113–126. Markgraf, V., Diaz, H.F., 2000. The past ENSO record: A synthesis. In: Diaz, H.F., Markgraf, V. (Eds.). El

320

Bert Rein et al.

Nin˜o and the Southern Oscillation – Multiscale variability and global and regional impacts, Cambridge, Cambridge University Press, 465–488. Martinson, D.G., Pisias, N.G., Hays, J.D., Imbrie, J., Moore, T.C., Shackleton, N.J., 1987. Age dating and the orbital theory of the ice ages: development of a high resolution 0 to 300 000-year chronostratigraphy. Quaternary Research 27, 1–29. McGregos, H.V., Gagan, M.K., 2004. Western Pacific Coral d18 O records of anomalous Holocene variability on the El Nin˜o-Southern Oscillation. Geophysical Research Letters 31, doi: 10.1029/2004 GL019972. Moy, C.M., Seltzer, G.O., Rodbell, D.T., Anderson, D.M., 2002. Variability of El Nin˜o/ Southern oscillation activity at millennial timescales during the Holocene epoch. Nature 420, 162–165. NorthGRIP Members, 2004. High resolution record of Northern Hemisphere climate extending into the last interglacial period. Nature 431, 147–151, doi: 10.1038./nature02805. Nu´n˜ez, L., Grosjean, M., Cartajena, I., 2002. Human occupations and climate change in the Puna de Atacama, Chile. Science 298, 821–824. Ortlieb, L., 2000. The documented historical record of El Nin˜o events in Peru: An Update of the Quinn record (sixteenth through nineteenth centuries), In: Diaz, H.F., Markgraf, V. (Eds.). El Nin˜o and the Southern Oscillation – Multiscale variability and global and regional impacts, Cambridge, Cambridge University Press, 207–295. Otto-Bliesner, B.L., Brady, E.C., Shin, S-I., Liu, Z., Shields, C., 2003. Modeling El Nin˜o and its tropical teleconnections during the last glacial-interglacial cycle, Geophysical Research Letters 30, 2198, doi:10.1029/2003GL018553. Philander, S.G.H., 1990. El Nin˜o, La Nin˜a, and the Southern Oscillation. San Diego, Academic Press, 293 p. Pieters, C.M., Englert, P.A.J., 1993. Topics in remote sensing – Remote geochemical analysis: Elemental and mineralogical composition, Cambridge, Cambridge University Press, 594 pp. Prahl, F.G., Muehlhausen, L.A., Zahnle, D.L., 1988. Further evaluation of long-chain alkenones as indicators of paleoceanographic conditions. Geochimica et Cosmochimica Acta 52, 2303–2310. Rein, B., 2003. In-situ Reflektionsspektroskopie und digitale Bildanalyse – Gewinnung hochauflo¨sender Pala¨oumweltdaten mit fernerkundlichen Methoden: Habilitation thesis, University of Mainz, 104 p. Rein, B., Sirocko, F., 2002/2003. In-situ reflectance spectroscopy – analysing techniques for high resolution pigment logging in sediment cores. International Journal of Earth Science 91, 950–954, Erratum v. 92, 143.

Rein, B., Lu¨ckge, A., Sirocko, F., 2004. A major Holocene ENSO anomaly during the Medieval period. Geophysical Research Letters 31, L17211, doi:10.1029/2004GL020161. Rein, B., Lu¨ckge, A., Reinhardt, L., Sirocko, F., Wolf, A., Dullo, C.-W., 2005. El Nin˜o variability off Peru during the last 20,000 years, Paleoceanography 20, PA4003, doi:10.1029/2004PA001099. Reinhardt, L., Kudrass, H.-R., Lu¨ckge, A., Wiedicke, M., Wunderlich, J., Wendt, G., 2002. High-resolution sediment echosounding off Peru: Late Quaternary depositional sequences and sedimentary structures of a current-dominated shelf. Marine Geophysical Researches 23, 335–351. Rodbell, D.T., Seltzer, G.O., Anderson, D.M., Abbott, M.B., Enfield, D.B., Newman, J.H., 1999. An 15 000year record of El Nin˜o-driven alluviation in Southwestern Ecuador. Science 283, 516–520. Rosenthal, Y., Oppo, D., Linsley, B.K., 2003. The amplitude and phasing of climate change during the last deglaciation in the Sulu Sea, western equatorial Pacific. Geophysical Research Letters 30, 1428. Sandweiss, D.H., Richardson, J.B., Reitz, E.J., Rollins, H.B., Maasch, K.A., 1996. Geoarchaeological evidence from Peru for a 5000 years B. P. onset of El Nin˜o. Science 273, 1531–1533. Sandweiss, D.H., Maasch, K.A., Burger, R.L., Richardson III, J.B., Rollins, H.B., Clement, A., 2001. Variation in Holocene El Nin˜o frequencies: Climate records and cultural consequences in ancient Peru. Geology 29, 603–606. Scheidegger, K.F., Krissek, L.A., 1982. Dispersal and deposition of eolian and fluvial sediments off Peru and northern Chile. Geological Society of American Bulletin 93, 150–162. Shackleton, N., 2000. The 100 000-year ice age cycle identified and found to lag temperature, carbon dioxide, and orbital eccentricity. Science 289, 1897–1902. Sigman, D.M., Boyle, E.A., 2000. Glacial/interglacial variations in atmospheric carbon dioxide. Nature 407, 859–869. Sirocko, F. 2003. What drove past teleconnections. Science 301, 1336–1337. Skilbeck, C.G., Goodwin, I., Gagan, M., Watson, M., Aiello, I.W., 2004. High-resolution palaeo-El Nin˜o records from Peru continental margin. 32nd International Geological Congress, 2004. Abs. Vol. pt. 2, abs. 227–4, 1033. Suess, E., von Huene, R., 1988. Scientific results - leg 112, Proceedings of the Ocean Drilling Program. 1015 pp. Tudhope, A.W., Chilcott, C.P., McCulloch, M.T., Cook, E.R., Chappell, J., Ellam, R.M., Lea, D.W., Lough, J.M., Shimmield, G.B., 2001. Variability in El Nin˜o-Southern Oscillation through a glacial– interglacial cycle: Science 291, 1511–1517.

Abrupt Change of El Nin˜o Activity off PERU Wefer, G., Heinze, P., Suess, E., 1990. Stratigraphy and sedimentation rates from oxygen isotope composition, organic carbon content, and grain size distribution at the Peru upwelling region; Holes 680b and 686b. In: Suess, E., von Huene, R., participants, c. (Eds.). Proceedings of the Ocean Drilling Program, Scientific Results, 355–367. Wells, L.E., 1990. Holocene history of the El Nin˜o phenomenon as recorded in flood sediments of northern coastal Peru. Geology 18, 1134–1137.

321

Welschmeyer, N.A., 1994. Fluorometric analysis of chlorophyll-a in the presence of chlorophyll-B and pheopigments. Limnology and Oceanography 39, 1985–1992. Wolf, A., 2003. Zeitliche variation im peruanischen Ku¨stenauftrieb seit dem Letzten Glazialen Maximum – Steuerung durch globale Klimadynamik, Doctoral Thesis, University of Kiel, Germany, 88 p. Zebiak, S.E., Cane, M.A., 1987. A model El Nin˜oSouthern Oscillation: Monthly weather review 115, 2262–2278.

This page intentionally left blank

22. Interglacial and Glacial Fingerprints from Lake Deposits in the Gobi Desert, NW China Bernd Wu¨nnemann1, Kai Hartmann2, Norbert Altmann2, Ulrich Hambach3, Hans-Joachim Pachur2 and Hucai Zhang4 1

Interdisciplinary Centre Ecosystem Dynamics in Central Asia, Freie Universitaet Berlin, Malteserstr. 74-100, 12249 Berlin, Germany 2 Institut fu¨r Geographische Wissenschaften, Freie Universitaet Berlin, Malteserstr. 74-100, 12249 Berlin, Germany 3 Lehrstuhl fu¨r Geomorphologie, Universitaet Bayreuth, Universitaetsstr. 3, 95447 Bayreuth, Germany 4 Nanjing Institute of Geography and Limnology (NIGLAS), Chinese Academy of Sciences, China

ABSTRACT A 230-m long sediment core from the centre of the Gaxun Nur Basin, Gobi Desert, NW China provides evidence for climate induced changes in water balance during the last glacial cycle. Millennial scale and short-term variations of geochemical precipitates and grain size show that freshwater fluxes from the Tibetan Plateau by surface run-off were the main controlling factors for lake evolution in the Tibetan dry forelands for about the last 250 kyr. Periods of positive water balance with strong lake extension and reverse developments generally coincide with changes in the global ice volume and with oxygen-18 records from Tibet and Greenland as well, documenting the close relationship between environmental conditions in remote desert regions of NW China and orbitally forced Northern Hemisphere high mountain mid-latitude and high-latitude climates on a regional and global scale. Our data imply that both the East Asian summer monsoon and the extra-tropical westerlies are the major feedback mechanisms for effective moisture supply over NW China. During the 10-kyr long interglacial warmmoist substage 5.5, summer monsoon moisture dominated owing to its strong northward shift beyond the modern

limit. At that time, a large and slightly saline lake filled the entire Gaxun Nur basin as a result of strong river inflow from the Tibetan catchment by melt water supply and by enhanced summer monsoon precipitation. Aeolian transport was weak. The Eemian interglacial in the Gaxun Nur region started at about 129 kyr, with warm and moist environmental conditions between 128 and 121 kyr BP and terminated around 119 kyr, documented by a strong climate shift towards dry conditions and enhanced mobilization of aeolian sand. During interstadial climates, contemporaneous with D/O events in Greenland ice cores, both wind systems most likely supplemented each other, while in transitional phases towards cold conditions, moisture supply by the westerlies seems to have dominated. Cold-dry stages, recorded in the Gaxun Nur core, are synchronous with the global climate. They induced strong lake-level declines and promoted aeolian transport of exposed lake sediments southwards due to the enhanced winter monsoon. Loess records from the Chinese Loess Plateau confirm that the temporal distribution of loess mobilization recorded in the Gaxun Nur sediments was synchronous with depositional phases on the Loess Plateau.

324

Bernd Wu¨nnemann et al.

22.1 INTRODUCTION Rapid climate fluctuations during the last glacial cycle are well documented in ice cores from Greenland (e.g. Dansgaard et al., 1993; NorthGRIP Project Members, 2004) and from Antarctica (Petit et al., 1999), the latter tracing back to the last eight cycles during Quaternary time (EPICA Members, 2004). They obviously have played a major role in reconstructing climate changes. Although it is widely accepted that the main driving force of climate shifts in orbital bands is strongly related to solar radiation (insolation) through time (e.g. Berger and Loutre, 1991; Clement et al., 2001; Leuschner and Sirocko, 2003), feedback mechanisms such as inland ice dynamics, thermohaline circulation and the atmospheric circulation pattern (e.g. the Northern Hemisphere lowand mid-latitude monsoon systems) may have affected land–ocean thermodynamic relations quite differently at regional scales. In particular, the climates over Asia seem to have been strongly influenced by the uplift of the Tibetan Plateau (An et al., 2001), co-controlling the evolution of the Asian monsoon system and the establishment of the Inner Asian dryland belt (Gobi Desert). However, palaeoclimate studies reveal the general high–frequency climate instability during the last glacial cycle, well known from Greenland ice cores and from North Atlantic sediment cores as stadials (Bond cycles and Heinrich events) and interstadials (Dansgaard–Oeschger cycles) on timescales of a few millennia (e.g. Dansgaard et al., 1993; Bond et al., 1997). The recognition of those climate cycles in marine sediments of the Arabian Sea (Schultz et al., 1998; Leuschner and Sirocko, 2003), South China Sea (e.g. Wang et al., 1999), in ice cores from Tibet (Thompson et al., 1997; Thompson, 2000) and in aeolian sediments from the Chinese Loess Plateau (e.g. Porter and An, 1995; Xiao et al., 1995) confirms a close relationship between low-latitude monsoonal climate variability and rapid temperature fluctuations of higher northern latitudes

(e.g. Schulz et al., 1998). A linkage between the Tibetan uplift, monsoon pattern and Northern Hemisphere glaciation seems to be evident (An, 2000; An et al., 2001). On the other hand, very little is known about feedback mechanisms controlling the expansion or shrinkage of arid regions in inner Asia during interglacial and glacial cycles and which of them might be dominant. The special significance of this dryland belt in strongly continental China north of the Tibetan Plateau with respect to climate reconstructions is due to the fact that two different northern hemispheric air masses – the East Asian Monsoon and the extra-tropical westerlies – intersect in this region. Both wind regimes are potential sources for water vapour transport, affecting the regional hydrological systems on the Tibetan Plateau and the desert forelands. Alternations in the summer monsoon strength as one prominent source of water vapour transport from low latitudes into the interior of China seem to have responded synchronously to major shifts in the climate system on a global scale, but on a regional scale the monsoonal effect varied considerably (An et al., 2001). At present, nearly nothing is known about the potential impact of the westerlies system, either compensating the lack of moisture transport by the monsoon system or even enhancing moisture availability along the arid zones of Central Asia. Previous palaeoclimate studies have shown that vast desert areas in northwestern China were covered by large lakes and swamps during marine isotope stage 3 (MIS) and during early and mid-Holocene times (Pachur et al., 1995; Wu¨nnemann et al., 1998; Zhang et al., 2001, 2002, 2004; Wu¨nnemann and Hartmann, 2002; Mischke et al., 2005) as a result of enhanced local precipitation and melt water discharge in periods of glacier advance in the high mountain catchments. All climate proxies in that region trace back to approximately 40 kyr. Here we present a sediment record from the Gaxun Nur Basin in arid north-western

Interglacial and Glacial Fingerprints from Lake Deposits

China which is assumed to cover the sedimentary history of the past 250 kyr. However, we focus on climate-controlled hydrological changes during the last glacial cycle. In this paper, we try to answer the questions (i) whether the water balance of the Gaxun Nur system reflects climate shifts from warm to cold stages and vice versa synchronously with global climate signatures and (ii) whether phases of aeolian activity respond to variations in the wind regime over NW China.

from the Qilian Mts. via the Hexi Corridor (Gansu Corridor, better known as a part of the former Silk Road) towards the Alashan Plateau (Molnar and Taponnier, 1975; Li et al., 1999; Hetzel et al., 2002; Hartmann, 2003), thus forming the ‘belt of left-lateral transpression’ (Cunningham et al., 1996). Consequently, the Gaxun Nur Basin developed as a large pull-apart-basin between these structural elements, now acting as the erosional base for the Hei River drainage system from the south (Tibetan Plateau). This endorheic basin covers an area of approximately 28 000 km2, while the total catchment of the Hei River system, connected with glaciers in the Qilian Mts., comprises roughly 130 000 km2. Along the distal part of the basin, three terminal lakes – Gaxun Nur, Sogo Nur and Juyanze – form a sickleshaped chain. At present, all lakes are dry. At present, climate in NW China is mainly controlled by the East Asian Monsoon transporting effective moisture during summer time up to its northern limit at about 40 N, parallel to the Yellow River course outside the Tibetan Plateau, and

22.2 GEOLOGICAL SETTING The Gaxun Nur Basin – part of the Alashan Plateau – forms an intramontane accumulation area between the Qilian mountains in the south and Gobi Altai mountains in the north (Fig. 22.1). Both mountain ranges are influenced by left-lateral plate motions accompanied by intense vertical displacements due to the ongoing collision of the Indian plate against the Eurasian continent. The sinistral pattern continues northwards 96°E

98°E

325

100°E

102°E

104°E

Gobi-

42°N

Juyanze

He ih e

agh

Altyn T

He

rri

Sh

do

Ha

r

g

an

an

n

Tengger Shamo

Co

He

xi

lia

iy

iyu

36°N

an

98°E

40°N

Badain Jaran Shamo

Qi

96°E

1000 km

D100

Gaxun Nur Basin

40°N

0

Gaxun Nur Sogo Nur

Beishan

100°E

0 102°E

Sh

42°N

Altay

100

200 Kilometers

38°N 104°E

Fig. 22.1 Satellite image (RGB 7-4-2) of the Gaxun Nur Basin north of the Tibetan Plateau with location of the deep drilling D100 and major structural elements (red lines correspond to faults).

326

Bernd Wu¨nnemann et al.

intersecting with the westerly waves, generating local rainfall as well as heavy rainfall events by overlying air masses of different temperature and water vapour content during summer time (Domro¨s and Peng, 1988). The intensity of summer rainfall strongly depends on the pressure gradient between the heat low over Tibet and the anticyclone over the Pacific ocean, resulting in highly variable annual precipitation between < 40 mm and roughly 100 mm within the desert regions, while in the upper catchment of the Hei River precipitation rises to 500 mm/year. The annual amplitude of monthly mean temperatures ranges between 28 C in summer and 12 C in wintertime. During wintertime, cold dry air masses from the anticyclone over Siberia are responsible for extremely low temperatures and strong winds from northern directions without appreciable precipitation. Owing to the southward shift of the westerlies in wintertime, rain or snowfall from that source is negligible. According to the distribution of annual precipitation, river discharges towards the Gaxun Nur Basin peak during spring and late summer. Owing to strong human impact, e.g. water storage in the upper reaches of the Hei River for irrigation purposes, the terminal lakes Gaxun Nur and Sogo Nur have desiccated since the 1960s, while the hydrologically separated Juyanze Lake dried up much earlier during the Han Dynasty (third century AD). Besides the well-developed drainage pattern of south–north-directed river channels, a dense network of presently dry channel systems originates north of the basin and indicates strong fluvial input from the Gobi Altai ranges mainly developed during early and mid-Holocene wet periods (Wu¨nnemann, 1999; Hartmann, 2003). They required much more local rainfall than the present amount. Previous investigations by Wu¨nnemann and Hartmann (2002) indicate an extended paludine environment of > 10 000 km2 in size during MIS 3 between 35 and 25 kyr,

evidenced by the distribution of dated near-surface lake sediments and palaeoshorelines of up to 25 m above the present dry lake floors at 900 m above sea level. The up to 300-m thick basin fills of Quaternary age are documented by a total of 80 drillings (Gansu Provincial Geological Bureau, 1980). The dominant facies is of lacustrine origin with grain-size diameters in the clay and silt fractions, locally interrupted by fluvial and aeolian sand. Coarse gravels were found only in the upper metres of some cores. During the terminal phase of the last glacial maximum (LGM) and most probably during early Holocene times, strong fluvial run-off promoted the accumulation of fluvial sand and coarse gravels from local catchment sources and covered large areas of the basin, protecting the finegrained lacustrine sediments against deflation. However, according to Pye and Zhou (1989), this region still serves as a prominent source area for loess transport. Recently published results by Hartmann (2003) and Becken et al. (2003) indicate a very young tectonically induced subsidence of the northern part of the basin by about 2 mm/yr during the last 30 kyr. In terms of hydrology, the three terminal lakes have been separated since then. This would explain the differences in sedimentation behaviour between the Gaxun Nur system and the eastern palaeolake Juyanze during Holocene time as suggested by Hartmann (2003). However, previously reported highest lake levels of the Gaxun Nur system during MIS 3 (Pachur et al., 1995; Wu¨nnemann and Hartmann, 2002) developed prior to these tectonic events. Hence, the sediment core D100, retrieved from a deep drilling in the centre of the Gaxun Nur Basin, remains suitable for climate reconstructions.

22.3 METHODS For the coring of D100 sediments, we used a Chinese rotational drilling system with 3 m long metal tubes 80–120 mm diameter in

Interglacial and Glacial Fingerprints from Lake Deposits

size. The drilling in the centre of the Gaxun Nur Basin at 42.1 N, 100.85 E, 940 m a.s.l. reached the pre-Quaternary sediments at 229 m depth. Once a core segment was retrieved, it was pressed out immediately, halved, described and finally documented by photographs. Sampling of 1-cm thick slices was done every 10 cm, depending on stratification. Where necessary, sampling was narrowed down to 1 cm, for example within laminated sequences. For analyses of micromorphological features, thin sections of selected samples were prepared and analysed under a polarization microscope. Samples for magnetic properties were taken as cubes of 8 cm3 size and measured in the laboratory for Palaeo- and Environmental Magnetism (PUM), Bayreuth University. All samples were subjected to standard palaeo- and rock magnetic laboratory protocols. However, here we focus on the inclination record of the characteristic remanent magnetization derived from the results of stepwise demagnetization in alternating magnetic fields. A record of the relative palaeointensity (RPI) of the Earth’s magnetic could not be determined because of the pronounced rock magnetic heterogeneity of the sediments, which did not allow for a proper normalization of the intensity of the characteristic remanent magnetization. Thus, the stratigraphic interpretation is based on the correlation of our inclination data with the stacked record of RPI of the Earth’s magnetic of Thouveny et al. (2004). Dried and powdered samples for geochemical analyses were performed as follows: – Total organic carbon (TOC) and CO3 by loss on ignition (550 and 900 C) and by LECO carbon analyser (infrared spectrometry) following standard procedures. – Multielement analyses of Ca, Mg, Fe, Mn, Na, K, S and PO4 by a Perkin Elmer optical emission spectrometer with inductive coupled plasma (ICP-OES) on pretreated samples by hydrochloric acid (HCl) digestion.

327

The mineral compositions of selected samples were detected by XRD (Philips PW1710, Cu-cathode, 36 kV, 24 mA). Grainsize analyses were performed by a laser diffraction particle sizer (Beckman Coulter LS 200) after removal of carbonate and organic compounds by HCl and H2O2 treatment respectively. To avoid coagulation, sodiumpyro-phosphate (Na4P2O7 * 10H2O) was added, and the sample was then shaken for several hours. Before measurement, every sample was divided into eight subsamples and, where necessary, dispersed in an ultrasonic bath. At least three subsamples were measured in each case. Thus, the final grain-size distribution of each sample is based on the average of three subsamples. Radiometric dating (AMS) on bulk samples was done at Leibniz Laboratory, Kiel, Germany, and Beta Analytics, Miami, USA. In several cases, twin dating of the organic carbon from the remaining lye fraction and humic acid was performed to estimate contamination within one sample. Additionally, thermoluminescence dating (IRSL, OSL) of sandy lake sediments was performed at the Geophysical Laboratories of the Chinese Academy of Science, Beijing, China, and at the Saxonian Academy of Sciences, Freiberg, Germany. Both laboratories used feldspar grains for dating. 22.4 AGE MODEL For reconstructing changes in water balance of the Gaxun Nur Basin and variations in dust mobilization through time, several independent age determinations were performed. Unfortunately, previously reported radiocarbon and TL/IRSL ages from the D100 core between 11 and 90 m depth (Wu¨nnemann, 1999; Table 22.1) do not correspond to OSL dates from a parallel core I70, but show age differences of several 10 000 years. According to geomorphological investigations and radiometric

Mean depth (m)

Laboratory number

Dated material

Core D100 Gaxun D100-1 D100-2 D100-1101 D100-1601 D100-IRSL1 D100-1901 D100-IRSL2 D100-IRSL3 D100-3402 D100-3904 D100-IRSL4 D100-10 D100-10 D100-IRSL5 D100-11 D100-11 D100-IRSL6 D100-12 D100-12 D100-12 D100-13 D100-13 D100-14 D100-14 D100-15 D100-8911 I70-16 I70-1470 I70-17 I70-2900 I70-18 I70-3850 I70-5050 HC8 cliff 96E27 HC21

Nur (42.1 N, 9.35 11.13 11.22 16.29 16.46 19.71 19.73 22.27 34.50 40.06 44.03 44.22 44.22 60.73 65.59 65.59 71.90 72.25 72.25 72.25 76.99 76.99 79.71 79.71 90.12 90.12 11.77 14.70 14.80 29.0 29.67 38.50 50.50 5.0 4.0 4.0

100.85 E, 940 m a.s.l.) Beij1 Lake sand Hv20383 Bulk carbonate beta 188774 Bulk organic Carbon beta 189674 Bulk organic Carbon Beij2 Lake sand beta 189675 Bulk organic Carbon Beij3 Lake sand Beij4 Lake sand beta 190783 Bulk organic Carbon beta 190784 Bulk organic Carbon Beij5 Lake sand KIA24208 Bulk organic Carbon KIA24208 Humic acid Beij6 Lake sand KIA24209 Bulk organic Carbon KIA24209 Humic acid Beij7 Lake sand KIA24848 Bulk organic Carbon KIA24848 Humic acid KIA24848 Bulk organic Carbon KIA25440 Bulk organic Carbon KIA25440 Humic acid KIA25441 Humic acid KIA25441 Bulk organic Carbon KIA25756 Bulk carbonate KIA25756 Bulk organic Carbon Beta 188775 Bulk organic Carbon Freiberg Lake sand Beta 189676 Bulk organic Carbon Freiberg Lake sand Beta 190786 Bulk organic Carbon Freiberg Lake sand Freiberg Lake sand Beta 190785 Bulk organic Carbon KIA 23627 Bulk organic Carbon KIA 24143 Ostracod shell

mg C

AMS age kyr

TL/IRSL/OSL age kyr 22

21.265 14.45 14.29 29.20 17.00 41.10 54.50 16.86 16.72 90 3.9 0.7

34.82 30.00

0.7 1

26.41 34.57

1.3 2.4 0.4 0.5 0.2 0.3 0.9 0.5 1.6

35.85 27.57 23.75 20.56 25.73 27.20 25.64 > 41:52 30.71 8.89

196

202

146 11.18 172 6.97 230 281 0.4 0.9

Locations of the last three dates, see Wu¨nnemann and Hartmann (2002) and Hartmann (2003).

15.060 17.390 22.790

Deviation þ

Deviation

1.8 0.5 0.1 0.1 3.7 0.1 2.9 5.0 0.1 0.1 6.2 0.4 0.8 15.0 0.5 1.0 17.0 0.9 0.3 0.7 0.4 1.7 1.3 0.4

1.8 0.5 0.1 0.1 3.7 0.1 2.9 5.0 0.1 0.1 6.2 0.4 0.7 15.0 0.5 0.9 17.0 0.8 0.3 0.6 0.3 1.4 1.1 0.4

0.4 0.08 10 0.05 12 0.06 16 23 0.08 0.25 0.15

0.4 0.08 10 0.05 12 0.06 16 23 0.08 0.25 0.15

13 C (‰)

3:9 23:6 23 22:9

23:2 22 25:03 25:02 22:86 23:86 22:78 21:55 26:83 25:14 30:77 27:08 20:09 4:03 19:08 23:9 21:1 23:0

22:8 22:49 1:66

Bernd Wu¨nnemann et al.

Sample

328

Table 1 Absolute age determinations from the cores D100 and I70 (parallel core) and adjacent sites, Gaxun Nur Basin, NW China

Interglacial and Glacial Fingerprints from Lake Deposits 60

0

70 80 Laschamp 90

50

100 ?

110 120 Age (kyr BP)

Depth (m)

datings of palaeoshorelines and nearsurface lake sediments of the Gaxun Nur system which revealed ages from the Holocene to Late Pleistocene MIS 3 (Wu¨nnemann and Hartmann, 2002; Hartmann, 2003), all IRSL and OSL dates have to be considered too old. A series of AMS dates from the core D100 between 11 and 90 m depth revealed much younger ages (Table 22.1), roughly correspondent with age determinations in the closer catchment of the lake. The AMS dates, however, do not increase with depth but scatter in a wide range between 20 and 40 kyr up to 90 m depth. Age differences between the remaining lye fraction and humic acid support the assumption of repeated reworking processes within phases of flooding during MIS 2 and 3, initiated by tectonically induced subsidence. As a result, extremely high redeposition of clastic material took place. Washing and sieving of bulk sediments for ostracod analyses revealed very few specimens of broken valves, thus confirming redepositional processes. This assumption is supported by the occurrence of wood remains > 1000 mm in size and charcoal particles which have been transported by single flooding events from various locations. This might be the reason why the AMS dates scatter irregularly through depth. Owing to uncertain absolute age determinations, we used palaeomagnetic datasets for further chronological approaches. Measurements from 65 to 156 m and from 187 to 198 m core depth revealed intervals with reverse inclinations at 67–93 m, 122–142 m and 190–196 m core depth respectively (Fig. 22.2). According to our AMS dates and to the palaeointensity record of Thouveny et al. (2004), the intervals with reverse inclinations of core D100 most likely correspond to relative lows of the Earth’s magnetic dipole moment at about 25 to 45, 95 to 125 and 185 to 195 kyr containing the Laschamp, post Blake, Blake and Icelandic Basin geomagnetic events, thus giving the tie points for our chronology so far. As the

329

130 140

Post-Blake

100

Blake

150 150 160

No data

170 180

Icelandic Basin 200

190 200 –90 –60 –30 0

30 60 90

Inclination (°)

0

0.5

1

1.5

2

2.5

R.P.I. (normal to mean)

Fig. 22.2 Palaeomagnetic results from core D100 show intervals of reverse inclinations of the characteristic remanent magnetization at 67–93 m, 122–142 m and at 190.5–196 m depth. Comparisons with the stacked record of Thouveny et al. (2004) indicate that the intervals with reverse inclinations in D100 core match the Laschamp, post-Blake, Blake and Icelandic Basin subchrones, hence giving major tie points for our chronology.

lithology of the core from 71 m depth downward consists of more or less uniform rhythmic silt-clay layers, we calculated a linear sediment accumulation rate (SAR) of 7.1 cm/century (63–128 m depth, Fig. 22.3) and 6.8 cm/century (144–228 m depth) except for parts between 128 and 144 m depth, where intercalated sandy layers of aeolian origin required a higher SAR of 8.5 cm/century (Fig. 22.4). With respect to the Blake and post-Blake reverse inclinations at 142–123 m core depth, we calculated for the reversal peak at 140 m depth an age of 115 kyr, thus matching the lowest RPI in the record of Thouveny et al. (2004) and the reverse inclination in core MD95-2042 from the Portuguese margin as well (Thouveny et al., 2004, Fig. 22.4b). Hence, the sediments

Bernd Wu¨nnemann et al.

330 0

20

40

60

SAR: 71 cm/kyr

100

120

Laschamp

Depth (m)

80

140

SAR: 85 cm/kyr Eemian

160

SAR: 68 cm/kyr

200

Icelandic Basin

Blake/post-Blake

180

220

0

20

40

60

80

100

120

140

160

180

200

220

240

260

280

Age (kyr BP) Dating methods AMS dates on humic acid, organic and carbonate fraction TL/IRSL dates (Beijing, China) OSL dates from a parallel core I70 (Freiberg, Germany)

Fig. 22.3 The age–depth relation of the core D100 based on palaeomagnetic data and AMS dates. Our model indicates nearly constant sediment accumulation rates of lacustrine sediments from 71 m downwards, except for the sequences between 128 and 144 m depth. Thermoluminescence dates (IRSL and OSL) are not used for this chronology as they are considered to be too old and do not confirm geomorphological investigations.

of the entire core D100 probably comprise the sedimentary history of the past 250 kyr, a much shorter time than previously expected (Wu¨nnemann and Hartmann, 2002). 22.5 RESULTS 22.5.1 Sediments of core D100, Gaxun Nur The 230-m thick sediments from the centre of the Gaxun Nur Basin were retrieved by 110 cores of maximum 3 m length. Recovery reached about 78% of the whole sediment

column. Missing parts are usually related to sandy sequences, difficult to retrieve by a rotational drilling technique without liners. However, strata with pelitic sediments had a coverage of nearly 100%. In general, the sediments of core D100 can be divided into three lithological units (Fig. 22.4A): gravels in the top 7 m, frequent alternations of sand and silt between 7 and 71.4 m depth and laminated clay-silt layers from 71.4 m downwards. The bottom sediments from 228.8 m to 230 m depth consist of pre-Quaternary red conglomerates in a clayey matrix.

Interglacial and Glacial Fingerprints from Lake Deposits

331

Lithology of core D100, Gaxun Nur, China A 0 10

B 1

Gravel Fluvial sand

20 30

Fluvial/ aeolian sand

2

40 50 60 70 80

3

Aeolian sand

Laminated lacustrine deposits

4

90

Depth [m]

100 110 120 130 140

5

6

Lacustrine/ aeolian deposits

150 Fine laminations

C

160 Core D100 175,00–175,032 m

170

Aragonite Aragonite

Dark layer

180 190

210

Lacustrine deposits fine-laminated

5 mm 4 3 2 1 0

220 230

Clay/silt, Carbonate Calcite

Sublamina 86 layers in 2 cm.

30°

30 μm

org. C

Light layer

200

Quartz

FeOOH

Cracks

D100, 175,03 m

28°

26°

°Φ

24°

Aragonite

Light sublayer Aragonite

Clay

Dark sublayer clay and organic matter

25.0 kV

90030

100 μm

Playa (Neogene)

Fig. 22.4 Lithology of core 100, Gaxun Nur Basin. A ¼ overview of the entire sediment core; B ¼ photographs and thin sections of sediment sequences: 1: coarse gravel 0–7 m depth; 2: laminated sand layers, 17.66–84 m depth; 3: laminated silt-clay layers with dark layers < 1 cm, 33.37–34.77 m depth; 4: laminated silt-clay layers with dark layers 1–2 cm thick, 72.12–55 m depth; 5,6: thin sections of laminated layers at 72 and 33.9 m depth. The dark layers contain more clay < 1 m in size, while only the light layers contain carbonatic fossil remains (here the brackish water tolerating ostracod Cyprideis torosa); C ¼ Thin section, scanning electron microscope photos and XRD analyses of laminated sediments from 175 m depth. Sublaminations show variations between clastic input and authigenic aragonite precipitation.

The coarse gravels of mainly metamorphic rocks (Figs. 22.4A, B1) up to 7 m depth indicate high-energy fluvial transport from local areas. They are widespread within the basin and cover lacustrine sediments. Stratified

and E-W- or N-S-dipping layers are indicative of surface erosion along the northern (Gobi Altai Tienshan range) and western (Beishan Mts.) catchments of the basin. Sources from the Tibetan Plateau can be

332

Bernd Wu¨nnemann et al.

excluded as the distance between the gravel plain and the Qilian Mts. is more than 350 km, too far away for coarse material to be transported into the interior of the basin. Poorly sorted, 2.5-m thick coarse-tomedium sands below the gravels complete the fluvial deposits and may indicate the initial phase of locally enhanced surface erosion. Limnic deposits beneath the fluvial components were AMS dated to 14 450 100 yr BP, 15 060 80; 17 390 250 and 22 790 150 yr (Table 22.1), indicating that the gravels and sands were formed after the LGM and probably during the Termination I and early Holocene. Between 9.5 and 71.4 m depth, frequent alternations of well-sorted fine sand and silt layers occur (Fig. 22.4B2-6), indicating a playa-like depositional environment. For the first time, discontinuously laminated sequences of clay and silt occur between 16.24 and 20.58 m depth. They consist of dark (< 1 cm thick) and light layers (1–3 cm thick), with a mean grain size of 11 mm (Fig. 22.4B3-4). They are typical of a lacustrine facies which fills the main part of the Gaxun Nur Basin. Mineral composition and extremely high percentages of pelitic components indicate a long-distance transport as suspended load from the Tibetan Plateau via the Hei River. The dark layers differ from the light ones by significantly higher percentages of fine clay (< 1 micron, Fig. 22.4 B5-6) and lower carbonate content. We assume that these laminations do not represent annual deposition behaviour but more likely periods of flooding and depositional events, keeping a shallow lake inexistence for centuries or even millennia. Similar sequences occur at 51.97–52.22 m, 65.04–67.30 m and 68.24–70.15 m depth. They are interrupted either by aeolian sand (e.g. 21.55–25 m, 61.69–63.25 m and 70.15– 71.46 m depth) or by semilacustrine sandy silt deposits. In some cases, alternations of sand and clayey silt (e.g. 50.18–52.60 m and 59.58–61.25 m depth) indicate that lacustrine sedimentation was probably overprinted by enhanced aeolian flux, thus

leading to comparable event layers (Fig. 22.4B 2) as stated above. From the general heterogeneity of the sediments down to 71.46 m depth, we assume that these different depositional processes require unstable hydrological conditions. Even complete desiccation of the playa system accompanied by stronger influence of aeolian processes might have occurred several times. From 71.46 m downwards to 228.8 m depth, the sediments consist of silt and clay. Most parts are fairly well laminated (Fig. 22.4). They indicate that a perennial lake existed during the entire period of sedimentation, mainly supplied by freshwater from the Hei River. Coarser sediments (up to 250 mm) of aeolian origin only occur in single layers between 117 and 151 m depth, indicating additional input of dune sand from the vicinity of the lake. Only selected sequences of the event layers have been counted so far. For example, 136 dark layers with a mean thickness of 1.8 cm appear between 73.7 and 78.6 m depth, indicating 27 flooding events per metre sediment. A similar value (26 events/m) accounts for 151.5–156.5 m depth, although the mean thickness of the dark layers is reduced to 1 cm, probably due to shorter durations of flooding events. On the other hand, variations in the thickness of light layers do not automatically indicate longer times of normal freshwater flux, rather, their thickness depends also on the amount of aeolian dust input (loess) from near-shoreline locations in phases of lake-level decline. In several sequences below 165 m depth, sublaminations within light layers occur. Sub-millimetre thick dark layers of clay and organic matter alternate with authigenic aragonite, revealed by SEM and XRD analyses (Fig. 22.4C). Counting of sublamina from a thin section at 175 m depth revealed 43 single layers per centimetre pointing to annual varves of authigenically precipitated aragonite during summertime and deposition of clay during the cold seasons. Similar couplets were documented in

Interglacial and Glacial Fingerprints from Lake Deposits

lake sediments from Lake Van, Turkey (Kempe and Degens, 1978). Carbonatic fossil remains, classifying the sediments as lacustrine, have been found in several parts of the core, although they are not always present. Up to now, higher abundances of the ostracod shells Limnocythere inopinata and Cyprideis torosa as well as the freshwater gastropods assigned to the superfamilies Rissooidea and Cerithioidea (determined by S. Mischke and F. Riedel, Berlin) have been found in sequences at 68–74 m, 100–103 m, 119–125 and below 165 m depth. 22.5.2 Geochemical properties For the interpretation of changes in the water balance of the Gaxun Nur Basin, we used the elements Ca, Mg and sulphur as the most appropriate ones to identify chemical precipitation of carbonates and sulphate in a lacustrine environment. They are indicative of increasing evaporation pressure over a water body and its residence time as well, both inducing rising salinity by the enrichment of ion concentration (Eugster and Hardie, 1978). SEM and XRD analyses of samples with higher concentrations of sulphur yielded evidence of gypsum formation when sulphur increased. TOC is used to identify phases of increased internal productivity of the lake. The generally low values of TOC (0.2–1.6% dry weight, maximum 2%) in core D100 are quite normal for arid regions, where bioproductivity is limited owing to warm/cold and dry climate conditions. However, algae (e.g. Charophytes, Pediastrum) have been detected in various parts of the sediments and therefore indicate lake-internal bioprocesses which depend on water quality, temperature and depth. On the other hand, detrital input of organic matter (e.g. charcoal, wood remains and pollen grains) from the catchment of the lake was recognized too, leading to a certain enrichment of TOC in phases of intense river discharge. The total amount of residual silicates after loss on ignition is used to calculate the basic

333

allochthonous sediment load transported either by the river from remote locations or by aeolian processes from the vicinity of the lake. The geochemical properties in the sediments which can be used as climate proxies are summarized in Fig. 22.4, plotted against the depth of core D100 and divided into five units: In unit 1 (160–144 m depth), relatively high values of up to 1.6% TOC are mainly derived from aquatic phytomass, developed during a period of positive water balance. Higher calcite (120–160 mg/g Ca) and gypsum ( 4 mg=g sulphur) precipitation points to a slightly saline water body, interrupted by a distinct freshwater pulse (154– 152 m depth). However, high calcite and moderate gypsum precipitation points to warm lake water as well as high phytomass production under mostly warm climate conditions. A reverse trend with more unfavourable conditions for lake-internal bioproductivity is detectable in the second half of unit 1. Unit 2 (145–110 m depth) is characterized by generally low TOC (< 0:5%), Ca (< 40 mg=g), sulphur (< 0:9 mg=g) concentrations and increased clastics, indicating a colder water body with reduced internal bioproductivity and external input of suspended load. Mg concentrations, indicative of Mg–calcite or even dolomite precipitation, remain constant, similar to unit 1. Two spikes of lowest Ca and sulphur at 143–141 m and at 135–132 m depth point to stronger changes in lake hydrology, resulting in a considerable lowering of the lake level, but without complete desiccation as suspended load appears to have been continuously accumulated. Close to the end of unit 2 (120–110 m depth) a return to more favourable conditions in terms of lake hydrology is detectable, with enhanced calcite, Mg–calcite and gypsum precipitation and reverse TOC contents. As Fig. 22.6 shows, we consider a general succession from freshwater inflow leading to rising lake levels, freshwater conditions and

Bernd Wu¨nnemann et al.

334

increasing TOC followed by a steady state with increased residence time of the water and the enrichment of ion concentration towards saline conditions. The main chemical precipitate at that period is gypsum, pointing to a stepwise shrinkage of the water body under dry-warm conditions. Low amounts of detrital silicates at 110–113 m depth (Fig. 22.5) support the arguments of a reduced water inflow. Rising salinity may also have caused successive reduction of internal biological productivity as the low TOC values suggest. Unit 3 (110–93 m depth) comprises a period of decreasing geochemical precipitates and TOC with minimum values of < 20 mg=g Ca and < 0:4% TOC at 106 and 102 m depth when clastic input is high.

Grain-size analyses (see below) classify the clastics as predominantly aeolian silt and sand pointing to a generally negative water balance, similar to certain periods in unit 2. Surface inflow starts to increase again towards the end of unit 3. Unit 4 (93–68 m depth) is characterized by frequent changes in TOC content between 0.4 and 2% as well as chemical precipitates (40–190 mg/g calcium, 6–40 mg/g magnesium and 6–20 mg/g sulphur) following the same succession as described above. These changes are due to short-term fluctuations in the water balance and evaporation intensity of generally unstable hydrological and climate conditions. For unit 5 (68–11 m depth), only TOC and Ca data are available so far. Both datasets

Ca (mg/g) 0

40

80

120

S (mg/g) 160

200

0.1

1

Unit 10

0 10 20 30

5 40 50

Depth (m)

60 70 80

4

90 100

3

110 120

2

130 140 150

1

160 0

0.4

0.8

0.2

TOC (%wt)

1.6

2

10

Mg (mg/g)

40

60

80

100

Silicate (%wt)

Fig. 22.5 Geochemical properties of the D100 core, Gaxun Nur Basin, NW China, between 7 and 160 m depth. The shaded areas in general mark periods of enhanced chemical precipitation and thus higher evaporation. Coloured graphs are 7-point low-pass filtered.

Interglacial and Glacial Fingerprints from Lake Deposits

335

Depth (m)

110

115

120 0

0.4

0.8

1.2

TOC (% wt)

1.6

2 0

40

80

120 160

Ca (mg/g)

200

1

10

S (mg/g)

Fig. 22.6 TOC, Ca and sulphur (gypsum) distribution in core D100, 108–120 m depth. Changes in TOC content are inverse to carbonate or sulphate concentrations indicating organic development in phases of positive discharge under freshwater conditions while long residence time and shrinkage of the lake induce enhancement of ion concentration due to evaporation.

indicate similar fluctuations as in the previous unit but with generally lower values and amplitude. This is due to the fact that sediment in this unit changed more frequently between a lacustrine and semilacustrine facies to fluvial types, both superimposed by aeolian sedimentation under changing hydrological conditions. Phases of enhanced chemical precipitation at 58–50 m, 32–26 m and 20–11 m depth accord with stable lacustrine environments, while the lowest values fall in periods of enhanced dune sand accumulation at the drilling site, pointing to extremely low lake levels. 22.5.3 Grain-size variations The classification in a triangular diagram (Fig. 22.7A) shows that the overwhelming majority of fine-grained sediments consists of silts. They are the main clastic components, either free of any sand fraction or with very small amounts of fine sand (63200 mm). The clay fraction sometimes amounts to maximum 46% of the total volume. The median of samples is 11 mm, similar to the quartz median (Qmd) of the

Luochuan loess section, reported by Xiao et al. (1995). The Gaxun Nur sediments cannot be characterized sufficiently by classical parameters such as mean, standard deviation, skewness and kurtosis, as many of the samples have bi- or even multimodal grainsize distributions. For example, the sample of Fig. 22.7B can be regarded as a typical unimodal clay-silt without any sandy component. The mode lies in the fine silt fraction, whereas the samples of Fig. 22.7C and 22.7D are typical multimodal silts with modes at 5 and 10 mm respectively, also containing variable proportions of coarse silt (modes at 25 and 53 mm) and fine sand (mode 160 mm). Especially in samples from dark layers (Fig. 22.7D), a further mode at 1 mm appears to be characteristic of higher clay components in those layers, probably accumulated after phases of flooding events. The enrichment of clay in dark layers may be due to coagulated clay particles which accumulated faster than noncoagulated ones. The variation of grain-size distributions versus depth of D100 core is documented in Fig. 22.8. Assuming that different

Bernd Wu¨nnemann et al.

336 0

(a)

1

(b)

4

77% silt 23% clay Median: 4.6 μm Mode: 5.9 μm

0.2

3 0.8

00

0.4

(63

) μm 0

Sa

nd

1

<2

...2

y( Cla

0.6

0μ

0.4

m)

2 0.6

0.6

2

B

6

20

60

200

600

2000

Particle diameter (μm)

0.8

0.2

C D

1

0.8

0.6

0

0.2

0.4

1

0

Silt (2...63 μm)

(d)

77% silt 19% clay 4% sand

3

Median: 7.8 μm Mode: 55 μm 2

77% silt 18% clay 11% sand Median: 8.4 μm Mode: 10.3 μm

2

Volume (%)

(c)

1

1

0 0.6

2

6

20

60

200

600

2000

Particle diameter (μm)

0 0.6

2

6

20

60

200

600

2000

Particle diameter (μm)

Fig. 22.7 Grain-size variations in sediments from the D100 core, Gaxun Nur Basin, NW China. a: Triangular diagram of all analysed samples. b, c: Grain-size distributions of three clayey silts: b ¼ unimodal silt, 77.34 m depth (light layer). c ¼ multimodal silt, 115.62 m depth (dark layer). d: multimodal silt, 148.525 m depth (dark layer). Coloured areas mark deviations between single measurements.

transport mechanisms are responsible for variations in grain size of clastic components that were deposited in a perennial but fluctuating lake, we can distinguish between two main depositional processes which contributed to the accumulation of the Gaxun Nur Basin: The fraction < 20 mm was transported as suspended load (allochthonous clastics) into the basin mainly via the Hei River. As it was deposited into a stillwater body far away from the river mouth and from the lake shoreline, we assume that coarser clastics could not reach the drilling location during phases of high lake levels. On the other hand, we cannot exclude a certain amount of fine fractions deposited as aeolian dust

within the lake. However, we assume that the majority of this grain-size fraction has been transported by fluvial processes. Its contribution of up to 80% of the whole sediment can be regarded as the basic load of clastic input. Even in sequences of coarse sand, up to 7% clay-silt is present. The grain-size fractions 2063 mm and 63200 mm indicate aeolian transport of loess or dune sand from the vicinity of the lake to the drilling site. In contrast to the Luochuan loess section of the loess zone in China – where 4416 mm fractions were mainly deposited during cold periods whereas the smaller fractions (165 mm) indicate deposition during interglacial climate conditions (Vandenberghe et al., 1997) – our data cannot

Interglacial and Glacial Fingerprints from Lake Deposits

337

<20 μm (%vol)

63. . .200 μm (%vol)

Grain-size index (20–50 μm/<20 μm)

20 40 60 80 100

0

0 0.4 0.8 1.2 1.6 2

20

40

60

Unit

60

60

70

70

80

80

90

90

100

100

110

110

120

120

130

130

140

140

150

150

4

Depth (m)

Depth (m)

5

3

2

1 160

160 1

10

100 1000

Mode (μm)

0

20

40

60

20...63 μm (vol%)

0

4

8

12

200...2000 μm (vol%)

Fig. 22.8 Grain-size distribution (7-point low-pass filtered) in sediments of the D100 core, Gaxun Nur Basin, NW China, versus depth. Shaded areas mark phases of positive water balance, relatively high lake levels and reduced aeolian transport.

be used in the same way, as our site is located several hundred kilometres north-west of the Loess Plateau, a prominent source area of loess transport, where coarser fractions up to medium sand are the major aeolian components. However, the grain-size index used by Rousseau et al. (2002) for the Nussloch section in Germany may serve as an indicator of aeolian variability in the Gobi Desert too (see Fig. 22.8). Although it cannot be decided whether the coarser components are always transported by wind or sometimes probably reworked fluvial material, both processes at least require a shrinking lake, promoting the exposure of near-shore lake sediments which could be easily removed by the wind and transported to the centre of the lake and elsewhere.

Conversely, in phases of enhanced river discharge, expansion of the lake and higher lake levels, the distance between the lake shoreline and the drilling site increased considerably, reducing coarse clastic input (> 70% of particles < 20 mm, shaded parts in Fig. 22.8 ). A relatively long period of stable/higher lake levels is also represented by sublaminated pelites (unit 1), only interrupted by short phases of aeolian input. At the transition from unit 1 to unit 2 and during the following unit 2 the input of aeolian components (loess and sandy loess) increases rapidly. In addition, medium and coarse sand (2002000 mm grain size) of fluvial origin at 135 m depth indicates that the river mouth extended lake-ward

Bernd Wu¨nnemann et al.

338

Ice volume

Solar forcing

Mid-June Insolation

Temperature

RSL

Lake temperature Evaporation salinity

NorthGRIP Guliya,Tibet Greenland 18 δ O (‰) –20 –16 –12

Sea level (m) –160–120–80 –40 0

Water balance lake-level suspended load

Winter monsoon strength aeolian flux (loess and dune sand)

Gaxun Nur D100 Clay-silt <20 μm (vol%) 60

70

80

Sand 63–2000 μm (vol%)

90 100

0

30

Luochuan Loess, China

60 0

0

A

B

C

D

E

F

G

H

I (n = 137)

10 20

(n =1758)

(n =1674)

(n =1674)

1

10

2

20

(n =1674)

42°N 2

30

30

40

40

50

50

60

60 70

70

3 Age (kyr BP)

65°N

Age (kyr BP)

Stages

4

C20 20

80 C22

90

80 90

22

5.1 5.2

100

100 110

5.3

23 24 25

C25

110

C26

120

LEAP ?

5.5

EEMIAN

130

130 400

450 500 2 (W/m )

550

–40 –44 –40 –36 –32 δ18O (‰)

5.4

120

0

40 80 120 160 200 Ca (mg/g wt)

0 0.4 0.8 1.2 1.6 2 Grain-size index (20–50 μm/<20 μm)

4

8 12 Qmd (μm)

16

6

Fig. 22.9 Comparison of climate records with proxies from the D100 core, Gaxun Nur Basin, Gobi Desert, NW China. A: Mid-June insolation for 65 N (solid line) and 42 N (dashed line) for the past 130 kyr. B: Relative sealevel (RSL) changes, adapted from Waelbroeck et al. (2002). C: NorthGRIP oxygen ice core record from central Greenland with selected numbers of cold stadials C26–C20 and interstadials (DO events), adapted from NorthGRIP Members (2004). D: Guliya ice core record from the central Tibetan Plateau, after Thompson et al. (1997). E: Calcium record from the D100 sediment core. E–H: Grain-size records (E ¼ suspended load, F ¼ loess component, G ¼ grain-size index and H ¼ aeolian sand) from the D100 sediment core. I: Quartz median record from the Luochuan loess section, adapted from Xiao et al. (1995). Data from the Gaxun Nur core D100 are 7-point low-pass filtered. Shaded areas mark warm-wet climate conditions, adapted from our data.

due to a strong shrinkage of the lake which enabled coarse material to be deposited in the centre close to the drilling site. At the end of unit 2 (117–110 m depth), grain-size variations are similar to those in unit 1, indicating a more positive water balance and a more stable water body, although short-term fluctuations are detectable by high-frequency alternations between clay and silt. In unit 3 (110–93 m depth), the input of aeolian sand and coarser clastics of probably fluvial origin between 109 and 102 m is high again, implying an extremely negative

water balance at that time. We assume that aeolian and probably fluvial inputs into a shrinking water body are the major processes at that time. Phases of higher suspended load indicating more stable lake conditions and thus higher lake levels throughout unit 3 and the following unit 4 appear at 96–94 m, 91–85 m, 83–77 m and 74.5–70.5 m depth. However, they are always interrupted by short-term phases of enhanced aeolian input into an existing lake. The change from predominantly fine-grained lacustrine deposits to sandy silt and finally dune sand from

Interglacial and Glacial Fingerprints from Lake Deposits

71.4 m depth upwards supports the arguments of a generally deteriorating water balance in the Gaxun Nur Basin. 22.6 DISCUSSION Lithology, geochemical properties and grain-size distributions of the D100 core from the Gaxun Nur Basin provide evidence for repeated changes in the water balance of an extended palaeolake within the Gobi Desert of north-western arid China. Our chronology indicates that the sediment records from the D100 core comprise the history of the past 250 kyr. However, we focus on the results of the upper 160 m of the core, i.e. the last 130 kyr (Fig. 22.9), thus covering the period from the last interglacial to the early part of the LGM ( 25 kyr). We compare our records with ice core data from Greenland (NorthGRIP Project Members, 2004) and Tibet (Thompson et al., 1997), with the relative sea-level record (Waelbroeck et al., 2002) and with the Luochuan loess record from the Chinese Loess Plateau as well (Xiao et al., 1995; Rousseau and Wu, 1997). Considering that the geochemical proxies (salinity and evaporation–precipitation ratio) and grain-size parameters (suspended load and aeolian transport) have a direct relation to changes in hygric conditions of the Gaxun Nur Basin and its catchment, we can set up following presumptions: Increased suspended load requires a substantial increase of river discharge towards the basin, triggered either by enhanced precipitation in the catchment and/or by melt water discharge, both resulting in the growth of lake size and water volume. As our drilling site in the centre of the basin is located far distant from the river mouth/shorelines, we assume that coarser sediment fractions were deposited by aeolian input as (sandy) loess. Fluvial input was only possible when the distance between the coring site and the river mouth/shorelines decreased considerably. The latter requires a strong shrinkage of

339

the lake in phases of reduced surface discharge. Furthermore, considering a substantial contribution of melt water towards the basin during the last glacial cycle, we also assume that melt water flux rose when snowmelt in the upper catchments was high due to increased temperatures and vice versa. This would mean that changing annual mean temperatures in the upper catchments (above and below the snowline), generally detectable by the oxygen-18 proxies from ice cores (Thompson et al., 1989, 1997; Thompson 2000), must have had an influence on the dynamics of the glaciers themselves, leading to either glacier advance in phases of colder and/or wetter conditions or glacier decay in periods of warming and/or decreasing precipitation. Hence, we hypothesize that glacier advance in Tibet requires a temporal shift of effective moisture supply from summer to autumn and wintertime to keep the glaciers growing. Such shift would also lead to a drop in the annual temperature, which is isotopically preserved in glaciers. Consequently, stable oxygen isotope signatures in glacier ice of Tibetan glaciers not only indicate temperature changes but also imply dynamic processes towards glacier growth or decay. 22.6.1 The last interglacial (MIS 5, 129–72 kyr BP) The warm stage of the last interglacial (MIS 5.5) appears in our records as a period of enhanced water discharge from southern regions (Tibetan Plateau) towards the basin which led to a rapid development of a large lake, coeval with a reduction of aeolian flux (Fig. 22.9G–I). According to our data, the transition from MIS 6 to MIS 5.5 is defined in the grain-size records by a simultaneous drop in aeolian activity and a rise in surface discharge at about 129 kyr, equal to 156 m core depth. Lowest sulphur and calcium values indicate a still reduced evaporation pressure and low water temperatures at that time. A rapid warming during the following

340

Bernd Wu¨nnemann et al.

1000 years is indicated by sudden increases in water temperature, residence time and evaporation, with rising salinities peaking at about 128 kyr. This phase of warmer and moister climate conditions comprises the time from the inception at 129 kyr to about 119 kyr, inferred from our age model (see Section 4). High calcite precipitation and surface discharge of nearly 10 000-year duration (Fig. 22.9E, F) indicate a stable aquatic environment, while a long residence time of the water body and strong evaporation may have caused increasing salinity under at least warm moist climate conditions. The intermittent distinct freshwater pulses in the middle part of this warm stage led to several refillings of the lake without a general change in climate conditions. Hence, it seems plausible to assume that these refillings towards freshwater conditions were due to enhanced rainfall in the catchment of the lake triggered by a strengthened SE summer monsoon. The lack of aeolian sand and a low grainsize index (Fig. 22.9G, H) in our record during that period support the assumption of reduced aeolian transport because most of the potential source area was covered by water and vegetation. This period of reduced dust flux and the development of a large lake within the Gaxun Nur Basin coincides with the formation of the S1 soil on the Loess Plateau (e.g. Kukla and An, 1989; An et al., 1991; An and Porter, 1997) and with lake formation in the Zoige Basin, north-eastern Tibetan Plateau (Chen et al., 1995). They all require warm-moist climate conditions in both high and low altitudes during substage 5.5, similar to the present climate or even warmer (Frenzel, 1994). Accordingly, the Guliya ice core record from central Tibet (Fig. 22.9D; Thompson et al., 1997) shows less negative oxygen-18 values, indicating higher temperatures over the Tibetan Plateau. Hence, we assume increased temperatures at high altitudes which may have caused high melt water flux during the summer season, resulting in negative glacier balances in the upper

catchments of the Gaxun Nur system. This is supported by the fact that the sea level of the South China Sea was equal to the present coastline (Yang, 1991) or even slightly higher (Jouzel et al., 2002; Waelbroeck et al., 2002; Shackleton et al., 2003). Both melt water and more frequent summer monsoon rainfall may have promoted the strong surface discharge towards the basins. Maher and Thompson (1995) calculated mean precipitation values on the basis of magnetic susceptibility data from loess sections and assume that rainfall north of the QinghaiTibetan Plateau during the last interglacial was two to three times higher than at present. Consequently, the summer monsoon front must have shifted some hundred kilometres northwards to back up sufficient rainfall in the presently dry desert, also assumed by An (2000). However, we are not able to quantitatively calculate the mean annual precipitation in the Gaxun Nur Basin during warm stage 5.5, but the existence of a large and biologically active water body supports our assumption of warm and moist climate conditions with potentially longer summer monsoon impact and shorter cold-dry winter seasons. According to our chronology, favourable climate conditions and a high lake level (but with a general decreasing trend) existed until 119 kyr. Small grain-size diameters (Fig. 22.9I) indicate a generally weak winter monsoon at that time. The most stable fresh-to-slightly brackish water conditions of about 7000-year duration occurred between 128 and 121 kyr, followed by a rapid desiccation and increase in dust flux (Fig. 22.9E–H) which culminated at about 119 kyr, equal to 144.7 m core depth. Owing to low data resolution, this environmental trend is not preserved in the Luochuan loess section. However, a comparison with the NorthGRIP ice core record (Fig. 22.9C) indicates that the change in lake hydrology towards a negative water balance corresponds to the rapid depletion of the oxygen-18 isotopes at the end of stage 5.5. We therefore assume a shift towards colder

Interglacial and Glacial Fingerprints from Lake Deposits

and drier climate conditions, also associated with a global sea-level drop of about 20 m at ca. 118 kyr BP (Fig. 22.9B) at least. As the severe drop in river discharge is accompanied by low carbonate precipitation and suspended load as well as high input of dune sand, we interpret this abrupt change as a signal for colder and drier climate conditions which may have prevailed for about 500–800 years. This short climate spell may correspond to the 468-year long late Eemian aridity pulse (LEAP), reported from central Europe by Sirocko et al. (2005) and terminates the warm-wet substage 5.5 in northwestern China. With respect to the timing and duration of this substage, our data imply that the warmwet climate phase between 129 and 119 kyr comprises the Eemian in north-western China. Compared with vegetation and marine records from western Europe, where the Eemian (MIS 5e) is considered to cover the period from 126 to 110 kyr (Shackleton et al., 2003; Sa´nchez Gon˜i et al., 1999, 2005), our time range for the Eemian is, however, strikingly different. In our opinion, this mismatch is primarily due to the fact that the desert regions in an extreme continental position north of the Tibetan Plateau responded immediately and very sensitively to changes in moisture availability, rather than to changes in local temperature. As the Gaxun Nur Basin is directly connected with glaciated high mountain ranges, both major changes in glacier budgets and alternations in moisture transporting air masses promptly affect the hydrological system in the Tibetan foreland, promoting either lake formation or aeolian activity. By contrast, vegetational changes (especially in trees) follow with a certain time lag, when plant-available water and the groundwater table have been stabilized and vice versa. However, we cannot decide whether the time lags of 3000 years for the onset and even 9000 years for the end of the Eemian can be explained by these relations alone. As the duration of the Eemian in northern Europe (126–115 kyr; Sa´nchez Gon˜i et al.,

341

2005) is much closer to our chronology, we assume that regional disparities of the Eemian not only depend on latitudinal climate differences but probably also on continentality. Palynological records from north-western China which could prove this assumption are still lacking. In terms of water availability, our data imply that during the succeeding substages of MIS 5 alternations from positive to negative balances do not differ strikingly from those during the warm-wet stage 5.5. However, the amplitudes between high and low river discharges increase considerably, likewise the amount of aeolian transport. Both indicate more frequent and intense changes between the extremes than before. Hence, we assume that the onset of glacier growth since the last glacial inception at about 118 kyr, most likely triggered by decreased insolation (Fig. 22.9A; Berger and Loutre, 2002; Calov et al., 2005), is the major reason for instable climate conditions in the catchment of the Gaxun Nur Basin since the termination of the Eemian. In phases of high river discharge and reduced aeolian transport (Fig. 22.9G-I), roughly correspondent with the D/O events 25–20 in the NorthGRIP ice core record (Fig. 22.9C), increased melt water flux and monsoon precipitation may have dominated to keep the water balance of the Gaxun Nur positive. According to our data, this happened during the first half of substage 5.4 (ca. 117–111 kyr) and during the substages 5.3 (106–101 kyr) and 5.1 (89–74 kyr). Small grain-size diameters in the Luochuan loess section (Fig. 22.9I) supplement our assumption of repeated millennial scale warmmoist climate pulses with high lake levels. The mollusc record from the Luochuan loess section even indicates that at about 88 kyr the climate on the Chinese Loess Plateau was warmer and wetter than today (Rousseau and Wu, 1997), but increasing chemical precipitates with enhanced gypsum formation afterwards point to a general trend towards aridification. According to Dansgaard et al. (1993), these phases of enhanced evaporation may correspond to

342

Bernd Wu¨nnemann et al.

DO events 23 (Brørup-Interstadial) and 21 (Odderade-Interstadial). Intermittent freshwater pulses from the south kept the lake in existence (see also Fig. 22.6). However, the climate might have been considerably different from the warm-wet Eemian substage which was characterized by locally enhanced summer monsoon rainfall buffering phases of reduced river discharge at that time. By contrast, lake phases between 118 and 74 kyr were mainly controlled by the run-off from the Tibetan catchments with increased rainfall and melt water flux there. Additional but limited water supply may have only resulted from local rainfall during the warm seasons, generated by a convective water vapour circle over a large water body, as previously assumed by Pachur et al. (1995). Summer monsoon rainfall in that region can be excluded, as the continuously increased albedo over the Tibetan Plateau has inhibited a considerable northward shift of the summer monsoon front. High-frequency alternations in the water balance of the Gaxun Nur Basin during the substages 5.3–5.1 also indicate that millennial scale periods of enhanced freshwater input are the major reasons for the deposition of laminated lacustrine sediments < 20 mm in size (Fig. 22.9F) into a perennial and probably deep lake. The laminated dark and light layers indicate periodical flooding events, which most likely occurred cyclically two to three times per century. The majority of such freshwater pulses are especially pronounced during the interstadial substages 5.3 and 5.1. Lowest discharge rates and enhanced aeolian flux within the Gaxun Nur Basin culminate during the substages 5.4 (second half, 108–106 kyr) and 5.2 (around 100 and 92–90 kyr), not exactly in phase with the Greenland and Tibetan ice core records. However, fluvial sand and high loess contents in the D100 core reflect lowest lake levels and vast areas of exposed lake sediments now easily deflated and transported southwards to the Chinese Loess Plateau,

owing to a strengthened winter monsoon (Xiao et al., 1995; Porter and An, 1995). In particular, coarser grains in the Luochuan loess section (Fig. 22.9I) indicate the coherence of exposed lacustrine sediments during phases of negative water balance and their transport to the loess regions of China by enhanced wind activity. 22.6.2 The cold stage 4 (74–60 kyr) The most severe cold phase during the last glacial cycle is well recorded in both Greenland and Tibetan ice cores (Fig. 22.9C, D). Cold and even dry conditions between 74 and 60 kyr are also detectable in the Gaxun Nur record by strong increases in aeolian flux, peaking at 72, 67 and 62 kyr, generally in agreement with the main occurrence of xerophilous mollusc taxa in the Luochuan loess section between 75 and 60 kyr, reflecting drier conditions than today (Rousseau and Wu, 1997). In contrast to the ice core and loess records, our data indicate an intermediate wet spell with enhanced river discharge between 70 and 68 kyr (Fig. 22.9, F–H), also identifiable in the Luochuan mollusc record (Rousseau and Wu, 1997, Fig. 22.3), which seems to have interrupted the cold-dry conditions north and east of the Tibetan Plateau for a 2000-year long period. One possible explanation for this mismatch with the ice core records might be the fact that parallel to the massive glacier advance in Tibet melt water flux towards the dry foreland increased dramatically and refilled the terminal lake. The coherence of glacier advance and high melt water flux in the catchment of the Gaxun Nur is only possible if we consider effective moisture supply at that time, not by the summer monsoon but more likely by the westerlies during late summer/early autumn. Snowfall in high altitudes then became most effective due to decreased summer radiation, while in lower altitudes (below the snowline) local precipitation and continuing ice melt contributed to increased surface run-off. This would also explain why enhanced wind

Interglacial and Glacial Fingerprints from Lake Deposits

velocities and dust transport from western directions resulted in loess accumulation on the Loess Plateau (Vandenberghe et al. 1997), while dust flux in the Gaxun Nur area remained low. From about 68 kyr onwards, the dramatic shrinkage of the Gaxun Nur lake promoted the exposure of fine-grained lacustrine deposits leading to deflation by wind erosion, transportation to south-eastern locations and deposition there as loess. This pathway of loess transport requires a stronger influence of the winter monsoon under cold-dry climate conditions which prevailed until about 60 kyr. Increased particle size in loess sections of the Loess Plateau (Fig. 22.9I, An et al., 1991; Porter and An 1995) confirms the intensive wind activity at that time. 22.6.3 The interstadial stage 3 (60–24 kyr) The climate conditions during MIS 3 are marked by strong alternations of warmwet and cold-dry pulses within the Gaxun Nur catchment, corresponding to fluctuations in the global sealevel (Fig. 22.9B). In terms of water budget, the temporal variations in the Gaxun Nur Basin reflect a close relationship to temperature-dependent isotopic changes of the NorthGRIP and Guliya ice cores, although not all cycles recorded there resulted in synchronous responses within the drylands of north-western China. However, phases of positive water balance with highest lake levels also in adjacent areas fall in periods of relatively warm and moist climate conditions (Pachur et al., 1995; Wu¨nnemann et al., 1998; Zhang et al., 2001, 2002) with enhanced Indian monsoon (Shi et al., 2001), when melt water discharge was effective owing to the melting of the Tibetan glaciers. Similar to the climate conditions during the substages 5.4–5.1, the water balance of the Gaxun Nur system during stage 3 was mainly controlled by surface discharge from the Tibetan Plateau. Evidence for enhanced local precipitation within the Gaxun Nur Basin at that

343

time is lacking so far. However, some information about the vegetation history (Ma et al., 2003) points to poorly developed desert grassland and gallery forests along the rivers, although the freshwater lake expanded to > 10 000 km2 at about 30 kyr ago (Wu¨nnemann, 1999; Hartmann, 2003). Prior to this event, phases of reduced lake levels at 57–55, 50–48 and 42–40 kyr coincide with enhanced loess transport (Fig. 22.9 F, G) and ice rafted debris (IRD) in the North Atlantic (McManus et al., 1999) On the other hand, mobile dune sand could not reach the centre of the lake, because the lake size remained large enough for sand to be transported by wind from the lake shore to the centre. Climate conditions were favourable for sand transport into the basin only just after phases of enhanced chemical precipitation of carbonates and sulphates (Fig. 22.8D), indicating a stepwise shrinkage of the lake after 47, 38, 32 and 26 kyr. Periods of warmer climate conditions and higher amounts of chemical precipitates roughly match interstadial phases of glacier decay in Tibet, whereas low lake levels and high loess transport generally match stadials of enhanced glacier advance and colder climate. As peaks of coarser loess in the Luochuan section (Fig. 22.9I) appear during phases of high lake levels in the Gaxun Nur Basin when most of the deflatable sediments were covered by water, we assume that loess transport at that time was probably more strongly affected by the westerlies, similar to the wind regime during the transitional phase from substage 5.1 to 4. 22.6.4 The LGM within stage 2 (LGM, 24–21 kyr BP) After the transition from stage 3 to the LGM at about 25 kyr, the aeolian influence increased remarkably, as inferred from aeolian deposits between 63 and 54 m core depth (Fig. 22.4A), while the water balance of the Gaxun Nur basin became as negative

344

Bernd Wu¨nnemann et al.

as in previous cold stages without complete desiccation. Increased aeolian flux since about 25 kyr coincides with the increase of coarse quartz grains in loess sections (Fig. 22.9I, Porter and An, 1995), both indicating increased wind strength under cold-dry climate conditions, leading to a strong expansion of the desert belt of about 20 longitude (Sun et al., 1998). Our proxy record ends at the transition from lacustrine to semilacustrine and fluvially dominated sediments after 24 kyr. However, previous investigations (Wu¨nnemann, 1999) show a continuation and even increase of the aeolian component under very shallow water conditions throughout stage 2. Fluvially transported sand and suspended load from northern and western catchments of the basin since the termination of stage 2 indicate a successive refilling of the lake as a result of locally enhanced rainfall by a strengthened summer monsoon after 14 kyr (Wu¨nnemann and Hartmann, 2002). Kohfeld and Harrison (2003) calculated that during stage 2 wind strength and mass accumulation rates were roughly 4.3 times higher than during stage 5. By contrast, our data indicate that wind intensity and aeolian flux during the cold substages of stage 5 and during stages 4 and 2 were probably not extremely different. Comparison with the Luochuan loess section supports our assumption of repeated dust fluxes with similar magnitudes during stages 4 and 2 and lower magnitudes during stages 5 and 3. More important is the fact that the duration of dust transport and the flow direction of loess-transporting air masses may have changed considerably through time. They probably had much stronger impact on depositional conditions than known so far. Further research on this topic is required. 22.6.5 Conclusion Our record from the Gaxun Nur Basin displays severe changes in the water balance during the last glacial cycle. On the assumption that our chronology is correct, phases of

positive water balance and thus the expansion of the terminal lake within a flat desert environment quite precisely follow global and regional patterns of highly variable climate conditions during the last interglacial and glacial cycle as documented in ice cores from Greenland and Tibet. During the warm interglacial stage 5.5, equal to the Eemian, the water balance of the Gaxun Nur system was controlled both by surface discharge from the northern and southern catchments of the lake and by sufficient local rainfall. It seems most likely that the necessary moisture supply was due to an intensified SE summer monsoon. This period of warm-wet climate conditions in Central Asian deserts during the last interglacial also prevailed in the Sahara as reported by Yan and Petit-Maire (1994) and Thiedig et al. (2000), indicating a marked decrease in desert environments along the Old World dryland belt. However, detailed palaeoclimate records which could prove the interglacial climates in the desert regions of China are still lacking. Conversely, periods of positive water balance in the Gaxun Nur Basin during the following cycles were mainly controlled by the surface discharge from the glaciated Tibetan Plateau. Glacier melt water and local rainfall in the upper catchment of the Hei River were the main sources of water supply towards the basin. Glacier advance in northern Tibet and enhanced rainfall in lower altitudes during phases of climate cooling may have been triggered by the influence of the westerly waves which provided the necessary moisture in late summer and early autumn to keep the glaciers growing. Consequently, the seasonal southward shift of the summer monsoon probably started earlier, while the Siberian high-pressure cell was not fully developed. This scenario would enable moist air masses to flow from western regions eastwards more frequently and use pathways along the west–east striking mountain barriers of the Tibetan Plateau. The periodically repeated climate changes in arid China are closely connected to global climate signatures mainly driven by solar

Interglacial and Glacial Fingerprints from Lake Deposits

forcing. Major climate changes throughout the last 130 kyr follow the insolation anomalies on a precessional band (Fig. 22.9A). However, our record shows that short-term variations of freshwater and aeolian fluxes respond to the East Asian Monsoon evolution (An et al., 1991; Porter and An 1995; An, 2000) and to the westerlies system as well (Pachur et al., 1995; Vandenberghe et al., 1997), both driven by a nonlinear pattern of thermohaline circulation and changes in ice volume (e.g. Ding et al., 1994). The synchronized coherence of loess transport from the Gaxun Nur Basin towards the Loess Plateau (Luochuan section) during cold events, comparable to those preserved in the Greenland ice cores and to IRD events (e.g. Xiao et al., 1995; McManus et al., 1999) indicates that cold dry air masses in the North Atlantic region may have fundamentally interacted with the winter monsoon and the westerlies over China. ACKNOWLEDGEMENT This research was funded by the German Ministry of Science and Technology (BMBF), grant UF-UFLD01097800-01LD0041 and by the Deutsche Forschungsgemeinschaft (DFG, Pa 161/16-4). We thank D. D. Rousseau, M. L. Siggaard-Andersen and M. F. Sa´nchez Gon˜i for their critical comments and suggestions to improve our manuscript. REFERENCES An, Z., 2000. The history and variability of the East Asian paleomonsoon climate. Quaternary Science Reviews 19, 171–187. An, Z.S., Porter, S.C., 1997. Millennial-scale climatic oscillations during the last interglaciation in central China. Geology 25, 603–606. An, Z.S., Kukla, G.J., Porter, S.C., Xiao, J., 1991. Magnetic susceptibility evidence of monsoon variation on the Loess Plateau of Central China during the last 130 000 years. Quaternary Research 36, 29–36. An, Z., Kutzbach, J.E., Prell, W., Porter, S.C., 2001. Evolution of Asian monsoons and phased uplift of

345

the Himalaya–Tibetan plateau since Late Miocene times. Nature 411, 62–66. Becken, M., Ho¨lz, S., Polag, D., Fiedler-Volmer, R., Burkhardt, H., 2003. Electromagnetic investigation in the Gaxun Nur Basin, Inner Mongolia, China. In: Mischke, S., Wu¨nnemann, B., Riedel, F. (Eds), International Symposium-Environmental Change in Central Asia. Department of Earth Sciences, Berlin, Abstracts, 8–10. Berger, A., Loutre, M.F., 1991. Insolation values for the climate of the last 10 million years. Quaternary Science Reviews, 10, 297–373. Berger, A., Loutre, M.F., 2002. An exceptionally long interglacial ahead? Science 297, 1287–1288. Bond, G., Showers, W., Cheseby, M., Lotti, R., Almasi, P., deMenocal, P., Priore, P., Cullen, H., Hajdas, I., Bonani, G., 1997. A pervasive millennial-scale cycle in North Atlantic Holocene and glacial climates. Science 278, 1257–1266. Calov, R., Ganopolski, A., Claussen, M., Petukhov, V., Greve, R., 2005. Transient simulation of the last glacial inception. Part I: Glacial inception as a bifurcation in the climate system. Climate Dynamics 24, doi: 10.1007/s00382-005-0007-6. Chen, F., Wang, S., Li, J., Shi, Y., Cao, J., Zhang, Y., Wang, Y., Kelts, K., 1995. Paleomagnetic record from RH lacustrine core in Zoige Basin of the Tibetan Plateau. Science in China, Ser. B 38 (12), 1513–1521. Clement, A.C., Cane, M.A., Saeger, R., 2001. An orbitally driven tropical source for abrupt climate change. Journal of Climate 14, 2369–2375. Cunningham, W.D., Windley, B.F., Dorjnamjaa, D., Badamgarov, J., Saandar, M., 1996. Late Cenozoic transpression in southwestern Mongolia and the Gobi Altai-Tien Shan connection. Earth and Planetary Science Letters 140, 67–81. Dansgaard, W., Johnsen, S.J., Clausen, H.B., DahlJensen, D., Gundestrup, N.S., Hammer, C.U., Hvidberg, C.S., Steffensen, J.P., Dveinbjo¨rnsdottir, A.E., Jouzel, J., Bond, G., 1993. Evidence for general instability of past climate from a 250-kyr icecore record. Nature 364, 218–220. Ding, Z.L., Yu, Z.W., Rutter, N.W., Liu, T.S., 1994. Towards an orbital time scale for Chinese loess deposits. Quaternary Science Reviews 13, 39–70. Domro¨s, M., Peng, G., 1988. The Climate of China. Springer, Berlin, 350 pp. EPICA Members, 2004. Eight glacial cycles from an Antarctic ice core. Nature 429, 623–628. Eugster, H.P., Hardie, L.A., 1978. Saline lakes. In: Lerman, A.B., (Ed.), Lakes-Chemistry, Geology, Physics, Springer, 237–293. Frenzel, B., 1994. Zur Pala¨oklimatologie der letzten Eiszeit auf dem Tibetischen Plateau. Go¨ttinger Geographische Abhandlungen 95, 115–141.

346

Bernd Wu¨nnemann et al.

Gansu Provincial Geological Bureau, 1980. Geological report on 1:200 000 scale geological survey (in Chinese), Lanzhou Press. Hartmann, K., 2003. Spa¨tpleistoza¨ne und holoza¨ne Morphodynamik im no¨rdlichen Gaxun Nur Becken, Innere Mongolei, NW- China. Dissertation FU Berlin, 1–123. Hetzel, R., Niedermann, S., Tao, M., Kubik, P.W., Ivyochs, S., Strecker, M.R., 2002. Low slip rates and long-term preservation of geomorphic features in Central Asia. Nature 417, 428–432. Jouzel, J., Hoffmann, G., Parrenin, F., Waelbroeck, C., 2002. Atmospheric oxygen 18 and sealevel change. Quaternary Science Reviews 21, 307–314. Kempe, S., Degens, E.T., 1978. Hydrographie, Warvenchronologie und organische Geochemie des Van-Sees, Ost-Tu¨rkei. In: Degens, E.T., Kurtmann, F. (Eds), The Geology of Lake Van, Ankara, 56–63. Kohfeld, K.E., Harrison, S.P., 2003. Glacial–interglacial changes in dust deposition on the Chinese Loess Plateau. Quaternary Science Reviews 22, 1859–1878. Kukla, G., An, Z., 1989. Loess stratigraphy in Central China. Palaeogeography, Palaeoclimatology, Palaeoecology 72, 811–814. Leuschner, D.C., Sirocko, F., 2003. Orbital insolation forcing of the Indian Monsoon – a motor for global climate changes? Palaeogeography, Palaeoclimatology, Palaeoecology 197, 83–95. Li, Y., Yang, J., Tan, L., Duan, F., 1999. Impact of tectonics on alluvial landforms in the Hexi Corridor, NW China. Geomorphology 28, 299–308. Ma, Y.Z., Zhang, H., Pachur, H.-J., Wu¨nnemann, B., Li, J., Feng, Z., 2003. Late Glacial and Holocene vegetation history and paleoclimate of the Tengger Desert, northwestern China. Chinese Science Bulletin 48, 1457–1463. Maher, B.A., Thompson, R., 1995. Paleorainfall reconstructions from pedogenic magnetic susceptibility variations in the Chinese loess and paleosols. Quaternary Research 44, 383–391. McManus, J.F., Oppo, D.W., Cullen, J.L., 1999. A 0.5million-year record of millennial-scale climate variability in the North Atlantic. Science 283, 971–975. Mischke, S., Demske, D., Wu¨nnemann, B., Schudack, M.E., 2005. Groundwater discharge to a Gobi desert lake during Mid and Late Holocene dry periods. Palaeogeography, Palaeoclimatology, Palaeoecology 225, 157–172. Molnar, P., Taponnier, P., 1975. Cenozoic tectonics of Asia: effects of a continental collision. Science 189, 419–426. NorthGRIP Project Members, 2004. High-resolution record of Northern Hemisphere climate extending into the last interglacial period. Nature 431, 147–151. Pachur, H.-J., Wu¨nnemann, B., Zhang H., 1995. Lake evolution in the Tengger Desert, Northwestern

China, during the last 40 000 years. Quaternary Research 44, 171–181. Petit, J.R., Jouzel, J., Raynaud, D., Barkov, N.I., Barnola, J.M., Basile, I., Bender, M., Davis, M., Delaygue, G., Delmotte, M., Kotlyakov, V.M., Legrand, M., Lipenkov, V.Y., Lorius, C., Pe´pin, L., Ritz, C., Saltzman, E., Stievenard, M., 1999. Climate and atmospheric history of the past 420 000 years from the Vostok ice core, Antarctica. Nature 399, 429–436. Porter, S.C., An, Z.S., 1995. Correlation between climate events in the North Atlantic and China during the last glaciation. Nature 375, 305–308. Pye, K., Zhou, L.P., 1989. Late Pleistocene and Holocene Eolian Dust Deposition in North China and the Northwest Pacific Ocean. Palaeogeography, Palaeoclimatology, Palaeoecology 73, 11–23. Rousseau, D.D., Wu, N., 1997. A new molluscan record of the monsoon variability over the past 130 000 yr in the Luochuan loess sequence, China. Geology 25, 275–278. Rousseau, D.D., Antoine, P., Hatte´, C., Lang, A., Zo¨ller, L., Fontugne, M., Ben Othman, D., Luck, J.M., Moine, O., Labonne, M., Bentaleb, I., Jolly, D., 2002. Abrupt millennial climatic changes from Nussloch (Germany) Upper Weichselian eaolian records during the Last Glaciation. Quaternary Science Reviews 21, 1577–1582. Sa´nchez Gon˜i, M.F., Eynaud, F., Turon, J.L., Shackleton, N.J., 1999. High resolution palynological record off the Iberian margin: direct land–sea correlation for the last interglacial complex. Earth and Planetary Science Letters 171, 123–137. Sa´nchez Gon˜i, M.F., Loutre, M.F., Crucifix, M., Peyron, O., Santos, L., Duprat, J., Malaize, B., Turon, J.L., Peypouquet, J.P., 2005. Increasing vegetation and climate gradient in Western Europe over the Last Glacial Inception (122–110 ka): datamodel comparison. Earth and Planetary Science Letters 231, 111–130. Schultz, H., Rad, U.von, Erlenkeuser, H., 1998. Correlation between Arabian Sea and Greenland climate oscillations of the past 110 000 years. Nature 393, 54–57. Shackleton, N.J., Sa´nchez Gon˜i, M.F., Pailler, D., Lancelot, Y., 2003. Marine Isotope Substage 5e and the EemianInterglacial. Global and Planetary Change 36, 151–155. Shi Y., Yu, G., Liu, X., Li, B., Yao, T., 2001. Reconstruction of the 30–40 ka BP enhanced Indian monsoon climate based on geological records from the Tibetan Plateau. Palaeogeography, Palaeoclimatology, Palaeoecology 169, 69–83. Sirocko, F., Seelos, K., Schaber, K., Rein, B., Dreher, F., Diehl, M., Lehne, R., Ja¨ger, K. Krbetschek, M., Degering, D., 2005. A late Eemian aridity pulse in

Interglacial and Glacial Fingerprints from Lake Deposits central Europe during the last glacial inception. Nature 436, 833–836. Sun, J., Ding, Z., Liu, T., 1998. Desert distributions during the glacial maximum and climatic optimum: Example of China. Episodes 21 (1), 28–30. Thiedig, F., Oezen, D., El Chair, M., Geyh, M.A., 2000. Absolute age of the Quaternary Lacustrine Limestone of the Al Mahruqh Formation-Muzuq Basin. In: Sola, M.A., Worsley, D. (Eds.), Geological Exploration in Murzuq Basin, 89–116. Elsevier, Rotterdam. Thompson, L.G., 2000. Ice core evidence for climate change in the Tropics: implications for our future. Quaternary Science Reviews 19, 19–35. Thompson, L.G., Mosley-Thompson, E., Bolzan, J., Dai, N., Gundestrup, N., Yao, T., Wu, X., Klein, L., Xie, Z., 1989. Holocen/Late Pleistocene climatic ice records from Qinghai-Tibetan Plateau. Science 246, 474–477. Thompson, L.G., Yao, T., Davis, M.E., Henderson, K. A., Mosley-Thompson, E., Lin, P.N., Beer, J., Synal, H.A., Cole-Dai, J., Bolzan, J.F., 1997. Tropical climate instability: The last glacial cycle from a Qinghai-Tibetan ice core. Science 276, 1821–1825. Thouveny, N., Carcaillet, J., Moreno, E., Leduc, G., Ne´rini, D., 2004. Geomagnetic moment variation and paleomagnetic excursions since 400 kyr BP: a stacked record from sedimentary sequences of the Portuguese margin. Earth and Planetary Science Letters 219, 377–396. Vandenberghe, J., An, Z.S., Nugteren, G., Lu, H.Y., Huissteden, K.V., 1997. New absolute time scale for the Quaternary climate in the Chinese loess region by grain size analyses. Geology 25, 35–38. Waelbroeck, C., Labeyrie, L., Michel, E., Duplessy, J.C., McManus, J.F., Lambeck, K., Balbon, E., Labracherie, M., 2002. Sealevel and deep water temperature changes derived from benthic foraminifera isotopic records. Quaternary Science Reviews 21, 295–305. Wang, L., Sarnthein, M., Erlenkeuser, H., Grimalt, J., Grootes, P., Heilig, S., Ivanova, E., Kienast, M., Pelejero, C., Pflaumann, U., 1999. East Asian monsoon climate during the Late Pleistocene: high-

347

resolution sediment records from the South China Sea. Marine Geology 156, 245–284. Wu¨nnemann, B., 1999. Untersuchungen zur Pala¨ohydrographie der Endseen in der Badain Jaran- und Tengger Wu¨ste, Innere Mongolei, NorthwestChina. Habilitation thesis, Berlin, 236 pp. Wu¨nnemann, B., Hartmann, K., 2002. Morphodynamics and Paleohydrography of the Gaxun Nur Basin, Inner Mongolia, China. Zeitschrift fu¨r Geomorphologie. N.F., Suppl.-Bd. 126, 147–168. Wu¨nnemann, B., Pachur, H.-J., Zhang H., 1998. Climatic and environmental changes in the deserts of Inner Mongolia, China, since the Late Pleistocene. In: Alsharhan, A.S., Glennie, K.W., Whittle, G.L., Kendall, C.G. St. C. (Eds.), Quaternary Deserts and Climatic Change, 381–394. Balkema, Rotterdam. Xiao, J., Porter, S.C., An, Z., Kumai, H., Yoshikawa, S., 1995. Grain size of quartz as an indicator of winter monsoon strength on the loess plateau of Central China during the last 130 000 yr. Quaternary Research 43, 22–29. Yan, Z., Petit-Maire, N., 1994. The last 140 ka in the Afro-Asian arid/semi-arid transitional zone. Palaeogeography, Palaeoclimatology, Palaeoecology 110, 217–233. Yang, Z., 1991. Evolution of Eastern Shelf of China in Quaternary and its environmental consequences. In: Liang, M., Zhang, J. (Eds.), Correlation of Onshore Quaternary in China, 1–22. Science Press, Beijing (in Chinese, english abstract). Zhang, H.C., Ma, Y.Z., Li, J.J., Qi, Y., Chen, G.J., Fang, H.B., Wu¨nnemann, B., Pachur, H.J., 2001. Palaeolake evolution and abrupt climate changes during last glacial period in NW China. Geophysical Research Letters 28 (16), 3202–3206. Zhang, H., Wu¨nnemann, B., Ma, Y.Z., Peng, J., Pachur, H.-J., Li J., Qi, Y., Chen, G., Fang, H., Feng, Z., 2002. Lake Leveland Climate Change between 40 000 and 18 000 14C Years BP in Tengger Desert, NW China. Quaternary Research 58, 62–72. Zhang, H.C., Peng, J.L., Ma, Y.Z., Chen, G.J., Feng, Z.D., Li, B., Fan, H.F., Chang, F.Q., Lei, G.L., Wu¨nnemann, B., 2004. Late Quaternary palaeolake levels in Tengger Desert, NW China. Palaeogeography, Palaeoclimatology, Palaeoecology 211, 45–58.

This page intentionally left blank

Section 4 Climate, Vegetation and Mammalian Faunas in Europe during Middle Pleistocene Interglacials (MIS 7, 9, 11) (ed. Thomas Litt)

This page intentionally left blank

23. Introduction: Climate, Vegetation and Mammalian Faunas in Europe during Middle Pleistocene Interglacials (MIS 7, 9, 11) Thomas Litt Institute for Palaeontology, University of Bonn, Nussallee 8, D-53115 Bonn, Germany

The basic principles used in subdividing the Quaternary into chronostratigraphic units are the same as for other Phanerozoic chronostratigraphic units which require boundary definitions and designations of boundary stratotypes (Salvador, 1994). However, in contrast to the rest of the Phanerozoic, the division of Quaternary sequences on the basis of climatic changes documented in sediment records is fundamental and has a long tradition. Classifications based on climatostratigraphic units such as ‘glacials’ or ‘interglacials’ are reasonably well established in different countries or areas and are accepted as regional standards (Gibbard and van Kolfschoten, 2004; Litt et al., 2005). Regarding the lower and upper boundary of the Middle Pleistocene, the responsible stratigraphic commission suggests to use the palaeomagnetically defined Brunhes/ Matuyama boundary (780 kyr, MIS 19) also as boundary between Lower and Middle Pleistocene (Richmond, 1996). There is a broad international consensus about this proposal (e.g. Pillans 2003). However, a formally defined boundary stratotype (GSSP) is still missing. The boundary between Middle and Upper Pleistocene is also not yet formally defined. Gibbard (2003) proposes to use the beginning of the last interglacial (around 130 kyr after U-series dating of calcite speleothems, see Spo¨tl et al., this volume) as Middle/Upper Pleistocene boundary. In addition, he suggests the newly processed cores from Amsterdam Terminal (parastratotype for the Eemian, see van Leeuwen et al., 2000) as boundary stratotype (GSSP). This proposal is currently under examination by a working group of the stratigraphical commission.

Whereas the basal chronostratigraphic units (stages/ages) were not formally defined so far internationally, cold and warm intervals are used as chronostratigraphic units in terms of regional stages of Middle Pleistocene in northern and central Europe (e.g. Elsterian Stage, Holsteinian Stage, see Gibbard and van Kolfschoten, 2004; Litt et al., 2005) (Fig. 23.1). The chronostratigraphic units include several complex stages (Cromerian, Saalian). Even in this case, the recognition of distinct criteria and of boundaries for the climatostratigraphic division is crucial. Numerous regional continental stratigraphic units and boundary stratotypes of the Quaternary were defined using palynostratigraphic criteria. The climatostratigraphic terms interglacial and interstadial were first defined by Jessen and Milthers (1928) for periods with characteristic records of nonglacial climate, as indicated by palaeobotanical evidence for major vegetation changes. Following these suggestions, interglacials in central Europe have been identified unequivocally as temperate periods with a climate optimum at least as the present interglacial (Holocene) in the same region. Interstadials have been described as periods that were either too short or too cold to reach the climate level of interglacial type in the same region. There is an increasing tendency of directly correlating terrestrial sequences with marine isotope records which is mainly based on curve matching. However, correlations must rely entirely on direct dating. Gibbard and West (2000) suspect that the extension of the use of isotope–stratigraphic subdivision and terminology of marine-core sequences (and

352

Thomas Litt

4

Holocene

2 4

0.1 6

0.2

5e 7

8

0.3 10

0.4 0.5

12

9 11 13

14 15

0.6

16

0.7

18

0.8

20 22

17 19 21

Holocene

Weichselian Eemian Drenthe Hoogeveen(W)

Weichselian Eemian Drenthe+Warthe Dömnitz (W) Fuhne (C) Holsteinian Elsterian

Holsteinian Elsterian

Poland

North Germany Ice advances

Holocene

Saalian Complex

4.5

Ice advances

Holocene

Ice advances

Vistulian Eemian Odranian+Wartanian Zbóinian (W) Liviecian (C) Mazovian Sanian 2

Ruhme/Bilshausen (W)

Cromerian Complex

δ18O

Late Pleistocene

5

Middle Pleistocene

5.5

Brunhes polarity Chron

Age (Ma) 0.0

The Netherlands

Saalian Complex

Marine isotope record

Noordbergum Glacial c Rosmalen Glacial b

Cold Stage Warm Stage Cold Stage Warm Stage Cold Stage

Westerhoven

Hunteburg (W) Cold Stage

Glacial a Waardenburg

Osterholz (W)

Ferdynandovian

Sanian 1 (C) Malopolanian (W) Nidamian (C) Augustovian 2 (W)

Fig. 23.1 Chronostratigraphical correlation table of the Middle and Upper Pleistocene in north-central Europe (compiled after Zagwijn, 1989; Gibbard and van Kolfschoten, 2004; Lindner et al., 2004; Litt et al., 2005). (C) – cold stage, (W) – warm stage.

ice-core sequences) on land might lead to a new stratigraphic Esperanto. Regional terrestrial stratigraphy, marine stratigraphy and ice-core stratigraphy are each a key to different parts of the climate system. For understanding the regional variability, it is essential to develop local and independent stratigraphies and chronologies. With respect to the terrestrial record, favourable geological conditions in southern Europe have in some cases led to the relatively undisturbed accumulation of thick Quaternary sedimentary sequences. Such sequences provide an opportunity to develop complete records of terrestrial events over multiple glacial–interglacial cycles. The linking of the longest pollen sequences from southern Europe (Tenagi Phillipon in Greece, Velle di Castiglione in Italy and Bouchet/Praclaux in France) has led to the emergence of a coherent stratigraphical framework of changes in vegetation for the last 450 kyr and has allowed tentative comparisons with the marine isotopic record (Tzedakis et al., 1997, 2001). This showed that the many stages and substages into which the marine isotopic sequence is divided into are also appropriate for viewing the continental record, although the marine and

terrestrial boundaries may not be precisely synchronous. The last five marine isotopic interglacials were investigated in Iberian margin deep-sea cores. The direct comparison between terrestrial pollen, benthic/planctic foraminifera and isotopes from the same cores shows that temperature changes are not in phase with ice volume variations (Desprat, this volume). Pronouced offsets between marine isotopic warm stage boundaries and forested intervals, as previously decribed by Sa´nchez Gon˜i et al. (2002) for MIS 5, are also documented for older interglacial stages (see Roucoux et al., this volume). Caused by biogeographical reasons, it is difficult to correlate interglacials from south European long sequences with warm stages from north-central Europe based on palynology alone. The way forward should be to date fixed events as accurately as possible based on new approaches in absolute dating methods. Based on these tie points, the astronomical cyclicity can provide a finer-scale chronology. A good geochronological frame can resolve long-standing disagreements such as whether the Holsteinian relates to MIS 9 or 11. Questions of the nature and duration of the Saalian, and

Climate, Vegetation and Mammalian Faunas during the Middle Pleistocene in Europe

precisely how many warm events occur within the Saalian cannot be resolved by ‘counting-backwards’ methods (Gibbard and West, 2000). The north-central European area known to have been affected by both the Elsterian and the Saalian glaciations provides good opportunity to establish a regional chronostratigraphy of the Middle Pleistocene based on stratigraphic superposition: the relationship of the Cromerian Complex to the Elsterian glaciation and the Holsteinian Interglacial to the Saalian Complex. In other parts of Europe from which interglacial pollen sequences are known, relationships to particular glaciations are sometimes difficult to elucidate. Therefore, the following brief overview about the Middle Pleistocene in north-central Europe should demonstrate the importance of interglacial sequences as stratotypes in relation to glacial deposits. 23.1 CROMERIAN COMPLEX STAGE This stage comprises several interglacials and is therefore believed to cover a much longer time span than the Cromerian Interglacial Stage (sensu stricto) as defined in East Anglia (West, 1980). The Cromerian Complex Stage of the Netherlands is defined by the recognition of at least four warm temperate and three cold substages (Zagwijn, 1985) indicating the climatic complexity of this time interval. As described previously, the Early–Middle Pleistocene boundary should be linked to the Brunhes–Matuyama palaeomagnetic boundary which has been recognized as falling within the end of Interglacial I (Waardenburg) of the Cromerian Complex or a bit later (780 kyr, MIS 19; see Turner, 1996). The upper boundary of the Cromerian Complex is widely agreed in north-central Europe. It is placed where major climate cooling heralds the onset of the Elsterial Glacial Stage. In northern Germany, the warm stages of the lower Middle Pleistocene

353

were palynostratigraphically defined based on a long continental record at Gorleben (Mu¨ller, 1986, 1992). The Gorleben sequence encompasses five warm phases in a stratigraphic superposition above the Bavelian Stage and below the Elsterian Stage. The oldest one is the mainly reversely magnetized Osterholz Interglacial (¼ Waardenburg). The youngest one is the Rhume (or Bilshausen) Interglacial, immediately preceding the Elsterian ice advance. It is the beststudied Cromerian interglacial in northern Germany. The duration of this interglacial is estimated as about 27 000 years, based on varve counts of Mu¨ller (1992). The pollen sequence shows several oscillations and two distinct forest declines, which may indicate fast climate deteriorations. At least the younger decline follows a tephra layer derived from the Eifel volcanic field. The pollen sequence as recorded for the Bilshausen Interglacial correlates well with the corresponding part of the Ka¨rlich Interlgacial in the Middle Rhine area (Bittmann and Mu¨ller, 1996) which represents the early Arvicola cantianus fauna (see von Koenigswald, this volume). The age of the so-called Brockentuff – a tephra layer related to the Ka¨rlich Interglacial – is about 400 kyr based on 40Ar/39Ar laser method (van den Boogard et al., 1998), which would correspond to MIS 11 (and to the Praclaux Interglacial in the Massif Central, southern France). 23.2 ELSTERIAN STAGE The term Elsterian named after the river Elster in central Germany first appeared on the geological maps (1:25 000) of the ‘Ko¨niglich Preußische Geologische Landesanstalt’ (i.e. Keilhack, 1911). It is the oldest glaciation represented by widespread till sheets throughout northcentral Europe (Eissmann et al., 1995; Ehlers et al., 2004). Two glacial cycles can be

354

Thomas Litt

identified in northern Germany during the Elsterian Stage. However, no intra-Elsterian interglacial has been identified. During both ice advances of the Elsterian Stage, major erosional structures were formed (subglacial channels and basins). In the erosional zones, lakes were formed which persisted into the Holsteinian Stage. 23.3 HOLSTEINIAN STAGE The term Holsteinian originates from Geikie (1894), who described interglacial marine sediments as Holsteinian beds. Hallik (1960) first defined Holsteinian sediments palynostratigraphically and correlated them with continental limnic interglacial records. Type sections of the Holsteinian are Hamburg-Dockenhuden (marine deposits) and Bossel, west of Hamburg (lacustrine deposits) (see Jerz and Linke, 1987). The vegetation succession of the Holsteinian warm period was later described by several authors who assigned regionally differing pollen assemblage zones (Erd, 1973 for eastern Germany; Mu¨ller, 1974 for north-western Germany; Krupinski, 2000 for Poland, where this interglacial stage is named Mazovian). The INQUA Subcommission on European Quaternary Stratigraphy defined the lower boundary of the Holsteinian as the transition from subarctic (still late Elsterian) to boreal conditions, and the upper boundary as the transition from boreal to subarctic (Saalian) conditions (Jerz and Linke, 1987). The duration of the Holsteinian is estimated as about 15–16 000 years, based on varve counts of Mu¨ller (1974) at Munster-Breloh, a milestone in the geochronological study of the Hosteinian stage in northern Germany. New Th/U datings based on peat deposits from the type section Bossel indicate an age of about 310–330 kyr BP (Geyh and Mu¨ller, 2005; and this volume). The consequence would be that the Holsteinian Interglacial in the type region as well as the Hoxnian Interglacial in England and the Landos Interglacial in the

Massif Central, France (Reille et al., 2000), are correlated with MIS 9. The interglacial vegetation development reconstructed by palynological data is very similar throughout north-central Europe and begins with a pine-birch forest. The immigration of thermophilous trees including alder, oak, elm, lime, esh, yew and hazel occurred more or less simultaneously. The early expansion of spruce is remarkable. Hornbeam and fir immigrated during the course of the interglacial. Particularly characteristic of numerous Holsteinian sites in north-central Europe is the appearance of Pterocarya and Azolla filiculoides. Our knowledge of the Holsteinian vegetation history is mainly based on pollen records obtained from interglacial lake sediments in glacial basins which were already formed during the Elsterian. In this respect, the discovery of a new Holsteinian sequence at Do¨ttingen in the Western Eifel volcanic field (Diehl and Sirocko, this volume) is a real progress, because it is situated outside the glacially affected region. Nevertheless, the palynological features are similar to those of the classical Holsteinian sites in north Germany. The biostratigraphical correlation and synchronization is obvious. An unsolved problem is the direct link between palynologically defined Holsteinian sites and mammalian faunas. From none of the typical north-central European Holsteinian deposits have mammalian remains been described so far. Middle Pleistocene interglacial faunas younger than the Elsterian were mainly found in sites with uncertain stratigraphical position (see von Koenigswald, this volume). The first half of the Holsteinian is characterized by temperatures somewhat lower than today. In the second half, the reconstructed mean temperatures are higher than today, in particular the July temperature (Ku¨hl and Litt, this volume). In addition, the Holsteinian seems to be less stable than the present interglacial (Holocene) or the last interglacial (Eemian) with some intrainterglacial coolings. The magnitude of the

Climate, Vegetation and Mammalian Faunas during the Middle Pleistocene in Europe

main cooling in the mid-Holsteinian is reconstructed as approximately 5 C for January temperature. No great change is reconstructed for July temperature during this episode. 23.4 SAALIAN COMPLEX STAGE Based on the definition of the Subcommission on European Quaternary Stratigraphy (Litt and Turner, 1993), the Saalian Complex Stage encompasses the period from the end of the Holstein Interglacial Stage (boundary between boreal and subarctic phase of the subsequent Fuhne cold phase) to the beginning of the Eemian Interglacial Stage (beginning of the birch zone). After these specifications, the Saalian proves to be a complex unit including several cold and warm fluctuations, whereas the latter may even reach the character of an interglacial. The Lower Saalian Complex Stage, i.e. the period between the end of the Holsteinian Stage and the first Saalian ice advance, is characterized by extensive valley-widening and intense accumulation of fluvial gravels. In the exposures (lignite open pits) of the type region in central Germany, a generally continuous, 5–20 m thick coarse sand and gravel terrace is observed (Eissmann et al., 1995). This important stratigraphic marker horizon separates the Elsterian and Saalian glacigenic sequences. In many profiles, several successive generations of ice wedge casts are found. Silty intercalations are often disturbed by cryoturbation. However, this terrace complex reflects changing climatic conditions during the lower Saalian substage. There is some evidence in north-central Europe of at least one pronounced warm event (Do¨mnitz warm Stage in north-eastern Germany after Erd, 1965 and as a synonym Wacken warm Stage in north-western Germany after Menke, 1968), possibly even of two warm periods after Urban (1995, this volume). However, these warm phases

355

(Reinsdorf, Scho¨ningen) are documented as incomplete pollen sequences. In no case are they separated by glacial sediments and stratigraphically they are situated before the first Saalian ice advance. At least for the Scho¨ningen Interglacial (Urban, 1995), which probably correlates with the Wacken/Do¨mnitz warm Stage, 230 Th/234U dates are available, suggesting a correlation with MIS 7 (see Urban, this volume), whereas the Fuhne cold Stage just after the Hosteinian, which is only documented with periglacial deposits, could be equivalent to MIS 8. Comparable temperate conditions in between a periglacial climate both before and afterwards have been documented in the Netherlands (Zagwijn, 1985; Vandenberghe, 1995). This temperate interval (Hoogeveen) also preceded the advance of the Saalian land ice. Similar to Wacken and Do¨mnitz, it is decribed rather as an interstadial with interglacial character than a full interglacial based on the absence of Abies and thermophilous genera such as Hedera and Buxus which do occur in the underlying Holsteinian beds. The equivalent of MIS 7 in the continental European scale is well developed in southern European long sequences from the Massif Central, France (Reille et al., 2000). This stage is characterized by three warm phases interrupted by stadials. It is interesting to note that none of these warm intervals in the Massif Central (named as Bouchet interstadials 1–3) reach the climatic level of a full-developed interglacial (such as in MIS 5e or 9 and 11). Several ice advances are known to have occurred in north-central Europe during the Upper Saalian. In northern Germany, the subdivision into two major ice advances has been used since Woldstedt (1954). The older Saalian ice advance (the so-called Drenthe) marks the maximum extent of the Saalian ice sheet. The younger Saalian ice advance is named Warthe. It mussed be stressed that between the maximum of the Saalian glaciation (i.e. the Drenthe phase) and the Eemian

356

Thomas Litt

no true interglacial intervened. With a high probability, both Drenthe and Warthe are correlated with MIS 6.

REFERENCES Bittmann, F., Mu¨ller, H., 1996. The Ka¨rlich interglacial site and its correlation with the Bilshausen sequence. In: Turner, C. (Ed.), The Early Middle Pleistocene in Europe, Balkema, Rotterdam, pp. 187–193. van den Boogard, C., Boogard, P. van den, Schmincke, H.-U., 1998. Quarta¨rgeologischtephrostratigraphische Neuaufnahme und Interpretation des Pleistoza¨nprofils Ka¨rlich, F.R.G. Eiszeitalter und Gegenwart 39, 62–86. Ehlers, J., Eissmann, L., Lippstreu, L., Stephan, H.-J., Wansa, S., 2004. Pleistocene glaciations in North Germany. In: Ehlers, J., Gibbard, P.L. (Eds.), Quaternary Glaciations – Extent and Chronology. Part I: Europe, Amsterdam, pp. 135–146. Eissmann, L., Litt, T., Wansa, S., 1995. Elsterian and Saalian deposits in their type area in central Germany. In: Ehlers, J. et al. (Eds.), Glacial Deposits in North-East Europe, Rotterdam, pp. 439–464. Erd, K., 1965. Pollenanalytische Gliederung des mittelpleistoza¨nen Richtprofils PritzwalkPrignitz. Eiszeitalter und Gegenwart 16, 252–253. Erd, K., 1973. Pollenanalytische Gliederung des Pleistoza¨ns der Deutschen Demokratischen Republik. Zeitschrift fu¨r geologische Wissenschaften 1, 1087–1103. Geikie, J., 1894. The Great Ice Age, and its relation to the antiquity of man. Stanfort, London, 850 pp. Geyh, M.A., Mu¨ller, H., 2005. Numerical 230Th/U dating and a palynological review of the Holsteinian/Hoxnian interglacial. Quaternary Science Reviews 24, 1861–1872. Geyh, M.A., Mu¨ller, H., 2006. Palynological and geochronological study of the Holsteinian/Hoxnian/ Landos Interglacial (this volume). Gibbard, P.L., 2003. Definition of the Middle-Upper Pleistocene boundary. Global and Planetary Change 36, 201–208. Gibbard P.L., West, R. G., 2000. Quaternary chronostratigraphy: the nomenclature of terrestrial sequences. Boreas 29, 329–336. Gibbard. P.L., van Kolfschoten, T., 2004. The Pleistocene and Holocene epochs. In: Gradstein, F., Ogg, J., Smith, A. (Eds.), A Geologic Time Scale 2004, Cambridge, pp. 441–452. Hallik, R., 1960. Die Vegetationsentwicklung der Holstein-Warmzeit in Nordwestdeutschland und

die Altersstellung der Kieselgurlager der su¨dlichen Lu¨neburger Heide. Zeitschrift der Deutschen Geologischen Gesellschaft 112, 326–333. Jessen, K., Milthers, V., 1928. Stratigraphical and palaeontological studies of freshwater deposits in Jutland and north-west Germany. Danmarks Geologiske Undersøgelse, II Raekke 48. Jerz, H., Linke, G., 1987. Arbeitsergebnisse der Subkommission fu¨r Europa¨ische Quarta¨rstratigraphie: Typusregion des Holstein-Interglazials (Berichte der SEQS 8). Eiszeitalter u. Gegenwart 37, 145–148. Keilhack, K., 1911. Geologische Karte von Preußen 1:25000. Erla¨uterungen zu Blatt Teltow. von Koenigswald, W., 2006. Mammalian Faunas from the interglacial periods in Central Europe and their stratigraphic correlation (this volume). Krupin´ski, K.M., 2000. Palynostratigraphic correlation of deposits of the Mazowian interglacial of Poland, Warsaw. Ku¨hl, N., Litt, T., 2006. Quantitative time series reconstructions of Holsteinian and Eemian temperatures using botanical data (this volume). Lindner, L., Gozhik, P., Marciniak, B., Marks, L., Yelovicheva, Y., 2004. Main climatic changes in the Quaternary of Poland, Belarus and Ukraine. Geological Quarterly 48, 97–114. Linke, G., Hallik, R., 1993. Die pollenanalytischen Ergebnisse der Bohrungen HamburgDockenhuden (qho 4), Wedel (qho 2) und Hamburg-Billbrock. Geologisches Jahrbuch A138, 169–184. Litt, T., Turner, C., 1993. Arbeitsergebnisse der Subkommission fu¨r Europa¨ische Quarta¨rstratigraphie: Die Saalesequenz in der Typusregion. Eiszeitalter und Gegenwart 43, 125–128. Litt, T., Ellwanger, D., Villinger, E., Wansa, S., 2005. Das Quarta¨r in der Stratigraphischen Tabelle von Deutschland 2002. Newsletters on Stratigraphy 41, 385–399. Menke, B., 1968. Beitra¨ge zur Biostratigraphie des Mittelpleistoza¨ns in Norddeutschland (pollenanalytische Untersuchungen aus Westholstein). Meyniana 18, 35–42. Mu¨ller, H., 1974. Pollenanalytische Untersuchungen und Jahresschichtenza¨hlungen an der holsteinzeitlichen Kieselgur von Munster-Breloh. Geologisches Jahrbuch A 21, 107–140. Mu¨ller, H., 1986. Altquarta¨re Sedimente im Deckgebirge des Salzstockes Gorleben. Zeitschrift der Deutschen Geologischen Gesellschaft 137, 85–95. Mu¨ller, H., 1992. Climatic changes during and at the end of the interglacials of the Cromerian Complex. In: Kukla, G.J., Went, E. (Eds.), Start of a Glacial. NATO ASI Series I (3), 51–69.

Climate, Vegetation and Mammalian Faunas during the Middle Pleistocene in Europe Pillans, B., 2003. Subdividing the Pleistocene using the Matuyama–Brunhes boundary (MBB): an Australasian perspective. Quaternary Science Reviews 22, 1569–1577. Reille, M., de Beaulieu, J.L., Svoboda, H., AndrievPonel, V., Goeudry, C., 2000. Pollen analytical biostratigraphy of the last five climatic cycles from a long continental sequence from the Velay region (Massif Central, France). Journal of Quaternary Science 15 , 665–685. Richmond, G.M., 1996. The INQUA-approved provisional Lower-Middle Pleistocene boundary. In: Turner, C. (Ed.), The Early Middle Pleistocene in Europe, Balkema, Rotterdam, pp. 319–326. Salvador, A. (Ed.), 1994. International Stratigraphic Guide: A Guide to Stratigraphic Classification, Terminology, and Procedure. 2. ed. XIX þ 214 pp; Trondheim, Boulder (International Union of Geology Sciences/Geology Society of America). Sa´nchez Gon˜i, M.F., Cacho, I., Turon, J-L., Guiot, J., Sierro, F.J., Peypouquet, J-P., Grimalt, J.O., Shackleton, N.J., 2002. Synchroneity between marine and terrestrial responses to millennial scale climatic variability during the last glacial period in the Mediterranean region. Climate Dynamics 19, 95–105. Turner, C., 1096. A brief survey of the early Middle Pleistocene in Europe. In: Turner, C. (Ed.), The Early Middle Pleistocene in Europe, Balkema, Rotterdam, pp. 295–317. Tzedakis, P.C., Andrieu, V., de Beaulieu, J-L., Birks, H.J.B., Crowhurst, S., Follieri, M., Hooghiemstra, H., Magri, D., Reille, M., Sadori, L., Shackleton, N.J., Wijmstra, T.A., 1997. Comparison of terrestrial and marine records of changing climate of the last 500,000 years. Earth and Planetary Science Letters 14, 967–982.

357

Tzedakis, P.C., Andieu, V., de Beaulieu, J-L., Birks, H.J.B., Crowhurst, S., Follieri, M., Hooghiemstra, H., Magri, D., Reille, M., Sadori, L., Shackleton, N.J., Wijmstra, T.A., 2001. Establishing a terrestrial chronological framework as a basis for biostratigraphical comparisons. Quaternary Science Reviews 20, 1583–1592. Urban, B., 1995. Palynological evidence of younger Middle Pleistocene interglacials (Holsteinian, Reinsdorf and Scho¨ningen) in the Scho¨ningen open cast lignite mine (eastern Lower Saxony, Germany). Mededelingen Rijks Geologische Dienst 52, 175–186. Urban, B., 2006. Interglacial pollen records from Scho¨ningen, north Germany (this volume). Vandenberghe, J., 1995. The Saalian Complex and the first traces of human activity in the Netherlands in a stratigraphic and ecologic context. Mededelingen Rijks Geologische Dienst 52, 187–194. West, R.G., 1980. The pre-glacial Pleistocene of the Norfolk and Suffolk coasts. Cambridge University Press. Woldstedt, P., 1954. Saaleeiszeit, Warthestadium und Weichseleiszeit in Norddeutschland. Eiszeitalter und Gegenwart 4/5, 34–48. Zagwijn, W.H., 1985. An outline of the Quaternary stratigraphy of the Netherlands. Geologie en Mijnbouw 64, 17–24. Zagwijn, W.H., 1989. The Netherlands during the Tertiary and the Quaternary: a case history of coastal lowland evolution. Geologie en Mijnbouw 68, 107–120. Zagwijn, W.H., 1996. The Cromerian Complex Stage of the Netherlands and correlation with other areas in Europe. In: Turner C. (Ed.), The Early Middle Pleistocene in Europe, Balkema, Rotterdam, pp. 145–172.

This page intentionally left blank

24. Fine-Tuning the Land–Ocean Correlation for the Late Middle Pleistocene of Southern Europe K.H. Roucoux1, P.C. Tzedakis1, L. de Abreu2 and N.J. Shackleton2 1

Earth and Biosphere Institute, School of Geography, University of Leeds, Leeds, LS2 9JT, UK 2 Department of Earth Sciences, Godwin Laboratory, University of Cambridge, New Museums Site, Pembroke Street, Cambridge, CB2 3SA, UK

ABSTRACT A marine pollen record from the Portuguese margin provides the means of correlating vegetation changes in southern Europe with North Atlantic sea-surface temperature and the marine isotope stratigraphy for the interval 180 to 345 kyr (marine isotope stages 7, 8 and 9). It reveals close correspondence between the patterns and timing of forest expansion and contraction and changes in North Atlantic sea-surface temperatures, but pronounced offsets between marine isotopic warm stage boundaries and forested intervals. Our study confirms the previous tentative observation that the latter vary in length from one stage to another and also highlights floristic differences between sites resulting from local climatic, geological and biogeographical factors. 24.1 INTRODUCTION The pattern of vegetation development in southern Europe during the last 450 kyr is well known from the long pollen records produced in sedimentary sequences from Greece (Wijmstra, 1969; Wijmstra and Smit, 1976), Italy (Follieri et al., 1988) and France (Reille and de Beaulieu, 1995; Reille et al., 1998, 2000) over the last 40 years. In contrast to the majority of sequences from northern Europe, these sites contain continuous records spanning multiple climatic cycles and thus provide an opportunity to study the response of vegetation at individual

sites to different combinations of environmental boundary conditions. They reveal a pattern of alternating periods of forest and more open vegetation on timescales of 104 to 105 years which are clearly a response to the Milankovitch-driven global climatic changes recorded in marine isotope records of global ice volume (Imbrie et al., 1984; Martinson et al., 1987). However, dating uncertainties mean that it has not been possible to determine the precise relationship between changes preserved in marine and terrestrial records on timescales of 103 to 104 years. If vegetation feedbacks are to be incorporated effectively into climate models, it is essential to establish the exact nature of this relationship (e.g. Crowley and Baum, 1997). Previous correlations have had to assume synchroneity between vegetational shifts and some feature of the marine record. The correlation scheme of Tzedakis et al. (1997) aligned four long pollen records from southern Europe to the SPECMAP benthic 18 O stack by assuming synchroneity between forest expansion and deglaciation, a necessary and reasonable assumption at the time. This study clarified the broad equivalence of terrestrial and marine signals, thus demonstrating that the marine isotope stratigraphy is a suitable framework within which to interpret terrestrial environmental records and suggested that the forested intervals, previously assumed to be of similar duration, in fact varied in length from one stage to another. We are now in a position to test the basis of the Tzedakis et al. (1997) correlation scheme

360

K.H. Roucoux et al.

using a pollen record generated recently in a marine sequence from the Portuguese margin. Combining high-resolution benthic and planktonic foraminiferal oxygen isotope data with pollen data, this sequence records the precise relationship between the marine isotope records of ice volume and sea-surface temperature, and vegetation development on land. Knowledge of this relationship enables fine-tuning of the correlation between the three key southern European pollen sequences which span the late Middle Pleistocene and the marine isotope stratigraphic scheme. 24.2 THE TERRESTRIAL RECORDS Cores from maar craters in the volcanic region of the Velay, in the French Massif Central, at sites Lac du Bouchet (44 559N, 3 479E, 1200 m a.s.l.) and Praclaux (44 499N, 3 509E, 1000 m a.s.l.) (Fig. 24.1) have produced a composite sedimentary sequence spanning the period from ca. 450 kyr to the present (Reille and de Beaulieu, 1995; Reille et al., 1995, 1998, 2000). The sequence

records an alternation of forest and open landscapes over time on Milankovitch time scales. Forested intervals tend to follow a typical succession, with Juniperus, Betula and Pinus in the early phases, followed by Quercus, Corylus, Carpinus and other mesic deciduous forest taxa, then Abies, Fagus and Picea (montane taxa with greater moisture requirements) and finally a period of Pinusdominated woodland. Between forested intervals, abundant Artemisia, Chenopodiaceae and Poaceae pollen indicate extensive steppe vegetation. Today mean annual precipitation reaches 1500 mm in this region. Winters are severe, with mean January temperatures of between 0 and 2 C, and summers relatively cool, with mean July temperatures of 16 to 18 C. This climatic regime accounts for the abundance of temperate/mesophilous taxa and presence of Picea (a tree of cool, wet and boreal climates) during past interglacial and interstadial episodes (Reille and de Beaulieu, 1995; Reille et al., 1995, 1998, 2000). In Italy, another ancient volcanic lake site, Valle di Castiglione situated 20 km east of Rome (41 539N 12 459E, 44 m a.s.l.), has

Fig. 24.1 Location of southern European pollen sites discussed in the text. Core MD01-2443 was retrieved at a depth of 2925 m, at 37 52.89N, 10 10.579W. Sites Lac du Bouchet (1200 m above sea level) and Praclaux (1000 m a.s.l.) are in the Velay region of the Massif Central, France, at 44 559N 3 479E and 44 499N 3 509E respectively (Reille and de Beaulieu 1995; Reille et al., 1998, 2000). The site of Valle di Castiglione (41 m a.s.l.) in Italy is at 41 539N 12 459E (Follieri et al., 1988). The site of Tenaghi Philippon (40 m a.s.l.) in northern Greece is at 41109N 24 209E (Wijmstra and Smit, 1976).

Fine-Tuning the Land–ocean Correlation

produced a long sedimentary sequence which spans the period from ca. 270 kyr to the present (Follieri et al., 1988) (Fig. 24.1). It records periods of high biomass, often diverse forest alternating with periods characterised by a dominance of steppe vegetation and much lower biomass (Follieri et al., 1988; Magri, 1989). Forest development typically follows a pattern similar to the succession in the Massif Central sequences, although Mediterranean tree taxa, including evergreen Quercus, Olea, Phillyrea, Fraxinus and Zelkova, were also present here. As in the Massif Central, forest intervals began and ended with a Pinus phase. The intervening open, herbaceous intervals were characterised by an abundance of Artemisia, Chenopodiaceae and Poaceae, indicating steppe vegetation. Today, mean annual precipitation at Valle di Castiglione is 800 mm, while mean temperatures are 7 C for January and 25 C for July. The chronologically longest Pleistocene sequence in southern Europe comes from the site of Tenaghi Phillippon (41109N 24 209E, 40 m a.s.l.) in northern Greece (Wijmstra, 1969) (Fig. 24.1). The sequence of peats and clays is 280 m long and spans the last ca. 900 kyr (Wijmstra and Groenhart, 1983). The section between 43 and 68 m represents MIS 7 to 9 (Wijmstra and Smit, 1976). The sequence records periods of forest vegetation, dominated by deciduous Quercus and Pinus, alternating with periods of steppe vegetation, dominated by Artemisia and Chenopodiaceae (ibid.). Located on the landlocked Drama plain, the site experiences a continental climate with a mean January temperature of 3.4 C in January and 23.9 C in July. Mean annual precipitation for the area is 600 mm. The region’s cool and dry climate probably accounts for the low abundance of moisture-requiring mesophilous taxa during previous warm stages and the relatively low abundance of Mediterranean taxa. A second terrestrial pollen sequence from Greece, the record from Ioannina, was also used in the Tzedakis et al. (1997) study, but as it is of lower

361

resolution than the other southern European sequences it is not considered here. A new high-resolution record from Ioannina is in the process of being generated (Tzedakis et al., 2002; Roucoux et al., in preparation). 24.3 THE MARINE RECORD Drilled in 2001 by the French research vessel Marion Dufresne, core MD01-2443 is one of several long deep-ocean sediment cores from the Portuguese margin that were collected with the aim of obtaining highresolution deep-sea records combined with a terrestrial pollen signal of contemporaneous vegetation on the adjacent continent. The benthic foraminiferal 18 O record in this core shows the classic succession of climatic cycles representing the accumulation and wasting of continental ice masses through marine isotope stages (MIS) 7, 8, 9 and the start of MIS 6 (Tzedakis et al., 2004a). The planktonic foraminiferal 18 O record at this site is thought to be dominated by seasurface temperature (Cayre et al., 1999; de Abreu, 2000; de Abreu et al., 2003), an inference supported by its close correspondence with temperature estimates based on planktonic foraminiferal faunal analyses (de Abreu, unpublished data). Abundance of the polar foraminifera Neogloboquadrina pachyderma (sinistral) is plotted in Fig. 24.3; it indicates the extent of polar water influence at the core site. There is some temporal offset between the benthic and planktonic 18 O curves, which is particularly pronounced at the deglaciations MIS 10/9 and MIS 8/7. The early shift in benthic 18 O values may be partly due to changes in deep-water temperatures and 18 O, reflecting changes in deep-water hydrography as shown for Termination I by Skinner and Shackleton (2005). The timescale for MD01-2443 (Tzedakis et al., 2004a) was developed by aligning the benthic 18 O record to the Antarctic Vostok deuterium (D/H) record (Petit et al., 1999) (Fig. 24.2). This is based on the implications of the study by Shackleton et al. (2000) which

362

K.H. Roucoux et al.

Fig. 24.2 Development of the age model for MD01-2443. Age control was established by alignment of the MD01-2443 benthic foraminiferal oxygen isotope record (d18 O ‰) to the Vostok deuterium record (D/H ‰) (Petit et al., 1999; Parrenin, personal communication). Inflections of the sediment accumulation rate (SAR) curve indicate positions of the tie-points.

showed strong similarity between the benthic 18 O record off Portugal and Antarctic temperatures. The implication that shifts in benthic 18 O are synchronous with shifts in Antarctic temperature allows the use of chronologies developed for the Antarctic ice cores to generate an age model for MD01-2443. The Vostok timescale used here has been developed by F. Parrenin (personal communication) through alignment with the EPICA Dome C record (EPICA Community Members, 2004) and is considered an improvement on the Vostok glaciological timescale (GT4) of Petit et al. (1999). This is because ice at the Vostok site originates upstream, where accumulation rates are poorly constrained, whereas ice at Dome C accumulates in situ, which facilitates derivation of a timescale based on an ice flow model. Figure 24.2 shows the alignment of the MD01-2443 record to Vostok D/H and the resulting sediment accumulation rates (SAR). Inflections of the SAR curve show the location of the control points, which

were chosen at the midpoints of transitions between MIS. Pollen analysis in MD01-2443 (see Tzedakis et al., 2004a for methodology) reveals phases of deciduous Quercus-dominated forest, ericaceous heathland and open steppe, but the temperate forest succession that typifies the terrestrial sequences is not present. Instead, forests were dominated by deciduous Quercus throughout, with small amounts of Corylus, Alnus, evergreen Quercus and Olea at times of maximum forest extent, followed by expansion of ericaceous heath as forest contracted (Fig. 24.3). This may be due to an absence of local refugia of temperate trees during glaciations combined with a tendency in Portugal for rapid soil leaching and acidification (a consequence of widespread granitic bedrock and relatively high precipitation). Today, mean annual precipitation ranges from 600 mm in the lowlands of the lower reaches of the river Tagus to 2000 mm in the mountains nearby. Mean temperatures are between 10 and 12 C in January and between 20 and 24 C in July.

Fine-Tuning the Land–ocean Correlation

363

Fig. 24.3 Pollen percentage data, benthic and planktonic oxygen isotopic data (Tzedakis et al., 2004a), and Neogloboquadrina pachyderma (sinistral) as a percentage of total planktonic foraminifera, for core MD012443. Pollen data are cumulative percentages: Mediterranean trees and shrubs; temperate or Eurosiberian trees and shrubs; Ericaceae; steppe taxa; pioneer taxa; other herbs (see legend). Forested intervals, where arboreal pollen percentages rise above 20%, are indicated in grey and labelled with their local stage name. Marine isotope stages (MIS) are also labelled and the isotopically defined temperate intervals are indicated in dark grey.

364

K.H. Roucoux et al.

24.4 CORRELATION Our correlation of the terrestrial sequences with the marine pollen record is based on two key assumptions: firstly, that climatic events originating in the North Atlantic are translated rapidly across southern Europe, since atmospheric circulation patterns result in dominantly westerly air flow over the region; and secondly, that vegetation responses to North Atlantic climatic warming were also rapid and effectively synchronous across southern Europe. Deciduous Quercus populations persisted at all of the southern European pollen sites through cold intervals (both glacial and stadial), thus ruling out the possibility of a migrational lag at the start of warm periods. The responsiveness of deciduous Quercus to millennial scale climatic oscillation during MIS 3 across the region (Allen et al., 1999; Sa´nchez Gon˜i et al., 2000, 2002; Tzedakis et al., 2004b; Roucoux et al., 2005) supports this conclusion. We correlate the terrestrial sequences with the marine pollen record using the midpoints of increases in arboreal pollen (AP) percentages, which represent the expansion of forest in response to climatic warming and increased moisture availability, as tie-points. We assume a constant SAR between tie-points.

A systematic definition of the beginning and end of forested intervals is clearly desirable. The value of 50% AP, or 40% AP excluding Pinus; has been used in terrestrial pollen sequences to define stage boundaries (Zagwijn, 1989). However, in the marine pollen record, a lower value is more appropriate because of the over-representation of herbaceous taxa in the marine pollen spectra (Roucoux, 2000; Tzedakis et al., 2004a). Even during peak interglacial conditions, maximum temperate tree values in MD01-2443 are ca. 50% compared with 80 to 90% in terrestrial records. Comparison with terrestrial Holocene sequences (van der Knaap and van Leeuwen, 1995, 1997; Roucoux, 2000) suggests that an appropriate AP value to represent the development of closed forest cover in the marine pollen record is 20% excluding Pinus; which is strongly over-represented in marine records (Heusser and Balsam, 1977; Tzedakis et al., 2004a). This value also coincides with the midpoint of the increases in AP in most instances. Forested intervals in MD01-2443 have been assigned local stratigraphic names (Table 24.1; Fig. 24.3). The correlation scheme for the three terrestrial pollen sequences and the marine isotope stratigraphy, established via correlation with the MD01-2443 pollen record, is

Table 24.1: Correlation scheme for temperate stages in the terrestrial and marine pollen sequences and the marine isotope stratigraphy. Marine stratigraphic nomenclature

Chronology in MD01-2443 (ka)

Local stratigraphic nomenclature

MIS

MIS lower boundaries

Onset of forest intervals

MD01-2443 (see Figure 3)

Bouchet / Praclaux

Valle di Castiglione

Tenaghi Philippon

7a 7c 7e (TIII) 9a 9c 9e (TIV)

200 216.8 246 291 318 338

198.5 215.3 243.2 290 315.5 336.6

Belem Cascais Estoril Mafra Queluz Lisboa

Bouchet 3 Bouchet 2 Bouchet 1 Amargiers Ussel Landos

Roma 3 Roma 2 Roma 1 n/a n/a n/a

H2–3 Symvolon H1 Symvolon Strymon Kavalla Krimenes Litochoris

Ages are given for the lower boundaries of the warm marine isotope stages, as identified in MD01-2443, and for the onset of forested intervals in this core (see Figure 2 and Tzedakis et al., 2004a for basis of chronology). Local stratigraphic names are assigned to forest intervals in MD01-2443 (see Figure 3). Local stage names for temperate intervals in the terrestrial sequences are from Reille et al. (2000), Follieri et al. (1988) and Wijmstra (1969).

Fine-Tuning the Land–ocean Correlation

shown in Table 24.1. When plotted on the same timescale (Fig. 24.4), it is clear that the four pollen sequences show very similar patterns of forest expansion and contraction. There follows a description of the relationship between conditions in the North Atlantic, the marine isotope stratigraphy and vegetation development in southwest Portugal based on the record in MD012443, together with a description of the vegetation changes taking place contemporaneously, according to our correlation, across southern Europe (Fig. 24.4). 24.4.1 MIS 9 At the start of MIS 9, MD01-2443 records the shift to lighter 18 O values approximately 1000 years earlier in the benthic record than in the planktonic, probably caused in part by changes in deep-water hydrography (Skinner and Shackleton, 2005). The associated forest expansion in southwest Portugal occurs ca. 1000 years later than the planktonic shift. Faunal planktonic foraminiferal assemblages (Fig. 24.3) indicate that this was probably due to the continued presence of polar water masses offshore; forest expanded abruptly as soon as polar waters retreated at 336 kyr. A similarly abrupt forest expansion is recorded in the Tenaghi Philippon (Wijmstra and Smit, 1976) and Praclaux (Reille et al., 2000) sequences. At all three sites, forest is dominated by deciduous Quercus in the earliest phase (337–333 kyr). The end of the Quercus phase marked the end of the forested interval in Portugal since Quercus populations did not recover and ericaceous heath expanded instead. This means that the forest interval was short, at only 3.6 kyr, compared with the duration of MIS 9e (14 kyr). At the terrestrial sites, there was a transient reduction in tree populations at ca. 333 kyr which, in contrast to the pattern recorded in MD01-2443, was followed by a recovery: at Praclaux, Quercus was replaced by Carpinus (Reille et al., 2000) and at Tenaghi Philippon, Quercus populations recovered (Wijmstra and Smit, 1976). Ericaceae extent in Portugal

365

reached a peak at 327 kyr, which coincides with the timing of Fagus and Abies expansion in France and Greece. Thus, the forested interval recorded in the French and Greek sequences appear to last somewhat longer (with durations of ca. 11 and 13 kyr respectively) than in Portugal and thus correspond more closely with the planktonic 18 O plateau of MIS 9e. This may of course be an artefact of varying sedimentation rate, but the fact that the duration corresponds closely with the isotopic plateau combined with the continued abundance of Ericaceae, which suggests relatively warm conditions in Portugal, supports this conclusion. Falling sea-surface temperatures and increasing global ice volume from 327 kyr coincide with shrinking ericaceous heathland in Portugal and contraction of forest populations recorded at Lac du Bouchet (Reille et al., 2000) and Tenaghi Philippon (Wijmstra and Smit, 1976). Relatively large tree populations, dominated by Pinus, persisted in the Massif Central through MIS 9d, and the degree of steppe population expansion was minor compared with the other stadial and glacial intervals. Pinus populations were more strongly reduced at Tenaghi Philippon, suggesting greater aridity in northern Greece than in France at this time. Both marine and terrestrial sites record moderate temperate forest extent during MIS 9c, dominated by deciduous Quercus with much smaller proportions of other temperate trees than during MIS 9e. At all four sites, the pattern of forest expansion and contraction corresponds to the pattern of planktonic 18 O values in MD01-2443. Planktonic 18 O records two clear warm intervals with a cooler phase in between, while the pollen records show two corresponding intervals of increased forest extent. This indicates that Atlantic influence on the climate in France and Greece was as strong as that experienced in southwest Portugal. Between 305 and 299 kyr, ericaceous heath was extensive in Portugal, while Picea was abundant at Lac du Bouchet (Reille et al., 2000), both suggesting high precipitation levels and low temperatures.

366

K.H. Roucoux et al.

Fig. 24.4 Comparison of southern European pollen records and the Portuguese margin pollen record for the interval 180 to 345 kyr. (a) Benthic and planktonic foraminiferal isotope ratios in MD01-2443; (b) Cumulative percentages of AP excluding Pinus (black) and Ericaceae (grey) in MD01-2443; (c) AP excluding (black) and including (grey) Pinus in the composite Lac du Bouchet/Praclaux sequence (after Reille et al., 2000); (d) same in Valle di Castiglione (after Follieri et al., 1988); (e) same in Tenaghi Philippon (after Wijmstra and Smit, 1976). Terrestrial sequences are aligned to the marine sequence using the midpoint of rapid increases in AP (excluding Pinus) as tie-points (positions indicated by a diamond). All are plotted on the Antarctic ice-core-derived timescale (see Fig. 24.2 and Tzedakis et al., 2004a).

Fine-Tuning the Land–ocean Correlation

In MD01-2443, MIS 9b records a pronounced tree population collapse in southwest Portugal coincident with a moderate and brief incursion of polar water offshore and a large accumulation of continental ice (more clearly recorded in other North Atlantic sequences; e.g. McManus et al., 1999; Shackleton 2000; Desprat et al., this volume). Sequences from both Tenaghi Phillipon and Praclaux also record very small residual temperate tree populations indicating cold, dry conditions across southern Europe (Wijmstra and Smit, 1976; Reille et al., 2000; Tzedakis et al., 2003). Forest expansion in southwest Portugal at the start of MIS 9a again corresponds closely with planktonic 18 O, and the offset with the marine isotope stage boundary is small. Temperate tree populations were intermediate in extent between those of 9c and 9e. At MD01-2443, Valle di Castiglione and Tenaghi Philippon, Quercus dominates, while at Lac du Bouchet Carpinus and Picea also contribute significantly to the vegetation (Wijmstra and Smit, 1976; Follieri et al., 1988; Reille et al., 2000). A gradual contraction of forest towards MIS 8 is recorded at all sites except Lac du Bouchet where temperate tree populations appear to collapse suddenly, although the pattern at this site may be distorted by hiatuses in the sequence (Reille et al., 1998). Increasing ice volume and decreasing sea-surface temperatures coincide with a decline in heath and forest populations after 276 kyr in southwest Portugal, and with a decline of Pinus and temperate tree populations at the other sites, indicating widespread cooling and drying. 24.4.2 MIS 8 During MIS 8, the MD01-2443 benthic 18 O record indicates greatest ice volume near the beginning of the stage, a pattern contrary to that of the last two glaciations (e.g. McManus et al., 1999; Shackleton, 2000). Likewise, the planktonic 18 O record shows lowest sea-surface temperatures in the early part of the stage with warmer conditions

367

in the latter half, although there is a short renewal of low temperatures just before the start of MIS 7e. The Portuguese vegetation reflects this climatic pattern, with the most extensive steppe populations and smallest tree populations near the beginning (276.6–264 kyr), and larger tree populations towards the end (264–252 kyr). The terrestrial sequences similarly show the smallest tree populations during the early part of MIS 8. Tree population expansion during the latter part of MIS 8 is experienced at all sites, characterised by deciduous Quercus (with Ericaceae) at MD01-2443, Pinus at Lac du Bouchet (Reille et al., 2000), Betula, deciduous Quercus, Corylus and Pinus at Valle di Castiglione (Follieri et al., 1988) and deciduous Quercus and Pinus at Tenaghi Philippon (Wijmstra and Smit, 1976). It may be that conditions were only warm and/or wet enough to allow the expansion of temperate trees at the most southerly sites, while further north, in France, only Pinus could expand. All the sites record a return to open vegetation at the end of MIS 8 (252–246 kyr). 24.4.3 MIS 7 The Portuguese forest expansion associated with MIS 7e coincides closely with the increasingly light values of planktonic 18 O. There is, however, a delay between forest onset and the MIS 7e boundary which, as in Termination IV, may be smaller than it appears if changes in deep-water hydrography are contained in the benthic 18 O record. The forested interval of MIS 7e in southwest Portugal is much shorter (6.2 kyr) than the marine isotopic warm stage (17 kyr), ending midway through the plateau in planktonic 18 O (which is between 233.5 and 241 kyr) in a similar way to the forested interval of MIS 9e, but with the important difference that in MIS 7e heath expands more gradually and does not dominate the vegetation until much later on in the stage. The three terrestrial sites also show a short forested interval during MIS 7e.

368

K.H. Roucoux et al.

The shortest is in France where a deciduous Quercus and Corylus phase is followed by a Carpinus phase ending at 239 kyr (Reille et al., 2000). The collapse of mesophilous forest populations in France coincides with the first major decline of Quercus populations in southwest Portugal. At Tenaghi Philippon, temperate forests were dominated by Quercus throughout (Wijmstra and Smit, 1976), while at Valle di Castiglione a complete forest succession is recorded (Follieri et al., 1988). Tree populations at these two sites appear to have persisted for slightly longer (by 1.3 kyr according to our correlation with MD01-2443) than those in southwest Portugal or the Massif Central, corresponding more closely to the plateau in benthic and planktonic 18 O, but still ending well before the marine isotopic warm stage (Fig. 24.4). The ensuing period of small (but oscillating) tree populations and expanding ericaceous heath in southwest Portugal coincided with falling sea-surface temperatures in the North Atlantic and a gradual increase in global ice volume towards the end of MIS 7e. This period saw brief expansions of tree populations at Valle di Castiglione (Quercus and Fagus) and Lac du Bouchet (Pinus) (Follieri et al., 1988; Reille et al., 2000), but the record from Tenaghi Philippon does not register such oscillations. The pronounced contraction of tree populations and expansion of steppe vegetation in southwest Portugal coinciding with MIS 7d indicates cold and dry conditions on land, similar to those of the preceding glaciation (MIS 8). Temperatures offshore, which had been decreasing gradually since the peak of the preceding temperate stage, reached glacial levels and polar water masses extended southwards to this latitude (Fig. 24.3). Records from Antarctica and the North Atlantic show that MIS 7d was an extremely cold stadial event (Ruddiman and McIntyre, 1982; McManus et al., 1999; Petit et al., 1999), the low global temperatures and large ice volume resulting from high eccentricity, low obliquity and high precession which combined to give

low June insolation in the Northern hemisphere at this time (Berger, 1978). Across southern Europe, all three terrestrial sequences record strong reductions in the extent of mesophilous tree populations and expansion of steppe vegetation (Wijmstra and Smit, 1976; Follieri et al., 1988; Tzedakis et al., 1997; Reille et al., 1998, 2000). During MIS 7c, forest expansion continued to track planktonic 18 O closely in southwest Portugal but lagged benthic 18 O producing an offset between the MIS boundary and the forested phase. MIS 7c saw the longest forested interval of MIS 7 (at 10.3 kyr), which reflects the long duration of warm seasurface temperatures (Tzedakis et al., 2004a). The three terrestrial sequences also record a long forested interval coincident with MIS 7c, which was the most floristically diverse of the temperate stages considered here (Follieri et al., 1988; Tzedakis et al., 2003). The initial deciduous Quercus peak in southwest Portugal coincided with the Quercus peaks at the other sites. Then, while in Portugal Quercus continued to dominate and Ericaceae populations were expanding, the other sites record a succession with Carpinus followed by Abies. At Valle di Castiglione, this was followed by a major expansion of Fagus (Follieri et al., 1988) and at Lac du Bouchet by moderate expansions of Fagus and Picea (Reille et al., 2000). Heath in Portugal contracted after 206 kyr which coincided with falling sea-surface temperatures and increasing global ice volume. An increase in the rate of tree population decline in the terrestrial sequences which, in our correlation scheme, appears contemporaneous with this, suggests cooling and/or drying across southern Europe at this time. During MIS 7b, Atlantic sea-surface temperatures declined only slightly (McManus et al., 1999). In MD01-2443, polar foraminifera remain rare through this interval (Fig. 24.3), while sea-level reconstructions indicate only minor accumulation of ice (Shackleton, 2000). Of all the stadial events of this sequence, this period has the least polar water influence at the latitude of southern

Fine-Tuning the Land–ocean Correlation

Portugal. Among the pollen sites, only the sequence from Lac du Bouchet shows a very pronounced tree population contraction to glacial levels, perhaps reflecting its relatively northerly position (Reille et al., 2000). Tree populations closely tracked temperatures offshore throughout MIS 7a. This warm interval is represented by a tree population expansion which, although shortlived, is more pronounced than in either of the previous MIS 7 warm stages. A similarly large expansion in response to MIS 7a warming is recorded at the terrestrial sites (Reille et al., 2000; Wijmstra and Smit, 1976; Tzedakis et al., 2003). This is unexpected since in Antarctica this was the coolest of the MIS 7 temperate stages (Petit et al., 1999), insolation was low relative to MIS 7c (Berger, 1978), and ice volume was similar to MIS 7c and MIS 7e (Shackleton, 2000). A possible explanation for this apparent discrepancy is the persistence of relatively large tree populations through the preceding stadial MIS 7b. Beginning from a larger starting population, forest expansion at the subsequent warming could lead to a largescale tree population expansion even in response to a relatively minor warming (Reille et al., 2000). The peak in deciduous Quercus populations in Portugal associated with MIS 7a corresponds to a complete forest succession at Lac du Bouchet and Valle di Castiglione (Follieri et al., 1988; Reille et al., 2000), while at Tenaghi Philippon forest remained dominated by Quercus (Wijmstra and Smit, 1976). Although all the sites register oscillations in tree populations as they shrank towards the transition to MIS 6, they appear to have remained extensive in Valle di Castiglione and Tenaghi Philippon for longer than in Portugal or the Massif Central. 24.4.4 MIS 6 The first 10 kyr of MIS 6 are recorded here. MD01–2443 records falling sea-surface temperatures and increasing global ice volume coincident with shrinking forests

369

and heathland, and the expansion of steppe in southwest Portugal. All three terrestrial sites record expansion of steppe vegetation, indicating the onset of extremely cold and dry conditions across southern Europe (Wijmstra and Smit, 1976; Follieri et al., 1988; Reille et al., 2000). 24.5 DISCUSSION The most important conclusion of the Tzedakis et al. (1997) correlation scheme is supported by our study, namely that there is broad correspondence between marine isotopically defined warm intervals (periods of low ice volume) and forested intervals on land. However, with the details afforded by the marine pollen record, this relationship can be seen to be complex. Vegetation and climate in southwest Portugal are more strongly influenced by North Atlantic sea-surface temperatures than by global ice volume, which makes sense a priori given that today terrestrial climate in southern Europe is largely determined by North Atlantic conditions. Furthermore, the relationship between benthic 18 O and global ice volume at any one site may be complicated by the influence of hydrography on deep-water temperatures and 18 O (Skinner and Shackleton, 2005). As a result of these various factors, forest expansion is often not synchronous with the deglaciation at the end of marine isotopic cold stages. For instance, at Termination IV the continued presence of polar water offshore caused tree population expansion to be delayed relative to the deglaciation; as soon as temperate waters arrived, temperate tree populations expanded immediately. There are two instances (the midwarm stage forest declines during MIS 7e and MIS 9e) when abrupt tree population changes cannot be accounted for entirely by conditions in the North Atlantic, nor indeed by shifts in global ice volume, since the changes taking place do not seem large enough to account for collapse of tree populations. An

370

K.H. Roucoux et al.

alternative suggestion has been made for the cause of these forest declines by Tzedakis et al. (2004a) who observed that maximum forest extent coincides with the period of maximum air temperatures and greenhouse gas concentrations (methane and carbon dioxide) in the Vostok ice core record (Petit et al., 1999). The ensuing forest collapse coincides with a rapid decrease in temperature and greenhouse gas concentrations. This correspondence suggests that the vegetation changes recorded for southwest Portugal are reflecting an abrupt, global scale shift in atmospheric and climatic conditions which is not registered in the planktonic or benthic 18 O curves. Thus, land– ocean correlations based on the actual relationship between vegetation and the benthic oxygen isotope record established in situ in a combined marine pollen record represent an improvement over the assumption of synchroneity that has underlain previous attempts. In addition, this correlation has enabled the assignment of a reliable age model to the terrestrial sequences, which provides a more accurate general chronological framework of vegetation changes than hitherto possible. However, the assumption of linear SARs between control points means that deviations from real age of specific events are possible. The patterns of vegetation change shown in the marine and terrestrial sequences share many similarities in their approximate duration and in the amplitude of changes in forest extent recorded. However, there are strong differences in the precise nature of the vegetation responses, i.e. the patterns of succession and taxa involved, between sites. These give rise to slight differences in the length of forest intervals between sites for many of the temperate periods, suggesting that we should not assume that periods of interglacial conditions necessarily result in forest vegetation everywhere, for the whole of their duration. Clearly, climate is modified by geography and topography, which in turn account for differences in vegetation response between sites. To take

one example, the MD01-2443 pollen record shows that in Portugal heathland expanded (perhaps as a consequence of soil degradation) as each warm stage progressed, while terrestrial sequences show that elsewhere in southern Europe forests continued to thrive. As another example, during cold intervals, drier sites, like Tenaghi Philippon, experienced more pronounced tree population contraction than more oceanic sites. When warming was very minor, for example, during the second half of MIS 8, the southerly sites saw expansion of temperate tree populations, while further north, in the Massif Central, only Pinus populations, being tolerant of cold and drought, could expand. Thus, there is no fundamental reason to expect the nature or duration of any forest interval associated with an interglacial to be the same everywhere in Europe. Peculiarities of geography, geology and biogeographical history result in a variable relationship between global climatic state and local vegetation. There are limitations to the conclusions we can draw about the relative timing of vegetation change, changes in sea-surface conditions and the positioning of MIS boundaries from the correlation presented here. Firstly, changes in deep-water hydrography contained in the 18 O of benthic foraminifera may distort the true phase relationship between the onset of the marine and terrestrial stages at Terminations. Secondly, our correlation assumes that the tree population response to warming was synchronous across southern Europe. Given the differences in the precise nature of the vegetation response and in forest duration between sites it may be that the underlying causes (differences in soils, local climate and biogeography for example) could cause differences in timing of forest expansion at the beginning of warm intervals as well. However, there are reasons to think otherwise. At all of the sites, deciduous Quercus expands first in response to warming; this taxon is known to have persisted at each of the sites throughout cold intervals; thus there is no

Fine-Tuning the Land–ocean Correlation

reason to expect any migrational lag in its response to climatic warming. This argument is supported by the demonstrable sensitivity of Quercus populations to millennial scale climatic fluctuations during MIS 3 (Allen et al., 1999; Sa´nchez Gon˜i et al., 2000, 2002; Roucoux et al., 2001; Tzedakis et al., 2004b) and by the fact that deciduous Quercus responded rapidly to warming at the start of the Holocene (e.g. Willis, 1994; Lawson et al., 2004). A consideration of the shape of the planktonic 18 O curve offers further support: the rapid rise in sea-surface temperature at the start of each warm period contrasts with the more gradual cooling towards the end. Even if local factors mean that the threshold between forest and nonforest vegetation corresponds to different sea-surface temperatures from site to site, the crossing of the various thresholds will occur closer together in time during a rapid change in SST than during a more gradual change. Hence, the vegetation response to rapid climatic warming at the start of interglacials is likely to be approximately synchronous across southern Europe, even allowing for variation in local conditions. 24.6 CONCLUSIONS The correlation of marine and terrestrial pollen records presented here confirms the broad correspondence between terrestrial forested, or temperate, intervals and marine isotopically defined warm intervals on orbital timescales across southern Europe. Closer inspection of the land–ocean relationship, at the millennial scale, reveals offsets between the MIS boundaries and the shifts in forest extent on land that define the terrestrial temperate stages. The main factors contributing to these offsets are probably (1) the overriding effect of North Atlantic sea-surface temperature in determining the timing of forest expansion, and (2) the effect of decreases in global temperature and greenhouse gas concentration, recorded early in MIS 7e and 9e in the

371

Vostok ice core to which the premature end of forested intervals in southern Europe may be related. In addition, hydrographic effects on benthic 18 O may lead to apparent earlier deglaciations (Terminations) and thus artificially stretch the offset between the MIS and terrestrial stage boundaries. Comparison between the pollen records reveals that forested intervals may vary considerably in length from one stage to the next, but differences in the length of forested intervals between sites are generally small. These differences in the length of forested intervals along with floristic differences result from local climatic, geological and biogeographical factors such as the effects of altitude and latitude on temperature, of distance from the sea and topography on precipitation levels, of the propensity of soils to become leached, and of the proximity of sites to glacial refugia for temperate trees. ACKNOWLEDGEMENTS This work was undertaken in association with the ‘POP Project’, EC Grant EVK22000-00089. Funding for KHR was provided by the Natural Environment Research Council. Funding for LA was provided by the Portuguese Foundation for Science and Technology under the fellowship contract: SFRH/BPD/1588/2000 and by the Calouste Gulbenkian Foundation, through a visiting fellowship to Woods Hole Oceanographic Institution. We are grateful to the French MENRT, TAAF, CNRS/INSU and especially to IFRTP for the coring operations aboard the Marion Dufresne II. We thank Mike Hall for isotopic measurements, John Corr for palynological sample preparation and Ian Lawson for comments on the manuscript. REFERENCES Allen, J.R.M., Brandt, U., Brauer, A., Hubberten, H.W., Huntley, B., Keller, J., Kraml, M., Mackensen, A.,

372

K.H. Roucoux et al.

Mingram, J., Negendank, J. F.W., Nowaczyk, N.R., Zolitschka, B., 1999. Rapid environmental changes in southern Europe during the last glacial period. Nature 400, 740–743. Berger, A., 1978. Long-term variations in caloric insolation resulting from the Earth’s orbital elements. Quaternary Research 9, 139–167. Cayre, O., Lancelot, Y., Vincent, E., Hall, M., 1999. Paleoceanographic reconstructions from planktonic foraminifera off the Iberian Margin: temperature, salinity and Heinrich events. Paleoceanography 14, 384–396. Crowley, T.J., Baum, S.K., 1997. Effect of vegetation on an ice-age model simulation. Journal of Geophysical Research 102 (D14), 16,463–16,480. de Abreu, L., 2000. High resolution palaeoceanography off Portugal during the last two glacial cycles. PhD thesis, University of Cambridge, 370 pp. de Abreu, L., Shackleton, N.J., Schonfeld, J., Hall, M., Chapman, M., 2003. Millennial-scale oceanic climate variability off the western Iberian margin during the last two glacial periods. Marine Geology 196, 1–20. Desprat, S., Sa´nchez Gon˜i, M.F., Naughton, F., Turon, J.-L., Duprat, J., Malaize´, B., Cortijo, E., Peypouquet, J.-P. (this volume) Climate variability of the last five isotopic interglacials: direct land–sea–ice correlation from the multiproxy analysis of north western Iberian margin deep-sea cores. EPICA Community Members, 2004. Eight glacial cycles from an Antarctic ice core. Nature 429, 623–628. Follieri, M., Magris, D., Sadori, L., 1988. 250,000-year pollen record from Valle di Castiglione (Roma). Pollen et Spores 30, 329–356. Heusser, L.E., Balsam, W.L., 1977. Pollen distribution in the NE Pacific Ocean. Quaternary Research 7, 45–62. Imbrie, J., Hays, J.D., Martinson, D.G., McIntyre, A., Mix, A.C., Morley, J.J., Pisias, N.G., Prell, W.L., Shackleton, N.J., 1984. The orbital theory of Pleistocene climate: support from a revised chronology of the marine 18 O record. In: Berger, A.L., Imbrie, J., Hays, J.D., Kukla, G., Saltzman, B. (eds.). Milanktovitch and Climate. D. Reidel, Dordrecht, pp. 269–306. Lawson, I.T., Frogley, M.R., Bryant, C., Preece, R., Tzedakis, P. C., 2004. The Lateglacial and Holocene environmental history of the Ioannina basin, north-west Greece. Quaternary Science Reviews 23, 1599–1625. Magri, D., 1989. Interpreting long-term exponential growth of plant populations in a 250 000-year long pollen record from Valle di Castiglione (Roma). New Phytologist 112, 123–128. Martinson, D.G., Pisias, N.G., Hays, J.D., Imbrie, J., Moore Jr., T.C., Shackleton, N.J., 1987. Age dating

and the orbital theory of ice ages: development of a high-resolution 0 to 300 000-year chronostratigraphy. Quaternary Research 27, 1–29. McManus, J.F., Oppo, D.W., Cullen, J.L., 1999. A 0.5-million-year record of millennial-scale climate variability in the North Atlantic. Science 283, 971–975. Petit, J.R., Jouzel, J., Raynaud, D., 1999. Climate and atmospheric history of the past 420 000 years from the Vostok ice core, Antarctica. Nature 399, 429–436. Reille, M., de Beaulieu, J-L., 1995. Long Peistocene records from the Praclaux crater, south-central France. Quaternary Research 44, 205–215. Reille, M., Andrieu, V., de Beaulieu, J-L., Guenet, P., Goeury, C., 1998. A long pollen record from Lac du Bouchet, Massif Central, France: for the period ca. 325 to 100 ka (OIS 9c to OIS 5e). Quaternary Science Reviews 17, 1107–1123. Reille, M., de Beaulieu, J-L., Svobodova, H., AndieuPonel, V., Goeury, C., 2000. Pollen analytical biostratigraphy of the last five climatic cycles from a long continental sequence from the Velay region (Massif Central, France). Journal of Quaternary Science 15, 665–685. Roucoux, K.H., 2000. Millennial scale vegetation and climate variability in north-west Iberia during the last glacial stage. PhD thesis, University of Cambridge, 270 pp. Roucoux, K.H., Shackleton, N.J., de Abreu, L., Scho¨nfeld, J., Tzedakis, P.C., 2001. Combined marine proxy and pollen analyses reveal rapid Iberian vegetation response to North Atlantic climate oscillations. Quaternary Research 56, 128–132. Roucoux, K.H., de Abreu, L., Shackleton, N.J., Tzedakis, P.C., 2005. The response of NW Iberian vegetation to North Atlantic climate oscillations during the last 65 kyr. Quaternary Science Reviews 24, 1637–1653. Ruddiman, W.F., McIntyre, A., 1982. Severity and speed of Northern Hemisphere glaciation pulses: the limiting case? Geological Society of America Bulletin 93, 1273–1279. Sa´nchez Gon˜i, M.F., Eynaud, F., Turon, J-L., Gendreau, S., 2000. European climatic response to millennial-scale climatic changes in the atmosphere-ocean system during the Last Glacial period. Quaternary Research 54, 394–403. Sa´nchez Gon˜i, M.F., Cacho, I., Turon, J-L., Guiot, J., Sierro, F.J., Peypouquet, J-P., Grimalt, J.O., Shackleton, N. J., 2002. Synchroneity between marine and terrestrial responses to millennial scale climatic variability during the last glacial period in the Mediterranean region. Climate Dynamics 19, 95–105. Shackleton, N.J., 2000. The 100 000-year ice-age cycle identified and found to lag temperature, carbon

Fine-Tuning the Land–ocean Correlation dioxide and orbital eccentricity. Science 289, 1897– 1902. Shackleton, N.J., Hall, M.A., Vincent, E., 2000. Phase relationships between millennial-scale events 64 000–24 000 years ago. Paleoceanography 15, 565–569. Skinner, L.C., Shackleton, N.J., 2005. An Atlantic lead over Pacific deep-water change across Termination I: implications for the application of the marine isotope stratigraphy. Quaternary Science Reviews 24, 571–580. Tzedakis, P.C., Andrieu, V., de Beaulieu, J-L., Birks, H.J.B., Crowhurst, S., Follieri, M., Hooghiemstra, H., Magri, D., Reille, M., Sadori, L., Shackleton, N.J., Wijmstra, T.A., 1997. Comparison of terrestrial and marine records of changing climate of the last 500 000 years. Earth and Planetary Science Letters 14, 967–982. Tzedakis, P.C., Frogley, M.R., Heaton, T.H.E., 2002. Duration of last interglacial conditions in northwest Greece. Quaternary Research 28, 53–55. Tzedakis, P.C., McManus, J.F., Hooghiemstra, H., Oppo, D.W., Wijmstra, T.A., 2003. Comparison of changes in vegetation in northeast Greece with records of climate variability on orbital and suborbital frequencies over the last 450 000 years. Earth and Planetary Science Letters 212, 197–212. Tzedakis, P.C., Roucoux, K.H., de Abreu, L., Shackleton, N.J., 2004a. The duration of forest stages in southern Europe and interglacial climate variability. Science 306, 2231–2235. Tzedakis, P.C., Frogley, M.R., Lawson, I.T., Preece, R.C., Cacho, I., de Abreu, L., 2004b. Ecological

373

thresholds and patterns of millennial-scale climate variability: the response of vegetation in Greece during the last glacial period. Geology 32, 109–112. van der Knaap, W.O., van Leeuwen, J.F.N., 1995. Holocene vegetation succession and degradation responses to climatic change and human activity in the Serra da Estrela, Portugal. Review of Palaeobotany and Palynology 89, 153–211. van der Knaap, W.O., van Leeuwen, J.F.N., 1997. Late Glacial and early Holocene vegetation succession, altitudinal zonation, and climate change in the Serra da Estrela, Portugal. Review of Palaeobotany and Palynology 97, 239–285. Wijmstra, T.A., 1969. Palynology of the first 30 metres of a 120 m deep section in northern Greece. Acta Botanica Neerlandica, 18, 511–527. Wijmstra, T.A., Smit, A., 1976. Palynology of middle part (30–78 meters) of 120 deep section in northern Greece (Macedonia). Acta Botanica Neerlandica 25, 297–312. Wijmstra, T.A., Groenhart, M.C., 1983. Record of 700 000 years vegetational history in eastern Macedonia (Greece). Revista de la Academia Colombiana Ciencias Exactas, Fisicas y Naturales 15, 87–98. Willis, K.J., 1994. The vegetational history of the Balkans. Quaternary Science Reviews 13, 769–788. Zagwijn, W.H., 1989. Vegetation and climate during warmer intervals in the Late Pleistocene of western and Central Europe. Quaternary International 3/4, 57–67.

This page intentionally left blank

25. Climate Variability of the Last Five Isotopic Interglacials: Direct Land–Sea–Ice Correlation from the Multiproxy Analysis of North-Western Iberian Margin Deep-Sea Cores S. Desprat1, M. F. Sa´nchez Gon˜i1, F. Naughton2, J.-L. Turon2, J. Duprat2, B. Malaize´2, E. Cortijo3 and J.-P. Peypouquet1 1

EPHE, DGO, UMR-CNRS 5805 EPOC, Universite´ Bordeaux 1, Avenue des Faculte´s, 33405 Talence, France 2 De´partement de Ge´ologie et Oce´anographie, Universite´ Bordeaux 1, Avenue des Faculte´s, 33405 Talence, France 3 Laboratoire des Sciences du Climat et de l’Environnement, LSCE-Valle´e, Baˆt. 12, avenue de la Terrasse, F-91198 Gif-Sur-Yvette cedex, France

ABSTRACT

25.1 INTRODUCTION

The last five isotopic interglacials (marine isotope stages 11, 9, 7, 5 and 1) were investigated in Iberian margin deep-sea cores, using terrestrial (pollen) and marine (planktic foraminifera assemblages, benthic and planktic oxygen isotopes) climatic indicators. This work shows that the climatic variability detected on the continent is contemporaneously recorded in the ocean, but temperature changes are not in phase with ice volume variations. The comparison of the different marine isotope stages highlights a common pattern within these stages. They are characterized by three major climatic cycles, related to orbital cyclicity, on which suborbital climatic fluctuations are superimposed. Particularly, suborbital events interrupt the deglacial warming associated with Terminations IV to I and the second major warm period of each isotopic interglacial as well as the transitions towards glacial marine isotope stages. MIS 7 displays a short first warm period ( 8 kyr) followed by a striking cold and dry period succeeded by a new strong warmth. In contrast, MIS 11 presents the longest period ( 31 kyr) of the last 450 000 years.

Forecasting the future climatic evolution of the current interglacial is a great challenge. Before that, it is necessary to determine the evolution of past interglacials and evaluate the response of different components of the Earth’s climatic system. Owing to the Earth’s astronomical configuration, marine isotope stage (MIS) 11 is the best candidate to be the analogue of MIS 1. However, characterizing the climatic evolution over different situations of insolation forcing will permit better understanding of climate dynamics during interglacial periods. The continental palaeoclimatic records covering the last 425 000 years are rare and often fragmentary, and their chronologies are difficult to establish. This impedes the comparison of the climatic changes detected on land with those identified in the oceanic and cryospheric realms. We present, here, the first direct land– sea–ice correlation for the last five isotopic interglacials (MIS 11, 9, 7, 5 and 1). The main purpose of this work was to document the climatic variability of these periods and to assess the phase relationship between the responses of different Earth’s environments – continent, ocean and ice – to climatic changes in order to discern analogies and

376

S. Desprat et al.

differences between them. For that, a multiproxy study (pollen, assemblages of planktic foraminifera and 18 O of planktic and benthic foraminifera) was performed from several NW Iberian margin deep-sea cores. By comparing the last five isotopic interglacials, we will highlight, on the one hand, the similarities of their climatic dynamics despite different astronomical forcing and, on the other hand, the dissimilarities concerning duration, warmth magnitude and forest succession of the warm periods. 25.2 PRESENT-DAY ENVIRONMENTAL SETTING AND POLLEN SIGNAL IN THE IBERIAN MARGIN The Iberian margin deep-sea cores were retrieved 100 km off the Galician coast at 2100 m of water depth (Fig. 25.1). This site is at present under the influence of the North Atlantic deep water. The northwestern Iberian climate is considered

temperate and humid as a result of the influence of dominant Atlantic winds over the year. Mean annual temperature is 12.5 C (mean temperature of the coldest month, MTCO ¼ 512 C; mean temperature of the warmest month, MTWA ¼ 1722 C) and mean annual precipitation is between 1000 and 2000 mm:an1 (Atlas Nacional de Espan˜a, 1992). This region, incised in the north by the Rias Baixas valleys (Galician coast basin) and the Min˜o-Sil river (Sil basin) (Atlas Nacional de Espan˜a, 1993) and crossed in the south by the Douro river, belongs to the Eurosiberian and subMediterranean regions (Ozenda, 1982). At present, deciduous oak woodlands (Quercus robur, Q. pyrenaica and Q. petraea), heaths (Ericaceae including Calluna), brooms (Genista) and gorses (Ulex) dominate the vegetal cover of north-western Iberia (Alcara Ariza et al., 1987). Studies on present-day pollen deposition in marine sediments show that pollen grains reach marine sites from the adjacent continent by both fluvial and aeolian transport

Fig. 25.1 Location of the studied deep-sea cores MD01-2447, MD99-2331 and MD03-2697.

Climate Variability of the Last Five Isotopic Interglacials

and subsequent sinking through the water column (Chmura et al., 1999; Dupont and Wyputta, 2003; Heusser, 1978; Heusser and Balsam, 1977). Further, it is suggested that cores located near continental regions with well-developed hydrographic basins and prevailing offshore winds, as it is the case of our cores, mainly recruit pollen from rivers (Heusser, 1978; Turon, 1984; Dupont and Wyputta, 2003). The north-western Iberian rivers, mainly the Douro and Min˜o and to a lesser extent the Rias Baixas, provide sediments to the shelf area. On the shelf, the fine-particle sediments in suspension are transported northwards by poleward currents, and some are deposited in the Douro and Galicia Mud Patches (Dias et al., 2002). Extreme storm events can produce resuspension of some sediment from the mud patches, and transport of sediments off the shelf can occur (Jouanneau et al., 2002; Vitorino et al., 2002). Pollen grains belonging to the fine-particle fraction have a similar behaviour as fine sediments during the sedimentary processes (Muller, 1959; Chmura and Eisma, 1995). This suggests

377

that pollen grains preserved in our Iberian margin cores mainly come from the Galician and Douro drainage basins. The comparison of marine and continental modern pollen samples with the present-day Iberian vegetation shows that our marine pollen records represent an integrated image of the regional vegetation of the north-western part of the peninsula (Naughton et al., in press). 25.3 MATERIAL AND METHODS Records of pollen and classical climate indicators (planktic foraminafera assemblages, benthic and planktic 18 O) for the last five isotopic interglacials derive from three north-western Iberian margin deep-sea cores (MD01-2447, MD03-2697 and MD992331) (Fig. 25.1). All these cores were retrieved at the same coordinates on board of the oceanographic vessel Marion Dufresne, using the giant corer CALYPSO. As shown in Fig. 25.2, the intervals corresponding to MIS 11, 9 and 7 were studied in

Fig. 25.2 Lightness record of the three Iberian margin deep-sea cores. Pollen analysis has been performed in the intervals represented by grey areas. Hatched area corresponds to the disturbed interval in core MD01-2447.

378

S. Desprat et al.

core MD01-2447. The beginning of stage 9, being unfortunately disturbed in this core, was studied in the twin core MD03-2697. We present therefore a composite record of MIS 9 built from the correlation of several marine proxies analysed in the twin cores (lightness L*, CaCO3 content, percentages of N. pachyderma s. and coiling ratio of Globorotalia truncatulinoides) (Desprat, 2005). MIS 1 and 5 records come from the third core MD99-2331 (Naughton et al., in press; Sa´nchez Gon˜i et al., 2005). 25.3.1 Pollen analysis Each interval corresponding to MIS 11, 9, 7, 5 and 1 was subsampled for pollen analysis at 10- or 5-cm intervals. The sample preparation technique followed the procedure described by de Vernal et al. (1996) improved at the De´partement de Ge´ologie et Oce´anographie, University Bordeaux I (Desprat, 2005). After chemical and physical treatments (cold HCl, cold HF and sieving through 10-mm nylon mesh screens), the residue was mounted unstained in glycerol. Pollen was counted using a Zeiss Axioscope light microscope at 500 and 1250 (oil immersion) magnifications. A minimum of 100 pollen grains excluding the over-represented Pinus grains in marine deposits (Heusser and Balsam, 1977; Turon, 1984) were counted in each of the 327 samples analysed. The pollen percentages for each taxon are based on a main pollen sum that excludes Pinus, aquatic plants, pteridophyte spores and indeterminable pollen grains. Pinus percentages were calculated from the main sum plus Pinus. Spores and aquatic pollen percentages were obtained from the total sum (pollen þ spores þ indeterminablesþ unknowns). 25.3.2 Isotopic analyses The sampling resolution interval oscillates between 20 and 2 cm for Cibicides wuellerstorfi and Melonis barleeanus benthic

foraminifera and Globigerina bulloides planktic foraminifera. Each specimen has been picked from the 250315 mm grain-size fraction and cleaned with distilled water. The preparation of each aliquot (four to eight specimens, representing a mean weight of 80 mg) has been done using the Micromass Multiprep autosampler, using an individual acid attack for each sample. The CO2 gas extracted has been analysed against NBS 19 standard, taken as an international reference standard. The isotopic analyses of core MD01-2447 and MD992331 have been carried out at the De´partement de Ge´ologie et Oce´anographie (UMR 5805 EPOC, Bordeaux I University, France), using an Optima Micromass mass spectrometer, and those of core MD03-2697 were performed at the Laboratoire des Sciences du Climat et l’Environnement (Gif-sur-Yvette, France), using a delta plus Finnigan isotope mass spectrometer. All the isotopic results are presented versus PDB. The mean external reproducibility of powdered carbonate standards is 0:05‰ for oxygen. The 18 O values for Cibicides wuellerstorfi and Melonis barleeanus were adjusted by þ0:64 per mil and þ0:36 per mil, respectively, to account for species-dependent departure from isotopic equilibrium (Shackleton and Opdyke, 1973; Graham et al., 1981; Duplessy et al., 1984; Jansen et al., 1988). 25.3.2 Chronological framework The age model of the intervals corresponding to MIS 11 and 9 is based on the graphical correlation of the benthic 18 O curve with the low latitude stack of Bassinot et al. (1994) (Desprat, 2005; Desprat et al., 2005). The chronology of MIS 7 section also derived from a graphical correlation but in this case using the benthic-stack of Martinson et al. (1987) (Desprat et al., 2006). The chronologies of MIS 5 and 1 are, in contrast, independent of the astronomical calibration. That of the last isotopic interglacial (Sa´nchez Gon˜i et al., 2005) is based on the correlation of the major climatic phases

Climate Variability of the Last Five Isotopic Interglacials

detected in core MD99-2331 with those identified and dated in the southern Iberian margin core MD95-2042 using the MD952042 chronology of Shackleton et al. (2002). For the interval corresponding to the last 25 000 years, the age model was established using the chronology of the climatic episodes identified in other North Atlantic records and the ages assigned to several well-dated botanical events in the Iberian Peninsula.

25.4 THE CLIMATIC VARIABILITY OF THE LAST FIVE ISOTOPIC INTERGLACIALS IN AND OFF NW IBERIA 25.4.1 General climatic dynamic During the last five isotopic interglacials, the warm periods in north-western Iberia are characterized by the development of the temperate and humid forest, principally deciduous oak. In turn, open vegetation dominated by Poaceae and Asteraceae, with some semidesert plants, or mainly composed of heathland, expands during cold periods. The recorded vegetation changes indicate that climate has strongly oscillated during the previous isotopic interglacials. The climatic evolution detected on the continent parallels the oceanic changes reflected by marine proxies (planktic foraminifera 18 O and percentages). Indeed, each cold episode is marked by an increase of the percentages of the polar planktic foraminifera N. pachyderma s. and heavier planktic 18 O values. The record of planktic foraminifera assemblages is only available for MIS 11, 9 and 7. It shows that the tropical and summer subtropical species generally reach their maximal development during the warm periods detected on the continent. During each of the Terminations I to IV, an abrupt cold event interrupts the development of the temperate and humid forest associated with the deglacial warming: Younger Dryas, post-Zeifen stadial,

379

MD47-7-S1 and MD47-9-S1 (Fig. 25.3). These cold events have a clear imprint in the oceanic realm, in particular that of Termination IV. These coolings appear to be of different magnitude and during Terminations II, III and IV, they occur at the onset of minimum ice volume. For example, MD47-8-I1/MD47-7-S1 cycle shows the strongest amplitude of vegetation changes. The occurrence of such an episode during Termination V is still unknown because our sedimentary core does not cover the whole MIS12/MIS11 transition. As suggested by long European pollen sequences (Reille et al., 1998, 2000; Tzedakis et al., 2001), our direct land–sea–ice correlation confirms that on the continent each isotopic interglacial is characterized by three major warm periods associated with low ice volume, in response to the astronomical forcing (Table 25.1, Fig. 25.3). Indeed, the major forested periods in north-western Iberia are associated with low ice volume contrasting with the open vegetation phases related with ice cap development. Nevertheless, our direct land–sea–ice correlation puts forward that the ice volume changes are not synchronous with the temperature shifts on the continent and in the ocean. As observed by Sa´nchez Gon˜i et al. (1999, 2005), Shackleton et al. (2002) and Tzedakis et al. (2004), the limits of isotopic substages do not correspond to those of the climatic phases detected in western Iberia. For example, the Eemian in Iberia does not correspond to the entire MIS 5e (Sa´nchez Gon˜i et al., 1999, 2005; Shackleton et al., 2002). During the isotopic interglacials MIS 11, 9, 7 and 5, the first major warm periods (Vigo, Pontevedra, Arousa and Eemian) are marked by a more developed forest than the following ones. This is particularly true for MIS 11. Indeed, the climate optimum of each stage, detected by the maximal expansion of oak forest, Mediterranean plants and the maximal contraction of pine woodlands, occurs during these first major warm periods. These climate optima are contemporaneous to the ice volume minimum of each

380

S. Desprat et al.

Fig. 25.3 Direct correlation of continental and marine proxies from Iberian margin deep-sea cores. From the left to the right: (1) Synthetic pollen diagram; (2) Percentages of warm planktic foraminifera (only for MIS 7, 9 and 11) and Neogloboquadrina pachyderma left coiling; (3) Planktic d18 O curve; (4) Benthic d18 O curves. For the last 25 000 years, the benthic isotopic data of core MD99-2331 being not available, we present those of core MD01-2447. The correlation between both cores has been performed using different marine proxies (lightness, CaCO3 content, abundance and coiling ratio of Globorotalia hirsuta and G. truncatulinoides and percentages of N. pachyderma s.) (Desprat, 2005); (5) June insolation at 65 N. Blue areas indicate cold periods.

Climate Variability of the Last Five Isotopic Interglacials

Table 25.1 Major warm periods detected in north-western Iberia during the last 425 000 years versus marine isotopic stratigraphy Marine Isotope Stages

Isotopic events

Iberian major warm periods

MIS 1 MIS 5

1.1 5.1 5.3 5.5

Holocene St. Germain II St. Germain I Eemian

MIS 7

7.1 7.3 7.5

Rianxo Ribeira Arousa

MIS 9

9a 9c 9e 11.1 11.23 11.3

Bueu Sanxenxo Pontevedra Cangas Moana Vigo

MIS 11

The time lags between the boundaries of isotopic substages and stages and those of forested phases are not indicated.

stage. However, MIS 7 presents another particularity most likely related to especially strong insolation maximum: the second major warm period (Ribeira) is also marked by strong expansion of the temperate and humid forest, development of warm planktic foraminifera and important ice volume decrease. This implies that Ribeira would be, at least, as warm as the Arousa interglacial. Therefore, MIS 7 displays two climatic optima both associated with low ice volume as shown by the benthic isotopic record. The Vigo, Pontevedra and Eemian interglacials are followed by a strong cold period (MD47-11-S1, MD97-9-S2 and Me´lisey I). In contrast, MIS 7 includes a suborbital cycle (MD47-7-S2/MD47-7-I2) between the Arousa interglacial and the strong cold period MD47-7-S3. MD47-7-S3 was the coldest and driest period on the continent of the last five isotopic interglacials, as indicated by the highest values of grassland and semidesert taxa of our pollen record. This phase is also marked by the strongest decrease of sea-surface temperatures as shown by the maximum percentages of N. pachyderma left coiling. This cold episode is also coeval with

381

the largest ice volume increase (MIS 7.4) of the last isotopic interglacials, being similar to the glacial maximum of MIS 8. These unusual very cold conditions and huge icesheet enlargement within an isotopic interglacial are also recorded in northern North Atlantic ODP sites 983 and 980 (Channell et al., 1997; McManus et al., 1999). It is remarkable that this episode occurs during the most important insolation minimum of the last 450 000 years. Other suborbital events are superimposed on this orbital climatic variability. The Vigo interglacial is marked by two cool events circa 417 and 40 kyr, having a clear influence on vegetation and planktic foraminifera. More especially, the temperate and humid forest developments associated with the second major warm periods of MIS 11, 9 and 5 (Moana, Sanxenxo, St-Germain I) are all interrupted by a cold episode (MD47-11-S1, MD97-9-S2 and Montaigu, respectively). After our age model, these cold events are short, between 2 and 4 kyr. Moreover, during these episodes, the benthic 18 O values become heavier, in particular during MD97-9-S2. This indicates a significant increase of ice volume during the cold event within MIS 9c, as shown by Tzedakis et al. (2004). These cold fluctuations are also clearly recorded in the Velay sequence (Reille et al., 2000). In contrast, MIS 7 does not show such a climatic event. During the interglacial–glacial transitions MIS 7–MIS 6 and MIS 5–MIS 4, another climatic cycle of minor order is detected. The cooling associated with the MIS 9–MIS 8 transition seems also to be interrupted by a warm oscillation which needs to be confirmed by supplementary analysis. Nonetheless, all these warm events are clearly recorded in the oceanic realm by light values of planktic 18 O and an increase of warm planktic percentages. In sum, in spite of the different astronomical configuration of the last five isotopic interglacials, a common climatic evolution pattern emerges. However, each warm phase is characterized by different duration, amplitude and forest succession.

382

S. Desprat et al.

25.4.2 Warmth amplitudes of the last 450 000 years At present, the different amplitudes of warmth during the last 450 000 years are discussed. The results are often contradictory, depending on the regions and proxies concerned. Some works suggest that the warmest phase of MIS 11 shows the highest temperatures of the last 500 000 years (Howard, 1997; Droxler and Farrel, 2000; Berstad et al., 2002). However, this idea is challenged by many works (Bauch et al., 2000; Hodell et al., 2000; Kunz-Pirrung et al., 2002; McManus et al., 2003) which have shown that MIS 11 was not warmer than today. The deuterium signal of Vostok ice-core records the highest temperatures for MIS 9 (Petit et al., 1999), but the recent results of EPICA-Dome C ice core do not confirm this idea (EPICA Community Members, 2004). These new data also show higher temperatures in Antarctica during MIS 11 than during the Holocene (EPICA Community Members, 2004). As previously shown, in our record, each isotopic stage displays its climate optimum during the first warm period, excepting MIS 7 which presents a second optimum during Ribeira. The first optimum of this isotopic interglacial, Arousa, is marked by the highest percentages of temperate and humid trees. However, Mediterranean plants are scantily represented during this interval and only by evergreen Quercus which does not reveal clear Mediterranean conditions. In contrast, although the Vigo, Ribeira, Eemian and Holocene interglacials are characterized by lower percentages of temperate and humid forest than those during the Arousa interglacial, they record true Mediterranean species such as Pistacia, Olea or Cistus. For this reason, it remains difficult to determine which period is the warmest of the last five climatic cycles in northern Iberia. The warm planktic foraminifera record of MIS 11, 9 and 7 indicates the highest seasurface temperatures during MIS 11 climatic optimum. However, the development of these warm foraminifera is only slightly

stronger than during the other climatic optima. Therefore, the Vigo interglacial may be the warmest period of the last 450 000 years, but the difference of temperature does not appear large. Moreover, the benthic isotopic signal does not display weaker ice volume during Vigo interglacial than during Pontevedra, Ribeira, Eemian or Holocene interglacials. The genesis of such a warm interglacial during MIS 11 remains still a mystery in palaeoclimatology taking into account the weak insolation forcing. 25.4.3 Duration of the forest phases MIS 11 is marked by a long first major warm period, lasting 31 kyr after our age model. The Vigo interglacial appears two times longer than the Eemian ( 16 kyr), and at least three times longer than the Pontevedra and Arousa interglacials ( 11 and 8 kyr, respectively). The Holocene began 10 000 years ago, and it is already longer than the Arousa interglacial. On the basis of the pollen analysis of the south-western Iberian margin deep-sea core MD01-2443, Tzedakis et al. (2004) and Roucoux et al. (this volume) suggest a very short forest phase during MIS 9e, lasting 3600 years, after which Ericaceae expand. To bypass the difference in duration inferred from the age models, we have tuned our planktic 18 O record to that of MD01-2443. After this exercise, the resulting duration of the Pontevedra period is 13 000 years. It is possible that the cold/arid event responsible for the shortening of the first warm period of MIS 9 in south-western Iberia is not shown by our sequence due to a too low-resolution analysis. However, our results do not suggest that this abrupt climatic deterioration detected in southwestern Iberia brings the first MIS 9 forest period to an end in the north-western part of the Peninsula. In the same way, the Praclaux sequence (Massif Central, France) also records a long warm period even if it includes a slight forest reduction (Tzedakis et al., 2004).

Climate Variability of the Last Five Isotopic Interglacials

Our observation confirms, despite the uncertainties associated with the age scale of our record, that the warmest period of MIS 11 would be longer than those of the following isotopic interglacial stages, and so far three times longer than the Holocene. 25.4.4 Forest successions The warm periods of our record are marked by the development of pioneer trees, principally Betula, followed by the expansion of deciduous oak and hazel. This is in agreement with the classical vegetational succession during interglacial periods in northern Europe, as described by van der Hammen et al. (1971). This ideal succession sees at the latest stages the expansion of hornbeam, beech, fir, finishing with the development of the boreal forest with spruce. In the north-western Iberian region, the boreal forest phase is never reached during the warm periods. Moreover, the expansion of latecomer trees is different from one period to another: . Abies strongly developed only at the end of the Vigo interglacial and St. Germain Ic and it was only present at the end of the Eemian and Bueu (the last forested period of MIS 9). . Carpinus betulus expanded in the second part of all first major warm phases when deciduous Quercus decreased except during the Holocene. However, the hornbeam expansion was very strong during the Eemian, weaker during the Pontevedra and Arousa interglacials and very weak during the Vigo interglacial. During the other warm periods, hornbeam had also its maximal expansion after that of deciduous oak. . Fagus never had an important expansion at the end of the first major warm phases. Beech is only sporadically recorded at the end of the Vigo and Pontevedra interglacials. In contrast, during the periods Bueu, Ribeira and Rianxo, Fagus plays an important role in the vegetal cover of the

383

north-western Iberian region, always associated with Carpinus betulus. It is also noteworthy that beech developed rapidly and strongly at the beginning of the Ribeira interglacial (Desprat et al., 2006). The different behaviour of these three trees, depending on the periods and regions, has been previously noticed by Tzedakis and Bennett (1995) and Tzedakis et al. (2001). Disentangling the factors responsible for the settlement of tree species in a given region and period is a difficult task. The late expansion of some tree species can be linked: (a) to their migration rate in relation to their own dispersion mechanisms such as reproduction or seed scattering and with their competencies to develop on more or less mature soils, (b) to the distance to the glacial refugial zone, (c) to the interspecific competition, (d) to the individual response of each species to climatic change, and (e) also to the direct effect of the climate (Huntley and Webb, 1989; Huntley, 1996). Climate can also play an indirect role in changing the interspecific relationships (Lischke et al., 2002). Small differences in climatic conditions at the beginning of a warm phase can also influence the development of the late expanding trees (i.e. Carpinus, Abies and Fagus) (Tzedakis et al., 2001). Moreover, reduced diversity of taxa such as Fagus, Carpinus and Abies may imply that they are more susceptible to disease or adverse climatic conditions (Tzedakis et al., 2001). During a short warm period such as the Arousa interglacial, the virtual absence of beech and fir in north-western Iberia may be associated with a too little time for migrating from faraway refugial areas to the Iberian Peninsula, likely in relation to their own migration mechanisms. Nevertheless, the biotic processes cannot explain the very late arrival of Fagus in north-western Iberia, approximately 25 000 years after the beginning of the Vigo interglacial, since it developed only 7000 years after the beginning of the

384

S. Desprat et al.

Pontevedra interglacial. Therefore, the different timing and magnitude of the expansion of the late succession trees is somehow linked to the inherent climatic conditions of each warm phase of the different isotopic interglacials, which are related to ice-sheet extension and to the orbital parameters. 25.5 CONCLUSIONS This work constitutes a new step in documenting the climatic variability of interglacial isotopic stages. It provides the first direct land–sea–ice correlation of the last five isotopic interglacials (MIS 1, 5, 7, 9 and 11) from the multiproxy analysis of three pollen-rich cores from the northwestern Iberian margin. This record puts forward the phasing, previously identified during MIS 5, between changes in oceanic surface conditions and continental climate during the previous isotopic interglacials. Despite the differences of astronomical forcing, several similarities between these isotopic stages emerge: (a) the occurrence of three major climatic cycles, related to orbital cyclicity, (b) a climatic optimum during the first major warm periods, associated with ice volume minimum, (c) a suborbital cold event interrupting the second major warm period and (d) a suborbital climatic instability during the glacial–interglacial and interglacial–glacial transitions. The largest insolation oscillations controlling MIS 7 may explain the discrepancies between the climatic variability of this isotopic interglacial and the observed general scheme: a second major warm period at least as warm as the first one and preceded by a very cold and dry episode associated with an unusual important ice volume. Another striking feature of this stage is the very short first warm period, the Arousa interglacial, which is even shorter than the Holocene. In contrast, MIS 11 presents the longest warm period (Vigo interglacial) of

the last 450 000 years, three times longer than our present interglacial.

ACKNOWLEDGEMENTS Financial support was provided by IPEV and PNEDC French programs. We thank logistics and coring teams on board of the R/V Marion Dufresne II during the Ginna, Geosciences and Picabia oceanographic cruises and Marie-He´le`ne Castera, Karine Charlier, Olivier Ther and Franc¸oise Vinc¸on for invaluable technical assistance. This paper is Bordeaux 1 University, UMRCNRS 5805 EPOC Contribution n 1591.

REFERENCES Alcara Ariza, F., Asensi Marfil, A., de Bolos y Capdevilla, O., Costa Tales, M., Arco Aguilar, M., Diaz Gonzales, T.E., Diez Garretas, B., Fernandez Prieto, J.A., Fernandez Gonzales, F., Izco Sevillando, J., Loidi Arregui, J., Martinez Parras, J.M., Navarro Andres, F., Ninot I Sugranes, J.M., Peinado Lorca, M., Rivas Martinez, S., Sa´nchez Mata, D., Valle Guitierrez, C., Vigo I Bonada, J., Wildpret de la Torre, W., 1987. La vegetacio´n de Espan˜a. Collection Aula Abierta. Universidad de Alcala de Henares, 544 pp. Atlas Nacional de Espan˜a, C., 1992. Ministerio de Obras Publicas y Transportes, Direccion General del Instituto Geografico Nacional, Madrid. Atlas Nacional de Espan˜a, H., 1993. Ministerio de Obras Publicas, Transportes y Medio Ambiente, Direccion General del Instituto Geografico Nacional, Madrid. Bassinot, F.C., Labeyrie, J.D., Vincent, E., Quidelleur, X., Shackleton, N.J., Lancelot, Y., 1994. The astronomical theory of climate and the age of the Brunhes-Matuyama magnetic reversal. Earth and Planetary Science Letters 126, 91–108. Bauch, H.A., Erlenkeuser, H., Helmke, J.P., Struck, U., 2000. A paleoclimatic evaluation of marine oxygen isotope stage 11 in the high-northern Atlantic (Nordic seas). Global and Planetary Change 24, 27–39. Berstad, I.M., Lundberg, J., Lauritzen, S.E., Linge, H.C., 2002. Comparison of the climate during Marine Isotope Stage 9 and 11 inferred from a speleothem isotope record from Northern Norway. Quaternary Research 58, 361–371.

Climate Variability of the Last Five Isotopic Interglacials Channell, J.E.T., Hodell, D.A., Lehman, B., 1997. Relative geomagnetic paleointensity and 18 O at site ODP 983 (Gardar Drift, North Atlantic) since 350 ka. Earth and Planetary Science Letters 153, 103–118. Chmura, G.L., Eisma, D., 1995. A palynological study of surface and suspended sediments on a tidal flat: implications for pollen transport and deposition in coastal waters. Marine Geology 128, 183–200. Chmura, G.L., Smirnov, A., Campbell, I.D., 1999. Pollen transport through distributaries and depositional patterns in coastal waters. Palaeogeography, Palaeoclimatology, Palaeoecology 149, 257–270. de Vernal, A., Henry, M., Bilodeau, G., 1996. Techniques de pre´paration et d’analyse en micropale´ontologie. Les cahiers du GEOTOP 3, 16–27. Desprat, S., 2005. Re´ponses climatiques marines et continentales du Sud-Ouest de l’Europe lors des derniers interglaciaires et des entre´es en glaciations. The`se doctorale, Universite´ Bordeaux I, Talence, 272 pp. Desprat, S., Sa´nchez Gon˜i, M.F., Turon, J.-L., McManus, J.F., Loutre, M.F., Duprat, J., Malaize´, B., Peyron, O., Peypouquet, J.-P., 2005. Is vegetation responsible for glacial inception during periods of muted insolation changes? Quaternary Science Reviews 24, 1361–1374. Desprat, S., Sa´nchez Gon˜i, M.F., Turon, J.-L., Duprat, J., Malaize´, B., Peypouquet, J.-P., 2006. Climatic variability of Marine Isotope Stage 7: direct land–sea–ice correlation from a multiproxy analysis of a northwestern Iberian margin deep-sea core. Quaternary Science Reviews 25 (9–10), 1010–1026. Dias, J.M.A., Jouanneau, J.M., Gonzalez, R., Araujo, M.F., Drago, T., Garcia, C., Oliveira, A., Rodrigues, A., Vitorino, J., Weber, O., 2002. Present day sedimentary processes on the northern Iberian shelf. Progress in Oceanography 52, 249–259. Droxler, A.W., Farrell, J.W., 2000. Marine Isotope Stage (MIS 11): new insights for a warm future. Global and Planetary Change 24, 1–5. Duplessy, J.-C., Shackleton, N.J., Matthews, R.K., Prell, W., Ruddiman, W.F., Caralp, M., Hendy, C.H., 1984. 13C record of benthic Foraminifera in the last interglacial ocean: implication for carbon cycle and the global deep water circulation. Quaternary Research 21, 225–243. Dupont, L., Wyputta, U., 2003. Reconstructing pathways of aeolian pollen transport to the marine sediments along the coastline of SW Africa. Quaternary Science Reviews 22, 157–174. EPICA Community Members, 2004. Eight glacial cycles from an Antarctic ice core. Nature 429, 623–628.

385

Graham, D.W., Corliss, B.H., Bender, M.L., Keigwin, D., 1981. Carbon and oxygen disequilibria of recent deep-sea benthic foraminifera. Marine micropaleontology 6, 483–497. Heusser, L., 1978. Spores and pollen in the marine realm. In: Haq, B. U., Boersma, A. (Eds.), Introduction to marine micropaleontology. Elsevier, New York, pp. 327–339. Heusser, L.E., Balsam, W.L., 1977. Pollen distribution in the N. E. Pacific Ocean. Quaternary Research 7, 45–62. Hodell, D.A., Charles, C.D., Ninnemann, U.S., 2000. Comparison of interglacial stages in the South Atlantic sector of the southern ocean for the past 450 kyr: implications for Marine Isotope Stage (MIS) 11. Global and Planetary Change 24, 7–26. Howard, W.R., 1997. A warm future in the past. Nature 388, 418–419. Huntley, B., 1996. Quaternary paleoecology and ecology. Quaternary Science Reviews 15, 591–606. Huntley, B., Webb III, T., 1989. Migration: species’ response to climatic variations caused by changes in the Earth’s orbit. Journal of Biogeography 16, 5–19. Jansen, E., Bleil, U., Henrich, R., Kringstad, L., Slettemark, B., 1988. Paleoenvironmental changes in the Norwegian sea and the northeast Atlantic during the last 2.8 m.y.: deep sea drilling project/ocean drilling program sites 610,642, 643 and 644. Paleoceanography 3, 563–581. Jouanneau, J.M., Weber, O., Drago, T., Rodrigues, A., Oliveira, A., Dias, J.M.A., Garcia, C., Schimdt, S., Reyss, J.L., 2002. Recent sedimentation and sedimentary budget on the western Iberian shelf. Progress in Oceanography 52, 261–275. Kunz-Pirrung, M., Gersonde, R., Hodell, D.A., 2002. Mid-Brunhes century-scale diatom sea surface temperature and sea ice records from the Atlantic sector of the Southern Ocean (ODP Leg 177, sites 1093, 1094 and core PS2089-2). Palaeogeography, Palaeoclimatology, Palaeoecology 182, 305–328. Lischke, H., Lotter, A.F., Fischlin, A., 2002. Untangling a Holocene pollen record with forest model simulations and independent climate data. Ecological Modelling 150, 1–21. Martinson, D.G., Pisias, N.G., Hays, J.D., Imbrie, J., Moore, T.C., Shackleton, N.J., 1987. Age dating and orbital theory of the Ice Ages: Development of a high-resolution 0 to 300,000-year chronostratigraphy. Quaternary Research 27, 1–29. McManus, J.F., Oppo, D.W., Cullen, J.L., 1999. A 0.5-million-year record of millennial-scale climate variability in the North Atlantic. Science 283, 971–975. McManus, J., Oppo, D., Cullen, J., Healey, S., 2003. Marine isotope stage 11 (MIS 11): analog for Holocene and future climate. In: Droxler, A.W., Poore, R.Z., Burckle, L.H. (Eds.), Earth’s climate and orbital

386

S. Desprat et al.

eccentricity: the marine isotope stage 11 question. Geophysical Monograph. American Geophysical Union, Washington, DC, pp. 69–85. Muller, J., 1959. Palynology of recent Orinoco delta and shelf sediments. Micropaleontology 5, 1–32. Naughton, F., Sa´nchez Gon˜i, M.F., Desprat, S., Turon, J.-L., Duprat, J., Malaize´, B., Joly, C., Cortijo, E., Drago, T., Freitas, M.C., in press. Present-day and past (last 25000 years) Marine pollen signal off Western Iberia. Marine Micropaleontology. Ozenda, P., 1982. Les ve´ge´taux dans la biosphe`re. Doin, Paris, 431 pp. Petit, J.R., Jouzel, J., Raynaud, D., Barkov, N.I., Barnola, J.-M., Basile, I., Bender, M., Chappellaz, J., Davis, M., Delaygue, G., Delmotte, M., Kotyakov, V.M., Legrand, M., Lipenkov, V.Y., Lorius, C., Pe´pin, L., Ritz, C., Saltzman, E., Stievenard, M., 1999. Climate and atmospheric history of the past 420 000 years from the Vostok ice core, Antarctica. Nature 399, 429–436. Reille, M., Andrieu, V., de Beaulieu, J.-L., Guenet, P., Goeury, C., 1998. A long pollen record from Lac du Bouchet, Massif Central, France: for the period ca. 325 to 100 ka BP (OIS 9c to OIS 5e). Quaternary Science Reviews 17, 1107–1123. Reille, M., de Beaulieu, J.-L., Svobodova, V., AndrieuPonel, V., Goeury, C., 2000. Pollen analytical biostratigraphy of the last five climatic cycles from a long continental sequence from the Velay region (Massif Central, France). Journal of Quaternary Science 15, 665–685. Roucoux, K.H., Tzedakis, P.C., de Abreu, L., Shackleton, N.J. (this volume). Fine tuning the land–ocean correlation for the late Middle Pleistocene of southern Europe. The climate of the past interglacials. Sa´nchez Gon˜i, M.F., Eynaud, F., Turon, J.-L., Shackleton, N.J., 1999. High resolution palynological record off the Iberian margin: direct land-sea correlation for the Last Interglacial complex. Earth and Planetary Science Letters 171, 123–137. Sa´nchez Gon˜i, M.F., Loutre, M.F., Crucifix, M., Peyron, O., Santos, L., Duprat, J., Malaize´, B.,

Turon, J.-L., Peypouquet, J.-P., 2005. Increasing vegetation and climate gradient in western Europe over the Last Glacial Inception (122–110 ka): models-data comparison. Earth and Planetary Science Letters 231, 111–130. Shackleton, N.J., Opdyke, N.D., 1973. Oxygen isotope and paleomagnetic stratigraphy of Equatorial Pacific core V28-238: oxygen isotope temperatures and ice volumes on a 105 year and 106 year scale. Quaternary Research 3, 39–55. Shackleton, N.J., Chapman, M., Sa´nchez Gon˜i, M.F., Pailler, D., Lancelot, Y., 2002. The classic marine isotope substage 5e. Quaternary Research 58, 14–16. Turon, J.-L., 1984. Le palynoplancton dans l’environnement actuel de l’Atlantique nord-oriental. Evolution climatique et hydrologique depuis le dernier maximum glaciaire. Me´moires de l’Institut de Ge´ologie du bassin d’Aquitaine 17. Universite´ de Bordeaux I, Bordeaux, 313 pp. Tzedakis, P.C., Bennett, K.D., 1995. Interglacial vegetation succession: a view from southern Europe. Quaternary Science Reviews 14, 967–982. Tzedakis, P.C., Andrieu, V., de Beaulieu, J.-L., Birks, H.J.B., Crowhurst, S., Follieri, M., Hooghiemstra, H., Magri, D., Reille, M., Sadori, L., Shackleton, N.J., Wijmstra, T.A., 2001. Establishing a terrestrial chronological framework as a basis for biostratigraphical comparisons. Quaternary Science Reviews 20, 1583–1592. Tzedakis, P.C., Roucoux, K.H., de Abreu, L., Shackleton, N.J., 2004. The duration of forest stages in southern Europe and interglacial climate variability. Science 306, 2231–2235. van der Hammen, T., Wijmstra, T.A., Zagwijn, W.H., 1971. The floral record of the Late Cenozoic of Europe. In: Turekian, K.K. (Ed.), The Late Cenozoic glacial ages. Yale University Press, New Haven, pp. 391–424. Vitorino, J., Oliveira, A., Jouanneau, J.M., Drago, T., 2002. Winter dynamics on the northern Portuguese shelf. Part 2: bottom boundary layers and sediment dispersal. Progress in Oceanography 52, 155–170.

26. Palynological and Geochronological Study of the Holsteinian /Hoxnian/ Landos Interglacial Mebus A. Geyh1 and Helmut Mu¨ller2 1

Formerly Leibniz Research Institute of Geosciences, Box 510153, 30 631 Hannover, Germany Formerly Federal Institute of Geosciences and Resources, 30 631 Hannover, Germany

2

ABSTRACT The discovery of a pollen assemblage in interglacial deposits comparable with that of the pollen assemblage zone (PAZ) MM 8 of the Holsteinian Interglacial at Munster Breloh allows us to distinguish coeval deposits from those of other pre-Eemian interglacials. Thus, there is palynological evidence for a reliable correlation of precisely investigated and numerically dated Holsteinian deposits with the Mazovian Interglacial (eastern Poland), the Hoxnian Interglacial (England and SW Ireland) as well as the Landos Interglacial in the Massif Central (France). Adapting the palynological correlation of the Praclaux Interglacial, Massif Central, with MIS 11, the Holsteinian/Hoxnian/Mazonian/Landos interglacial belongs to MIS 9. 230 Th/U dates from fen peat layers of the Holsteinian reference site at Bossel in northern Germany and from two sites with Hoxnian deposits in England date these interglacial deposits to about 320 kyr BP corresponding to MIS 9. This correlation is supported by supplementary 230Th/U dates of fen peat from several both pre-Eemian and non-Holsteinian sites in Germany. 26.1 HOLSTEINIAN DEPOSITS AND THEIR CORRELATION WITH THE MIS TIMESCALE Quaternary geologists identified three glaciations (Elsterian, Saalian and Weichselian) in the Quaternary of the lowlands of

northern Germany (Keilhack, 1896) and four in the northern Alps (Penck and Bru¨ckner, 1901/1909). The youngest glacial periods were correlated with each other. The interglacials in between were termed Holsteinian and Eemian, respectively. This simple concept was first truly questioned when Emiliani (1955) introduced stable isotope determinations into Quaternary marine geology. He discovered that variation of the oxygen isotope composition (given as 18 O values) in planktonic foraminifera reflects global climatic changes between warm and cold periods. He termed them marine isotope stages (MIS) and related odd and even numbers to warm and cold periods, respectively. According to the number of isotopically identified cold periods in Antarctic ice cores, many more than four glaciations have occurred during the Quaternary, with eight cycles recognized during the last 740 kyr (EPICA Community Members, 2004). The numerical timescale for the last 400 kyr was determined from the 18 O record of the marine core V28-238 (Shackleton and Opdyke, 1973). By tuning this timescale with the modelled insolation cycles of the Sun, which are controlled by orbital forcing, the present SPECMAP numerical timescale was developed (Imbrie et al., 1984). Its precision may be better than 5000 yr (Hays et al., 1976). Since the mid-1950s, one of the major tasks of Quaternary geologists has been to explain the discrepant number of terrestrial and marine glaciations and interglacials and to attempt a correlation of the

Mebus A. Geyh and Helmut Mu¨ller

388

well-studied terrestrial periods with the corresponding marine isotope stages. The favourite way includes both the unequivocal palaeontological and sedimentological identification of the deposits as pertaining to a specific glacial or interglacial and their numerical dating. The pollen assemblages of the Scho¨ningen and Do¨mnitz thermomers (senso stricto Lu¨ttig, 1965) are examples which cannot be yet unequivocally identified as only parts of the whole Holsteinian pollen sequence are preserved. Attempts to use ‘count-from-the-top strategy’ of strata and to compare the results with the SPECMAP in most cases yield ambiguous results and do not solve the chronological problem in such cases. However, comprehensive lithological, sedimentological, geological and palynological studies in restricted regions may yield strong correlations. One of the best examples is the Quaternary geologic survey of Saxony, Saxony-Anhalt, South Brandenburg and Thuringia in eastern Germany by Eissmann (2002). He places the Holsteinian Interglacial in MIS 9. A second very good example is the comprehensive palynological survey in the Netherlands by Zagwijn (1989). He proposed an age of approximately 320 kyr for the Holsteinian Interglacial based on a comparison of the oxygen isotope curve of core V28–238 with the Quaternary climatic stratigraphy of the Netherlands. The stratigraphic terms of the interglacials and interstadials used in this paper are compiled in Table 26.1.

26.2 PALYNOLOGICAL STUDIES 26.2.1 Unequivocal identification of the Holsteinian Interglacial The first detailed palynological investigation of the Holsteinian deposits in Germany was done by Hallik (1960). Turner (1970) correlated them with the Hoxnian deposits in England. The Holsteinian pollen profiles in north Germany reflect a characteristic vegetation sequence (Hallik, 1960; Meyer, 1974; Mu¨ller, 1974; Mu¨ller and Ho¨fle, 1994), with a late glacial period preceding the interglacial. Meyer (1974) and Mu¨ller (1974) elaborated the pollen sequence of the Holsteinian Interglacial in more detail and added a varve counted absolute timescale. The most important discovery was, however, a short but noticeable vegetation decline about 6000 years after the beginning of the Holsteinian Interglacial. It became the palynological key element to identify the Holsteinian Interglacial across Europe (Geyh and Mu¨ller, 2005; Fig. 26.1). At Munster Breloh and Bossel immediately after the spread of Carpinus and Abies in the upper part of the pollen assemblage zone MM VII (MM – Meyer 1974; Mu¨ller, 1974) at the transition of MM VII to MM VIII, thermophilous trees such as Quercus, Carpinus, Abies and Corylus decreased; Taxus completely disappeared. At the same time, Poaceae, Betula and Pinus spread to a larger extent. In the upper part of PAZ MM VIII thermophilous trees such as Quercus, Alnus and particularly Corylus increased again.

Table 26.1 Stratigraphic terms of the interglacials and interstadials used in this paper and correlation with the SPECMAP timescale Germany Odderade Eemian Scho¨ningen Holsteinian Gelkenbach Rhumian

England

Hoxnian

France

Thermomer

SPECMAP

Landos Jagonas Praclaux

Interstadial Interglacial Interglacial (?) Interglacial Interstadials Interglacial

MIS 5a MIS 5e MIS 7 (?) MIS 9 MIS 11 MIS 11

Palynological and Geochronological Study 25.4

25.3

25.2

25.1

25.0

24.9

24.8

24.7

24.6

389 Depth (m)

24.5

440 kyrs

Sample interval 0 Sediment rate (mm/yr) 0.5 1.0 0 Ericales 5 0 Poaceae 5 0 Betula 5 10 0

Pinus

10 20 30 40 50 60 70 80 0 Picea 5 10 0

Taxus

0

Carpinus

0 Alnus 5 10 15 20 25 0 5 10 15 0 5 0 3 0 3 0 3 VIII b

VIII c

IX b

180 kyrs

IX a

160 kyrs

VIII a

VII c

100 kyrs

Corylus

Quercus Ulmus Tilia Fraxinus PAZ

Fig. 26.1 Highly resolved pollen assemblage zone MM VII (dark grey), VIII (black) and IX (light grey) of the Holsteinian reference site at Munster-Brelow (Mu¨ller, 1974) as palynological key element of Holsteinian deposits.

Slightly later other thermophilous trees such as Carpinus, Ulmus and Tilia spread. Taxus returned at the beginning of MM IX and reached the former level long before the spread of Abies. The time marks of the spread or disappearance of specific pollen in PAZ VIII were precisely determined by varve counting of the Munster Breloh profile (Mu¨ller, 1974; Geyh and Mu¨ller, 2005). Slight temporal deviations from this reference timescale

can be expected between the rim and the centre of a lake, between profiles situated at large distances (Poland, Germany, Ireland and France) and in deviating climatic conditions. This is why we did not give an absolute timescale in Fig. 26.1. The correlation of the pollen assemblage zones used in the mentioned regions is shown in Fig. 26.2. It has to be kept in mind that the temporal correlation may not be better than 100 years.

Mebus A. Geyh and Helmut Mu¨ller

390

Tentative timescale (yr) 15 000

10 000

5000

0

ca. 7000 years (Müller 1974)

2206 years (Meyer 1974)

counts of annual layers

ca. 3000 years (this paper) ca. 4500 years (Turner 1970)

a

Gn III b

Gn II

Gn I

Gortian, SW Ireland

e Gi I

Ho IV

Ho III

Ho II

Ho I

L Lo

Hoxnian, E England

b

a

b

a

c

b

a

MM XV

c

b

a

c b a

c

b

a

c b a

MM XIV b MM XIII a

MM XII

MM XI

MM X

MM IX

MM VIII

MM VII

MM VI

MM V

MM IV MM III

MM II

MM I

Holsteinian, NW Germany

Mazovian, E Poland e

d

c

b

a

d

c

b

a

c b a

Taxus LPAZ

Pinus Alnus LPAZ

Pinus LPAZ Betula

Carpinus-Abies LPAZ

Pinus-Larix LPAZ

PAZ 33

PAZ 32

PAZ 31

PAZ 30

PAZ 29

PAZ 28

Landos Interglacial Massif Central, France

Fig. 26.2 Correlation of the pollen assemblage zones used in the following regions: Gortian in SW Ireland (Dowling and Coxon, 2004), Hoxnian in England (Turner, 1970), Marks Tey in Essex (Turner, 1970), Holsteinian in NW Germany (Meyer, 1974; Mu¨ller, 1974; Mu¨ller and Ho¨fle, 1994), Mazovian in E Poland (Binka and Nitychoruk, 1995) and Landosian in the Velay area of the France Massif Central (Reille et al., 2000). The tentative time scale was elaborated by counts of varve-like annual layers (Turner, 1970; Meyer, 1974; Mu¨ller, 1974) and recent counts from diatomite cores between the beginnings of PAZ MM VIII and of PAZ MM XI, i.e. between the sudden decrease of Taxus and Carpinus.

Applying a high temporal resolution of several decades at the transition from MM VII to MM VIII, the Holsteinian Interglacial can be identified palynologically across Europe. Pre-supposition for such high temporal resolution is that (a) a sufficient large number of pollen grains per sample is counted, (b) the sampling intervals in the profile are very small, (c) the pollen grains are well preserved, (d) the local pollen input does not dominate over the regional one and (e) the sediments are stratigraphically undisturbed.

The following examples provide evidence that these particular palynological characteristics are a regional phenomenon: 1. In the pollen diagram of the Mazovian Interglacial in eastern Poland (Binka and Nitychoruk, 1995), approximately 900 km east of the Holsteinian-type section, this characteristic decline of thermophilous vegetation is found in spite of a dominant continental climate. After the disappearance of Taxus, Quercetum mixtum and

Palynological and Geochronological Study

Alnus declined simultaneously with the increase of Betula, Pinus and Poaceae, but long before the distinct spread of Abies. Some Carpinus was present. 2. In eastern England, West (1956) and Turner (1970) found evidence for a similar evolution of the vegetation in the late part of the Hoxnian assemblage zone Ho IIc. After an increase of thermophilous trees such as Quercus, Taxus and Corylus, Taxus and Corylus suddenly disappeared. At the same time Poaceae, Betula and Pinus spread. In the upper part of this PAZ Ho IIc thermophilous trees such as Quercus and particularly Corylus also increased. Taxus returned at the beginning of PAZ Ho IIIa and reached the former level long before the spread of Abies. 3. Even 1100 km W of the Holsteinian type region at the profile Gortian CB 2 in Cork Harbour, SW Ireland, the assemblage zone Gn IIIb (at 25 m depth) reflects a sudden drop of Taxus, a subsequent slight peak of Betula and Picea and a strong spread of Pinus. At about 24.5 m depth, Taxus regained a frequency as before its decline (Dowling and Coxon, 2001). 4. In spite of a low temporal resolution of about 200 years, the samples from the Landos Interglacial (32.60–32.56 m depth at PAZ 30) of the composite pollen diagram of the Massif Central in France (Reille et al., 2000) reflect (a) the long-lasting decline of Taxus, (b) a short peak of Poaceae, (c) a limited spread of Pinus to low levels, (d) a fast decline of Carpinus and (e) an increase of Corylus and deciduous trees. In the upper part of the following PAZ 31 (32.32–32.00 m), Taxus returned. Abies spread strikingly in PAZ 32. 26.2.2 Evidence for a miscorrelation between the Holsteinian and Praclaux Interglacial The most complete palynological record of the European Quaternary was elaborated by Reille et al. (2000) spanning the last 420 kyr.

391

The corresponding profiles are located about 1000 km SW of the Holsteinian-type section. Reille et al. (2000) elaborated pollen diagrams of three sites with a tephra layer as very reliable link. This tephra occurs in the sediments of the Amargiers Interstadial (the second interstadial above the Landos Interglacial) and has a 40Ar/39Ar age of 275 5 kyr. The palynological miscorrelation of the Praclaux Interglacial preceding the Landos Interglacial in the French Massif Central (Reille et al., 2000) with the Holsteinian Interglacial explains why many Quaternary geologists and palynologists link the Holsteinian Interglacial with MIS 11. The following reasons strongly suggest considering this correlation as a miscorrelation: 1. The main argument of Reille et al. (2000) in correlating the Holsteinian and the Praclaux interglacials is the presence of Pterocarya and Fagus pollen in the last two PAZ Ho IIIb and the older Ho IV of the Hoxnian Interglacial. However, the occurrence of a few Pterocarya and Fagus pollen grains does not allow a reliable correlation between any of these interglacials. Small quantities of these pollen grains were also found in both the Rhumian Interglacial (at Bilshausen and Ka¨rlich; see Table 26.1) and at a slightly higher level in the Holsteinian Interglacial. 2. The Praclaux and the succeeding Landos interglacials (Table 26.1) are the oldest interglacials in the pollen sequence of the French Massif Central. Both contain a high amount of Abies pollen. This pollen already is present in the older part of the Praclaux Interglacial (MIS 11), which is similar to the Rhumian Interglacial (Table 26.1) in Germany (MIS 11 or older). But, Abies did not appear before the last third of the Landos/Holsteinian Interglacial (MIS 9). Hence, the temporal spread of Abies in the different interglacials supports a correlation of the Praclaux Interglacial with the Rhumian

392

Mebus A. Geyh and Helmut Mu¨ller

Interglacial rather than with the Holsteinian Interglacial. Moreover, the Praclaux Interglacial does not have a low frequency of Picea in its upper part while the Landos Interglacial has. 3. The glacial period after the Rhumian interglacial was followed probably 10 000 years later by the two relatively short Gelkenbach interstadials (Chanda, 1962), which most probably are similar in age to the two Jagonas interstadials (Reille et al., 2000) after the Praclaux Interglacial. Betula and Pinus appeared due to the climatic melioration besides some thermophilous trees such as Carpinus, Ulmus and Quercus. 4. According to Reille et al. (2000), the Praclaux Interglacial is palynologically younger than the Rhumian Interglacial (Table 26.1) at Ka¨rlich, Germany (Bittmann and Mu¨ller, 1996). The latter contains a tephra layer which has a 40 Ar/39Ar age of 396 20 kyr (van den Bogaard et al., 1989). For the correlation of the Praclaux Interglacial with MIS 11, the 40Ar/39Ar age of the Ka¨rlich Interglacial was proclaimed as too young. Methodical or physical reasons are not given for this. 5. An important argument for our correlation of the Holsteinian/Hoxnian Interglacial with MIS 9 and the Rhumian/ Praclaux Interglacial with MIS 11 is based on the different duration of the warm periods. Varves were counted from the Rhumian Interglacial at Bilshausen (Mu¨ller, 1992; Bittmann and Mu¨ller, 1996) and annual diatom layers from the Holsteinian Interglacial (Meyer 1974; Mu¨ller, 1974). The Rhumian Interglacial covered a period of about 25 000 years extended by a late glacial period of 8400 years. The Holsteinian Interglacial yielded only 15 000 to 16 000 varves preceded by the short Elsterian Late Glacial (Mu¨ller and Ho¨fle, 1994). The different durations of these two interglacial periods support a correlation of the Rhumian and Holsteinian interglacials with MIS 11

and 9, respectively, rather than with MIS 13 and MIS 11, if compared with the timescale of the Vostok ice core (Petit et al., 1999) and that of Dome C (EPICA Community Members, 2004).

26.3 230TH/ U DATING OF HOXNIAN AND HOLSTEINIAN DEPOSITS 26.3.1 Numerical dates from the Hoxnian deposits Most reliable and comparable numerical dates of Holsteinian and Hoxnian interglacials were obtained by means of the 230Th/U method applied to fen peat. Rowe et al. (1997) analysed samples from the Tottenhill Quarry, Nar Valley, NW Norfolk, UK, which is correlated palynologically with the Hoxnian Interglacial (Ventris, 1996). This latter peat is palynologically correlated with Ho I-Ho IIb (West, 1956; Turner, 1970; Ventris, 1985) corresponding to MM 1–MM 7 (Meyer, 1974; Mu¨ller, 1974) in the type region of the Holsteinian Interglacial in the lower Elbe area, north Germany. Therefore, the transgression of the Holsteinian Sea occurred at Tottenhill at the same time as in the type area of the Holsteinian Interglacial near Bossel (Mu¨ller and Ho¨fle, 1994) and Hamburg-Hummelsbu¨ttel (Hallik, 1960). Rowe et al. (1997) obtained a mean 230 Th/U age of 317 14 kyr corresponding to MIS 9. The reliability checks were positive. Geyh and Mu¨ller (2005) dated Holsteinian fen peat deposits from the Bossel site. The boreholes GE 1/85 (29 139E, 30 209N; 5.6 m above sealevel) and GE 00/1 were drilled in 1985 and 2001, respectively, in Bossel about 30 km west of Hamburg. These cores contained two organic layers, one above and one below the sediments of the Holsteinian Sea (Fig. 26.3). The first location became the reference site for the Holsteinian Interglacial in northern Germany as declared by the European Commission on Stratigraphy of the Quaternary (Jerz and Linke,

Palynological and Geochronological Study GE 1/85

GE 00/1

17.90 m

Uh 20.00 m 18.60 m

20.55 m 19.60 m

20.10 m

40.85 m 38.20 m 41.17 m 38.50 m

42.29 m 40.00 m

2430 2432 2435 2436 306 2181 2182 2183

m 19.15 19.15 19.25 19.25 19.45? 19.45 19.45 19.45

Uh

m

2426 2427 2428 2434 2153 329 2154 2155 2177

38.35 38.35 38.35 38.35 38.40 38.40 38.40 38.40 38.40

kyr 365 ± 45 330 ± 35 Open system 390 ± 70 360 ± >100 Open system Open system Open system

Sand Silt, sand Clayey sand, silt Gyttja Fen peat Peat fragments

kyr 310 ± 4 345 ± 25 310 ± 7 300 ± 20 300 ± 20 325 ± 40 300 ± 20 370 ± 25 Open system

41.00 m

42.00 m

Fig. 26.3 Correlation of the Holsteinian Interglacial deposits in cores GE 1/85 and GE 00/1 from Bossel showing sampled layers and corresponding detrituscorrected 230Th/ U ages. Depth scales are given for each core.

1987). In the lower part of the cores, humic, sandy-clayey glacial lacustrine/ marine sediments are present representing a cool phase at the beginning of the melting of the Elsterian ice corresponding to the onset of the marine Holsteinian Interglacial. These deposits are overlain by 1 m of clayey-to-sandy silt and about 1.5 m calcareous gyttja (PAZ MM I to MM V after Meyer, 1974 and Mu¨ller, 1974). Towards the top between 38.5 and 38.2 m (all depths refer to Core GE 00/1), fen peat (Picea-Alnus-PinusBetula) of PAZ MM VI is present. This peat is covered by approximately 16.2 m of compressed marine silty clay with plant remains mainly of a Salicornia facies. The uppermost part of the core contains more silty-to-sandy layers with fragments of marine shells. At a depth of approximately 22 m, the sediment changes to humic silt and sand

393

containing organic fragments. A fen peat layer between 19.6 and 18.6 m belongs to the Quercus-Abies-(Pinus) (PAZ MM XIII) and the Pinus-Alnus-Betula-Pterocarya occurrence (PAZ MM XIV) of the late Holsteinian. The lower part of this layer is fen peat. Between 19.0 and 18.6 m sandy ombrogenic peat and reworked Holsteinian material are present. Eight detritus-corrected 230Th/U dates between 298 and 347 kyr ð<30 kyrÞ were determined from the two peat layers (Fig. 26.3). The mean is 323 4 kyr. The seven detritus-corrected 230Th/U age between 293 and 369 kyr from the lower peat layer yielded a more reliable mean of 312 3 kyr. The four corresponding dates of the upper peat layer yielded a mean of the detrituscorrected 230Th/U ages of 327þ130 37 kyr. The reliability check was successful as the corresponding analytically controlled minimum 230Th/U sample ages of >360 to >650 kyr (Geyh and Mu¨ller, 2005) exceed the calculated 230Th/U ages. Other numerical 230Th/U dates support the 230Th/U age of the Holsteinian deposits. Seven sites in Germany containing interglacial fen peat layers were palynologically classified as either older or younger than the Holsteinian Interglacial but older than the Eemian Interglacial. Twenty-eight 230 Th/U dates were determined and the ‘isochron’ approach was applied. All dates scatter apart from those of MIS 9. The dispersion histogram of these dates (Fig. 26.4; Geyh, 1980) plotted together with the 230 Th/U dates from the Holsteinian Interglacial form three distinct peaks. The 230 Th/U dates of those samples which are either older or younger than the Holsteinian Interglacial represent MIS 7 and MIS 11, the dates of the Bossel section represent MIS 9. 230 Th/U dating of ‘authigenic calcium carbonate with very low uranium and thorium concentrations that has been precipitated from the former lake water’ was applied to the reference section of the Hoxnian at Marks Tey, UK (Rowe et al., 1999).

Mebus A. Geyh and Helmut Mu¨ller

394 –3

Flensburg MIS 9

MIS 7

MIS 11

N

Probability density (yr –1)/δ 2H ‰

–2

Kiel Rostock SPECMAP 5

7

–1

Hamburg

1 - Rahden 2 - Köln-Porz 3 - Krefeld 4 - Göttingen-Ottostr. 5 - Bossel Hannover 6 - Nachtigall 7 - Zetel 8 - Korschenbroich 6 Göttingen Leipzig 4

Bremen

0

1 Münster

+1

3

8

Köln 2

0

100

200

300 km

+2 100

150

200

250

300

350

400 230Th/U

450

age (kyr)

Fig. 26.4 left Dispersion histogram of the 230Th/ U isochron dates from (i) the Holsteinian section at Bossel (Geyh, 1980; dark grey) and from (ii) seven sites in Germany (map on right) palynologically assigned to nonHolsteinian and non-Eemian interglacials (light grey). The dispersion histogram supports the link between the Holsteinian Interglacial and MIS 9. The SPECMAP curve is also shown (Imbrie et al., 1984)

This material palynologically represents the whole interglacial (Rowe et al., 1999). A probability interpretation ends up with the statement that the Hoxnian Interglacial ‘indicates with 87% confidence that the Marks Tey sediments correlate with MIS 11 or some older stage’. But, the calculated analytically controlled minimum 230Th/U sample ages (Geyh and Mu¨ller, 2005) range from >165 to >245 kyr only which are smaller than the published ages. Hence, the 230Th/U ages from Marks Tey only allow the statement that the numerical age is older than MIS 7; they do not exclude, however, a chronological correlation of the Holsteinian Interglacial with MIS 9. 26.3.2 Other dating attempts of Holsteinian deposits Miller and Mangerud (1985) applied aminoracemization analysis to Holsteinian mollusc shells collected on various European coasts between northern France and

Norway and concluded that the Holsteinian Interglacial may belong to MIS 7 or less probably to MIS 9. ESR dating yielded ambiguous numerical results. Holsteinian mollusc shells collected from a Holsteinian profile near Hamburg yielded dates between 195 and 220 kyr (Linke et al., 1985), supporting a correlation with MIS 7. Sarnthein et al. (1986) published ESR dates >370 kyr from the same site and correlated the Holsteinian Interglacial with MIS 11. The reason for the discrepancy of the ESR dates of the same site was the use of different evaluation protocols. ESR dates of 319 38 kyr of two teeth, collected from the base of Stratum C at Hoxne, first placed the corresponding deposits in MIS 9 (Gru¨n et al., 1988; Schwarcz and Gru¨n, 1993). These ESR dates were later revised and model-corrected to fit correlation with MIS 11 (Gru¨n and Schwarcz, 2000). In view of the large number of uncertain parameters used for ESR dating by Gru¨n and Schwarcz (2000), we feel that the correlation of the Hoxnian could well be re-revised back to MIS 9.

Palynological and Geochronological Study

Misleading correlations are obtained with the ‘count-from-the-top strategy’ though referring to numerical dates of base or top layers. Examples are found by Urban (1995). She correlates the Holsteinian Interglacial with MIS 11 without palynological evidence and only based on two 230Th/U dates of peat of the Scho¨ningen Interglacial (Table 26.1).

26.4 CONCLUSION Both the Holsteinian and the Hoxnian interglacials are correlated with MIS 9 based on 230Th/U dates from the stratigraphic reference site of the Holsteinian Interglacial at Bossel in Germany and from the Tottenhill Quarry and Marks Tey sites of the Hoxnian Interglacial in England, respectively. This correlation is further supported by numerous 230Th/U dates from fen peat of non-Holsteinian and nonEemian interglacial sites in Germany. There is palynological evidence from high-resolution pollen diagrams in Poland, Germany, E England and SW Ireland that the Holsteinian Interglacial as well as the Hoxnian and Landos interglacials are correlated with MIS 9, and the Praclaux Interglacial with MIS 11.

ACKNOWLEDGEMENTS We thank Mrs. Gudrun Drewes, Sabine Mogwitz and Dipl.-Phys. Deniz Oezen for the analyses and measurements of the Bossel samples and the field campaigns. T. Litt, Bonn, contributed references. P. J. Rowe and D. H. Keen read the paper critically and made many valuable suggestions which substantially improved the manuscript. Our sincere thanks to Henry Toms, geologist and English translator, for his thorough and critical reading of our paper, corrections.

395

REFERENCES Binka, K., Nitychoruk, J., 1995. Mazovian (Holsteinian) lake sediments at Woskrzenice near Biala Podlaska. Geological Quarterly 39, 109–120. Bittmann, F., Mu¨ller, H., 1996. The Ka¨rlich Interglacial site and its correlation with the Bilshausen sequence. In: Turner, C. (Ed.), The Early Middle Pleistocene in Europe. Balkema, Rotterdam, pp. 187–193. Chanda, S., 1962. Untersuchungen zur plioza¨nen und pleistoza¨nen Floren- und Vegetationsgeschichte im Leinetal und im su¨dwestlichen Harzvorland (Untereichsfeld). Geologisches Jahrbuch 79, 783–844. Dowling, L.A., Coxon, P., 2001. Current understanding of Pleistocene temperate stages in Ireland. Quaternary Science Reviews 20, 1631–1642. Eißmann, L., 2002. Quaternary geology of eastern Germany (Saxony, Saxon-Anhalt, South Brandenburg, Thu¨ringia), type area of the Elsterian and Saalian stages in Europe. Quaternary Science Reviews 21, 1275–1346. Emiliani, C., 1955. Pleistocene temperatures. Journal of Geology 63, 538–573. EPICA Community Members, 2004. Eight glacial cycles from an Antarctic ice core. Nature 429, 623–628. Geyh, M.A., 1980. Holocene sealevel history: case study of the statistical evaluation of 14C dates. Radiocarbon 22, 695–704. Geyh, M.A., Mu¨ller, H., 2005. Numerical 230Th/U dating and a palynological review of the Holsteinian/Hoxnian Interglacial. Quaternary Science Reviews 24, 1861–1872. Gru¨n, R., Schwarcz, H.P., 2000. Revised open system U-series/ESR age calculations for teeth from Stratum C at the Hoxnian Interglacial type locality, England. Quaternary Science Reviews (Quaternary Geochronology) 19, 1151–1154. Gru¨n, R., Schwarcz, P., Chadam, J., 1988. ESR dating of tooth enamel: coupled correction for U-uptake and U-series disequilibrium. Nuclear Tracks and Radiation Measurements 14, 237–241. Hallik, R., 1960. Die Vegetationsentwicklung der Holstein-Warmzeit in Nordwestdeutschland und die Altersstellung der Kieselgurlager der su¨dlichen Lu¨neburger Heide. Zeitschrift der deutschen geologischen Gesellschaft 122 (2), 326–333. Hays, J.D., Imbrie, J., Shackleton, N.J., 1976. Variations in the earth’s orbit: pacemaker of the ice ages. Science 194, 1121–1132. Imbrie, J., Hays, J.D., Martinson, D.G., McIntyre, A., Mix, A.C., Morley, J.J., Pisias, N.G., Prell, W.L., Shackleton, N.J., 1984. The orbital theory of Pleistocene climate: Support from a revised chronology

396

Mebus A. Geyh and Helmut Mu¨ller

of the marine 18 O record. In: Berger, A.L., Imbrie, I, Hays, J., Kukla, G. (Eds.), Milankovitch and Climate. Reidel, Dordrecht, pp. 269–305. Jerz, H., Linke, G., 1987. Arbeitsergebnisse der Subkommission fu¨r Europa¨ische Quarta¨rstratigraphie. Typusregion des Holstein-Interglazials (Berichte der SEQS 8). Eiszeitalter und Gegenwart 37, 145–148. Keilhack, K., 1896. Die Geikiessche Gliederung der nordeuropa¨ischen Glacialablagerungen. Jahrbuch der Ko¨niglich Preußischen Geologischen Landesanstalten fu¨r 1895, 111–124. Linke, G., Katzenberger, O., Gru¨n, R., 1985. Description and ESR dating of Holsteinian interglaciation. Quaternary Science Reviews 4, 319–331. Lu¨ttig, G., 1965. Interglacial and interstadial periods. Journal of Geology 73, 579–591. Meyer, K.-J., 1974. Pollenanalytische Untersuchungen und Jahresschichtza¨hlungen an der holsteinzeitlichen Kieselgur von Hetendorf. Geologisches Jahrbuch A 21, 87–105. Miller, G.H., Mangerud, J., 1985. Aminostratigraphy of European marine interglacial deposits. Quaternary Science Reviews 4, 215–278. Mu¨ller, H., 1974. Pollenanalytische Untersuchungen und Jahresschichtenza¨hlungen an der holsteinzeitlichen Kieselgur von Munster-Breloh. Geologisches Jahrbuch A 21, 107–140. Mu¨ller, H., 1992. Climatic changes during and at the end of the interglacials of the Cromerian Complex. In: Kukla, G.J., Went, E. (Eds.), Start of the Glacial. Nato ASI Series 13, Springer, Berlin, Heidelberg, pp. 55–69. Mu¨ller, H., Ho¨fle, H.-Ch. 1994. Das HolsteinInterglazialvorkommen bei Bossel westlich von Stade und Wanho¨den no¨rdlich Bremerhaven. Geologisches Jahrbuch A 134, 71–116. Penck, A., Bru¨ckner, E., 1901/1909. Die Alpen im Eiszeitalter, Band I–III; Leipzig. Petit, J.R., Jouzel, J., Raynaud, D., Barkov, N.I., Barnola, J.M., Basile, I., Bender, M., Chappellaz, J., Davis, J., Delaygue, G., Delmotte, M., Kotlyakov, V.M., Legrand, M., Lipenkov, V., Lorius, C., Pe´pin, L., Ritz, C., Saltzman, E., Stievenard, M., 1999. Climate and atmospheric history of the past 420,000 years from the Vostok Ice Core, Antarctica. Nature 399, 429–436. Reille, M., de Beaulieu, J.L., Svoboda, H., AndrievPonel, V., Goeudry, C., 2000. Pollen analytical biostratigraphy of the last five climatic cycles from a long continental sequence from the Velay Region (Massif Central, France). Journal of Quaternary Science 15, 665–685.

Rowe, P.J., Richards, D.A., Atkinson, T.C., Bottrell, S.H., Cliff, R.A., 1997. Geochemistry and radiometric dating of a Middle Pleistocene peat. Geochimica et Cosmochimica Acta 61, 4201–4211. Rowe, P.J., Atkinson, T.C., Turner, C., 1999. U-series dating of Hoxnian interglacial deposits at Marks Tey, Essex, England. Journal of Quaternary Science 14, 693–702. Sarnthein, M., Stremme, H.E., Mangini, A., 1986. The Holsteinian Interglaciation: time-stratigraphic position and correlation to stable isotope stratigraphy of deep-sea sediments. Quaternary Research 26, 283–298. Schwarcz, H.P., Gru¨n, R., 1993. ESR dating of the lower industry. In: Singer, R., Gladfelder, B.G., Wymer, J.J. (Eds.), The Lower Paleolithic Site at Hoxne, England. University of Chicago Press, Chicago, pp. 203–205. Shackleton, N.J., Opdyke, N.D., 1973. Oxygen isotope and palaeomagnetic stratigraphy of equatorial Pacific core V28–238: oxygen isotope temperatures and ice volumes on a 105 year and 106 year scale. Quaternary Research 3, 39–55. Turner, C., 1970. The Middle Pleistocene deposits at Marks Tey, Essex. Philosophical Transactions of the Royal Society of London, Series B 257, 373–440. Urban, B., 1995. Palynological evidence of younger Middle Pleistocene interglacials (Holsteinian, Reinsdorf and Scho¨ningen) in the Scho¨ningen open cast lignite mine (eastern Lower Saxony, Germany. Mededelingen Rijks Geologische Dienst 52, 175–185. van den Bogaard, C., van den Bogaard, P., Schmincke, H.-U., 1989. Quarta¨rgeologischtephrostratigraphische Neuaufnahme und Interpretation des Pleistoza¨nprofils Ka¨rlich. Eiszeitalter und Gegenwart 39, 62–89. Ventris, P.A., 1985. Pleistocene environmental history of the Nar Valley, Norfolk. Ph.D. thesis, University of Cambridge. Ventris, P., 1996. Hoxnian interglacial freshwater and marine deposits in northwest Norfolk, England and their implications for sea level reconstruction. Quaternary Science Reviews 15, 437–450. West, R.G., 1956. The Quaternary deposits of Hoxne, Suffolk. Philosophical Transactions of the Royal Society of London, Series B 239, 265–356. Zagwijn, W.H., 1989. The Netherlands during the Tertiary and the Quaternary: a case history of coastal lowland evolution. Geologie en Mijnbouw 68, 107–120.

27. A New Holsteinian Pollen Record from the Dry Maar at Do¨ttingen (Eifel) Markus Diehl and Frank Sirocko Institute for Geosciences, University of Mainz, Becherweg 21, 55099 Mainz, Germany

ABSTRACT A new interglacial pollen sequence from the Do¨ttingen dry maar in the Eifel region of the Rheinish Schield is presented. Palynology is used to correlate with several classical north German Holsteinian sites. The lake sediments reveal the complete interglacial and also 60 m of laminated sediments from the glacial preceding the Holsteinian. The interglacial section indicates limnic conditions in its lower part and telmatic conditions in its upper part with an intermediate episode of peat formation. Ash layers document intensive volcanism during the interglacial in the Eifel region. Some of the north German Holsteinian sites reveal spikes of high abundance of Pinus, Betula and Poaceae and/or setbacks of more demanding taxa during the interglacial, often interpreted as cold events. The Do¨ttingen profile shows similar pattern, but with little response from the thermophilous pollen taxa. In the Do¨ttingen sequence, these vegetation ‘anomalies’ are preceded, or accompanied by phases of active volcanism. The role/interaction of climate and/or volcanism as a likely cause for these vegetation ‘anomalies’ is still to be quantified. 27.1 INTRODUCTION In the last decades, numerous pollen sequences assigned to the Holsteinian were investigated in Germany (see e.g. Gistl, 1928; Selle, 1954; Hallik, 1960; Majewski, 1961; Erd, 1969, 1970, 1987; Meyer, 1974; Mu¨ller, 1974b; Cepek and Erd, 1975; Erd and Mu¨ller, 1977; Dassow,

1987; Averdieck, 1992; Linke and Hallik, 1993; Mu¨ller and Ho¨fle, 1994; Eissmann et al., 1995; for the most detailed and/or most complete profiles). All these sites are located in the lowlands of north and eastern Germany on Elsterian glaciofluvial deposits. The Holsteinian vegetation history is thus well documented for northern Germany, but no Holsteinian profiles are known up to now from more southern localities and from greater altitudes with other climatic and edaphic conditions. Based on palynological investigations of Usinger (unpublished data), Stachel and Bu¨chel (1989) were the first to assign a Holsteinian age to interglacial sediments within the Do¨ttingen dry maar structure. The Do¨ttingen dry maar was thus cored in spring 2000 by the ELSA project (Eifel laminated sediment archive) at three locations (Figs. 27.1 and 27.2). Maar lakes provide deep sediment traps with anoxic bottom waters and yield perfect conditions for undisturbed conservation of organic remains like pollen grains and act as high-resolution archives for vegetation history. The Holocene and Eemian vegetation in the Eifel was already reconstructed from maar lakes (Litt, 1999; Sirocko et al., 2005), and the Do¨ttingen sequence is now the third interglacial to be studied in these sedimentary basins. 27.2 LOCATION The Eifel is part of the Variscan orogenic belt of central Europe. Rising magmas have produced a total of about 360 Quaternary eruption centres (Bu¨chel et al., 2000; Lorenz and

398

Markus Diehl and Frank Sirocko

N R

Dümpelmaar

he

Adenau Rieden Ormont

in

Wehr

Mayen

Döttingen dry maar Ulmen Gillenfeld

0

180–270 Koblenz

270–360

el

360–450

os

Daun

90–180

450–540

North Sea

M

Gerolstein

0–90

Neuwied

Laacher See

540–630 630–720

Cochem

>720 [m a.s.l.]

5 10 km

(Geobasisinformationen (DGM5) © Landesamt für Vermessung und Geobasisinformation Rheinland - Pfalz vom 28.06.2005, Az.: 26 722-1.401)

Quaternary volcanic activity centres of the Eifel: ‘Rammkernsondierung’ (up to 10 m depth) Investigated dry maars ‘Seilkernbohrung’ (up to 155 m depth) Maar lakes (dry maar) Phonolite eruptive centres (maar lake)

Fig. 27.1 Quaternary volcanoes of the West and East Eifel volcanic fields (after Schmincke, 2000, and van den Bogaard and Schmincke, 1990) with cores of the ELSA project (Eifel laminated sediment archive) and the Do¨ttingen dry maar.

Gravimetric Diatreme outcrop Magnetic Quaternary extrusives Tertiary extrusives Tertiary extrusives (assumed)

Fig. 27.2 Topographic map of dry maar Do¨ttingen (modified after Stachel and Bu¨chel, 1989) with magnetic and gravimetric outcrops of the diatreme and positions of the drillings.

A New Holsteinian Pollen Record

Zimanowski, 2000) with about 70 maar structures (Bu¨chel and Lorenz, 1982), most of them preserved as dry maar lakes today. The Do¨ttingen dry maar is located near the village Herresbach-Do¨ttingen (50.35 N, 7.12 E), southeast of the Nu¨rburgring within the Hocheifel Tertiary volcanic field. Geomagnetic and gravimetric survey has revealed its phreatomagmatic origin (Bu¨chel, 1984). The depression of the dry maar is today about 1–1.5 km in diameter and ca. 50 m in depth at around 500 m above sealevel. It is surrounded by several dikes and by the scoria cone Niveligsberg, which together with the tuff dike of Herresbach, represents the northeastern eruption centres of the Quaternary West Eifel volcanic field (Pirrung, 1998) (Figs. 27.1 and 27.2). However, Do¨ttingen is only 12–15 km away from the large Riedener Kessel and the East Eifel volcanic field (Fig. 27.1). Three drillings (P1/2, EB1) were undertaken by the ‘Wasserversorgungszweckverband’ of Maifeld-Mayen in 1993 (Fig. 27.2). Pumice layers within cold stage sediments, about 8 m below interglacial gyttjas, were assumed to correlate with tuffs from the Riedener Kessel eruption (Pirrung, 1998) dated to 380–430 kyr (van den Bogaard and Schmincke, 1990). Gyttjas and peat, uncovered by a trench (SB in Fig. 27.2), on top of the cold stage sediments, were already attributed to the Holsteinian (Stachel and Bu¨chel, 1989). Pirrung (1998) attributed the basal cold stage maar sediments to the Elsterian. 27.3 MATERIALS ¨ 1-3) were drilled in the Three cores (DO ¨ 1: centre of the Do¨ttingen dry maar (DO 25 55 25 ¨ 2 and DO ¨ 3: 71664/ 71682/ 79728, DO 55 79622, 460 m a.s.l.) (Fig. 27.2). Core ¨ 1 (Fig. 27.3) reached a depth of DO 89.5 m likely into the tephra of the original eruption, but the interglacial sediments on the top of the core were partly ¨ 2 and DO ¨ 3 (Fig. 27.4) reached eroded. DO 14.5 and 16.75 m depth, respectively, and

399

completely covered the interglacial deposits. All palynological analyses were ¨ 3. DO ¨ 3 shows greyishdone on core DO greenish silts from 16.75 to 13.75 m, calcareous in the uppermost 30 cm. The silts are overlain by a diatom-bearing interglacial gyttja with indistinct lamination. A fine-grained, bright, pumice-bearing ash layer occurs in about 13.21–13.17 m depth. A second ash layer lies in 12.90–12.77 m depth. The gyttja develops into a peat layer in 10.35–10.24 m depth, intercalated with a pumice- and bedrock-bearing ash (ca. 10.50–10.17 m depth), which is again overlain by a second detritus-rich diatom gyttja of bulky texture. Between 7.45 and 7.05 m, the composition is silt and between 7.05 and 6.25 m crumby mud, with more silt at 6.52–6.38 m depth. Above 6.25 m depth, greyish-greenish silts (in their upper parts with reddish striae) occur again with a 2-cm thick gyttja layer at about 6.05 m depth. Above 4.90 m, we find colluvium. The ash layers show idiomorphic, not weathered phenocrysts, pumice grains and nonrounded bedrock clasts. Therefore, these sediments are not, or only slightly, reworked. The occurrence of bedrock, a wide grain-size spectrum and broken phenocrysts indicate the phreatomagmatic character of these eruptions. Several thin ash stripes become only clearly visible in thin sections and occur at 13.86–13.82 m, 13.75–13.73 m, 13.075–13.055 m, around 12.30 m, at 11.995–11.985 m, 11.175–11.165 m, 9.79–9.78 m and 6.960–6.945 m depth (Figs. 27.5 and 27.6). Some of these ash stripes bear idiomorphic leucite, indicating again little or no reworking. All ash layers together document intensive volcanism during the interglacial.

27.4 METHODS Pollen preparation followed the techniques of Berglund and Ralska-Jasiewiczowa (1986) and Faegri and Iversen (1989),

400

Markus Diehl and Frank Sirocko

¨ 1 (depth scale: [m]). Interglacial sediments are from 3 to 7 m depth and truncated at Fig. 27.3 Photos of core DO their top. From 7 m depth downwards, there are glacial silts with pyroclastics, the basal volcanoclastics between 75 and 89 m depth are likely from the initial maar eruption.

A New Holsteinian Pollen Record

401

¨ 3 (depth scale: [m]) (rew.: reworked, d.-r.: detritus-rich). Fig. 27.4 Photos and lithology of core DO

respectively, modified after Litt (University of Bonn, Germany). Each pollen sample spanned a depth range of 1 cm and was of about 1 cm3 volume. The sediments were treated with potassium hydroxide solution (KOH), hydrochloric acid (HCl) and hydrofluoric acid (HF). For acetolysis, acetic acid (C2H4O2) and a mixture (9:1) of acetic anhydride (C4H6O3) and sulphuric acid (H2SO4) were used. The samples were archived with liquid anhydrous glycerol (C3H8O3). Centrifugation was done at

3000–3500 r.p.m. for 5 minutes. The samples were sieved through a 200-mm sifter and later filtrated through a 10-mm filter. Lycopodium-spore tablets (Department of Geology, Quaternary Sciences, So¨lvegatan 12, SE-223 62 LUND, Sweden) were added for calibration of absolute pollen concentration per ccm (Stockmarr, 1971). Pollen counting was done under maximum 600fold magnification. The pollen spectra were calculated using the sum of all terrestrial pollen as 100%. Hippophae¨,

402 Markus Diehl and Frank Sirocko

¨ 3 from the Do¨ttingen dry maar (not all counted Fig. 27.5 Pollen percentage diagrams with local pollen assemblage zones (LPAZ) and lithology of core DO nonarboreal pollen and no algae are displayed). Grey shaded curves are amplified tenfold. Nymphaea and Nuphar were subsumed to Nymphaeaceae. All taxa are in percentages of terrestrial pollen.

A New Holsteinian Pollen Record

Fig. 27.5 Continued

403

404

Markus Diehl and Frank Sirocko

Juniperus, Corylus, Hedera, Ilex, Viscum and Vitis were counted as arboreal pollen. Cyperaceae were treated as aquatic pollen. For pollen identification, literature (Erdtman, 1969; Faegri and Iversen, 1989; Moore et al., 1991; Beug, 2004) and a reference collection of pollen samples from living plants were used. 27.5 RESULTS The Do¨ttingen pollen assemblages can be subdivided into seven main local pollen assemblage zones (LPAZ) and sixteen subzones (Table 27.1) (Figs. 27.5 and 27.6). The pollen zones and boundaries with their palynological characteristics were defined on the pollen percentage diagram (Fig. 27.5). The LPAZ were then applied to the pollen concentration diagram of Fig. 27.6. In LPAZ 2, Picea and the mixed oak forest members Tilia and Acer occur contemporaneously, but clearly time delayed to Quercus, Ulmus and Fraxinus (Figs. 27.5 and 27.6). In LPAZ 7a, a phase of low arboreal pollen percentages with increased percentages of, for example, Poaceae, Ericaceae, Artemisia and Thalictrum occurs (Fig. 27.5). Therefore, these two pollen zones may be further subdivided. The defined pollen zones partly coincide with lithologic changes (Fig. 27.5). LPAZ 1 coincides with the calcareous part of the silts, reflecting carbonate precipitation during the warming of the lake at the interglacial onset. Pollen zones 2–6a consist of diatom-gyttjas which are indistinctly laminated within LPAZ 2–4 and detritus rich and of bulky texture from LPAZ 5b to LPAZ 6a, indicating deposition in a shallow lake. The LPAZ boundary 3/4 is preceded by an ash (partly mixed with gyttja remains) in 12.90–12.77 m depth. A leucite-bearing ash stripe of about 1–2 cm thickness occurs in LPAZ 4c. The LPAZ 5a/b boundary lies within the described peat layer at 10.35–10.24 m depth. The peat is enriched with conifer pollen grains and fern (Polypodiales) spores,

which may be the result of selective decomposition of less resistant pollen types (Straka, 1957, 1975). LPAZ 6b coincides with silts, whereas LPAZ 7a is conspicuously enriched in crumby mud and LPAZ 7b again coincides with silts. The thin gyttja layer in this LPAZ at 6.05 m depth is obviously reworked. Therefore, the remains of thermophilous taxa within this zone most likely originate from reworked interglacial sediments. This seems to be also the case for LPAZ 7a, which then can be interpreted as a boreal phase at the interglacial/glacial transition. A boreal phase between the afforesting period of LPAZ 1a and the LPAZ 2 with already thermophilous taxa could not be demarcated. Only if the pollen grains of Quercus (and Corylus) in LPAZ 1b are seen as allochthonic, this LPAZ could be interpreted as a boreal phase towards the temperate LPAZ 2. In this case, the Quercus (and Corylus) pollen may be reworked or originate from farther refuges through long-distance transport. The peat formation within LPAZ 5 is interpreted as the result of a change from limnic to telmatic conditions of the maar lake, beginning in LPAZ 5a, when Alnus is continuously decreasing. This is supported by a conspicuous increase of Sparganium within LPAZ 5b and the increase in Cyperaceae from LPAZ 5 on. The diatom-gyttjas on top of the peat thus show a rebounding lake level, during the later part of zone 5. Compared to the time before the peat formation, the enrichment in detritus and the bulky texture of the gyttja above the peat indicate higher sedimentation rates in a then shallow lake. The refilling is supported by Pirrung (1998) as well as by Stachel and Bu¨chel (1989) where the superposition gyttjapeat-gyttja was described first. The cause of this succession could be a fluctuation in groundwater level after a catastrophic event and/or by natural equilibrium change in precipitation/evaporation. Alternatively, draining of the lake by erosive cutting of the maar outlet and subsequent closure of the plughole seems possible but improbable. The peat

A New Holsteinian Pollen Record

¨ 3 from the Fig. 27.6 Simplified pollen concentration diagram (scale: [104 pollen/cm3]), with local pollen assemblage zones (LPAZ) and lithology of core DO Do¨ttingen dry maar. Grey shaded curves are amplified tenfold. 405

406

Markus Diehl and Frank Sirocko

¨ 3, their main palynological characterTable 27.1 Local pollen assemblage zones (LPAZ) of core DO istics and lower pollen zone boundaries, based on the pollen percentage diagram of Fig. 27.5 (NATP: nonarboreal terrestrial pollen) LPAZ 7

b a

6

b

a

5

c

b

Main palynological characteristics Betula–Pinus–Picea dominance

d c b a

3

b a

2

1

b

a

High NATP values ð 40%Þ

Final rise of NATP

Expansions of Betula and later on of Picea, sporadic pollen grains of Larix

Increase of Betula, decrease of Alnus and Corylus, disappearance of Abies

(Re)increase of Alnus, Setback of arboreal pollen, relative Decrease of Abies and mixed oak Corylus and Quercus maxima of Poaceae, Ericaceae, forest taxa, disappearance of (higher Corylus than Artemisia, Polypodiales and Buxus Quercus values), lower Sphagnum Abies values whilst high- Increase of grasses and other herbs, Onset of Alnus and Corylus er Betula, Pinus and Picea disappearance of Acer and Fagus (re)expansion values as in LPAZ 4, increase of NATP Strong setback of Alnus Dominance of coniferous pollen, Steep increase of Picea and less pronounced also Pinus and Picea maximum, mixed of Corylus, increase of oak forest minimum, increase of Betula, Pinus and Picea, Cyperaceae, Polypodiales and later on decreases of Sphagnum Carpinus, Taxus and Pterocarya maximum Sparganium End of Alnus decline mixed oak forest taxa high

a 4

Lower boundary

Decrease of Alnus, increase of Betula, Pinus, Picea

Onset of Alnus decline

Expansion of Carpinus and Abies, decline of Corylus, later on dominance of Abies and Alnus besides Quercus, Carpinus, Corylus, Buxus and Pinus, occurrence of Fagus, verification of Celtis and Pterocarya, minima of Picea and Betula within the profile

(Re)expansion of Carpinus, increase (Re)increase of Carpinus of Buxus, occurrence of Fagus

Maximum expansion of Taxus

Setback of Carpinus

Steep decrease of Carpinus

Carpinus maximum, increasing Alnus values

Steep increase of Abies

Betula–Pinus–Picea–Poaceae–peak, setbacks of Taxus and Corylus

Steep decrease of Corylus, occurrence of Carpinus

Corylus maximum

Occurrence of Abies

Initial high of Taxus

Increase of Taxus, decrease of Quercus

Maximum expansion of mixed oak forest members, accompanied with Picea, increasing Corylus and Alnus values whilst decreasing Pinus and Betula percentages

Occurrence of Alnus Ulmus, Fraxinus and Taxus

Betula–Pinus dominance

Transitional period with decreasing Betula and increasing Pinus, few Quercus and sporadic Corylus pollen grains

Beginning decrease of Betula

Afforesting time with Betula, Pinus, Salix, Juniperus and Hippophae¨, high amount of grasses and other herbs

Not present in the profile

A New Holsteinian Pollen Record

thickness allows an estimation of the duration of its forming. Pirrung (1998) reports 40 cm fen-peat and Stachel and Bu¨chel (1989) even 1.5 m of peat in the centre of the maar. Thus, the peat formation must have been of distinct duration rather than having been a short event. Pterocarya is one of the stratigraphically and environmentally indicative pollen taxa in the Do¨ttingen interglacial core section. It occurs today in regions with precipitation rates mostly above 1200 mm/a (Frenzel, 1968) or in lowlands, indicating its recent preference not only for warm conditions but also for high water availability. Thus, its maximum expansion (LPAZ 5b) above the peat within the refilling period of the lake may be evidence for more humidity. On the other hand, Pterocarya settled in the area even while the lake changed to telmatic conditions (LPAZ 5a). This may indicate that it grew only in proximity of the lake, where groundwater was always within reach. The presence of temperate taxa (mixed oak forest elements, Buxus, Pterocarya etc.), of Nymphaea or Nuphar (several pollen grains in LPAZ 5b and 5c) and in particular of Trapa natans (one pollen grain at 9.93–9.94 m, LPAZ 5b/c transition) shows that the rise of Picea, Pinus and even Betula throughout LPAZ 5 cannot be seen as an indicator for a (major) temperature drop. These taxa (Picea, Pinus and Betula) could have occupied earlier Alnus habitats. However, the interpretation of the pollen signal of the LPAZ 5a/b transition is complicated by the intercalated ash, which may have affected the vegetation. In LPAZ 5c, Pterocarya retrogrades together with Buxus, Carpinus, Taxus, Abies, Corylus and the typical mixed oak forest members. Increasing Polypodiales, Sphagnum and Cyperaceae indicate the establishment of telmatic habitats. Presence of Osmunda and Pterocarya indicate continuing temperate conditions. The high amount of coniferous pollen within LPAZ 5 points towards

407

selective corrosion of less resistant pollen types (Straka, 1957, 1975) due to oxic (and acid) depositional environments at the time of peat formation, i.e. when water depth was low. Therefore, the fluctuations of pollen percentages within LPAZ 5 are not interpreted as a sign for climate deterioration towards cooler temperatures. This is supported by low nonarboreal pollen percentages which do not point to a major decrease in forest cover within LPAZ 5. However, whether the lake-level fluctuation was associated with a phase of precipitation changes remains an open question. Percentages of grasses and other herbs increase in LPAZ 6a and level off around 10% of the total sum. Sedges (Cyperaceae) reach roughly 5–10%. They are interpreted as indicating a telmatic environment, rather than an open land vegetation. This is supported by the low percentages of heliophytes which indicate open vegetation, e.g. Artemisa, and because thermophilous species (Quercus, Carpinus, Taxus, Tilia, Abies, Buxus, Pterocarya and Ilex) are still present, although in lower percentages than before. Sphagnum, Polypodiales and Ericaceae indicate acid soils due to their telmatic habitats and/or due to increasing soil degradation. The latter may be enhanced by the dominance of conifers (Abies, Pinus, Picea and Taxus) in the area. Corylus and mixed oak forest taxa are slightly enriched in the later part of LPAZ 6a, as well as arboreal pollen itself. An explicit increase of grasses and herbs, especially of Artemisia (also presence of Thalictrum), is accompanied by decreases of thermophilous taxa in LPAZ 6b, indicating a climate deterioration towards the following LPAZ 7a. From LPAZ 6 on, the number of pollen grains affected by decomposition slightly increases, which could be explained by a partly aerobe sedimentation environment in the shallow lake and/or by enhanced entry of reworked material. Pollen number per cubic centimetre is generally lower in the second part of the Do¨ttingen sequence (Fig. 27.6), which indicates lower pollen influx and suggests a higher

408

Markus Diehl and Frank Sirocko

sedimentation rate. Poaceae pollen grains in the very pollen-poor LPAZ 7b are often crumbled and most of the nondeterminable pollen grains are probably grasses. Therefore and because of the fraction of nonautochthonic pollen grains within this LPAZ, arboreal pollen is surely overrepresented in this pollen subzone. 27.6 DISCUSSION 27.6.1 Palynological characteristics of the Holsteinian in Northern Germany The subcommission on European Quaternary stratigraphy (SEQS) considered the Lower Elbe area as the type area for the Holsteinian Interglacial (Jerz and Linke, 1987) and defined the boundaries of the Holsteinian between the transitions from subarctic (tundra) to boreal (taiga), and from boreal to subarctic conditions, respectively, at the profiles Eggstedter Holz, Hamburg-

Fig. 27.7 Map of Holsteinian sites in north-central Germany (W.: Wacken, E.H.: Eggstedter Holz, B.: Bossel, H.: Hamburg, Gra.: Granzin, P./P.: Pritzwalk Prignitz, N./W.: Neuruppin Wuthenow, M./ B.: Munster Breloh, H./B.: Hetendorf Bonstorf, O.O.: Ober-Ohe, K.: Klieken, Gro¨.: Gro¨bern, D./ W.: Delitzsch Wo¨lkau, R.: Rossendorf, D.: Do¨ttingen). Sites affected by the Holsteinian sea transgression are marked dark.

Dockenhuden, Bossel and Pritzwalk (lower boundary) and Munster/Breloh and Pritzwalk (upper boundary) (see Fig. 27.7 for site locations). Numerous pollen assemblage zones (PAZ) were defined, but they differ in number and definition from author to author. Szafer (1953) distinguished three, Selle (1955) six, Linke and Hallik (1993) five (with subzones eight) PAZ, Erd classified seven to eight main PAZ with several subdivisions (Erd, 1969, 1973, 1987; Erd and Mu¨ller, 1977), while Mu¨ller (1974b) defined the most complex division with up to fourteen zones (not including subzones) for the Holsteinian interglacial. However, the classifications developed by Erd and by Mu¨ller have been established for the Holsteinian vegetation (pollen) succession of northern Germany. Both show the typical, main pollen succession features of the Holsteinian for continental northern Europe: An early appearance of a mixed oak forest, a subsequent Taxus phase, both periods accompanied by Picea, later on a replacement of the two latter species by a Carpinus–Abies association together with Buxus, Pterocarya, Fagus and Celtis as indicative species. Erd defined a classification into seven zones on the profiles Granzin and Pritzwalk/Prignitz (Erd, 1969, 1973, 1987) (Fig. 27.8). Both localities are in the Saalian till plain in front of the Weichselian end moraine. In the profile Wuthenow/Strausberg (Cepek and Erd, 1975), pollen zones 5, 6 and 7 were subdivided into two parts each, and a following cold phase, the Fuhnian, was described. Finally on the profile Rossendorf (Erd, 1987), a temperate-to-interglacial pollen zone 8 between the subboreal-to-boreal PAZ 7 and the Fuhnian was added. Erd’s final subdivision of the Holsteinian vegetation history is reproduced in Table 27.2. An alternative zonation of the Holsteinian was given by Mu¨ller (1974b) for the Kieselgur (diatomite) deposits at Munster/ Breloh (Fig. 27.9). He established his classification on pollen counts from four cores and three outcrop sections. Mu¨ller

A New Holsteinian Pollen Record

409

Fig. 27.8 Pollen profiles from the sites Granzin (top) and Pritzwalk/Prignitz (bottom) (Erd, 1969, 1970), with pollen assemblage zones (PAZ). Both sites were affected by the Holsteinian transgression and show hiatae in their upper parts (PAZ 6 at Pritzwalk/Prignitz).

combined his counted sections with the work of Meyer (1974) on the nearby site of Hetendorf/Bonstorf for PAZ 1 to 5, which were not accessible at Munster/Breloh. Mu¨ller’s structuring was applied to the site of Bossel (Mu¨ller and Ho¨fle, 1994) and to Hamburg Hummelsbu¨ttel (Averdieck, 1992). Mu¨ller used fourteen main PAZ for the entire interglacial (Table 27.3). The difference to Erd’s classification is in particular in two Betula–Pinus peaks (BPP) (Mu¨ller’s

PAZ VIII and XI/XIIa). The older, immediately before the Carpinus–Abies period, marking the end of the initial Taxus expansion, coincided with a weak expansion of grasses and a short retreat of thermophilous pollen percentages. The second one is mainly characterized by a temporarily breakdown of Carpinus during the Carpinus–Abies phase. Mu¨ller’s older Betula–Pinus peak was also observed in the nearby Ober-Ohe

410

Markus Diehl and Frank Sirocko

Table 27.2 Holsteinian pollen assemblage zones (PAZ) after Erd (1987) and their main palynological characteristics PAZ

Main palynological characteristics FUHNIAN

H O L S T E I N I A N

8

‘Rewarming phase’ similar to PAZ 6, but with increased values of grasses, herbs, Calluna, Polypodiales, Sphagnum and presence of Hippophae¨

7 a, b

Subboreal to boreal zone with increased herbs and some Larix

6 a, b

Abies, Carpinus dominate associated with Buxus, lowest values of Picea throughout the interglacial, presence of Fagus

5 a, b

Occurrence of Pterocarya

Occurrence of Celtis, maximum spread of Buxus and Vitis

4

Picea–Taxus–Corylus zone, partially displacement of Taxus and Picea by Carpinus and later Abies, maximum spread of Tilia, first occurrence of Buxus and Vitis

3

High values of Quercus, Taxus, Corylus, Alnus, Picea, Carpinus begins to spread, Tilia, Ulmus, Acer, Ilex occur more frequently

2

Transitional period

1

Betula–Pinus time associated with Salix, Juniperus and herbs

Whereas PAZ 6b still belongs to the climatic optimum, PAZ 7a marks a transitional period towards boreal condition in PAZ 7b. Erd’s PAZ 8 could not be confirmed on any other Holsteinian profile yet.

diatomites (Gistl, 1928; Selle, 1954) as well as at Klieken (Majewski, 1961; Neumann, 2000) and Rossendorf (Erd, 1987) and less obvious at Delitzsch-Wo¨lkau (Dassow, 1987), but no similar pattern is visible at Gro¨bern (Eissmann et al., 1995). Mu¨ller’s second Betula–Pinus peak can be correlated only with the proximate site Ober-Ohe (Gistl, 1928; Selle, 1954), but also Gro¨bern (Fig. 27.10) shows an increase in Pinus (and less obvious in Betula) percentages at the respective profile section. Furthermore, an ‘irregular’ pattern with increasing Betula or Pinus (or even Larix) percentages, accompanied by a retreat of more demanding species, somewhat in the ‘middle’ of the interglacial profiles, is obvious at Rossendorf (Erd, 1987) and at Gro¨bern (Eissmann et al., 1995). The mentioned sites differ extremely in sampling resolution and profile completeness. Obviously, only within terrestrial sequences, features similar to the older Munster/Breloh BPP were found, but not at sites affected by the Holsteinian transgression (Fig. 27.7), as

documented by diatoms (see Fig. 27.8 for sites Granzin and Pritzwalk/Prignitz). Discrepancies are also apparent in the interpretation of the Betula–Pinus peaks. Whereas Mu¨ller (1974b) interpreted them as climatic setbacks, Cepek and Erd (1975) lead them back to local influences. The same interpretation problems occur with the Picea peak at Rossendorf or the Betula doublepeak at Gro¨bern. 27.6.2 The Do¨ttingen interglacial sequence and its correlation with the Holsteinian The Do¨ttingen pollen sequence with its early mixed oak forest, a subsequent phase with high Taxus values, both periods accompanied by Picea and later a Carpinus–Abies phase with Buxus, Pterocarya, Fagus and Celtis (Fig. 27.5) fits the main pollen succession features for the Holsteinian, as observed from the north German sites in the Holsteinian-type region.

A New Holsteinian Pollen Record

411

Fig. 27.9 Pollen profile from site Munster/Breloh after Mu¨ller (1974b) with his pollen assemblage zones (PAZ). The profile is schematic and resumes from the comparison of four cores and three outcrop sections. In the extremely similar pollen profile from the nearby site Ober-Ohe (Gistl, 1928; Selle, 1954), the early Betula– Pinus peak does not come with such catastrophic setbacks of Alnus, Corylus or even Quercus as in the Muster/Breloh profile.

The Do¨ttingen profile can be clearly distinguished from Holocene or Eemian pollen sequences. The Holocene pollen succession for the Eifel is documented from the maar lakes Holzmaar and Meerfelder Maar (Litt, 1999, 2000). Anthropogenic indicators like Secale, the absence of Abies, Taxus, Buxus, Pterocarya, the unimportance of Picea and a high abundance of Fagus are some characteristics within these profiles which preclude the Do¨ttingen sequence as belonging to the Holocene. The vegetation development in the Eemian across northern Germany is documented in Litt (1994); Menke and Tynni (1984) or Mu¨ller (1974a). Sirocko et al. (2005) published an Eemian pollen profile from the Eifel, see also Seelos and Sirocko (this volume). The Eemian succession distinguishes from that of Do¨ttingen especially by an increase of Picea and Tilia after (not while)

an initial Quercus-dominated mixed oak forest phase, a late occurrence of Abies, even while already decreasing percentages of Carpinus and a clearly time-delayed increase of Corylus and Alnus compared to Ulmus and Quercus. Pterocarya and Celtis are unknown from Eemian profiles. However, differences in the magnitude of the pollen percentages occur between the low mountain range site Do¨ttingen and the Holsteinian lowland sites of northern Germany. These differences could be due to regional peculiarities, and the Do¨ttingen profile therefore may represent a ‘low mountain range-type’ pollen profile for the Holsteinian to be distinguished from ‘lowlandtype’ profiles. For example, Pinus pollen percentages are comparatively low in the early temperate zones at Do¨ttingen, whereas they stay at high values throughout the

412

Markus Diehl and Frank Sirocko

Table 27.3 Holsteinian pollen assemblage zones (PAZ) after Mu¨ller (1974b) and their main palynological characteristics (": increase, #: decrease, >: higher, <: lower, : similar percentages) PAZ

Main palynological characteristics FUHNIAN

XIV

Pinus time (Pinus, Betula, Picea, Alnus, Ericales)

XIII

Quercus–Alnus decline time (Pterocarya, Fagus, Celtis present)

XII

C B

Quercus–Abies (Pinus–Alnus) time

Highest Quercus values throughout the interglacial Carpinus ", Buxus ", Vitis "

A H O L S T E I N I A N

XI

Younger Betula–Pinus peak (Alnus #, virtually disperse of Carpinus)

X

Carpinus–(Alnus–Pinus) time (Pinus, Alnus, Carpinus Betula, Abies)

IX

Corylus–Picea–(Alnus–Pinus) time (Pinus, Alnus, Corylus, Quercus, Carpinus ", Abies ")

VIII

Older Betula–Pinus peak (Corylus #, Quercus #, Alnus #, virtually disperse of Taxus) C

VII

B

Taxus–Corylus–Picea–(Pinus– Alnus) time

A

Abies, Carpinus, Osmunda occur Pinus < Alnus, Picea < Quercus, Corylus, Taxus Pinus > Alnus, Picea > Quercus, Corylus, Taxus

VI

Picea–Alnus–(Pinus–Betula) time (Alnus ", Picea, Pinus, Betula)

V

Ulmus–Pinus–Betula time, representatives of the mixed oak forest (Quercus, Ulmus, Tilia, Fraxinus, Acer) occur

IV

Pinus–Betula time (Pinus > Betula)

III

Betula–Pinus time (Betula Pinus)

II

Betula time (Betula > Pinus)

I

Afforesting time (Hippophae¨, Juniperus, Artemisia, Helianthemum, Thalictrum, tertiary sporomorphs)

Holsteinian in northern Germany. A main reason for this may be seen in edaphic conditions. In northern Germany, poor acid soils (podsols) ground on meltwater sands or tills, whereas loesses and brownearth soils are widespread in the Eifel and may be assumed also for pre-Holocene interglacials. On the other hand, higher values for Abies at mountain range sites in comparison with lowland sites have also been documented for the Eemian (Gru¨ger, 1995; Mu¨ller, 2000). The Betula–Pinus–Picea–Poaceae peak (Do¨ttingen LPAZ 4a) at the end of the Taxus period and the start of the Carpinus–Abies time allows comparison with the first Betula–Pinus peak from Munster/Breloh (Mu¨ller’s PAZ VIII). This likely equivalent to Mu¨ller’s older BPP is just about 5 cm in sediment thickness and was only detected through systematic searching of the relevant

core section. The peak in Poaceae pollen grains is only within sample 12.73–12.72 m, the Betula peak only within the following sample 12.72–12.71 m, whereas the Pinus peak occurs between 12.74 m and 12.69 m depth, centred at 12.73–12.71 m depth. The small depth range of this event shows that insufficient sampling resolution (compared to temporal resolution of the records) or even little reworking of the concerning sediment can conceal this feature. The observed ash layer (at 12.90–12.77 m depth) immediately below the LPAZ 3/4 boundary may explain the change in plant associations. In this case, the Betula–Pinus–Picea–Poaceae peak would represent pioneer vegetation after a catastrophic volcanic event. However, no charcoal or wood remains were detected, indicating that the location was not in close proximity to the corresponding

A New Holsteinian Pollen Record

413

Fig. 27.10 Simplified pollen profile from site Schmerz-Gro¨bern [originally in Eissmann et al., (1995), revised and redrawn after Ku¨hl and Litt, this volume] with pollen assemblage zones (PAZ) numbered after the classification of Erd. The setback of Carpinus after its initial expansion shall coincide with the Do¨ttingen LPAZ 4c and therefore with Mu¨ller’s second Betula–Pinus peak.

eruption. Beneath the ash layer, down to about 13.05 m depth, spanning the whole LPAZ 3b, the sediment layering is inclined (with ash remains at .075–13.055 m depth), indicating displacement likely in causal connection with the ash (volcanic tremor, earthquake, slump or ash load). This could also explain the conspicuous decrease in pollen concentration in that depth (Fig. 27.6). A destructive, disastrous temperature backstroke as a reason for the Betula–Pinus– Picea–Poaceae peak [as suggested by Mu¨ller (1974b)] for his first BBP seems unlikely, because thermophilous taxa like Quercus (and in general the mixed oak forest members) are unaffected. No significant signal can be obtained from cold-resistant, heliophilous herbs (e.g. Artemisa and Thalictrum), which would be expected if this event had a climatic cause. Furthermore, from the pollen concentration diagram (Fig. 27.7) it can be obtained that the Betula, Pinus and Picea

peaks in the pollen percentage diagram are not caused by extraordinary increases of the absolute pollen number of these taxa. In fact, their percentage peaks are mainly caused by a not proportionally steep drop of especially Corylus pollen concentration per volume. Only the Poaceae pollen percentage peak is accompanied by a steep increase in absolute pollen number. The presence of the Betula–Pinus peak both in the Eifel and in north Germany strengthens the demand for an over-regional connection. However, there are also differences between both sites. At Munster/Breloh, the event is characterized by a Betula–Pinus peak in combination with steep decreases of Alnus and Corylus, which both recover afterwards. In contrast to Munster/Breloh, the event at Do¨ttingen marks the beginning decrease of Corylus which does not show a (re)expansion after the event, even to higher percentages than before as at Munster/Breloh. Within the

414

Markus Diehl and Frank Sirocko

Do¨ttingen sequence, the Betula–Pinus–Picea– Poaceae peak is immediately followed by peaks of Corylus and Quercus. At Do¨ttingen, no setback of Alnus was found. The Do¨ttingen core reflects another feature that is also evident in several Holsteinian sites (Munster/Breloh, Klieken and Gro¨bern) and which might be correlated between these profiles. This is the contemporarily setback of Carpinus in LPAZ 4c after its initial expansion. At Munster/ Breloh, this event coincides with the second Betula–Pinus peak (Mu¨ller’s PAZ XI), at Gro¨bern with a small Pinus peak (between 19.00 and 19.50 m depth). The presence of thermophilous taxa (e.g. Tilia, Vitis, Buxus, Osmunda etc.) at Do¨ttingen seems to contradict a connection of this phase with climate deterioration, in particular because no signal comes Table 27.4 Tentative pollen zone correlation between Mu¨ller´s and Erd´s pollen assemblage zones (PAZ), and the local pollen assemblage zones (LPAZ) from Do¨ttingen (BPP: Betula– Pinus peak) PAZ After Mu¨ller

PAZ After Erd

¨3 LPAZ DO

XV

Fuhnian

7b

XIV

7

7a

XIII

5, 6 C

XII

6

4d

5

4c

B A (2. BPP)

XI (2. BPP) X IX

4b

VIII (1. BPP)

4

C VII

B

4a 3b

3

3a

2

2

A VI V IV

1b

III II I

1 1a

from Pinus, Betula or from the nonarboreal taxa. Instead, an ash stripe of about 2 cm thickness lies within this zone and may point again to an interaction with vegetation dynamics. ¨ 3 PAZ with A tentative correlation of the DO Erd’s and Mu¨ller’s palynological classification of the Holsteinian is done in Table 27.4. From the pollen zone correlation presented in Table 27.4, Mu¨ller’s second Betula–Pinus peak was at the same time as the Carpinus setback in Do¨ttingen’s LPAZ 4c, whereas Mu¨ller’s older BPP coincides with Do¨ttingen’s LPAZ 4a, respectively the Taxus decline in the beginning of Erd’s PAZ 4. Do¨ttingen’s LPAZ 6 must have been of comparatively short duration but with high sedimentation rates, indicated by its sediment thickness and detritus content. 27.7 CONCLUSION (1) The Do¨ttingen pollen sequence belongs to the Holsteinian interglacial. It represents the vegetation succession for the low mountain range site of the Eifel. (2) The Do¨ttingen profile offers a bipartite sequence, split into a limnic earlier and a telmatic later part with increasing nonarboreal pollen percentages. An episode of peat formation marks the boundary of this bipartition. (3) The older Betula–Pinus peak from Munster/Breloh can also be observed at Do¨ttingen. A volcanic eruption occurred immediately before this event. No signal for climate deterioration during the corresponding Do¨ttingen LPAZ 4a was found. (4) The Do¨ttingen profile documents a setback of Carpinus, correlatable with the younger Betula–Pinus peak from Munster/Breloh. No evidence for a causal connection to climate deterioration was found. Again a small ash stripe points to volcanic activity during this period. (5) The interglacial ends with an open boreal forest dominated by Betula, Pinus and Picea.

A New Holsteinian Pollen Record

In summary, the Holsteinian Do¨ttingen site reveals similar vegetation ‘anomalies’ as in northern Germany, but the Eifel record does not corroborate the existence of regionwide severe cold events within the interglacial. In contrast, the Eifel vegetation ‘anomalies’ develop subsequent or contemporarily to phases of volcanic activity. If this can be indeed interpreted as a climate/plate tectonic relation or as a response of central European vegetation to volcanic activity in the Eifel remains to be discovered.

ACKNOWLEDGEMENTS The study was funded by the German Ministery for Research and Education in the frame of the DEKLIM program. Discussions with Charles Turner, Helmut Mu¨ller and Thomas Litt provided insights into the Holsteinian vegetation dynamics. We thank in particular Thomas Litt, who introduced the first author into the technique of pollen preparation and the taxonomy of palynomorphs.

REFERENCES Averdieck, F.R., 1992. Das Holstein-Interglazial von Hamburg-Hummelsbu¨ttel. Meyniana 44, 1–13. Berglund, B.E., and Ralska-Jasiewiczowa., 1986. Pollen analysis and pollen diagrams. In: B.E. Berglund (Ed.), Handbook of Holocene palaeoecology and palaeohydrology, pp. 455–484. John Wiley, Chichester. Beug, H.-J., 2004. Leitfaden der Pollenbestimmung fu¨r Mitteleuropa und angrenzende Gebiete. Friedrich Pfeil Verlag, Mu¨nchen. Bu¨chel, G., 1984. Die Maare im Vulkanfeld der Westeifel, ihr geophysikalischer Nachweis, ihr Alter und ihre Beziehung zur Tektonik der Erdkruste. Unpublished PhD Thesis, Johannes GutenbergUniversita¨t. Bu¨chel, G., and Lorenz, V., 1982. Zum Alter des Maarvulkanismus der Westeifel. Neues Jahrbuch fu¨r Geologie und Pala¨ontologie 163, 1–22. Bu¨chel, G., Negendank, J.F.W., Wuttke, M., and Viereck-Go¨tte, L., 2000. Quarta¨re und tertia¨re Maare der Eifel, Enspel (Westerwald) und Laacher See: Vulkanologie, Sedimentologie und Hydrogeologie.

415

In: F.O. Neuffer, and H. Lutz (Eds.), Internationale Maar-Tagung. Mainzer Naturwissenschaftliches Archiv, Daun/Vulkaneifel. Cepek, A.G., and Erd, K., 1975. Das Holstein-Interglazial im Raum Neuruppin - ein neues pollenstratigraphisches Richtprofil und seine quarta¨rgeologische Bedeutung. Zeitschrift fu¨r geologische Wissenschaften 3, 1151–1178. Dassow, W., 1987. Neue Holstein-Interglazial-Profile aus dem Quarta¨r im Raum Leipzig. Zeitschrift fu¨r geologische Wissenschaften 15, 195–203. Eissmann, L., Litt, T., and Wansa, S., 1995. Elsterian and Saalian deposits in their type area in central Germany. In: J. Ehlers, S. Kozarski, and P.L. Gibbard (Eds.), Glacial deposits in North-East Europe. A.A. Balkema/Rotterdam/Brookfield, pp. 439–464. Erd, K., 1969. Das Holstein-Interglazial von Granzin bei Hagenow (Su¨dwestmecklenburg). Geologie 18, 590–599. Erd, K., 1970. Pollen-analytical classification of the middle Pleistocene. In: The German Democratic Republic. Palaeogeography, Palaeoclimatology, Palaeoecology 8, 129–145. Erd, K., 1973. Vegetationsentwicklung und Biostratigraphie der Do¨mnitz-Warmzeit (Fuhne/Saale 1) im Profil von Pritzwalk/Prignitz. Abhandlungen des Zentralen Geologischen Instituts 18, 9–48. Erd, K., 1987. Holsteininterglaziale Ablagerungen von Rossendorf bei Dresden. Zeitschrift fu¨r geologische Wissenschaften 15, 281–295. Erd, K., and Mu¨ller, A., 1977. Die Pleistoza¨nprofile Prellheide und Wildschu¨tz, Bezirk Leipzig, mit vollsta¨ndigem Holstein-Interglazial. Zeitschrift fu¨r geologische Wissenschaften 5, 745–765. Erdtman, G., 1969. Handbook of Palynology: An Introduction to the Study of Pollen Grains and Spores. Munksgaard, Copenhagen. Faegri, K., and Iversen, J., 1989. Textbook of Pollen Analysis. Wiley, New York. Frenzel, B., 1968. Grundzu¨ge der pleistoza¨nen Vegetationsgeschichte Nord-Eurasiens. Franz Steiner Verlag, Wiesbaden. Gistl, R., 1928. Die letzte Interglazialzeit der Lu¨neburger Heide pollenanalytisch betrachtet. Botanisches Archiv 21, 648–710. Gru¨ger, E., 1995. Correlation of Middle-European Late-Pleistocene pollen sequences of the Pfefferbichl and Zeifen types. Mededelingen Rijks Geologische Dienst 52, 97–104. Hallik, R., 1960. Die Vegetationsentwicklung der Holstein-Warmzeit in Nordwestdeutschland und die Altersstellung der Kieselgurlager der su¨dlichen Lu¨neburger Heide. Mitteilung aus dem Geologischen Landesamt Hamburg 32, 326–333. Jerz, H., and Linke, G., 1987. Arbeitsergebnisse der Subkommission fu¨r Europa¨ische Quarta¨rstratigraphie: Typusregion des Holstein-Interglazials

416

Markus Diehl and Frank Sirocko

(Berichte der SEQS 8). Eiszeitalter und Gegenwart 37, 145–148. Linke, G., and Hallik, R., 1993. Die pollenanalytischen Ergebnisse der Bohrungen Hamburg-Dockenhuden (qho 4), Wedel (qho 2) und Hamburg-Billbrook. Geologisches Jahrbuch A 138, 169–184. Litt, T., 1994. Pala¨oo¨kologie, Pala¨obotanik und Stratigraphie des Jungquarta¨rs im nordmitteleuropa¨ischen Tiefland. Dissertationes Botanicae 227. Litt, T., 1999. Bio- and chronostratigraphy of the lateglacial in the Eifel region, Germany. Quaternary International 61, 5–16. Litt, T., 2000. Vegetation history and palaeoclimatology of the Eifel region as inferred from palaeobotanical studies of annually laminated lake sediments. Terra Nostra 6: International Maar Conference, Daun/Vulkaneifel, 259–263. Lorenz, V., and Zimanowski, B., 2000. Vulkanologie der Maare der Westeifel. In: F.O. Neuffer, and H. Lutz (Eds.), Internationale Maar-Tagung. Mainzer Naturwissenschaftliches Archiv, Daun/Vulkaneifel. Majewski, J., 1961. Pollenanalytische Untersuchung der Kieselgur von Klieken. Geologie 32, 10–14. Menke, B., and Tynni, R., 1984. Das Eeminterglazial und das Weichselfru¨hglazial von Rederstall/Dithmarschen und ihre Bedeutung fu¨r die mitteleuropa¨ische Jungpleistoza¨n-Gliederung. Geologisches Jahrbuch 76. Meyer, K.-J., 1974. Pollenanalytische Untersuchungen und Jahresschichtenza¨hlungen an der holstein-zeitlichen Kieselgur von Hetendorf. Geologisches Jahrbuch A 21, 87–105. Moore, P. D., Webb, J. A., and Collinson, M. E., 1991. Pollen Analysis. Blackwell Scientific Publications, Oxford. Mu¨ller, H., 1974a. Pollenanalytische Untersuchungen und Jahresschichtenza¨hlungen an der eem-zeitlichen Kieselgur von Bispingen/Luhe. Geologisches Jahrbuch A 21, 149–169. Mu¨ller, H., 1974b. Pollenanalytische Untersuchungen und Jahresschichtenza¨hlungen an der holstein-zeitlichen Kieselgur von Munster-Breloh. Geologisches Jahrbuch A 21, 107–140. Mu¨ller, U. C., 2000. A Late-Pleistocene pollen sequence from the Jammertal, south-western Germany with particular reference to location and altitude as factors determining Eemian forest composition. Vegetation History and Archaeobotany 9, 125–131.

Mu¨ller, H., and Ho¨fle, H.-C., 1994. Die HolsteinInterglazialvorkommen bei Bossel westlich von Stade und Wanho¨den no¨rdlich Bremerhaven. Geologisches Jahrbuch 134, 71–116. Neumann, F., 2000. Pollenanalyse des Holsteinvorkommens in Klieken und palynostratigraphische Anwendung am Ostufer des Arendsees. Unpublished Diploma Thesis, Rheinische Friedrich Wilhelm-Universita¨t Bonn. Pirrung, B. M., 1998. Zur Entstehung isolierter alttertia¨rer Seesedimente in zentraleuropa¨ischen Vulkanfeldern. Mainzer Naturwissenschaftliches Archiv 20. Schmincke, H. U., 2000. Vulkanismus. Wissenschaftliche Buchgesellschaft Darmstadt. Selle, W., 1954. Die Vegetationsentwicklung des Interglazials von Ober-Ohe in der Lu¨neburger Heide. Abhandlungen vom Naturwissenschaftlichen Verein zu Bremen 33, 457–463. Selle, W., 1955. Die Vegetationsentwicklung des Interglazials vom Typ Ober-Ohe. Abhandlungen vom Naturwissenschaftlichen Verein zu Bremen 34, 33–46. Sirocko, F., Seelos, K., Schaber, K., Rein, B., Dreher, F., Diehl, M., Lehne, R., Ja¨ger, K., Krbetschek, M., and Degering, D., 2005. A Late Eemian Aridity Pulse in central Europe during the last glacial inception. Nature 436, 833–836. Stachel, T., and Bu¨chel, G., 1989. Das Do¨ttinger Maar: Fallstudie eines großen tertia¨ren (?) Tuffschlotes im Vulkanfeld der Hocheifel. Zeitschrift der deutschen geologischen Gesellschaft 140, 35–51. Stockmarr, J., 1971. Tablets with spores used in absolute pollen analysis. Pollen et Spores 13, 615–621. Straka, H., 1957. Pollenanalyse und Vegetationsgeschichte. A. Ziemsen-Verlag - Wittenberg Lutherstadt, Kiel. Straka, H., 1975. Pollen und Sporenkunde: Eine Einfu¨hrung in die Palynologie. In: Grundbegriffe der modernen Biologie. Fischer Verlag, Stuttgart, pp. 238. Szafer, W., 1953. Stratygrafia plejstocenu w Polsce na podstawie florystyczneiy (Pleistocene stratigraphy of Poland from the floristical point of view). Rocz. Pol. Tow. Geol. 22, 1–99. van den Bogaard, B., and Schmincke, H.-U., 1990. Die Entwicklungsgeschichte des Mittelrheinraumes und die Eruptionsgeschichte des Osteifel-Vulkanfeldes. In: W. Schirmer (Ed.), ‘‘Rheingeschichte zwischen Mosel und Maas.’’, Hannover, pp. 166–190.

28. Interglacial Pollen Records from Scho¨ningen, North Germany Brigitte Urban University of Lu¨neburg, Campus Suderburg, Herbert-Meyer-Str. 7, 29556 Suderburg, Germany

ABSTRACT

28.1 INTRODUCTION

The Pleistocene sequence of the Scho¨ningen lignite mine contains a number of interglacial and interstadial limnic and peat deposits, travertine tuff, soils, tills and fluvioglacial as well as loess deposits. There is evidence of four interglacials younger than the Elsterian glaciation and preceding the Holocene. The complex Pleistocene sequence contains six major cycles. The sequence begins with Late Elsterian glacial and interstadial deposits preceding the Holsteinian, followed by the Reinsdorf and Scho¨ningen interglacials, which represent the pre-Drenthe (Early Saalian Stadial) period. A pedocomplex developed in alluvial loess, younger than the Drenthe Stadial of the Saalian glaciation, is succeeded by a sequence of soft travertine and peat representing Eemian marine isotope stage 5e (MIS 5e) and MIS substages 5d, 5c. Channel sediments provide evidence of the Weichselian late glacial and the Holocene. The Scho¨ningen pollen record of MIS 5 and tentative correlatives of MIS 7, 9 and 11 as well as temperate interstadials of Late Elsterian and (intra) Saalian (s.l.) age are discussed and compared with other pollen records of North Germany, Western and Central Europe. Detailed pollen evidence for the Reinsdorf sequence, which is significant for its Lower Palaeolithic sites, including stadials and interstadials, is a major focus.

The Pleistocene sequence of Scho¨ningen (Lower Saxony, Germany) provides a key link between unglaciated and glaciated areas in Western Central Europe (Fig. 28.1). The complex Pleistocene sequence is of significance for the subdivision of the glaciated younger Middle Pleistocene part of Western Central Europe (Thieme et al., 1987; Urban et al., 1988, 1991a, 1991b; Urban 1995a, 1995b, 1996a, 1996b, 1999, 2002) and for archaeological evidence of early human occupation by Homo erectus (Thieme et al., 1987, 1992, 1993; Thieme and Maier, 1995). The investigations have focussed on exposed Quaternary deposits that are considered to span much of the last 500 000 years. The investigations occurred as excavation fronts progressed during mining of the underlying Eocene lignites. The Quaternary deposits are composed of various types of sediments including peaty, muddy and silty layers from former swamps, lakes, peat bogs and river flood plains, as well as hydromorphic soils that contain characteristic pollen assemblages. Fossil remains of molluscs, small and large mammals, fishes, reptiles and plant macro fossils are fairly abundant in some layers. The classical Holsteinian interglacial underlain by Elsterian glacial and Late Elsterian interstadial sediments was followed by a cold climatic deterioration interrupted by several temperate phases, the Mißaue 1, Mißaue 2 and SU A interstadial (Urban et al., 1991b). This cycle was later termed Channel 1 (Fig. 28.2) (Mania, 1998). A series of six such cycles of interglacial or

Keywords: Pleistocene, palaeobotany, palaeoclimate, 230Th/234U dating, interglacials

418

Brigitte Urban

and Channel III (Scho¨ningen Interglacial) represent warm climatic periods older than the Saalian ice advance. Channel IV, a pedosequence of pseudogleyic, alluvial loess, which has not yet been fully studied (Altermann, in preparation), overlies the older Saalian till (Drenthe stadium). The sediment in Channel V is a sequence of travertine, silts and peat, which has been correlated to the Eemian and to stage 5e of the marine isotope record (Urban et al., 1991a; Heijnis, 1992; Heijnis and Urban, 1995). Early Weichselian silts composed of loess, solifluction layers and fluvial deposits mark the onset of strong cooling. Late Weichselian Allero¨d peat with the Laacher See volcanic tuff layer and Younger Dryas silt had been identified underlying the Holocene sequence (Channel VI) (Urban et al., 1988) (Fig. 28.2). There is still debate about the stratigraphic position and correlation of the Reinsdorf and the Scho¨ningen interglacials and major interstadials found in Scho¨ningen with other pollen records and the marine

1 4

3 2

Weichselian Saalian 1- Wacken 2- Schöningen 3- Pritzwalk / Prignitz 4- Hoogeveen

Fig. 28.1 Maximum extension of Saalian and Weichselian ice sheets. Locations of Hoogeveen, Wacken, Scho¨ningen and Pritzwalk / Prignitz (Do¨mnitz).

interstadial and early glacial deposits in depressions (channels) have been identified, all considered to be climatically induced. Channel II (Reinsdorf Interglacial)

Exposure at the open-cast lignite mine, Schöningen NE

SW Esbeck Elm

Rim syncline

Salt dome

I II

III

IV

V

VI

b

1

4

2

5

3

6

~500 m

7

10

8

11

9

b

12

Fig. 28.2 Schematic section through the Quaternary sedimentary cycles I–VI (modified after Mania 1995; Mania, 1998; Thieme, 1997). The thickness of the geological deposits is not shown to scale. The Quaternary sediments reach a maximum thickness of ca. 45 m in the Esbeck mining field (Urban et al., 1991b). The actual distance between cycle VI Channel filling and the salt dome is about 2 km (this distance is not shown to scale). 1, Elsterian glacial deposits; 2, Saalian glacial deposits; 3, lacustrine deposits; 4, limnic telmatic sequences; 5, soil complexes; 6, loess deposits; 7, evaporites; 8 gypsum cap rock; 9, Buntsandstein; 10, Triassic limestone (Muschelkalk); 11, Triassic deposits (Keuper); 12 ð¼ bÞ, Tertiary deposits.

Interglacial Pollen Records from Scho¨ningen

istotope statigraphy. Of particular interest is the age and stratigraphic position of the Reinsdorf sequence which contains archaeological horizons with wooden throwing spears that are thought to be the oldest hunting weapons so far discovered (Thieme, 1996, 1997, 1998, 1999). 28.2 OPEN MINE ESBECK/ ¨ NINGEN SCHO The halokinetic depression of Scho¨ningen formed on the flanks of the HelmstedtStaßfurt Zechsteinian salt dome and was filled with thick limnic and marine sequences during Eocene times. At Scho¨ningen, these Tertiary strata are unconformably overlain by Quaternary sediments and soils of Middle and Late Pleistocene and Holocene age (Fig. 28.2). Mania (1998) described the Quaternary halokinetic processes as persisting mainly during interglacial periods, when shallow depressions, the channels, were formed, trending predominantly from northwest to southeast, corresponding to the trend of the salt dome. These channels have been filled with thick limnic, telmatic sediments or pedocomplexes (Fig. 28.2). Fennoscandian glaciers advancing from the northeast filled the depressions with glacial deposits. The study area has been covered twice by a Fennoscandian ice sheet, first during the Elsterian, the oldest ice advance of Northern Germany and Central Europe, and then the later Saalian advance (Fig. 28.1), and it was unglaciated during the last glaciation (Weichselian, MIS 4-2). 28.2.1 Elsterian The oldest Pleistocene sediments are glaciofluvial sands of an early Elsterian ice advance, which are overlain by two Elsterian tills. The younger till is capped by rhythmites, which in turn are overlain by three interstadial layers, formed in shallow basins, locally named Offleben 1, Offleben 2 and

419

Esbeck (Urban et al., 1988, 1991b; Elsner, 2003; Wansa, unpublished data) (Fig. 28.3). During the Offleben 1 and Offleben 2 interstadials, a boreal forest type prevailed, mainly dominated by birch, pine and spruce and locally elder in the late phases of a growing fen. The youngest of the investigated interstadials, the Esbeck Interstadial (Fig. 28.4), which is documented by nearly 3 m of organic mud and peat, is characterised by an early phase of open tundra-steppe with Juniperus, Salix, Pinus, Betula and some Alnus and Picea, followed by a succession of Juniperus and Salix during the Betula Zone (ESB3) and the expansion of Pinus and Picea during the local fen growth Zone ESB4, which is characterised by the occurence of Larix pointing to continental climatic conditions. Possibly three lithologically corresponding Late Elsterian interstadial phases have recently been identified in core sediments taken in the neighbouring Aller valley near Morsleben (Saxony-Anhalt), an exploration site for nuclear waste disposal, by Elsner (2003). 28.2.2 Holsteinian (Cycle I) During the initial phase of the brown coal excavations, limnic and telmatic deposits of a channel (Cycle I, Fig. 28.2) overlying the Elsterian sequence have been exposed adjacent to the Eemian Channel (Cycle V) and stratigraphically well positioned in relation to the Holocene (channel and Cycle VI) (Fig. 28.5). Because of its palynological composition in comparison with other sites in NW Germany (amongst others Munster-Breloh: Mu¨ller, 1974b; Hamburg-Dockenhuden: Linke and Hallik, 1993; Bossel: Mu¨ller and Ho¨fle, 1994), Urban et al. (1991b) assigned Cycle I to late parts of the Holsteinian interglacial. The terminal phase of the Holsteinian (Fig. 28.3) contains Abies, Pinus, Picea and Pterocarya as well as the water fern Azolla filiculoides, which is restricted to Early and Middle Pleistocene interglacials in

420

MITTELPLEISTOZÄN (SPÄT-) ELSTER OFFLEBEN-INTERSTADIAL I

STA- OFFLEBEN DIAL INTER?

ESBECK-STADIAL

STAD II

SAALE i. w. S.

HOLSTEIN

ESBECKINTERSTADIAL

BUSCH HAUS A

MISSAUE I I/II II

BUSCHHAUS B

INTERSTADIAL SU A

STADIAL SU A

HIATUS

Brigitte Urban

OF4/5 OF 1

OF 2

OF3

ESB 3 OFII1

ESB 1

ESB 2

ESB 4 ESB 5

Fig. 28.3 Elsterian and Holsteinian pollen sum curves.

SU8 SU9 SU10

SU 3 SU 1

SU 2

SU 4

SU 5

SU6 SU7

SU 11

SU 12

SU 13

Interglacial Pollen Records from Scho¨ningen

Fig. 28.4 Limnic telmatic sediments of Late Elsterian Esbeck interstadial overlaying silt and niveofluviatile sand.

Central Europe (Urban, 1997). A strong cooling is recorded by a significant increase of Artemisia and grasses during the following Buschhaus A Stadial, which is considered to mark the onset of the Saalian Complex sensu lato (penultimate glacial complex) (INQUA SEQS, 1992). That stadial period is followed by a twofold temperate phase, the Missaue I and II interstadials, characterized by pine, birch and some spruce. Buschhaus B Stadial had a steppe environment with dwarf shrub tundra and was followed by another temperate phase, interstadial SU A, with Pinus being the dominant tree genus (Fig. 28.3). Cycle I sediments are not recorded from the southern mining area of the Scho¨ningen open-cut mine, but here evidence for a Late Elsterian fire place of Homo erectus has been found, which is situated discordantly below gravel and Channel II (Cycle II) sediments. Richter (1998) dated a burned flint of the fire place by thermoluminescence (TL) to 450 40 kyr.

421

Fig. 28.5 Left (position of mining device): Cycle VI (late glacial and Holocene deposits composed of peat, alluvial loess, Tschernozem). Middle Part: Cycle V (Eemian calcareous tuff and peat and Early Weichselian loess palaeosol sequence). Right: Channel I (Holsteinian and successive early Saalian interstadials: peat and silty basin deposits). Photo from Thieme and Maier, 1995 (modified).

28.2.3 Reinsdorf (Cycle II) Channel II (Figs. 28.2, 28.7) is filled by sediments of the Reinsdorf Interglacial, a new biostratigraphic unit between the Elsterian and the Saalian sensu strictu (Urban, 1995a) and further interstadials and stadials. The sediment sequence of this Cycle II contains a series of five levels (levels 1–5, Cycle II-1 to Cycle II-5) represented by peat and organic, silty and calcareous muds, in places extremely rich in molluscs (Mania, 1998). These lacustrine sediments of Cycle II have been found to occur at archaeological sites Scho¨ningen 12 and 13 (Thieme et al., 1993; Thieme and Mania, 1993; Thieme, 1996, 1997, 1999; Urban, 1999), marking the extension of the basin (channel), which had a width of at least 1000 m. Recent investigations give evidence for at least 13 local pollen assemblage zones (LPAZ). Eleven of them are presented in

422

JUNGPLEISTOZÄN SAALE i. e. S.

SAALE i. w. S. BUDDENSTEDTINTERSTADIAL I

ELM. BUDDENSTEDT- ELMSTAD B INTERSTADIAL II STAD C

DRENTHE/ WARTHE

WEICHSEL

EEM I

II

III

FRUH-

BIOSTRATIGRAPHIE BIOSTRATIGRAPHY

III

TIEFE/DEPTH (cm) LITHOLOGIE/LITHOLOGY THERMOPHILE GEHOLZE THERMOPHILOUS TREES ABIES FIR

BAUMPOLLEN ARBOREAL POLLEN

? POAGEAE GRASSES

ARTEMISIA WORMWOOD

UBRIGE TERRESTRAISCHE KRAUTER FURTHER TERRESTRIAL HEADS

CYPERACEAE

ERICACEAE

S1 S3 S2 S4 S5 S6 S7

S8

EA1

BI1

EB1

BII1

BII2

EC1

Fig. 28.6 Scho¨ningen and Eemian pollen sum curves.

T1a T1b T2a T2b

T3

T4

T5

T6

SFW1

LOCAL POLLEN ASSEMBLAGE ZONES

Brigitte Urban

ESBECK/SCHÖNINGEN

ELM-STADIAL A

SCHÖNINGEN

Interglacial Pollen Records from Scho¨ningen

423

st ra

cl e cy en

ni ng

tig ra II-4c/5 S ph ch y ö

Middle Pleistocene

Saalian Complex s. l.

II-4b II-4a II-3

Reinsdorf stadial C Bio

Reinsdorf interstadial B Reinsdorf interstadial A

Reinsdorf stadial A

II-1

Reinsdorf interglacial

R 3a

R 3b

R 4/5

RS I1RS I2

0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 105 110 115 120 125 130 135 140 145 150 155 160 165 170

Reinsdorf stadial B

RI II2 RI I1

RI I2 RS II RI II1

0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 105 110 115 120 125 130 135 140 145 150 155 160 165 170 175 180 185 190

II-2

y Th Aberm i Ar es oph bo il r e ous al tr G p ra ol ees s le n Ar ses te m Fu is rth ia er C yp terr e er ac stri ea al he e Er rb ic s al es RS III1 LP AZ

lo g

th D ep 0 5 10 15 20 25 30 35 40 45 50

Li th o

(c

m )

Schöningen 13/94 profile 4 13/96 profiles 1, 2, 3

20

40

60

80

100

Silt

30 10 %

%

CaCO3

%

Peat

Loam

Molluscs

*Throwing spears

Fig. 28.7 Pollen sum curves of the Reinsdorf sequence.

this article on new pollen diagrams (Fig. 28.7, Figs. 28.9–28.12) that show a fivefold division of the middle and upper part of the interglacial and a sequence of

five climatic oscillations subdivided into eight LPAZ following the interglacial (Table 28.1). From the relative high values for grasses and herbs in the inferred forested

424

Brigitte Urban

Table 28.1 Pollen zones of the Reinsdorf sequence (Cycle II, levels 1–5) Local pollen assemblage zones (LPAZ)

Characteristic floral elements

Biostratigraphy

Schöningen cycle

? RS III 1

? Herbs– Poaceae– Betula Pinus– (Picea– Betula, cf. Larix)

? Reinsdorf Stadial C

II- 5 II- 4c/II-5

RI II2 RI II1 RS II RI I2

Betula–Herbs (cf. Larix) Juniperus– Poaceae–Herbs (cf. Larix) Pinus– Poaceae–Herbs (cf. Larix)

RI I1 RS I 2

Pinus–(Piceae– Alnus)–Herbs Poaceae–Herbs Cyperaceae–

RS I 1

Salix–Calluna

R 4/5

Pinus– Ericaceae

II- 4b Reinsdorf Interstadial B * thro wing spears made from spruce

Reinsdorf Stadial B

II- 4a II- 3

Reinsdorf Interstadial A Reinsdorf Stadial A

II-2

II-1

R4 Carpinus– Picea–Abies

Reinsdorf Interglacial

R3/R2 Corylus–Alnus R1 Quercus– Fraxinus–Tilia Earl y phases are lacking

periods of the interglacial, a warm dry forest steppe climate can be deduced. Level II-1 represents the maximum temperature and terminal phases of the Reinsdorf Interglacial (Fig. 28.8). The vegetation succession, described by LPAZ (see as well Urban, 1995a), is as follows: a Quercetum mixtum phase, LPAZ R3a, followed by a Corylus–Alnus phase, LPAZ R3a/b, a Carpinus–Picea–Abies phase, LPAZ R3b, and a Pinus–Ericaceae zone, LPAZ R4/5

(Figs. 28.9, 28.10). The water fern Azolla filiculoides occurs frequently during the Quercetum mixtum phase. The interglacial ends with an opening of the pine-birch forest and a strong increase of grasses, terrestrial herbs and Ericales (Reinsdorf Stadial A: LPAZ RS I1 and RS I2, Figs. 28.9, 28.10). Lake shore sediments of the Reinsdorf interglacial level II-1 (Figs. 28.9, 28.10) contain abundant wood (Schoch, 1999), plant macrofossils (Jechorek, 2000), bones of

Interglacial Pollen Records from Scho¨ningen

Fig. 28.8 Exposure of Cycle II limnic telmatic sediments overlying sand and gravel, sequence II-1: Reinsdorf Interglacial. 2004.

large mammals such as rhinoceros (Stephanorhinus kirchbergensis), straight tusk elephant (Palaeoloxodon antiquus), bear (Ursus spec.), horse (Equus mosbachensis), red deer (Cervus elaphus), deer (Capreolus capreolus) and the bovide Bos primigenius (Thieme et al., 1993; van Kolfschoten, 1995). Fossil remains of small and large rodents are also quite common in level II-1 as are remains of molluscs (Mania and Mai, 2001), birds, fish, reptiles (Bo¨hme, 2000), beetles and other insects. Numerous flint artifacts, wooden tools made of fir, and bones with cut marks found in level II-1 (Thieme, 1996, 1997, 1999) indicate early human occupation of the site during Lower Palaeolithic times. Level II-1 is characterised by Acer

425

campestre, Acer tataricum, Tilia cordata, Fraxinus excelsior, Prunus spinosa, Cornus sanguinea and Crategus monogyna, elements of slightly open deciduous forest, and by Abies alba, Taxus baccata, Carpinus betulus, Sorbus torminalis, Berberis vulgaris, Sambucus nigra, Cerasus avium, Lonicera xylosteum, elements of mesophilous mixed deciduous forests as demonstrated by karpological findings (Jechorek, 2000). The occurrence of the Pannonian floral element, Linum austriacum (Jechorek, 2000), points to 1.5–2 C higher annual temperatures compared to the present. The occurence of Zannichellia palustris indicates the presence of saline water, while occurrences of Potentilla anserina, Rumex maritimus and Chenopodium rubrum may point to slightly saline soil conditions. The presence of saline conditions is supported by chemical analysis of the sediments (Urban, unpublished data). Sediments of level II-2 are calcareous muds, which, in combination with a marked increase of grasses and a predominance of herbs and shrubs, denotes cool climatic conditions of Reinsdorf Stadial A (Figs. 28.9, 28.10; Table 28.1). Level II-3 sediments are organic mud and peat representing an interstadial period, named locally Reinsdorf Interstadial A (Figs. 28.10, 28.11). The vegetation was dominated by Pinus, with Betula, Alnus, and a few Picea (LPAZ RI I1 and RI I2). Trees indicative of a warm climate are almost absent. Level II-4 (Figs. 28.11–28.13) contains two stadials separated by an interstadial (Reinsdorf Stadials B and C and Reinsdorf Interstadial B), which comprise a transition into level II-5. Sediments are calcareous mud, organic mud and peat. The onset of climatic deterioration in the lower part of level II-4 is reflected by a dominance of herbs indicating a steppe environment (level II-4a, Reinsdorf Stadial B: LPAZ RS II). The upper part of level II-4 is characterised by a Pinus forest, with Betula (level II-4b, Reinsdorf Interstadial B, LPAZ: RI II1 Betula–Pinus zone and RI II2 Pinus–Betula

426

Schöningen 13/96 profile 1, x 668.00 m y 2.00 m Trees

Shrubs

Ericales

Grasses + Herbs pe

bs

r he

(m al th s s es tri p s rub cale ss res us e e a r i n D Tre Sh Er Gr Te Pi

ta e la /m eo jor c e a n p ty ype la m lia le t agotagoisia sum a t a re c an an tm tal Ce Se Pl Pl Ar To

AZ LP RS l2

99.35

s x ru a ies ari ipe ula us e L t n c n Pi Ab cf. Ju Be Al

pe ya ix ty ae lix m ar a us el a e m r e c h s t s e a u s s u c is o y c la s s u ce te etr na ini ra ea lu cu s in nu lu s elt er str bu gu a ulu ra rp ilia ory uer lmu alix cer raxi opu agu f. C . Pt f. O am ran yric um ede x rica rica mp allu acc ype oac C Q U S A F P F c cf c S F M H H lle E E E C V C P T Ca

99.26

RS l1

99.15 99.06 98.94

ra

st

o Bi

cle

y ph ra tig

cy

g

in

ön

en

h Sc

ll-2

L)

AS

ty

Reinsdorf stadial A

a di

98.55 98.45 98.35 R 3b

98.26 98.15 98.05

98 97.85 97.65 R 3a

97.55 97.45 97.36 97.26 20 40 60 80100

%

20 40 60 20 20 20 40 20 % % % %% % % %

20 20 20 40 % % % % % % % % % %% %% % %% %% % %%% % % % % %%% %

Fig. 28.9 Cycle II-1, II-2 (Reinsdorf Interglacial, Reinsdorf Stadial A, profile 1, Scho¨ningen 13/96).

500 1000 number

Brigitte Urban

98.65

ll-1

98.76

Reinsdorf interglacial

R 4/5

98.85

RS 12

102.52 102.27

RS 11

101.99 101.76

cy

hy

102.74

en ng ni hö

II-3 II-2

103.00

101.48 101.25

R 4/5

100.99 100.75 100.50 100.26

Interglacial Pollen Records from Scho¨ningen

R11

103.25

II-1

ae e pe ce ty o ia cea eae p e ula ae ula m a a a y i c s al ag i ri ac -t e n ce nd su re ant tem cho ter ter iac nu sa lpe tal Ce Pl Ar Ci As As Ap Ra Ro Fi To

cle

Terrestrial Herbs tig ra p

Grasses

Sc

he

pe e pe ty ae ty e typ m um s pe a e m . s y s a s f u s u m c t u l s u r s ae c a u e a r a i u u ix a s ar ibu pe ulu x ta rylu us xin erc us a rpin us ltis yria ngu rica der mul cac Eric pet llun cin pera ce a c g i . a u ce ie . L . V ni p li tu n li x a y Fr Q Um Ti Ca Fa Ce O Fr M He Hu Er cf Em Ca Va Cy Po Pi Ab cf cf Ju Po Sa Be Co Al

tra

l s s ria le se st s ir ca ras erre inu E G T P

Ericaies

os

r

sh

Shrubs

Bi

h pt es De Tre 103.49

d an

Trees

Reinsdorf interstadial A

A

s rb

Reinsdorf stadial A

(m

s ub

Reinsdorf interglacial

) SL

LP AZ

Schöningen 13/96 profile 2 x 653.00m y -988.00m

99.98 99.75 20 40 60 80 100 %

20 40 60 80 % % % % % % % %

20 % %

20 % % % % %

20 % % % % % % % % % % % % %

20 %

20 40 % % % % % % % % % % %

500 1000 number

Fig. 28.10 Cycles II-1, II-2, II-3 (Reinsdorf Interglacial, Reindorf Stadial A, Reinsdorf Interstadial A, profile 2, Scho¨ningen 13/96).

427

428

Schöningen 13/96 profile 3, x 729,00 m y –993,00 m le

RS II

102.00 101.86

RS II

101.62 101.50

RI 2

101.43 101.37 101.25 101.11

RI 1

100.99 100.87 100.75

en ng ni hö Sc

100.62 100.50

20 40 60 80100 20 40 60 80

%

%

%% %%%%

20 40 20 % % %% % % % % %% % % % % % % % % %

20 % % % % % % %% % %% % %

20 40 % %%

500 1000 1500

number

Fig. 28.11 Cycles II-3, II-4a, II-4b (Reinsdorf Interstadial A, Reinsdorf Stadial B, Reinsdorf Interstadial B, profile 3, Scho¨ningen 13/96).

Brigitte Urban

101.74

ll-4b

102.10

ll-4a

102.20

ll-3

RS l2

102.30

cy c

hy tig ra p

102.35

tra

s x ru s a s ari pe ulu la ce bie f. L uni op alix etu i P A c J P S B

os

l s ia le ses str s ica as re u Er Gr Ter Pin

pe ty e e e . e ea ea p ae e pec a/m e yp ae ty pe ac c um p t ce y t e e m aty o s m ia odi ylla em m cea ae la e m . s s p c e ae dul um a ty tru a iu a f la s us ia ag go is op ph th tru ria ce s ae ncu u u a l c e e l u r u u e n n a n e n c m a a l a n c s s i i s u c re ant nt tem en ryo lia alic cho ter liu iac nu ry nus xin er u lia rp gu ltis ang ric m ica ica pe llu cc ype ac sa lipe tal Ce Pl Pla Ar Ch Ca He Th Ci As Ga Ap Ra Co Al Fra Qu Ulm Ti Ca Fa Ce Fr My Hu Er Er Em Ca Va C Po Ro Fi To

Bi

bs

r he

Reinsdorf interstadial B

d

an

bs

ru

sh

Terrestrial herbs

Reinsdorf stadial B

102.40

es

e Tr

A

Grasses

Reinsdorf interglacial A

h

pt

De

(m

Ericales

LP AZ

Trees and shrubs

) SL

20 40 60 80 % % % % % 20 % % % % % % % % 20 % 20 % % % % % % % % % % % % 20 % % %

II-4c/5

II-4b

40

Reinsdorf Stadial C

20

Reinsdorf Interstadial B

10

Interglacial Pollen Records from Scho¨ningen

20 40 60 80 100 %

L)

e

cl

cy

y

ph

en

ng

ni

hö

Sc

ra

tig

tra

os

Bi

AZ

LP

C o Al rylu nu s C s ar Fr pin a u M ngu s yr la Er ica ic C ace al a l C una e yp er ac ea Po e ac e C ae er e Pl alia Ar ant typ te ag e C mis o m h ia a jo Th eno r/m a p ed C lictr odia ic u c ia ty As hor m ea pe e te iac C ra ea en ce e G tau ae al r R ium ea um Ap ex ia R ce an ae u R ncu os la Fi ace cea lip a e e To en ta du l s la um ty pe

Pi ce La a rix Ju n Sa ipe lix rus Be tu la

Te

Grasses

RS III1

sh

AS

Trees and shrubs

RI I2

d

an

m

(c

ru es bs rre s Pi t nu rial he s rb s

ss

ra

G

s

ee

Tr

th

ep

D

Schöningen 13/94 Profile 4, x 685.00m y 38.00m

Terrestrial herbs

1

50

200 400 number

Fig. 28.12 Cycles II-4b, II-4c/II-5 (Reinsdorf Interstadial B, Reinsdorf Stadial C, profile 4, Scho¨ningen 13/94). 429

430

Brigitte Urban

tartaricum, Acorellus pannonicus, Thymelaea passerina, Viola cf. alba, Ranunculus brutius, Ranunculus lateriflorus A. Dc. vl. nodiflorus L., Dychostylis micheliana and Potamogeton vaginatus. In agreement with the palynological findings, a warm climatic forest steppe type of vegetation indicating rather dry regional conditions can be inferred for the main part of the Reinsdorf Interglacial. Preliminary 230Th/234U age determinations of peat layers within the Reinsdorf interglacial deposits (level II-1) at site Scho¨ningen 12 (Heijnis and Urban, 1995) gave an approximate age of 320 kyr. 28.2.4 Scho¨ningen (Cycle III) Fig. 28.13 Cycle II, sequence II-4b: Reinsdorf interstadial B (strata of throwing spears) and sequence II-4c/II-5: Reinsdorf stadial C (cryoturbation of top sediments) 2004.

zone) (Figs. 28.12, 28.14). Pollen of Picea is very rarely found in Reinsdorf Interstadial B deposits. Jechorek (2000) identified macrofossils of Arctostaphylus uva-ursi, Carex aquatilis, Frangula alnus, Pinus sylvestris, Lonicera xylosteum, Rubus ideaus and Salix spec. in level II-4 sediments. Level II-4b also yielded numerous faunal remains, mainly horse, flint artifacts and well-preserved spears made from spruce, the oldest hunting spears so far discovered (Thieme, 1999). The pollen assemblage of level II-4b (Figs. 28.11, 28.12) contains only very rare Picea pollen, which might suggest transportation of the hunting weapons from scattered stands of spruce some distance from the site. Between level II-4c (Reinsdorf Stadial C, LPAZ: RS III1, nonArboreal zone) and II-5, within silty clays, frost structures occur and mark the onset of a periglacial environment and the definite end of cycle II (Fig. 28.13). In summary, the karpological investigations of the Reinsdorf sequence (Jechorek, 2000) reveal remains of 132 plant species including seven thermophilous exotics: Acer

Channel III contains a sedimentary sequence (Cycle III) that cross-cuts Channel II (Figs. 28.2, 28.6). The sequence is composed of silty muds and peats and represents the Scho¨ningen interglacial (Urban et al., 1991b; Urban, 1992, 1995a). The pollen assemblages are indicative of a warm, (sub)continental climate with high percentages of Pinus and Tilia with some Quercus. High components of Alnus found almost throughout the entire profile point to swampy environments. A Carpinus phase with Picea occurs near the end of the warm period. Abies is absent, apart from a single grain, while massulae of the water fern Azolla filiculoides are abundant in the Alnus-rich parts of the sequence. The Scho¨ningen interglacial is succeeded by the Elm A Stadial which is characterised by a marked increase of herbs and moderate increases of Artemisia, grasses and Ericales (Fig. 28.6). The Elm A Stadial is followed by two temperate periods, the Bu¨ddenstedt I and II Interstadials, with pollen assemblages indicative of Pinus–Betula forests. The Bu¨ddenstedt interstadials are separated by the Elm B Stadial. Increases of Betula and herbs mark the onset of another cold spell, the Elm C Stadial which caps the sequence (Fig. 28.9). The interglacial and stadial– interstadial sequence of Channel III (Cycle III) is overlain by glaciofluvial sands and till

Interglacial Pollen Records from Scho¨ningen

Holocene Early

Middle

Weichselian

~30.000

Eemian

~115.000

~130.000

Warm desiduous forest

Boreal forest Steppe forest

Shrub tundra

Detailed subdivision of climatic 230 units of the upper Quaternary of Schöningen TH/U * cycle Central Europe. TL ** dating

Tentaive correlation

MIS

kyr

basically adapted from the Schöningen record

Subatianticum Subboreal Atlanticum Boreal Preboreal

Late

~10.000

Tundra

Years BP

Steppe

Palaeoecological evolution Polar desert

subdivision of the Quaternary

Glaciated landscape

Time scale

431

1 C VI

Late Dryas Bølling / Allenød Early Dryas

2

Denekamp Hengelo

3

Moershoofd Glinde Oerel Odderade

4 5a

Brörup

5c CV 115 – 149*

Eemian interglacial

5e

Late

Hiatus Warthe

C IV

Glaciation

6

Drenthe Hiatus Elm C Büddenstedt II Elm B

C III

Büddenstedt I Elm A

Schöningen interglacial Middle

Saalian Complex

~200.000

177 – 234*

7

prox. 320

7/9

Hiatus Level 5

Reinsdorf - Stadial C Reinsdorf - Interstadial B Reinsdorf - Stadial B Reinsdorf - InterStadial A

Level 4

C II

Reinsdorf - Stadial A

Reinsdorf - interglacial

Level 3 Level 2 Level 1

Hiatus Stadial SU A Interstadial SU A Buschhaus B

Early

Missaue I

Holsteinian

Missaue II Buschhaus A

CI

Holsteinian interglacial

Late

Hiatus Esbeck

9/11 450 ± 40** Burn Flint

Pleni-

Glacial

Elsterian

Offleben II Hiatus Offleben I

Glaciation

Fig. 28.14 Synthesis of the Scho¨ningen pollen records, dating and tentative correlation with the marine isotope stages.

of the first Saalian ice advance (Drenthe Stadium). Peat of the Scho¨ningen interglacial gave uncorrected 230Th/234U ages of 180 and 227 kyr (Heijnis, 1992). Based on

the pollen record, correlation has been made with the Wacken (Menke, 1980) and Do¨mnitz (Erd, 1973) interglacials (Urban et al., 1991b; Urban, 1995a).

432

Brigitte Urban

28.2.5 Cycle IV Channel IV, which is eroded into Saalian glacial deposits, contains a pedocomplex developed in alluvial loess. Two pseudogleyic layers, presently being analysed by Altermann (personal communication) suggest the presence of at least one periglacial phase between the two major ice advances of the Saalian glaciation (Drenthe and Warthe Stadium, Figs. 28.2, 28.14), which is recorded with tundra-type vegetation and soils from the Red Cliff on the Isle of Sylt (North Western Germany) (Felix-Henningsen and Urban, 1982) 28.2.6 Eemian (Cycle V) The sequence in Channel V (Cycle V) is represented by either a Luvisol developed in loess, or, due to the influence of the palaeodrainage system, a soft travertine and peat of last interglacial age. In Scho¨ningen, interglacial sedimentation started concurrently with a Carpinus phase and continued to the first Early Weichselian interstadial (Figs. 28.5, 28.6). The Eemian peat and travertine layers are rich in Abies during the Pinus-Picea-Abies-Zone. The travertine sediments were deposited during a period of about 6000 years, deduced from the reconstructed pollen zones (Mu¨ller, 1974a). Local hydrological conditions during the late last interglacial and early glacial periods have been determined by pollen analysis and by plant macro remains, specifically moss analyses (Ho¨lzer in Urban et al., 1991a). The Eemian peaty layers reveal a thorium/uranium age of 132 17 (Heijnis, 1992) (Table 28.2, Fig. 28.14).

vegetation history, palaeo landscape and the degree of human impact on the area (Figs. 28.2, 28.5) (Thieme et al., 1987; Thieme and Maier, 1995).

28.3 DISCUSSION The sequence of Scho¨ningen gives geological and palaeoecological evidence for several temperate phases between the Holsteinian and the Eemian and reveals data on three interglacials and at least 10 interstadials between the end of the Elsterian and the beginning of the Saalian (Drenthian) glaciation (s.str.). Some authors state that, as (temperate) deposits investigated in Scho¨ningen do not occur in the same outcrop in perfect superposition, the succession of the warm stages will remain to some extent debatable (e.g. Turner, 1998). In contrast, we believe, from intensive geological field work, mapping and analysis of long geological transects since the opening of mine Esbeck in the 1980s, the superposition of stratigraphic units, the overlapping of sedimentation units which cut or underlay the strata at certain times of excavation and adjacent occurrences of channel fillings, that a stratigraphic sequence has been well established (Elsner, 1987; 2003; Hartmann, 1988; Urban et al., 1988, 1991a, 1991b; Lenhart, 1989; Tschee, 1991; Thieme et al., 1993; Mania, 1995, 1998). Moreover, palaeoecological and biostratigraphic evidence, archaeological findings and radiometric dating have resulted in a comprehensive stratigraphic scheme for the glaciated margin of Western Central Europe.

28.2.7 Holocene (Cycle VI) The youngest sedimentation cycle in the Scho¨ningen mine is comprised of Late Weichselian and Holocene deposits and soils (Channel VI sediments), and braided river deposits of the Mißaue and its tributaries. The variety of both sediments and soils has allowed a detailed reconstruction of the

28.3.1 Late Elsterian, Offleben I and II and Esbeck Interstadials The Offleben I, II and the Esbeck interstadials, which are related to late phases of the Elsterian glaciation, but not to its late glacial period, occurring on top of the youngest Elsterian till, have no known biostratigraphic equivalent

Interglacial Pollen Records from Scho¨ningen

433

Table 28.2 Tentative correlation of the Scho¨ningen sequence with the Velay record (Reille and de Beaulieu, 1995; Reille et al., 1998; de Beaulieu et al., 2001) Germany NE Niedersachsen

South Central France

Urban et al., 1988, 1991a, 1991b; Heijnis, 1992; Urban, 1995a, 1999, 2000; Urban and Heijnis, 1995

Eemian

230

Elm C Büddenstedt II Elm B Büddenstedt I Elm A Stadial Schöningen 230 Th/234U 180 and 227 kyr ? Harbke interstadial

Reinsdorf Stadial C

Le Bouchet 3 Belvezet Le Bouchet 2 Bonnefond Le Bouchet 1

(II-4/5) Charbonniers Stade (II-4)

Amergiers Interstade

(II-4)

Monteil Stade

Reinsdorf Interstadial A

(II-3)

Ussel Interstade

Reinsdorf Stadial A

(II-2)

Cayres Stade Landos

230

7?

? ?

Reinsdorf Stadial B Reinsdor f

5e

Ribains

Th/234U 132 +/- 17 kyr

Reinsdorf Interstadial B

MIS

Reille and de Beaullieu, 1995; Reille et al., 1998; de Beaullieu et al., 2001

Th/234U prox. 320 kyr?

40

Ar/39Ar 275 +/5kyr

Stadial SU A Bargette Stade Interstadial SU A Jagonas Interstade 2 Buschhaus B Stadial Pradelle 2 Missaue Interstadial (Missaue I, II) Jagonas Interstade 1 Buschhaus A Stadial Holsteinian Praclaux

elsewhere in Northern Germany and Western Europe. Attention should be paid to their relevance for correlation with the marine isotope record as they are preceding the classical Holsteinian interglacial and might be of significance for Late Elsterian climatic evolution. 28.3.2 Holsteinian, Missaue I and II Interstadials, Buschhaus B Stadial and Interstadial SU A In Scho¨ningen late phases of the Holsteinian interglacial and two following interstadials are documented (Urban et al., 1991a, 1991b; Urban, 1996a, 1996b). The terminal phases of the Holsteinian, which contain Abies, Pinus, Picea, Pterocarya and water fern Azolla filiculoides, have been

7/9 ?

9/11 ?

correlated with pollen zones XII, XIII and XIV (Mu¨ller, 1974b; Meyer, 1974). A strong cooling at the end of the Holsteinian marks the onset of the Saalian s.l.. This event, the Buschhaus A Stadial of the Scho¨ningen sequence, can be correlated with the Fuhne A Stadial at Pritzwalk/Prignitz (Erd, 1973). The twofold Mißaue interstadials I and II are equivalent to the Dockenhuden interstadial (Hallik and Linke, 1986; Urban, 1996a) and to the Pritzwalk interstadial A/B (Erd, 1973; Erd et al., 1987). In comparison with investigations at Bossel (northwest Niedersachsen), a site with sediments of the Holsteinian marine transgression (Mu¨ller and Ho¨fle, 1994), the early part of the twofold Mißaue interstadial may be correlated with zone XVI/XVII.

434

Brigitte Urban

Looking at the occurence, distribution and succession of certain taxa (Carpinus/Fagus, Pterocarya, Abies and Azolla filiculoides) in late Holsteinian sequences, not only of areas adjacent to the North Sea Basin (Zagwijn, 1973) but also of lower European latitudes, including the long sequence of the Velay (de Beaulieu et al., 2001) and La Cote in Vercors, France (Field et al., 2000), the sites of Samerberg II in Bavaria (Gru¨ger, 1983), Thalgut in Switzerland (Welten, 1988) and Krepiec, Zbojno and Losy in Poland (Lindner and Marciniak, 1998) are of significance. Long-distance correlation of the Praclaux interglacial with the Holsteinian has been proposed by de Beaulieu et al. (2001), and of the Mazovian by Lindner and Marciniak (1998), suggesting it a probable terrestrial correlative of stage MIS 11. Referring to Sarnthein et al. (1986), who 230 Th/234U and ESR dated marine molluscan shells from para-type and other Holsteinian interglacial deposits to > 350 and 370 kyr, Holsteinian beds have been correlated with MIS 11 or even an older interglacial event. There is ongoing debate on the exact age of the Holsteinian interglacial. Geyh and Mu¨ller (2005) recently presented 230Th/234U dates and a palynological review of the Holsteinian/Hoxnian interglacial. Their interpretation of 230Th/234U dates has led them to correlate the Holsteinian and the Hoxnian (Turner, 1970) with MIS 9. The authors present a brief review of correlations of the Holsteinian with MIS 11 citing mainly palynological work. Among those, their citation that, by using 230Th/234U dates of peat of the Scho¨ningen interglacial (Figs. 28.10, 28.12), Urban (1983, 1995a) ‘palynologically and indirectly correlates the Holsteinian interglacial to MIS 11’ is causing confusion and has to be corrected. As cited above, peat of the Scho¨ningen interglacial revealed uncorrected 230Th/234U ages of 180 and 227 kyr (Heijnis, 1992) and therefore can be related most appropriately to MIS 7 (Urban, 1995a). It can, therefore, not be used as an indirect correlation tool for dating the Holsteinian to MIS 11 (Fig. 28.14). There is further evidence that the Ka¨rlich interglacial sequence (Urban, 1983; Bittmann,

1992) might reveal two different interglacials (Urban, manuscript in preparation). Urban (1983), based on the palaeofloristic record, tentatively proposed a post-Holsteinian age for the Ka¨rlich interglacial s.str. Bittmann has correlated the Ka¨rlich interglacial sequence palynologically to the Bilshausen interglacial (Mu¨ller, 1965; Bittmann and Mu¨ller, 1996) and an 40Ar/39Ar age of about 400 kyr obtained from a tephra layer predating the Ka¨rlich interglacial (van den Boogaard et al., 1989: 396 þ 20 kyr) to the Cromerian (V). Refering to recent dating, van den Boogard (in Boenigk and Frechen, 1998) relates the eruption of the ‘Ka¨rlicher Brockentuff’ to the beginning of MIS 10. Boenigk and Frechen (1998) place the Ka¨rlich interglacial s.str. (Urban, 1983), by correlation with sequences from the Lower Rhine area, into the Saalian s.l., a correlation which had already been suggested by Urban (1983). The Holsteinian deposits of Scho¨ningen have not been dated so far. There is a TL date of a burnt silex from a prehistoric fire place (Richter, 1998) in Late Elsterian/Early Holsteinian deposits (Urban, unpublished data) available for the Scho¨ningen mine, revealing an age of 450 40 kyr (Fig. 28.14). The pollen record of those deposits point to an open tundra-(taiga) environment with pine and birch and indicate late glacial Elsterian environmental conditions. As there is dating and palaeoecological evidence of interglacial sequences following the Holsteinian, attention will be focussed on these superimposed strata of mine Scho¨ningen and their biostratigraphic correlation and preliminary dating. 28.3.3 Reinsdorf Interglacial, Reinsdorf Stadial A, Reinsdorf Interstadial A, Reinsdorf Stadial B, Reinsdorf Interstadial B, Reinsdorf Stadial C, Harbke Interstadial and Scho¨ningen Interglacial The term Reinsdorf Interglacial was introduced by Urban (1995a) for a new interglacial sequence at Scho¨ningen of post-Holsteinian and pre-Drenthe age (first

Interglacial Pollen Records from Scho¨ningen

Saalian ice advance), which yielded abundant fossil remains as well as archaeological evidence for the presence and activities of Homo erectus (Thieme and Maier, 1995). The archaeological site is still exposed, and the excavation and research continues under the supervision of Hartmut Thieme (Archaeological Survey of Lower Saxony, Hanover). As stated earlier, the Reinsdorf sequence contains interglacial and stadial–interstadial floras (Table 28.1) that are quite distinct from the Holsteinian (Urban, 1995a, 1995b, 1999). The main characteristics of the Reinsdorf Interglacial are a climatic optimum characterised by a forest phase with the spread of Tilia before Corylus; which is only represented by low values, and by the occurrence of a late and less pronounced Abies phase. Furthermore, the Reinsdorf sequence is characterised by two pronounced interstadials interrupted by phases of climatic deterioration (stadials), when the vegetation opened up to grass and herb-rich steppic environments (Table 28.1, Figs. 28.9–28.12, 28.14). Compared to the Holsteinian vegetation reconstructed from the same outcrop, a warm and distinctive continental regional type of climate can be inferred for the Reinsdorf Interglacial, indicated, for example, by occurrences of Acer tartaricum and Larix, as well as by low representation of Corylus. Larix also occurs during the interstadial and stadial phases. These climatic interpretations are supported by the mollusc assemblages (Mania and Mai, 2001). A 16-m profile covering the biostratigraphic units of the Reinsdorf sequence at the excavation site is presently under investigation. Thermal ionization mass spectrometry (TIMS) 230Th/U dating of peat taken from this profile is currently in progress (Frechen et al., this volume). The Scho¨ningen interglacial fen peat deposits (Fig. 28.6) contain a distinctive vegetational succession, dominated by Pinus and Alnus throughout the entire thermal part of the sequence. The Scho¨ningen

435

Interglacial terminates abruptly and is succeeded by a taiga-tundra type of vegetation and then two short boreal conifer phases. Urban (1995a) already discussed the main differences between the pollen zones, marker species and local as well as regional environments in the Holsteinian, Reinsdorf and Scho¨ningen and the probable correlatives of the latter, the Wacken (Menke, 1980), Do¨mnitz (Erd, 1973) as well as the Eemian interglacial floras in the Scho¨ningen mine in great detail. It was concluded that the Reinsdorf and Scho¨ningen interglacials differ strongly from each other in the following features which have proven to be of biostratigraphic value for long-distance correlation: 1. Absence of Abies (Scho¨ningen), 2. Occurence of an Abies phase during the Carpinus–Picea phase (Reinsdorf) 3. Tilia peaks during the Corylus and early Carpinus phases (Scho¨ningen) 4. Expansion of Corylus with low values and only after that of the Mixed-oak forest phase (Quercetum mixtum, QM), dominated by Tilia (Reinsdorf) 5. Reinsdorf interglacial characterised by a warm, continental regional forest steppe climate 6. Dominance by Alnus and Pinus in all observed pollen zones (Scho¨ningen). 7. An abundance of Pinus but only local importance of Alnus reflecting the moist hydrological conditions of the stands (Reinsdorf) There are only few sites with records of similar age or of similar biostratigraphic significance to Scho¨ningen from neighbouring areas. However, the Holsteinian and postHolsteinian deposits from Morsleben and Ummendorf (Aller Valley, Saxony-Anhalt) described by Strahl (1999), which are located less than 10–15 km distance of the Scho¨ningen mine with comparable geological, geogenetical origins, are of special interest as they might contain equivalents of the Reinsdorf Interglacial. Strahl (1999) has

436

Brigitte Urban

considered the Aller interglacial, which follows the Holsteinian, the deposits of which are overlain by the Morsleben Stadial A, Morsleben Interstadial B and Morsleben Stadial C, a most probable time correlative of the Reinsdorf, Wacken (Menke, 1968) and Do¨mnitz in the profile of Pritzwalk/ Prignitz (Erd, 1973). As stated earlier, the Holsteinian at Scho¨ningen is followed by the Buschhaus A stadial, which is considered to mark the onset of the Saalian Complex s. l. This is followed by a twofold temperate phase, the Missaue I and II Interstadials and the Buschhaus B Stadial characterised by a steppe environment which was followed by another temperate phase, Interstadial SU A (Figs. 28.3, 28.14). Though in Morsleben only one interstadial (Morsleben Interstadial B) is recorded, which might point to some discordances/hiatus between the post Holsteinian interstadials and the deposits of the Aller interglacial in that section, it has certain similarities with the Reinsdorf interglacial. Strahl (1999) records four pollen zones of the Aller interglacial, including a Carpinus–Abies–Picea–Alnus zone, which have correlatives in the Reinsdorf interglacial pollen zones (R3–R5). Unfortunately, only descriptions of pollen zones without diagrams have been published by Strahl so far, prohibiting more detailed comparison. The correlation of the Aller interglacial and the Wacken and Do¨mnitz interglacials still remains rather tentative as they differ palaeobotanically strongly from the Aller and Reinsdorf interglacials. There is no evidence from the Morsleben site of two interglacials following the Holsteinian and preceding the first Saalian ice advance. However, two interglacials are recorded by Mu¨ller (personal communication, 1999 and 2005) from pit Nachtigall (Niedersachsen). Mu¨ller found evidence of a first interglacial, which is rich in Tilia and characterised by a late Abies phase immediately following the Holsteinian and the early Saalian interstadials described in Scho¨ningen (Urban, 1999), and of a

subsequent interglacial, which is lacking Abies and which is dominated by Alnus and Pinus. In summary, it is most likely that the earliest interglacial of pit Nachtigall is a correlative of the Reinsdorf Interglacial and the second one probably of the Scho¨ningen Interglacial. Another probable botanical equivalent might be found in the cores drilled in Go¨ttingen recorded by Gru¨ger et al. (1994). The diagrams b, c and d of Go¨ttingen Ottostraße are reflecting an interglacial sequence, which is interpreted by Gru¨ger et al. (1994) as representing three interglacials interrupted by stadials and followed by interstadials. The whole sequence is characterised by several hiatuses and other disturbances, which is reflected by ‘noise’ in the pollen record. If the diagram zones DA 1–17 (after Gru¨ger et al., 1994) are synthesised, taking into account the hiatuses and some probable sediment contamination including the solifluction layer, the initial phases of one interglacial and at least one following interstadial could be identified as having definite similarities with the Reinsdorf interglacial sequence. The pollen diagram of the lacustrine deposit of Elsterian to Saalian age (Behre, 2004) at Surheide near Bremerhaven spans the lower and middle parts of an interglacial which is characterized by a pronounced peak of Corylus, a late maximum of Abies and a lack of Carpinus. Behre consequently suggests correlation with pre-Holsteinian rather than Holsteinian or early Saalian sequences of intra-Elsterian age, focussing on the Ferdinandow interglacial (JancykKopikowa, 1975; Jancyk-Kopikowa and Zarski, 1995). The interglacial deposits of Surheide do not have a correlative in the Scho¨ningen sequence at the present state of investigation. At Ro¨persdorf (Erd, 1987), the terminal part of the recorded interglacial is almost lacking. However, based on the pattern of spread and dominance of Tilia and Corylus, ¨ cker interseveral authors correlated the U glacial from Ro¨persdorf with postHolsteinian and pre-Eemian interglacial

Interglacial Pollen Records from Scho¨ningen

deposits. Based on recent results obtained within a drilling project of the Geological Survey of Brandenburg (Landesamt fu¨r Bergbau, Geologie und Rohstoffe Brandenburg) in the area of Prenzlau, Hermsdorf and Strahl (2005) point out major disturbances of the lithological setting of the interglacial deposits of Ro¨persdorf. Similar to Erd (1970), they correlate the interglacial deposits of Ro¨persdorf with the Eemian interglacial. Mania (1998) correlates the Reinsdorf sequence with Bilzingsleben II, which he considers to date to 400 kyr based on ESR and 230Th/234U determinations. Consequently, he relates it with MIS 11 and the Scho¨ningen Interglacial with Bilzingsleben III, which he correlates with the Do¨mnitz interglacial (Erd, 1973), citing an 230 Th/234U age of 320 kyr for the latter interglacial, which he therefore relates to MIS 9. As there are no comparable longer pollen records available for the travertine sequence of Bilzingsleben, a correlation with the Scho¨ningen sequence in our opinion is rather tentative. The Reinsdorf interglacial has been correlated by Urban (1995a) with the Zbojnian interglacial in Poland (Lindner and Marciniak, 1998), situated some 500 km east of Scho¨ningen. The Zbojnian interglacial is intercalated between two stadials that follow the Mazovian interglacial which is considered the equivalent of the Holsteinian in Poland. Correlation has been based on the great similarities of the climatic optimum characterised by Tilia dominance during the QM zone and the spread of Tilia before Corylus. Late, terminal phases of the Zbojnian interglacial are characterised by an Abies phase with lower values compared to the Mazovian interglacial, which is identical with the Abies distribution during the Holsteinian and Reinsdorf observed at Scho¨ningen. De Beaulieu et al. (2001) have correlated the Landos interglacial with the Zbojnian, ¨ cker (Erd, 1987) the Reinsdorf and the U Interglacials due to its stratigraphic position below the Bouchet 1 interglacial which is almost lacking Abies. Consequently, they

437

correlate Bouchet 1 with the Scho¨ningen, Wacken (Menke, 1968) and Do¨mnitz (Erd, 1970) interglacials. It should be noted that long-distance correlation based on vegetation changes and similarities inferred from pollen data and other plant remains between sites with latitudinal, altitudinal and edaphical differences have to take into account dissimilarities in occurrences of taxa or their relative representation (de Beaulieu et al., 2001). Considering these uncertainties, these authors state that marker taxa such as Pterocarya are very important. Fagus is also important in that it plays a key role in both the Holsteinian and in younger interglacials in the southern part of Central Europe and in contrast to Northern Germany where it appears with very low values during late phases of the Holsteinian, co-occuring with Pterocarya. With reservations and pointing out the tentative character of our correlation with the sequence of Velay (de Beaulieu et al., 2001) and La Coˆte, Val-de-Lans basin, France (Field et al., 2000), best-estimate comparison of Scho¨ningen with those long terrestrial records for the Middle Pleistocene of Central Europe is presented in Table 28.2. This palynological correlation is supported by the 40Ar/39Ar age of 275 5 kyr of the Armargier interstadial following the Landos interglacial that is tentatively correlated here with the Reinsdorf Interglacial. A recent reinvestigation of the Meikirch pollen record (Welten, 1982, 1988) in the Swiss Alpine Foreland by Preusser et al. (2005) has led to a reinterpretation of its correlation. Whereas Welten (1982, 1988) correlated the youngest of the three interglacial phases with the Eemian and the two older, Holstein 1 and Holstein 2, with the Holsteinian sensu strictu and the Do¨mnitz/ Wacken/Hoogeveen (Menke, 1968; Erd, 1970; Zagwijn, 1973) (Fig. 28.1), Preusser et al. (2005) favour a correlation of the entire Meikirch complex with MIS 7 mainly based on luminescence dating and comparison with marine climate records. In summary,

438

Brigitte Urban

the pollen record reveals three more or less complete interglacial phases of different climatic character, Meikirch 1, Meikirch 2 and Meikirch 3, which are separated from one another by the Birchli stadial, situated between Meikirch 1 and Meikirch 2, the latter followed by the long period of the Chutzen stadial. The Chutzen stadial is characterised by several colder and milder periods, Chutzen 1 by an open tundra, Gra¨chiswil 1 interstadial by a spread of Picea and Pinus, Chutzen 2 and Chutzen 3, the coldest phases, interrupted by the Gra¨chwil 2 interstadial which is also characterised by increases of Picea, Pinus and Betula. The Chutzen stadial is followed by the Bu¨tschwil interstadial which is characterised by an open Betula forest and remarkable amounts of Larix. It is considered to be part of the initial phase of reforestation of the Meikirch 3. Owing to the abrupt change in pollen composition, a hiatus is assumed to occur between the Bu¨tschwil interstadial and the Meikirch 3 interglacial. Concerning the interglacial sequences, Meikirch 2 is described as a temperate interglacial, Meikirch 1 and Meikirch 3 as well-developed warm periods. Preusser et al. (2005) state that Meikirch 2 might not be complete, and the climatic gradient between the Swiss alpine foreland and the Massif Central during the interglacials is not known. The sequence shows several similarities as well to the Jagonas interstadials, as well as to the Ussel and the Amargiers interstadial (Reille et al., 2000) (Table 28.2). Owing to the vegetation pattern of Meikirch 1, characterised amongst others by the lack of Fagus, a correlation with Holsteinian sequences is most unlikely. Taking all these observations as well as the unknown, perhaps even more pronounced, climatic gradient between the Alpine foreland and western central Germany into account, the Meikirch and the Scho¨ningen pollen records are tentatively compared (Table 28.3). It is not known whether the almost total lack of Abies in Meikirch 2 is due to a possible hiatus or whether it

reveals the degree of vegetation development. Its temperate character compared to that of the Meikirch 1 and 3 warm interglacials and its similarities with Middle Pleistocene interstadials of the French Central Massif, as stated above, suggests correlation with Reinsdorf Interstadials A and B. With this comparison, Meikirch 1 interglacial might correspond to the Reinsdorf interglacial. Part of the Chutzen stadial sequence might be correlated with Reinsdorf Stadial C and the cryoturbation horizon of the top sediments (Figs. 28.12, 28.13, Table 28.3). In Scho¨ningen, a peat of 1.5 m thickness topped by limnic sediments contains a pronounced interstadial, named locally the Harbke interstadial (Urban, 1996a). It has been found in a comparable lithostratigraphic position to the peat layer of the Scho¨ningen interglacial in the northern mining field Esbeck, but could not be correlated to date. The Harbke interstadial shows the following characteristics: a Betula–Pinus zone, followed by a Pinus–Betula–Picea zone and a Pinus–NAP zone with increasing amounts of Chenopodiaceae, Polygonaceae, Caryophyllaceae and Asteraceae. During pollen zone Esd V, the Pinus–Betula–Alnus– (Ericaceae–Sphagnum) zone, the growth of a local fen peat reached its climax. The Pinus–Betula–NBP zone marks the end of that biostratigraphic unit. Pollen of Larix occurs more or less continuously though with low values throughout the entire profile. If compared with the Meikirch record, the possibly truncated Bu¨tschwil interstadial might comprise part of the Harbke interstadial (Tables 28.2, 28.3). 28.3.4. Eemian (MIS 5e), Early Weichselian (MIS d, c), Late Glacial and Holocene In Scho¨ningen, the MIS 5e (Bassinot et al., 1994) equivalent travertine sediments were deposited during a period of about 6000 years, deduced from the reconstructed pollen zones (Mu¨ller, 1974a) which can be

Interglacial Pollen Records from Scho¨ningen

439

Table 28.3 Tentative correlation of the Scho¨ningen Middle Pleistocene sequence with the Meikirch record (Preusser et al., 2005) Germany NE Niedersachsen

Swiss alpine foreland

Urban et al. 1988, 1991a, 1991b; Heijnis, 1992; Heijnis and Urban, 1995; Urban, 1995a, 1999, 2002

Preusser et al., 2005

Elm C Büddenstedt II Elm B Büddenstedt I Elm A stadial Schöningen 230Th/234U 180 and 227 kyr ?

Hubel Meikirch 3 Hiatus ?

Harbke interstadial

Bütschwil

Reinsdorf stadial C

(II-4/5) Chutzen sequence

Reinsdorf interstadial B

(II-4)

Reinsdorf stadial B

(II-4)

Reinsdorf interstadial A

(II-3)

kirch 2

Reinsdorf stadial A

(II-2)

Birchli Meikirch 1

Reinsdorf

230Th/234U

prox. 320 kyr?

correlated with the Eemian and Early Weichselian site of Gro¨bern (Litt, 1990) and other Northern German Late Pleistocene sequences (Behre, 1974, 1989; Menke and Tynni, 1984; Behre and Lade 1986; Urban in Veil et al., 1992; Hahne et al., 1994; Caspers, 1997). In Scho¨ningen, the terminal part of the interglacial is represented by the Carpinus–Picea phase indicating rather uniform ‘oceanic’ climatic conditions and the following Pinus–Picea–Abies phase having a more boreal and suboceanic character in Northwestern and Central Europe (Aalbersberg and Litt, 1999). Hydrological and climatic development of the late last interglacial and early glacial periods have been determined by pollen analyses, plant macro remains and molluscs (Urban et al., 1991a). The Eemian peaty layers reveal a thorium/ uranium age of 132 17 kyr (Heijnis, 1992) (Fig. 28.14, Table 28.2). The youngest sedimentation cycle in Scho¨ningen contains late Weichselian and Holocene deposits and soils, which have allowed a detailed reconstruction of Late Glacial and Holocene

MIS

7?

Mei-

7/9 ?

physical environments and human impact (Figs. 28.5, 28.14) (Thieme et al., 1987, Thieme and Maier, 1995). 28.4. SUMMARY OF STRATIGRAPHIC ¨ NINGEN ASPECTS OF THE SCHO SEQUENCE In Northwest Europe, Middle Pleistocene post-Holsteinian sequences with a definite stratigraphical position below the Early Saalian till are rather rare. Besides the previously cited correlations based on biostratigraphic findings, several other authors have correlated the Scho¨ningen sequence with long Pleistocene records (e.g. Kukla and Cilek, 1996) and placed the Holsteinian into MIS 11 and the Reinsdorf interglacial into MIS 9. Jo¨ris and Baales (2003) attempted to correlate the Scho¨ningen sequence with the marine isotope chronology using the Vostok record (Petit et al., 1999) and came to the same conclusion. In their discussion of

440

Brigitte Urban

the stratigraphic position and the age of the throwing spears (Thieme, 1999), which the authors wrongly placed into the Reinsdorf interglacial instead of the following interstadial (Reinsdorf interstadial B of Cycle II-4; Urban, this volume; Thieme, 1999), the authors correlate the Scho¨ningen interglacial with MIS 9a, the early Saalian Drenthe with MIS 8 and the Warthe loess with stage 6. Both the 230Th/234U ages of 180 and 227 kyr for the Scho¨ningen interglacial and of 132 17 kyr for the Eemian interglacial peat (Heijnis, 1992; Urban, 1995a) suggest a correlation of the Scho¨ningen interglacial with MIS 7 rather than with MIS 9 and the Drenthe till with MIS 6. Recently, the Antarctic Vostok ice core (EPICA Community Members, 2004; McManus, 2004) provided detailed evidence of climate development over the past 420 000 years. Of particular interest is the ascertainment that the interglacial stage MIS 11, following Termination V, was about 28 000 years long. According to Meyer (1974), Mu¨ller (1974b) and Geyh and Mu¨ller (2005), the Holsteinian interglacial as determined by counting of annual diatom layers should have covered a period of about 16 000 years. In comparison, the Rhumian interglacial was determined to have been about 25 000 years long (Mu¨ller, 1992). Those observations of the particular length of the interglacial periods and botanical pattern are used as an important argument for correlation in addition to thorium/uranium dating of Holsteinian-type locality deposits by Geyh and Mu¨ller (2005). Recent approaches of direct correlation of land–sea records between terrestrial and marine climatic indicators and ice volume proxies from deep-sea core MD01 2447 (off northwestern Iberia) show that the warmest period of MIS 11 lasted about 32 000 years (426–394 kyr) and was followed by three warm/cold cycles (394–362 kyr) (Desprat et al., 2005). During those interstadial/stadial periods, deciduous forests prevailed, and heathland in transition to open grassland characterised the cold steppe-like

stadial environments. Of utmost interest in comparison with terrestrial records of early post-Holsteinian age is pollen zone MD47S2 of the second stadial between 384 and 382 kyr, which is abruptly intercalated between interstadials MD47-I2 and MD47I3. The same is observed in NW Europe, for example, at Scho¨ningen for the postHolsteinian twofold Mißaue interstadials I and II, which are equivalent to the Dockenhuden interstadial (Hallik and Linke, 1986; Urban 1996a) and to the Pritzwalk interstadial A/B (Erd, 1973; Erd et al., 1987). They all have a sudden and short cold spell in common, dividing the interstadial into two major parts. The long warm phase of MIS 11 in northwestern Iberia is named Vigo interglacial. Desprat et al. (2005), furthermore, found that the Vigo interglacial of MIS 11 in the marine pollen record off northwestern Iberia shows a floral succession and development similar to that of the Praclaux interglacial (Reille et al., 2000) and with certain features defining the Holsteinian interglacial of Western and Central Europe. Based on those observations, the authors correlate MIS 11 with the Praclaux interglacial and the Holsteinian. The summary of Fig. 28.14 presents a subdivision of the Quaternary of Western Europe based on the biostratigraphic units of the Scho¨ningen sequence. As the 230 Th/234U age determinations on peat, which were the pioneer research of Henk Heijnis (Groningen and Sydney) in the early 1990s, are still preliminary, and as different types of sediments and soils are currently the subjects of dating processes, I propose the scheme (Fig. 28.14) based on biostratigraphic correlatives. As the efforts of determining the exact age of the Holsteinian from the marine sediments of Bossel/Germany (Geyh and Mu¨ller, 2005) suggest a correlation with MIS 9, research is now focussing on testing this age determination at different sites and on further age determination of younger interglacial and interstadial peat deposits

Interglacial Pollen Records from Scho¨ningen

of definite post-Holsteinian sequences (Frechen et al., this volume). Amongst these, Scho¨ningen is of significance as it contains interglacial and interstadial peat deposits with an undisputable stratigraphical position below early Saalian till. ACKNOWLEDGEMENTS I am very indebted to Dr. Hartmut Thieme, Hannover, the archaeological excavator of the Scho¨ningen sites, for his sampling and provision of sediments as well as for fruitful and stimulating discussions, advice and financial support. I thank Christiane Hilmer, Suderburg, for valuable help with laboratory treatment of the samples and soil analyses and Barbara Albrecht for palynological work. I am very thankful to Katrin Becker and Mario Tucci, Suderburg, who helped draft the graphs and figures. I am very indebted to Helmut Mu¨ller, who gave me the opportunity to see the diagrams of pit Nachtigall and for the benefit of his comments and sharing of knowledge. Special thanks are given to Peter Kershaw, Clayton, Victoria, for critically reading the manuscript. I finally like to thank the two reviewers for their valuable advice.

REFERENCES Aalbersberg, G., Litt, T., 1999. Multiproxy climate reconstructions for the Eemian and Early Weichselian. Journal of Quaternary Science 13 (5), 367–390. Bassinot, F.V., Labeyrie, L.D., Vincent, E., Quidelleur, X., Shackelton, N., Lancelot, Y., 1994. The astronomical theory of climate and the age of the Brunhes-Matuyama magnetic reversal. Earth and Planetary Science Letter 126, 91–108. Behre, K.-E.,1974. Die Vegetation im Spa¨tpleistoza¨n von Osterwanna/Niedersachsen, Geologisches Jahrbuch A 18, 3–48. Behre, K.-E., 1989. Biostratigraphy of the last glacial period in Europe. Quaternary Science Reviews 8, 25–44. Behre, K.E., 2004. Das mittelpleistoza¨ne Interglazial von Surheide. Eiszeitalter und Gegenwart 54, 36–47.

441

Behre, K.E., Lade, U., 1986. Eine Folge von Eem und 4 Weichsel-Interstadialen in Oerel/Niedersachsen und ihr Vegetationsablauf. Eiszeitalter und Gegenwart 36, 11–36. Bittmann, F., 1992. The Ka¨rlich Interglacial, Middle Rhine Region, Germany; vegetation history and stratigraphic position. Vegetation History and Archaeobotany 1, 243–258, Berlin. Bittmann, F., Mu¨ller, H., 1996. The Ka¨rlich Interglaical site and its correlation with the Bilshausen sequence. In: Turner, C. (Ed.), The Early Middle Pleistocene in Europe. Balkema, Rotterdam, pp. 187–193. Bo¨hme, G., 2000. Reste von Fischen, Amphibien und Reptilien aus der Fundstelle Scho¨ningen 12 bei Helmstedt (Niedersachsen) Erste Ergebnisse. Praehistoria Thuringica 4, 18–27. Boenigk, W., Frechen, M., 1998. Zur Geologie der Deckschichten von Ka¨rlich/Mittelrhein. Eiszeitalter und Gegenwart 48, 38–49. Caspers, G., 1997. Die eem- und weichselzeitliche Hohlform von Groß Todtshorn (Kr. Harburg; Niedersachsen) – Geologische und palynologische Untersuchungen zu Vegetation und Klimaverlauf der letzten Kaltzeit. In: Freund, H., Caspers, G. (Eds.), Vegetation und Pala¨oklima der WeichselKaltzeit im no¨rdlichen Mitteleuropa, Hannover; Schriftenreihe der deutschen Geologischen Gesellschaft 4, 7–59. de Beaulieu, J.-L., Andrieu-Ponel, V., Reille, M., Gru¨ger, E., Tzedakis, C., Svoboda, H., 2001. An attempt at correlation between the Velay pollen sequence and the Middle Pleistocene stratigraphy from central Europe. Quaternary Science Reviews 20, 1593–1602. Desprat, S., Sa´nchez Gon˜i, M.F., Turon, J.-L., McManus, J.F., Loutre, M.F., Duprat, J., Malaize´, B., Peyron, O., Peypouquet, J.-P., 2005. Is vegetation responsible for glacial inception during periods of muted insolation changes. Quaternary Science Reviews 24, 1361–1374. Elsner, H., 1987. Das Quarta¨r im Tagebau Scho¨ningen der Braunschweigischen Kohlen-Bergwerke AG, Helmstedt. Diplomarbeit am Fachbereich Erdwissenschaften der Universita¨t Hannover. 126 p. unpublished. Elsner, H., 2003. Verbreitung und Ausbildung Elsterzeitlicher Ablagerungen zwischen Elm und Flechtinger Ho¨henzug. Eiszeitalter und Gegenwart 52, 91–116. EPICA Community Members, 2004. Eight glacial cycles from an Antarctic ice core. Nature 429, 623–628. Erd, K., 1970. Pollen-analytical classification of the Middle Pleistocene in the German Democratic Republic, Palaeogeography, Palaeoclimatology, Palaeoecology 8, 119–132. Erd, K., 1973. Vegetationsentwicklung und Biostratigraphie der Do¨mnitz-Warmzeit (Fuhne/Saale) im

442

Brigitte Urban

Profil von Pritzwalk/Prignitz, Abhandlungen des Zentralen Geologischen Instituts 18, 9–48. ¨ cker-Warmzeit von Ro¨persdorf Erd, K., 1987. Die U bei Prenzlau als neuer Interglazialtyp im SaaleKomplex der DDR. Zeitschrift fu¨r Geologische Wissenschaften 15, 297–313. Erd, K., Palme, H., Pra¨ger, F.,1987. Holsteininterglaziale Ablagerungen von Rossendorf bei Dresden. Zeitschrift fu¨r Geologische Wissenschaften 15, 281–295. Felix-Hennigsen, P., Urban, B., 1982. Paleoclimatic interpretation of a thick Intra-Saalian paleosol the ‘‘bleached loam’’ on the Drenthe moraines of Northern Germany. CATENA. 9, 1–8. Field, M.H., de Beaulieu, J.-L., Guiot, J., Ponel, P., 2000. Middle Pleistocene deposits at La Coˆte, Val-de-Lans, Ise´re department, France: plant microfossil, palynological and fossil insect investigations. Palaeogeography, Palaeoclimatology, Palaeoecology 159, 53–83. Frechen, M., Sierralta, M., Oezen, D., Urban, B., 2005. Uranium series dating of peat from Central and Northern Europe. In: The Climate of Past Interglacials. Frank Sirocko, Thomas Litt, Martin Claussen (Eds.) (this volume). Geyh, M.A., Mu¨ller, H., 2005. Numerical 230Th/U dating and a palynological review of the Holsteinian/Hoxnian interglacial. Quaternary Science Reviews 24, 1861–1872. Gru¨ger, E., 1983. Untersuchungen zur Gliederung und Vegetationsgeschichte des Mittelpleistoza¨ns am Samerberg in Oberbayern. Geologica Bavarica 84, 21–40. Gru¨ger, E., Jordan, H., Meischner, D., Schlie, P., 1994. Mittelpleistoza¨ne Warmzeiten in Go¨ttingen, Bohrungen Ottostraße und Akazienweg. Geologisches Jahrbuch A 134, 167–210. Hahne, J., Kemle, S., Merkt, J., Meyer, K.-D., 1994. Eem-, weichsel- und saalezeitliche Ablagerungen der Bohrung ‘‘Quakenbru¨ck GE2’’. In: K.-D. Meyer et al. (Eds.), Neuere Untersuchungen an Interglazialen in Niedersachsen; Geol. Jb. A 134, 9–69. Hallik, R., Linke, G., 1986. Die vegetationsgeschichtliche Entwicklung des Holstein-Interglazials nach Untersuchungen in der Region Hamburg. INQUA Subcommission on European Quaternary Stratigraphy. Abstract 8, Holstein Symposium, Hamburg. Hartmann, T., 1988. Elster- bis Saale-zeitliche Sedimente im Tagebau Scho¨ningen der Braunschweigischen Kohlen-Bergwerke AG, Helmstedt. Diplomarbeit am Fachbereich Erdwissenschaften der Universita¨t Hannover. 153 p. unpublished. Heijnis, H., 1992. Uranium/thorium dating of Late Pleistocene peat deposits in N.W. Europe. PhD thesis, Rijksuniversiteit Groningen, 149 pp. Heijnis, H., Urban, B., 1995. 230Th/234U dating of the middle and late Pleistocene organic deposits from the Scho¨ningen/Helmstedt area, Lower Saxony,

Germany. Schriften der Alfred Wegener Stiftung, 2/95, 109, INQUA, XIV Congress, Berlin. Hermsdorf, N., Strahl, J., 2005. Zum Problem der sogenannten ‘‘Ueckerwarmzeit’’ (Intrasaale) Untersuchungen an neuen Bohrkernen aus dem Raum Prenzlau. Tagung der Norddeutschen Geologen, 17.-20.5.2005, Lu¨beck. Abstract, 33–34. INQUA SEQS: Subcommission on European Quaternary Stratigraphy. The Saalian sequence in the type region (Central Germany) (Halle 1992). Convention: End of Holsteinian (top) is defined as base ¨ bereinkunft: of the Saalian Complex s.l. U Obergrenze des Holstein-Interglazials entspricht Untergrenze des Saale-Komplexes. Jancyk-Kopikowa, Z., 1975. Flora interglacjalu Marzowieckiego w Fernandowie. Biuletyn Institut Geologiczny 290, 1–94. Jancyk-Kopikowa, Z., Zarski, M, 1995. The Ferdinando´w interglacial at Stanislawice near Kozienice (Central Poland). Acta Palaeobotanica, 35, 7–13. Jechorek, H., 2000. Die fossile Flora des ReinsdorfInterglazials. Pala¨okarpologische Untersuchungen an mittelpleistoza¨nen Ablagerungen im Braunkohlentagebau Scho¨ningen. Praehistoria Thuringica 4, 7–17. Jo¨ris, O., Baales, M., 2003. Zur Altersstellung der Scho¨ninger Speere. Vero¨ffentlichungen des Landesamtes fu¨r Archa¨ologie 57, 281–287. Kukla, G., Cı´lek, V. 1996. Plio-Pleistocene megacycles:record of climate and tectonics, Palaeogeography, Palaeoclimatology, Palaeoecology 35, 121–144. Lenhart, R., 1989. Schichtlagerung und Zusammensetzung Elster- bis Saale-zeitlicher Sedimente im Baufeld Esbeck, Tagebau Scho¨ningen der Braunschweigischen Kohlen- Bergwerke AG, Helmstedt. Diplomarbeit am Fachbereich Erdwissenschaften der Universita¨t Hannover. 125 p. unpublished. Lindner, L., Marciniak, B., 1998. The occurrence of four interglacials younger than the Sanian 2 (Elsterian 2) Glaciation in the Pleistocene of Europe. Acta Geologica Polonica 48, 247–263. Linke, G., Hallik, R., 1993. Die pollenanalytischen Ergebnisse der Bohrungen Hamburg-Dockenhuden (qho4), Wedel (qho2) und Hamburg Billbrook. Geologisches Jahrbuch A 138, S. 169–184. Litt, T., 1990. Pollenanalytische Untersuchungen zur Vegetations- und Klimaentwicklung wa¨hrend des Jungpleistoza¨ns in den Becken von Gro¨bern und Grabschu¨tz. Altenburger naturwissenschaftliche Forschungen 5, 92–105. McManus, J.F., 2004. A great grand-daddy of ice cores. Nature 429, 611–612. Mania, D., 1995. Die geologischen Verha¨ltnisse im Gebiet von Scho¨ningen. In: Thieme, H., Maier, R., (Eds.), Archa¨ologische Ausgrabungen im Braunkohlentagebau Scho¨ningen, 33–43, Hahnsche Buchhandlung, Hannover.

Interglacial Pollen Records from Scho¨ningen Mania, D., 1998. Zum Ablauf der Klimazyklen seit der Elstervereisung im Elbe-Saalegebiet. Praehistoria Thuringica 2, 5–21. Mania, D., Mai, D.-H., 2001. Molluskenfaunen und Floren im Elbe-Saalegebiet wa¨hrend des mittleren Eiszeitalters. Praehistoria Thuringica 6/7, 46–92. Menke, B., 1968. Beitra¨ge zur Biostratigraphie des Mittelpleistoza¨ns in Norddeutschland. Meyniana 18, 35–42. Menke, B., 1980. Wacken, Elster-Glazial, marines Holstein-Interglazial und Wacken-Warmzeit. In: H.E. Stremme, B. Menke (Eds.), Quarta¨r-Exkursionen in Schleswig-Holstein, Geologisches Landesamt Schleswig-Holstein. Menke, B., Tynni, R., 1984. Das Eeminterglazial und das Weichselfru¨hglazial von Rederstall/ Dithmarschen und ihre Bedeutung fu¨r die mitteleuropa¨ische Jungpleistoza¨ngliederung. Geologisches Jahrbuch A 76, 120p. Meyer, K.-J., 1974. Pollenanalytische Untersuchungen und Jahresschichtenza¨hlungen an der holsteinzeitlichen Kieselgur von Hetendorf, Geologisches Jahrbuch A 21, 87–105. Mu¨ller, H., 1965. Eine pollenanalytische Neubearbeitung des Interglazial-Profils von Bilzhausen (UnterEichsfeld). Geologisches Jahrbuch 83, 327–352. Mu¨ller, H., 1974a. Pollenanalytische Untersuchungen und Jahresschichtenza¨hlungen an der eem-zeitlichenKieselgur von Bispingen/Luhe, Geologisches Jahrbuch A 21, 149–169. Mu¨ller, H., 1974b. Pollenanalytische Untersuchungen und Jahresschichtenza¨hlungen an der holstein-zeitlichen Kieselgur von Munster-Breloh, Geologisches Jahrbuch A 21,107–140. Mu¨ller, H., 1992. Climate changes during and at the end of the interglacials of the Cromerian complex, In Kukla, G.J., Wendt, E. (Eds.), Start of a Glacial, NATO ASI ser., 13, 51–69. Mu¨ller, H., Ho¨fle, H.-C., 1994. Das HolsteinInterglazialvorkommenbei Bossel westlich von Stade und Wanho¨den no¨rdlich Bremerhaven, Geologisches Jahrbuch, A 134, 71–116. Petit J.R., Jouzel, J., Barkov, N.I., Barnola, J.-M., Basile, I., Bender, M., Chappellaz, J., Davis, M., Delaygue, G., Delmotte, M., Kotlyakov, V.M., Legrand, M., Lipenkov, V.Y., Lorius, C., Pepin, L., Ritz, C., Saltzman, E, Stievenard, M., 1999. Climate and atmospheric history of the past 420000 years from the Vostok ice core, Antarctica. Nature 399, 429–436. Preusser, F., Drescher-Schneider, R., Fiebig, M., Schlu¨chter, C., 2005. Re-interpretation of the Meikirch pollen record, Swiss Alpine Foreland, and implications for Middle Pleistocene chronostratigraphy. Journal of Quaternary Science 20, 607–620. Reille, M., de Beaulieu, J.-L., 1995. Long Pleistocene Pollen Records from the Praclaux Crater, SouthCentral France. Quaternary Research 44, 205–215.

443

Reille, M., Andrieu, V., de Beaulieu, J-L., Guenet, P., Goeury, C., 1998. A long pollen record from Lac Du Bouchet, massif Central, France: For the Period ca. 325 to 100BP (OIS 9 c to OIS 5e), Quaternary Science Reviews 17, 1107–1123. Reille, M., de Beaulieu, J.-L., Svoboda, V., AndrieuPonel, V., Goeury, C., 2000. Pollen analytical biostratigraphy of the last five climatic cycles from a long continental sequence from the Velay region (Massif Central, France). Journal of Quaternary Science 15, 665–685. Richter, D., 1998. Thermolumineszenzdatierungen erhitzter Silices aus mittel- und jungpala¨olithischen Fundstellen. Anwendung und methodische Untersuchungen. PhD Thesis, Universita¨t Tu¨bingen. Sarnthein, M., Stremme, H.-E., Mangini, A. 1986. The Holsteinian Interglacial: time- stratigraphic position and correlation to stable-isotope stratigraphy of deep-sea sediments. Quaternary Research 26: 283–298. Schoch, W.H., 1999. Holz als Informationstra¨ger aus dem Pala¨olithikum. Praehistoria Thuringica 3, 98–106. Strahl, J., 1999. Biostratigraphische Untersuchungen im Bereich des Oberen Allertales (Raum Morsleben und Ummendorf). 66, Tagung AG Nordwestdt. Geologen, Tagungsband und Exkursionsfu¨hrer, 119–124. Halle. Thieme, H., 1996. Altpala¨olithische Wurfspeere aus Scho¨ningen, Niedersachsen – Ein Vorbericht, Archa¨ologisches Korrespondezblatt 26, 377–393. Thieme, H., 1997. Lower Paleolithic hunting spears from Germany. Nature 385, 807–810. Thieme, H., 1998. Altpala¨olithische Wurfspeere von Scho¨ningen, Niedersachsen. Praehistoria Thuringica 2, 22–31. Thieme, H., 1999. Altpala¨olithische Holzgera¨te aus Scho¨ningen, Lkrs. Helmstedt. Bedeutsame Funde zur Kulturentwicklung des fru¨hen Menschen. Germania 77, 451–487. Thieme, H., Mania, D., 1993. ‘‘Scho¨ningen 12’’ – ein mittelpleistoza¨nes Interglazialvorkommen im Nordharzvorland mit pala¨olithischen Funden, Ethnographisch-Archa¨ologische Zeitschrift 1993, 34: 610–619 Thieme, H., Maier, R., 1995. Archa¨ologische Ausgrabungen im Braunkohlentagebau Scho¨ningen, Landkreis Helmstedt. 191 pp. Braunschweigische Kohlen-Bergwerke AG Helmstedt, (Eds.), Verlag Hahnsche Buchhandlung Hannover. Thieme, H., Maier, R., Urban, B., 1987. Archa¨ologische Schwerpunktuntersuchungen im Helmstedter Braunkohlenrevier (ASHB) – zum Stand der Arbeiten 1983–1986. Archa¨ologisches Korrespondenzblatt 17, 445–462. Thieme, H., Maier, R., Urban, B., 1992. Neue Erkenntnisse zum urgeschichtlichen Siedlungsgeschehen. – Archa¨ologie in Deutschland, Heft 2, 26–30.

444

Brigitte Urban

Thieme, H., Mania, D., 1993. ‘‘Scho¨ningen 12’’ – ein mittelpleistoza¨nes Interglazialvorkommen im Nordharzvorland mit pala¨olithischen Funden, Ethnographisch-Archa¨ologische Zeitschrift 34, 610–619. Thieme, H., Mania, D., Urban, B., van Kolfschoten, T., 1993. Scho¨ningen (Nordharzvorland) eine altpala¨olithische Fundstelle aus dem mittleren Eiszeitalter, Archa¨ologisches Korrespondenzblatt 23, 147–163. Tschee, W., 1991. Die pleistoza¨ne Schichtfolge im Tagebau Scho¨ningen Baufeld Esbeck der Braunschweigischen Kohlen-Bergwerke AG, Helmstedt. Diplomarbeit am Fachbereich Erdwissenschaften der Universita¨t Hannover. 75p. unpublished. Turner, C., 1970. The Middle Pleistocene deposits of Marks Tey, Essex. Philosophical Transactions of the Royal Society of London, Series B 257, 3373–440. Turner, C., 1998. Volcanic maars, long Quaternary sequences and the work of the INQUA Subcommission on European Quaternary stratigraphy. In: Cavarretta, G., Fornaceri, M., Follieri, M., Girotti, Turner, C. (Guest Eds.), Quaternary Stratigraphy in Volcanic Areas. Quaternary International 47/48, 3–20. Urban, B. (1983). Biostratigraphic correlation of the Ka¨rlich Interglacial, Northwestern Germany. BOREAS, 12, pp. 83–90, Oslo. Urban, B., 1992. Interglacial/glacial transitions recorded from middle and young Pleistocene sections of eastern Lower Saxony-Germany. In: Kukla, G.J., Went, E. (Eds.), Start of a Glacial. NATO ASI Series, Vol. I 3, 37–50, Springer Verlag, Berlin. Urban, B., 1995a. Palynological evidence of younger Middle Pleistocene Interglacials (Holsteinian, Reinsdorf, Scho¨ningen) in the Scho¨ningen open cast lignite mine (eastern Lower Saxony/ Germany).Mededelingen Rijks Geologische Dienst 52, 175–186. Urban, B., 1995b. Vegetations- und Klimaentwicklung des Quarta¨rs im Tagebau Scho¨ningen. In: Thieme, H., Maier, R., (Eds.), Archa¨ologische Ausgrabungen im Braunkohlentagebau Scho¨ningen, pp. 44–56, Hahnsche Buchhandlung, Hannover. Urban, B., 1996a. Mittelpleistoza¨ne Waldzeiten im Tagebau Scho¨ningen: Spektren aus dem HolsteinInterglazial und dem Harbke-Interstadial. In: Spuren der Jagd – Die Jagd nach Spuren. Tu¨binger Monographien zur Urgeschichte 11, 487–495. Urban, B., 1996b. Zur Pala¨oo¨kologie und Stratigraphie des Mittelpleistoza¨ns im Tagebau Scho¨ningen/NO Niedersachsen. Landesamt fu¨r Natur und Umwelt des Landes Schleswig-Holstein: Bo¨den als Zeugen der Landschaftsentwicklung, 127–140. Urban, B., 1997. Grundzu¨ge der eiszeitlichen Klimaund Vegetationsgeschichte in Mitteleuropa. In: Wagner, G.A., Beinhauer, K.W. (Eds.), Homo

heidelbergensis von Mauer – Das Auftreten des Menschen in Mitteleuropa, 241–265, Universita¨tsverlag C. Winter, Heidelberg. Urban, B., 1999. Middle and Late Pleistocene biostratigraphy and paleoclimate of an open-pit coal mine Schoningen: Germany. Chinese Science Bulletin 44 Suppl., 30–37. Urban, B., 2002. Rekonstruktion pleistoza¨ner und holoza¨ner Landschafts- und Klimageschichte im no¨rdlichen Mitteleuropa mit Hilfe limnisch-telmatischer und terrestrischer Sediment- und Bodenabfolgen. In: Geo 2002 – Planet Erde: Vergangenheit, Entwicklung, Zukunft. Deutsche Geologische Gesellschaft 21, 336–337. Urban, B., Thieme, H., Elsner, H., 1988. Biostratigraphische, quarta¨rgeologische und urgeschichtliche Befunde aus dem Tagebau ‘‘Scho¨ningen’’, Landkrs. Helmstedt. Zeitschrift der deutschen geologischen Gesellschaft 139, 123–154. Urban, B., Elsner, H., Ho¨lzer, A., Mania, D., Albrecht, B., 1991a. Eine eem- und fru¨hweichselzeitliche Abfolge im Tagebau Scho¨ningen, Landkreis Helmstedt. Eiszeitalter und Gegenwart 41, 85–99. Urban, B., Lenhard, R., Mania, D., Albrecht, B., 1991b. Mittelpleistoza¨n im Tagebau Scho¨ningen, Ldkrs. Helmstedt. Zeitschrift der deutschen geologischen Gesellschaft 142, 351–372. van den Boogard, C., Boogard, P., v.d., Schmincke, H.-U.,1989. Quarta¨rgeologisch-tephrostratigraphische Neuaufnahme und Interpretation des Pleistoza¨nprofils Ka¨rlich. Eiszeitalter und Gegenwart 39, 62–86. van Kolfschoten, T., 1995. Faunenreste des altpala¨olithischen Fundplatzes Scho¨ningen 12 (Reinsdorf Interglazial). In: Thieme, H., Maier, R., (Eds.), Archa¨ologische Ausgrabungen im Braunkohlentagebau Scho¨ningen, Hahnsche Buchhandlung, Hannover, 85–94 Veil, S.; Breest, K.; Ho¨fle, H.-C.; Meyer, H.-H.; Plisson, H.; Urban-Ku¨ttel, B.; Wagner, G.A.; Zo¨ller, L., 1992. Ein mittelpala¨olithischer Fundplatz aus der Weichsel-Kaltzeit bei Lichtenberg, Lkr. Lu¨chow-Dannenberg. Germania 72, 1–66. Welten, M., 1982. Pollenanalytische Untersuchungen im ju¨ngeren Quarta¨r des no¨rdlichen Alpenvorlandes der Schweiz. Beitra¨ge zur Geologischen Karte der Schweiz – Neue Folge 156, 179pp. Welten, M., 1988. Neue pollenanalytische Ergebnisse u¨ber das ju¨ngere Quarta¨r des no¨rdlichen Alpenvorlandes der Schweiz (Mittel- und Jungpleistoza¨n). Beitra¨ge zur Geologischen Karte der Schweiz – Neue Folge 162, 40 pp. Zagwijn, W.H., 1973. Pollenanalytic studies of Holsteinian and Saalian beds in the northern Netherlands. Mededelingen Rijks Geologische Dienst 24, 139–156.

29. Mammalian Faunas From the Interglacial Periods in Central Europe and Their Stratigraphic Correlation Wighart von Koenigswald Institut fu¨r Pala¨ontologie der Universita¨t Bonn, Nussallee 8, D-53115 Bonn

ABSTRACT During the Pleistocene, climatic oscillations caused multiple faunal exchanges in Central Europe. The interglacial Elephas antiquus assemblage reinvaded Central Europe several times. The various Middle Pleistocene interglacial periods are very difficult to distinguish based on the fauna alone. The classical biostratigraphic tool, the different levels in a continuous evolutionary sequence, can only be applied in very few genera, e.g. Arvicola, to a limited degree. During the Middle and Late Pleistocene, a decline in faunal diversity can be observed. 29.1 INTRODUCTION The mammalian fauna of the Central European Pleistocene is characterized by drastic turnovers induced by climatic changes. During the Middle and Late Pleistocene, a continental fauna characterized by Mammuthus and Coelodonta characterizes the cold periods. During the warm periods, the interglacials, a fauna dominated by Elephas antiquus was present. In the Middle and Late Pleistocene, a Mammuthus assemblage can clearly be differentiated from an Elephas assemblage. Only very few herbivores occur in both assemblages (Fig. 29.1). With each climatic change, the fauna was exchanged. That means that the new fauna invaded, while the previous one became locally extinct (von Koenigswald, 2003). This faunal exchange of the Middle and Late Pleistocene is more significant than in most other areas of the world, due to the specific geographical situation of Central (and Western)

Europe. The West–East oriented mountain ranges of the Pyrenees and Alps stabilized the climate in the Mediterranean as well as in Central Europe. During cold periods, small oscillations of warmer climate were buffered, and thus the glacial fauna was not changed during most interstadials. Only interglacials led to an intensive faunal exchange, and the Mediterranean fauna expanded its area using the passages of the Rhone and the Danube valleys (Fig.29.2). Temperature and humidity are the main components of climate. The very significant difference between glacial and interglacial faunas in Central Europe is due to the shift from a highly continental biome during glacial periods to a more maritime-influenced climate during the interglacials. Thus, it is the changing precipitation, which has the greatest influence. Faunal lists for most of the sites mentioned here were given in von Koenigswald and Heinrich (1999), including the references to these local faunas, and thus have not been repeated here. 29.2 THE EARLY PLEISTOCENE The climatic fluctuations characterizing the Pleistocene had already started in the Pliocene, as documented in the pollen record. In the mammalian record, no typical cold faunas are yet known in the fossil record from the Pliocene or Early Pleistocene. This might be due to the very scarce fossil record, or to the fact that the mammalian fauna was not yet adapted to a cold environment. At least, there is no indication of faunal exchanges as in the Middle and Late Pleistocene.

446

Wighart von Koenigswald

Elephas - fauna

Mammuthus - fauna

Crocuta c. spelaea Ovibos moschatus

Dama dama Capreolus capreolus

Panthera leo spelaea

Bubalus murrenis

Rangifer tarandus

Coelodonta antiquitatis Bos primigenius

Ursus spelaeus

Hippopotamus amphibius Bison priscus

Saiga tatarica

Stephanorhinus kirchbergensis Cervus elaphus

Sus scrofa Elephas antiquus

Equus ferus

Megaloceros giganteus

Mammuthus primigenius

Fig. 29.1 The Elephas assemblage and the Mammuthus assemblage occurred alternatively in Central Europe during the Middle and Late Pleistocene. Only few herbivores but several carnivores occurred during interglacial and glacial environments (von Koenigswald, 2002).

The mammalian fauna of Untermaßfeld near Meiningen represents temperate climatic conditions of a very late phase of the Early Pleistocene. The rich fauna, containing many large herbivores and magnificent carnivores, was deposited in fluvial sediments after a great flood (Kahlke, 1997–2001) and includes Hippopotamus amphibius, a most indicative taxon for interglacial conditions. The specific ecological requirements of Hippopotamus will be discussed in the section on the Eemian. 29.3 MIDDLE PLEISTOCENE INTERGLACIAL FAUNAS BEFORE THE ELSTERIAN Among the faunas of the early Middle Pleistocene, two stratigraphic levels can be distinguished according to the occurrence of

two large vole taxa (Fig. 29.3). The genus Mimomys has rooted molars and characterizes the older faunas, while the more derived Arvicola characterized by rootless molars is present in the younger part of the Middle Pleistocene, the Late Pleistocene and Holocene. The genus Arvicola shows some progressive evolution during that time interval and thus is an important stratigraphic tool. The Mimomys savini faunas are well documented in Thu¨ringen, especially at Voigtstedt (Kahlke, 1965) and Su¨ssenborn (Kahlke, 1969). They are younger than the Matuyama/Brunhes boundary and thus belong to the early Middle Pleistocene. The fauna of Voigtstedt represents most probably a temperate environment, as indicated by the presence of Sus scrofa, Capreolus suessenbornensis and a flying squirrel Petauria voigtstedtensis. Voigtstedt is correlated with

Mammalian Faunas From the Interglacial Periods

?

Area of temporal occurence

Core

447

Core area of the glacial fauna

?

? area o f

the interg

lacial fauna 500 km

?

Fig. 29.2 Immigration routes of the Elephas assemblage from the South and the Mammuthus assemblage from the North East during the Middle and Late Pleistocene.

Geology

Small mammal stratigraphy

Holocene

Eemian

Arvicola cantianus-terrestris faunas

Elephas antiquus Hippopotamus Bubalus

5e

Lehringen Taubach

Warthe Drente Dömintz

Late Arvicola cantianus faunas

Elephas antiquus Bubalus

?7

Elephas antiquus

?9 ?11

W. Ehringsdorf Steinheim/Murr Schöningen Bilzingsleben

Hosteinian Elsterian Early Arvicola cantianus faunas Cromerian Complex

Lower Pleistocene

Important faunas

1

Arvicola terrestris-faunas

Weichselian Late Pleistocene

Middle Pleistocene

Significant Interglacials immigrants (and MIS)

Brunhes Matuyama

Arvicola Mimomys +

Elephas antiquus

Kärlich G Mosbach Mauer

Elephas antiquus Hippopotamus Arvicola

Mimomys savini-faunas

Mimomys savini-faunas with M. pusillus

Süssenbron Voigtstedt Hippopotamus

>19

Untermaßfeld

Fig. 29.3 Biostratigraphy of the Middle and Late Pleistocene in Central Europe (von Koenigswald and Heinrich, 1999).

448

Wighart von Koenigswald

the West Runton Freshwater Bed of the Cromer Forestbed Formation in East Anglia (Stuart, 1981). The fauna of Su¨ssenborn contains Ovibos and Rangifer, indicators of a cold climate as in later faunas and at present. Whether these genera represent an arctic environment in the early Middle Pleistocene remains uncertain since other taxa, especially the rhinos present in this fauna, are well known from warmer periods. The interglacial faunas from Jockrim in the Pfalz and from Wu¨rzburg-Schalksberg cannot be correlated stratigraphically, since neither Mimomys nor Arvicola is present. Both sites produced Hippopotamus amphibius, a faunal element occurring only during interglacial climatic conditions. At Mauer near Heidelberg, an interglacial period is represented, which is distinctly younger than Voigtstedt and Su¨ssenborn, since Arvicola cantianus has replaced Mimomys savini. Within the Arvicola cantianus faunas, Mauer represents an early stage since the diversity of insectivores and rodents is quite high. Together with Arvicola cantianus, another important faunal element, the straight-tusked elephant Elephas antiquus, occurs here for the first time. This elephant will be significant for all later interglacial faunas in Central Europe (except the Holocene). It should be mentioned that Mauer is the type locality of Homo heidelbergensis. The fully interglacial character of this site is stressed by the presence of Hippopotamus amphibius besides Sus scrofa and Capreolus capreolus priscus. The fauna of Mauer represents one of the interglacial periods within the later part of the Cromerian Complex, but it is not necessarily the latest interglacial. Another interglacial assemblage representing the early Arvicola cantianus faunas with a great diversity of insectivores and rodents, e.g. containing Talpa minor and Pliomys, was described from Ka¨rlich G. This site is important since the various layers are intercalated with tuffs. A tuff higher up in the profile was dated at 618 ka. A similar

and very rich early Arvicola cantianus fauna was found in the site Miesenheim nearby, covered by the same tuff. Distinctly younger is the so-called ‘Ka¨rlich Interglacial’ which was detected at the site Ka¨rlich Seeufer. The lake sediments produced remains of six individuals of Elephas antiquus together with artifacts (Bosinski, 1995). Similarly to the somewhat older fauna of Ka¨rlich H, it represents the early Arvicola cantianus fauna. The underlying ‘Brocken tuff’ provided an age of 396 ka. According to the pollen assemblages, Ka¨rlich Seeufer was regarded as late Cromerian and correlated convincingly with Bilshausen (Bittmann and Mu¨ller, 1996), which definitively underlies sediments of the Elsterian. Thus, the early Arvicola cantianus faunas antedate the Elsterian and represent part of the Cromerian Complex, but this faunal type might have occurred during several interglacial periods with cold or even glacial conditions in between. Some arctic faunal elements, Ovibos and Rangifer, occur in Mosbach 2, but typical glacial faunas are missing so far. The mammalian fauna of the Elsterian is more or less unknown. The occurrence of Coelodonta antiquus and Rangifer tarandus in Bornhausen near Seesen constitutes one of the faunas correlated with this glacial period. 29.4 MIDDLE PLEISTOCENE INTERGLACIAL FAUNAS BETWEEN ELSTERIAN AND SAALIAN Biostratigraphers struggle with the number of interglacial phases after the Elsterian and before the Eemian (Fig. 29.3). Mania and Thomae (2006) postulate four independent interglacial phases within the Holstein Complex and two additional interglacial periods within the Saalian. Sarntheim et al. (1986) and Schreve (2001) correlate the Holsteinian with MIS 11 and accordingly count three interglacials for this time period. Geyh and Mu¨ller (2005) dated the type locality of the Holsteinian and correlated it with MIS 9.

Mammalian Faunas From the Interglacial Periods

Litt et al. (2005) accepted one or two interglacial phases after the Holsteinian and before the first Saalian ice advance (Drenthe), rejecting any interglacial phase between the Drenthe and the Warthe. Thus besides the Holsteinian, only one (or two) additional interglacial phases are accepted as antedating the Eemian. Various sites have produced interglacial faunas, such as Bilzingleben II, Scho¨ningen 12, Scho¨ningen 13, Steinheim/Murr and Weimar-Ehringsdorf, but none of them shows a complete interglacial cycle in their pollen record. Bilzingsleben II is definitively younger than the Elsterian, but according to Mania (1997) and Mania and Thomae (2006), it is not the first interglacial period after the Esterian. The horizon, which produced the remains of Homo ‘erectus’ bizingslebenensis, provided a typical interglacial fauna with Elephas antiquus, Stephanorhinus kirchbergensis, Sus scrofa, and Capreolus. Among the rodents, the vole Arvicola cantianus is present, but some important elements from the early Arvicola cantianus fauna, e.g. Talpa minor, Drepanosorex and Pliomys, are missing although the fauna is fairly rich. Therefore, this faunal association is regarded as the younger Arvicola cantianus fauna. Another difference is that the rhino Stephanorhinus hundsheimensis is replaced by the more derived S. kirchbergensis. The faunas of Scho¨ningen 12 and 13 fit into the same faunal pattern, but based on the faunal content it is unclear whether they represent the same interglacial period or a slightly younger one. Differences seen in horses compared to those in Bilzingleben II are very difficult to evaluate, due to the limited material and the great tendency of horses to develop local forms. It is possible that Bos primigenius is present (E.v.Asperen, personal communication) which might be important for the comparison with Steinheim. The river deposits of Steinheim/Murr provided a very rich interglacial Elephas antiquus assemblage, which has often been

449

correlated with the Holsteinian (Adam, 1954, 2003). This was based on the old scheme with only three possible interglacials. Since the true number of interglacials during the Middle Pleistocene is known from the deep-sea record, a necessary correlation with the Holsteinian is obsolete. This Steinheim interglacial is older than the Eemian, and according to the geology it might correlate with MIS 7. The skull of Homo steinheimensis differs distinctly from the human remains found in Bilzingsleben II, and thus it belongs most probably to another human immigration. In the fauna, two important bovids occur for the first time: Bos primigenius and Bubalus murrensis. They are not present in Bilzingsleben II. The water buffalo is of great ecological significance since this animal requires open water during the winter to shelter from cold winds. The specific ecological requirements of Bubalus will be discussed together with those of Hippopotamus in the section on Eemian faunas. Weimar-Ehringsdorf is a travertine deposit, predominately of interglacial character. Human remains and a rich fauna were excavated from the lower travertine, which is of interest here. Traditionally, the site was regarded as Eemian, but several arguments indicate that this fauna represents an older interglacial period. Arvicola shows an evolutionary level intermediate between the one from Bilzingsleben and those of typical Eemian sites, for example Taubach and Burgtonna. Absolute dates indicate a correlation with MIS 7 (Schreve and Bridgland 2002). The difficulty of biostratigraphical correlation is caused by the similarity of the typical Elephas antiquus assemblages from the various interglacial periods of the late Middle Pleistocene and the Eemian. Differences in evolutionary level are restricted to few taxa. The mammalian fauna indicates that Bilzingsleben is earlier than Steinheim. Scho¨ningen may be the equivalent of Steinheim. Whether Weimar-Ehringsdorf belongs to the same interglacial cannot be decided from the faunal record.

450

Wighart von Koenigswald

Although one has to assume that the Elephas antiquus assemblages immigrated repeatedly, differences in the fauna are so limited that a biostratigraphic differentiation is not yet possible. 29.5 MAMMALIAN FAUNAS FROM THE EEMIAN After the end of the Saalian, the Elephas antiquus assemblage reinvaded from the Mediterranean. The mammalian fossil record is not rich enough to demonstrate the sequence of immigration similarly to that demonstrated among various taxa of the vegetation (Litt, 1994). The Elephas antiquus assemblage of the last interglacial does not show significant evolutionary changes compared to the previous interglacials. One of the few guide fossils is the vole Arvicola, which is characterized in its evolution by the transitional stage from A. cantianus to A. terrestris. Two rhinos Stephanorhinus kirchbergensis and S. hemitoechus are present. On the continent, S. kirchbergensis dominates over the other, while S. hemitoechus is the only rhino on the British Isles during this interglacial. In Northern Germany, the Eemian provides several localities with the typical Elephas antiquus assemblage including Stephanorhinus kirchbergensis, Sus scrofa, Dama dama, Cervus elaphus; Capreolus capreolus; Megaloceros giganteus and Bos primigenius. These faunas contain Equus ferus and sometimes Equus hydruntinus as well, species which are often claimed to represent a steppic environment. But obviously they occurred in forested biomes as well, since a dense forest can be assumed from the pollen record (Litt, 2000). In Thu¨ringen, the travertines of Burgtonna (Kahlke, 1978) and Taubach (Kahlke, 1977) provided important faunal remains of the Elephas antiquus assemblage. The stratigraphic position of the lakesite Neumark-Nord is debated. It is superimposed on the moraines of the Drenthe

glaciation (lower Saalian). Mania and Thomae (2006) postulate that the lake basin of Neumark-Nord 1 represents the first intra-Saalian interglacial phase covered by Warthe deposits (late Saalian glacial expansion). But Litt (in Mania et al., 1990) places Neumark-Nord into the Eemian according to the development of the vegetation and denies an intra-Saalian interglacial after the Drenthe. In the rodent fauna, the occurrence of Apodemus maastrichtiensis has to be mentioned. It is present at Weimar-Ehringsdorf but was not found in Eemian faunas of the area (Heinrich, 2001). But the stratigraphic value of this rodent is still open. Neumark-Nord 1 produced a number of well-preserved and entire skeletons of interglacial mammals, e.g. Elephas antiquus; Stephanorhinus kirchbergensis and Bos primigenius. The interglacial Dama dama geiselana is represented by a great number of male individuals with impressive antlers (Pfeiffer, 1999). Cervus elaphus; one of the few herbivores occurring during interglacial and glacial phases, is well represented, too. It is assumed that algal blooms of cyanobacteria poisoned at least some of these animals (Braun and Pfeiffer, 1994) (Tables 29.1, 29.2). Deposits of the Rhine River in the northern Oberrheinebene near Darmstadt are intensively quarried and produced a rich glacial fauna and several interglacial species (von Koenigswald, 1988). The stratigraphy of the pits cannot be observed directly since the profile is under the water table. Nevertheless, long-time observation indicates that the glacial fauna comes from the upper part while thick black trunks of oak (Quercus sp.) characterize the lower section. The interglacial fauna is quarried from the same depth. Most probably, this fauna belongs to the last interglacial, the Eemian. Typical and frequent faunal elements are those of Elephas antiquus, Stephanorhinus kirchbergensis, Dama dama and Bos primigenius. In addition, Bubalus murrensis and Hippopotamus amphibius

Mammalian Faunas From the Interglacial Periods

451

Rheinsande

Burgtonna

Taubach

MIS 1 Holocene

Weichselian

Saalian

Eemian MIS 5e

Lehringen

WeimarEhringsdorf

Steinheim/Murr

Schöningen II

MIS 9- MIS 7

Bilzingleben II

Kärlich G

Elsterian

Middle Pleistocene Mosbach II

Mauer

Voigtstedt

Untermaßfeld

Early Pleistocene

Table 29.1 Occurrence of selected large mammals in representative interglacial sites of Germany

Elephas antiquus Stephanorhinus etruscus Stephanorhinus hundsheimensis Stephanorhinus kirchbergensis

sp.

Stephanorhinus hemitoechus m.

Equus div. sp.

m.

sp.

st.

t.

Equus hydruntinus Hippopotamus amphibius Sus scrofa Dama reichenaui

sp.

sp.

Dama dama

sp.

sp.

( )

Cervus elaphus Megaloceros verticornis Megaloceros giganteus Alces latifrons

sp.

c.

sp.

Alces alces Capreolus suessenbornensis Capreolus capreolus Bison schoetensacki Bison priscus

sp.

m.

Bison bonasus Bos primigenius

?

Bubalus murrensis

occur in several of these sand pits. Their preservation and frequency exclude a redeposition from older sediments. A correlation with the last interglacial is plausible

due to the geological situation and the rich occurrence of Hippopotamus during the timeequivalent Ipswichian in the British Isles. In the same sediments, Alces latifrons and

452

Wighart von Koenigswald

MIS 1 Holozän

Weichselian Rheinsande

Burgtonna

Taubach

Eemian MIS 5e

Lehringen

WeimarEhringsdorf

Steinheim/Murr

Schöningen II

Bilzingleben II

MIS 9- MIS 7

Saalian

Elsterian Kärlich G

Mosbach II

Middle Pleistocene Mauer

Voigtstedt

Untermaßfeld

Early Pleistocene

Table 29.2 Occurrence of selected small mammals in representative interglacial sites of Germany

Talpa minor Talpa europaea Sorex (D.) savini

sp.

Mimomys savini Arvicola cantianus Arvicola cantianus/terrestris Arvicola terrestris Pliomys div sp. Trogontherium cuvieri

?

Castor fiber Apodemus sylvaticus Apodemus maastrichensis

sp. aff.

Trogotherium cuvieri were found (von Koenigswald and Menger, 1997). These species were often regarded as typical for much older deposits. But Alces latifrons is known from the interglacial of Weimar-Ehringsdorf, and the fossil limit of Trogontherium has become younger and younger during recent decades. Thus, the use of last appearance dates for stratigraphic purposes may be questionable, especially for rare species. The occurrence of Bubalus and Hippopotamus is of great ecological significance. Their present distribution in subtropical regions does not indicate similar conditions for periods when these animals were living in Central Europe. These animals can tolerate lower temperatures but submerge in the water to escape from cold winds. During winters, open water in rivers and lakes was only available in the Rhine valley, when the maritime influence was significantly

higher than today. That means mild winters but cooler summers if the annual mean temperature was not raised very much; and from the vegetation we know that the annual temperature was only 2 or 3 higher than today. Thus, a high maritime influence on the climate can be assumed for the Rhine area at least during part of the Eemian (von Koenigswald, 1988, 1991). The maritime influence was certainly less towards the East; at least Hippopotamus was not found further East during the Eemian. In this context, it is worthwhile noticing that the easternmost occurrence of this species is near Warsaw; thus it seems to avoid a very continental climate. This example indicates that it is not only the temperature, but even more the precipitation, which controls the faunal composition of the interglacial periods. In each Middle Pleistocene interglacial period and in the Eemian, the Elephas

Mammalian Faunas From the Interglacial Periods

antiquus assemblage was present in Central Europe. Since there was no major refuge north of the Alpine arc the species had to re-immigrate each time from the Mediterranean. The composition of the Elephas antiquus assemblage is fairly similar in various interglacial periods, but some species occur irregularly, or only in specific interglacial periods. The two genera Hippopotamus and Bubalus were mentioned because of their ecological significance. Finds of Macaca sylvanus are very rare. The faunal record of Central Europe does not allow us to decide whether two, three or even four different interglacial phases were intercalated between the Elsterian and the Saalian. 29.6 THE WEICHSELIAN AND THE HOLOCENE With the beginning of the last glaciation, the interglacial fauna became locally extinct in Central Europe. Most probably, the species survived longer in the Mediterranean, but their disappearance is not well documented, neither stratigraphically nor regionally. During the Weichselian, interglacial species such as Sus scrofa may have reinvaded Central Europe only sporadically (von Koenigswald and Heinrich, 1996). While Elephas antiquus and Stephanorhinus kirchbergensis became extinct, Dama and Hippopotamus survived in the eastern Mediterranean, but these species were not able to re-colonize Central Europe during the Holocene. Dama reinvaded Central and Western Europe in each interglacial period but not during the Holocene, although the climatic conditions were more favourable during the middle Holocene (Atlanticum) than today. However, when human reintroduced Dama in medieval times, this deer flourished and is now widespread again (von Koenigswald, 2002). This example indicates that it is not only the climate, through temperature and humidity, which

453

controls the faunal composition of interglacial periods. There are particular historical factors involved too.

REFERENCES Adam K.D., 1954. Die mittel-pleistoza¨nen Faunen von Steinheim an der Murr (Wu¨rtemberg). Quaternaria 1, 131–144. Adam, K.D., 2003. Der Homo steinheimensis im Spannungsfeld von Alt- und Neumensch. Vero¨ffentlichungen des Landesamtes fu¨r Archa¨ologie Sachsen-Anhalt - Landesmuseum fu¨r Vorgeschichte 57. Bittmann, F., Mu¨ller, H., 1996. The Ka¨rlich Interglacial site and its correlation with the Bilshausen sequence. In: Turner, C. (Ed.), The early Middle Pleistocene in Europe. Balkema, Rotterdem, 187–193. Bosinski, G., 1995. The Palaeolithic and Mesolithic of the Rhineland. In: Schirmer, W. (Ed.), Quaternary fieldtrips in Central Europe 2 Field trips on special topics. Pfeil, Mu¨nchen, 829–999. Braun, A., Pfeiffer, T., 1994. Cyanobacterial blooms as the cause of a Pleistocene large mammal assemblage. Paleobiology 28, 139–154. Geyh, M.A., Mu¨ller, H., 2005. Numerical 239Th/U dating and a palynological review of the Holsteinian/Hoxnian interglacial. Quaternary Science Reviews 24, 1861–1872. Heinrich, W.D., 2001. Kleinsa¨ugerreste aus interglazialen Ablagerungen von Neumark-Nord, Mitteldeutschland. Praehistoria Thuringica 6/7, 132–138. Kahlke, H.D. (Ed.), 1965. Das Pleistoza¨n von Voigtstedt. Pala¨ontologische Abhandlungen, Pala¨ozoologie II, 221–692. Kahlke, H.D. (Ed.), 1969. Das Pleistoza¨n von Su¨ssenborn. Pala¨ontologische Abhandlungen, Pala¨ozoologie III, 367–788. Kahlke, H.D. (Ed.), 1977. Das Pleistoza¨n von Taubach bei Weimar. Quarta¨rpala¨ontologie 2, 1–509. Kahlke, H.D. (Ed.), 1978. Das Pleistoza¨n von Burgtonna. Quarta¨rpala¨ontologie 3, 1–399. Kahlke, R.D. (Ed.), 1997–2001. Das Pleistoza¨n von Untermaßfeld bei Meiningen (Thu¨ringen) Teil 1– 3, Ro¨misch-Germanisches Zentralmuseum Mainz, Monographien 40/1–3. von Koenigswald, W. (Ed.), 1988. Zur Pala¨oklimatologie des letzten Interglazials im Nordteil der Oberrheinebene. Pala¨oklimaforschung 4, 1–327. von Koenigswald, W., 1991. Exoten in der Großsa¨ugerFauna des letzten Interglazials von Mitteleuropa. Eiszeitalter und Gegenwart 41, 70–84. ¨ kologie und Biosvon Koenigswald, W. 1992. Zur O tratigraphie der beiden pleistoza¨nen Faunen von

454

Wighart von Koenigswald

Mauer bei Heidelberg. In: Beinhauer, K.W., Wagner, G.A. (Eds), Schichten von Mauer - 85 Jahre Homo erectus heidelbergensis. – Mannheim, 101–110. von Koenigswald, W. 2002. Lebendige Eiszeit Klima und Tierwelt im Wandel. Wissenschaftliche Buchgesellschaft Darmstadt und Theiss Stuttgart, pp. 190. von Koenigswald, W. 2003. Mode and cause for the Pleistocene turnovers in the mammalian fauna of Central Europe. Deinsia 10, 305–312. von Koenigswald, W., Heinrich, W.-D., 1996. Kurze Charakterisierung der Vera¨nderungen in der Sa¨ugetierfauna des Jungquarta¨rs in Mitteleuropa. Tu¨binger Mongraphien zur Urgeschichte 11, 437–448. von Koenigswald, W., Heinrich W.-D., 1999. Mittelpleistoza¨ne Sa¨ugetierfaunen aus Mitteleuropa der Versuch einer biostratigraphischen Zuordnung. Kaupia 9, 53–112. von Koenigswald, W., Menger, F., 1997. Mo¨gliches Auftreten von Trogontherium cuvieri und Alces latifrons im letzten Interglazial der no¨rdlichen Oberrheinebene. Cranium 14, 2–10. Litt, T., 1994. Pala¨oo¨kologie, Pala¨obotanik und Stratigraphie des Jungquarta¨rs im nordmitteleuropa¨ischen Tiefland. Dissertationes Botanicae 227, pp. 185. Litt. T., 2000. Waldland Mitteleuropa – die Megaherbivorentheorie aus pala¨obotanischer Sicht. Berichte aus dem Bayerischen Landesamt fu¨r Wald und Forstwirtschaft 27, 49–64. Litt, T., Ellwanger D., Villinger, E., Wansa, S., 2005. Das Quarta¨r in der Stratigraphischen Tabelle von Deutschland 2002. Newsletter of Stratigraphy 41, 385–399. Mania D. (Ed.), 1997. Bilzingsleben V, Homo erectus – seine Kultur und Umwelt. Verlag AusbildungþWissen Bad Homburg Leipzig.

Mania, D., Thomae, M., 2006. Pleistoza¨nstratigraphie und Pala¨olithikum im mittleren Elbe-Saale-Gebiet. 73. Tagung der Arbeitsgemeinschaft Norddeutscher Geologen 6.–9. Juli 2006 Halle. Mania, D., Thomae, M., Litt, T., Weber, T., 1990. Neumark-Gro¨bern. Beitra¨ge zur Jagd des mittelpala¨olithischen Menschen. Vero¨fftlichungen des Landesmuseums fu¨r Vorgeschichte in Halle 43, pp. 319. Pfeiffer, T., 1999. Sexualdimorphismus, Ontogenie und innerartliche Variabilita¨t der pleistoza¨nen Cervidenpopulationen von Dama dama geiselana PFEIFFER 1998 und Cervus elaphus L. 1758 (Cervidae, Mammalia) aus Neumark-Nord (Sachsen-Anhalt, Deutschland). Berliner geowissenschaftliche Abhandlungen E 30, 207–313. Sarntheim, M., Stremme, H.E., Mangini, A. 1986. The Holsteinian interglacial: Time-stratigraphic position and correlation to stable-isotope stratigraphy of deep-sea sediments. Quaternary Research 26, 283–298. Schreve, D.E., 2001. Mammalian evidence from the Middle Pleistocene fluvial sequences for complex environment change at the oxygene isotope substage level. Quaternary International 79, 65–74. Schreve, D.E., Bridgland, D.R., 2002. Correlation of Emglish and German Middle Pleistocene fluvial sequences based on mammalian biostratigraphy. Netherlands Journal of Geoscienes 81, 357–373. Stuart, A. J., 1981. A comparison of the Middel Pleistocene mammal faunas of Voigtstedt (Thuringia, GDR) and West Runton (Norfolk, England). Quarta¨rpala¨ontologie 4, 155–163. Weber, T., Litt, T., Scha¨fer, D., 1996. Neue Untersuchungen zum a¨lteren Pala¨olithikum in Mitteldeutschland. Terra and Prehistoria – Beitra¨ge zur Ur- und Fru¨hgeschichte Mitteleuropas 9, 13–19.

30. MIS 5 to MIS 8 – Numerically Dated Palaeontological Cave Sites of Central Europe Wilfried Rosendahl1, Doris Do¨ppes2 and Stephan Kempe2 1

Reiss-Engelhorn-Museen, Zeughaus C5, D-68159 Mannheim, Germany Institut fu¨r Angewandte Geowissenschaften, Schnittspahnstr. 9, D-64287 Darmstadt, Germany

2

ABSTRACT Caves are among the most important sites preserving Quaternary fossils. Owing to the CaCO3-rich environment and the protection against erosion, even remains of early Pleistocene faunas are preserved in caves, while contemporaneous surface deposits have been lost. However, faunal remains cannot be linked to any interglacial or glacial period since no species exists which is characteristic of any specific period. Reliable dating of such remains is therefore required. This is now possible applying 230Th/U dating of speleothems. ESR dating of speleothems or 230Th/U dating of bones is, however, of disputable value. Dating of the base and top speleothem accretional layers permit assigning Pleistocene faunal remains to the MIS chronology. In this paper, we present for the first time an overview of all numerically dated palaeontological cave sites in Central Europe between MIS 5 and MIS 8. From twelve sites, a total of 31 strata were dated, most of them deposited during MIS 5; the rest belongs to MIS 6 and MIS 7; and only one sample representing MIS 8 provided reliable numerical dates. Numerically dated palaeontological cave sites older than MIS 8 are not known. 30.1 INTRODUCTION Caves are terrestrial depositories that may store a large variety of organic and inorganic remains. The latter may contain

important climatic and ecological information on Quaternary climate cycles. Caves are connected to the surface by joints, pits, sinkholes, dolines and horizontal passages. These serve to admit solid, fluid and gaseous matters directly or by way of intermediate deposits for final deposition, reflecting the changes of the environment above. Palaeontological remains are brought into the cave through a variety of processes. The excellent preservation of bones, molluscs and sometimes even other organic remains is due to the constant cave climate in which seasonal temperature changes are largely missing, and which has a rather constant humidity, wet in temperate climates and dry in desert settings. Even more important is the carbonate chemistry of the seepage waters and of the sediments which prevent dissolution of bones. Thus, only the soft tissue is lost due to bacterial decomposition, while the calcite saturated groundwater does not dissolve bones, teeth and shells. Conditions are also suitable for preserving pollen and charcoal (Bastin, 1978; Quinif and Bastin, 1994; McGarry and Caseldine, 2004) and even insect remains (Davis, 1999; Rosendahl and Kempe, 2002). Palaeontological remains from caves – specifically skeletal parts of Pleistocene large mammals – were among the first to be described scientifically (e.g. Esper, 1774; Rosenmu¨ller, 1794; Cuvier, 1805; Buckland, 1823). Their investigation marked the beginning of scientific cave research in Europe (e.g. Shaw, 1992). The dating of many species of Pleistocene mammals was established using bone deposits from caves or karst pits,

456

Wilfried Rosendahl, Doris Do¨ppes and Stephan Kempe

among them cave bear, cave hyena, cave lion, mammoth and woolly rhinoceros. Even though many mechanisms may cause deposition of bones in caves, it is generally possible to separate them into three groups (e.g. Zapfe, 1954; Rosendahl, 1995). The first group comprises the ‘cave dwellers’, animals which actively enter caves for temporary protection or to hibernate. These include a large variety of bat species and also large mammals such as the cave bear (Kurte´n, 1976; Rabeder et al., 2000; Rosendahl et al., 2000), whose skeletons accumulated in large quantities in the so-called ‘bear caves’. The second group of fossils is due mainly to the activity of the cave hyena (Crocuta crocuta spelaea), a distinctly larger subspecies of the still-existing African spotted hyena (Crocuta crocuta). Hyenas can feed on bones; but in order to do so, they need to retreat to a protected site. Therefore, large bones, specifically those of the large Pleistocene herbivores, were carried individually into caves and partly consumed (e.g. Zapfe, 1939). ‘Hyena den caves’ therefore contain a variety of large, partly consumed, individual bones of those mammals living in the vicinity of the cave, but they never contain complete skeletons. Because cave hyenas are climate indifferent, the bone deposits have been accumulated both in interglacial and in glacial times. Also human activity can be a reason for bone accumulations in caves. Sometimes hyena and human occupation of the same cave have been observed (e.g. Weinbergho¨hlen bei Mauern, Koenigswald et al., 1974). ‘Prey animal deposits’ can also contain bones of small mammals when they derive from owls. At their daytime roosts owls and other nocturnal birds of prey regurgitate undigested skeletal remains, which can form sizeable bone deposits over the millennia. Such roosts are often found in cave entrances or in niches under overhanging ledges (abris). These deposits are most important for Quaternary palaeontology because small mammals are much better suited for the reconstruction of palaeo-

climatic and palaeoecological conditions than the occurrence of large mammals (Koenigswald, 1973; Heinrich, 1982, 1987). Small mammals occupy very characteristic environmental niches, and they react faster to climatic alterations than large mammals. The third group of bone deposits is formed accidentally. Either animals fall through fissures, pits and sink holes into the caves or their bones are secondarily transported into the caves by water or sediment. Thus, they may occur in deposits without much matrix as bone beds, or they may be part of a larger volume of clastic sediments. In this group, the distribution of species best reflects the fauna in the vicinity of the cave. Thus, fossils from caves can deliver important palaeoecological clues. Since many species are highly sensitive to climate and specific environments, their presence is a proxy for past environmental conditions. Cave fossils or fossil communities, however, cannot be used as dating tools a priori, since the same communities or animals may reoccur several times as the climate shifts between colder and warmer conditions. This is specially a problem during the MIS 6 and MIS 4. Nevertheless, certain mammal species do show a pronounced evolution throughout the Pleistocene and changes in certain physical characteristics, such as the structure of their teeth that can be used for an age proxy. When comparing characteristics of a certain species between different localities, it is often possible to deduce a relative temporal succession. Specifically certain small mammals, which in general show a fast evolutionary adaptation, have been found to be very valuable in this respect (Koenigswald, 1992). The Arvicolides for example play an important role for the biostratigraphic dating of the Middle Pleistocene since they evolved rapidly during this time interval (e.g. Koenigswald, 1973, 1992; Koenigswald and Heinrich, 1999). But even with these tools, only a relative age determination within the Pleistocene can be obtained (Lower Pleistocene, older or younger Middle Pleistocene or Upper Pleistocene)

Numerically Dated Palaeontological Cave Sites

(Koenigswald and Heinrich, 1996, 1999). Paralleling certain fossils with specific glacials or interglacials or Marine Isotope Stages (MIS) is only possible by numeric dating. Within the discussed time frame, only the 230 Th/U methods (alpha-spectrometry or TIMS) (e.g. Edwards et al., 1986/87) and with very strict limitations, the ESR method (e.g., Rink, 1997) can deliver useful age dates. In the following overview, we therefore summarize only those Central European cave sites (Fig. 30.1) for which numeric age data covering the period MIS 5 to 8 are available. The information regarding the respective sites was taken from the cited literature. A detailed critical discussion of the dating methods and the validity of the respective dates are, however, beyond the scope of this paper. It is necessary to point out that the individual dates obtained from bones or teeth must be viewed critically due to the open-system problem inherent to all samples which

457

potentially suffered exchange with a fluid phase during their depositional history.

30.2 SITE DESCRIPTIONS 1 – Grotte Scladina Location: Belgium, Province of Namur, commune of Andenne, village of Sclayn Coordinates: 2 1893899E, 49 1794399N Altitude: 137.7 m a.s.l. (metre above sealevel) Local cave register number: Not registered Geographical position: The Scladina Cave is located in Sclayn, a village at the right bank of the Meuse Valley, ca. 60 km SE of Brussels. Site description: The 30 m long and 6 m high, tunnel-like cave was discovered in 1971; it is excavated scientifically since 1978 (Otte, 1992; Otte et al., 1998).

Fig. 30.1 Map of the reported Central European cave sites (numbers refer to the text).

458

Wilfried Rosendahl, Doris Do¨ppes and Stephan Kempe

Stratigraphy: The main profile is 5.5 m thick and consists of intercalated clastic sediments and flowstone. It can be divided into 19 units (Otte et al., 1983). From top to bottom, these units are CC1, 36, 37, 38, 39 and 40, 1A, 1B, 2A, 2B, 3, CC4, 4A with CC14, 4B, 5, 6, 7A and 7B. Dating: Several flowstone layers were dated by the 230Th/U method in several laboratories (Gewelt et al., 1992). Four dates for unit 3 yielded an average age of 83 23 kyr. Layer CC4 was dated seven times. Two cores yielded mean ages of 114 23 and 110 14 kyr, respectively. This layer was also dated by TL yielding at its top 117:2 11 kyr and 122 11 kyr at its base (Debenham, 1998). From unit 5, a burnt artifact was also dated by TL and yielded 130 20 kyr (Huxtable and Aitken, 1992). According to magnetostratigraphical results, unit 3 coincides with MIS 5b, unit 4 with MIS 5c and unit 5 with MIS 5d (Ellwood et al., 2004). Fauna: Unit 3: Panthera leo spelaea, Crocuta crocuta spelaea, Canis lupus, Vulpes vulpes, Ursus spelaeus, Ursus arctos, Sus scrofa, Cervus elaphus, Rangifer tarandus, Dama dama, Capreolus capreolus, Bovinae, Capra ibex, Rupicapra rupicapra, Equus caballus, Hystrix cristata, Lepus sp. (Simonet, 1992). Microtus arvalis/agrestis, Microtus sp., Arvicola terrestris, Clethrionomys glareolus, Apodemus cf. sylvaticus, Talpa europaea, Citellus sp. (Cordy, 1992). Unit 4: Panthera leo spelaea, Panthera pardus, Felis sylvestris, Crocuta crocuta spelaea, Canis lupus, Cuon sp.?, Vulpes vulpes, Alopex lagopus, Ursus spelaeus, Ursus arctos, Ursus sp., Meles meles, Mustela putorius, Martes martes, Sus scrofa, Cervus elaphus, Rangifer tarandus, Dama dama, Capreolus capreolus, Bos primigenius, Bison priscus, Bovinae, Capra ibex, Rupicapra rupicapra, Mammuthus primigenius, Coelodonta antiquitatis, Equus caballus, Hystrix cristata, Lepus sp., Castor fiber (Simonet, 1992). Lagurus lagurus, Citellus sp., Dicrostonyx gulielmi, Microtus gregalis, Microtus oeconomus, Microtus arvalis/agrestis, Pitymys subterraneus, Microtus sp., Arvicola terrestris, Clethrionomys glareolus, Apodemus

cf. sylvaticus, Talpa europaea, Lemmus lemmus Chiropetra (Cordy, 1992). Unit 5: Panthera leo spelaea, Crocuta crocuta spelaea, Canis lupus, Ursus spelaeus, Ursus arctos, Ursus sp., Mustela foina, Sus scrofa, Cervus elaphus, Rangifer tarandus, Dama dama, Capreolus capreolus, Capra ibex, Rupicapra rupicapra, Mammuthus primigenius, Coelodonta antiquitatis, Equus caballus (Simonet, 1992). 2 – Einhornho¨hle Location: Germany, Lower Saxony, Harz, village of Scharzfeld Coordinates: 10 2491099E, 51 3891299N Altitude: 370 m a.s.l. Local cave register number: 4328/04 Geographical position: The Einhornho¨hle (Unicorn Cave) is located in the southern Harz Mountains near Scharzfeld, ca. 100 km SE of Hannover. Site description: The cave is more than 600 m long and consists of a string of large halls and domes connected by low passages. The first written record dates back to 1541, but the cave itself has been known much longer. It was an important source of bones sold as Unicornu fossile for medical purposes. Virchow (1872) began scientific excavations that continued until recently (e.g. Nielbock, 2003). The most recent excavation lasted from 1984 to 1988 (Nielbock, 2002). Stratigraphy: The 1.5-m thick standard profile from the ‘Weißer Saal’ encompasses the following layers: The ‘cave bear loam’ is sandwiched between a younger flowstone layer and a fossiliferous clay layer (Nielbock, 1987). In the Jacob-Friesen Passage, a more than 2-m thick profile yielded nine layers (0 – upper unit, H – lowest unit) (Nielbock, 1987). Dating: Cave bear bones of the ‘Weißer Saal’ yielded ESR dates between 95 and 104 kyr (Nielbock, 1987). Cave bear bones from the Jacob-Friesen Passage (units D to H) yielded 230Th/U dates of 126 þ10=9 kyr and 173 þ19=16 kyr (Wild et al., 1988). Fauna: ‘Weißer Saal‘; Canis lupus, Ursus spelaeus, Panthera leo spelaea (Nielbock, 1987).

Numerically Dated Palaeontological Cave Sites

Jacob-Friesen Passage (units D to H): Talpa europaea, Sorex araneus, Arvicola terrestris, Microtus nivalis, Microtus arvalis, Microtus agrestis, Microtus oeconomus, Canis lupus, Ursus spelaeus (Nielbock, 1987).

3 – Hunas Location: Germany, Bavaria, Fra¨nkische Alb, village of Hartmannshof Coordinates: 11 3294199E, 49 3091799N Altitude: 520 m a.s.l. Local cave register number: A 236 Geographical Position: In a limestone quarry near the Village of Hartmannshof, ca. 40 km E from Nuremberg. Site description: The cave was discovered in 1956 and investigated in the following eight years (Heller, 1983). The initial excavation opened the youngest part of an extensive stratigraphic sequence. In anticipation of the complete destruction of the site by the quarry, new excavations were started in 1983 and still go on. Stratigraphy: The cave is completely filled with layered sediments. The roof itself collapsed, covering the sediment-fill and sealing the cave entrance. About 12 m of sediments were investigated since 1983. The sediment stack can be divided into 22 layers. From top to bottom, these are the units A, B, C, D, E, F1, F2, G1, G2, G3, H, J, Koben, Kmitte, Kunten, Loben, Lmitte, Lunten, M, N, O and P (Rosendahl et al., 2006). Dating: In 2002, a flowstone layer was discovered at the base of the section (unit P). The layer is clearly connected with the cover sediment series without interruption. A stalagmite from this layer was mass spectrometrically dated (TIMS) by the 230Th/U method. The base yielded an age of 79 8 kyr and the top an age of 77 10 kyr (Rosendahl et al., 2005, 2006). Fauna: Unit O–P: Various genera and species of smaller and larger mammals occur, which have not been determined in detail as yet (Hilpert, personal communication).

459

4 – Conturines-Ho¨hle Location: Italy, South Tyrol, Dolomites, village of S.Vigilio-Marebbe Coordinates: 11 599E, 46 349N Altitude: 2775 m a.s.l. Local cave register number: Not registered Geographical position: The ConturinesHo¨hle is located at the eastern slope of Piz dles Conturines, ca. 65 km E of Bozen. Site description: Fossil bone remains in the 160 m long Conturines-Ho¨hle are known since 1987 (Rabeder, 1991). Excavations were conducted from 1988 to 1990 (Rabeder, 1991), from 1996 to 1998 and in 2001. Stratigraphy: The floor of the upper parts of the cave is covered by thick flowstone. It is overlain in turn by fossiliferous dolomitic sand that is buried below large blocks (Rabeder, 1991). Dating: The basal flowstone is older than 350 kyr, beyond the range of dating by the 230 Th/U method (Frisia et al., 1993). The bone-bearing sands are much younger; the two oldest dates obtained by dating the bones are 87 5 kyr and 108 þ8=7 kyr (Withalm, 1995). Fauna: Marmota marmota, Ursus spelaeus, Panthera leo spelaea (Rabeder, 1991). 5 – Ramesch-Knochenho¨hle Location: Austria, Upper Austria, Totes Gebirge, village of Spital am Pyhrn Coordinates: 14 159E, 47 399N Altitude: 1960 m a.s.l. Local cave register number: 1636/8 Geographical position: The RameschKnochenho¨hle is located at the northface of the Ramesch, a peak in the Warscheneckgruppe/Totes Gebirge, W of Spital am Pyhrn, ca. 75 km S of Linz. Site description: Behind a wide entrance, a large, 30 m long hall with an almost level floor opens. The total length of the cave amounts to 310 m. First palaeontological excavations were conducted between 1979 and 1984 (Draxler et al., 1986).

460

Wilfried Rosendahl, Doris Do¨ppes and Stephan Kempe

Stratigraphy: Undisturbed deposits are found only in the entrance hall (Draxler et al., 1986). Below a Holocene layer with gastropods (unit A) a typical cave loam with cave bear bones occurs (units B–E). Further down sterile loams fill the interstices between large blocks below dark sediments with cave bear remains, allochthonous pebbles (unit G) and fossil free pebblebearing sands (so-called ‘Augensteinsande’, unit H). Dating: Cave bear bones from unit G yielded 230Th/U dates between 117 þ11=10 kyr and 150 þ25=19 kyr (Draxler et al., 1986). Fauna: Unit G; Ursus spelaeus (Draxler et al., 1986).

6 – Schwabenreith-Ho¨hle ¨ tscherLocation: Austria, Lower Austria, O Lunzersee-Hochkar, village of Lunz am See Coordinates: 14 5893899E, 47 5093399N Altitude: 959 m a.s.l. Local cave register number: 1823/32 Geographical position: The Schwabenreith Ho¨hle is located S of the farm of Schwabenreith, W of the village of Lunz/ See, and ca. 120 km SE of Linz. Site description: The 134-m long cave has two nearly horizontal passages. The cave has been known locally for a long time but was just investigated only at the end of the 1960s (Hartmann and Hartmann, 1969). Excavations lasted from 1990 to 2000 (Fladerer, 1992; Pacher, 2000). Stratigraphy: The terminal hall yielded most of the fossils (Site 2) where stalagmites sit directly on the rock floor. Above them a fossil-free loamy sand (unit 6) was deposited followed by a 1.3-m thick fossiliferous layer (unit 5), which is covered by a layer of flowstone (unit 2). Dating: The basal and top (unit 2) flowstone layers yielded 230Th/U ages of 116 5 kyr and 78 þ 30=23 kyr, respectively (Frank and Rabeder, 1997b). Fauna: Ursus spelaeus (Pacher, 2000).

7 – Herdengelho¨hle Location: Austria, Lower Austria, ¨ tscher-Lunzersee-Hochkar, village of O Lunz am See Coordinates: 14 5893899E, 47 5092599N Altitude: 878 m a.s.l. Local cave register number: 1823/4 Geographical position: The cave is situated SW of the farm Herdengel, W of the village of Lunz am See, ca. 120 km SE of Linz. Site description: The cave is 129 m long. Digs conducted in 1927 and 1928 remained without success but in 1935 fossils were discovered (Abrahamczik, 1936). Modern scientific excavations were conducted from 1983 to 1989 (Nagel and Rabeder, 1991; Frank and Rabeder, 1997a). Stratigraphy: The 8-m thick sediments contain six layers. The lowest layer (down to 750 cm) is a sterile, fine sand covered by flowstone. Unit 1 (380–430 cm) above it is a black yellow loam which contained bones stained black. Unit 2 (360–380 cm), a flowstone layer, shows well-developed stalagmites. Unit 3 (330–360 cm) and unit 4 (300–330 cm) contained partially-to-well preserved brown-stained bones. Unit 5 (280–300 cm) is composed of blocks and reddish loam. Unit 6 (200–280 cm) forms a 2-m thick cover composed of fossil-free, light yellow loam (Leitner-Wild et al., 1994). Dating: The flowstone layer of unit 2 was 230 Th/U dated to 111þ11=10 kyr. The cave bear bones of unit 1 date back to the period from 135 þ11=10 kyr to 127 7 kyr (Leitner-Wild et al., 1994). Fauna: Unit 1; Ursus spelaeus (Frank and Rabeder, 1997a). 8 – Repolustho¨hle Location: Austria, Styria, Mittelsteirischer Karst, village of Frohnleiten Coordinates: 15 2095199E, 47 1893599N Altitude: 525 m a.s.l. Local cave register number: 2837/1 Geographical position: The Repolustho¨hle is located in the valley of Badl, ca. 20 km N of Graz.

Numerically Dated Palaeontological Cave Sites

Site description: The cave was discovered in 1910 and is 66 m long. It ends in a natural filled pit. Excavations were conducted between 1947 and 1955 (Mottl, 1951; Mottl and Murban, 1955). The sediments of the pit were investigated by Temmel (1996), and the material of old excavations was revised by Fu¨rnholzer (1997). Stratigraphy: The sedimentary profile in the pit (from the bottom to the top) shows loam and clay followed by rust-coloured phosphate-rich sediments with manganese streaks, grey sands and a rust-coloured phosphate-rich soil (Mottl and Murban, 1955). Dating: A cave bear bone from the lowest layers in the pit yielded a 230Th/U age of 230 þ 13=12 kyr (Fu¨rnholzer, 1997). Fauna: Lower rust-coloured phosphaterich sediment; Aves, Talpa europaea, Sorex cf. araneus, Myotis bechsteini, Plecotus auritus, Marmota marmota, Spermophilus cf. citellus, Cricetus major, Apodemus sylvaticus, Apodemus flavicollis, Clethrionomys glareolus, Arvicola hunasensis, Microtus arvalis, Hystrix cf. vinogradovi, Lepus sp., Canis lupus ?, Canis mosbachensis, Vulpes vulpes, Cuon alpinus ssp., Ursus arctos, Ursus deningeri, Martes martes, Meles meles, Mustela nivalis, Putorius sp., Felis silvestris, Lynx lynx, Panthera pardus, Panthera leo spelaea, Sus scrofa, Cervus elaphus, Megaloceros giganteus, Capreolus capreolus, Rangifer tarandus, Bison priscus, Rupicapra rupicapra, Capra ibex, Elephantidae indet (Rabeder and Temmel, 1997). 9 – Divje babe I Location: Slovenia, Innerkrain (Notranjska), village of Reka Coordinates: 13 549E, 46 019N Altitude: 450 m a.s.l. Local cave register number: 812 Geographical position: The Divje Babe I site is located in the Idrija valley near the village of Reka, at the western slope of the Sˇebrelje plateau, ca. 60 km W of Ljubljana. Site description: The cave is approximately 45 m long and 45 m wide. Systematic excavations lasted from 1980 to 1986 (Turk

461

et al., 1989) and from 1989 to 1995 (Turk et al., 2002). Stratigraphy: This Upper Pleistocene site contains a long stratigraphic sequence spanning the period from approximately 120 to 35 kyr (Turk et al., 2001). The 11.5-m thick profile is divided into 21 units; the oldest is unit 21 (Turk et al., 2002). Dating: 230Th/U dating of a bear bone yielded 81 10 kyr (unit 20) and 84 7 kyr for unit 19 (Nelson, 1997). Fauna: Unit 20–17; Ursus spelaeus (99%) (Turk et al., 1989). 10 – Vindija Location: Croatia, Hrvatsko Zagorje, village of Ivanec Coordinates: 16 2093899E, 46 199N Altitude: 275 m a.s.l. Local cave register number: Not registered Geographical position: Vindija is located on the southwest side of the Kriznjak Peak, 55 km N of Zagreb. Site description: The cave chamber is approximately 50 m long, 28 m wide and more than 10 m in height. Vindija was first mentioned in 1878, and excavation was started in 1928 (Malez, 1979). Extensive archaeological and palaeontological excavations lasted from 1974 to 1986 and from 1993 to 1994 (Malez et al., 1984; Karavanic´, 1995). Stratigraphy: The 12-m thick stack of sediments in Vindija cave can be divided into 13 layers designated from unit A (youngest) to unit M (oldest). Dating: 230Th/U dates from cave bear bones of unit J range from 156 2 kyr to 196 þ20=15 kyr and of the underlying unit K from 150þ16=13 kyr to 212þ17=13 kyr (Wild et al., 2001). Fauna: Unit J: Marmota marmota, Canis lupus, Cuon alpinus europaeus, Ursus spelaeus, Panthera leo spelaea, Cervus elaphus, Capra ibex and Bovidae indet. (Malez and Ullrich, 1982). Unit K: Canis lupus, Ursus spelaeus, Panthera leo spelaea, Panthera pardus, Crocuta crocuta spelaea, Dicerorhinus kirchenbergensis,

462

Wilfried Rosendahl, Doris Do¨ppes and Stephan Kempe

Sus scrofa, Megaloceros giganteus, Cervus elaphus, Dama dama, Capreolus capreolus, Bos primigenius (Malez and Ullrich, 1982). 11 – Krapina Location: Croatia, Hrvatsko Zagorje, village of Krapina Coordinates: 15 529E, 46 109N Altitude: 120 m a.s.l. Local cave register number: Not registered Geographical position: Krapina (Husˇnjakovo rock shelter) is located on the western side of the Husˇnjak hill, W of Krapina, 42 km NW of Zagreb. Site description: The rock shelter was ca. 12 m high. The site was excavated from 1899 to 1905 (Gorjanovic´-Kramberger, 1906). Stratigraphy: The 12-m thick stack of sediments in Krapina can be divided into 10 layers designated unit I (oldest) to unit 9 (youngest). Dating: The age of tooth enamel of hominids from unit 9 yielded a 230Th/U date of 113 10 kyr and an ESR date of 87 7 kyr. The ages of teeth from units 1–8 were indistinguishable, with a mean of 130 10 kyr (Rink et al., 1995). Tooth enamel of hominids from unit 9 to 6 and unit 1 was dated by U-series dates and ESR (Rink et al., 1995). Fauna: Unit 9: Lepus sp., Marmota marmota, Canis lupus, Ursus spelaeus, Lynx lynx, Cervus elaphus, Rupicapra rupicapra (Malez, 1970) Units 7 and 8: Aves, Amphibia, Castor fiber, Myoxus glis, Canis lupus, Ursus spelaeus, Ursus arctos priscus, Mustela putorius, Martes martes, Felis silvestris, Panthera pardus, Sus scrofa, Cervus elaphus, Capreolus capreolus, Bos primigenius, Stephanorhinus kirchbergensis, Homo neanderthalensis (Malez, 1978). Units 5 and 6: Cricetus cricetus, Canis lupus, Vulpes vulpes, Ursus spelaeus, Ursus arctos priscus, Mustela cf. eversmanni, Lynx lynx, Cervus elaphus, Alces alces, Bison priscus, Equus cf. germanicus, Homo neanderthalensis (Malez, 1978). Unit 1: Castor fiber, Panthera pardus, Stephanorhinus kirchbergensis, Homo neanderthalensis (Malez, 1970).

12 – Bis´nik Jaskinia Location: Poland, Province of Ło´dz´, Krakow Cze˛stochowa Uplands, village of Pilica Coordinates: 19 559E, 50 289N Altitude: 395 m a.s.l. Local cave register number: Not registered Geographical position: The Bis´nik Jaskinia is located near the village of Pilica in the central part of the NiegowonickoSmolen´skie Hills, ca. 50 km N of Cracow. Site description: The cave is about 73 m long. It consists of a rock shelter and a cave proper connected to it. Excavation lasted from 1991 to 2000 (Mirosław-Grabowska, 2002). Stratigraphy: The more than 7-m thick clastic sediments can be divided into 18 layers. The lowest series consists of layers 8 to 18 (Mirosław-Grabowska, 2002). Dating: 230Th/U dating of bones from layers 12 and 13 yielded an age range of 115–128 kyr, for layer 14 from 128 to 200 kyr, for layer 15 from 200 to 250 kyr and for layers 16 and 17 from 250 to 270 kyr (MirosławGrabowska, 2002). Fauna: Layers 12, 13: Ursus spelaeus, Cervus elaphus, Capreolus capreolus, Rangifer tarandus, Bos primigenius, Equus caballus, Clethrionomys glareolus (Mirosław-Grabowska, 2002). Layer 14: Canis lupus, Ursus spelaeus, other carnivores, Sus scrofa, Megaloceros giganteus, Cervus elaphus, Rangifer tarandus, Bison priscus, Equus caballus, Arvicola terrestris, Microtus oeconomus, Clethrionomys glareolus (Mirosław-Grabowska, 2002). Layer 15: Canis lupus, Vulpes vulpes, Ursus spelaeus, other carnivores, Sus scrofa, Cervus elaphus, Capreolus capreolus, Rangifer tarandus, Bison priscus, Equus caballus, Arvicola terrestris, Microtus oeconomus, Apodemus sylvaticus, Clethrionomys glareolus (Mirosław-Grabowska, 2002). Layers 16, 17: Ursus spelaeus, Crocuta crocuta spelaea, Vulpes vulpes, other carnivores, Cervus elaphus, Megaloceros giganteus, Capreolus capreolus, Alces alces, Rangifer tarandus, Bison priscus, Coelodonta antiquitatis, Equus caballus (Mirosław-Grabowska, 2002).

Numerically Dated Palaeontological Cave Sites

30.3 CONCLUSIONS In spite of the fact that numerous palaeontologically important cave sites are known in Central Europe which may be dated into the time period MIS 5 to MIS 9 (e.g. Koenigswald and Heinrich, 1996, 1999; Do¨ppes and Rabeder, 1997), we found only twelve sites in the literature for which numerical dates have

463

been published. A total of 31 layers have been dated in these sites (Fig. 30.2). Most of them yielded dates corresponding to MIS 5, a few represent MIS 6 and 7 and only one can be correlated with MIS 8. Older numerically dated sites have not yet been found. More than half of the faunal strata have been dated because they are part of archaeologically important sites. Numerical dates of purely

Fig. 30.2 Timetable and assignment of cave deposits to the MIS chronology. Dot date without standard deviation, dot and line date with a standard deviation, solid line time range of strata with several dates. Explanation of abbreviations: BC Bis´nik Jaskinia (12 – number of site description), Cu Conturines-Ho¨hle (4), DB ¼ Divje babe I (9), Hd Herdengelho¨hle (7), Hu Hunas (3), Kr Krapina (11), Re Repolustho¨hle (8), Rk Ramesch-Knochenho¨hle (5), SC Grotte Scladina (1), Sw Schwabenreith-Ho¨hle (6), Uc Einhornho¨hle (2), Vi Vindija (10).

464

Wilfried Rosendahl, Doris Do¨ppes and Stephan Kempe

palaeontological sites are rare even though the archive of caves and their rich bone beds offer enough potential for a systematic study. There are several reasons why this has not been done yet: First of all, cave sites are not easily accessible (compared to open air sites); second, the mechanisms of deposition are complicated and can only be understood by investigating the genesis of the respective cave and its sediments in general. Cave studies therefore require additional knowledge (i.e. speleology) before the specific site can be interpreted correctly. The most important point, why so few dates exist is related to the dating itself. Beyond the reach of the 14C method, bones, as mentioned above, can only be dated by the 230Th/U or ESR methods. The 230Th/U method is the one most commonly used. But ESR dates of bones are highly problematic and should not be used (e.g. the dates of the Einhornho¨hle, no. 2). Only ESR dates of tooth enamel appear to be correct (e.g. Krapina, no. 11). TL dating, used for speleothems (e.g. Grotte Scladina, no.1) exclusively, is methodologically also of doubtful quality, and TL dates should today be regarded with caution and their usage should be discontinued. Even the 230Th/U dating of bones is methodologically problematic due to the fact that bones very often prove to be open systems (for a discussion see Bischoff et al., 1995). The unusual standard deviations of the dates of unit J of Vindija (no. 10; Wild et al., 2001) may be caused by exactly this open-system problem. Therefore, it is essential that prior to the dating both the excavator and the dating geochemist discuss the chronostratigraphy, palaeoecology and palaeoclimatology of a site in detail. Even though, there are isotopists who view all 230Th/U bone dates critically and suggest they be discarded all together (e.g. Geyh, 2005). The 230Th/U-dating technique of speleothems has been improved substantially, resulting in more reliable results since the 1990s. These have been used for the reconstruction of Middle to Upper Pleistocene climate and environment (e.g. Winograd

et al., 1992; Kempe et al., 2002; Genty et al., 2003; Holzka¨mper et al., 2005), the dating of bones is lagging behind. Only a few laboratories (e.g. Vienna and Warsaw) are currently applying it. It would therefore be profitable if the technique of dating bones with 230Th/U could also be improved. Interesting suggestions in this direction have been made by Hercman and Gorka (2000), Pike et al. (2002) and Eggins et al. (2005). Hoffmann and Mangini (2003) also describe an interesting method to date teeth and perhaps bones from open systems. Even though speleothem dating with the TIMS 230Th/U method is not entirely free of methodological problems, TIMS speleothem dates are the best dates available today to establish cave-based chronologies. Flowstone layers above or below the faunal strata can thus be dated, bracketing the ages of the bones (e.g. Schwabenreith-Ho¨hle, no. 6). Those sites, which do not have speleothem supported age models, should therefore be revisited, and additional samples should be dated to give the currently available bone dates further credibility. For climatic and ecological investigations of speleological faunas, we should therefore target those cave sites which can be dated via speleothems. Additionally, bones could be dated with TIMS 230Th/U and teeth with ESR in order to advance dating techniques in general. Since the open-system problem induces a substantial inaccuracy regarding dates of bones and teeth, a critical assessment of specific faunal assemblages with respect to their exact stratigraphical position remains difficult. In addition, numerical dates have a certain standard deviation caused by methodological problems. This deviation can be substantial with the consequence that only two of the faunal assemblages suitable for ecological discussion can be attributed to either a glacial or an interglacial. These two faunas are layers 12–13 of the Bisnik Jaskinia (attributed to MIS 5e) and the fauna recovered from the pit of the Repolust Cave (attributed to MIS 7).

Numerically Dated Palaeontological Cave Sites

MIS 5 Insectivora Sorex araneus Talpa europaea Chiroptera Myotis bechsteinii Plecotus auritus Lagomorpha Lepus sp. Rodentia Marmota marmota Spermophilus sp. Spermophilus citellus Castor fiber Cricetus cricetus Cricetus major Lemmus lemmus Dicrostonyx gulielmi Clethrionomys glareolus Lagurus lagurus Arvicola terrestris Arvicola hunasensis Microtus sp. Microtus agrestis Microtus arvalis Microtus agrestis /arvalis Microtus gregalis Microtus oeconomus Microtus subterraneus Chionomys nivalis Apodemus flavicollis Apodemus sylvaticus Glis glis Hystrix cristata Hystrix cf. vinogradovi Carnivora Canis aureus Canis lupus Canis mosbachensis Cuon sp.? Cuon alpinus Alopex lagopus Vulpes vulpes Ursus sp. Ursus arctos Ursus spelaeus Ursus deningeri Mustela sp. Mustela eversmanii Mustela nivalis Mustela putorius Martes foina

Common shrew Common mole

+ +

465

MIS 6

MIS 7

+ +

cf. +

Bechstein’s bat Brown long-eared bat

+ +

Hare

+

Alpine marmot Suslik European suslik Eurasian beaver Common hamster Giant hamster Norway lemming Collared lemming Bank vole Steppe lemming Nothern water vole Water vole Vole Field vole Common vole Field/common vole Narrow-headed vole Tundra vole Pine vole Snow vole Yellow-necked mouse Wood mouse Edible dormouse Crested porcupine Porcupine

+ +

+

+ +

+ +

Golden jackal Grey wolf Tautavel wolf Dhole Indian dhole Arctic fox Red fox Bear Brown bear Cave bear Deninger bear Weasels Steppe polecat Weasel Western polecat Beech marten; Stone marten

MIS 8

+ + cf.

+ + + + + + + + + + + + +

+

+

+

+ +

+

+

+

+ cf. + +

+ + + cf.

+

+

? +

+

+

+ + + + + +

+ + +

cf.

cf.

+ +

+ +

+ + + + +

+

+

+

Fig. 30.3 (table) Faunal distribution of the reported cave sites from MIS 5 to MIS 8 without critical reflection on the numerical dates.

Wilfried Rosendahl, Doris Do¨ppes and Stephan Kempe

466

Martes martes Meles meles Felis silvestris Panthera leo spelaea Panthera pardus Lynx lynx Crocuta crocuta spelaea Artiodactyla Sus scrofa Dama dama Cervus elaphus Megaloceros giganteus Alces alces Rangifer tarandus Capreolus capreolus Bovidae indet. Bovinae Bos primigenius Bison priscus Rupicapra rupicapra Capra ibex Proboscidae Elephantidae indet. Mammuthus primigenius Perissodactyla Equus caballus Equus cf. germanicus Coelodonta antiquitatis

Pine marten Badger European wild cat Cave lion/Fig. 30.4, a Leopard European lynx Cave hyena Wild boar Fallow deer Red deer Giant deer Elk, Moose Reindeer Roe deer Bovid Bovid Aurochs Steppe wisent Alpine chamois Alpine ibex

MIS 5 + + + + + + +

MIS 6 +

+ + +

+ + + + + + + +

+

+ + + + +

+ + + +

+ + + +

+

+

+ + + +

+ + + +

+

+

+ + +

Wood elephant?/Fig. 30.4, b Woolly mammoth

Wild horse Horse Woolly rhinoceros/Fig. 30.4, c Stephanorhinus kirchbergensis Merck’s rhinoceros Primate Homo neanderthalensis Neanderthalian/Fig. 30.4, d

+ + + + +

MIS 7 + + + + + + +

+ + + +

MIS 8

+

+ + + + +

+

+

+

+ +

+

Fig. 30.3 Continued

In case of Bisnik Jaskinia, only the Rangifer tarandus component is in contrast to its Eemian age, but the faunal remains of the Repolust Cave combine both glacial and interglacial species, i.e. Rangifer tarandus and Megaloceros giganteus occur together with Capreolus capreolus and Sus scrofa. Thus, the presence of cold climate species is in contrast to the numerical interglacial date (Fig. 30.3, 30.4). The accuracy of the numerical dates of all other sites and layers do not permit to attribute the faunas into a specific glacial or interglacial. Thus the faunas cannot be evaluated regarding their ecological and climatic character. Contradicting occurrences of glacial and interglacial faunal elements cannot be resolved as long as the numerical dates allow for both possibilities.

In conclusion, the now available numerical dates of palaeontological sites in Central Europe do not allow – with the exception of Bisnik Jaskinia and the Repolustho¨hle Cave – a critical discussion of their faunal assemblages as to their ecological climatic distribution. But even those two sites are not without contradicting faunal elements, and it remains doubtful if they represent either glacial or interglacial faunas. In spite of all these problems, palaeontological cave sites represent a rich archive that can deliver important contributions to the reconstruction of the Middle and Upper Pleistocene palaeoclimate of Central Europe, provided many additional dates can be obtained to verify results obtained from other terrestrial archives.

Numerically Dated Palaeontological Cave Sites

(a)

(b)

(c)

(d)

467

Fig. 30.4 Lifestile reconstructions of different faunal elements named in Fig. 30.3: a=Panthera leo spelaea (climate indifferent); b = Elephas antiquus (interglacial); c = Coelodonta antiquus (glacial) d = Homo neanderthalensis (climate indifferent). All photographs W. Rosendahl, Mannheim.

ACKNOWLEDGEMENTS The authors thank the reviewers for critical and helpful remarks and Dr. M.S. Werner (Hilo/Hawaii) for editorial suggestions. REFERENCES Abrahamczik, W., 1936. Karsterscheinungen in der Umgebung von Lunz am See (mit besonderer Beru¨cksichtigung der Ho¨hlen). Unpublished Ph. D., University Vienna, 100 pp.

Bastin, B., 1978. L’analyse pollinique des stalagmites: une nouvelle possibilite´ d’approche des fluctuations climatiques du Quaternaire. Annales de la Socie´te´ Ge´ologique de Belgique 101, 13–19. Bischoff, J., Rosenbauer, R.J., Moench, F., 1995. Useries age equations handicap uranium assimilation fossil bones. Radiochimica Acta 69, 127–135. Buckland, W., 1823. Reliquiae Diluvianae; or, observations of the organic remains contained in caves, fissures, and diluvial gravel, and on other geological phenomena, attesting the action of an universal deluge. J. Murray, London, 2nd ed., 303 pp. Cordy, J.-M., 1992. Bio- et chronostratigraphie des de´pots quaternaires a` partir des

468

Wilfried Rosendahl, Doris Do¨ppes and Stephan Kempe

micromammife`res. E´tudes et recherches arche´ologiques de l’Universite´ de Lie`ge 27, 79–125. Cuvier, G.L., 1805. Sur les ossemens fossiles d’hye`nes. Annales du Muse´um d’Histoire Naturelle 6, 127–144. Davis, O.K., 1999. Pollen and other microfossils in Pleistocene speleothems, Kartchner Caverns, Arizona. Journal of Cave and Karst Studies 61, 89–92. Debenham, N.C., 1998. Thermoluminescence dating of stalagmitic calcite from la Grotte Scladina at Sclayn (Namur). E´tudes et recherches arche´ologiques de l’Universite´ de Lie`ge 79, 39–43. Do¨ppes, D., Rabeder, G. (Eds.) 1997. Plioza¨ne und ¨ sterreichs. Mitteilungen der pleistoza¨ne Faunen O ¨ sterreiKommission fu¨r Quarta¨rforschung der O chischen Akademie der Wissenschaften 10, 1–411. Draxler, I., Hille, P., Mais, K., Rabeder, G., Steffan, I., Wild, E., 1986. Pala¨ontologische Befunde, absolute Datierungen und pala¨oklimatologische Konsequenzen der Resultate aus der Ramesch-Knochenho¨hle. In: Hille, P., Rabeder, G. (Eds.), Die RameschKnochenho¨hle im Toten Gebirge. Mitteilungen der ¨ sterreiKommission fu¨r Quarta¨rforschung der O chischen Akademie der Wissenschaften 6, 7–66. Edwards, R.L., Chen, J.H., Wasserburg, G.J., 1986/87. 238U-234U- 230Th- 232Th systematic and the precise measurement of time over the past 500 000 years. Earth and Planetary Science Letters 81, 175–192. Eggins, S.M., Gru¨n, R., McCulloch, M.T., Pike, A.W.G., Chappell, J., Kinsley, L., Mortimer, G., Shelley, M., Murray-Wallace, C.V., Spo¨tl, C., Taylor, L., 2005. In situ U-series dating by laser-ablation multi-collector ICPMS: new prospects for Quaternary geochronology. Quaternary Science Reviews 24, 2523–2538. Ellwood, B.B., Harrold, F.B., Benoist, F.L., Thacker, P., Otte, M., Bonjean, D., Long, G.J., Shahin, A.M., Hermann, R.P., Grandjean, F., 2004. Magnetic susceptibility applied as an age-depth-climate relative dating technique using sediments from Scladina Cave, a Late Pleistocene cave site in Belgium. Journal of Archaeological Science 31, 283–293. Esper, J.F., 1774. Ausfu¨hrliche Nachricht von neuentdeckten Zoolithen unbekannter vierfu¨ssiger Thiere, und denen sie enthaltenden, so wie verschiedenen anderen, denkwu¨rdigen Gru¨ften der Obergebu¨rgischen Lande des Marggrafthums Bayreuth, Georg Wolfgang Knorrs seelige Erben, Nu¨rnberg, 148 pp. Fladerer, F., 1992. Erste Grabungsergebnisse aus der Schwabenreithho¨hle bei Lunz am See (Niedero¨sterreich). Die Ho¨hle 43, 84–92. Frank, C., Rabeder, G., 1997a. Herdengelho¨hle. In: Do¨ppes, D., Rabeder, G. (Eds.), Plioza¨ne und ¨ sterreichs. Mitteilungen der pleistoza¨ne Faunen O ¨ sterreiKommission fu¨r Quarta¨rforschung der O chischen Akademie der Wissenschaften 10, 181–185.

Frank, C., Rabeder, G., 1997b. Schwabenreith-Ho¨hle. In: Do¨ppes, D., Rabeder, G. (Eds.), Plioza¨ne und ¨ sterreichs. Mitteilungen der pleistoza¨ne Faunen O ¨ sterreiKommission fu¨r Quarta¨rforschung der O chischen Akademie der Wissenschaften 10, 227–231. Frisia, S., Bini, A., Quinif, Y., 1993. Morphologic, crystallographic and isotopic study of an ancient flowstone (Grotta di Conturines, Dolomites): implications for palaeoenvironmental reconstructions. Speleochronos 5, 3–18. Fu¨rnholzer, J., 1997. Repolustho¨hle (Kat. – Nr. 2837/1) – Revision der Grabungen von 1947 bis 1955. Jubi¨ sterreichischen Nationalbank la¨umsfonds der O Projekt Nr. 5691, 64 pp. Genty, D., Blamart, D., Ouhadi, R., Gilmour, M., Baker, A., Jouzel, J., Van-Exter, S., 2003. Precise dating of Dansgaard–Oeschger climate oscillations in Western Europe from stalagmite data. Nature 421, 833–837. Gewelt, M., Schwarcz, H.P., Szabo, B.J., 1992. Datations 230Th/234U et 14C de concre´tions stalagmitiques. E´tudes et recherches arche´ologiques de l’Universite´ de Lie`ge 27, 159–172. Geyh, M.A., 2005. Handbuch der physikalischen und chemischen Altersbestimmung. Wissenschaftliche Buchgesellschaft, Darmstadt, 211 pp. Gorjanovic´-Kramberger, K., 1906. Der diluviale Mensch von Krapina in Kroatien. C.W. Kreidels Verlag, Wiesbaden, 218 pp. Hartmann, H., Hartmann, W., 1969. Neue Ho¨hlen im Scho¨pftaler Wald. Ho¨hlenkundliche Mitteilungen des Landesvereins fu¨r Ho¨hlenkunde in Wien und Niedero¨sterreich 25, 113–115. Heinrich, W.-D., 1982. Zur Evolution und Biostratigraphie von Arvicola (Rodentia, Mammalia) im Pleistoza¨n Europas. Zeitschrift fu¨r Geologische Wissenschaften 10, 683–735. Heinrich, W.-D., 1987. Neue Ergebnisse zur Evolution und Biostratigraphie von Arvicola (Rodentia, Mammalia) im Quarta¨r Europas. Zeitschrift fu¨r Geologische Wissenschaften 15, 389–406. Heller, F. (Ed.) 1983. Die Ho¨hlenruine Hunas bei Hartmannshof (Landkreis Nu¨rnberger Land) – Eine pala¨ontologische und urgeschichtliche Fundstelle aus dem Spa¨t-Riß. Quarta¨r-Bibliothek 4, 408 pp. Hercman, H., Gorka, P., 2000. U-series dating of bones from Bisnik Cave – open system dating models. Climate Changes: the Karst Record II, Krakow, Abstract Volume, pp. 67. Hoffmann, D., Mangini, A., 2003. A method for coupled ESR/U-series dating of teeth showing post-depositional U-loss. Quaternary Science Reviews 22, 1367–1372. Holzka¨mper, S., Spo¨tl, C., Mangini, A., 2005. High Alpine flowstone provides new insights in the timing and progression of MIS 7, 5 and 3. Earth Planetary and Science Letters 236, 751–764.

Numerically Dated Palaeontological Cave Sites Huxtable, J., Aitken, M.J., 1992. Thermoluminescence dating of burned flint and stalagmitic calcite. E´tudes et recherches arche´ologiques de l’Universite´ de Lie`ge 27, 175–178. Karavanic´, I., 1995. Upper Paleolithic occupation levels and late-occurring Neandertal at Vindija Cave (Croatia) in the context of Central Europe and the Balkans. Journal of Anthropological Research 51, 9–35. Kempe, S., Rosendahl, W., Wiegand, B., Eisenhauer, A., 2002. New Speleothem Datation from Caves in Germany and their importance for the Middleand Upper-Pleistocene Climate Reconstruction. Acta Geologica Polonica 52, 55–61. Koenigswald, W. v., 1973. Vera¨nderungen in der Kleinsa¨ugerfauna von Mitteleuropa zwischen Cromer und Eem (Pleistoza¨n). Eiszeitalter und Gegenwart 23/24, 159–167. ¨ kologie und BiostratiKoenigswald, W. v., 1992. Zur O graphie der beiden pleistoza¨nen Faunen von Mauer bei Heidelberg. In: Beinhauer, K.W., Wagner, G.A. (Eds.), Schichten von Mauer – 85 Jahre Homo erectus heidelbergensis, Braus Mannheim, 101–110. Koenigswald, W. v., Heinrich, W.-D., 1996. Kurze Charakterisierung der Vera¨nderungen in der Sa¨ugerfauna des Jungquarta¨rs in Mitteleuropa. Tu¨binger Monographie zur Urgeschichte 11, 441 pp. Koenigswald, W. v., Heinrich, W.-D., 1999. Mittelpleistoza¨ne Sa¨ugetierfaunen aus Mitteleuropa – der Versuch einer biostratigraphischen Zuordnung. Kaupia 9, 53–112. Koenigswald, W. v., Mu¨ller-Beck, H., Pressmar, E., 1974. Die Archa¨ologie und Pala¨ontologie in den Weinbergho¨hlen bei Mauern (Bayern), Grabungen 1937–1967. Archaeologica Venatoria 3, 1–152. Kurte´n, B., 1976. The Cave Bear Story, Life and Death of a Vanished Animal. Columbia University Press, 163 pp. Leitner-Wild, E., Rabeder, G., Steffan, I., 1994. Determination of the evolutionary mode of Austrian alpine cave bears by uranium series dating. Historical Biology 7, 97–104. Malez, M., 1970. Die Ergebnisse der Revision der pleistoza¨nen Fauna aus Krapina. Krapina 1899– 1969, 45–56. Malez, M., 1978. Stratigraphische, pala¨ofaunistische und pala¨olithische Verha¨ltnisse des Fundortes Krapina. Krapinski pracovjek I evolucija hominida, Rad Jugoslavenske akademije znanosti i umjetnosti 61–102. Malez, M., 1979. Paleolitsko i mezolitsko doba u Hrvatskoj, Praistorija jugoslavenskih zemalja I, pp. 227–295. Malez, M., Simunis, An., Simunis, Al., 1984. Geoloki, sedimentoloki i paleoklimatski odnesi spilje Vindje i blize okolice. Rad Jugoslavenske akademije znanosti i umjetnosti 411, 231–264.

469

Malez, M., Ullrich, H., 1982. Neuere pala¨oanthropologische Untersuchungen am Material aus der Ho¨hle Vindija (Kroatien, Jugoslawien). Palaeontologia Jugoslavica 29, 1–44. McGarry, S.F., Caseldine, C., 2004. Speleothem palynology: an undervalued tool in Quaternary studies. Quaternary Science Reviews 23, 2389–2404. Mirosław-Grabowska, J., 2002. Geological value of Bis´nik Cave sediments (Cracow-Cze˛stochowa Upland). Acta Geologica Polonica 52, 97–110. Mottl, M., 1951. Die Repolustho¨hle bei Peggau (Steiermark) und ihre eiszeitlichen Bewohner (mit einem Beitrag von V. Maurin). Archaeologia Austriaca 8, 1–78. Mottl, M., Murban, K., 1955. Neue Grabungen in der Repolustho¨hle bei Peggau in der Steiermark. Mitteilungen des Museums fu¨r Bergbau, Geologie und Technik am Landesmuseum ‘Joanneum’ Graz 15, 77 – 87. Nagel, D., Rabeder, G., 1991. Exkursionen im Plioza¨n ¨ sterreichs. O ¨ sterreichische und Pleistoza¨n O Pala¨ontologische Gesellschaft, 44 pp., Wien. Nelson, D.E., 1997. Radiocarbon dating of bone and charcoal from Divje babe I cave. In: Turk, I. (Ed.), Mousterian ‘‘bone flute’’ and other finds from Divje babe I cave site in Slovenia. Opera Instituti Archaeologici Sloveniae 2, 51–65. Nielbock, R., 1987. Holoza¨ne und jungpleistoza¨ne Wirbeltierfaunen der Einhornho¨hle/Harz. Unpublished Ph.D., University of Technology, Clausthal, 197 pp. Nielbock, R., 2002. Die Einhornho¨hle – Forschungsstand und -perspektiven. In: Rosendahl, W., Morgan, M., Lo´pez Correa, M. (Eds.), Cave-BearResearches/Ho¨hlen-Ba¨ren-Forschungen. Abhandlungen zur Karst- und Ho¨hlenkunde 34, 5–11. Nielbock, R., 2003. Die Suche nach dem diluvialen Menschen – oder: Die Erforschungsgeschichte der Einhornho¨hle. Die Kunde N.F. 53, 9 pp. Otte, M. (Ed.) 1992. Recherches aux grottes de Sclayn. Volume 1. Le contexte. E´tudes et recherches arche´ologiques de l’Universite´ de Lie`ge 27, 178 pp. Otte, M., Leotard, J.-M., Schneider, A.-M., Gautier, A., 1983. Fouilles aux grottes de Sclayn (Namur). Helinium 23, 112–142. Otte, M., Patou-Mathis, M., Bonjean, D. (Eds.) 1998. Recherches aux grottes de Sclayn. Volume 2. L’arche´ologie. E´tudes et recherches arche´ologiques de l’Universite´ de Lie`ge 79, 426 pp. Pacher, M., 2000. Taphonomische Untersuchungen der Ho¨hlenba¨renfundstellen in der Schwabenreith-Ho¨hle bei Lunz am See (Niedero¨sterreich). Beitra¨ge zur Pala¨ontologie 25, 11–85. Pike, A.W.G., Hedges, R.E.M., Van Calsteren, P., 2002. U-series dating of bone using the diffusionadsorption model. Geochimica et Cosmochimica Acta 66, 4273–4286.

470

Wilfried Rosendahl, Doris Do¨ppes and Stephan Kempe

Quinif, Y., Bastin, B., 1994. Datation uranium/thorium et analyse pollinique d’une se´quence stalagmitique du stade isotopique 5 (Galerie des Vervie´tois, Grotte de Han-sur-Lesse; Belgique). Comptes rendus de l’Acade´mie des Sciences de Paris 318, 211–217. Rabeder, G., 1991. Die Ho¨hlenba¨ren der Conturines. Athesia Verlag, Bozen, 124 pp. Rabeder, G., Nagel, D., Pacher, M., 2000. Der Ho¨hlenba¨r. Species 4, Jan Thorbecke Verlag, Stuttgart, 111 pp. Rabeder, G., Temmel, H.J., 1997. Repolustho¨hle. In: Do¨ppes, D., Rabeder, G. (Eds.), Plioza¨ne und pleis¨ sterreichs. Mitteilungen der Komtoza¨ne Faunen O ¨ sterreichischen mission fu¨r Quarta¨rforschung der O Akademie der Wissenschaften 10, 181–185. Rink, W.J., 1997. Electron Spin Resonance (ESR) dating and ESR applications in Quaternary science and archaeometry. Radiation Measurements 27, 975–1025. Rink, W.J., Schwarcz, H.P., Smith, F.H., Radovcic, J., 1995. ESR ages for Krapina hominids. Nature 378, 24. Rosendahl, W., 1995. Zur taphonomischen Differenzierung quarta¨rer Großsa¨ugerfunde aus Ho¨hlen. Mitteilungsblatt der Gesellschaft fu¨r Urgeschichte 3, 5–8. Rosendahl, W., Darga, R., Ku¨hn, R., Pacher, M., 2000. Der Ho¨hlenba¨r in Bayern. Pfeil Verlag, Mu¨nchen, 48 pp. Rosendahl, W., Kaulich, B., Reisch, L., Ambros, D., 2005. Hunas Cave: 50 ky (OIS 5b – OIS 3) climate and environment history in Southern Germany. DEKLIM/PAGES Conference Mainz: ‘‘Climate Change at the very end of a warm stage’’, Abstract Volume, pp. 190–191. Rosendahl, W., Kempe, S., 2002. Erstnachweis von mittelpleistoza¨nen Insektenresten aus einem Ho¨hlensinter in Deutschland. Neues Jahrbuch fu¨r Geologie und Pala¨ontologie, Monatsheft 11, 693–704. Rosendahl, W., Wiegand, B., Kaulich, B., Reisch, L., 2006. Zur Altersstellung der mittelpala¨olithischen Ho¨hlenfundstelle Hunas/Ldkr. Nu¨rnberger Land – Ergebnisse und Interpretationen alter und neuer Sinterdatierungen. Germania 84, 1. Halbband, 1–18. Rosenmu¨ller, J.C., 1794. Quaedam de ossibus fossilibus animalis cujusdam, historiam ejus et cognitionem accuratiorem illustrantia, disertatio, quam d, 22. Octob. 1794. Ad disputandum proposuit Ioannes Christ. Rosenmu¨ller Heßberga-Francus, LL. AA.M. in Theatro anatomico Lipsiensi Prosector assumto socio Io. Chrs. Heinroth Lips. Medicinae Studiosus Cum tabula aenea, Leipzig, 34 pp. Shaw, T.R., 1992. History of Cave Science, the Exploration and Study of Limestone Caves to 1900. Sydney Speleological Society, New South Wales, Australia, 2nd ed., 338 pp.

Simonet, P., 1992. Les associations des grands mammife`res. E´tudes et recherches arche´ologiques de l’Universite´ de Lie`ge 27, 127–151. Temmel, H.J., 1996. Die mittelpleistoza¨nen Ba¨ren (Ursidae, Mammalia) aus der Schachtfu¨llung der Repolustho¨hle bei Peggau in der Steiermark ¨ sterreich). Unpublished Ph.D., University (O Vienna, 258 pp. Turk, I., Dirjec, J., Strmole, D., Kranjc, A., Car, J., 1989. Stratigraphy of Divje babe I. Results of the excavations 1980–1986. Razprave, Slovenska Akademija Znanosti in Umetnosti, Razred za Naravoslovne Vede, Classis 4, 30, 161 pp. Turk, I., Skaberne, D., Blackwell, B.A.B., Dirjec, J., 2001. Morphometric and chronostratigraphic sedimentary analyses and palaeoclimatic interpretations for the profile at Divje babe I, Slovenija. Archeoloski vestnik 52, 221–247 (in Slovenian, summary in English). Turk, I., Skaberne, D., Blackwell, B.A.B., Dirjec, J., 2002. Assessing humidity in Upper Pleistocene karst environment palaeoclimates and palaeomicroenvironments at the cave Divje babe I, Slovenija. Acta carsologica 31, 139–175 (in Slovenian, summary in English). ¨ ber bewohnte Ho¨hlen der VorVirchow, R., 1872. U zeit, namentlich der Einhornho¨hle im Harz. Zeitschrift fu¨r Ethnologie der Deutschen Gesellschaft fu¨r Vo¨lkerkunde und der Berliner Gesellschaft fu¨r Anthropologie, Ethnologie und Urgeschichte 4, 251–258. Wild, E.M., Paunovic, M., Rabeder, G., Steffan, I., Steiner, P., 2001. Age determination of fossil bones from the Vindija Neanderthal site in Croatia. Radiocarbon 43, 2B, 1021–1028. Wild, E.M., Steffan, I., Rabeder, G., 1988. Uraniumseries dating of fossil bones. Progress report Institut fu¨r Radiumforschung und Kernphysik Wien 53, 53–56. Winograd, I.J., Coplen, T.B., Landwehr, J.M., Riggs, A.C., Ludwig, K.R., Szabo, B.J., Kolesar, P.T., Revesz, K.M., 1992. Continuous 500 000-year climate record from vein calcite in Devils Hole, Nevada. Science 258, 255–260. Withalm, G., 1995. Vergleichend ro¨ntgenologischmethodische Untersuchungen an den Tibien von Ursus spelaeus und Ursus arctos. Unpublished Diploma Thesis, University Vienna, 30 pp. Zapfe, H., 1939. Lebensspuren der eiszeitlichen Ho¨hlenhya¨ne. Die urgeschichtliche Bedeutung der Lebensspuren knochenfressender Sa¨ugetiere. Palaebiologica 7, 111–146. Zapfe, H., 1954. Beitra¨ge zur Erkla¨rung der Entstehung von Knochenlagersta¨tten in Karstspalten und Ho¨hlen. Beitra¨ge zur Geologie, Staatliche Geologische Kommission der Deutschen Demokratischen Republik 12, 3–60.

31. The Last and the Penultimate Interglacial as Recorded by Speleothems From a Climatically Sensitive High-Elevation Cave Site in the Alps Christoph Spo¨tl1, Steffen Holzka¨mper2 and Augusto Mangini3 1

Institut fu¨r Geologie und Pala¨ontologie, Leopold-Franzens-Universita¨t Innsbruck, Innrain 52, 6020 Innsbruck, Austria 2 Department of Physical Geography and Quaternary Geology, Stockholm University, 10691 Stockholm, Sweden 3 Forschungsstelle Radiometrie, Heidelberger Akademie der Wissenschaften, Im Neuenheimer Feld 229, 69120 Heidelberg, Germany

ABSTRACT Within the greater Alpine region, absolutely dated climate records of the penultimate and the last interglacial are exceptionally rare. Speleothems offer an important and still underutilized source of information about the timing and duration of warm periods during the Middle and Upper Quaternary. The focal point of intense research is Spannagel Cave, a large high-altitude (ca. 2200 to 2500 m a.s.l.) cave system in the Zillertal Alps of Austria. The presently low (1.4 to 2.5 C) cave temperature provides a natural threshold for speleothem growth, i.e. the cave acts as a climatically sensitive archive. U-series dating of calcite speleothems, facilitated by exceptionally high U content in combination with highresolution stable isotope analyses allow identifying warm climate periods. Calcite growth at 236 to 229 kyr, 211 to 206 kyr and 199 to 192 kyr is in good accordance with U-series dated sea-level records and marine sediments whose chronology was tuned to orbital parameters. Oxygen isotope data show that the climate in the Alps was consistently cooler during the penultimate interglacial than during the last interglacial. Carbon isotope data also show a major difference between the two interglacials: while alpine soil and vegetation was apparently well developed during the last interglacial (and similar as today), high C

isotope values testify the lack of pedogenic C input during the penultimate warm periods. Accordingly, the area above the cave was either barren or – more likely – covered by a warm-based glacier. Previously regarded as evidence of ice-free conditions early during the penultimate deglaciation speleothem deposition at 136 kyr is now seen as an indication of a major change of the glacier’s thermal state most likely as a result of the collapse of ice-stream network at the end of the penultimate glacial maximum. Following a return to stadial conditions not conducive to speleothem formation and marked by a hiatus in speleothem growth, fully interglacial conditions did not commence until 130 kyr and prevailed until 119 kyr.

31.1 INTRODUCTION The Alps are probably the most thoroughly studied mountain range on Earth as far as climate history is concerned. This is particularly true for the instrumental period (i.e. since the middle of the eighteenth century). Recent studies, for instance, have shown that the Alps warmed by about 1.5 C since the end of the ‘Little Ice Age’, which is more than twice as much as the mean Northern Hemisphere warming and similar to the temperature trend observed in the Arctic

472

Christoph Spo¨tl, Steffen Holzka¨mper and Augusto Mangini

(Bo¨hm et al., 2001, personal communication 2005; Moritz et al., 2002; Arctic Climate Impact Assessment, 2004; Climate Research Unit, 2004). The reason for this strong regional increase in temperature during the past 150 years remains to be fully understood. In the light of record high temperatures in the past two decades (e.g. 1998, 2002, 2003) – absent in even the longest instrumental records – a long-term perspective of the natural climate variability in the Alps is sought. Recent studies on wood fragments released by receding alpine glaciers demonstrate that during the first part of the Holocene, the glaciated area of the Alps was repeatedly substantially smaller than today. These studies also suggest a pattern of repeated rises and falls of the equilibrium line altitude during the Holocene (Hormes et al., 2001; Nicolussi and Patzelt, 2000a, 2000b) that may possibly be related to the millennial scale climate variability of the North Atlantic (Bond et al., 2001; Broecker, 2001). Previous interglacials can provide additional insights into climate and environmental changes in the Alps. Until now, however, climatic records from pre-Holocene interglacials are extremely sparse and fragmentary within the Alps, owing to the destructive action of ice during the glacials. Therefore, our current understanding of climate change during the last interglacial relies on often nondated palaeovegetation data from low-lying sites in the foreland of the Alps. Hardly any reliable record is available from an interglacial prior to the last interglacial, and even the latter successions are mostly dated by means of biostratigraphic correlation only (e.g. Gru¨ger, 1979; Drescher-Schneider, 2000; Mu¨ller et al., 2003; Mu¨ller and Kukla, 2004). In recent years, a new type of palaeoclimate information has been retrieved from the shallow subsurface from speleothems in caves. These secondary carbonates have seldom been affected by erosion processes and may contain a wealth of palaeoclimate proxy information that can be precisely dated using state-of-the-art mass spectrometric U-series methods (see

Dorale, 2004; White, 2004 for recent reviews of this subject). Speleothem deposition requires the presence of groundwater supersaturated with respect to calcite. In regions such as the Alps, the limiting factor for speleothem formation is temperature. Consequently, speleothem growth intervals obtained from studies of presently cool alpine caves provide important temporal constraints on the timing of warm climate periods in the past. This article provides an overview of recent efforts of our group working in high-alpine cave sites to retrieve, analyse and evaluate this palaeoclimate information. Focusing on interglacial growth periods from Spannagel Cave in Tyrol, it considers and re-examines both previously published work and new, hitherto unpublished palaeoclimatic proxy data. 31.2 CAVE SETTING Spannagel Cave is located in the Central Alps of Austria and comprises a network of slightly more than 10 km of passages and short shafts located between 2524 and 2195 m (Fig. 31.1). It is the largest out of a series of more than 30 caves that developed within the Jurassic calcite marble that forms a tectonically deformed, 20-m thick slab dipping towards N and NNW beneath granitic gneiss. A crucial aspect is the proximity of the cave to the Hintertux Glacier (Fig. 31.1). Developed partly beneath a broad ridge separating two adjacent glacially shaped valleys about half of the cave system was in a subglacial position as recent as during the ‘Little Ice Age’, which was the most extensive glacier advance during the Holocene in the Alps (Maisch et al., 2000; Hormes et al., 2001). Sharp crested, poorly vegetated lateral moraines mark the 1850 advance of the Hintertux Glacier (Fig. 31.1). During periods when the ice extent was larger than this advance, most if not all of the cave was buried beneath

The Last and the Penultimate Interglacial

473

(a)

Karstifiable carbonate rocks in Austria

VIENNA

LINZ

SALZBURG

BREGENZ

INNSBRUCK GRAZ

KLAGENFURT

Till

(b)

500 m

Marble Granitic gneiss Paragneiss Glacier tongue during ‘Little Ice Age’

N

Spannagel Cave

Spannagel hut

Hintertux Glacier

(c) Lateral moraine ridge (‘Little Ice Age’ advance) 100 m

Spannagel hut Cave entrance

Fig. 31.1 Simplified geologic map of the area near Spannagel Cave, western Zillertal Alps, Tyrol. The plan view of the cave is superimposed on the map. Other, smaller caves present in this region are omitted for clarity. The location of the cross section (C) below is indicated by the two green arrows in B. The cave in bound to the thin slab of marble. Note different scales.

Christoph Spo¨tl, Steffen Holzka¨mper and Augusto Mangini

474

ice, reaching a maximum thickness of ca. 250 m during the last glacial maximum (van Husen, 1987). The present-day air temperature in the cave is constant during the year and slightly above the freezing point (Fig. 31.2) allowing water–rock interactions and (slow) formation of speleothems. Contrary to many other high-elevation caves in the Alps – which contain none or ancient speleothems only– several passages of Spannagel Cave provide evidence of modern dripstone formation, such as soda straws, stalactites, stalagmites and flowstones (Figs. 31.2 and 31.3). According to multiannual cave water monitoring dripwater feeding, speleothems is thermodynamically supersaturated with respect to calcite (either year around or seasonally; Spo¨tl, unpublished data), and 230Th/U dates demonstrate a Holocene age of these commonly active speleothems (Fig. 31.2). Other passages of Spannagel Cave contain speleothems – locally abundant – that are clearly inactive and partially broken and/or

corroded. 230Th/U dates document several episodes of speleothem growth during the past few hundred thousand years, including some dates exceeding the limit of the 230Th/ U method (Spo¨tl et al., 2004; Fig. 31.2). We identified a set of flowstones and stalagmites that grew during the penultimate interglacial (marine isotope stage 7, MIS 7) and the last interglacial (MIS 5.5). These samples were collected from different parts of the large cave system (Fig. 31.2). The dates allow validating previous findings based on isotope analyses of single speleothem samples. 31.3 METHODS The chronology of speleothem samples from Spannagel Cave was established by 230Th/U TIMS dating at Heidelberg University. For analytical details, see Frank et al. (2000). As a result of spike recalibration during the second half of 2004 and calibration against the ‘HU-1’ uraninite standard solution SPA11,52 SPA50,51

N 1.5

1.8

100 m

SPA129 1.9

Speleothems

1.6

Holocene and modern Last interglacial Penultimate interglacial

2.5

Entrance 1.4

1.7

SPA59

Fig. 31.2 Plan view of the cave network. Symbols indicate 230Th/U dated flowstone and stalagmite samples of Holocene, last interglacial and penultimate interglacial age. Samples examined in this study are labelled. Not shown are speleothems older than MIS 7 as well as samples formed during MIS 3. Yellow labels refer to cave air temperatures (constant year around) in degrees Celsius based on multiannual monitoring.

The Last and the Penultimate Interglacial

475

31.4 THE SPELEOTHEM RECORD

Fig. 31.3 This currently dry passage of Spannagel Cave shows evidence of Holocene speleothem formation (white carbonate deposits on ceiling and walls) and of former presence of a cave stream which left a sandy rubble blanket on the passage floor.

(assuming that uraninite is at secular equilibrium for the 230Th-234U-238U sequence), previously published 230Th/U dates were systematically too young. Revised 230Th/U dates of these results published by Spo¨tl et al. (2002) and Holzka¨mper et al. (2004) are provided by Holzka¨mper et al. (2005). Additional hitherto unpublished 230Th/U dates are presented in this article. A micromill technique was used to obtain continuous, high-resolution stable carbon and oxygen isotope transects along 230Th/Udated speleothem samples. The spatial resolution was 0.10–0.15 mm, about one order of magnitude higher than in our previous studies. The isotopic compositions were measured using a DeltaplusXL mass spectrometer equipped with an automated carbonate preparation system (Gasbench II) optimized for high sample throughput. Results are reported relative to the VPDB standard, and standardization was accomplished using NBS 19. The long-term precision of 13 C and 18 O values expressed as the 1-sigma standard deviation is 0:065 and 0:075‰, respectively (Spo¨tl and Vennemann, 2003). The internal structure of these speleothems was examined using standard thin-section petrography including epifluorescence microscopy.

We previously regarded the presence of speleothems as evidence of ice-free conditions above this high alpine cave, i.e. the equilibrium line altitude (ELA) was similar as today or higher. Based on age dating and stable isotope examination of several additional samples, it became evident that this assumption is too simplified. Spannagel Cave is exceptional, inasmuch as it apparently permitted local speleothem deposition also during times when most if not all of the cave system was covered by a (temperate) glacier, i.e. the ELA was at least as low as during the peak of the ‘Little Ice Age’. Two fundamentally different processes of water–rock interactions give rise to speleothem deposition in this cave system, and the stable C and O isotopes are key to identify them: (a) partitioning of soil-derived carbon dioxide into the seepage water during periods of moderately to well-developed vegetation cover, giving rise to speleothems with low 13 C values. (b) Dissolution of marble by protons released by the oxidation of disseminated sulfides (mostly in the tectonically overlying gneiss) resulting in high 13 C values in speleothems very similar to those of the hostrock. While the first mechanism is incompatible with the presence of ice above the cave, the latter process only requires the presence of liquid water in the aquifer and may continue to operate even when the cave is buried beneath a warm-based glacier. Studies of Holocene speleothems and modern dripwaters show that both processes operate hand in hand, but the pedogenic carbon dioxide source clearly being the dominant one (Spo¨tl, unpublished data). Pre-Holocene speleothem samples demonstrate that this was not the case for several older growth periods when 13 C values reflect the composition of the hostrock. These samples therefore suggest that the ground above the cave was either barren or possibly covered by ice. The only parameter that directly registers changes in atmospheric temperature above an alpine cave is the O isotopic composition.

476

Christoph Spo¨tl, Steffen Holzka¨mper and Augusto Mangini

Qualitatively speaking, speleothems showing high 18 O values formed from dripwaters that can be traced back to meteoric precipitation at comparably warmer temperatures than speleothems showing low 18 O values. Applying the two isotope systems in conjunction with 230Th/U dates allows us to identify the following scenarios: (a) Warm interglacial climate, high ELA, alpine vegetation above the cave: high 18 O and low 13 C values, e.g. Holocene (b) Cool interglacial or interstadial climate, ELA slightly lower than during the peak of the ‘Little Ice Age’, barren karst landscape and large parts of the cave in subglacial position: still relatively high 18 O, high 13 C values, e.g. Greenland Interstadial 14 (Spo¨tl et al., 2006). (c) Cold, stadial climate, ELA lower by several hundred meters, ice covers entire cave: low 18 O and high 13 C values, e.g. stadial preceding Greenland Interstadial 12 (Spo¨tl et al., 2006). (d) Cold, full glacial climate, ELA lower by ca. 1 km: no speleothem deposition probably because the glacier at this altitude was cold based, e.g. LGM. The predominantly interglacial records discussed in this article were obtained bothfrom flowstones, i.e. sheet-like accumulations of layered speleothems that form beneath a spatially variable dripwater source, and from stalagmites which are fed by a rather constant point source. A common feature of the former type of speleothems is the lateral variability of individual growth layers, reflecting changes in the amount of water flowing over different parts of its growing surface. As a result, the thickness of individual growth intervals as recorded by a single flowstone sample may not be representative of the overall growth rate of the flowstone. In addition, growth interruptions and related minor hiati may be the result of flow-route switching on the flowstone surface and hence

were not related to (external) hydrological changes. None of the sampled speleothems were currently active, and most of them were removed from their original growth position by former cave streams, most likely associated with flooding when the cave acted as a subglacial meltwater conduit (NB: there are no high-discharge streams in the cave today). While growth rates of speleothems from low-lying caves beneath densely vegetated areas in central Europe are typically approaching and exceeding 100 mm=yr (e.g. McDermott et al., 1999; Niggemann et al., 2003), resulting in Holocene stalagmites 50 to over 100 cm high, speleothems from Spannagel Cave are dwarfed and commonly contain multiple-growth discontinuities, similar to speleothems from high-latitude cave sites (e.g. Linge et al., 2001; Berstad et al., 2002). While local processes in the cave (e.g. spatial variability of dripwater supply) may account for minor discontinuities, prominent hiati are associated with major environmental changes outside the cave causing calcite precipitation to stop. Owing to such growth interruptions reliable depth versus age relationships are difficult to establish, a feature common in other flowstones as well (e.g. Baker et al., 1995; Wang et al., 2004). We therefore mostly restrict our discussion to individual growth intervals defined by 230Th/U dates. 31.4.1 Flowstone SPA 52 This piece of flowstone actually consists of three samples, a larger piece (SPA 52A) and two lateral equivalents of its upper part (SPA 52B, Fig. 31.4) and SPA 11, about 30 cm and 40 cm apart from the first one, respectively. Figure 31.5 shows a high-resolution stable isotope transect of SPA 52A together with the updated set of 28 230Th/U dates. According to the 230Th/U dates, about two-thirds of the samples were formed during MIS 7.3 to 7.1. The corresponding warm phase in the Alps – indicated by high 18 O values of the calcite – lasted from 211 and 189 kyr.

The Last and the Penultimate Interglacial

SPA52A

H4

477

SPA52B H4

H3

H3

H2 H2

H1

4 cm

Fig. 31.4 Slab of flowstone SPA 52A and lateral subsample SPA 52B (SPA 11 not pictured). The black lines indicate the locations of the high-resolution stable isotope traverses. Hiati are marked by H1 to H4.

The 18 O record reveals three intervals with consistently higher values separated by periods of lower values (Fig. 31.5). In one instance (100 mm above the base of the sample), this 18 O shift is associated with a macroscopic hiatus (H1 in Fig. 31.4). The periods of 18O-enriched calcite between MIS 7.3 and 7.1 can be compared to warm climate periods known from other climate archives. We chose the highly resolved alkenone record from sediment core ODP-977A retrieved from the western Mediterranean (Alboran Sea, Martrat et al., 2004). Its chronology is tied to the NorthGRIP ice core (North Greenland Ice Core Project Members, 2004). The chronology of the older section of the sediment is based on orbital tuning of the planktonic 18 O isotope curve. The lowermost 39 mm of the calcite in SPA 52A – dated at 211 to 206 kyr – records a period of rather high 18 O values depicting a slight trend towards lower values up section. The timing of this warm interval coincides with MIS 7.3, dated to be between 218 and 203 kyr in the Alboran Sea core (Fig. 31.5), which also shows an overall cooling trend. MIS 7.3 calcite precipitation abruptly ended in Spannagel Cave at 206 kyr associated

with a 2.5‰ drop in 18 O, consistent with the sharp 5 C cooling of the sea-surface temperatures (SSTs) in the Mediterranean. Despite this global cooling (MIS 7.2), calcite deposition continued in Spannagel Cave registering significantly lower oxygen isotope values. Following the sharp 2‰ rise at 65 mm above the base, warm conditions in Spannagel Cave (i.e. high 18 O values) were re-established at 195 kyr, which compares well to the MIS 7.1 signal of the Alboran Sea record ( 199 to 192 kyr). Even the (bifurcate) structure of this interglacial is similar in both records (Fig. 31.5). This period of high sea level (MIS 7.1) has been constrained to range from 202 to 190 kyr in the coastal Argentarola Cave (Italy; Bard et al., 2002) which is consistent with the chronology of SPA 52A. We note that MIS 7.1 did not reach the same high 18 O values as the preceding interglacial which may indicate an overall cooler temperature in the Alps during MIS 7.1 as compared with 7.3. By comparing sea-level data from Argentarola Cave and from a Bahamas flowstone (Li et al., 1989), Bard et al. (2002) concluded that sea level was higher during MIS 7.3 than during MIS 7.1.

Christoph Spo¨tl, Steffen Holzka¨mper and Augusto Mangini

478

Distance from base (mm) 0

20

40

60

80

100

120

140

160

180

200

90 H1

Age (kyr)

110

H2

H4

H3

130 150 170 190 210 1 6

0

δ13C (‰, VPDB)

4

–1

2

–2

δ13C

0

–3

–2

–4

–4

–5

–6

–6 –7

–8

–8

–10

–9 –10

–14

–11

–16

–12

–18

–13

–20

?

δ18O (‰, VPDB)

δ18O

–12

–14 –15

–22 –23

19

AI-26 (MIS 5.5)

AI-12′ (MIS 7.3)

AI-15′ (MIS 7.5)

AI-9′

AI-11′ (MIS 7.1)

15 13

AS-12′ (MIS 7.2)

11

AI-1′ AS-1′

17

AI-10′

SST (°C)

21

9 240

230

220

210

200

190

180

170

160

150

140

130

120

110

Age (ODP-977A; kyr)

Fig. 31.5 230Th/ U ages and the stable carbon and oxygen isotope record of flowstone SPA 52A as a function of distance from base. The chronology is based on a recalibration of the original 230Th/U dates of Spo¨tl et al. (2002) which resulted in a shift towards older ages (e.g., 3 kyr at 130 kyr). One data point at 148 kyr was discarded because it fell right at the major hiatus at 135 mm and most likely represents a mixture. The new high-resolution stable isotope track (150 m increments – Fig. 31.4) was micromilled a few centimetres off the location of the previous low-resolution profile. This new profile essentially confirms all major isotope features of the latter (Spo¨tl et al., 2002), but there are variations in the relative thickness of segments within this large flowstone sample, e.g. the thickness of the last interglacial calcite (see Fig. 31.4). The positions of the individual 230 Th/ U dates – shown with their 2-sigma errors – were plotted onto the new isotope profile by correlating salient features in both isotope tracks. Intervals of speleothem growth are highlighted by the horizontal grey bars. The growth phase during MIS 7 was discontinuous as indicated by the short, dashed, horizontal lines within the grey bar. Solid and vertical lines represent major growth hiati (H1–H4). The bottom curve shows alkenone-derived sea-surface temperature variations of the Alboran Sea (core ODP-977A; Martrat et al., 2004) for the period between MIS 7.5 and 5.5, and dashed lines suggest correlations between the 18 O profile of SPA 52 and the Mediterranean record. Events labelled AI and AS refer to warm and cold intervals, respectively (Martrat et al., 2004).

The Last and the Penultimate Interglacial

The remaining part of the MIS 7 section in sample SPA 52A is more difficult to interpret because of the presence of two discontinuities (H1 at 100 and H2 at 125 mm). There is clear evidence for a second cold period (reaching 18 O values lower than those during MIS 7.2) followed by a last warm interval with considerable internal isotope variability dated to 190 to 189 kyr. This period of high 18 O values may correspond to the short interstadial (AI-109 in Martrat et al., 2004) in core ODP977A immediately postdating MIS 7.1 ( 190 kyr; Fig. 31.5). Alternatively, this last warm period in Spannagel Cave could also correspond to the next interstadial (AI-99 in Martrat et al., 2004) centred at 186 kyr. In any case, no calcite was deposited subsequent to 189 kyr and during MIS 6, the penultimate glacial. Speleothem formation commenced again 53 kyr later, with a complexly structured layer of greyish calcite and high 18 O values compared to those of the warm MIS 7 periods. It follows a marked 3.5‰ decline and a recovery towards intermediate 18 O values (Fig. 31.5). This period is chronologically well bracket by 10 230Th/U dates ranging from 136 to 133 kyr. The Alboran Sea SST record exhibits a significant and prolonged interstadial between 138 and 131 kyr (AI-19) and a Younger Dryas-like cold reversal afterwards (AS-19 at 130 kyr; Fig. 31.5). This pattern of an early warming has also been reported from other archives, including the Devils Hole calcite (Winograd et al., 1997), a pollen record from Portugal (Sa´nchez Gon˜i et al., 2005), corals from Huon Peninsula (Esat et al., 1999) and Barbados (Gallup et al., 2002), aragonitic marine sediments off the Bahamas (Henderson and Slowey, 2000) and more recently from a speleothem record of Hulu Cave in China (Cheng et al., 2006). The marked drop in 18 O in flowstone SPA 52A is apparently slightly older than the Younger Dryas-type cold phase in core ODP-977A, which falls right within the major growth hiatus of SPA 51A at 170 mm above base.

479

The main phase of speleothem deposition during the last interglacial (MIS 5.5) lasted from 125 to 119 kyr, consistent with the highest SST values in the western Mediterranean Sea. The oxygen isotope data are the highest within the past 85 kyr in Spannagel Cave only rivalled by the earliest part of MIS 7.3 (Fig. 31.5). In contrast to the MIS 7 interglacials, the MIS 5.5 calcite is characterized by a surprisingly low variability in 18 O. The topmost calcite layer in this sample – overlying hiatus H4 – records a short interval of flowstone accretion during MIS 5.3. Figure 31.5 also shows a high-resolution 13 C record of flowstone SPA 52A. There are intervals of covariant 18 O and 13 C values alternating with segments of little or even negative correlation. We note a clear difference in the carbon isotopic composition of calcite deposited during the warm periods of MIS 7 and calcite formed during MIS 5.5. The latter depicts the lowest 13 C value (down to 10:3‰) at the time of high 18 O values (Fig. 31.5). This anticorrelation is consistent with the establishment of vegetation above this high-alpine cave site. In contrast, the covariant swings in both isotopes during MIS 7 and the generally high 13 C values suggest little if any input of carbon from the soil zone. Figure 31.6 compares the MIS 5.5 part of sample SPA 52A to that of the lateral sample SPA 52B. Both high-resolution records show remarkably stable 18 O values which decline at the end of the MIS 5.5 segment (marked by H4). In contrast, the 13 C values are highly variable and show an early trend of decline followed by a return to slightly higher values near the top (Fig. 31.6). We did not attempt matching the individual isotope peaks among the two adjacent flowstone samples. It is obvious however that the basic isotopic information provided by sample SPA 52A is representative for most parts of the ca. 4 m2 large flowstone. A lowresolution isotope profile is also available from a second lateral piece, SPA 11 (not shown graphically), and the first-order trends are essentially the same.

Christoph Spo¨tl, Steffen Holzka¨mper and Augusto Mangini

480

Distance from base (mm)

Age (kyr)

165

170

115 117 119 121 123 125 127 129 131 133 135

175

180

H3

185

190

195

200

H4 92.8 ± 0.5 kyr

2 0

δ13C

–2 –4 –6

–7

–8

–8

–10

–9

δ18O

–10 –11 –12

δ18O (‰, VPDB)

δ13C (‰, VPDB)

4

–2 –4 –6

δ13C

–8 –8

–10

–9 –12

δ18O

–10 –11 –12

δ18O (‰, VPDB)

δ13C (‰, VPDB)

–13

Fig. 31.6 18 O and 13 C profiles of the last interglacial portion of flowstone SPA 52. The upper two isotope tracks are a close-up of sample SPA 52A (see Fig. 31.4), while the two lower isotope traverses were analysed across the lateral equivalent SPA 52B (Fig. 31.4). The chronology is based on a recalibration of the original 230 Th/U dates by Spo¨tl et al. (2002). No 230Th/ U dates are available from SPA 52B. The stable isotope sampling interval was 150 m. Solid vertical lines represent growth discontinuities indicated by petrography and inferred from 230Th/ U dates. The top calcite layer formed significantly later than the MIS 5.5 calcite.

31.4.2 Flowstone SPA 59 This inactive flowstone piece ranges in thickness between 6 and 11 cm and was collected in the south-central part of Spannagel Cave (Fig. 31.2). The flowstone encompasses several intervals of calcite deposition separated by well developed hiati (Fig. 31.7). Only a summary of its salient features during MIS 7 and 5 is given here. Thirty-seven 230Th/U dates were determined from this flowstone which is presented in detail elsewhere (Holzka¨mper

et al., 2005). The flowstone record reveals a short growth phase between 236 and 229 kyr, with high 18 O values coinciding with the second half of MIS 7.5 as shown in the Alboran Sea SST record. The oxygen isotopic composition of this interglacial calcite is comparable to that of MIS 7.3 and 7.1 calcite of sample SPA 52A. 230Th/U ages derived from Bahamas slope sediments suggest that full interglacial conditions of MIS 7.5 lasted from 237 to 228 kyr (Robinson et al., 2002), which fits remarkably well to

The Last and the Penultimate Interglacial

MIS 5.5 MIS 7.1 MIS 7.5

2 cm

Fig. 31.7 Cross-section of flowstone SPA 59 showing the location of the stable isotope traverse (black line) and the stratigraphic position of calcite layers deposited during the penultimate and the last interglacial.

this growth phase of SPA 59. No calcite was precipitated during MIS 7.3 in sample SPA 59, which we attribute to site-specific hydrological changes (see above). The presence of a thick portion of MIS 7.3 calcite in sample SPA 52A demonstrates that this absence is indeed not palaeoclimatologically controlled. Overlying a hiatus, SPA 59 records a layer of calcite deposited during MIS 7.1 ( 192 to 199 kyr), consistent with the chronology obtained from flowstone SPA 52A (Fig. 31.5) and with a high sea level between 202 and 190 kyr (Bard et al., 2002). It is instructive to compare the isotope pattern of the MIS 7.1 calcite between the two flowstone samples SPA 59 and 52A (Holzka¨mper et al., 2005; Fig. 31.5): both 18 O records show a double peak, whereby the older peak is consistently larger than the younger one and the absolute 18 O values are comparable. The precision of the 230Th/U dates does not permit to verify or falsify the synchronicity of these isotope signals between

481

the two samples, but it is an intriguing possibility. No calcite was deposited after 192 kyr and during MIS 6, again consistent with the results from sample SPA 52A. Speleothem growth commenced again at 137 to 135 kyr, comparable to the first flowstone sample. 18 O values remained at an intermediate level nearly identical to that of sample SPA 52A ( 12‰). A thin layer of isotopically enriched calcite dated at 131–130 kyr records the onset of full interglacial conditions, and the peak 18 O value is precisely the same as those recorded in flowstone SPA 52A (and lateral equivalents, i.e. 8:5‰). Most of the MIS 5.5 calcite, however, appears to have been dissolved subsequently marked by a highly porous, residue-rich dark grey layer (Holzka¨mper et al., 2005; Fig. 31.7). It is unknown when this dissolution took place, but observations in Spannagel Cave reveal that different parts of the cave show evidence of active speleothem deposition and features of active corrosion (corroborated by hydrochemical data – Spo¨tl, unpublished data) during today’s interglacial. This situation arises from the wide range of hydrological flow systems, ranging from slow seepage flow (producing dripwaters supersaturated with respect to calcite) to fast fissure flow (permitting rapid access of rather aggressive surface waters (e.g. snow meltwater) to host rock and speleothem surfaces in the underground. 31.4.3 Stalagmites SPA 50 and SPA 51 Located only a few meters apart from flowstone SPA 52, these two stalagmites were found detached from their substrate (Figs. 31.8 and 31.10). They were probably broken off by high-discharge floods during the last glacial period but not transported more than a couple of meters as shown by their well-preserved original surface. Both stalagmites formed subsequent to the penultimate glacial and show comparable isotope records. Growth of SPA 50 started at 134 kyr (Holzka¨mper et al., 2004, 2005;

Christoph Spo¨tl, Steffen Holzka¨mper and Augusto Mangini

482

H3

4 cm

Fig. 31.8 Polished slab of stalagmite SPA 50. Note unconformity H3 near the base. The black line marks the micromilled stable isotope traverse to the right of a previously (manually drilled) low-resolution profile.

Fig. 31.9). In terms of both its formation period and the low 18 O values, this calcite is directly comparable to the early growth of calcite at the nearby flowstone SPA 52A (Fig. 31.5), as well as to the more distant flowstone SPA 59. Following a marked hiatus (H3; Fig. 31.8), calcite deposition recommenced at 129 kyr associated with 18 O values similar to that of flowstone SPA 52A. This MIS 5.5 calcite comprises the majority of the stalagmite and yielded a 230Th/U age of 121 kyr at the top. Because of its rather high growth rate, stalagmite SPA 50 could be examined in considerable detail (see Holzka¨mper et al., 2004). The results of stalagmite SPA 51 are reported here for the first time. This stalagmite is smaller (Fig. 31.10) and lacks the early growth phase of SPA 50. 230Th/U dates constrain its formation period from 127 to 119 kyr (Table 31.1), whereby the growth rate increased after 125-124 kyr similar to that of stalagmite SPA 50 (Fig. 31.9).

Age (kyr)

Distance from base (mm) 119 121 123 125 127 129 131 133 135 6

0

30

60

90

120

150

180

210

240

H3

δ13C (‰, VPDB)

4 2 0

δ13C

–2 –4 –6 –8

δ18O

–8 –9 –10 –11 –12 –13 –14

δ18O (‰, VPDB)

–10

Fig. 31.9 18 O and 13 C profiles of stalagmite SPA 50. The chronology is based on a recalibration of the original 230Th/ U dates (Holzka¨mper et al., 2004, 2005). The stable isotope sampling interval was 150 m. Dashed vertical lines represent changes in growth rate inferred from 230Th/U dates.

The Last and the Penultimate Interglacial

H4

483

interpretation. The latter stalagmite also records a much younger final calcite growth phase ( 56 kyr). Growth during MIS 3 is also replicated in flowstone SPA 59 (Holzka¨mper et al., 2005) as well as in two stalagmites from a neighbouring cave (Spo¨tl and Mangini, 2002; Spo¨tl et al., 2006).

31.5 PALAEOCLIMATIC DISCUSSION 31.5.1 MIS 7

2 cm

Fig. 31.10 Polished slab of stalagmite SPA 51 and location of stable isotope profile (black line). Hiatus H4 separates MIS 5.5 calcite from a thin layer of calcite formed during MIS 3. The apparent boundary ca. 5.5 cm above the base is not a hiatus.

The high-resolution stable oxygen isotope profile resembles that of its neighbour SPA 50 as well as that of flowstone SPA 52A with respect to both the low amplitude of variation and the absolute value (Fig. 31.11). The carbon isotope record of these three samples show similar overall features, i.e. 13 C values as low as 10‰ and a much higher amplitude as compared with that of the 18 O vales. The lack of an obvious match between the three carbon isotope curves suggests that this parameter reacts sensitively to minor differences between individual dripwater sites. Excursions towards high values are attributed to stochastic variability (decrease) in drip rate and progressive departure from equilibrium conditions. Parallel increases in 18 O – seen best in stalagmite SPA 51 – corroborate this

Given the fact that reliably dated palaeoclimate records from surface sediments of the penultimate interglacial in the Alps are unavailable, the flowstone records retrieved from Spannagel Cave have far-reaching significance. Flowstones are inherently more vulnerable to small changes in the seepage water supply than stalagmites. Nevertheless, both flowstones SPA 52A and 59 (Holzka¨mper et al., 2005) provide remarkable insights into the timing and progression of warm periods during MIS 7. As there is no Northern Hemisphere ice record going back to MIS 7, we chose the SST record from the Alboran Sea as a marine archive for comparison. Although this record is continuous and highly resolved, it lacks direct age control (Martrat et al., 2004). This alkenone record displays large, up to 10 C, temperature changes in the surface water of the Mediterranean Sea between warm and cold periods within MIS 7 (Martrat et al., 2004). The overall climate evolution seen in this record strongly resembles reconstructions from pollen data (Tzedakis et al., 2004), albeit at considerably higher detail. Periods during which calcite enriched in 18O was deposited in Spannagel Cave coincide with MIS 7.5 (SPA 59), 7.3 (SPA 52A) and 7.1 (SPA 52A and 59). This is also seen in the marine and terrestrial (pollen) records and provides first-time evidence of these three isotopically dated interglacial periods in the Alps. More important, the timing is consistent with other U-series

484

Lab #

2478 3253 3794 3254 3255 3829 3258 3793 3792 3830 3791

Distance from top (cm)

0.3 3.0 4.5 6.5 8.3 13.0 13.5 14.6 16.0 17.0 18.7

234 U

Concentration238U

Concentration232Th

Concentration230Th

Age

(‰)

(‰)

ðmg=gÞ

ðmg=gÞ

(ng/g)

(ng/g)

(pg/g)

(pg/g)

(kyr)

(kyr)

23.9 26.6 27.1 22.3 20.9 14.9 20.2 14.2 11.9 11.9 13.5

1.8 2.2 1.9 1.6 1.6 1.7 1.7 1.8 1.4 1.4 1.9

109.14 2.7110 3.5610 5.4210 4.0670 3.2493 3.7050 4.6840 5.5250 4.4753 6.3610

0.13 0.0027 0.0036 0.0054 0.0041 0.0032 0.0037 0.0047 0.0055 0.0045 0.0064

<0.05 0.9863 0.2972 1.4387 0.3120 0.6290 4.1840 0.2846 0.4321 0.7146 3.6440

0.0036 0.0018 0.0046 0.0013 0.0038 0.0230 0.0027 0.0013 0.0044 0.0310

701.531 30.3686 40.1100 61.3169 45.6227 36.57 41.5687 52.7200 62.3500 51.03 72.8400

3.087 0.2429 0.3000 0.2698 0.1779 0.32 0.2868 0.5800 0.1900 0.41 0.6300

56.35 119.1 120.2 122.3 120.8 123.1 121.0 123.3 124.5 126.9 127.5

0.37 1.8 1.7 1.1 1.0 2.1 1.6 2.5 0.8 2.0 2.2

Errors are quoted as 2-sigma standard deviations.

Christoph Spo¨tl, Steffen Holzka¨mper and Augusto Mangini

Table 31.1 Mass spectrometric 230Th/U dates of stalagmite SPA 51

The Last and the Penultimate Interglacial

485

Distance from base (mm) 0

30

60

90

120

150

180

115 H4

117

56.4 ± 0.4 kyr

Age (kyr)

119 121 123 125 127 129 131 133 135 6

δ13C (‰, VPDB)

4 2 0 –2 –4

δ13C

–6 –7

–8

–8 –9

δ O 18

–10 –11 –12 –13 –14

δ18O (‰, VPDB)

–10

–15

Fig. 31.11 18 O and 13 C profiles of stalagmite SPA 51 which formed next to stalagmite SPA 50 (see Fig. 31.5). The chronology is based on 230Th/U dates in Table 31.1. The stable isotope sampling interval was 150 m except for the uppermost 18 mm where it was reduced to 100 m. The solid line represents the H4 growth discontinuity indicated by petrography and inferred from U/Th dates (the topmost calcite layer formed during MIS 3).

dated chronologies from the Mediterranean (Bard et al., 2002) and the Bahamas (Robinson et al., 2002). Interestingly, the 18 O values of the speleothems record evidence of drops in 18 O of precipitation in the Alps by up to a few per mil at the end of MIS 7 warm periods (Fig. 31.5). This is surprising as atmospheric cooling should have also affected the cave system – currently slightly above the freezing point (Fig. 31.2) – and cause freezing of the karst fissure aquifer and hence shutdown of calcite deposition. The fact is that flowstone was not only deposited during the transition into the cold stages (e.g., MIS 7.3/7.4 transition; Fig. 31.5), but growth apparently continued during the latter as shown by low 18 O values (e.g. MIS 7.2; Fig. 31.5). Hence, liquid water was – at least temporarily –

available which allowed calcite deposition to continue also during cooling events. This raises the question of how speleothem deposition occurred at this rather extreme cold cave site. The carbon isotope data help to solve this question. Inspection of the isotope records presented in this study reveals two modes of stable isotope behaviour over time: (a) intervals of covariant isotope trends (most of MIS 7), and (b) periods of missing correlation between 13 C and 18 O (MIS 5.5., as well as during the Holocene; Spo¨tl, unpublished data). This suggests a fundamental difference in isotope fractionation and ultimately in the growth mechanism of speleothems at this site. Antipathetic isotope behaviour is the expected mode on millennial and longer time scales, when warm

486

Christoph Spo¨tl, Steffen Holzka¨mper and Augusto Mangini

climate gives rise to high 18 O values in conjunction with low 13 C values as a result of increased biological activity in the soil zone. It is interesting to note that presentday climate only supports alpine meadows above Spannagel Cave, and the average soil thickness does not exceed 10–20 cm. Some areas are even nearly vegetation free (recently deglaciated glacier forefield in the west; Fig. 31.1). Similar surface environmental conditions can be inferred for MIS 5.5 (see below), but apparently not for MIS 7 warm phases. None of the three MIS 7 interglacials reveals low 13 C values approaching 10‰; on the contrary, relatively high 18 O values indicating warm atmospheric conditions coincide with high 13 C values. Although some of this covariance may be due to kinetically controlled fractionation, it is more reasonable to attribute these high 13 C values to low or even negligible input of soil-derived organic carbon into the dripwater. In other words, the C isotope data unanimously suggest very little if any vegetation at this cave site during MIS 7 warm periods. This is consistent with MIS 7 interglacial 18 O values that are lower than MIS 5.5 18 O values, qualitatively suggesting overall lower atmospheric temperatures. Estimates of the global sea level during that time (Waelbroeck et al., 2002; Antonioli et al., 2004) indicate that the MIS 5.5 sea level was higher than that of the penultimate interglacial again implying more fully developed interglacial climate during MIS 5.5. Lower atmospheric temperatures during MIS 7 interglacials argue for a drop in the ELA below Holocene values. Given the fact that already half of Spannagel Cave was buried beneath ice during the ‘Little Ice Age’, this further drop in the ELA inevitably results in a more extensive glacier expansion likely covering most of the cave system. It is this blanket of apparently warm-based ice that is reflected by the lack of a pedogenic 13 C signal registered in the speleothems. More importantly, the presence of a temperate glacier

provides a plausible explanation why speleothem precipitation evidently continued during cold periods. Had the area above the cave been ice free, strong atmospheric cooling as indicated by the drop of calcite 18 O values at the end of MIS 7 interglacials by several per mil would inevitably have resulted in permanent freezing conditions in the cave, thus shutting off dripwater supply. 31.5.2 MIS 6 and 5 There are no 230Th/U-dated samples from Spannagel Cave of MIS 6 age, implying that conditions during full glacials are not conducive to calcite deposition. The temperature in the cave was most likely below the freezing point when the glacier reached a similar maximum thickness as during the LGM, i.e. ca 250 m (van Husen, 1987). Calcite deposition commenced early during Termination II, demonstrating a warming of the subsurface. Our updated 230 Th/U dates for this event suggest a time range between 136 and 133 kyr (SPA 52A, SPA 50 and SPA 59). We previously interpreted this as evidence of an early deglaciation in the Alps (Spo¨tl et al., 2002) in line with a series of other studies suggesting high global sea level at that time (e.g. Henderson and Slowey, 2000; Gallup et al., 2002; Antonioli et al., 2004). This early growth phase, however, did not continue into MIS 5.5, but was terminated by a hiatus in all samples from Spannagel Cave. Our records also reveal a marked difference in the oxygen and carbon isotopic compositions of this early calcite growth and the subsequent MIS 5.5 interval: the former is strongly depleted in 18O and shows 18 O values even lower than those that characterize MIS 7 warm periods (Fig. 31.5). These low 18 O values coupled with high 13 C values strongly indicate that temperatures – albeit increasing – were still significantly lower than during MIS 5.5. In analogy to the modern ELA position relative to the cave system,

The Last and the Penultimate Interglacial

we now regard it as very likely that ice still covered the area. We speculate that the MIS 6 dendritic glacier network had collapsed and given way to isolated, apparently warm-based valley glaciers at 136 kyr, which is reflected by slow calcite deposition on flowstones and stalagmites. This early growth phase was followed by a hiatus (until 130129 kyr) during which conditions for speleothem growth apparently deteriorated to the point where calcite precipitation came to a standstill. Recently, new high-resolution climatic proxy records from a variety of regions (Mediterranean – Martrat et al., 2004; China – Cheng et al., 2006, Iberian margin – Sa´nchez Gon˜i et al., 2005; California – Cannariato and Kennett, 2005) have shown a similar oscillation at Termination II, whereby an early warm (and/or wet) period was followed by a Younger-Dryas-type cold (and/or dry) reversal after which the full interglacial warming started. These studies are consistent with our cave records with respect to the absolute timing of the early warming phase (although some of them lack a direct age control). These records also suggest that temperature (and humidity) were below the MIS 5.5 level. It is therefore reasonable to regard this early warm phase as a pronounced interstadial with high sea level rather than the early part of the last interglacial. In any case, deglaciation was probably underway by 136 kyr in the Alps, but was interrupted shortly after by a cold spell during which ice spread again over the cave’s recharge area. Fully interglacial conditions including soil development and vegetation cover were only established at 129–130 kyr, consistent with speleothem data from Italy (Drysdale et al., 2005), Israel (Bar-Matthews et al., 2003) China (Yuan et al., 2004; Cheng et al., 2006). It is instructive to look at the last deglaciation of the alpine ice sheet for an analogue. Alpine piedmont glaciers retreated from their maximum extent probably as early as 1921 kyr (Ivy-Ochs et al., 2004). Major inneralpine valleys were definitely ice free,

487

and mountain glaciers were already confined to upper reaches of deep alpine valleys by 1718 kyr (Lister, 1988; Denton et al., 1999; van Husen, 2000), i.e. ca. 3 kyr prior to the abrupt warming of the Bølling interstadial and more ca. 6 kyr prior to the onset of the Preboreal. These observations suggest that the alpine glacier network – because of its comparably small size – reacted sensitively to early warming during Termination I. A decrease in seasonality concomitant with this warming may explain the obvious mismatch between glacier records in the Alps and the temperature record from Greenland ice cores (Denton et al., 2005). With respect to its oxygen isotope composition, the calcite formed during MIS 5.5 has higher 18 O values than that formed during the three MIS 7 growth phases (with exception of the calcite of SPA 52A formed during early MIS 7.3). 18 O values cannot be translated into absolute palaeotemperatures without additional constraints (e.g. McDermott, 2004). But our data qualitatively suggest higher atmospheric temperatures during MIS 5.5 than during MIS 7 warm periods provided the local temperature–18 O precipitation gradient was similar than that of today (e.g. Rozanski et al., 1992; Kohn and Welker, 2005). 18 O values of Eemian and Holocene speleothems from this cave site are comparable, suggesting similar overall temperatures in these two interglacials. The end of MIS 5.5 is recorded rather consistently in different samples from Spannagel Cave: 118:5 2:0 kyr (SPA 52), 119:5 1:9 kyr (SPA 11), 120:8 0:9 kyr (SPA 50) and 119:1 1:8 kyr (SPA 51). These four samples – from one chamber near the northeastern corner of the cave (Fig. 31.2) – indicate the kind of intersample precision that can be achieved even when using both stalagmites and flowstones. Unfortunately, the end of MIS 5.5 is not registered in flowstone SPA 59 (marked by a high-porosity dissolution layer) from the southern segment of the cave. A preliminary top age of 117:9 1:3 kyr, however, was recently obtained from a small

488

Christoph Spo¨tl, Steffen Holzka¨mper and Augusto Mangini

Eemian stalagmite (SPA 129 – Fig. 31.2), which is consistent with the ages of the four samples listed above. The cessation of speleothem growth in Spannagel Cave at 119 kyr – nearly synchronous with the timing of glacial inception (Calov et al., 2005; Sa´nchez Gon˜i et al., 2005) and with a cold and dry event recently identified in varve-counted maar lake sediments from Germany (Sirocko et al., 2005) – highlights the high sensitivity and reliability of the alpine speleothem archive. 31.6 CONCLUSIONS Speleothems from Spannagel Cave are currently the only absolutely dated climatic archive in the Alps covering the penultimate interglacial and the last interglacial including the penultimate deglaciation. The chronology of the MIS 7 warm periods – indicated by 18O-enriched calcite – is in accordance with other U-series-dated marine and continental records. Climate was consistently cooler than during MIS 5.5, and carbon isotope data suggest little if any vegetation above the cave, which was probably covered by a warm-based glacier during most of MIS 7. The new dates support our previous interpretation of speleothem growth at 136 kyr following the penultimate glaciation. Low 18 O coupled with high 13 C values, however, now lead us to believe that this highalpine area was probably still buried beneath ice at 136 kyr, although the presence of speleothems demonstrates warmbased conditions beneath the ice and hence a major change of the glacier consistent with a collapse of the MIS 6 icestream network. We therefore regard this growth phase early during Termination II as a result of a prolonged interstadial warming. A growth discontinuity, probably due to a Younger Dryas-type glacier advance, separates this growth phase from 18 O-rich and 13C-depleted calcite of MIS 5.5, suggesting high soil activity and

well-developed alpine vegetation. The last interglacial’s calcite deposition came to an end after 119 kyr, i.e. in response to the glacial inception. ACKNOWLEDGEMENTS Funding for this study was provided by the DEKLIM program of the German Federal Ministry for Education and Research and the Austrian Science Fund (START Y122GEO). We are grateful to the owner for unlimited access to Spannagel Cave and to many colleagues, students and cavers for logistic on-site support during numerous caving tours over the years. R. Eichsta¨dter and M. Wimmer prepared and analysed several of the samples discussed in this paper. We thank two referees for constructive comments which improved the final version of this article.

REFERENCES Antonioli, F., Bard, E., Potter, E-K., Silenzi, S., Improta, S., 2004. 215-ka history of sea level oscillations from marine and continental layers in Argentarola Cave speleothems (Italy). Global and Planetary Change 43, 57–78. Arctic Climate Impact Assessment, 2004. Impacts of a Warming Arctic – Arctic Climate Impact Assessment. Cambridge University Press, Cambridge, 144 pp. Baker, A., Smart, P.L., Edwards, R.L., 1995. Paleoclimate implications of mass spectrometric dating of a British flowstone. Geology 23, 309–312. Bar-Matthews, M., Ayalon, A., Gilmour, M., Hawkesworth, C.J., 2003. Sea–land isotopic relationships from planktonic foraminifera and speleothems in the Eastern Mediterranean region and their implications for paleorainfall during interglacial intervals. Geochimica et Cosmochimica Acta 67, 3181–3199. Bard, E., Antonioli, F., Silenzi, S., 2002. Sea level during the penultimate interglacial period based on a submerged stalagmite from Argentarola Cave (Italy). Earth and Planetary Science Letters 196, 135–146. Berstad, I.M., Lundberg, J., Lauritzen, S.E., Linge, H.C., 2002. Comparison of the climate during

The Last and the Penultimate Interglacial marine isotope stage 9 and 11 inferred from a speleothem isotope record from Northern Norway. Quaternary Research 58, 361–371. Bo¨hm, R., Auer, I., Brunetti, M., Maugeri, M., Nanni, T., Scho¨ner, W., 2001. Regional temperature variability in the European Alps: 1760–1998 from homogenized instrumental time series. International Journal of Climatology 21, 1779–1801. Bond, G., Kromer, B., Beer, J., Muscheler, R., Evans, M.N., Showers, W., Hoffmann, S., Lotti-Bond, R., Hajdas, O., Bonani, G., 2001. Persistent solar influence on North Atlantic climate during the Holocene. Science 294, 2130–2136. Broecker, W. S., 2001. Was the medieval warm period global? Science 291, 1497–1499. Calov, R., Ganopolski, A., Claussen, M., Petoukhov, V., Greve, R., 2005. Transient simulation of the last glacial inception. Part l: glacial inception as a bifurcation in the climate system. Climate Dynamics 24, 545–561. Cannariato, K.G., Kennett, J.P., 2005. Structure of the penultimate deglaciation along the California margin and implications for Milankovitch theory. Geology 33, 157–160. Cheng, H., Edwards, R.L., Wang, Y., Kong, X., Ming, Y., Kelly, M.J., Wang, X., Gallup, C.D., 2006. A penultimate glacial monsoon record from Hulu Cave and two-phase glacial terminations. Geology 34, 217–220. Climate Research Unit (2004): http:// www.cru.uea.ac.uk/ Denton, G. H., Heusser, C.J., Lowell, T.V., Moreno, P.I., Andersen, B. G., Heusser, L. E., Schlu¨chter, C., Marchant, D.R., 1999. Interhemispheric linkage of paleoclimate during the last glaciation. Geografiska Annaler 81A, 107–153. Denton, G.H., Alley, R.B., Comer, G.C., Broecker, W.S., 2005. The role of seasonality in abrupt climate change. Quaternary Science Reviews 24, 1159–1182. Dorale, J.A., Edwards, R.L., Alexander, E.C., Shen, C.C., Richards, D.A., Cheng, H., 2004. Uraniumseries dating of speleothems: current techniques, limits, and applications. In: Sasowsky, I.D. Mylroie, J. (Eds.), Studies of Cave Sediments. Physical and Chemical Records of Paleoclimate, Kluwer, New York, 177–197. Drescher-Schneider, R., 2000. Die Vegetations- und Klimaentwicklung im Riß/Wu¨rm- Interglazial und im Fru¨h- und Mittelwu¨rm in der Umgebung von Mondsee. Ergebnisse der pollenanalytischen Untersuchungen. In: Klimaentwicklung im Riss/ Wu¨rm Interglazial (Eem) und Fru¨hwu¨rm (Sauerstoffisotopenstufe 6–3) in den Ostalpen. – Mitteilungen der Kommission fu¨r Quarta¨rforschung der ¨ sterreichischen Akademie der Wissenschaften O 12, 39–92.

489

Drysdale, R.N., Zanchetta, G., Hellstrom, J.C., Fallick, A.E., Zhao, J.-x., 2005. Stalagmite evidence for the onset of the Last Interglacial in southern Europe at 129 1 ka. Geophys. Res. Letters 32, doi: 10.1029/2005GL024658. Esat, T.M., McCulloch, M.T., Chappell, J., Pillans, B., Omura, A., 1999. Rapid fluctuations in sealevel recorded at Huon Peninsula during the penultimate deglaciation. Science 283, 197–201. Frank, N., Braum, M., Hambach, U., Mangini, A., Wagner, G., 2000. Warm period growth of travertine during the Last Interglaciation in southern Germany. Quaternary Research 54, 38–48. Gallup, C.D., Cheng, H., Taylor, F.W., Edwards, R.L., 2002. Direct determination of the timing of sealevel change during Termination II. Science 295, 310–313. Gru¨ger, E., 1979. Spa¨triß, Riß/Wu¨rm und Fru¨hwu¨rm am Samerberg in Oberbayern – ein vegetationsgeschichtlicher Beitrag zur Gliederung des Jungpleistoza¨ns. Geologica Bavarica 80, 5–64. Henderson, G.M., Slowey, N.C., 2000. Evidence from U–Th dating against northern hemisphere forcing of the penultimate deglaciation. Nature 404, 61–66. Holzka¨mper, S., Mangini, A., Spo¨tl, C., Mudelsee, M., 2004. Timing and progression of the Last Interglacial derived from a high alpine stalagmite. Geophysical Research Letters 31, doi: 10.1029/ 2003GL019112. Holzka¨mper, S., Spo¨tl, C., Mangini, A., 2005. Highprecision constraints on timing of Alpine warm periods during the middle to late Pleistocene using speleothem growth periods. Earth and Planetary Science Letters 236, 751–764. Hormes, A., Mu¨ller, B.U., Schlu¨chter, C., 2001. The Alps with little ice: evidence for eight Holocene phases of reduced glacier extent in the central Swiss Alps. The Holocene 11, 255–265. Ivy-Ochs, S., Scha¨fer, J., Kubik, P.W., Synal, H.A., Schlu¨chter, C., 2004. Timing of deglaciation on the northern Alpine foreland (Switzerland). Eclogae Geologicae Helvetiae 97, 47–55. Kohn, M.J., Welker, J.M., 2005. On the temperature correlation of O in modern precipitation. Earth and Planetary Science Letters 231, 87–96. Li, W.X., Lundberg, J., Dickin, A.P., Ford, D.C., Schwarcz, H.P., McNutt, R., Williams, D., 1989. High-precision mass-spectrometric uraniumseries dating of cave deposits and implications for palaeoclimatic studies. Nature 339, 534–536. Linge, H., Lauritzen, S.E., Lundberg, J., 2001. Stable isotope stratigraphy of a late Last Interglacial speleothem from Rana, northern Norway. Quaternary Research 56, 155–164. Lister, G.S., 1988. A 15 000-year isotopic record from Lake Zu¨rich of deglaciation and climatic

490

Christoph Spo¨tl, Steffen Holzka¨mper and Augusto Mangini

change in Switzerland. Quaternary Research 29, 129–141. Maisch, M., Wipf, A., Denneler, B., Battaglia, J., Benz, C., 2000. Die Gletscher der Schweizer Alpen. 2 ed., Zu¨rich (Hochschulverlag), 373 pp. Martrat, B., Grimalt, J.O., Lopez-Martinez, C., Cacho, I., Sierro, F.J., Flores, J.A., Zahn, R., Canals, M., Curtis, J.H., Hodell, D.A., 2004. Abrupt temperature changes in the Western Mediterranean over the past 250 000 years. Science 306, 1762–1765. McDermott, F., Frisia, S., Huang, Y., Longinelli, A., Spiro, B., Heaton, T.H.E., Hawkesworth, C.J., Borsato, A., Keppens, E., Fairchild, I. J., van der, K. Borg, Verheyden, S., Selmo, E., 1999. Holocene climate variability in Europe: evidence from 18 O, textural and extension-rate variations in three speleothems. Quaternary Science Reviews 18, 1021–1038. McDermott, F., 2004. Palaeo-climate reconstruction from stable isotope variations in speleothems: a review. Quaternary Science Reviews 23, 901–918. Moritz, R.E., Bitz, C.M., Steig, E.J., 2002. Dynamics of recent climate change in the Arctic. Science 297, 1497–1502. Mu¨ller, U.C., Pross, J., Bibus, E., 2003. Vegetation response to rapid change in central Europe during the past 140,000 yr based on evidence from the Fu¨ramoos pollen record. Quaternary Research 59, 235–245. Mu¨ller, U.C., Kukla, G.J., 2004. North Atlantic Current and European environments during the declining stage of the last interglacial. Geology 32, 1009–1012. Nicolussi, K., Patzelt, G., 2002a. Untersuchungen zur holoza¨nen Gletscherentwicklung von Pasterze und Gepatschferner (Ostalpen). Zeitschrift fu¨r Gletscherkunde und Glazialgeologie 36, 1–87. Nicolussi, K., Patzelt, G., 2002b. Discovery of earlyHolocene wood and peat on the forefield of the Pasterze glacier, Eastern Alps, Austria. The Holocene 10, 191–199. Niggemann, S., Mangini, A., Richter, D.K., Wurth, G., 2003. A paleoclimate record of the last 17,600 years in stalagmites from the B7 cave, Sauerland, Germany. Quaternary Science Reviews 22, 555–567. North Greenland Ice Core Project Members, 2004. High-resolution record of Northern Hemisphere climate extending into the last interglacial period. Nature 431, 147–151. Robinson, L.F., Henderson, G.M., Slowey, N.C., 2002. U–Th dating of marine isotope stage 7 in Bahamas slope sediments. Earth and Planetary Science Letters 196, 175–187. Rozanski, K., Aragua´s-Aragua´s, L., Gonfiantini, R., 1992. Relation between long-term trends in

oxygen-18 isotope composition of precipitation and climate. Science 258, 981–985. Sa´nchez Gon˜i, M.F., Loutre, M.F., Crucifix, M., Peyron, O., Santos, L., Duprat, J., Malaize´, B., Turon, J.F., Peypouquet, J.P., 2005. Increasing vegetation and climate gradient in western Europe over the Last Glacial Inception (122–110 ka): data-model comparison. Earth and Planetary Science Letters 231, 111–130. Sirocko, F., Seelos, K., Schaber, K., Rein, B., Dreher, F., Diehl, M., Lehne, R., Ja¨ger, K., Krbetschek, M., Degering, D., 2005. A late Eemian aridity pulse in central Europe during the last glacial inception. Nature 436, 833–836. Spo¨tl, C., Mangini, A., 2002. Stalagmite from the Austrian Alps reveals Dansgaard-Oeschger events during isotope stage 3: Implications for the absolute chronology of Greenland ice cores. Earth and Planetary Science Letters 203, 507–518. Spo¨tl, C., Vennemann, T., 2003. Continuous-flow IRMS analysis of carbonate minerals. Rapid Communications in Mass Spectrometry 17, 1004–1006. Spo¨tl, C., Mangini, A., Frank, N., Eichsta¨dter, R., Burns, S.J., 2002. Start of the last interglacial period at 135 ka: Evidence from a high alpine speleothem. Geology 30, 815–818. Spo¨tl, C., Burns, S.J., Frank, N., Mangini, A., Pavuza, R., 2004. Speleothems from the high-Alpine Spannagel Cave, Zillertal Alps (Austria). In: Sasowsky, I.D., Mylroie, J. (Eds.), Studies of Cave Sediments. Physical and Chemical Records of Paleoclimate, Kluwer, New York, 243–256. Spo¨tl, C., Mangini, A., Richards, D.A., 2006. Chronology and paleoenvironment of marine isotope stage 3 from two high-elevation speleothems, Austrian Alps. Quaternary Science Reviews 25, 1127–1136. Tzedakis, P.C., Roucoux, K.H., de Abreu, L., Shackleton, N.J., 2004. The duration of forest stages in southern Europe and interglacial climate variability. Science 306, 2231–2235. van Husen, D., 1987. Die Ostalpen in den Eiszeiten, Vienna, Geologische Bundesanstalt. van Husen, D., 2000. Geological processes during the Quaternary. Mitteilungen der osterreichischen Geologischen Gesellschaft 92, 135–156. Waelbroeck, C., Labeyrie, L., Michel, E., Duplessy, J.C., McManus, J. F., Lambeck, K., Balbon, E., Labracherie, M., 2002. Sea level and deep water temperature changes derived from benthic foraminifera isotopic records. Quaternary Science Reviews 21, 295–305. Wang, X., Auler, A.S., Edwards, R.L., Cheng, H., Cristalli, P. S., Smart, P.L., Richards, D.A., Shen, C.C., 2004. Wet periods in northeastern Brazil over the past 210 kyr linked to distant climate anomalies. Science 432, 740–743.

The Last and the Penultimate Interglacial White, W. B., 2004. Paleoclimate records from speleothems in limestone caves. In: Sasowsky, I.D., Mylroie, J. (Eds.), Studies of Cave Sediments. Physical and Chemical Records of Paleoclimate, Kluwer, New York, 243–256. Winograd, I.J., Landwehr, J.M., Ludwig, K.R., Coplen, T.B., Riggs, A.C., 1997. Duration and

491

structure of the past four glaciations. Quaternary Research 48, 141–154. Yuan, D., Cheng, H., Edwards, R.L., Dykoski, C.A., Kelly, M.J., Zhang, M., Qing, J., Lin, Y., Wang, Y., Wu, J., Dorale, J.A., An, Z., Cai, Y., 2004. Timing, duration, and transitions of the Last Interglacial Asian Monsoon. Science 304, 575–578.

This page intentionally left blank

Section 5 Modelling Past Interglacial Climates (ed. Martin Claussen)

This page intentionally left blank

32. Climate System Models – A Brief Introduction Martin Claussen Meteorological Institute, University Hamburg, and Max Planck Institute for Meteorology, Bundesstr. 53, D-20146 Hamburg, Germany

Models are, generally speaking, descriptions of nature. There are several possibilities to describe nature. There is the artist’s point of view, there are narrative models and there are physical and mathematical models. The artist’s point of view is often helpful as it brings specific aspects of nature to the attention of scientists. The history of cloud paintings is an illustrative example (Wehry and Ossing, 1997). Narrative models, or story lines, are written concepts to qualitatively reconstruct causal relationships. Such narrative, conceptual models are often used in geology (e.g. Haug and Tiedemann, 1998). Examples of physical models are wind tunnels which simulate atmospheric boundary-layer flow or rotating tanks (e.g. Greenspan, 1980) in which rotating flows as proxy for atmospheric and oceanic motion are investigated. Physical models allow performing well-defined control experiments which, in nature, are impossible to do. So far, physical models only highlight specific processes in the climate system. In this part of the book, only mathematical models of the climate system are discussed. Mathematical models are a set of differential and diagnostic equations. These equations describe the dynamics of components of the climate system with varying degree of approximation. Most equations can be derived from the fundamental laws of conservation of energy, momentum and mass, and such a set of equations is referred to as deductive model. Climate models are generally quasi-deductive (Saltzman, 1985), because they include some empirical parameterisation of processes which cannot be deduced from fundamental laws such as

the so-called subgrid-scale motion discussed below. In parameterisation, only the effect of a process on the system is mathematically described. In contrast to deductive models, inductive models are not derived from first principles. Instead, the dynamics of climate processes, which are assumed to be important, are formulated in mathematical terms without explicit consideration of fundamental physical constraints – such as the models by Calder, Imbrie, Paillard, Paul and Berger discussed in Chapter 3. Inductive models are used to demonstrate the plausibility of climate processes and to explore the consequences of assumptions imposed on the system. There are several ways to classify climate system models according to their complexity and degree of approximation. For example, McGuffie and HenderssonSellers (1997) chose the picture of a pyramid in which the vertices represent mainly atmospheric components which are then coupled together to form the apex of a comprehensive model of atmospheric, oceanic and, in some cases, biospheric dynamics, a so-called general circulation model. Claussen et al. (2002) proposed an alternative view on this ‘classical’ hierarchy by introducing an indicator which characterises the number of interacting components of the climate system being explicitly described in a climate system model (see Fig. 32.1). In the following chapters, mainly general circulation models (in Chapters 33, 34, 35, 37) and the so-called Earth system models of intermediate complexity (EMICs – in chapters 36, 38, 39) are discussed.

496

Martin Claussen

11

8

2

CSM 4

7 9

1

I “Simple” Box-M.

7 6

3 10

A0-GCM A-GCM

D 6

5

5

4

4

3 2

5

6

3 2

1 1 1

2

3

4

5

6

G

Fig. 32.1 The three-dimensional spectrum of climate system models. The coordinates are: I, integration, i.e., the number of interacting components of the climate system (see Chapter 1.1) explicitly described in a model; G, geographical detail of description, here defined as the order of magnitude of the number of grid cells when counting atmosphere and ocean modules only; D, dynamic processes considered, here taken as the spatial dimension of atmospheric and of the oceanic modules (some models, for example, predict the atmospheric temperature on average over the depth of the atmosphere, or on average over a latitudinal band; hence their atmospheric module is considered two-dimensional in which the dynamics of atmospheric circulation is described in a more parameterised way than is the case in three-dimensional models). The numbers refer to several models of intermediate complexity, or so-called Earth system models of intermediate complexity (EMICs), while CSM, AO-GCM, A-GCM represent typical general circulation models (GCM) referred to as climate system model (CSM) or atmosphere–ocean GCMs (AOGCM) or atmospheric GCMs (A-GCM). For details see Claussen et al. (2002). This figure is taken from Claussen et al. (2002) with the permission of Springer-Verlag.

In relation to data, mathematical models can be used in two ways. Firstly, mathematical models are used in a prognostic, or predictive, mode. In this mode, models are driven by forcing (by changes in the Earth’s orbit, for example; see Chapter 1), and the results of the model are compared with data or proxy data to (a) interpret data and (b) to validate the model. Secondly, mathematical models can be run in an assimilation mode, i.e. they are ‘tuned to data’. Such tuning can

be done by nudging, for example, which means that some terms are added to the equations of energy, momentum and mass conservation which push the model towards observed data. Other methods require that models in an assimilation mode minimise the distance between model results and data points which are scattered in time and space. In the assimilation mode, models do not predict climate variations, but they are meant to interpolate sparse data in space and time in a physically consistent manner. Moreover, by means of assimilation, it is possible to explore the relative importance of processes in the observed climate variations. In this chapter, all models are used in the prognostic mode. When comparing data and model results, it is important to realise that climate is regarded as a stochastic process. Motion in the climate system consists of many scales ranging from the formation of cloud droplets, wind gusts, drift of pollen, to swirling ocean currents and the waxing and waning of ice sheets. Therefore, it is not possible to predict all this motion in a deterministic way. Instead, the complexity of the problem is reduced by averaging the variables which characterise the state of the system (such as wind speed, temperature, concentration of carbon in plants, etc.) over some region and some period in time. This averaging procedure introduces some degree of randomness, firstly, because the deviation from the average, the so-called subgrid-scale motion is not predictable deterministically and, hence, it is treated as a stochastic variable, and secondly, because subgrid-scale motion affects the averaged, grid-scale motion. Most differences between models can be attributed to the differences in parameterising subgrid-scale processes, i.e. in describing the effect of the subgrid-scale motion on the grid-scale motion. To explore the uncertainty of model results, it is useful to perform multimodel simulations. Therefore, the results of not just one model, but of a suite of models will be presented in Section 5.

Climate System Models

But even if the subgrid-scale motion were precisely known, the dynamics of the averaged climate variables are still hard to describe; in order to forecast any climate change, the state of the climate system in the past from which the forecast is started has to be reconstructed – which, of course, can be done only to some limited extent. Also the forcing of the climate system is known more or less accurately. In this situation, the averaged quantities are predicted by starting the simulations many times from a number of equally likely, but not precisely known, initial states and by using a number of equally likely, or plausible, external forcing. The former situation is referred to as an initial-value problem and the latter, a boundary-value problem. Actually, the separation between these problems depends on the memory of the system. Weather forecast is a typical initial-value problem, while future scenarios of climate change due to a hypothesised increase in atmospheric CO2 concentration are a sort of boundary-value problem. In some cases, a well-known, strong external forcing dominates climate dynamics such that all simulations yield similar results. In fact, often only one simulation is done, and it is implicitly assumed that other simulations would yield similar results – an assumption that will be critically assessed in Chapter 39. The randomness of climate motion has an important implication for simulations of interglacials and glacial inceptions which will be discussed in the following chapters. Some processes, like the calving of icebergs and subsequent flushing of fresh water into the ocean, depend on small-scale processes

497

which appear to be random. Therefore, it is impossible to precisely predict the occurrence of the first Heinrich event or first stadial at the last glacial inception. Instead, one can only explore the conditions which favour the occurrence of such events, and one can only expect to simulate the gross features seen in palaeoclimate reconstructions. ACKNOWLEDGEMENT The author thanks Susanne Weber, KNMI and Claudia Kubatzki, AWI-Bremerhaven, for constructive comments. REFERENCES Claussen, M., Mysak, L.A., Weaver, A.J., Crucifix, M., Fichefet, T., Loutre, M.-F., Weber, S.L., Alcamo, J., Alexeev, V.A., Berger, A., Calov, R., Ganopolski, A., Goosse, H., Lohman, G., Lunkeit, F, Mokhov, I. I., Petoukhov, V., Stone, P., Wang, Zh., 2002. Earth system models of intermediate complexity: Closing the gap in the spectrum of climate system models, Climate Dynamics 18, 579–586. Greenspan, H.P., 1980. The theory of rotating fluids. Cambridge University Press, 328 pp. Haug, G.H., Tiedemann, R., 1998. Effect of the formation of the isthmus of Panama on Atlantic Ocean thermohaline circulation. Nature 393, 673–676. McGuffie, K., Henderson-Sellers, A., 1997. A climate modelling primer. Wiley, Chichester, 253 pp. Wehry, W., Ossing, F.J. (eds.), 1997. Wolken – Malerei – Klima. Deutsche Meteorologische Gesellschaft, Berlin, 191 pp. Saltzman B., 1985. Paleoclimatic modeling. In: Hecht AD (ed), Paleoclimate analysis and modeling. Wiley , Chichester, 341–396.

This page intentionally left blank

33. Simulations of the Eemian Interglacial and the Subsequent Glacial Inception with a Coupled Ocean–Atmosphere General Circulation Model Frank Kaspar1,2 and Ulrich Cubasch2 1

Max Planck Institute for Meteorology, Model and Data Group, Bundesstr. 53, D-20146 Hamburg, Germany 2 Institute for Meteorology, Freie Universita¨t Berlin, Carl-Heinrich-Becker-Weg 6-10, D-12165 Berlin, Germany

ABSTRACT A coupled ocean-atmosphere general circulation model was used to perform multicentennial climate simulations of the Eemian interglacial and the subsequent glacial inception. The simulations are performed as equilibrium experiments with orbital parameters and greenhouse gas concentrations set to values of 125 000 and 115 000 years before present (BP). These dates represent periods with enhanced and weakened seasonal cycles of insolation on the northern hemisphere. Significant changes in seasonal temperatures are simulated in particular for the continental areas of the northern hemisphere. Comparisons with pollen-based reconstructions of European temperatures show that the model simulates realistic spatial temperature patterns for the warm phase of the Eemian. For 115 000 years BP, the reduction in summer insolation leads to a perennial snow coverage over parts of North America, which is continuously expanding. Together with a continuous increase of Arctic sea ice volume, this results in a long-term cooling trend. Consistent with geological records, the snow accumulation starts in north-eastern Canada. In this region, southward winds transport cold Arctic air onto the continent. 33.1 INTRODUCTION Among the variety of models applied in climate research, atmospheric general

circulation models (AGCMs) have the most complex representation of atmospheric processes. They are derived from models used in numerical weather prediction. Coupled ocean–atmosphere general circulation models (OAGCMs) consider in addition the strong interactions between oceans and atmosphere by coupling a threedimensional ocean GCM that typically also contains a sea–ice model. One frequent application of the models is the analysis of climate change scenarios under the influence of anthropogenic greenhouse gas emissions. The projected changes of climate towards the end of the 21st century are large compared to changes that occurred during the last centuries (Cubasch et al., 2001). Therefore, simulations of interglacial climates provide opportunities for evaluating how models respond to stronger changes in forcing (Mearns et al., 2001). If the models prove to be able to reproduce the observed climatic patterns of these periods, they can also be used to analyse the mechanisms that are responsible for these patterns. By comparing climate simulations of the current climate with observations, it has been shown that the current generation of OAGCMs is capable of simulating surface air temperature distributions particularly well. The simulation of clouds is a major source of uncertainty in these models. Consequently, precipitation is simulated less well than temperature, but still reasonable (McAvaney et al., 2001). However, the parameters simulated by OAGCMs may contain

500

Frank Kaspar and Ulrich Cubasch

systematic errors. To reduce the relevance of these errors in comparisons between model results and palaeoclimatic reconstructions, one typically looks at differences between two simulations, i.e. between the period of interest and the current or preindustrial period. Here, we present simulations of the Eemian interglacial, which was the last interglacial before the present one (approximately 130 000 yr BP–116 000 yr BP; Kukla et al., 2002a). Climate variations during the last 500 000 years are dominated by interglacial–glacial cycles, which are believed to be driven by changes in insolation as a result of variations in Earth’s orbit around the Sun (Berger and Loutre, 2002; see also chapter 1, this volume), but many examples can be found that contradict that assumption (Rind, 2002). The simulations used in this study represent periods within the Eemian interglacial with approximately maximum and minimum summer insolation on the northern hemisphere. They are connected with different states of the climate system that are visible in palaeodata. Approximately 127 000 yr ago, mixed forests were established in Europe. The minimum of global ice volume occurred at approximately 125 000 yr BP. We therefore consider this date as representative for the warm phase of the Eemian. After approximately 115 000 yr BP, open vegetation replaced forests in north-western Europe. At Vostok (Antarctica), air temperatures dropped sharply (Kukla et al., 2002a). We use this date to represent the start of the last glaciation. First attempts to apply GCMs to conditions at 115 kyr BP were performed in the 1980s with atmosphere-only GCMs (e.g. Royer et al., 1984). These experiments have been run as time-slice experiments with constant boundary conditions for short intervals (e.g. one year in the study of Royer et al., 1984). Sea-surface temperatures (SSTs) were fixed to modern conditions, or the models were coupled to mixed layer oceans. Although cooling over the high northern

latitudes occurred, most of these early experiments failed to simulate glacial inception as they were not able to generate perennial snow coverage (see Yoshimori et al., 2002; Vettoretti and Peltier, 2004 for detailed reviews). More recent studies applied coupled ocean–atmosphere GCMs (e.g. Khodri et al., 2001) or included vegetation feedbacks (e.g. Gallimore and Kutzbach, 1996). These studies showed that the inclusion of additional feedback mechanisms allows the occurrence of perennial snow coverage at 115 kyr BP on the North American continent, which is consistent with geological records. Currently, Earth system models are developed that are based on OAGCMs coupled to terrestrial vegetation models as well as to marine and terrestrial carbon cycle models (for an example see Gro¨ger et al., this volume). Coupling of an interactive vegetation model can cause substantial amplifications of feedbacks. Experiments with fixed boundary conditions have also been performed for the warm phase of the Eemian interglacial. One example is the study of Montoya et al. (1998, 2000) who used a coupled ocean– atmosphere model and compared the results with ocean temperatures reconstructed by CLIMAP Project Members (1984). The atmospheric component was a predecessor of the atmospheric model used in our experiments, but with a coarser spatial resolution. They found an agreement in the mean anomaly of the SST within the uncertainties, but the data from the CLIMAP Project indicate that SSTs were only little different from today. Felis et al. (2004) used the same model that is used in our study with a 100 times accelerated orbital forcing and compared the results for 124 000 years BP with temperature records from corals from the Red Sea. They concluded that during the last interglacial, the North Atlantic Oscillation had a tendency towards a high index state in comparison with today and that it is therefore responsible for warmer winters in Europe.

Simulations of the Eemian Interglacial

In this chapter, we analyse the simulated seasonal reactions of important climate variables for 125 000 years BP and 115 000 years BP and the mechanisms of the long-term cooling in the simulation for 115 000 years BP. In the following chapter, Widmann et al. (this volume) analyse teleconnections in the Eemian based on these simulations.

33.2 THE MODEL AND THE BOUNDARY CONDITIONS OF THE SIMULATIONS 33.2.1 Model description The ECHO-G model (Legutke and MaierReimer, 1999; Legutke and Voss, 1999) consists of the ECHAM4 atmosphere model (Roeckner et al., 1992) coupled to the HOPE-G ocean model. The atmospheric component is a spectral model with a horizontal resolution of T30 ð 3:75 Þ and 19 vertical hybrid sigma-pressure levels with the highest level at 10 hPa. The ocean model HOPE-G is the global version of the Hamburg ocean primitive equation model (Wolff et al., 1997) and includes a dynamic– thermodynamic sea–ice model with a viscous-plastic rheology (Hibler, 1979). A gaussian T42 Arakawa E-Grid (approximately 2.8 ) is used with a gradual meridional refinement reaching 0.5 in the tropical ocean between 10 S and 10 N. The vertical resolution is given by 20 horizontal levels, with eight levels within the top 2000 m from the ocean surface. The atmospheric and oceanic components are coupled with a flux correction in order to minimize a climate drift of the coupled system away from the climatology of the uncoupled models. The model time steps are 30 minutes for ECHAM4 and 12 hours for HOPE-G. The resolution of the model is appropriate for palaeoclimatic experiments as it allows simulating time slices of several centuries on modern computing facilities. The analysis of a 1000-year control simulation shows an overall good skill in simulating today’s climatology and

501

interannual variability (Min et al., 2005a). The El Nin˜o Southern Oscillation (ENSO) and the North Atlantic Oscillation (NAO) are also simulated reasonably well (Min et al., 2005b). A detailed evaluation of OAGCMs was published by McAvaney et al. (2001). 33.2.2 The experimental setup The simulations are performed as equilibrium experiments with constant boundary conditions. Therefore, the simulated temporal climate variations within each simulation are only caused by the nonlinear internal dynamics of the climate model. The orbital parameters and greenhouse gas concentrations have been set to values of 125 and 115 kyr BP (hereafter EEM and GI). Orbital parameters have been calculated according to Berger (1978). Greenhouse gas concentrations (CO2, CH4, N2O) have been adapted to values obtained from Vostok ice cores (Petit et al., 1999; Sowers, 2001). Concentrations of chlorofluorocarbons (CFCs) are set to zero. The model’s default values are retained for all the remaining boundary conditions, i.e. present-day conditions are used. A third simulation with preindustrial conditions is used for comparison (hereafter PI). The preindustrial phase has been selected because greenhouse gas concentrations are very similar to the values in the simulations for the Eemian, and therefore the observed differences in simulated climate can be attributed to insolation change. All parameters are shown in Table 33.1. All simulations have been initialized with the same oceanic state, i.e. potential temperature and salinity calculated from the Levitus et al. (1994) climatology. Three parameters are responsible for the seasonal distribution of insolation: the eccentricity of the Earth’s orbit, the angle of the axis of rotation (obliquity) and the position of the equinoxes relative to the perihelion. A general description of the long-term behaviour of these parameters and their impact on seasonal insolation is given in Chapter 2 (Berger et al., this volume). Figure 33.1a

502

Frank Kaspar and Ulrich Cubasch

Table 33.1 Orbital parameters and greenhouse gas concentrations of the simulation runs

Eccentricity Obliquity ( ) Perihelion ( ) CO2 (ppmv) CH4 (ppbv) N2O (ppbv) CFCs (ppbv) Length of the simulation (years)

125 kyr BP EEM

115 kyr BP GI

Preindustrial PI

0.0400 23.79 127.3 270 630 260 0 2000

0.0414 22.41 290.9 265 520 270 0 3000

0.0167 23.44 282.7 280 700 265 0 2000

Orbital parameters are calculated following Berger (1978). The angle of perihelion refers to vernal equinox. Greenhouse gas concentrations are based on Vostok ice cores [CO2 and CH4: Petit et al. (1999); N2O: Sowers (2001)]. (a) 90N –5 60

–10 –15 –20 –25 –30

60N

30N

50 EQ 40 35 30 25 20 15

30S –50 60S

0

10

90S –60

–30

0

Jan

Feb

Mar

180

210

240

270

Apr May Jun Jul Aug Sep True longitude from Vernal Equinox

30

60

90

Oct

Nov

Dec

–60 –50 –40 –35 –30 –25 –20 –15 –10 –5

120

0

5

150

10 15 20 25 30 35 40 50

60

(b) 90N 5 60N

–40 –35

10

5

–30 30N

–25 –20

EQ

–15 –10

30S 20 15 60S

–5

10

–5 5

90S –60

–30

0

Jan

Feb

Mar

180

210

240

270

Apr May Jun Jul Aug Sep True longitude from Vernal Equinox

30

60

90

Oct

Nov

Dec

–60 –50 –40 –35 –30 –25 –20 –15 –10 –5

120

0

5

150

10 15 20 25 30 35 40 50 60

Fig. 33.1 Insolation anomaly in W/m2 at (a) 125 kyr BP and (b) 115 kyr BP relative to current conditions as a function of latitude and the position on the Earth’s orbit, calculated according to Berger (1978). Position on the orbit is plotted as true longitude, which is the angle relative to vernal equinox (21 March).

Simulations of the Eemian Interglacial

illustrates the seasonal distribution of insolation on the different latitudes and seasons as anomalies from today for the Eemian at 125 000 yr BP. The combined effect of greater obliquity and eccentricity, together with the fact that perihelion occurred in northern hemisphere summer, caused an amplification of the northern hemispheric seasonal cycle of insolation. At 65 North, the insolation at mid-month June was 11.7% higher than today (531 W/m2 instead of 475 W/m2). The insolation of mid-month December was 1.9 W/m2 compared to 3.0 W/m2 today. Figure 33.1b shows the insolation anomalies for 115 000 yr BP. At 115 000 yr BP, the angle of perihelion was almost opposite compared to 125 000 yr BP and similar to today. The perihelion occurring in northern hemisphere winter combined with greater eccentricity and lower obliquity caused a weakening of the seasonal cycle of insolation on the northern hemisphere. The insolation of mid-month June at 65 North was 7.6% lower (439 W/m2) than today. At mid-month December, it was 6.8 W/m2.

503

33.3 RESULTS 33.3.1 Initial trends and the temporal evolution of the global mean temperature The simulations have been performed over at least 2000 years. During the first 100 years, all simulations are dominated by similar initial trends. Global mean near-surface temperature rises by approximately 0.4 K (Fig. 33.2). After this initial phase, the simulations start to differ. Especially in case of the simulation for 115 kyr BP, a significant long-term cooling trend occurs which in detail is examined in section 3.3. The two other simulations also show a slight cooling trend in the order of 0.1 K per 1000 years. The initial trends are caused by adaptation of the ocean circulation. Most features of the oceanic circulation reach quasi-stationary conditions after approximately 150 years in all three simulations, e.g. the maximum North Atlantic overturning rate or the Atlantic inflow of Antarctic Bottom Water at 30 S. Some features, especially the Pacific

286.3 286.2 286.1 286 285.9 T (K)

285.8 285.7 285.6 285.5 285.4 285.3 285.2 285.1

Solid: Preindustrial Dots: 125 kyr BP Dashes: 115 kyr BP

285 300

600

900

1200

1500

1800

2100

2400

2700

3000

t (yr)

Fig. 33.2 Time series of the annual mean near-surface temperature for the three simulations as 50 years running mean.

504

Frank Kaspar and Ulrich Cubasch

inflow of Antarctic Bottom Water at 30 S need up to 1000 years to reach their final equilibrium, but also in these cases the fastest changes occur during the first 150 years. 33.3.2 Spatial patterns In this section, we discuss seasonal spatial patterns of near-surface temperature (2 m temperature), precipitation, winds at 10 m and sea-level pressure (SLP). For the analysis, we use mean values of a 100-year interval starting 1700 years after the beginning of the simulations. We did not use a longer interval to avoid averaging over a period including a significant cooling trend as it is the case for the GI simulation. In case of the PI and the EEM simulation, the selected interval is within the stable phase of the simulation. Choosing different intervals

does not significantly change the results. In cases of the 115 kyr BP simulation, it should be considered that the long-term cooling would lead to a further decrease in temperature, but the regional structure of the simulated anomalies does not change significantly after this period. 33.3.2.1 Northern summer at 125 kyr BP The increased summer insolation at 125 kyr BP causes a strong increase in temperature especially on the continents of the northern hemisphere (Fig. 33.3a). Values þ4 K greater than in the preindustrial simulation occur on almost all continental areas north of 30 N. Values greater than þ10 K are reached over Siberia and parts of Central Asia. The mean temperature increase over the northern hemispheric continents during

(a) 2 m – Temperature (K)

(b) Sea-level pressure (hPa)

60N

60N

30N

30N

EQ

EQ

30S

30S

60S

60S

180

120W

60W –4

–2

0 –1

60E 1

2

120E

180

180

120W

4

–4

(c) 10 m – Winds (m/s)

60N

30N

30N

EQ

EQ

30S

30S

60S

60S 120W

60W

0

60E

–2

0 –1

60E 1

2

120E

180

120E

180

4

(d) Precipitation (mm/day)

60N

180

60W

120E

180

180

120W

60W

60E

0

10 –4

–2

–1

1

2

4

Fig. 33.3 Mean simulated difference (EEM – PI) in northern summer (JJA): (a) near-surface temperature (K), (b) sea-level pressure (hPa), (c) 10-m winds (m/s), (d) precipitation (mm/day).

Simulations of the Eemian Interglacial

505

Table 33.2 Difference in simulated temperature and precipitation for land, sea and total area of the globe and both hemispheres Global Land

Sea

Ann DJF JJA

0.20 1.14 2.08

0.10 0.26 0.37

Ann DJF JJA

0.88 0.00 2.06

Ann DJF JJA Ann DJF JJA

Northern hemisphere Total

Land

Sea

Total

Southern hemisphere Land

Sea

Total

Difference in near-surface temperature ðEEM PIÞ (K) 0.13 0.20 0.23 0.22 0.22 0.51 1.18 0.38 0.69 1.06 0.86 2.57 0.78 1.47 1.06

0.01 0.16 0.06

0.04 0.33 0.25

0.61 0.42 0.74

Difference in near-surface temperature ðGI PIÞ (K) 0.68 0.94 0.67 0.77 0.75 0.30 0.21 0.39 0.16 0.45 1.12 2.64 0.82 1.52 0.86

0.56 0.43 0.69

0.60 0.44 0.72

0.08 0.22 0.42

0.00 0.25 0.28

Difference in precipitation ðEEM PIÞ (mm/day) 0.02 0.13 0.04 0.08 0.02 0.12 0.10 0.20 0.08 0.45 0.08 0.53 0.28 0.03 0.20

0.04 0.29 0.27

0.04 0.15 0.18

0.04 0.11 0.20

0.01 0.13 0.10

Difference in precipitation ðGI PIÞ (mm/day) 0.02 0.07 0.03 0.04 0.01 0.06 0.07 0.09 0.03 0.17 0.01 0.28 0.12 0.04 0.03

0.00 0.16 0.08

0.00 0.10 0.06

Annual values as well as values for northern winter and summer are shown for the EEM (125 kyr BP) and the GI (115 kyr BP) simulation as difference to the preindustrial simulation. The values are calculated for an interval of 100 years starting 1700 after the beginning of each simulation.

summer (JJA) is 2.6 K (compare Table 33.2). It is strongest in the higher northern latitudes. The increase is about 3.7 K when averaged over 30 N–90 N and 3.9 K when averaged over 60 –90 N. The northern hemisphere’s oceans also react with a temperature increase, but due to the greater heat capacity of the oceans the simulated change is much smaller (0.8 K). A belt with reduced temperatures occurs at around 20 North over Africa and India, which is caused by significantly higher cloud coverage in the summer months. The simulated difference in the percentage of cloud coverage is approximately þ10% over Africa at 20 N and approximately þ25% in the region from the Arabian peninsula to India at 25 N. Especially in these regions significantly enhanced westerly winds cause an increase in transport of oceanic moisture into the continents and therefore also increased summer precipitation (Fig. 33.3c and 33.3d). The intertropical convergence zone (ITCZ) is shifted northwards by approximately 5 over Africa in summer. This contributes to an

extension of precipitation farther north. These results are consistent with the study of Rohling et al. (2002) who also found a northward shift of the ITCZ during Eemian summer based on the interpretation of proxy data. Similar results have been found in the model experiments by Montoya et al. (2000). The same mechanisms are visible in the simulations of Gro¨ger et al. (this volume), who showed that this effect is even enhanced when an interactive vegetation model is coupled to the system. The southern hemisphere also reacts with increased temperatures consistent with the increased insolation in northern summer (JJA). The insolation anomaly is not as strong as on the northern hemisphere (Fig. 33.1a), but causes also changes in temperature greater than 2 K in the region from approximately 15 S to 25 S on all continents, with central areas where values of more than þ4 K are reached. The mean temperature increase over the southern hemisphere land areas in summer is about 1.1 K. The simulated increase in temperature over

506

Frank Kaspar and Ulrich Cubasch

the southern hemisphere oceans is less than 0.1 K. The change in SLP is the direct response to the temperature anomalies (Fig. 33.3b). SLP is lower on the warmer continental areas (except Australia) and it is higher on most oceanic areas. The enhanced SLP gradient between land and oceans is responsible for increased winds in several coastal regions, in particular at the eastern coasts of North America and Asia (Fig. 33.3c). The changes in SLP are also responsible for stronger westerly winds at approximately 20 N over Africa and the Arabian region. The enhanced moisture transport leads to additional precipitation in this region of Africa and over India. The increase is about þ4 mm=day in large areas (Fig. 33.3d). Mean summer precipitation on the northern (southern) hemispheric continental areas is 0.53 mm/day (0.20 mm/day) higher

than in the preindustrial simulation and 0:28 mm=day (0:27 mm=day) lower on the oceanic areas (Table 33.2). 33.3.2.2 Northern winter at 125 kyr BP The decrease in winter insolation at 125 kyr BP is most significant south of 30 N. Again, this decrease is clearly visible in simulated winter temperatures (Fig. 33.4a). Negative temperature anomalies are simulated for almost all continental areas south of this latitude. A decrease in temperature of more than 4 K is simulated for the Sahara, Arabia, India, the Tibetan plateau and Australia. North of 30 N contrasting patterns are simulated: A strong decrease in temperature is simulated for the northern part of North America, whereas increased temperatures are simulated for the Arctic Ocean and the north-eastern part of the European

(a) 2 m – Temperature (K)

(b) Sea-level pressure (hPa)

60N

60N

30N

30N

EQ

EQ

30S

30S

60S

60S

180

120W

0

60W –4

–2

–1

60E 1

2

120E

180

180

120W

4

–4

(c) 10 m – Winds (m/s)

60N

30N

30N

EQ

EQ

30S

30S

60S

60S

120W

60W

0

60E

–2

0 –1

60E 1

2

120E

180

120E

180

4

(d) Precipitation (mm/day)

60N

180

60W

120E 7

180

180

120W

0

60W –4

–2

–1

60E 1

2

4

Fig. 33.4 Mean simulated difference (EEM – PI) in northern winter (DJF): (a) near-surface temperature (K), (b) sea-level pressure (hPa), (c) 10-m winds (m/s), (d) precipitation (mm/day).

Simulations of the Eemian Interglacial

continent. The change in winter mean temperature is around 1:1 K for the land areas of both hemispheres and significantly smaller for the ocean areas. SLP is higher over almost all land areas, with the exception of Greenland and Northern Europe (Fig. 33.4b). The increase is most distinct in Southern Europe, Asia, below 50 N, and Alaska. It is decreased over large parts of the oceans, except for the North Atlantic. Over the North Atlantic, a dipolelike change in SLP is simulated, with a significant decrease north of 60 N and an increase in the region between approximately 30 N and 50 N. The spatial structure of this dipole is comparable to a high index state of the NAO in the current climate (for a detailed discussion of the Arctic and North Atlantic Oscillation in these simulations see Widmann et al., this volume). It is responsible for a distinct increase in westerly winds over the Atlantic at 55 N to 60 N (Fig. 33.4c). These westerlies cause additional advection of oceanic heat from the Atlantic into the European continent and are therefore responsible for the simulated winter warming over Europe. The strong warming over the Arctic Ocean is related to a significantly reduced sea ice coverage which is a result of the increase summer insolation (Fig. 33.6). No significant change in precipitation is simulated for the continental areas of the northern hemisphere. On the southern hemisphere, a decrease in precipitation is simulated on all continents in the region between 0 and 30 S. The mean change in the southern hemisphere’s land area is 0:45 mm=day. In total, the global annual mean temperature is 0.13 K higher than in the preindustrial simulation. 33.3.2.3 Comparison with palaeoclimatic reconstructions and other modelling studies Ku¨hl (2003) published a perature reconstructions and macro-remains from It contains information

dataset with tembased on pollen 47 European sites. on summer and

507

winter temperatures for two phases of the Eemian (the so-called Corylus and Carpinus phase). The Corylus phase represents the early Eemian and is therefore comparable to our results for 125 000 yr BP. The general large-scale patterns of simulated summer and winter temperatures are in good agreement with this dataset: Higher temperatures than today are reconstructed for summer over whole Europe with the exception of a few locations in Southern Europe. This is in agreement with the simulated homogenous summer temperature increase. Furthermore, the dataset of Ku¨hl (2003) shows lower winter temperatures in Western Europe, but significantly higher winter temperatures than today in north-east Europe with increasing anomalies towards Scandinavia reaching values greater than þ6 K in Finland. This behaviour is in contrast to the reduced winter insolation, but does consistently occur in the dataset as well as in the simulation. A detailed comparison of this dataset and the model results is published by Kaspar et al. (2005). Velichko et al. (2005) also reconstructed Eemian temperatures in Central and Eastern Europe based on pollen data. They also found greater positive deviations in winter temperatures than in summer temperatures. The positive deviations in winter temperatures were more considerable for the eastern sites of their analysis (Il’inskoye, European part of Russia, 56 159N, 37 309E). The good agreement between simulated and reconstructed temperature patterns allows the conclusion that the model simulates a realistic Eemian climate and that orbitally induced changes in insolation are sufficient to explain the observed patterns at least for the European region. For winter, the temperature change is not a direct reaction to insolation change, but an indirect effect due to changes in atmospheric circulation and sea–ice. The small changes in temperature over the oceans (see Table 33.2) are consistent with the findings of the CLIMAP Project Members (1984), who concluded, based on oxygen isotope analysis and biotic census

508

Frank Kaspar and Ulrich Cubasch

counts in 52 cores across the world ocean, that the last interglacial ocean was not significantly different from the modern ocean. The winter warming in Eastern Europe is also visible in the simulation of Felis et al. (2004). However, in contrast to our results, they get negative winter temperature anomalies for northern Scandinavia. This difference is probably caused by different patterns in sea ice coverage in the Arctic Ocean due to the accelerated orbital forcing (100 times) in the experiment of Felis et al. (2004). The reconstructions of Ku¨hl (2003) show strong positive anomalies in this region. Montoya et al. (2000) also got a positive winter temperature anomaly for parts of Europe in their simulation, but only for the southern part of Europe and therefore not spatially consistent with the reconstructions. This spatial mismatch is probably caused by the coarser resolution of their atmospheric

model ð 5:6 Þ. At a larger scale, the spatial patterns of temperature, SLP, winds and precipitation are in agreement with the results of Montoya et al. (2000). 33.3.2.4 Spatial patterns for 115 kyr BP The insolation anomaly at 115 kyr BP has the almost opposite structure as at 125 kyr BP (compare Fig. 33.1a versus 33.1b). This leads to an almost opposite reaction in several model results. Summer temperatures are significantly decreased in most continental areas, in particular in the northern hemisphere north of 20 N, where a decrease of more than 4 K is simulated for large parts (Fig. 33.5a). Consistently, a strong increase in SLP is simulated for these areas (Fig. 33.5b). A strong cooling is also simulated for the Arctic Ocean. Reduced westerly winds occur north of the equator at 15 N (Fig. 33.5c, 33.6). As a

(a) 2 m – Temperature (K)

(b) Sea-level pressure (hPa)

60N

60N

30N

30N

EQ

EQ

30S

30S

60S

60S

180

120W

60W –4

–2

60E

0 –1

1

2

120E

180

180

120W

–4

4

(c) 10 m – Winds (m/s) 60N

30N

30N

EQ

EQ

30S

30S

60S

60S

120W

60W

0

60E

–2

0 –1

60E 1

2

120E

180

120E

180

4

(d) Precipitation (mm/day)

60N

180

60W

120E 7

180

180

120W

60W –4

–2

0 –1

60E 1

2

4

Fig. 33.5 Mean simulated difference ðGI PIÞ in northern summer (JJA): (a) near-surface temperature (K), (b) sea-level pressure (hPa), (c) 10-m winds (m/s), (d) precipitation (mm/day).

Simulations of the Eemian Interglacial

consequence, reduced precipitation in the order of 2 mm/day occurs over Africa and India. Reduced precipitation is also simulated for parts of North America. On the remaining parts of Asia, Europe and South America south of the equator, no significant change in precipitation is simulated (Fig. 33.5d). Consistent with the increase in winter insolation on the lower northern latitudes, higher winter temperatures are simulated for northern Africa (10 N–30 N) and southern Asia south of 45 N (not shown). Again, these regions also show the strongest decrease in SLP. A strong cooling of more than 4 K occurs in the Arctic Ocean and the neighbouring continental areas of Asia and north-eastern Europe. An increase in precipitation of approximately 2 mm/day is simulated for Africa between 0 N and 10 N. For the other land areas of the northern hemisphere, the precipitation does not differ significantly from the preindustrial results. (a) Preindustrial

509

Mean temperature changes are shown in Table 33.2. As in the case of the EEM simulation, the strongest change in temperature is simulated for the northern hemispheric land areas in summer ð2:6 KÞ. The change in summer temperature is lower than 1 K for the oceanic areas in both hemispheres and the land area in the southern hemisphere. For winter, the change is below 0:5 K for both hemispheres and both seasons. The strongest change in precipitation is also simulated for the land areas of the northern hemisphere in summer ð0:28 mm=dayÞ. 33.3.3 Mechanisms of snow accumulation on the North American continent A continuous expansion of perennial snowcovered areas on the North American continent occurs during the progression of the simulation for 115 kyr BP (Fig. 33.7). The (b) 125 kyr BP

80 60 40 20 10 0

(c) 115 kyr BP/100–199

80 60 40 20 10 0

(d) 115 kyr BP/1700–1799

80 60 40 20 10 0

80 60 40 20 10 0

Fig. 33.6 Minimum sea ice cover (%) in the Arctic. For all simulations, the monthly minimum of the average annual cycle for the years 1700–1799 is shown. In case of the GI simulation, the result for an additional interval at the beginning of the simulation (years 100–199) is shown.

510

Frank Kaspar and Ulrich Cubasch 80N 75N 70N 65N 60N 55N 50N 45N 40N 35N 30N 160W

150W 200

140W

130W 500

120W

110W 1000

100W 1500

90W

80W 2000

70W

60W 2990

Years after start of the simulation (10-year average)

Fig. 33.7 Expansion of permanently snow-covered areas over North America in the 115 kyr BP simulation. On marked areas, snow cover occurs in all months of the average annual cycle of the selected decade. This decade is centred round the years mentioned in the legend. The size of the grid boxes corresponds to the spatial resolution of the atmosphere model ( 3:75 ).

permanently snow-covered area on the northern hemisphere is increasing from 4:0 million km2 (mainly Greenland) to 7:6 million km2 after 3000 years. The increase proceeds approximately linear. The accumulated volume of snow on the North American continent is increasing by about 89 000 km3 during the first 1000 years, by about 276 000 km3 during the following 1000 years and by about 527 000 km3 during the last 1000 years of the simulation. This increase in the accumulation rate is caused by the fact that accumulation over areas which are already snow covered continues at an almost constant rate but the total area is also increasing. This long-term accumulation of snow denotes the initial process for a glaciation. This glacial nucleation starts in the Canadian Arctic archipelago and continues into north-eastern Canada. The ECHO-G model does not contain an inland ice model; therefore the build-up of the glaciers is not simulated in detail. The northern hemispheric sea ice volume is increasing from 22 000 to 57 000 km3 almost linearly during

the 3000 simulated years. This is mainly caused by increasing average ice thickness. The area permanently covered by sea ice is increasing from 6.6 to 8.6 million km2 during the first 1000 years of the simulation. Owing to the spatial limitation of the Arctic basin by the surrounding continents, the increase proceeds much slower afterwards, reaching a total area of 9.1 million km2 at the end of the simulation. All values refer to the annual minima based on monthly values. Snowfall over North America during the summer months is stronger in the GI simulation than in the other simulations. Figure 33.8a shows the annual cycle averaged over North America (130 W–60 W; 50 N–80 N). During the progression of the GI simulation, snowfall in summer is increasing significantly. This is caused by the reduced summer temperatures, but not connected with an increase in total precipitation. The annual precipitation in that area is 433 mm/yr (years 1700–1799) compared to 475 mm/yr in the PI and 495 mm/yr in the EEM simulation. As Fig. 33.8b shows, the decrease in

Simulations of the Eemian Interglacial

precipitation in the GI simulation is caused by a strong reduction in the second half of the year. The differences in net precipitation (precipitation minus evaporation), which is more relevant for snow accumulation, are (a)

45

125 kyr BP Preindustrial 115 kyr BP (1700–1799) 115 kyr BP (100–199)

Snowfall (mm/month)

40 35 30 25 20 15 10 5 0 1

2

3

4

5

6

7

8

9

10

11

12

8

9

10

11

12

Month (b) 60 125 kyr BP Preindustrial 115 kyr BP (1700–1799) 115 kyr BP (100–199)

Precipitation (mm/month)

55 50 45 40 35 30 25 20 1

2

3

4

5

6

7

Month

Precipitation – evaporation (mm/month)

(c)

40

125 kyr BP Preindustrial 115 kyr BP (1700–1799) 115 kyr BP (100–199)

35 30 25 20 15 10 5 0 1

2

3

4

5

6

7

Month

8

9

10

11

12

511

less distinct. Annual net precipitation in the GI simulation is 242 mm/yr (years: 1700–1799; rsp. 250 mm/yr at the beginning of the simulation during the years 100–199) compared to 246 mm/yr in the PI and 238 mm/yr in the EEM simulation. In the summer months, net precipitation is significantly higher in the GI simulation (Fig. 33.8c). Several authors (e.g. Khodri et al., 2001; Kukla et al., 2002b; Vettoretti and Peltier, 2003) discuss the relevance of an increase in northward atmospheric moisture transport from the equator to the pole as one important factor for the build-up of a perennial snow cover. As in our simulation precipitation over the relevant areas is not increased, this mechanism does not seem to be of major importance. However, an increase in northward moisture transport is simulated for the neighbouring ocean in northern summer, but not for the North American land area (Fig. 33.9). This is consistent with the findings of Khodri et al. (2001), who also found an increased northward moisture transport over the Atlantic in their simulation. In the simulations of Vettoretti and Peltier (2003) for 116 kyr BP, precipitation is also reduced compared to modern climate. However, due to enhanced reduction in evaporation compared to precipitation, they found an increase in net precipitation (P-E), which is also consistent with our results. The simulated start of the glaciation in north-eastern Canada is consistent with Fig. 33.8 (a) Annual cycle of snowfall in the three simulations (125 kyr BP, 115 kyr BP and preindustrial) over the northern part of North America (130 W–60 W; 50 N–80 N). In case of the simulation for 115 kyr BP, values for two intervals (years 100–199 and 1700–1799) are shown to illustrate the increase in snowfall in the course of the simulation. For the remaining two simulations, the values are averaged over the interval 1700–1799. (b) Annual cycle of precipitation in the three simulations for the same region as in Fig. 33.8 (a). (c) Annual cycle of net precipitation (precipitation minus evaporation) in the three simulations for the same region as in Fig. 33.8 (a).

Frank Kaspar and Ulrich Cubasch

geological records (Clark et al., 1993). During the 3000 years of the GI simulation, glacial nucleation does not occur in the region of the Rocky Mountains. This different behaviour between the eastern and the western part of the continent is in the simulation caused by different directions of heat transport in summer. Figure 33.10 shows the wind directions at 850 hPa in summer. The region where accumulation of snow occurs first is dominated by winds with a strong southward component. They transport cold air from the Arctic region into the continent. In the western part of the continent, northward winds prevail and cause advection of warm air from the Pacific. The warm air prevents the development of a perennial snow cover in this region. In the region of the Rocky Mountains, the northward summer winds are even enhanced in the GI simulation. Snow accumulation on land can be converted to sea-level reduction by just dividing the accumulated volume by the surface area of the Earth’s oceans. The volume accumulated over North America during the 3000 years of the simulation is equivalent to a reduction in sealevel of 2.44 m. Owing to the increase in the accumulation rate

Northward moisture transport (m/s g/kg)

512

0.8 0.6 0.4 0.2 0 –0.2 –0.4 –0.6 –0.8 –1 JJA DJF

–1.2 –1.4 180W

120W

60W

0

60E

120E

18

Fig. 33.9 Northward meridional transport of moisture in the northern hemisphere for summer and winter ðGI PIÞ. Moisture transport [m/s g/kg] is calculated as product of the meridional wind component v and specific humidity q (Peixoto and Oort, 1992). The values are integrated over all vertical levels (mass-weighted) and the northern hemisphere latitudes (0 –90 N). A distinct increase in northward moisture transport occurs in summer at the Pacific and Atlantic longitudes.

towards the end of the simulation, the rate of sea level change is also increasing and reaches a value of 18 cm per century. An additional reduction in sealevel is caused

90N 80N 70N 60N 50N 40N 30N 20N 10N EQ 180

160W

140W

120W

100W

80W

60W

40W

20W

0

10

Fig. 33.10 Winds at 850 hPa in summer (JJA) in the simulation for 115 kyr BP (averaged over the years 1750–1799).

Simulations of the Eemian Interglacial

by density changes of the oceanic water. Based on the equation of state, which depends on temperature, pressure and salinity (Gill, 1982), a reduction of 93 cm is calculated for the simulation period of 3000 years, i.e. 3 cm per century (net effect, reduced by the trend of the preindustrial simulation). The calculation is based on the full ocean, i.e. upper and deep ocean are considered. Additional snow accumulation occurs over Antarctica and Greenland, but for an accurate assessment of the net contribution of these regions an explicit glacier model should be coupled to the climate model. As this is currently not included, we did not consider these effects here. The total decrease of sea level during the last interglacial–glacial transition is of the order of more than 50 m during a period of 10 000 years (Lambeck and Chappell, 2001) and therefore in the order of 50 cm per century. During the progression of the simulation, an increase in the snow accumulation rate occurred, which suggests a further increase, in particular when glaciation also starts on the other continents. Furthermore, it has to be considered that the process is presumably accelerated when greenhouse gases are significantly reduced, as it was the case during the transition. Changes in vegetation could also contribute to an acceleration. In summary, orbitally induced changes in insolation are sufficient to cause a long-term expansion of snow-covered areas in those parts of the North American continent that are dominated by southward summer winds from the Arctic region. 33.4 CONCLUSIONS The simulations have shown that the coupled ocean–atmosphere model ECHO-G is capable of simulating features of the climate of the last interglacial and the subsequent glacial inception in agreement with geological records. For the warm phase of the Eemian at 125 000 years BP, orbitally induced changes in insolation are sufficient

513

to simulate seasonal temperature patterns in Europe which are in agreement with pollenbased temperature reconstructions. For summer, the direct reaction to increased insolation is responsible for higher temperatures, whereas winter changes in circulation cause increased temperatures in central and northeastern Europe. As the climate model does not contain a dynamic vegetation model, these results indicate that changes in the dynamics of the atmosphere–ocean system are sufficient to explain the spatial structure of the reconstructed European temperatures. However, simulated and reconstructed anomalies do not exactly agree, and additional mechanisms may be relevant. For a detailed discussion of vegetation–climate feedbacks, see Gro¨ger et al. (this volume). Differences in insolation as at 115 000 years BP are sufficient to initiate the process of glaciation on the North American continent. ACKNOWLEDGEMENTS This study has been performed within the German Climate Research Program DEKLIM of the German Ministry for Education and Research (BMBF). The simulations have been performed at the German Climate Computing Center (DKRZ, Hamburg). We thank Stephanie Legutke and Stephan J. Lorenz for their support concerning the ECHO-G model. The comments of two anonymous reviewers helped to improve the manuscript.

REFERENCES Berger, A.L., 1978. Long-term variations of daily insolation and Quaternary climate changes. Journal of the Atmospheric Sciences 35, 2362–2367. Berger, A., Loutre, M.F., 2002. An exceptionally long interglacial ahead? Science 297, 1287–1288. Clark, P.U., Clague, J.J., Curry, B.B., Dreimanis, A., Hicock, S.R., Miller, G.H., Berger, G.W., Eyles, N., Lamothe, M., Miller, B.B., Mott, R.J., Oldale, R. N., Stea, R.R., Szabo, J.P., Thorleifson, L.H., Vincent, J.S., 1993. Initiation and development of the Laurentide

514

Frank Kaspar and Ulrich Cubasch

and Cordilleran ice sheets following the last interglaciation. Quaternary Science Reviews 12, 79–114. CLIMAP Project Members, 1984. The last interglacial ocean. Quaternary Research 21(2), 123–224. 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: Climate Change 2001: The Scientific Basis, edited by J.T. Houghton et al., Cambridge University Press, New York, 525–582. Felis, T., Lohmann, G., Kuhnert, H., Lorenz, S.J., Scholz, D., Pa¨tzold, J., Al-Rousan, S.A., AlMoghrabi, S. M., 2004. Increased seasonality in Middle East temperatures during the last interglacial period. Nature 429, 164–168. Gallimore, R.G., Kutzbach, J.E., 1996. Role of orbitally induced changes in tundra area in the onset of glaciation. Nature 381, 503–505. Gill, A.E., 1982. Properties of Seawater. In: Atmosphere–Ocean Dynamic, Appendix 3, edited by A.E. Gill, Academic Press, New York. Hibler, W.D. III., 1979. A dynamic thermodynamic sea ice model. Journal of Physical Oceanography 9, 817–846. Kaspar, F., Ku¨hl, N., Cubasch, U., Litt, T., 2005. A model-data-comparison of European temperatures in the Eemian interglacial. Geophysical Research Letters 32, L11703. Khodri, M., Leclainche, Y., Ramstein, G., Braconnot, P., Marti, O., Cortijo, E., 2001. Simulating the amplication of orbital forcing by ocean feedbacks in the last glaciation. Nature 410, 570–574. Ku¨hl, N., 2003. Die Bestimmung botanischklimatologischer Transferfunktionen und die Rekonstruktion des bodennahen Klimazustandes in Europa wa¨hrend der Eem-Warmzeit. Dissertationes Botanicae 375. Kukla, G.J., Bender, M.L., de Beaulieu, J.L., Bond, G., Broecker, W.S., Cleveringa, P., Gavin, J.E., Herbert, T.D., Imbrie, J., Jouzel, J., Keigwin, L.D., Knudsen, K.-L., McManus, J.F., Merkt, J., Muhs, D.R., Mu¨ller, H., Poore, R.Z., Porter, S.C., Seret, G., Shackleton, N.J., Turner, C., Tzedakis, P.C., Winograd, I.J., 2002a. Last interglacial climates. Quaternary Research, 58, 2–13. Kukla, G.J., Clement, A.C., Cane, M.A., Gavin, J.E., Zebiak, S.E., 2002b. Last interglacial and early glacial ENSO. Quaternary Research 58, 27–31. Legutke, S., Maier-Reimer, E., 1999. Climatology of the HOPE-G Global Ocean–Sea Ice General Circulation Model, Technical Report 21, Deutsches Klimarechenzentrum, Hamburg, Germany. Legutke, S., Voss, R., 1999. The Hamburg Atmosphere–Ocean Coupled Circulation Model ECHO–G, Technical Report 18, Deutsches Klimarechenzentrum, Hamburg, Germany. Lambeck, K., Chappell, J., 2001. Sealevel change through the last glacial cycle. Science 292, 679–686.

Levitus, S., Burgett, R., Boyer T.P., 1994. World Ocean Atlas. Vol. 3, Salinity and Vol. 4, Temperature. NOAA Atlas NESDIS 3/4, U. S. Government Printing Office, Washington, DC. McAvaney, B.J., Covey, C., Joussaume, S., Kattsov, V., Kitoh, A., Ogana, W., Pitman, A.J., Weaver, A.J., Wood, R.A., Zhao, Z.-C., 2001. Model evaluation. In: Climate Change 2001: The Scientific Basis, edited by J.T. Houghton et al., Cambridge University Press, New York, pp. 471–523. Mearns, L.O., Hulme, M., Carter, T.R., Leemans, R., Lal, M., Whetton, P., 2001. Climate scenario development. In: Climate Change 2001: The Scientific Basis, edited by J. T. Houghton et al., Cambridge University Press, New York, 739–768. Min, S.-K., Legutke, S., Hense, A., Kwon, W.-T., 2005a. Internal variability in a 1000-year control simulation with the coupled climate model ECHO-G. Part I: near surface temperature, precipitation, and mean sealevel pressure. Tellus A 57(4), 605–621. Min, S.-K., Legutke, S., Hense, A., Kwon, W.-T., 2005b. Internal variability in a 1000-year control simulation with the coupled climate model ECHO-G. Part II: ENSO and NAO. Tellus A 57(4), 622–640. Montoya, M., Crowley, T.J., von Storch, H., 1998. Temperatures at the last interglacial simulated by a coupled ocean–atmosphere model. Paleoceanography 13, 170–177. Montoya, M., von Storch, H., Crowley, T.J., 2000. Climate simulation for 125 kyr BP with a coupled ocean–atmosphere general circulation model. Journal of Climate 13, 1057–1071. Peixoto, J.P., Oort, A.H., 1992. Physics of climate. American Institute of Physics, New York. Petit, J.R., Jouzel, J., Raynaud, D., Barkov, N.I., Barnola, J.M., Basile, I., Bender, M., Chappellaz, J., Davis, J., Delaygue, G., Delmotte, M., Kotlyakov, V.M., Legrand, M., Lipenkov, V.M., Lorius, C., Pe´pin, L., Ritz, C., Saltzman, E., Stievenard, M., 1999. Climate and atmospheric history of the past 420,000 years from the Vostok ice core – Antarctica. Nature 399, 429–436. Rind, D., 2002. The sun’s role in climate variations. Science 296, 673–677. Roeckner, E., Arpe, K., Bengtsson, L., Brinkop, S., Du¨menil, L., Esch, M., Kirk, E., Lunkeit, F., Ponater, M., Rockel, B., Sausen, R., Schlese, U., Schubert, S., Windelband, M., 1992. Simulation of the present day climate with the ECHAM model: Impact of model physics and resolution. Report No. 93, Max-PlanckInstitute for Meteorology, Hamburg, Germany. Rohling, E.J., Cane, T.R., Cooke, S., Sprovieri, M., Bouloubassi, I., Emeis, K.C., Schiebel, R., Kroon, D., Jorissen, F.J., Lorre, A., Kemp, A.E.S., 2002. African monsoon variability during the previous interglacial maximum. Earth Planetary Science Letters 202, 61–75.

Simulations of the Eemian Interglacial Royer, J.F., Deque, M., Pestiaux, P., 1984. A sensitivity experiment to astronomical forcing with a spectral GCM: Simulation of the annual cycle at 125 000 BP and 115 000 BP. In: Milankovitch and Climate, Part 2, edited by A.L. Berger et al., D. Reidel, Dordrecht, pp. 733–763. Sowers, T., 2001. N2O record spanning the penultimate deglaciation from the Vostok ice core. Journal of Geophysical Research – Atmospheres 106(D23), 31, 903–31,914. Velichko, A.A., Novenko, E.Y., Pisareva, V.V., Zelikson, E.M., Boettger, T.,Junge, F.W., 2005. Vegetation and climate changes during the Eemian interglacial in Central and Eastern Europe: comparative analysis of pollen data. Boreas 34(2), 207–219.

515

Vettoretti, G., Peltier, W.R., 2003. Post-Eemian glacial inception. Part II: elements of a cryospheric moisture pump. Journal of Climate 16, 912–927. Vettoretti, G., Peltier, W.R., 2004. Sensitivity of glacial inception to orbital and greenhouse gas climate forcing. Quaternary Science Reviews 23, 499–519. Wolff, J.O., Maier-Reimer, E., Legutke, S., 1997. The Hamburg Ocean Primitive Equation Model HOPE, Technical Report No. 13, Deutsches Klimarechenzentrum, Hamburg, Germany. Yoshimori, M., Reader, M.C., Weaver, A.J., McFarlane, N.A., 2002. On the causes of glacial inception at 116 kaBP. Climate Dynamics 18, 383–402.

This page intentionally left blank

34. Simulated Teleconnections During the Eemian, the Last Glacial Inception and the Preindustrial Period Martin Widmann, Nikolaus Groll and Julie M. Jones Institute for Coastal Research, GKSS Research Centre, Max-Planck-Straße 1, 21502 Geesthacht, Germany

ABSTRACT

34.1 INTRODUCTION

In this chapter, it is shown that the relationship between regional temperatures and large-scale circulation can change through time. The analysis is based on three 1000-year-long quasiequilibrium simulations with the coupled atmosphere–ocean general circulation model ECHO-G for the Eemian interglacial, the last glacial inception and the preindustrial period, which differ with respect to the Earth’s orbit, and as a consequence with respect to the incoming solar radiation. The temperature signals of the Arctic Oscillation (AO) are compared, and in some regions noticeable differences are found. For instance, during winter, there is a clear signal in Western Europe in the preindustrial but not in the Eemian simulation, whereas in Siberia, a strong signal is only found in the Eemian simulation. The results indicate in which regions an AO signal can be expected to be seen in temperaturesensitive proxy data with annual to multidecadal temporal resolution. An analysis of the mean and the variability of the atmospheric circulation suggests that the differences in the AO-temperature signal are mainly due to stronger mean westerly winds over Europe and western Siberia in the Eemian compared to the preindustrial period. Based on the same simulations, it is also shown that the spatial representativity of annual to decadal regional temperature variability differs between the three periods. For regions with a strong AO temperature signal, the temperature representativity pattern strongly resembles the AO temperature signal, and variations in the two patterns are closely related.

In Chapter 33, the spatial structure of the mean temperature during the Eemian was discussed. It was shown that the mean temperature field in a general circulation model (GCM) quasi-equilibrium simulation is consistent with temperature estimates derived from proxy data networks with millennial-scale temporal resolution and dating accuracy. The Eemian simulation and a simulation for the last glacial inception also allowed to asses to what extent the local mean temperature response to the orbital forcing is due to local changes in the incoming radiation, and to what extent it is due to an altered mean atmospheric and ocean circulation, or to changes in the cryosphere (Cubasch et al., 2006; Kaspar et al., 2005). Terrestrial proxy records for Eemian temperatures with a potentially multidecadal or higher resolution of the climate signal have recently become available (e.g. Seelos, 2004; Seelos and Sirocko, 2005; Seelos and Sirocko, this volume). On these timescales, local temperatures, and thus also temperature-sensitive proxy data, are strongly influenced by atmospheric circulation variability, which may be either internally generated or caused by fast varying forcing factors such as changes in volcanic aerosol concentrations or in insolation. As on longer timescales, local temperatures are also directly affected by the forcings through changes in the local radiative balance. In the first part of this chapter, we investigate which large-scale circulation signals can be expected to be seen in multidecadal Eemian temperature proxy data. In the second part, we discuss the spatial

518

Martin Widmann, Nikolaus Groll and Julie M. Jones

representativity of regional high-resolution temperature estimates. For the Late Holocene, large-scale circulation variability has been estimated from proxy data and long instrumental records by means of regression-based methods (e.g. Cook et al., 2002; D’Arrigo et al., 2003; Jones and Widmann, 2003), which are also known as upscaling models, because they estimate large-scale climate from local or regional data. In these studies, the statistical models were fitted during the overlap period of proxy or long instrumental records and high-density meteorological measurements. This approach is not directly applicable to earlier periods, such as the Eemian, mainly because the statistical relationship between local and large-scale climate, and as a consequence between proxy data and large-scale climate, may have been substantially different from the relationship under present conditions. The unknown relationship between multidecadal large-scale climate variability and local proxy data can also not be investigated through the analysis of proxy data at different sites, because the absolute dating accuracy is currently not high enough for a joint analysis of a proxy network on these timescales. Moreover, the number of temporal high-resolution records is still very limited, and records that contain signals from variables other than temperature, which would be needed to investigate circulationtemperature links, are scarce. We therefore use model simulations as a surrogate for the real climate for analysing large- to local-scale relationships. We investigated 1000 years of the two quasiequilibrium simulations with the atmosphere–ocean GCM ECHO-G presented in Chapter 33 and in Kaspar et al. (2005). One simulation is for 125ka BP and represents a period where deglaciation was complete and a relative stable climate was reached, the other is for 115 kyr BP and represents the last glacial inception. These simulations were compared with 1000 years of a quasiequilibrium simulation for preindustrial

conditions with the same model (Lorenz and Lohmann, 2004). The orbital parameters and greenhouse gas concentrations that were used for these simulations are given in Chapter 33, Table 33.1. GCM-based upscaling models have been used already for the Holocene for aiding the interpretation of proxy-based climate reconstructions (e.g. Zorita and Gonzalez-Rouco, 2002; Rutherford et al., 2003; von Storch et al., 2004), but not yet for the Eemian. This approach is known as pseudo proxy studies. In Section 2, the simulated mean temperature and circulation, and the circulation variability for 125 kyr BP, 115 kyr BP and for the preindustrial period are compared. With respect to variability, we focus on the Northern Hemisphere, as this is where most of the Eemian proxy data are available. We also focus on the Arctic Oscillation (AO), which is the dominant mode of circulation variability, and which is closely linked to the North Atlantic Oscillation (NAO). In Section 3, the AO temperature signals in the different simulations are investigated. Much of this material has been published by Groll et al. (2005) along with an analysis of the circulation anomalies that are linearly most closely linked to European temperatures. A discussion of methodological aspects related to the interpretation of regression maps, which are frequently used in our work, has been published by Widmann (2005). Section 4 presents an analysis of the link between regional and large-scale temperatures. 34.2 MEAN AND VARIABILITY General features of the mean climate of the Eemian and the glacial inception simulation are described in Chapter 33. Here we consider the differences in mean sea-level pressure (SLP) and temperature between the simulations. In northern winter, SLP is lower in the Eemian than in the preindustrial simulation over the ocean basins and higher over most parts of the land surface, with the exception of Northern and Eastern

Simulated Teleconnections During the Eemian

Europe (Fig. 34.1). Stronger westerlies over the latter region are evident in the Eemian simulation (not shown), consistent with the lower pressure here amplifying the pressure gradient. Temperatures in this region are higher in the Eemian than in the preindustrial simulation (Fig. 34.1d), consistent with the stronger westerlies advecting maritime air. Winter pressure differences between the glacial inception and the preindustrial simulation are smaller than differences for

519

the Eemian simulation. The higher pressure over Northwest Europe in the glacial inception simulation results in weakened westerly winds in this region. In order to investigate the main patterns of circulation variability in the simulations, empirical orthogonal functions (EOFs) were calculated. EOF analysis, which is also known as principal component analysis, is a standard technique, in which a timedependent spatial field is represented as a

DJF (a) PI SLP (hPa)

(b) PI 2m−TMP (K)

0

243

1

26325 8

2

01 161 10

293

12 1008

243

3

25

268 233

(d) EEM–PI 2m–TMP (K)

1 2

–5 –3

–6

3

–4 –2

1

0

–1

2

2

–1

273

258

248

283

278

273

(c) EEM–PI SLP (hPa) 0

283 278

263

298

288 293

288

1000 100 8

1004 1012

298

30 3

298

298

2

12

996

10

283

293 93 2

298

10

8

29

1008

1004 1000 996 992

273 8 28

293

1016

1016

25

8

26

268

4 02

101

1012

233

253 248

8

102

102 8 20 10

273

26

1012

248

288

0

32

12

102

10

16

10

10

1000

1016

16 10

3 2 23 53 8

243 1004

1 0

–3

–1

–1

–4

4

0

1

1

–2

–1

0

–2

0

0

–1

–1

0

0

0

–2

3

2

0

0

–1

0

1

–1

1

0

0

0

–1 0

1

0 –2

0 –1

0–1

1

(e) GI–PI SLP (hPa)

(f) GI−PI 2m−TMP (K) –2–1

–4

–2

76

0

3

2

45

–3

1

0 1

–1 –1

–2

1

0 1

0

–1

3

0

0

2

0

1

0

0

0

0

0

0

0

0

0

0

0

0

1

0

0

–1

–1 –1

–1

Fig. 34.1 Mean sea-level pressure (a) and temperature (b) for DJF in the preindustrial simulation (PI). Mean DJF SLP (c) and 2m temperature difference (d) between the Eemian and the preindustrial simulation (EEM-PI), and the mean DJF SLP (e) and 2m temperature difference (f ) between the glacial inception and the preindustrial simulation (GI-PI).

520

Martin Widmann, Nikolaus Groll and Julie M. Jones

superposition of spatial patterns with timedependent amplitudes. The patterns are calculated as the normalised eigenvectors of the covariance matrix of the data. The most important property is that the first eigenvector (which is associated with the largest eigenvalue) represents the pattern with the largest variance of the amplitudes among all possible patterns, the second pattern has the largest amplitude variance among all patterns orthogonal to the first one, and so on. Because of this property, an approximation of the data as an expansion using only a few of the leading EOFs captures the maximal amount of the true variance that can be described with a given number of patterns, and thus the first EOF can be interpreted as the dominant variability pattern. The AO is usually defined as the first EOF of extratropical SLP, which is shown for winter and summer in Fig. 34.2. The EOF values at the individual grid cells (the so-called EOF loadings) are scaled to be the pressure changes associated with a positive change of one standard deviation in the AO index. The patterns are similar in all three

simulations. Particularly in winter, a pronounced dipole over the North Atlantic corresponding to the NAO and a weaker dipole over the North Pacific are evident. There are also some small differences. The EOF loadings are slightly larger in the Eemian simulation during winter and spring (not shown) in the Atlantic sector, and smaller during winter in the Pacific. In the other seasons, there are only very small differences between the Eemian and preindustrial simulation. In the glacial inception simulation, the Pacific centre is stronger during winter than in the Eemian and preindustrial simulation. There is also a slight strengthening in the Atlantic centre in the glacial inception simulation compared to the other two simulations. In summer, the AO in the glacial inception simulation is slightly stronger in the Arctic centre than in the other two simulations. The fraction of variance explained by EOF1 is very similar in all runs and seasons. The second and third EOFs (not shown) do not differ much between the simulations. To determine the combined effect of the above changes in the SLP mean and in

PI

EEM

30.54%

29.44%

32.38% 0 1 2

3

2

3 4

1

−3

−4 0

0

–3

0 −1

3

–1 1

−4 −3 1−1

1 1 –

2

2

23.42%

0

−4 −1

0

−2

0

−2

0

−2

DJF

1

−1

2

24.46%

1

1

GI

24.91%

0

–1

1

0

−1

0

0

−2

−2

1

2

JJA

0

−1

0

0

1

0

0

Fig. 34.2 First simulated EOF for seasonal SLP, upper ( lower) panels: DJF (JJA). Left panels: preindustrial simulation (PI); middle panels: Eemian simulation (EEM); right panels: glacial inception simulation (GI).

Simulated Teleconnections During the Eemian

EOF1, we calculated SLP composite maps for situations with an AO index larger than plus or minus two standard deviations. The distribution of the AO index is approximately Gaussian in all runs, so that for every composite plot a set of about 25 seasons out of 1000 in the entire simulation is used. The results for winter are shown in Fig. 34.3. Consistent with the stronger westerly mean

521

flow during winter over Eurasia in the Eemian simulation compared to the other two simulations, the westerly flow in both AO phases extends further into Eurasia in the Eemian simulation. In some areas, this changes the character of the difference between positive and negative AO index situations. While the flow over Western Europe in the positive AO state is westerly

2*Std(AO+)

Mean

2*Std(AO−) 1016

1012

0

102

2

20 10

101

1008

10 28

20

08

1012

1 99 96000 92

1016

10

PI

20 10 24 10

1028

100

4

10

10

102

08

0

1016

1028

24 10

4 102

1020

6

6

101

1004

00 10

1012

24 10

0

101

1016

102

1016

1016

8

8

100

1032

10

6 101

0

100

1028

1028

24

24 10

6

101

0

1008 1012

10

102

24

4

2

08 10 2 101

102

1020

101

992

996

1032

1016

1016

24

1000 1020

1000

10 16

1020

10

1016

1028

10 20

1016

1020

10 36 103 2

10

1024

20 10

1008

100

1012

EEM

996

20

10 1 10

12

1016

2

0

32 10

10 20

10 24

1032

10 16

8 100

8

1016

102

00

10

10

12

1016

102

1024

4 1080 100

28

10

1024

1020

1012

996 24

32

1000 1004

102 8

102

4

1016

1012 16

992

10

1000

1020

100

996

4

102

0

1016

103

20

10 24

10

28

1016

1008 1016

10

1016

GI

2

1008

10

10

32

10

10

6 101

1020

1020

1024 28

1008

1000

1012

1004 24

10

1008 1020 1020

24 10

102

8

1020

Fig. 34.3 Mean DJF SLP (left panels) and composites of DJF SLP for situations with high (middle panels) and low (right panels) AO indices (two standard deviations). Upper panels: preindustrial simulation (PI); middle panels: Eemian simulation (EEM); lower panels: glacial inception simulation (GI).

522

Martin Widmann, Nikolaus Groll and Julie M. Jones

in all simulations, the winds are weak in the negative AO state in the preindustrial and the glacial inception simulation, but still somewhat westerly in the Eemian simulation. Over western Siberia and the Lake Baikal region positive AO situations are associated with southwesterly flow in the Eemian simulation, but not in the other two simulations. For negative AO index, there is no westerly flow over this area in all simulations. In summary, it appears that in winter the increased westerly mean flow in the Eemian simulation pushes the area in which the AO index variability leads to a clear switch between westerly and nonwesterly flow further into the continent. The positive spring AO state (not shown) is associated with weaker westerly flow over the Atlantic and over Europe in the Eemian than in the preindustrial simulation. For negative spring AO phases, there is no large difference in the southern and central parts of Europe, but a higher SLP gradient between Greenland and the Norwegian Sea in the Eemian simulation may be associated with a temperate climate in northern Europe during negative phases. In summer and autumn (not shown), there are no substantial differences in AO states between the simulations. Thus in winter, the results suggest that the difference in the flow characteristics between positive and negative AO phases over Europe (Western Siberia) is smaller (larger) in the early Eemian than in the preindustrial period and during the glacial inception. Although the similarity of SLP EOF1 implies that the SLP difference between positive and negative AO index situations, which is independent of the mean SLP pattern, is similar in all simulations, a change from very weak to moderately strong westerly flow can be expected to make a larger difference in the regional climate than a change from moderate to strong westerly flow, even if the difference in the wind speed is the same. In other words, the potentially nonlinear relationship between flow and regional climate may lead to an effect of a changed mean flow on the relationship between the AO index and the

local climate, even if the AO pattern does not change. In particular, this effect could lead to a weaker (stronger) relationship between European (western Siberian) temperatures and the AO in the early Eemian. 34.3 THE AO-TEMPERATURE SIGNAL In order to directly analyse the relationship between the AO index and temperatures, we now calculate correlation and regression coefficients between the standardised AO index and the 2m temperature in the three simulations, using a 31-year Hamming filter. A Hamming filter is similar to a Gaussian filter, but the response in the frequency domain is slightly more optimal (smaller side lobes). The findings indicate in which locations an AO signal may be expected in temperature-sensitive proxy data on multidecadal timescales during the Eemian. The regression (coloured) and correlation (contour lines) coefficients for northern winter are shown in Fig. 33.4. The overall spatial structure of the winter AO-temperature signal in the Eemian simulation is similar to that of the preindustrial simulation; however the coefficients are lower in most regions. In Central Europe, where most of the terrestrial proxy data with high temporal resolution are available, the regression coefficients drop by about 40%, and the correlation drops proportionally from 0.4 in the preindustrial to 0.2 in the Eemian simulation. Owing to the smoothing with the 31-year filter and the resulting autocorrelation, these changes are not significant at the 5% level. However, correlation differences with no filtering or with a 5- and 11-year filter are similar and significant, which suggest that the changes in the 31-year filtered data are also not a random sampling effect. In Western Europe, there is a weak AOtemperature signal in the preindustrial simulation, which completely vanishes in the Eemian simulation. In the glacial inception simulation, the AO-temperature signal is much more similar to the preindustrial

Simulated Teleconnections During the Eemian

523

Fig. 34.4 Winter AO-temperature signal for the preindustrial simulation (PI, left panel), the Eemian simulation (EEM, middle panel) and the glacial inception simulation (GI, right panel). Colours show the temperature change associated with a positive change in the AO index of one standard deviation. Contour lines show the correlations between the AO index and grid-cell temperatures. All values are derived from 31-year filtered (Hamming) data.

than to the Eemian simulation, with positive coefficients over Central and Western Europe, and the highest values over eastern Fennoscandia. Thus, in large parts of Europe the AO explains in the Eemian simulation less of the multidecadal winter temperature variability than in the other two simulations. Only in parts of Siberia is the Eemian AO signal stronger by up to 0.3 K, and the correlation higher by about 0.2. These findings corroborate the hypothesis on the link between the AO index and regional temperatures formulated at the end of the previous section.

34.4 SPATIAL REPRESENTATIVITY OF REGIONAL TEMPERATURES In this section, we investigate how local temperatures on annual-to-decadal timescales at one location are related to those at other locations. The motivation for this analysis is to aid the understanding of the spatial representativity of high temporal resolution temperature estimates from proxy data. We will derive the links between temperatures at

different locations from the same GCM simulations used in the previous sections. As these are quasi-equilibrium simulations and thus no changing forcings which could influence temperatures are used, we can only investigate the part of the link that is due to the circulation variability. If the simulated circulation variability is realistic, the simulated temperature teleconnections can be expected to partly account for teleconnections in empirical data because of the internally generated variability in the real climate system, and also because of forced circulation variability with the same structure as the internally generated variability. We will show that the simulated teleconnections in the temperature field are closely linked to the AO-temperature signal. As the AO is a largescale circulation feature, it is reasonably well represented in GCMs. Modelling studies suggest that solar forcing can influence the state of the AO (Shindell et al., 2001). The correlation between the temperatures at two sites in the real world is also influenced by the spatial patterns of the forced local temperature changes due to altered local radiative balances, which are usually similar over

524

Martin Widmann, Nikolaus Groll and Julie M. Jones

large areas and thus can be expected to increase correlations at many locations, as well as by forced circulation changes that have a different structure than the internally generated variability. Our work thus is just a first step towards a comprehensive understanding of the temperature teleconnections in the real climate system. We focus on the spatial representativity of temperatures from three regions from which Eemian terrestrial proxy data are available: Western Europe, Eastern Europe and the Lake Baikal area. We analyse the representativity for a given month of the year by regressing the simulated temperature at all model grid cells against the temperature at the area of interest, which is defined as the spatial mean over six grid cells that cover this area. As a compromise between annual and decadal timescales Fig. 34.5 shows results for January based on 5-year filtered (Hamming) values. Using annual data or filters of different length up to 11 years changes the results only slightly. The January patterns contain in all three simulations and for all three regions positively correlated regions around the centre, which are somewhat stretched in the direction of the prevailing westerly flow, and which are only slightly different in the three simulations. In addition to this, Western Europe shows in January marked teleconnections that are similar in the preindustrial and the glacial inception simulation and different in the Eemian simulation. A comparison with the AO-temperature signal presented in Fig. 34.4 reveals for the preindustrial and the glacial inception simulation a similar structure, with negative values over Greenland and the Bering Strait and positive values over Canada. This is physically plausible as Western Europe is in the preindustrial and the glacial inception simulation located in an area with a strong AOtemperature signal, and it thus can be expected that the teleconnections of Western European temperatures are dominated by the AO. In the Eemian simulation, there is

no strong AO-temperature signal in Western Europe, and the respective representativity pattern shows indeed little similarity with the AO-temperature signal. The January regression coefficients for Eastern Europe are small; however the correlations indicate in all simulations temperature teleconnections that resemble the AO-temperature signal. Moderate teleconnections can be found for Lake Baikal temperatures. This region is located in a maximum of the AO-temperature signal in the Eemian simulation, and consistent with this, the representativity pattern (Fig. 34.5, middle right panel) resembles the AOtemperature signal in the Eemian simulation, as both show for instance negative correlation and regression coefficients west of Greenland, and over Alaska and northwestern Africa, and positive correlations over North America. These teleconnections for Lake Baikal are much smaller or not present in the other two simulations. The other seasons are dominated by the regional positive correlations and show only weak teleconnections. 34.5 SUMMARY By analysing quasi-equilibrium GCM simulations for 125 kyr BP, 115 kyr BP and the preindustrial period, it was shown that the different orbital forcing in these periods substantially affects the mean atmospheric flow. For the Northern Hemisphere, which is the focus of our study, this effect is particularly strong during winter. The spatial structure of the dominant patterns of circulation variability relative to the individual means of the simulated periods differs only slightly between the simulations. Composite maps for positive and negative AO index situations showed that, despite the similar AO patterns, the more westerly mean flow in the Eemian simulation compared to the other two simulations shifts the area where the state of the AO index distinguishes between westerly and nonwesterly winds from Europe to western

Simulated Teleconnections During the Eemian

525

Fig. 34.5 One-point regression and correlation maps for 5-year filtered January 2 m temperature from several regions where Eemian proxy data are located. Left panels: Western Europe (WE); middle panels: Eastern Europe (EE); right panels, Lake Baikal area (LB). Top panels: preindustrial simulation (PI); middle panels: Eemain simulation (EEM); lower panels: glacial inception simulation (GI). Colours show the temperature change that is linearly related to a positive temperature change of 1 K in the reference area. Correlations are shown as contour lines, with an isoline spacing of 0.2.

Siberia. These differences are in turn linked to changes in the winter AO-temperature signal, primarily to a shift of the Eurasian maximum of the AO-temperature signal from Europe to western Siberia.

Teleconnections between simulated northern hemispheric temperatures at different locations are particularly strong in winter. When the centre region is located in a maximum area of the AO-temperature

526

Martin Widmann, Nikolaus Groll and Julie M. Jones

signal, the temperature teleconnections closely resemble the AO-temperature signal. The differences between the winter AO-temperature signal in the different simulations are associated with consistent differences in the temperature teleconnections. These results have implications for the interpretation of proxy data that contain a climate signal with high temporal resolution. The changes in the simulated AOtemperature signal showed that the relationship between regional temperature and large-scale circulation can be different at different times, and thus relationships between temperature or temperature-sensitive proxy data and large-scale circulation derived during the instrumental period may not be valid in other periods. The fact that the component of spatial representativity of regional temperatures that is caused by internally generated circulation variability can also depend on the period means that high-frequency correlations between temperature-sensitive proxy data at different locations may change through time. ACKNOWLEDGMENTS We thank the two anonymous reviewers for their valuable comments which helped us to improve the manuscript. We thank Frank Kaspar and Stefan J. Lorenz for providing their model results, Stephanie Legutke for her support with the ECHO-G model. This work was funded within the EEM project (climate change at the very end of a warm stage) by the Federal Ministry of Education and Research under the DEKLIM program (Deutsches Klimaforschungsprogramm).

REFERENCES Cook, E.R., R.D.D’ Arrigo, and M.E. Mann, 2002. A very-well verified multiproxy reconstruction of the winter North Atlantic Oscillation Index since A.D. 1400. J. Clim., 15, 1754–1764.

Cubasch, U., E. Zorita, F. Kaspar, F. Gonzales-Rouco, H. von Storch and K. Pro¨mmel, 2006. Simulation of the role of solar forcing on climate. Adv. space Res., 37, 1629–1634, doi:10.1016/j.asr.2005.04.07. D’Arrigo, R.D., E.R. Cook, M.E. Mann, and J.C. Jacoby, 2003. Tree–ring reconstructions of temperature and sea level pressure variability associated with the warm–season Arctic Oscillation since A.D. 1650. Geophys. Res. Lett., 30, 1549, doi:10.1029/2003GL017250. Groll, N., M. Widmann, J.M. Jones, F. Kaspar, and S. Lorenz, 2005. Simulated relationships between regional temperatures and large-scale circulation: 125 kyr BP (Eemian) and the preindustrial period. J. Clim., 18, 4035–4048. Jones, J.M., and M. Widmann, 2003. Instrument- and tree-ring-based estimates for the Antarctic Oscillation. J. Clim., 16, 3511–3524. Kaspar, F.,N. Ku¨hl, U. Cubasch, and T. Litt, 2005. A model-data-comparison of European temperatures in the Eemian interglacial. Geophys. Res. Lett., 32, L11703, doi:10.1029/ 2005GL022456. Lorenz, S.J., and G. Lohmann, 2004. Acceleration technique for Milankovtich type forcing in a coupled atmosphere–ocean circulation model: method and application for the Holocene. Clim. Dyn., 23, 727– 743, doi: 10.1007/s00382-004-0469-y. Rutherford, S., M.E. Mann, T.L. Delworth, and R.J. Stouffer, 2003. Climate field reconstructions under stationary and non-stationary forcing. J. Clim., 16, 462–479. Seelos, K., 2004. Entwicklung einer numerischen Partikelanalyse auf Basis digitaler Du¨nnschliffaufnahmen und Anwendung der Methode auf die ELSA-HL-2 Kernsequenz 66-41 m. Ph.D. thesis, University of Mainz, Germany, 173 pp. Seelos, K., and F. Sirocko, 2005. RADIUS – Rapid particle analysis of digital images by ultra-high resolution scanning of thin sections. Sedimentology, 52, 669–681, doi:10.1111/j.1365-3091.2005.00715.x. Shindell, D.T., G. A. Schmidt, M.E. Mann, D. Rind, and A. Waple, 2001. Solar forcing of regional climate change during the Maunder Minimum. Science, 294, 2149–2152. von Storch, H., E. Zorita, J.M. Jones, Y. Dmitriev, F. Gonzales-Rouco, and S. Tett, 2004. Reconstructing past climate from noise data. Science, 306, 679–682. Widmann, M., 2005. One dimensional CCA and SVD, and their relationship to regression maps. J. Clim., 18, 2785–2792. Zorita, E., and F. Gonzales-Rouco, 2002. Are temperature-sensitive proxies adequate for North Atlantic Oscillation reconstructions? Geophys. Res. Lett., 29, 481–484.

35. Orbital Forcing on Atmospheric Dynamics During the Last Interglacial and Glacial Inception Gerrit Lohmann1,2 and Stephan J. Lorenz3 1

Alfred Wegener Institute for Polar and Marine Research, Bussestrasse 24, D-27570 Bremerhaven, Germany 2 Research Center Ocean Margins, University of Bremen, Klagenfurterstr. 2, 28359 Bremen, Germany. 3 Max-Planck-Institute for Meteorology, Model & Data Group, University of Hamburg, Bundesstr. 53, D-20146 Hamburg, Germany ABSTRACT

Large-scale atmospheric patterns are examined on orbital timescales using the ECHO-G climate model which explicitly resolves the atmosphere–ocean–sea ice dynamics. It is shown that in contrast to boreal summer where the climate follows mainly the local radiative forcing, the boreal winter climate is strongly determined by modulation of circulation modes linked to the Arctic Oscillation/North Atlantic Oscillation and the El Nin˜o Southern Oscillation. We find that a positive phase of the Arctic Oscillation/ North Atlantic Oscillation is linked to below normal convection in the tropical Pacific. The related atmospheric circulation patterns induce nonuniform temperature anomalies, much stronger in amplitude than by the direct solar insolation. In concert with the direct solar insolation, this provides for a temperature drop over the Northern Hemisphere continents 115 000 years before present. We argue that the large-scale teleconnection pattern is important for the interpretation of proxy data as well as for the mechanisms responsible for the last interglacial, glacial inception and millennial climate variability. 35.1 INTRODUCTION A major factor in the pacing of climate change on timescales of tens of thousands of years seems to be the change in seasonal sunlight

distribution or insolation induced by variations in the Earth’s orbital parameters (e.g. Berger et al., 1993; Imbrie et al., 1993). The shape of the Earth’s orbit shifts from more elliptical to more circular at periodicities of about 100 000 years. Other rhythms are linked to the tilt of the Earth’s axis (periodicities of 29 000, 41 000 and 54 000 years) and the timing of the seasons relative to the Earth’s closests position to the Sun on the orbit (periodicities of 19 000 and 23 000 years). Palaeoclimatic modelling studies, aiming at reconstruction of past climate states, are usually performed on the basis of either time-slice simulations or transient simulations using models of reduced complexity (see also Claussen, this volume). Here, we want to concentrate on the dynamic evolution of the Northern Hemisphere atmospheric circulation during the past 140 000 years with special emphasis on the last interglacial temperature evolution making use of a coupled atmosphere–ocean–sea ice general circulation model. The basic assumption in the model setup is that the timescales of the astronomical forcing at the abovementioned frequencies can be separated from the much shorter timescales of the atmosphere–ocean–sea ice system, and that variations of the thermohaline circulation during interglacials played a minor role compared to other periods. These assumptions justify our procedure in the model configuration, where the astronomical forcing is accelerated (Lorenz and Lohmann, 2004).

528

Gerrit Lohmann and Stephan J. Lorenz

Various terrestrial and marine records show that the Holocene was marked by coherent patterns of climate variability at regional or global scales. These patterns of past climate variability are often related to changes in major phenomena that influence the last century climate variability like the El Nin˜o Southern Oscillation (ENSO) (Clement et al., 1999; Tudhope et al., 2001; Kitoh and Murakami, 2002) or the Arctic Oscillation/ North Atlantic Oscillation (AO/NAO) (Keigwin and Pickart, 1999; Rimbu et al., 2003a; Kim et al., 2004). Winter atmospheric circulation, surface air temperature and surface precipitation over the Northern Hemisphere are strongly influenced by the AO (Thompson and Wallace, 1998). The AO is the dominant mode of the atmospheric circulation variability in the Northern Hemisphere, and it is characterized by a meridional dipole in atmospheric sea-level pressure (SLP) between polar region and mid-latitudes. The climatic impacts are seen over the North Pacific, the North Atlantic and over the Northern Hemisphere continents, particularly as milder winters and lengthening of the growing season in Europe and northern Asia. The AO includes the North Atlantic Oscillation (NAO) phenomenon (Glowienka-Hense, 1990; Hurrel, 1995), which may be considered a different view of the same phenomenon (Thompson et al., 2000). The SLP anticorrelation between the Azores high and Icelandic Low has been used as a circulation index to characterize the atmospheric circulation in the North Atlantic realm (Defant, 1924; Kutzbach, 1970; Glowienka, 1985). We address the question whether these modes of variability are relevant on orbital timescales. The signature of the AO/NAO during the Holocene was identified in lacustrine data from the northeastern United States (Noren et al., 2002) as well as in the North Atlantic sediment cores (Keigwin and Pickart, 1999; Kim et al., 2004). Recent coral data from the northern Red Sea suggest a modulation of AO/NAO for the Eemian

and late Holocene climate (Felis et al., 2004). Here, we investigate the spatial heterogeneity of the climate evolution around the last interglacial applying a general circulation model and address the modulation of largescale climate modes on orbital timescales.

35.2 METHODOLOGY 35.2.1

The global circulation model

For the long-term climate simulations, we apply the coupled atmosphere–ocean general circulation model ECHO-G (Legutke and Voss, 1999). The atmospheric part of this model is the general circulation model ECHAM4 (Roeckner et al., 1996). The prognostic variables are calculated in the spectral domain with a triangular truncation at wave number 30 (T30), which corresponds to a Gaussian longitude–latitude grid of approximately 3:8 3:8 . The vertical domain is represented by 19 hybrid sigmapressure (terrain following) levels with the highest level at 10 hPa. The ECHAM4 model is coupled to the Hamburg ocean primitive equation model (HOPE) (Wolff et al., 1997). The ocean model includes a dynamic–thermodynamic sea ice model with snow cover. It is discretized on an Arakawa-E grid, with a resolution of approximately 2:8 2:8 . In the tropics, its meridional resolution is increased to 0.5 . The model consists of 20 irregularly spaced vertical levels with 10 levels covering the upper 300 m. The ECHAM model has been modified with respect to the standard version in order to account for subgrid scale partial ice cover (Gro¨tzner et al., 1996) which is also considered in the HOPE ocean model. 35.2.2

Orbital forcing

The ECHO-G model has been adapted to account for the influence of the annual distribution of solar radiation due to the slowly varying orbital parameters (Lorenz and Lohmann, 2004): the eccentricity of the

Orbital Forcing on Atmospheric Dynamics

Earth’s orbit, the angle between the vernal equinox and the perihelion on the orbit and the obliquity, i.e. the angle of the Earth’s rotation axis with the normal on the orbit. These parameters define the astronomical forcing of the climate system (Milankovitch, 1941). The calculation follows Berger (1978) providing the incoming solar radiation at the outer boundary of the atmosphere.

529

Figure 35.1 shows the changing solar irradiance due to the slowly evolving orbital parameters from 140 kyr BP until 100 kyr BP at the boreal summer (Fig. 35.1a) and the winter solstice (Fig. 35.1b), respectively. The insolation achieves its maximum between 130 and 125 kyr BP and around 105 kyr BP at Northern Hemisphere summer solstice. This is due to both a larger tilt of the Earth’s

Insolation at boreal summer solstice (Wm–2)

(a)

600

80N

560

580 560

450

540

Latitude

60N

40N 540 450

450 520

520

20N

500

500 400

400

450

450

EQ 400

350

400

300

350

300

350

250

300

350

20S

300

250

250

140

135

130

125

250

120

115

110

105

100

kyr BP Insolation at boreal winter solstice (Wm–2)

(b)

250

250

20N 300 350 400

250

300

300

350

350

350

450

450

400

400 500

500

520

450

520

450

500

Latitude

300

400

EQ

20S

250

40S

450

60S 540 560

140

560

580

135

130

125

540

80S

0 50 0 52

540 500

120

115

110

105

100

kyr BP

Fig. 35.1 Evolution of the latitudinal distribution of solar radiation for 140 to 100 kyr BP, following Berger (1978). Shown is the zonal average of insolation (a) at the summer solstice and (b) at the winter solstice in Wm 2 . Note the different range of latitudes: regions poleward of 20 C of the respective hemisphere with polar night are omitted, where the radiation keeps less than 200 Wm 2 and no significant change occurs.

530

Gerrit Lohmann and Stephan J. Lorenz

rotation axis and the precession cycle, moving the passage of the Earth through its perihelion from boreal summer in the early Eemian to begin of January today. At the winter solstice (Fig. 35.1b), a lack of insolation during the last interglacial is strongest at about 127 kyr BP and is centred round the equator. This is due to the precession cycle, since the distance to the Sun is then at maximum in boreal winter.

climate trends. It is found that the magnitude of orbitally forced Holocene trends is largely independent of the chosen acceleration factor. Throughout the whole experiments, fixed greenhouse gas concentrations (latest Holocene values: 280 ppm CO2, 700 ppb CH4, 265 ppb N2O) and modern values for vegetation, land–ocean and continental ice distribution and sea level were used.

35.2.3 Experimental setup

35.3 RESULTS

For the control run, we utilize constant greenhouse gas concentrations typical of the preindustrial era of the latest Holocene (end of 18th century): 280 ppm CO2, 700 ppb CH4 and 265 ppb NO2. Other boundary conditions are kept constant throughout the simulation, and modern irradiation is prescribed. This experiment has been integrated over 3000 years of model simulation into a climate state,whichisregardedasthequasi-equilibrium response of the model to preindustrial boundary conditions. It serves as a baseline and initialization of the simulations. Computer resources for running a complex model-like ECHO-G over the time period of the last interglacial and the Holocene are very demanding. Therefore, the timescale of the orbital forcing has been shortened by an acceleration factor of 100. The underlying assumptions are that the astronomical Milankovitch-type forcing operates on much longer timescales (millennia) than those inherent in the coupled atmosphere– ocean system, and that changes in deep temperatures have a negligible impact on the surface properties. For our long-term simulations, we neglect possible changes in deep-ocean circulation and concentrate on the atmospheric dynamics. The long runs over the last 140 000 years are continued to the next 30 000 years into the future. The last 140 000 years are therefore represented in 1400 model years. For a detailed description, we refer to Lorenz and Lohmann (2004) where the effect of the acceleration factor is evaluated on Holocene

Figure 35.2 displays the temperature evolution of the Northern Hemisphere from 140 kyr BP until 30 kyr after present under astronomical forcing. The only forcing in the experiment is given by astronomical forcing without changing greenhouse gas concentrations. The temperature evolution for the boreal summer and winter is dominated by the precessional cycle (modulated by eccentricity) and are mainly out of phase. For the boreal summer, a long-term cooling trend for 15 to 0 kyr BP is detected. At 3 kyr BP, the insolation at the top of the atmosphere almost reached the present energy level. In the next millennia, an increase in the boreal summer and a lowered boreal winter insolation is linked to the precessional cycle. For comparison, the red lines in Fig. 35.2 show the temperature evolution for the period 1900–2000 AD as evaluated with the same model, but with increased greenhouse gas concentrations (Lorenz and Lohmann, 2004). The model simulations can be interpreted as climate evolution for the interglacial periods, where changes in ice sheets played a minor role compared to the last glacial period. The amplitudes of insolation changes are much larger for the last interglacial, than for the Holocene, because the eccentricity – affecting the precessional cycle – is around 4.1% compared to roughly 1.7% in the Holocene. The changes in the future are going to further decrease, since the eccentricity, following its 100- and 40-kyr cycles, decreases near to zero during the following 50 kyr (compare Berger et al., this volume).

Orbital Forcing on Atmospheric Dynamics Boreal summer (JJA). ECHOG model forced by insolation

(a) 21

NH Temperature (°C)

20.5 20 19.5 19 18.5 18 17.5 17

−120

−100

−80

−60

−40

−20

0

20

Time (kyr) (b)

Boreal Winter (DJF). ECHOG model forced by insolation

NH Temperature (°C)

9 8.5 8 7.5 7 6.5 6 5.5 5

−120

−100

−80

−60

−40

−20

0

20

Time (kyr)

Fig. 35.2 Simulated Northern Hemisphere surface air temperature evolution for boreal summer (JJA) and winter (DJF). Periods where the climate is close to interglacials without huge North American and Eurasian ice sheets are indicated by grey bars. For comparison, the red lines show the temperature evolution for the period 1900–2000 AD as evaluated with the same model.

Figure 35.3 displays the absolute surface temperature simulated for boreal winter (December–January–February, DJF), summer (June–July–August, JJA) and annual mean at 130 kyr BP. As discussed by Joussaume and Braconnot (1997), a precise comparison of astronomical seasons under different orbital parameters would be to define the length of the months on the basis of their current angle on the Earth’s orbit.

531

Owing to Kepler’s law on the elliptical orbit the track speed is not constant. For example, the respectively defined time interval for the today’s JJA season (1st of June to 30th of August concerning the model’s year of 360 days) lies for the 130 kyr BP time slice from 25th of May until 19th of August (85 days in 130 kyr BP instead of 90 days today). We recalculate the new JJA and DJF surface temperature field for 130 kyr BP on this basis and find that the maximum error is less than 1 C with much lower values at low latitudes (not shown). Therefore, we stick to the simple calculation by using the fixed 90 days for the DJF and JJA seasons in the following analysis. For the time 125 kyr BP relative to 130 kyr BP, a warming over the Northern Hemisphere continents during DJF is partly compensated by a JJA cooling over most parts over the Northern Hemisphere (Fig. 35.4). The DJF cooling over the Greenland Sea and warming northeastern Europe is induced by an anomalous low-pressure system over Scandinavia (Fig. 35.5a). The DJF warming over northern North America is induced by advection of warm air that stems from the south (Fig. 35.5a). In our model, the surface temperature trends, especially during DJF, show pronounced spatial heterogeneity. For the boreal winter (120 minus 125 kyr BP), the Labrador Sea warms up to 8 C, whereas large parts of Europe tend to cool (Fig. 35.6a). The summer warming in the Labrador Sea area is due to the stored winter signal in the oceanic mixed layer (Fig. 35.6a). Such an anomaly pattern is detected for weakening of the Icelandic Low or the negative phase of the AO/NAO. The boreal summer cooling trend for 120 kyr BP (Fig. 35.6b) is very pronounced (up to 10 C over the Northern Hemisphere continents) which is linked to the insolation forcing (Fig. 35.1a). Figure 35.7 displays the anomalous sea-level pressure (SLP) where the weakening of the Icelandic Low is detected. For 115 kyr BP, a cooling of up to 6 C over the Northern Hemisphere continents during boreal winter is accompanied by an

532

Gerrit Lohmann and Stephan J. Lorenz

Fig. 35.3 Surface air temperature at 130 kyr BP, for winter (DJF), summer (JJA) and annual mean. The values represent an average over 20 years of the simulation (representing 2000 years) centred at 130 kyr BP. Units are C.

insolation-induced cooling during JJA (Fig. 35.8). The trend towards the weakening of the Icelandic Low or the negative phase of the AO/NAO is strongly enhanced

(Fig. 35.9). The weakened westerlies over the northern North Atlantic associated with the AO/NAO in its negative phase (Fig. 35.9a) are associated with the spatial

Orbital Forcing on Atmospheric Dynamics

533

Fig. 35.4 As in Fig. 35.3, but temperature difference for 125 minus 130 kyr BP.

heterogeneity seen in the regional temperatures (Fig. 35.8a). At southern high latitudes, a polar warming is accompanied by reduced westerlies (Fig. 35.9).

The modelled winter temperature difference between last interglacial and 115 kyr BP reveals a cooling pattern over the North Atlantic and adjacent continental areas

534

Gerrit Lohmann and Stephan J. Lorenz

Fig. 35.5 Sea-level pressure anomaly for 125 minus 130 kyr BP. Units are hPa. The anomaly is calculated over 20 years of the simulation (representing 2000 years) centred at 130 and 125 kyr BP, respectively.

that cannot be explained by direct insolation forcing, but that resembles the spatial signature of the AO/NAO. For 115 kyr BP, the winter surface temperature

anomaly indicates a tendency towards a low AO/NAO index state, with cold winters in central Europe due to reduced advection of warm oceanic air from the

Orbital Forcing on Atmospheric Dynamics

535

Fig. 35.6 As in Fig. 35.4, but for 120 minus 125 kyr BP.

west. The boreal winter circulation changes cannot be explained by local insolation, but with atmospheric circulation anomalies.

Motivated by this finding, we calculated an Icelandic Low circulation index for the experiment (Fig. 35.10). This index is calculated from the SLP region located at

536

Gerrit Lohmann and Stephan J. Lorenz

Fig. 35.7 As in Fig. 35.5, but for 120 minus 125 kyr BP.

Iceland (30 –0 W; 55 –80 N). The results do not sensitively depend on the chosen region of the index. In order to suppress the noise in the system, the first 10 vectors of a

singular spectrum analysis (Ghil et al., 2002) of the index were taken (lag: 100 years, the results are independent of the exact value). In our case, this procedure is

Orbital Forcing on Atmospheric Dynamics

537

Fig. 35.8 As in Fig. 35.4, but for 115 minus 120 kyr BP.

equivalent to a low-pass filter of the signal. Corresponding to this index (Fig. 35.10), the composite maps of geopotential heights at 500 and 200 hPa are calculated,

indicating an annular circulation pattern in the model (not shown). Figure 35.11 shows the composite map of the modelled boreal winter precipitation related to the Icelandic

538

Gerrit Lohmann and Stephan J. Lorenz

Fig. 35.9 As in Fig. 35.5, but for 115 minus 120 kyr BP.

Low index (Fig. 35.10). Interestingly, the strongest correspondence occurs in the tropical Pacific and Indian Oceans. Precipitation in the region (150 W–150 E; 5 S–5 N)

is a sensitive measure of the convection and strength of the Hadley Cell. The corresponding time series is included into Fig. 35.10.

Orbital Forcing on Atmospheric Dynamics

539

4

6

Above normal convection AO/NAO –

4

Above normal convection El Nino-like

3 2

SLP (hPa)

2 1 0

0 −1

−2

−2 −4

AO/NAO +

−120

−100

−80

−60

−40

−20

0

La Nina-like Below normal convection

−3

Below normal convection −6

20

Time (kyr)

−4

−120

−100

−80

−60

−40

−20

0

20

Time (kyr)

Fig. 35.10 Black line: Strength of the Icelandic Low during boreal winter (DJF). High values correspond to anomalously high pressure and negative phase of AO/NAO. The index is based on SLP for the region (30 –0 W; 55 –80 N). Red line: Normalized precipitation index (DJF) for the region (150 W–150 E; 5 S–5 N). The indices were low-pass filtered, see text.

We evaluate furthermore a normalized ENSO index from the zonal pressure difference between Darwin (80 –130 E; 5 S–5 N)and Tahiti (90 –140 W; 5 S–5 N)

Fig. 35.12 Blue line: ENSO index on orbital timescales based on the zonal sea-level pressure difference between Darwin (80 –130 E; 5 S–5 N) and Tahiti (90 –140 W; 5 S–5 N). Positive values in Fig. 35.12 indicate El Nin˜o-like, negative values La Nin˜a-like conditions. Red line as in Fig. 35.10.

in the model, the negative of the Southern oscillation index. Positive values in Fig. 35.12 indicate El Nin˜o, low values La Nin˜a conditions. The figure reveals a modulation on orbital timescales, mainly driven by precession.

Fig. 35.11 Composite map of the DJF precipitation related to the strength of the Icelandic Low shown in Fig. 35.10. Units are mm/month.

540

Gerrit Lohmann and Stephan J. Lorenz

35.4 DISCUSSION This paper discusses some results of a coupled atmosphere–ocean general circulation model experiment performed to study the effect of orbital forcing on climate. An acceleration technique is applied that has enabled them to run the equivalent of 140 000 years. With a normal setup of such models, this would not have been feasible due to the considerable computing requirements of these models. Here, we focus on the effect of orbital forcing on atmospheric dynamics during the last interglacial and the glacial inception. The main finding is that the orbital forcing regulates AO/NAO variability, which determines changes of boreal winter Northern Hemisphere temperatures. We find that the temperature reduction from 120 to 115 kyr BP in northern and central Europe is up to 6 C for boreal winter and up to 3 C for boreal summer, respectively (Fig. 35.8). The effect is strongest where the sea ice cover changed north of Sibiria (not shown), a similar amplifying mechanism is also detected for the Holocene (Lohmann et al., 2005). The simulated temperature changes are matched by pollenbased temperature reconstructions that reveal an approximately 3 C summer cooling and a 10 C winter cooling at the end of the Eemian in central Europe (Ku¨hl and Litt, 2003). The strong winter cooling in the model is due to the change in circulation (Fig. 35.9) associated with a weakened Icelandic Low. The mechanism for the Northern Hemisphere dynamics is related to tropical convection (Fig. 35.11) and strength of the Hadley Cell. The convection anomaly and its associated diabatic heating directly drive the atmospheric circulation (Fig. 35.10). The persistent tropical anomalies modulate the planetary-scale waves induced by persistent tropical circulation anomalies. Consistently, analysis of observational data indeed shows that the Pacific storm track is displaced poleward prior

to the onset of the negative NAO phase (Franzke et al., 2004, and references therein). This suggests that the latitudinal position of the Pacific storm track plays an important role for determining the NAO. A similar mechanism has been reported for quasi-decadal climate variability that stems from tropical Atlantic convection anomalies (Dima et al., 2001; Terray and Cassou, 2002; Dima and Lohmann, 2003). We note that the ENSO–NAO relationship is nonstationary for the instrumental record (Rimbu et al., 2003b) and can have opposite trends for the last decades than for the entire observational period (Hoerling et al., 2001; Rimbu et al., 2003b). However, negative SST anomalies in the tropical Pacific (i.e. La Nin˜a conditions) are statistically significantly related to an SLP pattern in the North Atlantic similar to the positive phase of NAO when the entire observational period is considered in the analysis (Pozo-Va`zquez et al., 2001). The applied acceleration technique assumes that the impact of changes in the deep ocean on surface temperatures is small compared to the direct effect of orbital forcing. This assumption is probably valid during interstadials when no external forcing mechanisms like meltwater releases are present. In our earlier paper (Lorenz and Lohmann, 2004), we presented arguments that the overturning circulation is only minor affected through the insolation forcing during the Holocene. Sea-surface temperature variations in the Holocene as obtained by alkenone data are mainly linked to the trend in AO/NAO, and possible variations in the overturning circulation seem to be of minor importance explaining temperature variations (Rimbu et al., 2004). With our method, we have excluded effects like the surface conditions in the Southern Ocean experiencing a delayed response to radiative forcing due to the influence of the deep-ocean circulation (Goosse and Renssen, 2001). The fact that part of the winter SST signal is seen in the summer anomaly pattern

Orbital Forcing on Atmospheric Dynamics

(Figs. 35.4, 35.6, 35.8) is due to the rectifying effect of the ocean: Upper-ocean temperature anomalies created over a deep mixed layer in winter may be preserved in the following summer and reappear in the following fall and winter. This could have implications for the documentation of the temperature signal in proxies. The winter signal is preserved for long-term climate variability, and the documented signal is biased towards the winter climate signal, even when not identical with the season when blooming takes place. This finding is complementary to the usual assumption of the astronomical theory that the summer insolation forcing dominates over the winter forcing. The dominant winter SST signal is furthermore in contrast to other model simulations (Hall et al., 2005) where a mixed layer model has been applied for representing the ocean. Our analysis indicates that on orbital timescales part of climate variability is related to the modulation of AO/NAO and ENSO modes. We have neglected variations in many subsystems of the Earth climate system relevant for the past 140 kyr, e.g. the ice sheets, carbon cycle and vegetation. The modelling concept is therefore different to some of the Earth system modelling approaches in this chapter (Gro¨ger et al.; Loutre et al.; Kagayama and Ramstein; Kubatzki et al., this volume). Here, the effect of insolation forcing on atmospheric dynamics was isolated in the model set-up. The results should be interpreted in a climatic sense only for the last interglacial and the Holocene. In the composite analysis, we have taken the whole period from the model output in order to get a proper statistics and to better understand the associated mechanisms. It is very interesting to compare our results for the Holocene branch of our simulation with other findings. Predominant La Nin˜a conditions in the tropical Pacific during the early Holocene are suggested also by other model simulations. A simulation with an intermediate complexity ENSO model (Zebiak and Cane, 1987) forced by variations in heating due to orbital variations in

541

seasonal insolation shows weaker ENSO activity in the early to mid-Holocene than in the late Holocene (Clement et al., 1999). Another model simulation shows that tropical Pacific temperature and circulation patterns during 6 kyr BP are similar to those observed at the present day La Nin˜a period (Kitoh and Murakami, 2002). A sizable tropical SST response to the precessional component of Milankovich variations is also seen in the three-dimensional global coupled ECBilt model (Tuenter et al., 2003), thus indicating that the rectification effect seen by Clement et al. (1999) is not model dependent. It has been further argued that the precessional component of Milankovich variations may be relevant also for deglaciation processes through low-to-high latitude teleconnections (Rodgers et al., 2003). Our finding of AO/NAO modulation on orbital timescales is consistent with alkenone-based SST reconstructions for the Holocene (Lorenz et al., 2005). A positive phase of the AO/NAO at 6 kyr BP relative to present day is accompanied by low insolation in the tropical region. Furthermore, modelling experiments reveal that changes in the regional circulation in the Nordic Seas during the Holocene are compatible with the Milankovitch forcing (Lohmann et al., 2005). Additional high-resolution spatiotemporal proxy datasets are necessary to examine the spatial and temporal patterns of climate variability during the last interglacial. For this purpose, a new piston core with laminated sediments has been recently recovered off Peru in order to trace ENSO variation (Rein et al., this volume). Based on our results and the analogy with the instrumental period, we argue that the modulation of convection in the tropical Pacific Ocean on orbital timescales is compatible with an enhanced positive phase of AO/NAO during the Eemian and early Holocene. The positive phase of the AO/NAO during the Eemian is furthermore consistent with proxy records indicating mild winters in Europe (Zagwijm, 1996; Aalbersberg and Litt, 1998; Klotz et al., 2003; see

542

Gerrit Lohmann and Stephan J. Lorenz

also Chapter 13 this volume) and associated cold winters in the northern Red Sea (Felis et al., 2004). The astronomically driven AO/ NAO in the model is in line with the analysis of the seasonal signal in an atmospheric circulation model (Hall et al., 2005). Future studies should include the interactions of atmospheric dynamics with other climate components such as vegetation and land ice. Vegetation seems to amplify the climate signal during the glacial inception (see Groeger et al., Kubatzki et al., this volume). 35.5 CONCLUSIONS Applying an acceleration technique for Milankovitch forcing (Lorenz and Lohmann, 2004), the modulation of largescale climate modes on orbital timescales is addressed. Simulated Northern Hemisphere winter temperatures at mid-to-high latitudes are strongly linked to variations in the atmospheric dynamics on orbital timescales. We find that a reduced (increased) strength of the Icelandic Low and the negative (positive) phase of AO/NAO is associated with an increase (decrease) in tropical convection and El Nin˜o-like (La Nin˜a-like) conditions. Interestingly, similar behaviour is found for the instrumental record when the entire observational period is considered, i.e. negative SST anomalies in the tropical Pacific are related to the positive phase of NAO (PozoVa´zquez et al., 2001; Rimbu et al., 2004). In our simulation, the transition from a more positive to a more negative phase of AO/NAO is related to the increase of tropical insolation during the winter season associated with the Earth’s precession cycle. In contrast to boreal winter, the summer is largely affected by the local radiative forcing except for regions with a large mixed-layer depth where the winter signal is stored. How can the AO/NAO variations on orbital timescales be understood? The fundamental processes determining the AO/NAO are linked to internal variability (on interannual timescales), ocean dynamics

in the North Atlantic (decadal timescales), coupling with the stratosphere, and/or to interactions with other climate components. Three different concepts exploring such variability can be classified (Franzke et al., 2004): planetary-scale instabilities, interactions among synoptic-scale transients and remotely forced stationary waves. Our phenomenon belongs therefore to the third concept as proposed by Hoskins and Karoly (1981) which can be viewed as a boundaryvalue problem. Given the relatively short timescale of the atmospheric variability, our definition of the atmospheric modes is related to the modulation of the modes on timescales longer than that of the life cycles of the teleconnection patterns. For the Eemian, the changed atmospheric circulation reflects a temperature pattern similar to the positive phase of the AO/ NAO. Consistent with pollen-based temperature reconstructions (Ku¨hl and Litt, 2003), we find a temperature drop from 120 to 115 kyr BP in northern and central Europe for boreal winter and summer, respectively. A large part of the strong winter cooling is due to the change in Northern Hemisphere circulation associated with a weakened Icelandic Low. It can be speculated that the glacial inception is due to the direct effect of insolation and atmospheric dynamics, i.e. summer cooling in combination with a negative phase of the AO/NAO during winter. As a logical next step, future studies should include the interactions of atmospheric dynamics with other climate components such as vegetation and land ice, possibly amplifying the climate response as detected in our simulation. Recently, Sirocko et al. (2005) presented evidence for multicentennial cooling events in the central European continent and attributed this to changes in the oceanic meridional overturning circulation. Here, we provide an alternative hypothesis that abrupt cooling events during the last glacial inception (Sirocko et al., 2005) may be linked to strong shifts in the Northern Hemisphere atmospheric circulation. More high-resolution proxy data are

Orbital Forcing on Atmospheric Dynamics

necessary in order to evaluate the spatial pattern related to such events. Future studies will also examine common and different mechanisms of climate variations during the Holocene and the last interglacial. Given the apparent involvement of the atmospheric dynamics on orbital timescales, mechanisms triggering the AO/NAO and ENSO should therefore not only provide interpretation of proxy data but also encompass and connect remote regions, and identify their effects for long-term climate variability.

ACKNOWLEDGEMENTS We appreciate constructive comments by two anonymous referees. This work was supported by Bundesministerium fu¨r Bildung und Forschung through DEKLIM, and by Deutsche Forschungsgemeinschaft through DFG Research Centre Ocean Margins at Bremen University

REFERENCES Aalbersberg, G., Litt, T., 1998. Multiproxy climate reconstructions for the Eemian and Early Weichselian. J. Quaternary Sci. 13, 376–390. Berger, A.L., 1978. Long-term variations of daily insolation and Quaternary climatic changes, J. Atmos. Sci. 35, 2362–2367. Berger, A., Loutre, M-F., Tricot, C., 1993. Insolation and Earth’s Orbital Periods. J.Geophys. Res. 98(D6), 10341–10362. Clement, A.C., Seager, R., Cane, M.A.,1999. Orbital controls on ENSO and the tropical climate. Paleoceanography, 14, 441–456. Defant, A., 1924. Die Schwankungen der atmospha¨rischen Zirkulation u¨ber dem nordatlantischen Ozean im 25-jahrigen Zeitraum 1881–1905. Geogr. Annaler 6, 13–41. Dima, M., Rimbu, N., Stefan, S., Dima, I., 2001. Quasi-decadal variability in the Atlantic Basin involving tropics-midlatitudes and ocean– atmosphere interactions, J.Climate 14, 823–832. Dima, M., Lohmann, G., 2004. Fundamental and derived modes of climate variability. Application to biennial and interannual timescale. Tellus 56A, 229–249.

543

Felis, T., Lohmann, G., Kuhnert, H., Lorenz, S.J., Scholz, D., Pa¨tzold, J., Al-Rousan, S.A., AlMoghrabi, S.M., 2004. Increased seasonality in Middle East temperatures during the last interglacial period. Nature 429, 164–168. Franzke, C., Lee, S., Feldstein, S.B., 2004. Is the North Atlantic Oscillation a breaking wave? J. Atmos. Sci. 61, 145–160. Ghil, M., Allen, R.M., Dettinger, M.D., Ide, K., Kondrashov, D., Mann, M.E., Robertson, A., Saunders, A.,Tian, Y., Varadi, F., Yiou, P., 2002. Advanced spectral methods for climatic time series, Rev. Geophys. 40(1), 1–41, 1029/ 2000GR000092. Glowienka, R., 1985. Studies on the variability of Icelandic Low and Azores High between 1881–1983. Contrib. Atmos. Phys. 58, 160–170. Glowienka-Hense, R., 1990. The North Atlantic Oscillation in the Atlantic-European SLP. Tellus 42A, 497–507. Goosse, H., Renssen, H., 2001. A two-phase response of the Southern Ocean to an increase in greenhouse gas concentrations. Geophys. Res. Letters 28, 1023–1026. Gro¨tzner, A., Sausen, R., Claussen, M., 1996. The impact of sub-grid scale sea–ice inhomogeneities on the performance of the atmospheric general circulation model ECHAM3, Climate Dyn. 12, 477–496. Hall, A., Clement, A., Thompson, D.W.J., Broccoli, A. Jackson, C., 2005. Atmospheric dynamics govern northern hemisphere wintertime climate variations forced by changes in earth’s orbit. J. Climate 18, 1315. Hoerling, M.P., Hurrell, J.W., Xu, T., 2001. Tropical origins for recent North Atlantic climate change, Science 292, 90–92. Hoskins, B.J., Karoly, D.J., 1981. The steady-state linear response of a spherical atmosphere to thermal and orographic forcing. J.Atmos. Sci. 38, 1175–1196. Hurrell, J.W., 1995. Decadal trends in the North Atlantic Oscillation: Regional temperatures and precipitation, Science 269, 676–679. Imbrie, J., Berger, A., Boyle, E.A., Clemens, S.C., Duffy, A., Howard, W.R., Kukla, G., Kutzbach, J., Martinson, D.G., McIntyre, A., Mix, A.C., Molfino, B., Morley, J.J., Peterson, L.C., Pisias, N.G., Prell, W.L., Raymo, M.E., Shackleton, N.J., Toggweiler, J.R., 1993. On the structure and origin of major glaciation cycles: 2. the 100,000-year cycle, Paleoceanography 8, 701–738. Joussaume, S., Braconnot, P., 1997. Sensitivity of paleoclimate simulation results to season definitions. Journal of Geophysical Research 102(D2), 1943–1956.

544

Gerrit Lohmann and Stephan J. Lorenz

Keigwin, L.D., Pickart, R.S., 1999. Slope water current over the Laurentian fan on interannual to millennial time scales, Science 286, 520–523. Kim, J.-H., Rimbu, N., Lorenz, S.J., Lohmann, G., Nam, S.-I., Schouten, S., Ru¨hlemann, C., Schneider, R.R., 2004. North Pacific and North Atlantic seasurface temperature variability during the Holocene. Quat. Sci. Rev. 23, 2141–2154. doi:10.1016/j.quascirev.2004.08.010. Kitoh, A., Murakami, S., 2002. Tropical Pacific climate at the mid-Holocene and the Last Glacial Maximum simulated by a coupled atmosphere–ocean general circulation model, Paleoceanography 17, doi:10.1029/2001PA000, 19–1–19–13. Klotz, S., Guiot, J., Mosbrugger, V., 2003. Continental European Eemian and early Wuermian climate evolution: Comparing signals using different quantitative reconstruction approaches based on pollen. Global and Planetary Change 36(4), 277–294. Ku¨hl, N., Litt, T., 2003. Quantitative time series reconstruction of Eemian temperature at three European sites using pollen data. Veget. Hist. Archaeobot. 12, 205–214. Kutzbach, J.E., 1970. Large-scale features of monthly mean Northern Hemisphere anomaly maps of sealevel pressure. Mon. Wea. Rev. 98, 708–716. Legutke, S., Voss, R., 1999. The Hamburg atmosphere–ocean coupled circulation model ECHO-G, Technical report No. 18, Deutsches Klimarechenzentrum, Hamburg. Lohmann, G., Lorenz, S.J., Prange, M., 2005. Northern high-latitude climate changes during the Holocene as simulated by circulation models, in The Nordic Seas: An Integrated Perspective, H. Drange, T. Dokken, T. Furevik, R. Gerdes, and W.Berger (eds.), Geophysical Monograph 158, American Geophysical Union, Washington DC, pp. 273–288. doi:10.1029/ 158GM18. Lorenz, S.J., Lohmann, G., 2004. Acceleration technique for Milankovitch type forcing in a coupled atmosphere–ocean circulation model: method and application for the Holocene. Climate Dyn., 23, 727–743. doi:10.1007/s00382–004–0469-y. Lorenz, S.J., Kim, J.-H., Rimbu, N., Schneider, R.R., Lohmann, G., 2005. Orbitally driven insolation forcing on Holocene climate trends: evidence from alkenone data and climate modeling, Paleoceanography (in press). Milankovitch, M., 1941. Kanon der Erdbestrahlung und seine Anwendung auf das Eiszeitenproblem, 133, Royal Serb. Acad. Spec. Publ., Belgrad. Noren, J.A., Bierman P.R., Steig J.E., Lini A., Southon J., 2002. Millenial-scale storminess variability in

the northeastern United States during the Holocene epoch, Nature 419, 821–824. 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. Clim. 14, 3408–3420. Rimbu, N., Lohmann, G., Kim, J.-H., Arz, H.W., Schneider, R., 2003a. Arctic/North Atlantic Oscillation signature in Holocene sea surface temperature trends as obtained from alkenone data. Geophys Res. Lett. 30. doi:10.1029/ 2002GL016570. Rimbu, N., Lohmann, G., Felis, T., and Pa¨tzold, J., 2003b. Shift in ENSO teleconnections recorded by a Red Sea coral. J. Climate 16, 1414–1422. Rimbu, N., Lohmann, G., Lorenz, S.J., Kim, J.-H., Schneider, R., 2004. Holocene climate variability as derived from alkenone sea surface temperature reconstructions and coupled ocean–atmosphere model experiments. Climate Dyn. 23, 215–227. doi:10.1007/s00382–004–0435–8. Rodgers, K., Lohmann, G., Lorenz, S., Schneider, R., Henderson, G., 2003. A Tropical Mechanism for Northern Hemisphere Deglaciation. Geochem., Geophys., Geosyst. 4(5), 1046, doi: 10.1029/ 2003GC0000508. Roeckner, E., Arpe, K., Bengtsson, L., Claussen, C.M., Du¨menil, L., Esch, M., Giorgetta, M., Schiese, U., Schulzweida, U., 1996. The atmospheric general circulation model ECHAM4:Model description and simulation of the present day climate, Report 218, Max-Planck-Institut fu¨r Meteorologie, 90 pp. Sirocko, F., Seelos, K., Schaber, K., Rein, B., Dreher, F., Diehl, M., Lehne, R., Ja¨ger, K., Krbetschek, M., Degering, D., 2005. A late Eemian aridity pulse in central Europe during the last glacial inception. Nature 436, 833–836. doi: 10.1038/nature03905. Terray, L., Cassou, C., 2002. Tropical Atlantic sea surface temperature forcing of quasi-decadal variability over the North Atlantic European region. J. Climate 22, 3170–3187. Thompson, D.W.J., Wallace, J.W., 1998. The Arctic Oscillation signature in the wintertime geopotential height and temperature fields, Geophys. Res. Lett. 25, 1297–1300. Thompson, D.W.J., Wallace, J.W., Hegerl, G.C., 2000. Annular modes in the extratropical circulation. Part II Trends, J. Climate 13, 1018–1036. Tudhope A.W., Chilcott, C.P., McCulloch, M.T., Cook, E.R., Chappell, J., Ellam, R.M., Lea , D.W., Lough, J.M., Shimmield, G.B., 2001. Variability in the El Nin˜o-Southern Oscillation through a glacial–interglacial cycle. Science 291, 1511–1517.

Orbital Forcing on Atmospheric Dynamics Tuenter, E., Weber, S.L., Hilgen, F.J., Lourens, L.J., 2003. The response of the African summer monsoon to remote and local forcing due to precession and obliquity, Global Planet. Change 36, 219–235. Wolff, J.-O., Maier-Reimer, E., Legutke, S., 1997. The Hamburg ocean primitive equation model

545

HOPE, Technical Report 13, Deutsches Klimarechenzentrum, Hamburg, Germany. Zagwijn, W.H., 1996. An analysis of Eemian climate in Western and Central Europe. Quat. Sci. Rev. 15, 451–469. Zebiak, S.E., Cane, M.A., 1987. A model for El Nin˜oSouthern Oscillation. Mon. Wea. Rev., 115, 2262–2278

This page intentionally left blank

36. Interglacials as Simulated by the LLN 2-D NH and MoBidiC Climate Models M.F. Loutre1, A. Berger1, M. Crucifix2, S. Desprat3 and M.F. Sa´nchez Gon˜i3 1

Universite´ catholique de Louvain, Chemin du Cyclotron, 2 BE-1348 Louvain-la-Neuve, Belgium 2 Hadley Centre for Climate Prediction and Research, Met Office, Fitzroy Road, Exeter, EXI 3PB, U.K. 3 EPHE, EPOC, UMR-CNRS 5805, Universite´ Bordeaux 1, Avenue des Faculte´s, 33405 Talence, France

ABSTRACT Two Earth system models of intermediate complexity (EMICs), i.e. LLN 2-D NH and MoBidiC, were designed in Louvain-laNeuve to test the astronomical theory of palaeoclimate. The purpose was to see whether the astronomically driven insolation is the main driver of climate change over the last glacial–interglacial cycles. It also aims at identifying the major processes and feedbacks at work in the climate system. Here, we report on the results obtained for the interglacial periods of the last 800 kyr and for the future over the next 100 kyr. Major processes governing the response of the modelled climate system to insolation and/or CO2 changes are related to the albedo-temperature and water vapour– temperature feedbacks, to the taiga-tundra direct and indirect impacts on highlatitudes surface albedo, to the altitude and continental effects on the precipitation over the ice sheets, to the lagging lithospheric response to the ice-sheet loading and to the mechanical destabilisation of the ice sheets through the rapid melting of their southern front as compared to the northern. 36.1 MILANKOVITCH THEORY OF PALAEOCLIMATES As discussed elsewhere (insolation; this volume), Milankovitch (1941) argued that

insolation changes in the high northern latitudes during the summer season were critical to the waxing and waning of continental ice sheets. According to the mathematics of insolation, a minimum in the Northern Hemisphere caloric summer insolation at high latitudes is required for entering into glaciation. This is the case for (1) a Northern Hemisphere summer occurring at the aphelion; (2) a maximum eccentricity, which leads to a large distance between the Earth and the Sun at the aphelion; and (3) a minimum obliquity implying a weak seasonal contrast. Given these Milankovitch conditions, the northern high latitudes would remain cool enough in summer for preventing snow and ice from melting, thereby generating a positive feedback mechanism. Moreover, mild winters would allow substantial evaporation in tropical and temperate latitudes, an intensified general circulation due to an intensified latitudinal gradient of insolation in the Northern Hemisphere, and, as a consequence, abundant snowfalls in polar latitudes (Berger, 1980). A simple linear version of the Milankovitch hypothesis would therefore predict that the total ice volume and climate would vary with the same regular pattern as the insolation. However, the climate system is far from being linear, and models are needed to test its response to the astronomical forcing. Such time-dependent coupled climate models were designed in Louvain-la-Neuve to answer that problem.

548

M.F. Loutre et al.

36.2 LLN 2-D NH AND MoBidiC CLIMATE MODELS The first LLN climate model (LLN 2-D NH model) describes the Northern Hemisphere (NH) atmosphere, ocean mixed layer, sea ice, ice sheets and snow-covered and snow-free land and their interactions (Galle´e et al., 1991). It is a two-dimensional (latitude–altitude) sectorially averaged model. In each latitudinal belt, the surface is divided into at most seven oceanic or continental surface types, each of which interacting separately with the subsurface and the atmosphere. Special attention is paid to the albedo of snow, of vegetation in the northern regions and of sea ice. The atmosphere–ocean model is asynchronously coupled to a model of the three main NH ice sheets and their underlying bedrock. The coupled climate model is then forced by the astronomically derived insolation for each day and latitude (Berger, 1978) and by the atmospheric CO2 concentration (this model did not contain an interactive carbon cycle). More details on the model are given in Galle´e et al. (1991), and also in Berger et al. (1990b) for the ice sheet – lithosphere model, in Berger et al. (1989) for the upper ocean and in Berger et al. (1994) for the radiative convective scheme. Simulation of the present climate shows that the model is able to reproduce the main features of the atmospheric general circulation and the seasonal cycles of the oceanic mixed layer, of the sea ice and of the snow cover (Galle´e et al., 1991). The second version of the LLN climate model, MoBidiC, covers both the Northern and Southern Hemispheres. It includes several components not represented in the first model: the deep ocean (Hovine and Fichefet, 1994; Crucifix et al., 2001b; Crucifix and Loutre, 2002), the continental biosphere (Brovkin et al., 1997) and the ocean carbon cycle. The atmosphere–ocean–ice–sheet coupling includes a run-off scheme and iceberg calving in order to account for the possible impact of meltwater on ocean salinity (Crucifix and Berger, 2002).

Both models have been extensively used for transient simulations of all or part of the most recent glacial–interglacial cycles and are particularly useful in understanding the interglacials. 36.3 THE HOLOCENE (9 kyr BP– PRESENT) In the transient response of MoBidiC to insolation and CO2 variations over the last 9 kyr (Crucifix and Loutre, 2002), both oceans and continents are cooling monotonously, a result that confirms the early simulation with the LLN 2-D NH model (Berger et al., 1990a). Globally averaged annual mean temperature decreases by 0.15 C. North of 60 N, this cooling is largest in spring and summer (up to 6 C over the continents). The cooling is faster between 4 and 1 kyr Before Present BP. Consequently, the northern tree line shifts southwards by about 600 km; most of this shift takes place between 4 and 1 kyr. Moreover, reorganisations of the boreal forest introduce a lag of about 200 years in the system between transient and equilibrium simulations (Crucifix et al., 2002). At the same time, arctic sea ice increases which enhances the cooling effect. The thermohaline circulation exhibits a slight slowdown throughout the Holocene. The feedback analysis reveals that the ocean and vegetation taken separately have opposite impacts on the summer continental temperature trends. Owing to its thermal inertia, the ocean reduces the summer continental cooling trend. On the contrary, the southward shift of the tree line (i.e. expansion of tundra) resulting from shorter and colder growing seasons enhanced the summer cooling trend over both the oceans and the continents. This counteracting influence subsists when both vegetation and ocean temperatures are calculated; the dominant impact is, by far, from vegetation (see Fig. 12 in Crucifix et al., 2002). This illustrates the strong

Interglacials as Simulated by the LLN 2-D NH and MoBidiC

*GDD0 is the annual sum of the continental surface air temperature for days which temperature exceeds 0 C.

in Mediterranean was higher than today during the early Holocene. 36.4 OUR FUTURE Several CO2 scenarios were used as forcing in LLN 2-D NH to simulate the climate of the next 100 kyr, a time interval of low amplitude in the insolation variations. Amongst them, constant CO2 concentrations and Vostok CO2 (Jouzel et al., 1993; Petit et al., 1999) shifted by either 131 kyr or 414 kyr were used. Most of the scenarios of variable CO2 concentration (Loutre and Berger, 2000; Berger and Loutre, 2002) lead to an exceptionally long interglacial from 5 to 50 kyr After Present AP (Fig. 36.1); after 60 kyr AP, the fast continental ice volume increase is briefly interrupted by a short reversal, which is followed by a slower increase towards the next glacial maximum reached shortly after 100 kyr AP (Loutre, 2003). An early glacial inception can take place only if the CO2 concentration remains lower than 210 ppmv, and a large increase in the ice volume after 50 kyr AP is simulated

CO2 (ppmv)

300

250

200 16

0

NH temperalure (°C)

NH ice volume (105 km3)

impact of the vegetation shift on both oceans and continents, especially in spring and early summer, but the synergy is weak between vegetation and ocean throughout the Holocene. This differs substantially from the conclusions drawn by Ganopolski et al. (1998), a difference which can probably be best explained by their higher sea ice sensitivity, especially in winter. Sensitivity studies allowed us also to assess the respective influence of the different components of the external forcing: precession, obliquity and CO2. The impact of precession on annual mean temperatures results from a change in the annual mean surface albedo of the continents, due to changes in both the snow field and tundra, which are both sensitive to seasonal climatic variations. The high latitude cooling under a full astronomical forcing is much larger than the sum of the individual contributions of precession and obliquity. The effect of CO2 is enhanced when orbital forcing is taken into account, suggesting that throughout the Holocene, the highlatitude climate may have been temporarily highly sensitive to CO2 (mainly from 4 to 1 kyr BP). Part of these results is related to the nonlinear relationship between degree-days growing above 0 C (GDD0*) and tree fraction in high latitudes. During the Holocene, both precession and obliquity contribute to decrease GDD0, leading to values for which vegetation is much more sensitive to climate. In particular, this explains why small changes in the CO2 concentration have an important impact on the tree line at about 2 kyr BP. In partial agreement with the model, palaeobotanic reconstruction suggested a maximum temperature some 6 kyr ago (Huntley and Prentice, 1988). Foraminifera and diatom-based reconstruction (Koc¸ et al., 1993; Kerwin et al., 1999; Perez-Folgado et al., 2004) suggested that the sea-surface temperature (SST) in the North Atlantic and

549

10 20 30 40 50

100

50

0

Time (kyr)

–50

15 14 13 12 100

50

0

–50

Time (kyr)

Fig. 36.1 Simulated continental ice volume (bottom left) and annual mean temperature (bottom right) of the Northern Hemisphere using LLN 2-D NH climate model forced by insolation and different scenarios of CO2 concentration (top left) for the future. The CO2 scenarios are the Vostok values shifted by either 131 (dark blue) or 414 kyr (cyan) and a scenario from a threshold model (magenta) (Paillard, personal communication).

550

M.F. Loutre et al. 6

3

Northern hemisphere ice volume (10 km ) 0

10 0

20 5

30 10 40

210 ppmv 220 ppmv 230 ppmv 240 ppmv 250 ppmv

40

30

20

10

Time (kyr AP)

260 ppmv 270 ppmv 280 ppmv 290 ppmv

50

100

50

0

Time (kyr AP)

Fig. 36.2 Simulated ice volume of the Northern Hemisphere using LLN 2-D NH climate model forced by insolation and different constant CO2 concentration ( from 210 to 290 ppmv). The integrations start with present-day conditions for the ice sheets (3.2 106 km3 of ice) and the present-day bedrock depression (left) over the next 130 kyr, (right) enlargement from 10 to 40 kyr AP (Loutre and Berger, 2000) (credit: Kluwer Academic Publishers, Climatic Change, 46, 2000, p. 67, Future climatic change: are we entering an exceptionally long interglacial? Fig. 3).

only if CO2 remains lower than 240 ppmv (Fig. 36.2). An important and robust feature of the simulated future is thus the expected very long interglacial that already started 10 kyr ago. This feature is independent of human activity. Indeed the simulations were only forced by the insolation and CO2 forcings of the last 10 kyr, without taking into account any anthropogenic forcing. Over this period, as during the marine isotopic stage 11 (MIS 11, see Section 6), CO2 remains high over thousands of years, counteracting the direct effect of a declining insolation. This result was actually confirmed by Peltier and Vettoretti (2001): under the present-day insolation regime and preindustrial CO2 concentration, the Canadian climate general circulation model is indeed not able to simulate a glacial inception, although it did for the end of the Eemian. The human impact on climate amplifies this long interglacial with a complete melting of the Greenland ice sheet if the CO2 concentration becomes larger than

750 ppmv in our model (Berger, 2001). But when did human activities start to modify the natural behaviour of the climate system? Recently, Ruddiman (2003, 2005) suggested that anthropogenic emissions of greenhouse gases due to land-use practices may have altered the atmospheric CO2 concentration thousands of years ago and not only since the Industrial Revolution, as it is generally thought. He hypothesised that the increasing CO2 trend since 8 kyr BP years is ‘anomalous’ and must be related to human activities. Starting from this hypothesis, we tested whether an early (about 8 kyr BP) decrease in CO2 concentration could have led to a glacial inception. Three scenarios of CO2 evolution were tested (Fig. 36.3 and Crucifix et al., 2005). In the first scenario, CO2 starts to decrease at 11 kyr BP and from 7 kyr BP follows the CO2 behaviour observed during MIS 9, i.e. CO2 continues to decrease although it undergoes a large reversal between 6 and 3 kyr AP. In

Interglacials as Simulated by the LLN 2-D NH and MoBidiC

Insolation 65N June (Wm–2)

response to this forcing, the model still simulates a long interglacial with a small amount of ice in the Northern Hemisphere over the next 50 kyr. In the second scenario, CO2 is assumed to decrease linearly

550

500

450

CO2 (ppmv)

275

250

225

200

NH ice volume (106 km3)

0

10

20

30

40

50

100

50

0

–50

Time (kyr)

Fig. 36.3 Simulated continental ice volume ( bottom panel) of the Northern Hemisphere using LLN 2-D NH climate model forced by insolation (top) and different scenarios of CO2 concentration (middle). The purpose of the simulation is to test whether an alternative CO2 concentration over the Late Holocene could have induced an early glacial inception. The CO2 forcing for the future corresponds to the values measured in the Vostok core over the last glacial–interglacial cycle (black, red and green) and over MIS 9 (orange).

551

from 263 ppmv at 11 kyr BP to 206 ppmv at 4 kyr AP. Then, it increases up to 228 ppmv at 11 kyr AP and finally follows the Vostok CO2 values shifted by 111 kyr. In this case, the interglacial still remains long although there is a re-growth of the Northern Hemisphere ice sheet, which culminates to 9 106 km3 at 22 kyr AP. In the third scenario, the CO2 decreases very quickly after 11 kyr BP. It reaches 210 ppmv at 7 kyr BP and remains at this low level for 13 kyr. Then, it follows the previous scenario. An early re-growth of the Northern Hemisphere ice sheets is simulated, starting already before present, with an ice volume culminating at almost 30 106 km3 as early as 23 kyr AP. These results confirm the major importance of the CO2 concentration over the last 10 000 years, during which eccentricity is low. They also suggest that, although a different CO2 trend during the Late Holocene, namely a fast and large decrease of the CO2 concentration, may have prevented the climate to remain in an interglacial state, as claimed by Ruddiman (2003), the deviation from the measured CO2 values is so large that it is highly unlikely that human activities were responsible for it. Moreover, these simulations are not accounting for human-related land-use changes, which are responsible for high CO2 concentration during the Late Holocene, according to Ruddiman (2003). However, it might be expected that deforestation in the high latitudes would induce a cooling through the vegetation–temperature–albedo feedback. Indeed, if forest is replaced by grass or bare soil during winter, the winter surface albedo of this snow-covered land will be increased, inducing a temperature cooling that will start a feedback loop. Consequently, it might be expected that deforestation would favour a cooling in the high latitudes.

36.5 THE LAST INTERGLACIAL, MIS 5e The simulation of the climate of the last interglacial starts from an equilibrium state,

552

M.F. Loutre et al.

where orbital parameters are computed for 126 kyr BP and the CO2 concentration is fixed at 261 ppmv. The topography of Antarctica is as today, while the Greenland ice sheet is reduced to half its present-day volume. The transient simulation is forced by the variation of the orbital parameters from 126 to 115 kyr BP. First, the CO2 concentration and ice sheets are kept to their initial values (Crucifix and Loutre, 2002). While CO2 concentration hardly varied during the Eemian, there is evidence of global ice volume variations of the order of 10 to 15 106 km3 . Therefore, this simulation can only be used as a first-order approximation to study climate processes at work during the Eemian. A second experiment was also performed for which the ice-sheet component was fully coupled and the CO2 concentration varied according to the Vostok values (Sa´nchez Gon˜i et al., 2005). Both experiments are qualitatively similar, although the amplitude and the absolute values of the climate response might slightly differ. 36.5.1 Insolation and ice volume Insolation at 65 N in June, taken here as a guideline, is minimum around 140 kyr BP and peaks at 128 kyr BP before reaching another minimum at 116 kyr BP. The maxima of the simulated (LLN 2-D NH) Northern Hemisphere ice volume are located at 134 and 109 kyr BP, lagging behind the insolation by about 6000 years. The minimum ice volume is reached at 126 kyr BP and lasts 10 000 years, covering the whole period during which insolation is decreasing from a maximum of about 550 Wm 2 to a minimum of 440 Wm 2 , a change of 20% (Berger et al., 1996a). 36.5.2 Sea Ice cover Sea ice component belongs to the fast response part of the climate system and is only slightly influenced by the ice sheets. Indeed, the simulated minimum area of

sea ice cover, which occurs during summer, is parallel to the 65 N insolation in June, and the simulated maximum area of sea ice cover has a tendency to remain high at a time when the insolation is already increasing (as it is the case between 139 and 134 kyr BP, for example). 36.5.3 Sea-surface temperature At the transition between isotopic stages 6 and 5, the warming simulated by the LLN 2-D NH model occurs roughly 4000 years before the reconstructed one by Cortijo et al. (1994), both in the 70–75 N – Norwegian Sea and in the 50–55 N – North Atlantic sectors, while the SST maximum occurs roughly at the same time in the data and in the model ( 127 kyr BP), at least in high latitudes. In the 70–75 N band, the problem is more complicated because sea ice can there cover the whole oceanic sector, e.g. from 140 to 132 kyr BP and from 120 to 105 kyr BP, periods during which the temperature of the sea at the surface is at the freezing point. That zone is partly free of sea ice only during summers from 132 to 120 kyr BP (i.e. over 12 kyr). Maximum SST (at 127 kyr BP) lags behind insolation maximum by 1000 years only; it occurs during a minimum of the sea ice extent and it coincides with a time when most of the ice sheets have disappeared over the northern continents. The zonal mean temperature at the surface in that zonal band (including both the oceans and the continents) is clearly reflecting the direct influence of insolation, of sea ice and of the ice sheets. Its minima are located between 139 and 134 kyr BP and around 111 kyr BP, 5000 years later than the insolation minima. Its maximum is reached at 127–126 kyr BP, almost in phase with insolation because, at that time, the ice sheets have disappeared and there is no sea ice in that latitudinal belt in late summer. Between 126 and 115 kyr BP, the simulated (by MoBidiC, Fig. 36.4) annual mean temperatures over the continents north of 60 N decrease by 5 C, and SSTs decrease

Interglacials as Simulated by the LLN 2-D NH and MoBidiC

°C

–116 –0.4 –0.6 –0.8 –1 SST –1.2

–118

–120

–122

–124

–126

–10 Continental temperature (Eurasia)

–12 °C –14

0.8 0.6 0.4 0.2

% of forest 0 –116 –118

–120

–122

–124

–126

Time (cal years BP)

Fig. 36.4 Simulated evolution of annual mean Atlantic ocean temperature North of 60 N (top), continental (Eurasia) surface temperature North of 60 N (centre), fraction of continent covered by forest North of 60 N (bottom).

by 0.8 C (Crucifix and Loutre, 2002). Both winter and summer temperatures decrease, but the largest changes occur in summer with up to 14 C over the continents and 2.6 C over the oceans. Most of the cooling occurs between 122 and 120 kyr BP, in reasonable agreement with those estimated from pollen data for southwestern Europe (Sa´nchez Gon˜i et al., 2000). Summer temperature of sea water at the surface, in both the 50–55 N and 70–75 N latitudinal bands, starts to decrease well before (actually 11 000 years before) the ice sheet starts to grow over the continents (as happened at 116 kyr BP). In the 50–55 N band, the simulated SST starts to rise again (first slowly, then more rapidly) when the ice sheet starts to form (116 kyr BP). In that 50–55 N zonal band, surface temperature change is therefore out of phase (by 180 degree) with the ice volume over the continents: cooling arises at the end of the melting phase of the ice sheets and warming occurs at the beginning of their growing phase. This is in agreement with the hypothesis of Ruddiman and McIntyre (1979), which claimed that the ice sheets were growing under warm SST conditions. An intensified Atlantic thermohaline circulation is simulated during the last inception phase in Wang and Mysak (2002), which further supports the hypothesis of Ruddiman and McIntyre (1979).

553

Although the timing of what corresponds to isotopic substages 5a, 5b and 5c is well reproduced, the model simulates a total melting of the Northern Hemisphere ice sheets between 126 and 117 kyr BP, 100 and 97 kyr BP and 83 and 74 kyr BP. Although this is not realistic [the Greenland ice sheet having survived at least over the last two to three glacial–interglacial cycles (Dansgaard et al., 1993)], it does not prevent the simulated ice sheets to grow again, leading to a 100-kyr quasi-cyclicity similar to the one seen in geological data. 36.5.4 Feedbacks at work A deeper analysis of these experiments also confirms the importance of the processes governing the response of the modelled climate system to insolation and/or CO2 changes. These are fundamentally related to the albedo and water vapour–temperature feedbacks (Berger et al., 1993b), to the taigatundra direct and indirect impacts on high latitudes surface albedo (Berger, 2001), to the altitude and continental effects on the precipitation over the ice sheets, to the lagging lithospheric response to the icesheet loading (Crucifix et al., 2001a) and to the mechanical destabilisation of the ice sheets through the rapid melting of their southern front as compared with the northern front during the deglaciation (Berger et al., 1992, 1993a). These mechanisms can be best documented between 130 and 110 kyr BP. When insolation decreases, surface temperature decreases, which delays the melting of snow fields at high latitudes. This cooling over the continents, in turn, causes a gradual southward shift of the boreal tree line (by 1400 km). This vegetation shift (Fig. 36.4) is a fundamental component of the response of the climate system to the orbital forcing during interglacial periods and a key component involved in the glacial inception (at about 118 kyr BP, according to Imbrie et al., 1984). The large amplitude of the response to vegetation is due to the

554

M.F. Loutre et al.

positive feedback loop that takes place between the tree line shift and surface temperature (Otterman et al., 1984; Harvey, 1988). Changes in vegetation distribution affect mainly the albedo of the snowcovered lands and impact the winter and spring temperatures, whereas the vegetation distribution is essentially sensitive to summer temperatures. The length of the snow-cover season is the factor that closes the amplifying feedback loop. Indeed, as a lower tree fraction cools down the surface in spring, thaw is delayed. Consequently, the snow-free season is shorter, which is unfavourable to tree growth. This mechanism is particularly effective in MoBidiC, as shown by the separation factor technique (Crucifix et al., 2002) and in an intercomparison between several EMICs (Brovkin et al., 2003). As a consequence, taiga is replaced by tundra, which increases the albedo of the vegetated surface covered by snow. Moreover, the transition from taiga-dominated to tundra-dominated vegetation cover takes place very quickly, between 120 and 118 kyr BP, although the astronomical forcing varies slowly. Both snow field and tundra are therefore leading to an increase of the surface albedo, creating a positive feedback. This is reinforced by the subsequent decrease of water vapour content in the atmosphere that results in a decrease in the downward infrared radiation at the surface. These positive feedbacks amplify strongly the original cooling. As a result, the Northern Hemisphere ice sheets start to develop and grow (Berger, 2001). Negative feedbacks will progressively develop such as the reduced snowfall at the top of the ice sheets because of the decreased surface temperature caused by the increasing altitude; in the interior of the continents, snowfall on top of the ice sheets decreases even more due to the continental climate effect. These processes progressively slow the build-up of the ice sheets up to the time they start to retreat. When insolation starts to increase (at 116 kyr BP, 7 kyr before the ice maximum is reached;

Berger et al., 1996a), the ice-sheet surface warms, which increases the ablation. This lowers the zonal surface albedo and induces a further replacement of tundra by taiga, which positively feedbacks directly on the albedo, and, indirectly, by reducing the albedo of the continental surfaces covered by snow. The resulting warming is further amplified through the water vapour feedback. As a consequence, the ice sheets start to melt, the melting being accelerated by the subsequent sea-level rise (Berger, 2001). 36.5.5 Precession and obliquity Three simulations were performed by MoBidiC to address the respective role of obliquity and precession in the transient behaviour of climate between 126 and 115 kyr BP (later called Eemian). In the standard simulation, the model is forced by the computed insolation (Berger, 1978) and a fixed CO2 concentration (261 ppmv). The transient response of the atmosphere, the ocean–sea–ice system, the vegetation and the snow cover is computed interactively. The topography of the Antarctic ice sheet is as today, while the Greenland ice sheet is reduced to half of its present-day size. In the simulation, named O, precession is kept fixed to its 126 kyr BP value and only obliquity varies. As a consequence, the annual mean insolation is varying at all latitudes. Simulation P corresponds to a varying precession, but a constant obliquity. In this case, the annual mean insolation is constant throughout the simulation, but this does not prevent changes in the insolation distribution during the year to occur. The annual mean temperature change simulated during the Eemian in the standard simulation is roughly the sum of the responses in the simulations P and O. However, the contribution of precession is the largest. This result implies that, although precession does not influence annual mean insolation, it controls most of the changes in annual mean temperature and it has a strong impact on summer insolation.

Interglacials as Simulated by the LLN 2-D NH and MoBidiC

Therefore, it causes significant variations of the snow cover and sea ice, which feedbacks on summer temperature, which is critical for the northern latitude vegetation. In turn, changes in vegetation cover lead to changes in the surface albedo mostly in spring and autumn. Thus, precession modifies the annual mean shortwave balance at the top of the atmosphere through its impact on the planetary albedo. On the contrary, changes in summer insolation due to obliquity are much smaller than those caused by precessional changes. Therefore, obliquity has almost no impact on vegetation changes and on the summer sea ice area. The impact of obliquity on the annual mean shortwave balance is then mainly achieved through the decline in annual mean insolation at the top of the atmosphere and is only marginally amplified by changes in the planetary albedo. This explains why the cooling in simulation O is gradual, while simulation P exhibits acceleration in the cooling trend between 122 and 120 kyr BP. Annual mean perennial sea ice gets thicker in the three experiments (standard, O and P). This is because the annual mean vertical heat flux at the sea ice surface actually decreases, even in simulation P, because the surface cooling reduces both sensible and latent heat fluxes towards the atmosphere as well as the infrared emission. Equilibrium therefore requires a thickening of the sea ice that reduces the annual mean conduction heat flux in the sea ice layer. The slight slowdown of the thermohaline circulation throughout the Eemian is due to the winter warming in intertropical areas, which leads to a density decrease of midlatitude subsurface water that reduces the North Atlantic overturning because it weakens the meridional density gradient. The ultimate causes of the warming are an increase in winter insolation and an increase in the winter atmospheric temperature, which reduces the sensible and latent heat flux from the ocean to the atmosphere. The effect of subsurface water warming in the

555

Atlantic is however partly compensated for by a salinity increase at high latitudes caused by a decrease in the freshwater balance in the Arctic. The simulated slight slowdown of the Atlantic overturning is thus mainly caused by precession, because mid-temperature insolation changes which is its ultimate cause, are mostly driven by precession at these latitudes (Crucifix and Loutre, 2002). As a conclusion, in MoBidiC, precession is the main contributor to climatic changes (continental and ocean cooling, decline in tree cover, thickening of the sea ice and weakening of the ocean circulation) simulated during the Eemian. 36.5.6 Validation of the simulation Qualitatively speaking, the results of our simulations are compatible with palaeoclimate reconstructions (Cheddadi et al., 1998; Cortijo et al., 1999; Heusser, 2000). Both LLN models simulate full interglacial conditions from about 125 to 117 kyr (with a slight difference of a few thousands years between the models), which corresponds with the warmest phase of the Eemian recorded in southern Iberia between 126 and 117 kyr. A first winter cooling is documented in northern Germany at around 123 kyr (Ku¨hl and Litt, 2003). By 121–120 kyr, deciduous forest persisted in Iberia as far north as 42 N (Sa´nchez Gon˜i et al., 2005), but coniferous forest began to expand in the northernmost regions (52 N). Seven pollen cores from France and Poland suggested that winters abruptly cooled by about 6 to 10 C around 4 to 5 kyr after the beginning of the Eemian. This is coherent with the rapid southward displacement of the tree line simulated by the model, starting from 120 kyr (Sa´nchez Gon˜i et al., 2005). The SST reconstructions by Cortijo et al. (1999) suggest a decrease by 2.5 C of August temperature for the northern North Atlantic between approximately 124 and 120 kyr BP (dates inferred from benthic 18 O data), whereas temperature at lower latitudes

556

M.F. Loutre et al.

remained more stable or even experienced a slight increase. This is in agreement with the SST cooling simulated in the North Atlantic. Starting from 119 kyr BP, the model simulates the re-growth of the ice sheets. Although the first slight decrease of temperatures in Iberia is detected, on the basis of the expansion of carpinus betulus, at around 121 kyr BP, quantitative climate reconstruction reveals that a substantial cooling, by 2–3 C, took place at around 118 kyr BP (Sanchez Go´ni et al., 2005). The maximum ice volume is displayed at the same time (109 kyr) in the model and in the data, as indicated by the benthic isotopic record of the Iberian margin deep cores (Shackleton et al., 2003). The agreement between model and data for precipitation is less good, probably because of the too simple structure of MoBidiC and too rough representation of the mechanisms leading to precipitation (Sa´nchez Gon˜i et al., 2005). In MoBidiC, the general circulation of the ocean does not experience any fundamental reorganisation throughout the Eemian, with the simulated circulation being similar to that observed today, a result well supported by reconstructions based on biogeochemical tracers (Boyle and Keigwin, 1982; Duplessy and Shackleton, 1985). 36.6 MIS 11 In addition to computed insolation, the LLN 2-D NH model was forced by several CO2 scenarios, either based on the Vostok CO2 concentration or based on multiple regression between CO2 and climate proxy from deep-sea cores (Berger et al., 1996b; Li et al., 1998; Loutre, 2003), to cover the last four climatic cycles, including MIS 11. The simulations are actually started at 575 kyr BP with no ice sheet in the Northern Hemisphere. This assumption was made because this epoch corresponds to interglacial periods. The atmospheric CO2 concentration from Li et al. (1998) is used up to stage 11 to allow the ice sheets to adjust to climate

variations. The simulated climate is only discussed for the last four climate cycles. Using the Vostok CO2 leads to an ice volume that remains lower than 5 106 km3 from 400 to 350 kyr BP, while with the regression-based scenarios, the simulated continental ice volume remains at its minimum for only 10 kyr. This feature of the Vostok experiment is similar to what might happen to our stage 1, especially if the future CO2 level is kept high over a sufficiently long period of time. Actually, in the Vostok experiment, amplitude in the insolation variation remains small, while atmospheric CO2 concentration remains high from 405 to 340 kyr BP. This prevents ice sheets from growing, leading to a very long interglacial at that time (Loutre, 2003). Also, a less pronounced MIS 10 glacial is generated with 22 106 km3 of ice to be compared with 31 106 km3 in the other experiments. This confirms the importance of (1) the previous state of the climate system in a transient experiment as underlined by Berger et al. (1998) and (2) the phase relationship between insolation and CO2 forcings on the length of the simulated stage 11 ice volume (Loutre, 2003). For example, if the Vostok CO2 is made older by 10 kyr, insolation and CO2 decrease in phase between 410 and 400 kyr BP, resulting in a much shorter interglacial lasting only about 10 kyr (Fig. 36.5). In order to test these results further, MoBidiC has recently been used to simulate MIS 11. For this simulation, the model starts from an equilibrium state, where orbital parameters are computed for 420 kyr BP and the CO2 concentration is fixed at 264 ppmv. The initial state of the ice sheet and of the ocean circulation is taken as for the last glacial maximum. The transient simulation is then forced by both the variation of the orbital parameters and the atmospheric CO2 concentration from 420 to 350 kyr BP. Although a full set of sensitivity experiments are not yet performed, some interesting results are already available. The simulation suggests that MIS 11 could be divided into two major subintervals. The

Interglacials as Simulated by the LLN 2-D NH and MoBidiC

557

(b)

(a) Insolation (Wm–2)

Insolation (Wm–2)

(1)

550

550

500

500

450

450

Shifted GT4 chronology

(2)

CO2 (ppmv)

CO2 (ppmv) 275

275

250

250

225

225

220

220

GT4 chronology 0

0

(3)

–1

10

–1 10

20

0

30

20

0

30 1

40

NH ice volume (106 km3)

1 40

SPECMAP δ O18 (‰) 50

–350

–375

–400

–425

–450

Time (kyr BP)

50

NH ice volume

(106 km 3) SPECMAP δO18 (‰)

–350

–375

–400

–425

–450

Time (kyr BP)

Fig. 36.5 Insolation (top panel) and CO2 (middle panel) forcings and Northern Hemisphere continental ice volume ( bottom panel) simulated by the LLN 2-D NH model over stages 12 to 10. Different CO2 scenarios are used during MIS 11. The reference simulation, indicated by the green line, uses the CO2 values measured in the Vostok core (Petit et al., 1999). In the left panel, CO2 values during MIS 5 (magenta) and MIS 9 (cyan) are also used. In the right panel, the MIS 11 CO2 series is made either older (dashed line) or younger (dash-dotted line) by 10 kyr. SPECMAP (Imbrie et al., 1984) curve is given for comparison.

first one, from 416 to 396 kyr BP, is characterised by the warmest temperature, both in the Northern Hemisphere (annual mean value over the interval of 14.9 C) and in the North Atlantic (annual mean value over the interval of 0.5 C). The end of this interval is characterised by a rapid southward shift of the tree line in northern latitudes and by important increase in the arctic sea

ice extent. During the second phase (396–362 kyr BP), climate is still warm although cooler than in the first phase (14.5 C for the Northern Hemisphere and 0.2 C for the North Atlantic). At the beginning of this period, northern latitude vegetation largely recovers and arctic sea ice retreats. At its end, climate is marked by a degradation characterised by sea ice

558

M.F. Loutre et al.

extension and northern tree line retreat. The whole period is characterised by small amplitude variability in temperature. These results seem to be in agreement with proxy records. Pollen in northwestern Iberia suggest a long, warm, period from 426 to 394 kyr BP (Desprat et al., 2005). SSTs imply a long warm stable period in northern latitudes (McManus et al., 2003). Isotopic data show that ice actually accumulated in northern high latitudes as soon as 400 kyr BP. Further, pollen data indicate a southward migration of vegetation belts involving albedo increase and glacial inception as early as 405 kyr BP. The model indicates a similar succession of events. The second part of MIS 11 in northwestern Iberia (394 to 362 kyr BP) is characterised by a highfrequency variability of about 2 to 3 C in the mean temperature of the coldest month (Desprat et al., 2005). Such a high variability can also be inferred from high-resolution long European pollen sequences (Velay from France, Reille et al., 2000 and Tenaghi Philippon from Greece, Wijmstra and Smit, 1976) and from biogenic silica content of Lake Baikal (Karabanov et al., 2003). 36.7 CONCLUSIONS Two different Earth system models of intermediate complexity (EMICs), i.e. LLN 2-D NH and MoBidiC, developed in Louvain-laNeuve, were used to study the climate of the last interglacials and the transition from interglacials to glacials. The critical role of two external forcings, i.e. astronomically driven insolation and atmospheric CO2 concentration, were discussed, as well as sensitivity studies on the role of the feedbacks. During the Holocene, vegetation and ocean had a counteracting influence on surface temperature, i.e. ocean reduces the summer cooling induced by insolation variations while vegetation change results in a further summer cooling. A similar conclusion can also be drawn for the Eemian interglacial.

Although annual insolation does not contain any precession component, climatic precession is the dominant orbital factor driving annual mean temperature. Indeed, summer insolation, mostly precessionally driven, induces vegetation changes, which feedback on temperature over the whole year. The modelling work confirms that insolation is the major driver of long-term climate change. When insolation changes are large, CO2 is only amplifying the temperature response to the insolation. However, the contribution of CO2 is becoming more important when insolation variations are small, as it is the case during MIS 11 and the present-day interglacials. Indeed, during these interglacials, an early decrease in CO2 concentration might lead to an early glacial inception while the interglacial can be very long (several tens of thousands of years) if CO2 concentration does not decrease soon enough.

REFERENCES Berger, A., 1978. Long-term variations of daily insolation and Quaternary climatic changes. Journal of Atmospheric Science 35(12), 2362–2367. Berger, A., 1980. The Milankovitch astronomical theory of paleoclimates: a modern review. In: Beer, A., Pounds, K., Beer, P. (Eds.), Vistas in Astronomy. Pergamon Press Ltd, Great Britain, pp. 103–122. Berger, A., 2001. The role of CO2, sea level and vegetation during the Milankovitch forced glacial– interglacial cycles. In: Bengtsson, L., Hammer, C.U. (Eds.), Geosphere–Biosphere Interactions and Climate. Cambridge University Press, Cambridge (UK), New York (USA), pp. 119–146. Berger, A., Loutre, M.F., 2002. An exceptionally long interglacial ahead? Science 297(5585), 1287–1288. Berger, A., Fichefet, T., Galle´e, H., Tricot, C., Marsiat, I., van Ypersele, J.-P., 1989. Astronomical forcing of the last glacial–interglacial cycle. In: Crutzen, P., Ge´rard, J.C., Zander, R. (Eds.), Our Changing Atmosphere. Universite´ de Lie`ge, Institut d’Astrophysique, Cointe-Ougre´e, pp. 353–382. Berger, A., Fichefet, T., Galle´e, H., Marsiat, I., Tricot, C., van Ypersele, J.-P., 1990a. Physical interactions within a coupled climate model over the last glacial interglacial cycle. Philosophical

Interglacials as Simulated by the LLN 2-D NH and MoBidiC Transactions of the Royal Society of Edinburgh: Earth Sciences 81(4), 357–369. Berger, A., Fichefet, T., Galle´e, H., Marsiat, I., Tricot, C., van Ypersele, J.-P., 1990b. Ice sheets and sealevel change as a response to climatic change at the astronomical time scale. In: Paepe, R., Fairbridge, R.W., Jelgersma, S. (Eds.), Greenhouse Effect, Sea Level and Drought. Kluwer Academic Publishers, Dordrecht, The Netherlands, pp. 85–107. Berger, A., Fichefet, T., Galle´e, H., Tricot, C., van Ypersele, J.P., 1992. Entering the glaciation with a 2–D coupled climate model. Quaternary Science Reviews 11(4), 481–493. Berger, A., Galle´e, H., Tricot, C., 1993a. Glaciation and deglaciation mechanisms in a coupled 2–D climate–ice sheet model. Journal of Glaciology 39(131), 45–49. Berger, A., Tricot, C., Galle´e, H., Loutre, M.F., 1993b. Water-Vapor, CO2 and insolation over the last glacial interglacial cycles. Philosophical Transactions of the Royal Society of London Series B-Biological Sciences 341(1297), 253–261. Berger, A., Tricot, C., Galle´e, H., Fichefet, T., Loutre, M.F., 1994. The last two glacial–interglacial cycles simulated by the LLN model. In: Duplessy, J.-C., Spyridakis, M.-T. (Eds.), Long Term Climatic Variations, Data and Modelling. Nato ASI series, Serie I: Global Environmental Change. Springer-Verlag, Berlin, pp. 411–452. Berger, A., Galle´e, H., Li, X. S., Dutrieux, A., Loutre, M.F., 1996a. Ice–sheet growth and high–latitudes sea–surface temperature. Climate Dynamics 12(7), 441–448. Berger, A., Loutre, M.F., Galle´e, H., 1998. Sensitivity of the LLN climate model to the astronomical and CO2 forcings over the last 200 ky. Climate Dynamics 14(9), 615–629. Berger, W.H., Bickert, T., Yasuda, M.K., Wefer, G., 1996b. Reconstruction of atmospheric CO2 from ice-core data and the deep-sea record of Ontong Java plateau: the Milankovitch chron. Geol. Rundsch. 85, 466–495. Boyle, E.A., Keigwin, L.D., 1982. Deep circulation in the North Atlantic over the last 200,000 years: geochemical evidence. Science 218, 784–787. Brovkin, V., Ganopolski, A., Svirezhev, Y., 1997. A continuous climate–vegetation classification for use in climate–biosphere studies. Ecological Modelling, 101(2–3), 251–261. Brovkin, V., Levis, S., Loutre, M.F., Crucifix, M., Claussen, M., Ganopolski, A., Kubatzki, C., Petoukhov, V., 2003. Stability analysis of the climate–vegetation system in the northern high latitudes. Climatic Change 57(1–2), 119–138. Cheddadi, R., Mamakowa, K., Guiot, J., de Beaulieu, J.-L., Reille, M., Andrieu, V., Granoszewski, W., Peyron, O., 1998. Was the climate of the Eemian

559

stable? A quantitative climate reconstruction from seven European pollen records. Palaeogeography Palaeoclimatology Palaeoecology 143(1–3), 73–85. Cortijo, E., Duplessy, J.C., Labeyrie, L., Leclaire, H., Duprat, J., van Weering, T.C.E., 1994. Eemian cooling in the Norwegian Sea and North Atlantic ocean preceding continental ice-sheet growth. Nature 372(6505), 446–449. Cortijo, E., Lehman, S., Keigwin, L., Chapman, M., Paillard, D., Labeyrie, L., 1999. Changes in meridional temperature and salinity gradients in the North Atlantic Ocean (30 degrees-72 degrees N) during the last interglacial period. Paleoceanography 14(1), 23–33. Crucifix, M., Berger, A., 2002. Simulation of ocean– ice sheet interactions during the last deglaciation. Paleoceanography 17(4), 1054 (doi: 10.1029/ 2001PA000702). Crucifix, M., Loutre, M.F., 2002. Transient simulations over the last interglacial period (126–115 kyr BP): feedback and forcing analysis. Climate Dynamics 19(5–6), 417–433. Crucifix, M., Loutre, M.F., Lambeck, K., Berger, A., 2001a. Effect of isostatic rebound on modelled ice volume variations during the last 200 kyr. Earth and Planetary Science Letters 184(3–4), 623–633. Crucifix, M., Tulkens, P., Berger, A., 2001b. Modelling abrupt events in glacial climate. In: Haupt, B.J., Maslin, M.A. (Eds.), The Oceans and Rapid Climate Changes: Past, Present and future. AGU, pp. 117–134. Crucifix, M., Loutre, M.F., Tulkens, P., Fichefet, T., Berger, A., 2002. Climate evolution during the Holocene: a study with an Earth system model of intermediate complexity. Climate Dynamics 19(1), 43–60. Crucifix, M., Loutre, M.F., Berger, A., 2005. Commentary on ‘‘The anthropogenic greenhouse era began thousands of years ago’’. Climatic Change 69(2–3), 419–426. Dansgaard, W., Johnsen, S.J., Clausen, H.B., Dahl-Jensen, D., Gundestrup, N.S., Hammer, C.U., Hvldborg, C.S., Steffensen, J.P., Sveinbjo¨rnsdottir, A.E., Jouzel, J., Bend, G., 1993. Evidence for general instability of past climate from a 250-kyr ice-core record. Nature 364, 218–220. Desprat, S., Sa´nchez Gon˜i, M.F., Turon, J.-L., McManus,J.F., Loutre, M.F., Duprat, J., Malaize´, B., Peyron, O., Peypouquet, J.-P., 2005. Is vegetation responsible for glacial inception during periods of muted insolation changes? Quaternary Science Reviews 24(12–13), 1361–1374. Duplessy, J.C., Shackleton, N.J., 1985. Response of global deep–water circulation to earth’s climatic change 135,000–107,000 years ago. Nature 316, 500–507. Galle´e, H., van Ypersele, J.-P., Fichefet, T., Tricot, C., Berger, A., 1991. Simulation of the last glacial cycle

560

M.F. Loutre et al.

by a coupled, sectorially averaged climate–ice sheet model. Part I: The climate model. Journal of Geophysical Research–Atmospheres 96, 13139–13161. Ganopolski, A., Kubatzki, C., Claussen, M., Brovkin, V., Petoukhov, V., 1998. The influence of vegetation–atmosphere–ocean interaction on climate during the mid-Holocene. Science 280(5371), 1916–1919. Harvey, L.D.D., 1988. On the role of high latitude ice, snow and vegetation feedbacks in the climatic response to exteral forcing changes. Climatic Change 13, 191–224. Heusser, L.E., 2000. Rapid oscillation in western North America vegetation and climate during oxygen isotope stage 5 inferred from pollen data from Santa Barbara Basin (Hole 893A). Palaeogeogr., Palaeoclimatol., Palaeoecol 161(3–4), 407–421. Hovine, S., Fichefet, T., 1994. A zonally averaged, three–basin ocean circulation model for climate studies. Climate Dynamics 10, 313–331. Huntley, B., Prentice, I.C., 1988. July temperature in Europe from pollen data, 6000 years before present. Science 241, 687–690. Imbrie, J., Hays, J., Martinson, D.G., McIntyre, A., Mix, A.C., Morley, J.J., Pisias, N.G., Prell, W.L., Shackleton, N.J., 1984. The orbital theory of Pleistocene climate : support for revised chronology of the Marine 18O record. In: Berger, A., Imbrie, J., Hays, J.D., Kukla, G., Saltzman, B. (Eds.), Milankovitch and Climate. Reidel, Dordrecht, pp. 269–305. Jouzel, J., Barkov, N.I., Barnola, J.M., Bender, M., Chappellaz, J., Genthon, C., Kotlyakov, V.M., Lorius, C., Petit, J.R., Raynaud, D., Raisbeck, G., Ritz, C., Sowers, T., Stievenard, M., Yiou, F., Yiou, P., 1993. Extending the Vostok ice-core record of paleoclimatic to the penultimate glacial period. Nature 364(6436), 407–412. Karabanov, E., Prokopenko, A., Williams, D., Khursevich, G., Kuzmin, M., Bezrukova, E., Gvozdkov, A., 2003. High-resolution MIS 11 record from the continental sedimentary archive of Lake Baikal, Siberia. In: Droxler, A.W., Poore, R.Z., Burckle, L.H. (Eds.), Earth’s climate and orbital eccentricity: the marine isotope stage 11 question. Geophysical Monograph. American Geophysical Union, Washington, DC, pp. 223–230. Kerwin, M.W., Overpeck, T., Webb, R.S., de Vernal, A., Rind, D., Healey, R., 1999. The role of oceanic forcing in mid-Holocene Northern Hemisphere climatic change. Paleoceanography 14(2), 200–210. Koc¸, N., Janssen, E., Haflidason, H., 1993. Paleoceanographic reconstructions of sea ocean conditions in the Greenland, Iceland and Norwegian Seas

through the last 14 ka based on diatoms. Quaternary Science Reviews 12, 115–140. Ku¨hl, N., Litt, T., 2003. Quantitative time series reconstruction of Eemian temperature at three European sites using pollen data Vegetation History and Archaeobotany 13, 205–214. Li, X.S., Berger, A., Loutre, M.F., 1998. CO2 and northern hemisphere ice volume variations over the middle and late Quaternary. Climate Dynamics 14(7–8), 537–544. Loutre, M.F., 2003. Clues from MIS 11 to predict the future climate – a modelling point of view. Earth and Planetary Science Letters 212(1–2), 213–224. Loutre, M.F., Berger, A., 2000. Future climatic changes: Are we entering an exceptionally long interglacial? Climatic Change 46(1–2), 61–90. McManus, J., Oppo, D., Cullen, J., Healey, S., 2003. Marine isotope stage 11 (MIS 11): analog for Holocene and future climate. In: A.W. Droxler, A.W., Poore, R.Z., Burckle, L.H. (Eds.), Earth’s climate and orbital eccentricity: the marine isotope stage 11 question. Geophysical Monograph. American Geophysical Union, Washington, DC, pp. 69–85. Milankovitch, M., 1941. Kanon der Erdbestrahlung und seine Anwendung auf das Eiszeitenproblem. Spec. pub. 132, Mathematical and natural Sciences, 33. Royal Serbian Sciences, Belgrade, 633 pp. (‘‘Canon of Insolation and the Ice Age Problem’’, English Translation by Israe¨l Program for the U.S. Department of Commerce and the National Science Foundation, Washington DC, 1969, and by Zavod za Udzbenike I nastavna Sredstva in cooperation with Muzej nauke I technike Srpske akademije nauka I umetnosti, Beograd, 1998). Otterman, J., Chou, M.-D., Arking, A., 1984. Effects of nontropical forest cover on climate. Journal of Applied Meteorology 23, 762–767. Peltier, W. R., Vettoretti, G., 2001. Glacial inception in the greenhouse: coupled A–O GCM simulation of the 5e–5d transition and the next. In: Chylek, P., Lesins L. (Eds.), First International Conference on Global Warming and the Next Ice Age, Dalhousie University, Halifax, Nova Scotia, Canada, pp. 256. Perez-Folgado, M., Sierro, F.J., Flores, J.A., Grimalt, J.O., Zahn, R., 2004. Paleoclimatic variations in foraminifer assemblages from the Alboran Sea (Western Mediterranean) during the last 150 ka in ODP Site 977. Marine Geology 212(1–4), 113–131. Petit, J.R., Jouzel, J., Raynaud, D., Barkov, N.I., Barnola, J. M., Basile, I., Bender, M., Chappellaz, J., Davis, M., Delaygue, G., Delmotte, M., Kotlyakov, V.M., Legrand, M., Lipenkov, V. ., Lorius, C., Pe´pin, L., Ritz, C., Saltzman, E., Stievenard, M., 1999. Climate and atmospheric history of the past

Interglacials as Simulated by the LLN 2-D NH and MoBidiC 420,000 years from the Vostok ice core, Antarctica. Nature 399(6735), 429–436. Reille, M., de Beaulieu, J.-L., Svobodova, V., Andrieu-Ponel, V., Goeury, C., 2000. Pollen analytical biostratigraphy of the last five climatic cycles from a long continental sequence from the Velay region (Massif Central, France). Journal of Quaternary Science 15(7), 665–685. Ruddiman, W.F., 2003. The anthropogenic greenhouse era began thousands of years ago. Climatic Change 61, 261–293. Ruddiman, W.F., 2005. Cold climate during the closest stage 11 analog to recent millennia. Quaternary Science Reviews 24(10–11), 1111–1121. Ruddiman, W.F., McIntyre, A., 1979. Warmth of the subpolar North Atlantic Ocean during Northern Hemisphere ice-sheet growth. Science 204(4389), 173–175. Sa´nchez Gon˜i, M.F., Eynaud, F., Turon, J.-L., Shackleton, N.J., Cayre, O., 2000. Direct land–sea correlation for the Eemian and its comparison with the Holocene: a high resolution palynological

561

record off the Iberian margin. Netherland Journal Geoscience 79, 345–354. Sa´nchez Gon˜i, M.F., Loutre, M.F., Crucifix, M., Peyron, O., Santos, L., Duprat, J., Malaize´, B., Turon, J.-L., Peypouquet, J.-P., 2005. Increasing vegetation and climate gradient in Western Europe over the Last Glacial Inception (122–110 ka): datamodel comparison. Earth and Planetary Science Letters 231: 111–130 (doi: 10.1016/j.epsl.2004. 12.010). Shackleton, N.J., Sa´nchez Gon˜i, M.F., Pailler, D., Lancelot, Y., 2003. Marine isotope substage 5e and the Eemian interglacial. Global and Planetary Change 757, 1–5. Wang, Z.M., Mysak, L.A., 2002. Simulation of the last glacial inception and rapid ice sheet growth in the McGill Paleoclimate Model. Geophysical Research Letters 29(23), 2102. Wijmstra, T.A., Smit, A., 1976. Palynology of the middle part (30–78 metres) of the 120 m deep section in Northern Greece (Macedonia). Acta Botanica Neerlandica 25, 297–312.

This page intentionally left blank

37. Vegetation–Climate Feedbacks in Transient Simulations Over the Last Interglacial (128 000–113 000 yr BP) M. Gro¨ger1, E. Maier-Reimer1, U. Mikolajewicz1, G. Schurgers1, M. Vizcaino1 and A. Winguth2 1

Max-Planck-Institut fu¨r Meteorologie, Bundesstraße 53, D-20146 Hamburg Center for Climatic Research, Department of Atmospheric and Oceanic Sciences, 1225 W, Dayton St Madison, WI 53706, USA

2

ABSTRACT The presently developed MPI/UW 3D – Earth system model for long-term integrations is applied to simulate the climate of the last interglacial. The model consists of an atmospheric and oceanic general circulation model, a dynamical terrestrial vegetation model and a marine carbon cycle model. The model was forced with time-varying insolation from 129 000 to 113 000 years before present (yr BP), revealing substantial feedbacks from the land biosphere on climate. These turned out to be important both for the simulated temperature in high northern latitudes and for the precipitation in the northern hemisphere monsoon belt. During the Eemian warm period, the simulated boreal forest extends in many places till the Arctic Ocean, and during the following cold period the transition between tundra and taiga migrated further south. Furthermore, associated albedo changes strongly amplify the simulated temperature changes. The intensified summer insolation during the Eemian leads to a higher precipitation over continents in the northern hemisphere. The strongest response is seen in the tropics and the African-Asian monsoon belt due to increased land–sea temperature contrasts. Vegetation is established in the Western Sahara desert. Compared to simulations with land vegetation prescribed at presentday pattern, the amount of precipitation in the Sahara is more than twice as large. The simulations show strong impacts on the

time-transient climate response by triggering nonlinear delays and accelerations seen in various atmospheric and oceanic temperature time series. The simulated storage of carbon in the terrestrial biosphere is relatively large. The carbon storage in land vegetation is increased by more than 10% during the Eemian compared to the following cold period. The associated changes in storage in soil and litter account for more than 100 GtC.

37.1 INTRODUCTION The last interglacial period provides a complex pattern of climatic changes that culminated in the growth of large-scale continental ice sheets over North America and Scandinavia, motivating comprehensive and multidisciplinary modelling strategies. During this period, the northern hemisphere summer insolation declined from values larger than today in the early Eemian to a minimum at around 115 000 yr BP (see Chapter 2, Berger et al., this volume). Model studies covering this period have applied either less comprehensive models (typically coupled ocean–atmosphere or atmosphereonly models) or Earth system models of intermediate complexity that are higher parameterized. During the Eemian warm period, the main features found in these studies is a strong warming north of 30 N (Harrison et al., 1995; Montoya et al., 1998, 2000; Kubatzki et al., 2000; Crucifix

564

M. Gro¨ger et al.

and Loutre, 2002) and an intensified summer monsoon in the northern hemisphere (Kubatzki et al., 2000; Montoya et al., 2000). By contrast, cooling especially in boreal regions was simulated in studies focusing on the late Eemian when large-scale continental ice sheets started to grow (Crucifix and Loutre, 2002; Meissner and Gerdes, 2002; Yoshimori et al., 2002; Meissner et al., 2003). However, the climate response to orbital forcing is further complicated by internal climate variability, nonlinear feedbacks and interactions between individual climate components that damp or amplify the response to orbital forcing. In particular, the glacial inception must have been heralded and paralleled by important feedback loops between individual climate components including changes in ocean circulation, vegetation, albedo and greenhouse gases. Recent work has already emphasized important feedbacks for the glacial inception. Albedo-temperature feedbacks north of 60 N were found to be driven by changes in snow coverage, sea–ice and vegetation (Crucifix and Loutre, 2002; Meissner et al., 2003). The role of oceanic circulation changes was highlighted by Meissner and Gerdes (2002). Another climate phenomenon highly sensitive to variations in insolation is the global monsoon system. Many studies found a strengthening of the monsoon precipitation in the northern hemisphere during periods when insolation was enhanced. This is suggested by a large number of model simulations (Claussen and Gayler, 1997; Levis et al., 2004) and corroborated by palaeoclimatic reconstructions (Larrasoana et al., 2003; Yuan et al., 2004). In this study, we will focus on the preconditions for glacial inception and the monsoon intensity in the northern hemisphere. Emphasis is laid on feedbacks arising from the terrestrial biosphere in time-transient and insolation-forced simulations. The terrestrial biosphere can influence climate via two major approaches. In a biophysical line

of evidence, vegetation alters the land surface, which affects the lower boundary condition for the atmosphere primarily by controlling the surface albedo and absorption of short-wave radiation. In addition to this, on a global scale, the terrestrial biosphere can act as source or sink for greenhouses gases in climate change scenarios. Although the biophysical and biochemical feedbacks operate contemporaneously in the real world, only few model studies did account for them simultaneously and in a full prognostic mode (Claussen et al., 2001). This is a drawback, since the two feedbacks can have opposite impact on climate giving rise to a complex, nonlinear pattern of climate response. Hence in this study, attention is focused on feedbacks associated with terrestrial vegetation and carbon cycle by using an Earth system model in which the dynamics of these feedbacks are explicitly described. 37.2 MODEL, EXPERIMENTAL SETUP The atmospheric component of the ESM used in this study consists of the ECHAM3 GCM with a horizontal resolution of 5:6 degrees and 19 vertical levels (Roeckner et al., 1992). The physical ocean component is an improved version of the Hamburg large-scale geostrophic ocean model (Maier-Reimer et al., 1993) including a thermodynamic sea–ice model. The oceanic component run on a 64X64 E grid (Arakawa and Lamb, 1977), resulting in a horizontal resolution of approximately 4 degrees and 22 vertical levels. Continental ice sheets are prescribed at present-day pattern. The terrestrial biosphere is the Lund-Potsdam-Jena (Sitch et al., 2003) dynamic global vegetation model (DGVM) running with the same horizontal resolution as used for the atmosphere. The LPJ models vegetation as 10 different plant functional types (PFT) that fractionally inhabit the grid space. The distribution of PFT is related to climatic boundary conditions for plant growth,

Climate Feedbacks in Transient Simulations

regeneration and plant-specific parameters that control plant competition for light and water. Carbon storage is calculated in living biomass, litter and soils. Climatic feedbacks from the terrestrial biosphere are provided to the atmosphere via monthly means of surface background albedo, vegetation cover, forest cover and roughness length. The model used here holds an interim position between higher resolution complex AOGCMs (used primarily for relatively short-term simulations (Chapter 33 Kaspar et al., Chapter 34, Widmann) and more economic models of intermediate complexity with less explicitly modelled processes (used for long-term integrations, Chapter 36 Loutre et al., Chapter 38 Kageyama et al., Chapter 39, Kubatzki et al.). Since the typical adjustment times of individual climate components are in the same order as their characteristic timescales, this allows to integrate the highly dynamic (and computational consumptive) model components only from time to time. Therefore, in order to save computational resources, a periodically synchronous coupling scheme is applied after Voss and Sausen (1996). Accordingly, the atmospheric component is switched off after integrating for 24-month periods. In the following asynchronous periods the length of which is determined interactively (10 years on average), the fluxes from previous synchronous periods are used to drive the other model components. A nonlinear energy balance model is used to damp SST and sea–ice anomalies during the asynchronous periods. For the vegetation model LPJ, an archive consisting of input parameters of the last four synchronous periods was used in order to generate some variability. This was necessary to avoid a model drift. The constant forcing applied in equilibrium simulations is appropriate for the study of internal climate feedbacks but neglects the specific adjustment times of different climate components. Kubatzki et al. (this volume) demonstrate in chapter 39 that time-slice experiments differ significantly

565

from transient runs simulating the glacial inception. Therefore, in this study a timetransient insolation forcing has been applied. All experiments presented here were started from restart files derived from a 10 000-year integration of the fully coupled model under present-day boundary conditions. In order to asses the biophysical impact from the terrestrial biosphere on climate, two sets of experiments were carried out. In one set, the land–surface conditions for the atmosphere and glacier mask were fixed according to present-day conditions (ECHAM3 land surface, without biophysical feedback). In the second set of experiments, the dynamic changes in vegetation simulated by the LPJ were used to update the surface boundary condition for the atmosphere (with biophysical feedback). Each set consists of a control run with present-day insolation and an Eemian run in which the model has been forced with time-varying insolation according to an orbital configuration corresponding to the period from 129 000 to 113 000 yr BP. The first thousand years are excluded from consideration, because of possible spinup effects. In the Eemian experiments, the forcing has been accelerated by a factor of eight. This tightens the carbon fluxes between the atmosphere and the sources of disturbance and thus, limits the interpretation of pCO2 to the detection of long-term trends rather than permitting quantitative examinations. In the following, a brief description of the forcing is given. In Section 3, internal climate feedbacks will be investigated by difference fields between selected time slices that represent the minimum and maximum climatic extremes. Emphasis is laid on climate feedbacks from dynamic vegetation changes. Finally, the time-transient evolution is analysed using time series from selected climate parameters. The daily values of incoming solar radiation at the top of the atmosphere have been calculated from the obliquity, eccentricity

M. Gro¨ger et al.

566

and the longitude of the perihelion using the algorithm of Berger and Loutre (1991) and a solar constant of 1365 W/m2. The results have been adapted to the 360 dayyear used in the model. The time-transient changes in the orbital parameters cause only minor changes in the yearly mean insolation. In the course of the Eemian, the insolation decreases only by about 3 W/m2. In the early Eemian, however, the seasonal contrast was amplified in the northern hemisphere. As a consequence, the northern hemisphere received considerably more insolation in the early part of the Eemian (Fig. 37.1a). Because of the increased (a) W/m2

540 510 480 450

11

SST

12

(b)

10 0.8

(c)

0.6 0.4 0.2

0

(d)

T

–15

–18

128

124

120

116

kyr BP

Fig. 37.1 Time series of (a) yearly maximum of daily mean insolation at 60 N and 60 S (dashed curve). Straight lines indicate present-day values. ( b) North Atlantic sea-surface temperatures north of 30 N. Dashed line indicates experiment without feedback. (c) vegetation cover [fraction] averaged over 60 to 90 N. (d) Near-surface temperatures overland points north of 60 N. Dashed lines indicate the simulations without feedback. Thin lines indicate control experiments. Displayed are 1000year averages.

obliquity, the insolation changes are more pronounced in the higher latitudes in the early parts of the Eemian. For a detailed description about the Eemian insolation pattern, we refer to Chapter 2 (Berger et al., this volume) 37.3 RESULTS To investigate internal feedbacks triggered by changes in insolation, atmospheric climatologies are averaged between 127 000125 000 yr BP and 116 000114 000 yr BP. These time intervals roughly correspond to the maximum and minimum extremes in the temperature response along the transient simulations. The difference in climatology between the two time slices (115 000–126 000 yr BP) therefore reflects the total response to insolation in the transient experiments. Accordingly, the following results are being interpreted as warming/cooling anomalies with reference to the 115 000 yr BP climate. Figure 37.2 displays the 2-m temperature difference between 115 000 and 126 000 yr BP. In both experiments, a strong cooling is simulated in the northern hemisphere which is more intense over continents than over oceans. The cooling over the ocean is strongest in the North Pacific with anomalies of more than 4 K in the western area. The weaker cooling in the North Atlantic is related to convective deep mixing in this area which transports warm surface waters to depth. In external forced warming (cooling) scenarios, this will decrease (increase) the net oceanic heat loss to the atmosphere. Accordingly, this convective damping leads to lower cooling anomalies in the North Atlantic compared to the North Pacific. Strongest cooling is observed over the Arctic area where the increase in sea–ice thickness (not shown) amplifies the cooling. Additionally the advance in sea–ice extent at 115 000 yr BP leads to thermal insulation of the ocean and thus decreases the oceanic heat loss to the atmosphere.

Climate Feedbacks in Transient Simulations (a)

567

126 k

Control

115 k

0.8

60N

0.7 30N

0.6 EQ

0.5 0.4

30S

0.3

60S

0.2 180

120W –7

–6

60W –5

–4

0 –3

–2

–1

60E 1

2

3

120E 4

5

180

90

180

270

360

0

90

180

270

360

6

Fig. 37.3 Seasonal cycle of simulated land surface albedo north of 60 N. Left (right) panel shows experiment with (without) feedback.

(b) 60N 30N EQ 30S 60S

180

0

120W –4

60W

–3.5 –3 –2.5 –2

0

60E

–1.5 –1 –0.5 0.5

1

120E 1.5

2

180

3

Fig. 37.2 (a) Yearly mean 2-m temperature difference between 126 and 115 kyr BP displayed for the experiment with feedback. (b) Difference between (a) and experiment without feedback.

Differences between the two Eemian experiments are predominantly found in boreal regions (Fig. 37.2b). In these regions, the temperature response is already large in the model with prescribed vegetation. This is the direct effect of expanded snow covered areas that translates into a higher surface albedo (not shown), leading to reduced absorption of short-wave radiation. In the fully coupled model, however, cooling anomalies exceed 4 K over wide areas in the northeast of Siberia and Alaska and even 6 K over eastern Canada and Scandinavia. The underlying principle for these strong local cooling patterns is the replacement of boreal forests by tundra during the late Eemian. The resulting decrease in vegetation cover has an additional impact on the surface albedo. Figure 37.3 illustrates the yearly cycles of land surface albedo averaged between 60 and 90 N for 126 and 115 000 yr BP. The

vegetation feedback is seen throughout the whole year but becomes largest during winter. This is because the influence of snow on surface albedo is weaker in forest-covered areas compared to grasslands that predominates the fixed ECHAM3 vegetation over most of the boreal regions. At 126 000 yr BP, the maximum winter (summer) land albedo of the model with nondynamic vegetation exceeds that of the fully coupled model by 0:11 ð 0:07Þ. This corresponds to a relative increase of about 21% in winter and 29% in summer. The replacement of boreal forest by tundra at 115 000 yr BP has the consequence that the maximum winter (summer) albedo of the fully coupled model exceeds that of the model with nondynamic vegetation by 0:09 ð 0:02Þ which corresponds to a lowering of 12% (5%). This vegetation-related albedo effect enhances (lowers) the temperatures at 126 000 yr BP (115 000 yr BP) in the experiment with included feedback which explains the stronger response seen in temperature (Fig. 37.2). Besides this, also the seasonal length with very low summer land albedo below 0.4 varies with the choice of the coupling setup. At 115 000 yr BP, this season is 24 days shorter in the simulation using interactive vegetation whereas at 126 000 yr BP it is 30 days longer. In agreement with the forcing (Fig. 37.1a), the southern hemisphere experiences slight warming ð< 1 KÞ at 115 000 yr BP. Only in the subtropics, a weak cooling is simulated

M. Gro¨ger et al.

due to a higher cloud cover and increased evaporative cooling. This is a consequence of an overall strengthening of the southern hemisphere monsoon system. Strong warming anomalies seen in the northern African tropics, over India, and Mexico (Fig. 37.2a) indicate a negative feedback simulated in both experiments. In these regions, a lower cloud cover during northern hemisphere summer decreases the surface radiation. Here, the largest reductions in precipitation are found, which also decreases evaporative cooling (Fig. 37.4a). Precipitation anomalies exceed 25 mm/month in Mexico, 75 mm/month in northwest Africa and 100 mm/month in southeast India in the experiment with interactive vegetation. These changes reflect a weakened monsoon as a consequence of lower land–sea temperature contrasts during the late Eemian. Again, the changes in precipitation are weaker when applying fixed vegetation (Fig. 37.4b). The difference is extreme in the Sahara desert where precipitation decreases during the monsoon season by about 170 mm/month in the full model, whereas with prescribed vegetation it is decreased by only 72 mm/month. This difference refers to a strong albedo feedback due to the vanishing vegetation in this region. Over wide areas, the darker temperate grasslands are replaced by lighter desert surfaces which increase the local surface albedo by up to 0.10. As a result, the area that benefits from monsoon precipitation is restricted to the south (Fig. 37.4). The time-transient evolution of yearly mean precipitation averaged over the Sahara desert (10 W–20 E; 10–30 N) is shown in Fig. 37.5. A maximum is reached between 126 000 and 125 000 yr BP in both model versions, but the changes simulated by the fully coupled model are more than twice as strong. In the experiment without feedback, a monotonous decrease follows indicating a more or less linear response to the forcing. By contrast, when the vegetation feedback is included, a rapid transition to drier conditions is seen between 125 500

(a) 60N 30N EQ 30S 60S

180

120W

60W

–90 –60 –30 –15 –10

0 –5

60E 5

10

5

10

15

120E 30

60

30

60

180

90

(b) 60N 30N EQ 30S 60S

180

120W

60W

–90 –60 –30 –15 –10

0 –5

60E 15

120E

180

90

Fig. 37.4 (a) Yearly mean precipitation difference between 126 and 115 kyr BP displayed for the experiment with feedback. (b) Difference between (a) and the experiment without feedback. 50

40

mm/month

568

30

20

10

0 128

124

120

116

kyr

Fig. 37.5 Time series of annual mean precipitation over the Sahara Desert [10 W–20 E; 10 –30 N]. Displayed are 1000-year averages. Dashed lines indicate the experiments with prescribed vegetation. Thin lines indicate the control experiments.

and 124 500 yr BP and from 123 500 to 120 500 yr BP followed by a slight decreasing trend. The steep decreases in precipitation are accompanied (with a little phase lag) by drastic reductions in vegetation

Climate Feedbacks in Transient Simulations

cover (not shown) which then accelerate the transition to drier conditions. From about 118 000 yr BP, the two curves are virtually indistinguishable from the two control experiments. Higher precipitation is seen in South America, South Africa and over Australia (Fig. 37.4a) which is consistent with the intensified southern hemisphere monsoon. An extraordinary increase is observed over the tropical western Atlantic, resulting from a stronger atmospheric convergence in this area as indicated in the velocity potential fields (not shown) during boreal summer.

569

60N 30N EQ 30S 60S

180

120W

60W

0

120W

60W

0

60E

120E

180

60E

120E

180

60N 30N EQ 30S

37.3.1 Changes in vegetation The changes in temperature and precipitation have a large impact on the terrestrial biosphere. The most prominent changes are seen in regions with strong climatic feedback. The entire boreal belt suffers under the strong cooling in the high northern latitudes and shows reductions in vegetation cover larger than 75%. These changes monitor mainly the southward retreat of boreal forests which (in the vicinity of the North Atlantic warm anomaly and a more northern position of the westerlies at 126 000 yr BP) formerly reached the Arctic Ocean in northern Europe and northwestern Asia (Fig. 37.6a). Changes of similar magnitude are found in the African tropics and south-east Asia. In Africa, the tropical forests and adjacent savanna retreat southward and temperate grasslands vanish in the western part of the Sahara desert. In South-east Asia, tropical forests retreat to the south and completely disappear in India (Fig. 37.6). In semiarid regions between 30 and 50 N that are not affected by changes in the monsoon system vegetation, cover is slightly increased. This refers mainly to the appearance of forests. These continental regions are characterized by strong cooling anomalies especially during the growing season but only moderate decreases in rainfall (< 5 mm=month over most of these areas).

60S

180 desert

savanna

trop.for.

temp.for.

temp.gr.

bor.for.

tundra

ice

Fig. 37.6 Simulated ecosystem distribution for (a) 126 kyr BP and ( b) 115 kyr BP.

Such conditions decrease the stress for water-demanding PFT, thus favouring rather humid ecosystems. Slight increases in vegetation cover are found also in the southern continents in areas that benefit from intensified monsoon precipitation (Fig. 37.4). The changes in the terrestrial biosphere and climate have only minor consequences for the terrestrial carbon storage which varies by only 107 GtC (Fig. 37.7) along the last interglacial. In case of the vegetation (litter) stock, an upper limit is reached in the course of the Eemian at roughly 1050 (640) GtC which is reduced by more than 10% ð 15%Þ at the end of the experiment. Intensified decomposition dominates the soil carbon in the early warmer part of the Eemian which reduces this stock and counteracts the increase in vegetation and litter. The total carbon storage which reflects the net effect of all processes thus reaches its maximum when the climate has cooled down moderately to improve soil carbon preservation but is still warm enough to maintain boreal forests in the high latitudes.

M. Gro¨ger et al.

570

Total biosphere

3900

Gt

3800 3700 3600

Gt

1100

Vegetation

1000 900

Gt

Soil 2200 2100

Gt

640

Litter

600 560 520 128 000

124 000

120 000

116 000

kyr

Fig. 37.7 Time series of carbon storage for the experiment with vegetation feedback.

37.4 TIME-TRANSIENT SENSITIVITY OF TEMPERATURE TO CHANGES IN INSOLATION Figure 37.1d shows the temperature response averaged over landpoints between 60 and 90 N for the two experiments in increments of thousand-year averages. In the experiment with interactive vegetation, the near-surface temperature falls by 4.8 K from its Eemian maximum to 19:5 C at the end of the experiment. In the experiment with fixed vegetation, the decrease is only 3.0 K. This corresponds to 60% of the 4.8 K decrease in the fully coupled model, while the two control experiments are statistically indistinguishable. The higher sensitivity of the fully coupled model mainly reflects the dynamic changes in vegetation cover and associated albedo changes described above. The difference between the two model versions fades towards 121 500 yr BP when the insolation becomes similar to that of the control run. They are virtually indistinguishable up to 117 500 yr BP which is relatively long considering the steeply decreasing insolation at that time (Fig. 37.1a). This reflects the fact that the

prescribed ECHAM3 vegetation has already an overall lower vegetation cover in boreal regions compared to the control run of the fully model. Because of the same reason, the difference between the two curves is slightly lower at the end of the experiments than during the Eemian, but nevertheless it is significant. The standard deviation of differences between the two control experiments shown in Fig. 37.1d (thin lines) is 0.26 K, while the maximum difference between the Eemian experiments amounts to 0.96 K at 113 500 yr BP which is more than the threefold standard deviation. The time-transient evolution of seasurface temperatures in the North Atlantic shows a significant amplification of the warming when using the fully coupled model (Fig. 37.1b). This is related to stronger atmospheric warming over land (Fig. 37.1d) which translates into the ocean mainly via the ocean–atmosphere heat fluxes. In case of the North Atlantic, the heat loss to the atmosphere is up to 0.1 PW lower in the experiment with dynamic vegetation during the maximum warming period. Another effect is seen in the shape of the following cooling period in the North Atlantic. When using prescribed vegetation, cooling starts at 125 500 yr BP proceeding to the end of the experiment indicating a nearly linear response to the forcing. In the fully coupled model, the beginning of the cooling period is clearly delayed. Again, the delay is linked to the evolution over land where anomalous high temperatures (compared to the model without feedback) prevail until 122 500 yr BP and go against the astronomical cooling signal. After 122 500 yr BP, vegetation is progressively diminished (1c) and the cooling accelerates over land.

37.5 DISCUSSION AND CONCLUSIONS The results demonstrate a substantial amplification of insolation-forced climate feedbacks when using interactive dynamical

Climate Feedbacks in Transient Simulations

vegetation instead of nondynamic prescribed vegetation. The overall rationale behind these feedbacks is changes in surface albedo due to land surface dynamics that further affect temperature and precipitation. Two main feedbacks that are amplified by vegetation dynamics have been demonstrated: 1) boreal vegetation–temperature feedback, and 2) low-latitude vegetation–monsoon feedback The simulated climate and vegetation pattern agree with results from other model studies (an overview can be found in Kabat et al., 2003) and are well supported by palaeoclimatic reconstructions. Pollen spectra and macrofossil assemblages from the Noatak Basin indicate extended distribution of boreal forests in Alaska during the last interglacial (Muhs et al., 2001; Edwards et al., 2003). Similar changes are preserved in deposits from the St Lawrence estuary in Quebec where a shift from boreal forests to tundra vegetation at the end of the last interglacial is recorded by pollen spectra (Clet and Occhietti, 1995). Likewise, the simulated intensification of the northern hemisphere monsoon system is in excellent agreement with proxy data. In particular, the greening of the western part of the Sahara desert during the Eemian is supported by reconstructions from pollen records (van Andel and Tzedakis, 1996). Jahns et al. (1998) provided evidence for drier conditions during glacial periods compared to interglacials in the African tropics. Jiang and Ding (2005) recently presented pollen records from the South China Sea that indicate humid climatic conditions to be more abundant in the area of the Chinese Loess Plateau in interglacial and interstadial periods during the last 130 kyr. Furthermore, the distribution of deserts and sandy land in China was greatly reduced during the last interglacial period, and the mobile dune area was about two-thirds of that of

571

today (Chen et al., 2004). Moreover, lighter interglacial planktonic 18 O values found in long-term records from core locations in the South China Sea point to a generally stronger East Asian summer monsoon during the Quaternary warm periods (Jun et al., 2004). In contrast, the simulated weaker Northern Australia monsoon seems to be in conflict with palaeoenvironmental data reported in Veeh et al. (2000), indicating generally wetter conditions in this region during interglacials. The northern hemisphere temperatures over land and even in the uppermost ocean layer showed a more linear response to insolation forcing when using prescribed vegetation instead of dynamic vegetation. With regard to the last glacial inception, the boreal temperature feedback at 115 000 yr BP together with changes in sea ice cover appears to be an essential precondition for the inception of continental ice sheet growth.

REFERENCES Arakawa, A., and Lamb, V.R., 1977. Computational design of the basic dynamical processes of the UCLA general circulation model. Methods Comput. Phys. 16, 173–283. Berger, A., and Loutre, M.F., 1991. Insolation values for the climate of the last 10 million years. Quat. Sci. Rev. 10, 297–317. Chen, H.Z., Su, Z.Z., Yang, P., and Dong, G.R. 2004. Preliminary reconstruction of the desert and sandy land distributions in China since the last interglacial period. Science in China Series D-Earth Sciences 47, 89–100. Claussen, M., and Gayler, V., 1997. The greening of the Sahara during the mid-Holocene: Results of an interactive atmosphere-biome model, Global Ecology and Biogeography Letters V6(N5), 369–377. Claussen, M., Brovkin, V., Petoukhov, V., and Ganopolski, A., 2001. Biogeophysical versus biogeochemical feedbacks of large-scale land-cover change. Geophys. Rev. Letters 26(6), 1011–1014. Clet, M., and Occhietti, S., 1995. Pollen content of Sangamonian interglacial deposits, Ile-auxCoudres, Middle St-Lawrence-estuary, Quebec. Ge´ographie Physique et Quaternaire 49(2), 291–304.

572

M. Gro¨ger et al.

Crucifix, M., and Loutre, M.F., 2002. Transient simulation over the last interglacial period (126–115kyr BP): feedback and forcing analysis. Clim. Dyn. 19, 417–433. Edwards, M.E., Hamilton, T.D., Elias, S.A., Bigelow, N.H., and Krumhardt, A.P., 2003. Interglacial extension of the boreal forest limit in the Noatak Valley, northwest Alaska: Evidence from an exhumed river-cut bluff and debris apron. Arctic Antarctic and Alpine Research 35(4), 460–468. Harrison, S.P., Kutzbach, J.E., Prentice, I.C., Behling, P.J., and Sykes, M.T., 1995. The response of northern hemisphere extratropical climate and vegetation to orbitally induced changes in insolation during the last interglaciation. Quat. Res. 43, 174–184. Jahns, S., Huls, M., and Sarnthein, M., 1998. Vegetation and climate history of west equatorial Africa based on marine pollen record off Liberia (site GIK 16776) covering the last 400,000 years. Rev. Paleobot. Palynol. 102, 277–288. Jiang, H., and Ding, Z., 2005. Temporal and spatial changes of vegetation cover on the Chinese Loess Plateau through the last glacial cycle: evidence from spore-pollen records. Review of Palaeobotany and Palynology 133, 23–37. Jun, T., Wang, P.X., and Cheng, X.R., 2004. Development of the East Asian monsoon and Northern Hemisphere glaciation: oxygen isotope records from the South China Sea. Quat. Sci. Rev. 23(18–19), 2007–2016. Kabat, P., Claussen, M., Dirmeyer, P.A., Gash, J.H.C., Bravo de Guenni, L., Meybeck, M., Pielke, R.S., Vo¨ro¨smarty, C.J., Hutjes, R.W.A., Lu¨tkemeier, S. (Eds.), 2003. Vegetation, Water, Humans and the Climate: A New Perspective on an Interactive System. Global Change – The IGBP Series, 566 pp, Springer Verlag, Berlin, Heidelberg. Kubatzki, C., Montoya, M., Rahmstorf, S., Ganopolski, A., and Claussen, M., 2000. Comparison of the last interglacial climate simulated by a coupled global model of intermediate complexity and an AOGCM. Clim. Dyn. 16, 799–814. Larrasoana, J.C., Roberts, A.P., Rohling, E.J., Winklhofer, M., and Wehausen, R., 2003. Three million years of monsoon variability over the northern Sahara. Clim. Dyn. 21(7–8), 689–698. Levis, S., Bonan, G.B., and Bonfils, C., 2004. Soil feedback drives the mid-Holocene North African monsoon northward in fully coupled CCSM2 simulations with a dynamic vegetation model. Clim. Dyn. 23, 791–802. Maier-Reimer, E., Mikolajewicz, U., and Hasselmann, K., 1993. Mean circulation of the Hamburg LSG OGCM and its sensitivity to the thermohaline

surface forcing. Journal of Physical Oceanography 23, 731–757. Meissner, K.J., and Gerdes, R., 2002. Coupled climate modelling of ocean circulation changes during ice age inception. Clim. Dyn. 18, 455–473. Meissner, K.J., Weaver, A.J., Matthews, H.D., and Cox, P.M., 2003. The role of land surface dynamics in glacial inception: a study with the UVic Earth system model. Clim. Dyn. 21, 515–537. Montoya, M., Crowley, T.J., and von Storch, H., 1998. Temperatures at the last interglacial simulated by a coupled ocean–atmosphere climate model. Paleoceanography 13(2), 170–177. Montoya, M., Crowley, T.J., and von Storch, H., 2000. Climate simulation for 125 kyr BP with a coupled ocean–atmosphere general circulation model. J. Clim. 13, 1057–1072. Muhs, D.R., Ager, T.A., and Beget, J.E., 2001. Vegetation and paleoclimate of the last interglacial period, central Alaska. Quat. Sci. Rev. 20(1–3), 41–61. Roeckner, E., Arpe, K., Bengtsson, L., Brinkop, S., Do¨menil, L., Esch, M., Kirk, E., Lunkeit, F., Poneter, M., Rockel, B., Sausen, R., Schlese, U., Schubert, S., and Windelbrand, M., 1992. Simulation of present day climate with the ECHAM model: impact of the model physics and resolution. Max Planck Institute for Meteorology, Hamburg, Rep. 93. Sitch, S., Smith, B., Prentice, I.C., Arneth, A., Bondeau, A., Cramer, W., Kaplan, J., Levis, S., Lucht, W., Sykes, M., Thonicke, K. and Venevski, S., 2003. Evaluation of ecosystem dynamics, plant geography and terrestrial carbon cycling in the LPJ dynamic vegetation model. Global Change Biology 9, 161–185. van Andel, T.H., and Tzedakis, P.C., 1996. Paleolithic landscapes of Europe and environments: 150 000– 25 000 years ago: an overview. Quaternary Sci. Rev. 15, 481–500. Voss, R., and Sausen, R., 1996. Techniques for asynchronous and periodically synchronous coupling of atmosphere and ocean models Part II impact of variability. Clim. Dyn. 12, 605–614. Veeh, H.H., McCorcle, D.C., and Heggie, D.T., 2000. Glacial/interglacial variations of sedimentation on the West Australian continental margin: constraints from excess 230Th. Marine Geology 166, 11–30. Yoshimori, M., Reader, M.C., Weaver, A.J., and McFarlane, N. A., 2002. On the causes of glacial inception at 116 kaBP. Clim. Dyn. 18, 383–402. Yuan, D.X., Cheng, H., Edwards, R.L., Dykoski, C.A., Kelly, M.J., Zhang, M.L., Qing, J.M., Lin, Y.S., Wang, Y.J., Wu, J.Y., Dorale, J.A., and An, Z.S., Cai, Y.J., 2004. Timing, duration, and transitions of the last interglacial Asian monsoon science 304, 575–578.

38. Mechanisms Leading to the Last Glacial Inception over North America: Results From the CLIMBER-GREMLINS Atmosphere– Ocean–Vegetation Northern Hemisphere Ice-Sheet Model Masa Kageyama1, Sylvie Charbit1, Catherine Ritz2, Myriam Khodri3 and Gilles Ramstein1 1

Laboratoire des Sciences du Climat et de l’Environnement/IPSL (UMRI572 CEACNRS-UVSQ), CE Saclay, L’Orme des Merisiers, Baˆtiment 701, 91191 Gif sur Yvette Cedex France 2 Laboratoire de Glaciologie et Ge´ophysique de l’Environnement, Saint-Martin-d’He`res, France 3 Geophysical Fluid Dynamics Laboratory, U.S. Department of Commerce, National Oceanic and Atmospheric Administration, Princeton, New Jersey, USA, Now at: UR PALEOTROPIQUE (IRD) and LSCE, 32 Avenue Henri Varagual, 93143 Bondy Cedex France

ABSTRACT The CLIMBER-GREMLINS model is an atmosphere–ocean–vegetation–northern hemisphere ice-sheet model able to simulate ice-sheet growth in response to the transient forcings (insolation and CO2 changes) of the period 126–106 kyr BP. In the present version of the model, this growth mainly occurs over North America and reaches an equivalent of 17 m in sea-level drop. To quantify the role of the vegetation, ocean and icesheet feedbacks in this glaciation of North America, we have conducted sensitivity experiments in which the feedback of each of these components is sequentially switched off. These experiments show that, in this model. (1) glacial inception does not occur when vegetation is fixed to its interglacial state (experiment testing the response of the atmosphere–ocean–boreal land–ice system), (2) glacial inception occurs faster than in the standard experiment when the ocean surface characteristics (surface temperature and sea ice extent) are prescribed to their interglacial seasonal cycle (experiment testing the sensitivity of the atmosphere–vegetation–boreal land–ice system), (3) the ice-sheet albedo and altitude feedbacks are not crucial for starting the

glaciation, but the albedo feedback doubles the ice volume growth rate, (4) the potential effect of a reduction in the thermohaline circulation is tested via an additional experiment in which we have forced it to its ‘off’ mode. The results show that this mode favours the fastest land ice growth of all our experiments.

38.1 INTRODUCTION The last glacial inception occurred at a time (ca. 115 000 years ago) when the northern high-latitude summer insolation was close to a minimum (Berger, 1978; Berger et al., this volume). The Milankovitch theory (Berger, 1988) relates weak summer insolation values (typically under a certain threshold) to cold summers favouring perennial snow at high latitudes and therefore, eventually, ice-sheet growth. The study of palaeoclimatic records has suggested that this relationship between summer insolation and northern hemisphere ice-sheet growth follows many pathways. For instance, on the one hand, the oceanic temperatures in the Greenland–Iceland–Norwegian seas significantly decrease at around 120 kyr BP, much

574

Masa Kageyama et al.

before the minimum in summer insolation (Cortijo et al., 1999). On the other hand, vegetation changes, consisting of the southward retreat of the boreal tree line, help amplify the snow albedo feedback which is somewhat inhibited in a forest environment (Sa´nchez Gon˜i et al., 2005). Numerical experiments have confirmed that the feedbacks from the vegetation and the ocean, in addition to the ice-sheet albedo feedback, play an important role in amplifying the initial insolation signal and creating the climatic conditions ideal for ice-sheet growth (see Vettoretti and Peltier (2004) for a review). However, at present, there are only few models in which the feedbacks from all three of these components can be compared within a single framework. Here we present the results from one of these models: CLIMBER-GREMLINS. This model simulates significant ice-sheet growth over North America in response to the insolation and CO2 changes occurring between 126 and 106 kyr BP (Kageyama et al., 2004). In the present paper, we present the results from three sensitivity experiments designed to evaluate the feedbacks from ice sheet, vegetation and ocean. The fourth sensitivity experiment was inspired by the evidence for abrupt climate change as early as 115 kyr BP (see for instance NGRIP Project Members, 2004). One of the possible explanations for these abrupt climate changes over Greenland is the sudden switch of the thermohaline circulation from an active mode (with the formation of deep water in the North Atlantic and large heat and mass transport to the North in the Atlantic Ocean) to a ‘sluggish’ or ‘off’ mode in which these nprocesses are nearly stopped. We have not attempted to mimic each of these events (see Kubatzki et al., in this volume, for another approach in the simulation of these events). Rather, we have forced the thermohaline circulation (THC) in its off mode (in the sense of Ganopolski and Rahmstorf 2001) to examine the maximum effect of an extremely slow THC on northern hemisphere ice-sheet growth.

38.2 MODEL DESCRIPTION The CLIMBER-GREMLINS model has been developed by coupling the atmosphere– ocean–vegetation model CLIMBER2.3 (Petoukhov et al., 2000) to the northern hemisphere ice-sheet model GREMLINS (GREnoble Model for Land Ice in the northern hemisphere: Ritz et al., 1997). Here we provide a brief summary of the model characteristics. The atmosphere in CLIMBER2.3 is represented on a coarse grid (seven sectors in longitude and 18 10 -wide latitudinal bands) by a statistical-dynamical model which includes a description of the atmospheric hydrological cycle (moisture transport by the mean circulation and transport by mid-latitude eddies which is parameterised). The ocean is described by three latitude–depth basins for the Atlantic, Indian and Pacific oceans. The vegetation is described by two vegetation types: grass and trees, and provides a simple representation of the modulation of the snow albedo by vegetation. The ice-sheet model GREMLINS is a threedimensional thermomechanical model for grounded ice, which is forced by temperature and mass balance fields at its surface. In the present version of the model, the ablation is computed from the mean annual and summer surface temperatures via a positivedegree-day method. Therefore, the forcing fields for GREMLINS are the annual and summer temperatures and annual snowfall. GREMLINS distinguishes three surface types: ice sheets, ice-free land and ocean. Grid points are attributed to the ocean surface type if they are not covered by ice sheets and if the altitude of the grid point is below sea level, which is given to the model as an input. In all the experiments presented here, the climate and ice-sheet models exchange information every 200 years. A downscaling procedure (described in Kageyama et al., 2004; Charbit et al., 2005) has been developed in order to transfer the results from the CLIMBER coarse atmosphere to the fine

Mechanisms Leading to the Last Glacial Inception

grid ð45 km 45 kmÞ of GREMLINS. This procedure is based on the calculation, in the atmosphere component of CLIMBER, of the surface radiative balance on 15 different levels, as if they were the grid box level. These levels are regularly spaced every 300 m between 0 and 4200 m. Thus, within each CLIMBER grid box, we obtain the surface temperatures for 15 different altitudes. The total precipitation vertical profiles are computed accounting for the free atmosphere temperature vertical profiles computed in CLIMBER, and snowfall is obtained on the 15 vertical levels as a fraction of the total precipitation which is itself a function of the temperature at these 15 different levels. Furthermore, these calculations are performed for each CLIMBER surface type (ocean, sea ice, ice sheets, forests, grassland and desert) and averaged over the three surface types of GREMLINS (ocean, ice-free land and ice sheets). The mean annual and summer temperatures, as well as the annual accumulation, are then computed on the GREMLINS grid by trilinear interpolation and serve as forcing for GREMLINS. In turn, GREMLINS computes the fraction of ocean, land ice and ice-free land as well as the altitude over the continents, which are given back to CLIMBER as new boundary conditions. Thus, the altitude of the ice sheet is taken into account in two different ways in the model. The effect of the altitude on the atmospheric dynamics is computed using the average altitude of each grid box in CLIMBER. The dependence of temperature and snowfall on altitude is taken into account via our downscaling procedure and the calculation of these fields on the 15 vertical levels defined above. In our sensitivity experiments, when we refer to the altitude feedback of the ice sheet on the climate, we have only isolated the first effect (effect of the altitude on the atmospheric dynamics), not the second one. In the present version, this model has some shortcomings which are currently being corrected. As far as the glacial

575

inception is concerned, it does not simulate a fast enough ice-sheet growth, probably partly because no inception occurs, even after 115 kyr BP, over Fennoscandia. However, ice growth over North America is substantial, which provides the opportunity to study the role of the feedbacks from the different components of the climate system during the glacial inception. 38.3 EXPERIMENTAL DESIGN In the standard experiment (STD), the model is simply forced by the evolving insolation from 126 to 106 kyr BP (Berger, 1978; Berger et al., this volume, Fig. 5) and the CO2 concentration measured in the Vostok ice core (Petit et al., 1999), which fluctuates between 260 and 280 ppm from 126 to 112 kyr BP and decreases from 260 to 230 ppm from 112 to 106 kyr BP. Furthermore, the northern ice-sheet model is forced by the time-dependent sea-level values from Bassinot et al. (1994). For simplicity of the experimental design, the initial ice-sheet and bedrock altitudes are set to their present values: the only ice sheet in the domain is Greenland. This is a simplification in the sense that the Greenland ice sheet is known to have been smaller than present during the Eemian. However, its extent and volume are not very well constrained (see for instance Cuffey and Marshall, 2000). In the experiments presented here, the ice-sheet model adjusts to the Eemian external forcings within the first thousand years of the experiment (Kageyama et al., 2004). The climatic (atmosphere–ocean–vegetation) initial state is in equilibrium with this ice-sheet distribution and 126 kyr BP insolation and CO2. In particular, the THC is its active mode, with a maximum overturning at around 20 Sv. From this STD experiment, four sensitivity experiments have been conducted, which are summarised in Table 38.1. In the first one (125VEG), the vegetation is fixed to its 125 kyr value, consisting of extensive forests over northern North America. The

576

Masa Kageyama et al.

Table 38.1 Summary of the numerical experiments used in the present study Experiment

Atmosphere

Ocean and sea ice

Vegetation

NH ice sheets

STD 125VEG

Interactive Interactive

Interactive Interactive

Interactive Interactive

122ICE

Interactive

Interactive

Interactive Fixed to 125 kyr BP value Interactive

125OCN

Interactive

THCoff

Interactive

Seasonal cycle of ocean and sea ice characteristics fixed to 125 kyr BP values Interactive, but with thermohaline circulation in off mode

Interactive

Fixed to 122 kyr BP value. Ice sheet model then used as a diagnostic Interactive

Interactive

Interactive

The atmosphere, ocean, sea ice and vegetation models are the components of the CLIMBER2.3 model. The northern hemisphere ice-sheet model is the GREMLINS (GREnoble Model for Land Ice in the Northern hemisphere) model.

atmosphere, ocean, sea ice and ice sheets are interactive. In the second one (122ICE), the ice-sheet feedback is inhibited from 122 kyr BP onwards: from this date, the ice-sheet extent and thickness and land– sea distribution computed by GREMLINS are not transferred to the CLIMBER model, which is forced, until the end of the run, by the 122 kyr BP values (i.e. minimal ice sheets) of these variables. The extent and volume of the Greenland ice sheet at this stage of the simulation are in broad agreement with reconstructions (see for instance Tarasov and Peltier, 2003). On the other hand, GREMLINS is still forced by the climatic output from CLIMBER in order to evaluate the strength of the ice-sheet feedback. All three of these experiments have been analysed by Kageyama et al. (2004). To investigate the role of the ocean, we have performed a third and fourth sensitivity run. The third sensitivity experiment (125OCN) mimics the approach we followed to study the role of the vegetation: the ocean and sea ice models are disconnected from the earth system model from 125 kyr BP onwards, and from this date the atmosphere–vegetation–ice-sheet model is forced by the 125 kyr BP sea-surface temperature and sea ice cover seasonal cycle. This experimental design has two main consequences: (1) The ocean and sea ice cannot react anymore to changes in incoming

insolation and do not play any role in the system equilibration to the evolving incoming insolation and changing CO2; (2) From 125 kyr BP onwards, the northern SSTs in summer are warmer and sea ice extent smaller than if they had been computed, with a maximum difference at 115 kyr BP. The difference in winter SSTs is not as large since winter insolation at high latitudes is very small anyway. The fourth sensitivity experiment (THCoff) lets the ocean and sea ice adjust to the evolving incoming insolation but with the constraint of the THC being in its ‘off’ mode. This state has been obtained by imposing a permanent 0.06 Sv freshwater flux to the north Atlantic ocean in the fashion of Ganopolski and Rahmstorf (2001)’s northern North Atlantic scenario (i.e. between 50 and 70 N). In the present case, we obtain an ‘off’ mode in which the THC is completely shut down, rather than a slow mode which can be obtained by perturbing a glacial state. In the palaeorecords, there is some evidence for THC instabilities as early as 110 kyr BP (Lehman et al., 2002), but these are short-lived compared to the duration of our experiments. However, they could have a important impact on ice-sheet growth by modifying the northern high-latitude climate and therefore the possibilities of ice sheet growth. The THCoff experiment is a simple experiment which is not intended to

Mechanisms Leading to the Last Glacial Inception

be realistic but is designed to test the sensitivity of ice-sheet growth to the state of the THC in a drastic (and exaggerated) scenario. 38.4 COMPARISON OF THE SIMULATED ICE-SHEET EVOLUTIONS The evolutions of the volume (upper panels) and area (lower panels) of land ice over North America in each experiment is shown in Fig. 38.1 as a function of time (left panels) and of June–July–August average insolation (right panels). In the STD experiment, inception occurs at 121 kyr BP, and the glaciated area (Fig. 38.1, lower left panel, black curve) increases steeply to reach 3.5 million km2 in 3.5 kyr. From 117.5 kyr BP, the icesheet area increases slowly and levels at the

577

value of 4.5 million km2 in the last two to three thousand years of the simulation. In terms of land–ice area, the inception is therefore fast and concomitant with the large decrease in insolation, which is at its minimum at 114:4 kyr BP. In terms of volume, the inception is more gradual but also reaches an apparent equilibrium during the last thousands of years of the simulation (Fig. 38.1, upper left panel, black curve). In fact, the ice-sheet extent and volume increase as long as the summer insolation decreases (Fig. 38.1, right panels, black curves), but do not decrease when the June-July-August insolation increases, from 114.4 kyr BP onwards. Rather, their growth rate drastically decreases and reaches zero by the last thousand years of the simulation. The importance of the ice-sheet feedback suggested by this result is further shown 20

18

18

16

16

Ice volume (1015 m3)

Ice volume (1015 m3)

North America 20

14 12 10 8 6 4

12 10 8 6 4 2

2

0

0 –125

–120

–115

–110

380 10

10

Ice area (106 km2)

8

Ice area (106 km2)

14

6

4

2

400

420

440

460

480

420

440

460

480

8

6

4

2

0

0 –125

–120

–115

Time (kyr BP)

–110

380

400

JJA insolation (W/m2)

Fig. 38.1 (top left) Evolution of the ice-sheet volume covering North America in all the simulations: STD (black), 125VEG (green), 125OCN (blue), 122ICE (cyan), THCoff (red); (bottom left) same for the ice-sheet area; (top right) ice-sheet volume as a function of June–July–August insolation (in W/m2). The simulations start at the r.h.s. of the graph with ice volume equal to zero. The minimum insolation is reached at 114 400 yr BP; (bottom right) same graph for the ice-sheet area.

578

Masa Kageyama et al.

by the 122ICE sensitivity experiment (Fig. 38.1, cyan curve). In this experiment, the atmosphere–ocean–vegetation model does not ‘know’ about the inception over North America. As a result, the glaciated area and volume, computed off-line, increase smoothly from 121 to 110 kyr BP and then react to the increase in summer insolation. Despite the increase of summer snow cover from the beginning of the run, with a summer snow fraction of 20 to 25% at 75 N, to 115 kyr BP, with a summer snow fraction of 55% in the same latitude band, the summer albedo increase is much smaller than in the STD run, in which summer snow fraction equals 100% at 75 N at 115 kyr BP. The snow albedo feedback alone is therefore not strong enough, in our model, to initiate a significant ice sheet, in terms of both volume and area of land–ice. In the 122ICE experiment, the ice-sheet extent, with a maximum of 3 million km2, is always smaller than the corresponding STD value. The ice-sheet feedback (and not only the snow albedo feedback) is therefore important in determining the ice-sheet growth rate, maximum ice-sheet area and volume, and as a consequence, the delay of the ice-sheet response to the evolving external forcings. The influence of vegetation in simulating a glacial inception with our model is crucial (experiment 125VEG, Fig. 38.1, green curves). When vegetation is prescribed to its interglacial value, the glaciated area only reaches 1 million km2, and the ice-sheet volume is negligible compared to the results of the other simulations. To disentangle the role of the vegetation from that of the ice sheets, we have conducted experiments equivalent to 125VEG and STD, but without any coupling to the ice-sheet model. In the latter experiment, the summer surface air temperature over northern North America is up to 2.4 C cooler than in the experiment with fixed vegetation, with a surface albedo up to 0.12 larger and a summer snow fraction up to 0.18 larger. Once coupled to the ice-sheet model, this temperature difference translates into a positive snow mass balance

and therefore inception over northern North America for the experiment with interactive vegetation. Without the transition from forest to grass over northern North America, the summer albedo is not high enough and the summer temperature does not decrease enough to allow for a positive snow mass balance over these regions and inception does not occur. The ice-sheet feedback further amplifies the difference between the experiments with interactive vegetation and with fixed vegetation. Although designed in a very different fashion, both experiments testing the role of the ocean result in a much larger glaciated area than in STD. In the case of 125OCN, the inception occurs 1.5 kyr earlier than in STD, and the continental ice area grows very fast until 120 kyr BP, reaching 4.5 million km2 at that time. It then increases smoothly to reach more than 6 million km2 by the end of the run. The ice area and volume in this simulation are therefore larger than in STD for the whole simulated period. This is unexpected: prescribing warm summer SSTs and less extensive sea ice in the mid- and high northern latitudes could have resulted in warmer temperatures over land as well. This is not what is observed in this experiment. As summer insolation decreases, summer temperature over land and over Arctic sea ice decreases more than in the STD experiment. This is also the case when we compare two equivalent experiments in which the ice-sheet model is not coupled to CLIMBER2.3, showing that this over-reaction of the land and sea ice surfaces to insolation changes when the surface ocean is prescribed compared to the interactive ocean case occurs primarily in the atmosphere–ocean–vegetation model. In the prescribed SST and sea ice area experiment, the atmosphere appears to compensate for the fact that the SSTs cannot adjust to the insolation changes, by being colder than in the STD experiment in summer and warmer in winter. A direct result of these colder land temperatures is decreased ablation and therefore faster ice-

Mechanisms Leading to the Last Glacial Inception

sheet growth. Also, more precipitation falls as snow, increasing the ice-sheet growth rate. Lastly, in summer, the equator-to-pole temperature gradient is reinforced in 125OCN, which results in slightly amplified heat and moisture transport to high latitudes. However, the corresponding total precipitation changes are rather small when compared to the annual snow fall increase. Therefore, it is the local temperature decrease over land and sea ice, in 125OCN, compared to STD that is primarily responsible for faster ice-sheet growth, and not accumulation changes due to increased moisture fluxes. In the experiment in which we constrain the THC in its ‘off’-mode, the ice-sheet growth is even faster than in the previous experiment: glaciation starts earlier, at 124 kyr BP, the fast growth phase ends at 121 kyr BP with an area of 5 million km2, followed by a slower growth phase until the end of the simulation. The growth rate during this second phase is constant and much larger than in any other simulation. The final area is more than 8 million km2. In this run, the SSTs are free to adjust to the evolving external forcings, with the constraint of the THC being in its off mode. The resulting climate over both land and ocean at high latitude is very cold, both in summer and in winter. The largest climate differences compared to STD are in the mid-latitudes (not shown), which corresponds to the reaction to the THC documented in Ganopolski and Rahmstorf (2001). This suggests that as soon as the THC slows down, the ice-sheet growth can be much faster because of a decrease in ablation, and an increased accumulation related to larger precipitation, of which a larger fraction falls as snow because of the lower temperatures. The increase in precipitation could itself be related to an increase in mid-latitude northward moisture fluxes, simulated both in summer and in winter in the 40 to 60 N latitude band. Thus, periods of slow THC, which have been documented for the last inception (Lehman et al., 2002), will therefore be included in more realistic experiments in the future.

579

The final distribution of glaciated areas in STD is shown on the top graph of Fig. 38.2. The southern tip of the Greenland ice sheet has melted as a result of the interglacial high summer insolation values and has not recovered by the end of the simulation. Inception has occurred over Alaska, the Canadian Rockies, the western Canadian plains and the Canadian Archipelago. This distribution is not accurate but could be a result of our rather crude horizontal interpolation of the CLIMBER results onto the GREMLINS grid. Here, the final distribution of the ice sheets in the different runs is given for the purpose of comparing the experiments. In all of them, inception occurs over the Canadian Rockies, because of the altitude effect on temperatures and relatively high precipitation there, which is related to our bilinear interpolation of precipitation fields onto the GREMLINS grid. For an inception to occur over the Canadian archipelago, both the ice-sheet and vegetation feedbacks are needed, as shown by the 125VEG and 122ICE experiments. The inception occurs separately over this area and on the western Canadian plains, where inception starts afterwards. In the case of 125OCN and THCoff runs, the final North American ice sheet extends from Hudson Bay to Alaska. The southern tip of Greenland is ice-covered, and so is eastern Siberia. In our model, as long as Hudson Bay is below sea level in the ice-sheet model, the computed mass balance over this area is based on CLIMBER’s results over the corresponding ocean (see section 2), which induce relatively high values of ablation over this area. Moreover, the ice-sheet model does not include any representation of ice shelves which could help getting a glacial inception over this region. The glaciation over Alaska and Siberia could be related to each other, in the sense that when an ice sheet develops over Alaska, the surrounding region gets colder, which results in an ice sheet growing over Siberia. Specific experiments exploring the climate–cryosphere dynamics over this region are currently being analysed.

580

Masa Kageyama et al.

STD

125VEG

125OCN

125ICE

THCoff

10

510

1010

1510

2010

2510

3010

3510

4010

Fig. 38.2 Ice-sheet thickness (in m) at the end of the STD and sensitivity experiments.

38.5 DISCUSSION AND PERSPECTIVES The CLIMBER-GREMLINS model, in its present version, is able to simulate significant land ice over North America as a result of the variations in incoming insolation and

CO2. Even if the location and size of the ice sheets are far from being perfect, such a model, which includes the atmosphere, the ocean, the vegetation and the northern hemisphere ice sheets, allows for a comparison, within a single framework, of the role of each of these components in the last glacial

Mechanisms Leading to the Last Glacial Inception

inception. Our sensitivity experiments show that both vegetation and ice-sheet feedbacks are essential to simulate significant ice-sheet growth. The role of the ocean is more ambiguous. The experiment in which SSTs and sea ice are prescribed to their 125 kyr BP seasonal cycle shows that glaciation can occur even when summer SSTs are warm. This, however, could be an artefact of the experimental design, which forces the model to adjust to the decreasing incoming insolation by decreasing the temperature over land, which is logical since ocean temperatures cannot be adjusted in this experiment. Whether the same behaviour occurs in other models, and in particular in general circulation models, remains to be investigated. The second experiment testing the role of the ocean shows that if SSTs are allowed to adjust, but with an ‘off’-mode THC, ice-sheet growth is even faster. This experiment, in which the THC has been switched off during the whole duration of the run, is unrealistic, but there is evidence of some periods of abrupt THC variations within the inception that would be worth being investigated. Current and future work on our model includes developing a new downscaling procedure to account for precipitation redistribution as a function of topography and land–sea mask. This appears to be essential to correct the locations of glacial inception and could be crucial to simulate a glaciation over Fennoscandia. In the present version, our model results are different from those obtained at PIK (Calov et al., 2005a, 2005b). A comparison between our ice-sheet models and coupling procedure would be very informative about the dominant processes leading to glacial inception.

ACKNOWLEDGEMENTS We thank the PIK Climber team for providing us their CLIMBER2.3 model and the organisers of the 2004 DEKLIM-EEM workshop for stimulating discussions. We are also grateful to Marie-France Loutre and

581

Michel Crucifix, whose reviews greatly helped in clarifying the initial manuscript. REFERENCES Bassinot, F. C., L. D. Labeyrie, E. Vincent, X. Quidelleur, N. J. Shackleton and Y. Lancelot, 1994. The astronomical theory of climate and the age of the BrunhesMatuyama magnetic reversal. Earth and Planetary Science Letters 126, 91–108. Berger, A., 1978. Long-term variations of daily insolation and quaternary climatic changes. Journal of the Atmospheric Sciences 35, 2362–2367. Berger, A., 1988. Milankovitch theory and climate. Reviews of Geophysics 26, 624–657. Calov, R., A. Ganopolski, M. Claussen, V. Petoukhov and R. Greve, 2005a. Transient simulation of the last glacial inception. Part I: glacial inception as a bifurcation in the climate system. Climate Dynamics 24, 545–561. Calov, R., A. Ganopolski, V. Petoukhov M. Claussen, V. Brovkin and C. Kubatzki, 2005b. Transient simulation of the last glacial inception. Part II: sensitivity and feedback analysis. Climate Dynamics 24, 563–576. Charbit, S., M. Kageyama, D. Roche, C. Ritz and G. Ramstein, 2005. Investigating the mechanisms leading to the deglaciation of past continental Northern Hemisphere ice sheets with the CLIMBER-GREMLINS coupled model. Global and Planetary Change 48, 253–273. Cortijo, E., S. Lehman, L. Keigwin, M. Chapman, D. Paillard and L. Labeyrie, 1999. Changes in meridional temperature and salinity gradients in the North Atlantic ocean (30 –72 N) during the last interglacial period. Paleoceanography 14, 23–33. Cuffey, K. M. and S. J. Marshall, 2000. Substantial contribution to sea level rise during the last interglacial from the Greenland ice sheet. Nature 403, 591–594. Ganopolski, A. and S. Rahmstorf, 2001. Rapid changes of a glacial climate simulated in a coupled climate model. Nature 409, 153–158. Kageyama, M., S. Charbit, C. Ritz, M. Khodri and G. Ramstein, 2004. Quantifying ice sheet feedbacks during the last glacial inception. Geophysical Research Letters 31, doi:10.1029/2004GL021339. Lehman, S. J., J. P. Sachs, A. M. Crotwell, L. D. Keigwin and E. A. Boyle, 2002. Relation of subtropical Atlantic temperature, high-latitude ice rafting, deep water formation, and European climate 130,000– 60,000 years ago, Quaternary Science Reviews 21, 1917–1924. North Greenland Ice Core Project Members, 2004. High–resolution record of Northern Hemisphere

582

Masa Kageyama et al.

climate extending into the last interglacial period. Nature 431, 147–151. Petit, J.-R., J. Jouzel, D. Raynaud, N. I. Barkov, J.-M. Barnola, I. Basile, M. Benders, J. Chappellaz, M. Davies, G. Delaygue, M. Delmotte, V. M. Kotlyakov, M. Legrand, V. Y. Lipenkov, C. Lorius, L. Pe´pin, C. Ritz, E. Saltzman and M. Stievenard, 1999. Climate and atmospheric history of the past 420,000 years from the Vostok ice core, Antarctica. Nature 399, 429–436. Petoukhov, V., A. Ganopolski, V. Brovkin, M. Claussen, A. Eliseev, C. Kubatzki and S. Rahmstorf, CLIMBER-2: a climate system model of intermediate complexity. Part I: model description and performance for present climate, 2000. Climate Dynamics 16, 1–17. Ritz, C., A. Fabre and A. Letre´guilly, 1997. Sensitivity of a Greenland ice sheet model to ice flow and

ablation parameters: consequences for the evolution through the last climatic cycle. Climate Dynamics 13, 11–24. Sa´nchez-Gon˜i, M. F., M.-F. Loutre, M. Crucifix, O. Peyron, L. Santos, J. Duprat, B. Malaize´, J.-L. Turon and J.-P. Peypouquet, 2005. Increasing vegetation and climate gradient in Western Europe over the Last Glacial Inception (122–110 ka): datamodel comparison. Earth and Planetary Science Letters 231, 111–130. Tarasov, L. and W. R. Peltier, 2003. Greenland glacial history, borehole constraints and Eemian extent. Journal of Geophysical Research 108(B3), 2143, doi:10.1029/2001JB001731. Vettoretti, G. and W. R. Peltier, 2004. Sensitivity of glacial inception to orbital and greenhouse gas climate forcing. Quaternary Science Reviews 23, 499–529.

39. Modelling the End of an Interglacial (MIS 1, 5, 7, 9, 11) Claudia Kubatzki1, Martin Claussen2, Reinhard Calov2 and Andrey Ganopolski2 1

Alfred Wegener Institute for Polar and Marine Research, Bussestr. 24, 27570 Bremerhaven, Germany 2 Meteorological Institute, University of Hamburg, and Max Planck Institute for Meteorology, Bundesstr. 53, 20146 Hamburg, Germany

ABSTRACT With the CLIMBER-2 model, the last glacial inception is simulated as a rapid ice-sheet expansion over northern North America, with only little ice sheets in Scandinavia. We present sensitivity studies that more deeply investigate the climatic feedbacks at the end of an interglacial. (i) An ice-free Greenland during the Eemian appears as a second stable state in the model. The initial size of the Greenland ice sheet, however, has only little effect on the subsequent glacial inception. (ii) North American glaciation can be reduced or even suppressed using preindustrial or Eemian vegetation and/or ocean surface conditions. (iii) Timing and amplitude of the last glacial inception cannot be estimated by the use of time-slice simulations. (iv) Changes in precession (perihelion) are crucial for the ice-sheet growth, obliquity and CO2 merely act as amplifiers. (v) Cold events at the end of the interglacial can be reproduced by the introduction of freshwater disturbances into the North Atlantic. (vi) The model is able to simulate earlier glacial inceptions as well, although difficulties exist with respect to the strength of the glaciation. Future glacial inception in the model happens only in about 50 000 years from now. 39.1 INTRODUCTION A number of climate model experiments exist that attempt to simulate the climatic changes at the last glacial inception, in

particular GCM time-slice experiments of about 115 kyr BP. These simulations typically investigate individual processes within the climate system that led to or supported the development of ice sheets at the end of the Eemian interglacial. Here, we attempt to perform systematic and consistent model analyses of the environmental and climatic changes at that time. The Earth system model of intermediate complexity CLIMBER-2 is used to simulate the climatic changes at the end of different interglacials, with a focus on the end of the Eemian (or the end of marine isotopic stage MIS 5e). Our simulations extend results obtained in earlier simulations of the last glacial inception and performed by different authors, in particular those described in Calov et al. (2005a, 2005b). The aim of our project was threefold, (i) deepening of the understanding of processes and mechanisms that led to the last glacial inception, (ii) improvement of the agreement of the model results with the geological record, and (iii) application of the model to other interglacials and prognosis for the natural, anthropogenically undisturbed end of the current interglacial, the Holocene. CLIMBER-2 couples dynamic models of atmosphere, ocean (including sea ice), terrestrial vegetation (in terms of potential fractions of trees, grass and bare soil) and inland ice. The models for atmosphere and land surface have a spatial resolution of 10 in latitude and about 51 in longitude. The ocean resolves three ocean basins with 2.5 latitudinal resolution and 20 vertical unevenly spaced levels, and the northern

Claudia Kubatzki et al.

39.2 RAPID ICE-SHEET EXPANSION AT THE LAST GLACIAL INCEPTION Sea-level reconstructions from different sources, like corals or oxygen isotopes (e.g. Chappell et al., 1996; Waelbroeck et al., 2002; Siddall et al., 2003), indicate an increasing global ice volume between about 120 and 115 kyr BP. At that time, large ice sheets developed, especially over North America (e.g. Clark et al., 1993). We simulated the end of the last interglacial and the last glacial inception, in a transient model run. For this purpose, we forced the model by prescribing changes in insolation and atmospheric CO2 concentration over the time period from 128 000 to 100 000 years before present (128 to 100 kyr BP). The model was driven by changes in the orbital parameters following Berger (1978). From this, the latitudinally and seasonally varying insolation was computed. Work in progress (not shown here) suggests that the mass balance of inland ice reveals the highest correlation with the maximum summer insolation at 65 N. Maximum insolation at that latitude declines from the Eemian onwards and

reaches its minimum at around 116 kyr BP (see chapters 2/3). From that time on, it increases towards 105 kyr BP and then again decreases. The atmospheric CO2 variations were prescribed according to Petit et al. (1999), see chapter 6. In our fully coupled transient run (Kubatzki et al., 2006a, 2006b), glacial inception in terms of a rapid ice-sheet expansion takes place at about 117.5 kyr BP (see Fig. 39.1a). Inland ice develops mainly over northern North America. Glaciation also takes place in Scandinavia but to a much smaller extent (see Fig. 39.2). Whereas the Laurentide ice sheet remains after 116 kyr BP, the Fennoscandian ice sheet disappears

Area (106 km2)

(a) 14 12 10 8 6 4 2 0

(b) Area (106 km2)

hemisphere inland ice model is run with a resolution of 0.75 in latitude and 1.50 in longitude. A detailed model description can be found in Petoukhov et al. (2000), model validation and sensitivity tests in Ganopolski et al. (2001), the terrestrial vegetation model is described in Brovkin et al. (2002), the inland ice model in Greve (1997) and its coupling to the climate model in Calov et al. (2005a). In the simulations presented in the following, the model is driven only by changes in insolation (in terms of variations of the Earth’s orbital parameters) and atmospheric CO2 concentration. The model simulates a large number of variables describing the state of the climate, ocean, vegetation and inland ice which then can be compared with the geological record (see also general description of climate system models in chapter 32).

0.6 0.5 0.4 0.3 0.2 0.1 0 –128

–124

–120

–116

–112

–108

–104

–100

Time (kyr BP)

Fig. 39.1 (a) Northern hemispheric ice area (106 km2) and (b) area of the Fennoscandian ice sheet (106 km2) in the transient simulation of the fully coupled model over the Eemian and last glacial inception.

Latitude

584

80N

4000

70N

3000 2000

60N

1500

50N

1000 40N

500

30N 180

50 120W

60W

0

60E

120E

180

(m)

Longitude

Fig. 39.2 Thickness of inland ice (m) for the transient simulation of the last glacial inception at 113 kyr BP.

Sea level (m)

Modelling the End of an Interglacial 10 0 –10 –20 –30 –40 –128

–124

–120

–116

–112

–108

–104

–100

Time (kyr BP)

Fig. 39.3 Sea-level difference as compared with the preindustrial climate (m) in the transient simulation of the last glacial inception (solid) and in a transient simulation assuming no inland ice in the northern hemisphere at the beginning of the run (dashed).

after the increasing insolation went beyond a certain level (see Fig. 39.1b). The global annual cooling in the course of our model run amounts to roughly 1.6 C. The sea-level change corresponding to the developing ice sheets amounts to about 40 m as maximum in our simulation (see Fig. 39.3). Timing and amount compare reasonably with the reconstructions (e.g. Chappell et al., 1996; Waelbroeck et al., 2002; Siddall et al., 2003). The last glacial inception as found in CLIMBER-2 can be described as a bifurcation transition in the climate system, its rapidity is caused by the strong positive snow albedo feedback (see also Calov et al., 2005a, 2005b). 39.3 AN ICE-FREE GREENLAND AT THE EEMIAN AS A SECOND STABLE STATE Berger et al. (1998) found in their simulations that the initial size of the Greenland ice sheet used in a model run was relevant for the simulation of the last glacial– interglacial cycle. In their work, the size of northern hemispheric ice sheets assumed for the Eemian determined the amount of inland ice at 100 kyr BP. We started our transient simulation in the highest boreal summer insolation phase of the Eemian interglacial. Although according to the data, agreement exists that the Greenland ice sheet during the warmest part of the Eemian was smaller than

585

at present (e.g. North Greenland Ice Core Project, NGRIP Members, 2004), it is not clear to what extent the Greenland and the West Antarctic ice sheet, respectively, contributed to the observed relative sea-level rise (e.g. Cuffey and Marshall, 2000). We varied the initial size of the Greenland ice sheet and ran time slice as well as transient simulations (Kubatzki et al., 2006a). First, we performed two different time-slice simulations for the insolation and CO2 forcing at 128 BP. The one simulation was started with the simulated preindustrial size of the Greenland ice sheet; the other one was started with no northern hemispheric ice sheets. In both simulations, the model was run for 200 kyr. At that time, climate and ice sheets had reached equilibrium with the forcing. When we start the time-slice simulation for 128 kyr BP with a preindustrial Greenland ice sheet, the ice sheet at the end of the model run is about 1:3 106 km3 smaller in volume than the Greenland ice sheet simulated for the preindustrial climate. The ice thickness is reduced mainly at the ice-sheet margins and in the southern parts of Greenland. This compares reasonably well with the estimates of the NGRIP Members (2004). Likewise, the simulated temperature anomalies for that region are consistent with the NGRIP reconstruction. When we start our second time-slice simulation for 128 kyr BP with no inland ice in the northern hemisphere, the size of the Greenland ice sheet amounts to only about 0:1 106 km2 even after 200 kyr of model integration time. Thus, CLIMBER-2 results suggest that two different equilibrium states were possible for the insolation and CO2 forcing of 128 kyr BP. Results of both time-slice experiments were then used as initial conditions for the transient simulations. Starting the transient run from the time-slice simulation with the reduced preindustrial Greenland ice sheet at 128 kyr BP gives the transient run we discussed in the previous section. Starting the transient run from the extreme

586

Claudia Kubatzki et al.

assumption of no inland ice in the northern hemisphere during the Eemian affects the simulation of the last glacial inception only marginally, as can be seen in Fig. 39.3. In conclusion, the size of the Greenland ice sheet at the Eemian has nearly no impact on the course of the last glacial inception. 39.4 SUPPRESSING OCEAN AND VEGETATION CHANGES CAN SUPPRESS GLACIAL INCEPTION Several model studies found that when ocean or vegetation were allowed to react on the climatic changes at the end of the last interglacial, the related (albedo) feedbacks amplified the cooling trend (e.g. Dong and Valdes, 1995; De Noblet et al., 1996b; Gallimore and Kutzbach, 1996; Khodri et al., 2001; Crucifix and Loutre, 2002; Wang and Mysak, 2002; Yoshimori et al., 2002; Meissner et al., 2003; Kageyama et al., 2004; Vettoretti and Peltier, 2004). In many of these and other (e.g. De Noblet et al., 1996a; Dong et al., 1996; Lorenz et al., 1996) simulations of the last glacial inception, the ocean and/or vegetation states were kept fixed. Various assumptions were made on the ocean state and vegetation cover. The surface was prescribed to its present day either observed or simulated state, to its preindustrial or to its Eemian state.

With CLIMBER-2, we tested the climate impact of a changing ocean and/or vegetation and the related synergistic effects (Kubatzki et al., 2006a). We repeated our transient simulation of the last glacial inception but alternatively prescribed the ocean surface (in terms of sea-surface temperatures (SSTs), sea ice fraction and thickness), vegetation (in terms of fractions of trees, grass and bare soil and leaf-area indexes for trees and grass) or both to their simulated either preindustrial or Eemian (128 kyr BP) state. On global annual average, the two states differ in that the Eemian climate was about 0.6 C warmer, and annual sea ice cover was diminished by about 2:6 106 km2 in the northern hemisphere and the boreal tree line was shifted to the north. The results of these sensitivity studies can be seen in Fig. 39.4. From Fig. 39.4a, it becomes clear that keeping the vegetation and/or the ocean surface fixed to its preindustrial state weakens the last glacial inception, i.e. glacial inception still takes place but the inland ice cover is strongly reduced as compared with the fully coupled model run. The main reason is that the albedo feedback (and synergism) resulting from the expansion of either sea ice, tundra or both and amplifying the cooling trend is suppressed when prescribing alternatively the ocean surface, vegetation or both. Figure 39.4b shows that when we keep the Earth’s

(a) Preindustrial b.c.

(b) Eem b.c. 12

Area (106 km2)

Area (106 km2)

12 10 8 6 4 2

10 8 6 4 2

–128 –124 –120 –116 –112 –108 –104 –100

Time (kyr BP)

–128 –124 –120 –116 –112 –108 –104 –100

Time (kyr BP)

Fig. 39.4 Northern hemispheric inland ice area (106 km2) over the Eemian and last glacial inception in the transient run and when fixing the Earth’s surface to its (a) preindustrial and (b) Eemian state, respectively; ocean and vegetation interactive (black), ocean interactive but vegetation fixed (blue), vegetation interactive but ocean fixed (green), fixed ocean and vegetation (red).

Modelling the End of an Interglacial

surface fixed to its Eemian state (reflecting a much warmer than preindustrial climate), glacial inception can be even suppressed in our model. Accordingly, the simulated climatic changes differ strongly between the different model realisations, and also the size of the Greenland ice sheet is affected. The relative role of ocean and vegetation changes depends also on the atmospheric CO2 reconstruction applied. The earlier the CO2 concentration drops (for example, in Barnola et al., 1987 as compared with Petit et al., 1999, see chapter 6) the larger is the impact of a changing ocean on the simulation of glacial inception (Kubatzki et al., 2006a). 39.5 TIME-SLICE SIMULATIONS CANNOT REPRODUCE A TRANSIENT SIMULATION OF THE LAST GLACIAL INCEPTION A number of simulations of the last glacial inception were time-slice simulations of about 115 kyr BP with fixed modern inland ice (e.g. Dong and Valdes, 1995; De Noblet et al., 1996a, 1996b; Dong et al., 1996; Gallimore and Kutzbach, 1996; Lorenz et al., 1996; Khodri et al., 2001; Yoshimori et al., 2002; Meissner et al., 2003; Vettoretti and Peltier, 2004). Some transient simulations with or without interactive ice-sheet model exist as well (DeBlonde and Peltier, 1991; Galle´e et al., 1992; Marsiat, 1994; Tarasov and Peltier, 1997; Berger et al., 1998; Crucifix and Loutre, 2002; Wang and Mysak, 2002; Kageyama et al., 2004). In the time-slice experiments, different atmospheric CO2 levels were assumed, ranging from 280 to 240 ppmv. Many of these simulations underestimated the extent of perennial snow cover (as representative of ice-sheet cover), especially in North America, as compared with the ice-sheet reconstructions and with that the related albedo feedback. The application of time-slice simulations implies that the climate at that time was in equilibrium with the forcing. Even though such

587

an assumption seems reasonable for long, stable interglacial periods, it can hardly be assumed for the highly transient climate during the last glacial inception. With CLIMBER-2, we tested to what extent time-slice simulations can reproduce the results of our transient model run (Kubatzki et al., 2006b). For this purpose, we performed two time-slice simulations for 115 kyr BP assuming an atmospheric CO2 level of 271 ppmv corresponding to 115 kyr BP (Petit et al., 1999). One simulation assumed a fixed preindustrial inland ice cover and in the other one, the inland ice model was interactively coupled. Figure 39.5a shows the temperature change in our transient simulation at 115 kyr BP as compared with the preindustrial climate. Figure 39.5b and 5c reveal that the temperature changes as resulting from the two time-slice experiments differ substantially from the transient run. The timeslice simulation with fixed ice sheets gives only a slight increase in the area of perennial snow cover at 115 BP as compared with the present. Thus, the global annual cooling simulated here is roughly 0.6 C smaller than that found in the transient run at 115 kyr BP. In the time-slice simulation with interactive ice sheets, the model climate and ice sheets have time to reach equilibrium with climate. Thus, ice sheets grow much larger than in the transient run, and the simulated cooling is substantially stronger. Agreement between the time-slice simulation with fixed preindustrial ice sheets and the transient run of the last glacial inception at 115 kyr BP can be apparently improved if two things are applied at the same time; firstly, the use of an atmospheric CO2 concentration of 240 instead of 271 ppmv (according to a second reconstruction of CO2 concentration, Barnola et al., 1987) and secondly, the inclusion of a high-resolution orography in the calculation of snow mass. Although this way, in the time slice run a perennial snow cover can be simulated that is comparable to the inland ice cover of the transient run at

588

Claudia Kubatzki et al.

sheets develop already under the permanent external forcing (low boreal summer insolation and CO2) corresponding to 121 kyr BP. Note that in the transient run, glacial inception appears only at about 117.5 kyr BP. In conclusion, time-slice simulations cannot be used to estimate either the amplitude or the timing of the last glacial inception.

(a) AOVItr (at 115 kyr BP) – Preindustrial 60

Latitude

30 0 –30 –60

–180

–120

–60

0

60

120

180

39.6 ORBITAL PARAMETERS AND CALENDAR DEFINITION

(b) AOV115k – Preindustrial 60

Calov et al. (2005b) demonstrated that the changing insolation is most crucial for the last glacial inception in CLIMBER-2. Transient changes in the atmospheric CO2 concentration amplify the growth of the ice sheets, and the choice of CO2 reconstruction used to drive the transient simulation does have an impact on the timing and magnitude of the glacial inception (Kubatzki et al., 2006a). But CO2 changes are neither necessary to simulate the last glacial inception (see Fig. 39.6, black curves) nor are they alone capable of initiating any substantial ice-sheet growth at that time.

Latitude

30 0 –30 –60

–180

–120

–60

0

60

120

180

60

120

180

(c) AOVI115k – Preindustrial 60

Latitude

30 0 –30 –60

–120

–60

0

10

Longitude –20

–5

–1

0

1

[°C]

Fig. 39.5 Differences in annual temperature ( C) as compared with the preindustrial climate for (a) the transient simulation of the fully coupled model at 115 kyr BP, ( b) a time-slice simulation of 115 kyr BP with fixed preindustrial ice sheets, and (c) a time-slice simulation of 115 kyr BP with interactive inland ice model.

115 kyr BP, still climatic differences exist between the two simulations (Kubatzki et al., 2006b). A series of runs with an interactive icesheet model reveals that in time-slice experiments, large northern hemispheric ice

Area (106 km2)

–180

8 6 4 2 0 –128

–124

–120

–116

–112

–108

–104

–100

Time (kyr BP)

Fig. 39.6 Ice area (106 km2) in North America in the transient simulation of the last glacial inception (applying changes of all three orbital parameters as well as a changing CO2; black solid), and in simulations where CO2 is fixed and changes are applied only for perihelion (green dashed), eccentricity ( blue dashed), obliquity (red dashed), precession, i.e. perihelion and eccentricity (green solid), perihelion and obliquity ( blue solid), eccentricity and obliquity (red solid) and all three orbital parameters ( black dashed) over the Eemian and last glacial inception.

Modelling the End of an Interglacial

Insolation changes in terms of changes in the orbital forcing are determined by the three orbital parameters (see chapters 2/3), each of them varying with its typical frequency, the tilt of the Earth’s axis, the eccentricity of the Earth’s orbit and the time of the perihelion (e.g. Berger, 1978). Crucifix and Loutre (2002) investigated the role of obliquity vs. precessional changes in simulations of the last glacial inception and found precession playing the major role. With CLIMBER-2, we tested the impact of the individual orbital parameters on our simulated climate and ice sheets. We repeated our transient simulation several times but each time allowing just one or two orbital parameters to vary with time. As a result, we find that the precession (and in fact, even only the perihelion) variation leads to a small ice sheet in North America (see Fig. 39.6, green curves). Changes in obliquity amplify the ice-sheet expansion, although obliquity changes alone are not sufficient to initiate any noticeable ice-sheet growth. Orbital parameter changes are clearly depicted in variations of the size of the Greenland ice sheet as well as in global climate variations. Climate model results are often presented in terms of monthly or seasonal averages. However, the calculation of these monthly or seasonal (not annual, i.e. calendar-independent) averages depends on the way the months are defined, either based on ‘classical’ days, i.e. equal time steps, or on equal angles on the Earth’s orbit. In terms of a comparison of incoming radiation at different geological times, the definition of monthly averages based on 30 angles on the orbit seems better (e.g. Berger, 1978; Kutzbach and Otto-Bliesner, 1982; Joussaume and Braconnot, 1997). To estimate the impact of the calendar definition on the calculation of monthly and seasonal averages for the last glacial inception, we averaged the results of our transient simulation using the angular definition (instead of applying the ‘classical’ definition based on days).

589

The interpretation of the model results at the end of the Eemian in terms of monthly and seasonal averages based on either calendar definition differs in two respects. Firstly, the transient minima and maxima of various seasonally averaged climate characteristics like northern hemispheric seasonal temperatures differ by up to 2 kyr. As we fixed the vernal equinox, this is true especially for the transient behaviour of fall averages. Secondly, the amplitude and timing of the seasonal cycle of, for example, the North African monthly averaged monsoon precipitation differs by up to ten per cent and one month, respectively, depending on what calendar definition is chosen to derive the averages (not shown here). The effect of the calendar definition on the ‘classical’ summer and winter means is only minor, as was already demonstrated by Montoya et al. (2000), e.g., for the Eemian. Of course, the choice of the calendar definition cannot affect calendar-independent data, like ice volume, forest area, annual temperature, maximum and minimum temperature or sum of positive degree days. Thereby, the latter are more preferable for the analysis of the simulations with varying orbital forcing. 39.7 ATLANTIC FRESHWATER DISTURBANCES CAUSE COLD EVENTS DURING THE LAST GLACIAL INCEPTION Palaeorecords show that during the last glacial inception, rather abrupt cooling and warming events have taken place over Greenland (e.g. NGRIP Members, 2004). The cold events are also seen in different proxies in the Atlantic and in Europe (as for example compiled in Mu¨ller and Kukla, 2004) and seem to be partly linked to the abundance of ice-rafted debris in the North Atlantic record. This suggests that the cold events might be at least partly explained by an increased iceberg calving and thus freshwater pulses into the Nordic Seas.

590

Claudia Kubatzki et al.

Temperature (°C)

In CLIMBER-2, we test this hypothesis of freshwater disturbances in the North Atlantic triggering cold events at the end of the last interglacial the following way. Prior to the last glacial inception, the decreasing insolation results in an expansion of northern hemispheric sea ice in the model. We assume that the larger sea ice area results in an increased sea ice export with the gyre circulation (see Ganopolski and Rahmstorf, 2001). After ice sheets started to grow, two effects are assumed. First a decreasing Arctic freshwater export caused by the shallowing and narrowing of the Denmark Strait and the closure of the Canadian archipelago. And secondly, an increased freshwater input into the Arctic Seas related to calving events from the ice sheets as they are displayed in the record of ice-rafted debris. The timing of the freshwater disturbances introduced into our model was chosen in accordance with the timing of the cold events in the NGRIP record. The purpose of this study was not to investigate the timing but the climatic consequences of these freshwater disturbances. Preliminary results of our simulation are shown in Fig. 39.7. The introduction of additional freshwater into the Nordic Seas leads to a weakening of the ocean overturning circulation and a southward shift of the convection sites and thus to a reduced heat transport into the north. As a result, the climate gets colder not only over Greenland (as was seen in the NGRIP record) but also in Europe. In particular, (deciduous) tree –4 –5 –6 –7 –8 –9 –10 –11 –128

–124

–120

–116

–112

–108

–104

–100

Time (kyr BP)

Fig. 39.7 Annual temperature ( C) over Greenland in the transient run (dashed) and in the transient run after modification of the freshwater balance in the North Atlantic as described in the text (solid).

cover in central Europe is reduced during these cold episodes in the model, a feature that is also evident in the geological record (see chapter 4). When the individual cold events come to an end, the Atlantic Ocean overturning circulation recovers completely. As ice-sheet growth, in particular in northern America, is enhanced during the cold events, climate remains cooler, also over Greenland. 39.8 EARLIER GLACIAL INCEPTIONS AND THE END OF THE HOLOCENE We also used our model to investigate the end of earlier interglacials, namely of MISs 7e, 9e, and 11 and to make a forecast on how natural climate would develop in the future (the term MIS is used here for convenience, to be more precise, 7e and 9e should be called substage; alternatively, terrestrial stages could be used). Again, CLIMBER-2 is driven by insolation as well as atmospheric CO2 changes (for MIS9 e/d, the CO2 changes of MIS 5e/d were assumed to be more plausible); for the future, a constant CO2 concentration of 280 ppmv was applied. Preliminary results show that our model is able to simulate a glacial inception at the end of MISs 7e, 9e and 11 (see Fig. 39.8a, b c). Whereas the scale of sea-level change at the end of MISs 9e and 11 can be simulated reasonably in comparison with the geological evidence (e.g. Siddall et al., 2003), the simulated magnitude of ice volume increase at the end of MIS 7e corresponds to a sealevel drop of about 100 m and is thus unrealistically large. To reduce inland ice growth at MIS7e/d to a reasonable amount, we have to apply interglacial CO2 concentrations of about 280 ppmv during the whole course of the model run. The insolation drop at the end of MIS 7e is larger than at the end of MIS 5e which explains the strong model response. The snow albedo feedback plays an important role in the simulation of the last

Modelling the End of an Interglacial (b) MIS 9e/d Sea level (m)

Sea level (m)

(a) MIS 7e/d

591

0 –20 –40 –60 –80 –100 –120 –244 –240 –236 –232 –228 –224 –220

0 –20 –40 –60 –80 –100 –120

(CO2 MIS5e/d)

–332 –328 –324 –324 –316 –312 –308

Time (kyr BP)

Time (kyr BP)

Sea level (m)

(c) MIS 11/10 0 –20 –40 –60 –80 –100 –120

–410

–400

–390

–380 –370 Time (kyr BP)

–360

–350

Fig. 39.8 Sea-level change (m) at the end of earlier interglacials, (a) MIS 7e/d, (b) MIS 9e/d, (c) MIS 11/10. The CO2 concentration in the model runs is chosen according to the reconstruction of Petit et al. (1999). As at the end of MIS 9e, the CO2 concentration remains at a high (interglacial) level in these data, we repeated the simulation for MIS 9e/d but using the CO2 reconstruction for MIS 5e/d (dashed line).

glacial inception. We are also able to simulate a reasonable increase in ice volume at the end of MIS 7e by decreasing the albedo of snow in our model by only a few percent which means making the model ‘warmer’. However, when we assume a snow albedo value in such a way that we are able to get a reasonable inception at MIS 7e (i.e. use the ‘warmer’ model), the simulated inland ice growth at MISs 9e/d and 11/10 becomes too weak. In this case, glaciation at the end of MIS 11 appears too late, and glacial inception at the end of MIS 9e can be simulated only when applying a glacial CO2 concentration of 210 ppmv during the whole model run. Using the higher value of snow albedo (and thus the ‘colder model’) in a timeslice simulation of the preindustrial climate leads to large-scale glaciation. As demonstrated before, time-slice simulations might not always properly display the timing and strength of glacial inceptions. In a transient run starting at the warmer conditions of about 10 kyr BP, noticeable glacial inception takes place only at about 50 kyr in the future (not shown here), comparable with results of Berger and Loutre (2002).

ACKNOWLEDGEMENTS We warmly thank Ralf Greve for providing us with the ice-sheet model SICOPOLIS. The authors would like to thank Alexandra Jahn for technical assistance. The work was funded by a subcontract to project 01LD0041 (DEKLIM-EEM) of the Bundesministerium fu¨r Bildung und Forschung (BMBF). Reinhard Calov was funded by the Deutsche Forschungsgemeinschaft (DFG) projects CL 178/2-1 and -2. REFERENCES Barnola, J.M., Raynaud, D., Korotkevich, Y.S., Lorius, C., 1987. Vostok ice core provides 160,000-year record of atmospheric CO2. Nature 329, 408–414. Berger, A., 1978. Long-term variations of daily insolation and Quaternary climatic changes. Journal of the Atmospheric Sciences 35, 2362–2367. Berger, A., Loutre, M.F., 2002. An exceptionally long interglacial ahead? Science 297, 1287–1288. Berger, A., Loutre, M.F., Galle´e, H., 1998. Sensitivity of the LLN climate model to the astronomical and CO2 forcings over the last 200 ky. Climate Dynamics 14, 615–629. Brovkin, V., Bendtsen, J., Claussen, M., Ganopolski, A., Kubatzki, C., Petoukhov, V., Andreev, A., 2002. Carbon cycle, vegetation, and climate dynamics in

592

Claudia Kubatzki et al.

the Holocene: Experiments with the CLIMBER-2 model. Global Biogeochemical Cycles 16 (4), 1139, DOI 10.1029/2001GB001662. Calov, R., Ganopolski, A., Claussen, M., Petoukhov, V., Greve, R., 2005a. Transient simulation of the last glacial inception. Part I: Glacial inception as a bifurcation in the climate system. Climate Dynamics 24 (6), 545–561, DOI 10.1007/s00382-005-0007-6. Calov, R., Ganopolski, A., Petoukhov, V., Claussen, M., Brovkin, V., Kubatzki, C., 2005b. Transient simulation of the last glacial inception. Part II: Sensitivity and feedback analysis. Climate Dynamics 24 (6), 563–576, DOI 10.1007/s00382005-0008-5. Chappell, J., Omura, A., Esat, T., McCulloch, M., Pandolfi, J., Ota, Y., Pillans, B., 1996. Reconciliation of late Quaternary sea levels derived from coral terraces at Huon Peninsula with deep sea oxygen isotope records. Earth and Planetary Science Letters 141, 227–236. Clark, P.U., Clague, J.J., Curry, B.B., Dreimanis, A., Hicock, S.R., Miller, S.R., Miller, G.H., Berger, G.W., Eyles, N., Lamothe, M., Miller, B.B., Mott, R.J., Oldale, R.N., Stea, R.R., Szabo, J.P., Thorleifson, L.H., Vincent, J.-S., 1993. Initiation and development of the Laurentide and Cordilleran ice sheets following the last interglaciation. Quaternary Science Reviews 12, 79–114. Crucifix, M., Loutre, M.F., 2002. Transient simulations over the last interglacial period (126–115 kyr BP): feedback and forcing analysis. Climate Dynamics 19, 417–433. Cuffey, K.M., Marshall, S.J., 2000. Substantial contribution to sealevel rise during the last interglacial from the Greenland ice sheet. Nature 406, 591–594. DeBlonde, G., Peltier, W.R., 1991. A one-dimensional model of continental ice volume fluctuations through the Pleistocene: implications for the origin of the mid-Pleistocene climate transition. Journal of Climate 4, 318–344. De Noblet, N., Braconnot, P., Joussaume, S., Masson, V., 1996a. Sensitivity of simulated Asian and African summer monsoons to orbitally induced variations in insolation 126, 115 and 6 kBP. Climate Dynamics 12, 589–603. De Noblet, N.I., Prentice, I.C., Joussaume, S., Texier, D., Botta, A., Haxeltine, A., 1996b. Possible role of atmosphere–biosphere interactions in triggering the last glaciation. Geophysical Research Letters 23 (22), 3191–3194. Dong, B., Valdes, P.J.,1995. Sensitivity studies of northern hemisphere glaciation using an atmospheric general circulation model. Journal of Climate 8, 2471–2496. Dong, B., Valdes, P.J., Hall, N.M.J., 1996. The changes of monsoonal climates due to earth’s orbital perturbations and ice age boundary conditions. Palaeoclimates 1, 203–240.

Galle´e, H., van Ypersele, J.P., Fichefet, T., Marsiat, I., Tricot, C., Berger, A., 1992. Simulation of the last glacial cycle by a coupled, sectorially averaged climate–ice sheet model 2. Response to insolation and CO2 variations. Journal of Geophysical Research 97 (D14), 15713–15740. Gallimore, R.G., Kutzbach, J.E., 1996. Role of orbitally induced changes in tundra area in the onset of glaciation. Nature 381, 503–505. Ganopolski, A., Rahmstorf, S., 2001. Rapid changes of glacial climate simulated in a coupled climate model Nature 409, 153–158. Ganopolski, A., Petoukhov, V., Rahmstorf, S., Brovkin, V., Claussen, M., Eliseev, A., Kubatzki, C., 2001. CLIMBER-2: a climate system model of intermediate complexity. Part II: model sensitivity. Climate Dynamics 17, 735–751. Greve, R., 1997. A continuum-mechanical formulation for shallow polythermal ice sheets. Philosophical Transactions of the Royal Society of London A355, 921–974. Joussaume, S., Braconnot, P., 1997. Sensitivity of paleoclimate simulation results to season definitions. Journal of Geophysical Research 102 (D2), 1943–1956. Kageyama, M., Charbit, S., Ritz, C., Khodri, M., Ramstein, G., 2004. Quantifying ice-sheet feedbacks during the last glacial inception. Geophysical Research Letters 31, L24203, DOI 10.1029/ 2004GL021339. Khodri, M., Leclainche, Y., Ramstein, G., Braconnot, P., Marti, O., Cortijo, E., 2001. Simulating the amplification of orbital forcing by ocean feedbacks in the last glaciation. Nature 410, 570–574 Kubatzki, C., Claussen, M., Calov, R., Ganopolski, A., 2006a. Sensitivity of the last glacial inception to initial and surface conditions. Climate Dynamics, 27, 333–344. Kubatzki, C., Calov, R., Claussen, M., Ganopolski, A., 2006b. On the problem of time-slice experiments in simulations of the end of the last interglacial. Climate Dynamics, submitted. Kutzbach, J.E., Otto-Bliesner, B.L., 1982. The sensitivity of the African-Asian monsoon climate to orbital parameter changes for 9000 years B.P. in a lowresolution general circulation model. Journal of the Atmospheric Sciences 39, 1177–1188. Lorenz, S., Grieger, B., Helbig, Ph., Herterich, K., 1996. Investigating the sensitivity of the atmospheric general circulation model ECHAM 3 to paleoclimatic boundary conditions. Geologische Rundschau 85, 513–524. Marsiat, I., 1994. Simulation of Northern Hemisphere continental ice sheets over the last glacial– interglacial cycle: experiments with a latitude– longitude vertically-integrated ice sheet model coupled to a zonally-averaged climate model. Palaeoclimates 1, 59–98.

Modelling the End of an Interglacial Meissner, K.J., Weaver, A.J., Matthews, H.D., Cox, P.J., 2003. The role of land–surface dynamics in glacial inception: a study with the UVic Earth System Model. Climate Dynamics 21 (7–8), 515–537. Montoya, M., von Storch, H., Crowley, T.J., 2000. Climate simulation for 125 000 years ago with a coupled ocean–atmosphere general circulation model. Journal of Climate 13, 1057–1072. Mu¨ller, U.C., Kukla, G.J., 2004. North Atlantic Current and European environments during the declining stage of the last interglacial. Geology 32 (12), 1009–1012, DOI 10.1130/G20901.1. North Greenland Ice Core Project Members, 2004. High-resolution record of northern hemisphere climate extending into the last interglacial period. Nature 431, 147–151. Petit, J.R., Jouzel, J., Raynaud, D., Barkov, N.I., Barnola, J.-M., Basile, I., Bender, M., Chappellaz, J., Davis, M., Delaygue, G., Delmotte, M., Kotlyakov, V.M., Legrand, M., Lipenkov, V.Y., Lorius, C., Pe´pin, L., Ritz, C., Saltzman, E., Stievenard, M., 1999. Climate and atmospheric history of the past 420,000 years from the Vostok ice core, Antarctica. Nature 399, 429–436. Petoukhov, V., Ganopolski, A., Brovkin, V., Claussen, M., Eliseev, A., Kubatzki, C., Rahmstorf, S., 2000. CLIMBER-2: A climate system model of

593

intermediate complexity. Part I: Model description and performance for present climate. Climate Dynamics 16, 1–17. Siddall, M., Rohling, E.J., Almogi-Labin, A., Hemleben, Ch., Meischner, D., Schmelzer, I., Smeed, D.A., 2003. Sealevel fluctuations during the last glacial cycle. Nature 423, 853–858. Tarasov, L., Peltier, W.R., 1997. Terminating the 100 ky ice age cycle. Journal of Geophysical Research 102 (D18): 21665–21693. Vettoretti, G., Peltier, W.R., 2004. Sensitivity of glacial inception to orbital and greenhouse gas climate forcing. Quaternary Science Reviews 23, 499–519. Waelbroeck, C., Labeyrie, L., Michel, E., Duplessy, J.C., McManus, J.F., Lambeck, K., Balbon, E., Labracherie, M., 2002. Sea level and deep water temperature changes derived from benthic foraminifera isotopic records. Quaternary Science Reviews 21: 295–305. Wang, Z., Mysak, L.A., 2002. Simulation of the last glacial inception and rapid ice sheet growth in the McGill Paleoclimate Model. Geophysical Research Letters 29 (23), 2102, DOI 10.1029/2002GL015120. Yoshimori, M., Reader, M.C., Weaver, A.J., McFarlane, N.A., 2002. On the causes of glacial inception at 116 kaBP. Climate Dynamics 18, 383–402.

This page intentionally left blank

Section 6 Synthesis

This page intentionally left blank

40. Chronology and Climate Forcing of the Last Four Interglacials Frank Sirocko, Martin Claussen, Thomas Litt, Maria Fernanda Sa´nchez Gon˜i, Andre´ Berger, Tatjana Boettger, Markus Diehl, Ste´phanie Desprat, Barbara Delmonte, Detlev Degering, Manfred Frechen, Mebus A. Geyh, Matthias Groeger, Masa Kageyama, Frank Kaspar, Norbert Ku¨hl, Claudia Kubatzki, Gerrit Lohmann, Marie-France Loutre, Ulrich Mu¨ller, Bert Rein, Wilfried Rosendahl, Katy Roucoux, Denis-Didier Rousseau, Klemens Seelos, Mark Siddall, Denis Scholz, Christoph Spo¨tl, Brigitte Urban, Maryline Vautravers, Andrei Velichko, Stefan Wenzel, Martin Widmann and Bernd Wu¨nnemann

The last four interglacials (intervals during which global ice volume was similar to, or less than, that of our current warm stage) correspond to the warmest parts of the marine oxygen isotope stages MIS 5, 7, 9, 11. These interglacials followed the 100-kyr rhythm of eccentricity, but each had different insolation regimes, different durations, different ice volumes and different sea-level heights, but atmospheric greenhouse gas concentrations were similar and reached values which, by and large, were close to those of the current interglacial (Holocene or MIS 1) before the industrial revolution led to the artificial enrichment of the atmosphere’s greenhouse gas concentrations via the burning of fossil fuels. The Holocene is addressed in a few papers, but an intercomparison of the ongoing interglacial with the past interglacials is not the focus of this book. This final paper of the book will summarize the evidence presented and discussed in the research articles. It is intended to be comprehensible to the lay reader and thus does not go into detail. Every paper is represented by a list of three statements which sum up the key findings. This is followed by a synthesis of the current state of knowledge on each of the climatic warm intervals discussed in the book. There is almost complete agreement on several themes, in particular on the subject of insolation forcing, while other topics, such as the correlation of sequences dated by different techniques, will need to be discussed and evaluated further before a consensus is reached. At the end of this paper, we make some

suggestions for future research themes, which will need to be answered soon if the climate of the past is to be of use for the prediction of climatic and environmental evolution in the future when the natural forcing of climate has to interact with the human influence on the atmosphere and land surface.

STATEMENTS BY FIRST AUTHORS OF RESEARCH ARTICLES Martin Claussen, University Hamburg and Max Planck Institute for Meteorology, Hamburg, Germany (1. Introduction to Climate Forcing and Climate Feedbacks) 1. Commonly, climate is defined as statistics (mean, variance, . . .) of the atmosphere. In climate physics, however, a wider definition has proven to be useful in which climate is viewed as the state and statistics of the climate system which encompasses the atmosphere, the ocean, the ice and the land including the living world, i.e. the marine and terrestrial biosphere as well as carbon and nutrient cycles. 2. Climate varies on all timescales, not only because of changes in climate forcing, but also because of internal dynamics and feedbacks between the components of the climate system which are seemingly unrelated to variations in forcing.

598

Frank Sirocko et al.

3. Not all climate variations on timescales longer than 30 years are directly driven by oscillation forcing; some of these variations could arise because of the sluggish dynamics of the deep ocean or the ice sheets or because of a strongly nonlinear, disproportional response of the climate system to subtle variations in forcing. Andre´ Berger, Universite´ Louvain, Belgium

catholique

de

(2. Insolation During Interglacials) 1. The spectrum of the long-term variations of daily insolation is dominated by precession everywhere over the Earth and for any day except close to the polar night. 2. The energy received by the Earth over a given time slice during the year defined in terms of the longitude of the Sun (an astronomical season for example) is a function of obliquity only, the length of such astronomical seasons being a function of precession only. 3. The amplitude of the variations of insolation (in particular during the interglacials) is a function of the amplitude of the variations of precession. Martin Claussen, University Hamburg, and Max Planck Institute for Meteorology, Hamburg, Germany (3. A Survey of Hypotheses for the 100-kyr Cycle) 1. Even some 160 years after first geological evidence, the ice-age riddle is not yet fully solved. However, we have some clues on which elements should constitute a theory of Quaternary Earth system dynamics, regarding concepts and model structure. 2. It is certain that the ice-age riddle cannot be explained in terms of a single dominating process. Instead, a systems approach involving a number of feedback processes and the nonlinear nature of the climate system is expected to lead to a solution.

3. It is likely that changes in insolation caused by changes in orbital parameters trigger fast internal feedbacks such as the water vapour–temperature feedback and the snow–albedo feedback. Some feedbacks, like the snow–albedo feedback, amplify climate changes very rapidly once a certain threshold in the system is crossed. Initial changes are then further amplified by slower feedbacks such as biogeochemical and biogeophysical feedbacks and finally, the isostatic response of the lithosphere to icesheet loading. Andre´ Berger, Universite´ Louvain, Belgium

catholique

de

(4. Modelling the 100-kyr Cycle – An Example From LLN EMICs) 1. The LLN model is a model of intermediate complexity which takes into account, in a simplified way, the atmosphere, the hydrosphere, the cryosphere, the biosphere, the lithosphere, their mutual interactions and internal feedbacks. Under the forcing of insolation and a progressively decreasing atmospheric CO2 concentration, it simulates the transition between the 41-kyr and the 100-kyr worlds around 850 kyr. 2. The model also simulates the spectrum of the northern hemisphere ice volume over the last 450 kyr, with the strongest period at 100 kyr. 3. It fails to reproduce the reduced amplitude of the 100-kyr cycles before 450 kyr with cool interglacials and cold glacials. Frank Sirocko, University of Mainz, Germany (5. Introduction: Palaeoclimate Reconstructions and Dating) 1. The beginning, end and duration of the past interglacials do not appear to be synchronous all over the world, i.e. parts of the climate system have been in an interglacial state for longer than others.

Chronology and Climate Forcing

599

2. Beginning and end of interglacials in the low latitudes and in the Antarctic lead to respective changes on the northern hemisphere. 3. Time-transgressive climate shifts are also strong over Europe, where the sea-surface temperature (SST) changes of the North Atlantic drift were associated with a stepwise shift of the vegetation zones, at least at the end of the past interglacial, with interglacial conditions persisting for longer in southern Europe than in the north.

1. The last nine interglacial periods differ not only in height and variability of sea level, but also in timing relative to northern summer insolation peaks. 2. Sea levels during MIS 5e, 9c and 11 were close to or slightly higher than modern sea level but sea level during MIS 7 may have been slightly lower than present day. 3. Some interglacials have a single peak close to modern sea level (MIS 5e, 9c) and others have several (MIS 7), while MIS 11 persisted with little variation for at least 30 kyr.

Barbara Delmonte, DISAT, Milano-Bicocca, Milano, Italy.

Manfred Frechen, Leibniz Institute for Applied Geosciences (GGA-Institut) Hannover, Germany

University

(6. Late Quaternary Interglacials in East Antarctica from Ice-Core Dust Records) 1. Aeolian dust records from deep East Antarctic ice cores preserve evidence for extremely low dust fluxes during the last five interglacials (10 to 25 times lower than in glacial periods). This is related to reduced primary production and mobilization of dust on the Southern Hemisphere continents, and to changes in atmospheric transport and the hydrological cycle. The data show no evidence for pronounced cold events within the last five interglacials (back to MIS 11.3). 2. The Sr–Nd isotope fingerprint of aeolian dust in Antarctica suggests a dominant southern South American provenance during Quaternary glacial times, but the first geochemical data for stage 5.5 and the Holocene show significant differences and opens up the possibility for different provenance mixing. 3. Dust-size variability in the EPICA-Dome C ice core suggests shorter transport time for dust or more direct air mass penetration to the site during interglacials with respect to cold periods. Mark Siddall, University of Bern, Switzerland (7. Eustatic Interglacials)

Sea

Level

During

Past

(8. Uranium-Series Dating of Peat from Central and Northern Europe) 1. Fen peat is suitable for uranium-series dating under the asumption of complete fractionation of uranium and thorium during formation and no gain or loss of U and Th since the time of formation. 2. Uranium-series dating provides more reliable and precise absolute dates of interstadial and interglacial peat layers in Central Europe. Examples of uranium dates for MIS 5, 7 and 9 are: 91 2 kyr for lignite (MIS 5c) from Zell in Switzerland, 106 11 kyr for the fen peat (MIS 5e) from Allt Odhar in Scotland, 214 8 kyr for peat (MIS 7) from Groß-Rohrheim in Germany and 317 14 kyr for fen peat from Tottenhill Quarry in Norfolk, England. 3. TIMS 230Th/U-dating results provide a chronological frame for terrestrial sediments such as fen peat, cave sinter and travertine covering the past 500 000 years. Denis Scholz, Heidelberger Akademie der Wissenschaften, Germany (9. U-Redistribution in Fossil Reef Corals from Barbados, West Indies, and Sea-Level Reconstruction for MIS 6.5)

600

Frank Sirocko et al.

1. Postdepositional remobilisation produces significantly wrong U-series ages of fossil reef corals and may, therefore, have consequences for the precise determination of the timing and duration of past interglacials. 2. A small degree of U-redistribution can only be detected by analysis of a large number of samples but not by the conventional reliability criteria. 3. MIS 6.5 sea level was between 50 11 and 47 11 m relative to the present sea level from 176:1 2:8 to 168:9 1:4 kyr BP. Bert Rein, Johannes Gutenberg-University, Mainz, Germany (10. Holocene and Eemian Varve Types of Eifel Maar Lake Sediments) 1. During the last interglacial, multicentennial periods of increased dust storm activity and aeolian dust deposition occurred at 126, 118 and 112 kyr BP. So far, no evidence exists for comparable extreme, multicentennial dry periods within the Holocene as observed in the deposits of the last interglacial palaeolake. 2. Besides the Younger Dryas sediments, which where deposited immediately before the Holocene period, no indication could be found for climatically induced dust storm activity during the Holocene. 3. The lithic maximum in the sediments of the Little Ice Age (290–630 yr BP) does not necessarily require a climatological explanation since it can be explained by anthropogenically induced soil erosion due to increasing population density around the lake when medieval villages on the surrounding plateau deserted. Detlev Degering, Saxon Academy of Sciences, TU Freiberg, Germany (11. Dating of Interglacial Sediments by Luminescence Methods) 1. Luminescence methods were successfully applied in the age determination of

interglacial sediments. The dated event is the last optical bleaching of the sediment and the maximum determinable ages are currently of the order of some hundred thousands of years, depending on sediment properties and the luminescence technique applied. A sufficiently long exposure to sunlight prior to sedimentation is essential for correct age determination; an incomplete reset of the luminescence signal will lead to age overestimation. 2. Eemian deposits can be dated with adequate precision by conventional luminescence methods. These include multiple aliquot infrared stimulated (IRSL) techniques (using polymineral fine-grain and coarse-grain potassium feldpar samples) and, in some cases (of low radioactivity), quartz optically stimulated luminescence (OSL) single-aliquot regeneration techniques. However, higher precision is reached by the infrared radiofluorescence method (IR-RF). Reliable luminescence dating of older interglacials (MIS 7 and 9) is only possible by the IR-RF method. 3. Dating of interglacial sediments requires, furthermore, the consideration of special dose rate-related problems: (i) the occurrence of radioactive disequilibria and (ii) layered sediments with varying radioisotope content. In both cases, high analytical effort and model calculations are necessary to minimize the influence of these sources of errors. Stefan Wenzel, Schloß Monrepos, Neuwied, Germany (12. Neanderthal Presence and Behaviour in Central and Northwestern Europe During MIS 5e) 1. The majority of the few archaeological sites of Central and Northwestern Europe which can be dated more precisely within MIS 5e are attributed to the first third of this interglacial (up to the Quercetum mixtum–Corylus phase).

Chronology and Climate Forcing

2. So far, there is no evidence of hominid occupation in the Carpinus phase of the Eemian, and only a few archaeological sites are known from younger biozones of the Eemian. 3. In contrast to the rarity of archaeological sites from the last two-thirds of the last interglacial, which could indicate a deterioration of living conditions, the early Eemian sites and the sites solely attributed to the Eemian evidence that the lifeways of the Neanderthals living then differed little from those of their ancestors and their successors occupying a more open landscape: they hunted big game; they used a similar lithic technology, they practised long-distance transport of lithic artefacts (perhaps indicative of social networks) and they performed symbolic behaviour. Maria Fernanda Sa´nchez Gon˜i, Universite´ Bordeaux 1, France (13. Introduction to Climate and Vegetation in Europe During MIS 5) 1. The complexity, both temporal leads and lags and geographical variability, of the nonlinear climatic signal in response to insolation changes during the penultimate deglaciation, last interglacial and last glacial inception is clearly shown by the collection of papers presented in this chapter. 2. Problems of nomenclature and stratigraphy for the last interglacial, reflecting the history of the discipline, are on the way to being solved, although uncertainties in the chronology of a number of MIS 5 records require further investigation. 3. Our scientific community should concentrate its efforts in accurately correlating the available records. It is only through this approach that we will be able to document the climatic variability of MIS 5 in an integrated way, link the processes reflected in different parts of the Earth system and propose reliable scenarios

601

for the mechanisms underlying the climatic variability of the last interglacial. Klemens Seelos, Johannes University, Mainz, Germany

Gutenberg-

(14. Aprupt Cooling Events at the Very End of the Last Interglacial) 1. The LEAP (late Eemian aridity pulse) is detectable in loess records of the Eifel region and in the northern German record of Rederstall. 2. Taiga vegetation characterizes Rederstall in the first phase after the LEAP (after 118 kyr) and develops gradually into tundra after 115 kyr. At the same time, Carpinus-dominated temperate forest spread in the Eifel region and in France, and finally deteriorated at 111 kyr. 3. Loess and pollen records in the Eifel region and northern Germany show the first drastic and fast cooling, associated with widespread aridity, during the C24 cold event (111 kyr). Denis-Didier Rousseau, Universite´ Montpellier II, France (15. Estimates of Temperature and Precipitation Variations During the Eemian Interglacial: New Data From the Grande Pile Record (GP XXI)) 1. We applied a new method on new pollen data from Grande Pile, the inverse mode to the Biome4 vegetation model. The method utilizes 13 C, measured in parallel to the pollen samples as a constraint for the model. First the biomes and the 13 C simulated by the model are compared with the biome allocation of the pollen data. The 13 C to be simulated takes into account the degradation effect on the preserved organic matter. 2. This study highlights variation ranges of annual precipitation and/or temperature narrower using an inverse modelling procedure with 13 C than without. These

602

Frank Sirocko et al.

narrow ranges allow to reveal expected climatic trends that were noticed in marine sediments but not precisely reconstructed on the continent. 3. The variations in temperature appear to be related to SST oscillations in the North Atlantic region and are also in agreement with the timing of ice-sheet build-up in the Northern Hemisphere. Seasonal variations are also identified in the estimated temperatures of the warmest and coldest months. Norbert Ku¨hl, University of Bonn, Germany (16. Quantitative Time-Series Reconstructions of Holsteinian and Eemian Temperatures Using Botanical Data) 1. Reconstruction of Eemian and Holsteinian temperatures shows uninterrupted interglacial conditions for both intervals. However, the Holsteinian seems to be less stable than the Eemian with some intra-interglacial coolings. 2. The course of temperature change within the Eemian and Holsteinian interglacial stages differs. Holsteinian January and July temperatures were higher in the later part of the interglacial, while the Eemian had its temperature optimum during the early phase. 3. The temperature decline at the very end of the Holsteinian interglacial resembles in magnitude the decrease at the end of the Eemian. Reconstructions reveal a drop in mean January temperature by 10–15 C and in mean July temperature by about 3 C. Andrei Velichko, Institute of Geography RAS, Moscow, Russia (17. Comparative Analysis of Vegetation and Climate Changes During the Eemian Interglacial in Central and Eastern Europe) 1. As follows from comparison of pollen data, the main phases in the evolution

of vegetation during the Eemian Interglacial were similar in the Central and Eastern Europe. 2. We demonstrate the regional pattern of vegetation in the Eemian and Early Weichselian. Mixed broad-leaved forests in Central Europe included species that characteristically require a certain oceanicity of climate (e.g. Ilex aquifolium, Hedera helix and Taxus baccata). The participation of these plants decreased eastward. Of those species, only Tilia platyphyllos and Viscum album are found in the Eemian pollen assemblages in the East European Plain. Plant communities also differed noticeably from west to east during the cooler intervals at the beginning and end of the interglacial, primarily in the proportion of broad-leaved species in zonal vegetation formations. 3. Significant contrasts in environmental and climatic fluctuations mark the Saalian/Eemian boundary (transition from MIS 6 to MIS 5e). Two substages of vegetation development can be identified in the pollen diagram. Pine and spruce forest with shrubs occurred in the earliest phase. Then open birch woodlands and steppe-like communities occupied the area. Vegetation dynamics at this boundary resemble those detected at the transition from Weichselian to Holocene (Allerød and Younger Dryas). Tatjana Boettger, UFZ Centre for Environmental Research Leipzig-Halle, Germany (18. Indications of Short-Term Climate Warming at the Very End of the Eemian in Terrestrial Records of Central and Eastern Europe) 1. Geochemical and palynological investigations of lacustrine sediments from Central and Eastern Europe document at least two warming events during the transition from the Eemian to the Early Weichselian on a broad European transect.

Chronology and Climate Forcing

2. The first pronounced warming phase takes place towards the very end of the Eemian. The second climatic amelioration was detected within the first Weichselian Stadial (Herning). 3. Correlations between Eemian European terrestrial sequences and their possible connection to the NGRIP record are discussed. Generally, it appears that warming phases towards the end of the last interglacial preceded the final transition to glacial conditions. Ulrich Mu¨ller, Johann Wolfgang Goethe University Frankfurt, Germany (19. Vegetation Dynamics in Southern Germany During Marine Isotope Stage 5 ( 130 to 70 kyr Ago)) 1. The Eemian interglacial in southern Germany is characterized by dense thermophilous deciduous forests during the early and middle part and coniferous forests in the late part of the interglacial. 2. The exact timing and persistence of Eemian forests in southern Germany is a matter of debate. Varve chronologies suggest a duration of Eemian forests from 126 to 115 kyr BP in northern Germany and from 127 to 109 kyr BP in southern Italy. Possibly, Eemian forests existed in southern Germany from 126 to 110 kyr BP. 3. A comparison of vegetation reconstructions across Europe points to a steepening of meridional vegetation gradients during the declining stage (115 to 109 kyr BP) of the last interglacial. Presumably, the steepening of vegetation gradients was associated with a southward displacement of the warm North Atlantic current. Maryline Vautravers, University of Cambridge and British Antarctic Survey, Cambridge, UK (20. Subtropical NW Atlantic Surface Water Variability During the Last Interglacial)

603

1. In the Gulf Stream area, eight periods can be recognized between 108 and 134 kyr related to changes in the intensity of summer stratification and late winter mixing. 2. At the end of MIS 5e (122–116 kyr), winter SST are at their maximum while summer SST start to decrease. Superimposed on this trend, we found several high-frequency cooling events. 3. Most of these small coolings are associated with lithics peaks testifying for iceberg incursions in the subtropical area during the last interglacial. Bert Rein, Johannes Gutenberg University, Mainz, Germany (21. Abrupt Changes of El Nin˜o Activity off Peru During Stage MIS 5e-d) 1. El Nin˜o activity sharply dropped during the last interglacial, as it did uring the middle of the Holocene when perihelion occurred in late summer. However, during MIS 5d, the strength of El Nin˜o activity did not recover with more favourable insolation conditions as was observed during the late Holocene. 2. The strong El Nin˜o activity that was indicated by the Zebiak and Cane ENSO model according to orbital forcing did not occur in Peru at the beginning of the last glaciation. The transition into a glacial world apparently changed critical boundary conditions, which are linearized around a current mean climatology in the ZC ENSO model. Bernd Wu¨nnemann, Freie Universita¨t Berlin, Germany (22. Interglacial and Glacial Fingerprints from Lake Deposits in the Gobi Desert, NW China) 1. Hydrological changes in the Chinese Gobi Desert are strongly interlinked with climate dynamics on the Tibetan Plateau during the last interglacial–glacial cycle

604

Frank Sirocko et al.

2. The Eemian interglacial stage between 129 and 119 kyr appears to have been a period of positive water balance within the Gaxun Nur Basin as a result of warm and moist climate conditions with enhanced summer monsoon moisture. 3. The subsequent rapid increase of climate instability with phases of colder and drier conditions coincides with strong shrinkages of lake size and enhanced aeolian transport, frequently influenced by the extra-tropical westerlies and the winter monsoon. Katy Roucoux, University of Leeds, UK (24. Fine Tuning the Land–Ocean Correlation for the Late Middle Pleistocene of Southern Europe) 1. There are pronounced phase offsets between forested intervals in Portugal and marine isotope-defined warm intervals. For example, MIS 7e lasts from 246 to 229 kyr, while the forested interval associated with it lasts from 243.2 to 237 kyr. 2. Forested intervals varied in length from one warm stage to the next, and correlation of marine and terrestrial pollen sequences indicates that this pattern applies across southern Europe. In the marine pollen record of MD01-2443, the shortest forested period, at 3.5 kyr long, is the Lisboa forest interval associated with MIS 9e, while the longest uninterrupted forested period, at 10 kyr long, is the Cascais forest interval associated with MIS 7c. 3. Forested intervals also varied in floristic character across southern Europe as a result of local climatic, geological and biogeographical factors. For example, it appears that while tree populations were replaced by Ericaceous heath in Portugal at the end of the forested intervals associated with MIS 7e and MIS 9e, coniferous forest thrived elsewhere in southern Europe. Hence, we cannot

assume that periods of interglacial conditions necessarily resulted in forest vegetation everywhere for their whole duration. Ste´phanie Desprat, Universite´ Bordeaux1, France (25. Climate Variability of the Last Five Isotopic Interglacials: Direct Land–Sea–Ice Correlation from the Multiproxy Analysis of North-western Iberian Margin Deep Sea Cores) 1. The last five isotopic interglacials (MIS 1, 5, 7, 9 and 11) were investigated in NW Iberian margin deep-sea cores, using terrestrial (pollen) and marine (planktic foraminifera, benthic and planktic foraminifera oxygen isotopes) climatic indicators. 2. This work shows that the climatic variability detected on the continent is contemporaneously recorded in the ocean. Although minima and maxima of ice volume and marine and terrestrial temperatures in the NW Iberian region appear synchronous, temperature changes are not in phase with ice volume variations. Particularly, substantial ice accumulation at high latitudes, associated with the glacial inception of these past interglacials, generally lags NW Iberian temperature decrease by some millennia. 3. The comparison of the different marine isotope stages highlights a common pattern of climatic dynamics within these isotopic interglacials, which is characterized by three major climatic cycles, related to orbital cyclicity, on which suborbital climatic fluctuations are superimposed. Mebus A. Geyh, Leibniz Research Institute of Geosciences, Hannover, Germany (26. Palynological and Geochronological Study of the Holsteinian/Hoxnian/Landos Interglacial)

Chronology and Climate Forcing

1. Based on new TIMS 230Th/U dates from two fen peat layers of the Holsteinian reference site at Bossel, from several Holsteinian and non-Holsteinian profiles in northern Germany as well as on reevaluated numerical 230Th/U dates from two sites with Hoxnian deposits in England, the Holsteinian Interglacial has an 230Th/U age of about 320 kyr and therefore is correlated with MIS 9. 2. There is palynological evidence for a reliable correlation between precisely analysed Holsteinian and Hoxnian deposits in Poland, Germany, England, SWIreland and France. 3. It became obvious that the Holsteinian interglacial is correlated with the Landos interglacial rather than with the Praclaux interglacial. The latter belongs to MIS 11 and is linked with the Rhume interglacial. Markus Diehl, Johannes Gutenberg-University, Mainz, Germany (27. A New Holsteinian Pollen Record from the Dry Maar at Do¨ttingen (Eifel)) 1. The Do¨ttingen pollen sequence is the first Holsteinian pollen profile from middlesouthern Germany. It represents a ‘low mountain range-type’ Holsteinian vegetation succession, to correlate with that of the north German lowlands, but with different pollen percentage values, showing that the pine-birch dominance of the north German profiles cannot be seen as a overregional, typical Holsteinian signature. 2. The first spike of pine-birch dominance (sometimes interpreted in Northern Germany as a cooling event) is also visible at Do¨ttingen, but is preceded by a volcanic ash layer. The Do¨ttingen sequence thus points towards a causal connection between volcanic activity and the Holsteinian vegetation development. 3. The transition of the Holsteinian to the subsequent cold stage is characterized by

605

a change from a temperate forest to an open boreal forest within centuries. Brigitte Urban, University of Lu¨neburg, Germany (28. Interglacial Pollen Records from Scho¨ningen, North Germany) 1. The complex Pleistocene sequence of the Scho¨ningen browncoal mine, Germany, contains six major cycles (I–VI) providing biostratigraphical evidence of four interglacials (Holsteinian, Cycle I, Reinsdorf, Cycle II, Scho¨ningen Cycle III, Eemian, Cycle V) and a number of interstadials younger than the Elsterian glaciation and preceding the Holocene (Cycle VI) being tentatively correlated with MIS 5, 7, 9 and 11 respectively. 2. The position of the Lower Palaeolithic spruce throwing spears of Scho¨ningen can be assigned to the ultimate and already cool and dry Reinsdorf interstadial B of Cycle II (MIS 9/8?), characterized by a pine-birch open forest. 3. MIS 7 and MIS 9 (tentatively correlated with the Scho¨ningen and Reinsdorf interglacials, respectively) differ strongly in their moisture regime as shown by the low presence of fir (Abies) in terminal phases of the Reinsdorf and its total lack in the Scho¨ningen interglacial, pointing to an increasing dryness of the forested periods (interglacials) following the Holsteinian in northern Germany. Wighart von Koenigswald, University of Bonn, Germany (29. Mammalian Faunas From the Interglacial Periods in Central Europe and Their Stratigraphic Correlation) 1. The MIS 5e, 7, and 9 can be distinguished only vaguely by the mammalian fauna. Their diversity in insectivores and rodents is smaller than in the pre-Elsterian faunas. The only taxon showing an evolutionary trend is large vole Arvicola changing from cantianus to terrestris at about the Eemian.

606

Frank Sirocko et al.

2. The multiple climatic changes during the middle and upper Pleistocene caused an almost complete exchange of the mammalian fauna. In each interglacial period, the Mammuthus assemblage adapted to the cold climate disappeared and the Elephas assemblage immigrated from the Mediterranean region. Most mammalian lineages do not show a continuous evolution in Central Europe. Immigration and local extinction is the normal pattern. 3. The occurrence of the exotic Hippopotamus amphibius, known in recent times only from Africa, does not indicate very high temperatures or during the Eemian, but more likely a high maritime influence with mild winters. Wilfried Rosendahl, Reiss-EngelhornMuseen, Mannheim, Germany. (30. MIS 5 to MIS 8 – Numerically Dated Palaeontological Cave Sites of Central Europe) 1. The numerical dates (MIS 5 - MIS 8) now available for palaeontological cave sites in Central Europe (12 sites with 31 dated strata) do not allow a critical discussion of their faunal assemblages with regard to their ecological-climatic distribution, with the exception of two sites. But even these two sites are not without contradicting faunal elements, and it remains uncertain whether they represent glacial or interglacial faunas. 2. In spite of all these problems, palaeontological cave sites represent a rich archive that can deliver important contributions to the reconstruction of the Middle and Upper Pleistocene palaeoclimate of Central Europe, provided many additional dates can be obtained to verify results obtained from other terrestrial archives. Christoph Spo¨tl, Leopold-Franzens-Universita¨t Innsbruck, Austria (31. The Last and the Penultimate Interglacial as Recorded by Speleothems From a

Climatically Sensitive High-Elevation Cave Site in the Alps) 1. Changes in climate during MIS 5 to 7 identified by speleothem growth periods and oxygen isotope data are synchronous within the precision of the U/Th method with coral records of sea-level changes. 2. During MIS 7, the climate in the Alps was cooler, glaciers were larger and the timberline was lower than during MIS 5e. Speleothem growth during MIS 7 commenced 236 kyr and ended 190 kyr ago. 3. Full interglacial conditions during MIS 5e were reached 130 kyr ago and terminated 118–119 kyr ago. Martin Claussen, University Hamburg and Max Planck Institute for Meteorology, Hamburg, Germany (32. Climate System Models – A Brief Introduction) 1. In climate physics, climate models are a set of mathematical equations which are derived from physical principles and which are used in a prognostic mode to predict climate variations as function of some external forcing or which are tuned to data to interpolate between sparse data in time and space in a physically consistent manner. 2. When comparing data and model results, it is important to realize that climate is regarded as stochastic processes; therefore, it is not possible to predict all climate variations such as the precise course of a glacial interception in a deterministic way. 3. Only the interpretation of past climate variations by using mathematical climate models, which are validated against palaeoclimate evidence, will lead to picture of climate which is consistent with the physical understanding of our world. 4. In this chapter, models of different complexity are used. In chapters 33, 34

Chronology and Climate Forcing

and 35, results of a comprehensive atmosphere ocean model or even an atmosphere ocean–vegetation model (in chapter 37) are discussed. In chapters 36, 38 and 39, models of intermediate complexity are used which interactively simulate the dynamics of the atmosphere, the oceans, vegetation and ice sheets. Frank Kaspar, Max Planck Institute for Meteorology, Hamburg, Germany (33. Simulations of the Eemian Interglacial and the Subsequent Glacial Inception with a Coupled Ocean-Atmosphere General Circulation Model) 1. European seasonal patterns of pollenbased temperature reconstructions for the early Eemian (at 125 kyr) are in good agreement with simulation results of a state-of-the-art coupled ocean atmosphere general circulation model (ECHO-G) which was driven by orbitally induced changes in insolation. 2. When the same model is driven with insolation patterns of the last glacial inception at 115 kyr, the significantly reduced summer insolation of the northern latitudes leads to a distinct cooling of the northern hemisphere and the occurrence of a perennial snow cover over parts of North America. 3. The snow cover initially occurs in the region of the Canadian Archipelago, where cool southward winds from the Arctic prevail, and it is continuously expanding into the continent during the simulated period of several millennia. Martin Widmann, GKSS Research Centre, Geesthacht, Germany, and University of Birmingham, UK (34. Simulated Teleconnections During the Eemian, the Last Glacial Inception and the Preindustrial Period) 1. Simulations with a coupled atmosphere– ocean general circulation model (the same

607

as used in the previous section) indicate that orbital forcing can change relationships between different climate variables. 2. Simulated temperature signals of multidecadal variations in the Arctic oscillation index are weaker in Europe and stronger in Siberia during the Eemian compared to the preindustrial period. 3. Simulated teleconnections between annual to decadal temperature variability at different locations are related to the Arctic Oscillation temperature signal, and can be somewhat different in the Eemian and in the preindustrial period. Gerrit Lohmann, Alfred Wegener Institute for Polar and Marine Research, Bremerhaven, Germany. (35. Orbital Forcing on Atmospheric Dynamics During the Last Interglacial and Glacial Inception) 1. Atmospheric dynamics plays an important role on orbital timescales, e.g. the modulation of the Icelandic Low. 2. The atmospheric teleconnections provide a bridge between low and high latitudes and transport the precessional forcing to high latitudes. 3. The related atmospheric circulation patterns induce nonuniform temperature anomalies which can have a greater amplitude than those resulting from direct solar insolation forcing. 4. Transient simulations are performed using the coupled atmosphere–ocean– sea ice general circulation model ECHO-G by applying a novel acceleration technique. Andre´ Berger, Universite´ Louvain, Belgium

catholique

de

(36. Interglacials as Simulated by MOBIDIC and the LLN 2-D Climate Models) 1. The LLN models succeed to reproduce the interglacials of the last 450 kyr as a response to insolation and CO2 provided

608

Frank Sirocko et al.

feedbacks are correctly taken into account, in particular those related to water vapour, surface albedo and isostatic rebound. Sensitivity analyses show that glacial–interglacial cycles cannot be simulated under CO2 forcing alone. Using the insolation forcing alone, glacial–interglacial cycles are simulated but only under low constant CO2 values (under 230 ppmv). 2. Prediction for the next hundreds of thousands of years leads to a very long interglacial MIS 1. This length is similar to what was simulated for MIS 11. In both cases, these extremely long interglacials seem to be related to the very low eccentricity value which prevails during these times, the orbit of the Earth being almost circular. 3. In mid- and high latitudes during MIS 5e, sea-surface temperature in summer starts to decrease 11 kyr before the ice sheets start to grow over the continents. Matthias Groeger, Max-Planck-Institut fu¨r Meteorologie, Hamburg, Germany (37. Vegetation – Climate Feedbacks in Transient Simulations Over the Last Interglacial (128 000–113 000 yr BP)) 1. During the Eemian, considerable changes in the land surface properties are simulated with a coupled ocean atmosphere vegetation marine biogeochemistry model (ECHAM3.6 – LSG2 – LPJ – HAMOCC); among these the northward expansion of boreal forests, and the greening of the Sahara desert are the most pronounced ones. 2. On land, the areas most sensitive to insolation are the northern high latitudes and the northern hemisphere monsoon belt. 3. In these regions, vegetation feedbacks can amplify the insolation-forced climate change by a factor of 2. Masa Kageyama, Laboratoire des Sciences du Climat et de l’Environnement, Gif sur Yvette Cedex, France

(38. Mechanisms Leading to the Last Glacial Inception over North America: Results From the CLIMBER-GREMLINS AtmosphereOcean-Vegetation-Northern Hemisphere IceSheet Model) 1. The CLIMBER-GREMLINS model, in which the atmosphere, the ocean, the vegetation and the northern hemisphere ice sheets and the interactions between those components are represented, simulates a glacial inception over northern North America under the transient insolation and CO2 forcings of the period 126–106 kyr. 2. Glacial inception does not occur when vegetation is fixed to its interglacial state and is twice as slow, in terms of volume, when the icesheet albedo feedback is not taken into account. 3. The role of the ocean and sea ice are more ambiguous. A first experiment in which the seasonal cycle of the ocean surface characteristics is fixed to its interglacial state shows that this yields a faster inception than with an interactive ocean. On the other hand, when the Atlantic meridional overturning circulation is forced to stop, this yields an even faster inception. Claudia Kubatzki, Alfred Wegener Institute for Polar- and Marine Research, Bremerhaven, Germany (39. Modelling the End of an Interglacial (MIS 1, 5, 7, 9, 11)) 1. The CLIMBER-SICOPOLIS model, which describes the dynamics of atmosphere, ocean, vegetation and northern hemispheric inland ice, simulates a glacial inception over northern North America at about 117.5 kyr as a result of mainly changes in precession (perihelion), obliquity as well as CO2 changes act as amplifiers. 2. Glacial inception in North America at the end of the Eemian might not have happened with vegetation and ocean surface remaining at their interglacial (Eemian)

Chronology and Climate Forcing

state, i.e. the feedback of the changing Earth’s surface on climate is needed (compare with 38). 3. Cold events at the end of the interglacial can be reproduced, by the introduction of freshwater disturbances into the model North Atlantic. Synthesis on the chronology and forcing of climate change during the last interglacials 1. There is clear evidence that insolation changes are the primary driver of past climate variations. The 100-kyr rhythm of eccentricity dominates the occurrence of past interglacials, but it is still not fully understood why interglacials have occurred even at times when insolation on the northern hemisphere was quite low in absolute terms (for example, during MIS 11). Presumably, it is not the amplitude of insolation, but the change of the amplitude, which drives the climate system to cross some thresholds. 2. Some model studies indicate that precession is the main driver of glacial inception. 3. North Atlantic SST patterns, mainly a function of deep-water formation in the far north Atlantic, are an important mechanism for the temperature and availability of moisture in central Europe. 4. Also changes in vegetation cover, mainly a shift in Artic tree line, and changes in ocean dynamics appear to amplify glacial–interglacial climate changes. In some models, a glacial inception does not occur if vegetation is set to interglacial values. 5. Ice melting (Termination) and ice accumulation (glacial inception), primarily controlled by 100 000-year insolation changes, occurred in the high latitudes of the northern hemisphere and were synchronous by definition. What did not occur simultaneously was the response of the different regions in the world to these insolation changes. The southern hemisphere and low-latitude regions

609

appear to respond to orbital forcing earlier than the north. MIS 5 1. The last interglacial MIS 5 began earlier than 135 kyr in the low latitudes with a humid phase in Equatorial Africa, i.e. at a time when the seasonal contrast between spring and fall insolation was at a maximum, which is a condition favourable for strong El Nin˜o, which apparently increased suddenly in strength at the end of the MIS 6 ice age. 2. Sea level began rising also around 135 kyr, leading to Termination II with the main transgression during the time of strong increase of northern summer insolation. The first well-dated evidence of interglacial conditions in central Europe come from speleothems in the Spannagel cave at 130 kyr, which implicates that the mean annual temperature in the Alps were on a modern level well before the northern insolation maximum at 127 kyr. This is consistent with speleothem data from Italy, where the onset of the last interglacial has recently been determined at 129 1 kyr. This date should be the beginning of the Eemian sensu stricto on the continent, but this date is in conflict with a marine age model which places the beginning of the Eemian sensu stricto at 126 kyr 2, established on botanical grounds in a marine core off Portugal. 3. The Eemian, sensu strictu (Zagwijn, 1961), is characterized by a typical succession of pollen zones (E1/E2: Betula– Pinus forest, E3: Quercetum mixtum, E4: Quercus–Corylus forest, E5: Carpinus forest, E6: Picea–Abies forest), which can be correlated across Europe and has a duration of about 10 000 years according to varve counts in northern Germany and in the Eifel. 4. A direct correlation between pollen and marine proxies conducted on Iberian margin deep-sea cores indicates that a

610

Frank Sirocko et al.

time lag exists between ice volume decay/growth and forest development in Europe during the last interglacial sensu strict,: (1) the minimum in ice volume is reached at 128 kyr, 2000 years before the onset of the Eemian forest. (2) The substantial accumulation of ice in high northern latitudes (MIS5e/MIS5d transition) occurs at 115 kyr, i.e. 5000 years before the demise of the temperate and conifer forests in Iberia, France and the Eifel. The 11 000-year duration of the Eemian in northern European latitudes above 50 N suggests that tundra vegetation expanded during the first 5000 years of MIS 5d when forest still occupied southern latitudes. 5. Earth models of intermediate complexity (EMIC) indicate that the shift in Arctic tree line played a major role in the last glacial inception, mainly via the albedo changes around 122–120 kyr. Pollen data are in agreement with this prediction as they identify a replacement of the temperate forest by conifers at 52 N as early as 120 kyr. This suggests a southward displacement of the boreal vegetation belt of around 10 by comparison to its location (60 N–70 N) at the beginning of the Eemian. 6. There are no indications of abrupt cooling during the Eemian sensu stricto. Air temperature reconstructions for central Europe based on botanical fossils reveal a slight summer cooling (2 C/10 kyr) and a little stronger winter cooling (3 C/ 10 kyr) for the Eemian sensu stricto. However, climatic reconstructions from a new pollen dataset of La Gande Pile using an inverse modelling procedure with 13 C to constrain pollen data-derived estimations would identify substantial climatic changes within the Eemian. 7. Stalagmite growth at Spannagel terminates at 118–119 kyr, which leaves us with a time of about 10 000–12 000 years at which annual temperature in the Alps were quite similar to modern conditions.

8. About 118 kyr is identified by several model experiments with ice-sheet dynamics as the time when snow cover in North America became perennial and the ice sheet over North America began to expand significantly. The cause for the perennial snow and expansion of ice sheets is most likely a threshold in the insolation regime, with summer insolation being no more sufficient to melt the snow in the northern Arctic of North America. The further increase in ice-sheet extent is caused by a positive feedback, i.e. further cooling over the snow and ice with a high albedo and growing ice sheets with higher and cooler surfaces. 9. During the entire length of MIS5 (132–74 kyr), seven episodes of ice rafting, C19 to C25, occurred in the midlatitude North Atlantic related not only to periods of maximum ice volume but also to episodes of ice-sheet growth and decay. The beginning of the substantial ice-sheet growth, MIS5e/MIS5d transition (115 kyr), is associated with a 2–3 C reduction in surface water temperature and labelled C26 event, but no ice-rafted detritus have been recorded. 10. The Laurentian ice sheet became first unstable at around 112 kyr (C25 event), and the surface water of the mid-North Atlantic experienced a widespread substantial cooling by at 6-7 C. 11. Major icebergs surged for the first time not before 110 kyr (C24), i.e. after a growth period of 8000 years, which would be enough time to reach a height of more than 1 km. This pronounced cold event coincides, based on the direct correlation between marine proxies and pollen, with the Me´lisey I steppic period on land. 12. Sea-level drop was about 50 m between the last glacial inception around 118 kyr and C24, corroborating the build-up of a massive ice sheet. 13. CO2 was still at an interglacial level during these first seven millennia of

Chronology and Climate Forcing

14.

15.

16.

17.

the new ice age (118–111 kyr), indicating that the global carbon cycle responded to changing insolation later than the atmosphere or even later than the icesheets in their initial growth phase. Vegetation around the Mediterranean, France and into the Eifel region showed a continuation of almost interglacial conditions from 118 to 111 kyr. Vegetation in northern Europe at this time was already cold adapted. The major vegetation change in the Mediterranean and also in the Eifel region occurred during C24, when the late interglacial forest changed abruptly to tundra in central Europe, to steppe in western Europe between 50 N and 40 N and to a semidesert in the Mediterranean region. Severe cold and aridity is most likely the effect of severe SST lowering during C24 over the North Atlantic, caused by a spread of icebergs and a cold meltwater lid that inhibited deep-water formation and advection of warm waters from the Carribean via the North Atlantic drift (Gulf Stream). The final successive minor cold episode of MIS 5 was the C21 event (86 kyr), Me´lisey II on land, which was also characterized by massive ice discharges. C24 and C21 were coincident with the most extreme glacial events MIS 5d and MIS 5b, respectively. Icebergs did not reach the southern latitudes below 40 N during any of these seven events.

MIS 7 1. Similar to MIS 5, a direct link between terrestrial pollen and marine benthic/ planktic foraminifera as well as oxygen isotopes has been obtained only in W Iberian margin deep-sea cores for MIS 7 and for older interglacial stages. In this region, the first warm period during MIS 7 was only half the length of the Eemian and does not appear substantially warmer than the successive ones occurring during

611

MIS 7.3 and MIS 7.1. Generally, the magnitude of the different climatic changes of MIS 7 is large and mirrored by the volume of ice in concert with the large amplitude of insolation variations which characterizes this stage. MIS 7.4 shows similar climatic conditions as the previous glacial: icecaps were particularly developed and related with the coldest and driest climate on the continent and the lowest sea surface temperatures of MIS 7. 2. The equivalent of MIS 7 in continental European is well developed in southern European long sequences from the Velay region (Massif Central, France), Valle di Castiglione (Italy) and Tenagi Phillipon (Greece). This stage is characterized by three warm phases interrupted by stadials. It is interesting to note that none of these warm intervals in the Massif Central (named as Bouchet interstadials 1–3) reach the climatic state of a fully developed interglacial (such as in MIS 5e or 9e and 11c), while they did in sequences from Greece and Italy. 3. There is no direct link between the northcentral European terrestrial records and the marine isotope stages. In addition, the correlation with long continental sequences in southern Europe is under debate caused by uncertainties of absolute dating older than the last interglacial. 4. There is some evidence in north-central Europe of at least one warm event that could be correlative with the warm intervals of MIS 7: the Do¨mnitz warm stage in northeastern Germany, the possibly synonymous Wacken warm stage in northwestern Germany and the Reinsdorf and Scho¨ningen warm stages. These warm phases are not separated by glacial sediments and are situated stratigraphically above the Holsteinian, and before the first Saalian ice advance. MIS 9 1. New 230Th/U datings based on peat deposits from the type section of the

612

2.

3.

4.

5.

6.

Frank Sirocko et al.

Holsteinian stage at Bossel (near Hamburg, Germany) indicate an age of about 310–330 kyr BP which would correspond to MIS 9 (and therefore to the Landos interglacial in the Massif Central). The INQUA Subcommission on European Quaternary Stratigraphy defined the lower boundary of the Holsteinian as the transition from subarctic (still late Elsterian) to boreal conditions and the upper boundary as the transition from boreal to subarctic (Saalian) conditions. The interglacial vegetation development reconstructed by palynological data is very similar throughout north-central Europe and begins with a pine-birch forest. The immigration of thermophilous trees including alder, oak, elm, lime, esh, yew and hazel occurred more or less simultaneously. The early expansion of spruce is remarkable. Hornbeam and fir immigrated during the course of the interglacial. Particulary characteristic of the Holsteinian Stage in north-central Europe is the appearance of Pterocarya and Azolla filiculuides. The first half of the Holsteinian is characterized by temperatures somewhat lower than today. In the second half, the reconstructed mean temperatures are higher than today, in particular the July temperature. In addition, the Holsteinian seems to be less stable than the present interglacial (Holocene) or the last interglacial (Eemian) with some intra-interglacial coolings. The magnitude of the main cooling in the midHolsteinian is reconstructed as approximately 5 C for January temperature. No great change is reconstructed for July temperature during this episode. The duration of the Holsteinian in central Europe is estimated as about 15 000– 16 000 years, based on varve counts at Munster-Breloh. This duration disagrees with the new evidence from the multiproxy study of Iberian margin records, allowing the establishment of the timing of botanical events through the benthic isotopic

chronology. The first forested phase of MIS 9 lasted 12 000 years in NW Iberia, France and Greece but only at around 3600 years in southwestern Iberia due to the occurrence of a dry cold event particularly affecting the forest in this region. MIS 11 and earlier interglacials 1. MIS 11 is an exceptional interglacial complex. Weak changes in insolation during this stage are surprisingly associated with substantial climatic and greenhouse gas variations. The duration of its first warm period, 30 000 years, is much longer than the succeeding interglacial periods. For instance, direct correlation between marine and terrestrial signals indicates that this warm period is twice as long as the Eemian in the Iberian Peninsula. 2. The continental equivalent of the MIS 11 in central Europe could be the Rhume interglacial documented at Bilshausen where 27 000 varve years have been counted. The pollen record shows that this interglacial stage was climatically unstable. Several fast climatic deteriorations are documented. 3. If this dating is correct, the consequence would be that the Cromerian Complex terminated with MIS 11. This would be in agreement with the 230Th/U datings of the Holsteinian-type section above the Elsterian stage, but contradicts the often-used correlation of MIS 11 to be the Holsteinian. 4. The ‘Cromerian Complex’ stage of the Netherlands is defined by the recognition of at least four warm temperate (Waardenburg, Westerhoven, Rosmalen and Noordbergum interglacials) and three (unnamed) cold substages indicating the climatic complexity of this time interval. 5. The Early–Middle Pleistocene boundary, the begin of the Cromerian Complex, is linked to the Brunhes–Matuyama palaeomagnetic boundary which has been recognized within the first Cromerian interglacial (Waardenburg) at 780 kyr (MIS 19).

Chronology and Climate Forcing

6. The above discussion shows that the correlation of interglacial pollen for all older interglacials than MIS 5 between northcentral Europe and southwestern Europe is still uncertiain because there are no established chronozones and few absolute dates, but also because of biogeographical reasons. Based on the long continental sequence from the Massif Central, a correlation of the Praclaux interglacial (MIS 11) and the Holsteinian appeared to establish, now several authors argue, on the basis of U/Th dates that the Holsteinian represents MIS 9, and for a synchronization between the last Cromerian interglacial (Bilshausen) with MIS 11 including the Praclaux interglacial. Open questions and recommendations for future research The majority of the open questions concern problems of dating. In the following, we will summarize the background for these problems and try to recommend steps towards solving these gaps in our knowledge on past interglacials. . U/Th dates from stalagmites match in general the chronology derived from orbitally tuned marine records; in detail, however, dates from terrestrial sites in central Europe show differences up to 4000 years for the beginning and end of the last interglacial in comparison with the marine records off Portugal tuned to the SPECMAP timescale and coral 230Th/U age. Whether this reflects uncertainties in the 230 Th/U dating or the SPECMAP chronology, or whether these offsets are related to time-transgressive behaviour of vegetation zones across Europe, cannot be finally decided yet. Thus, there is an urgent need for a precise correlation of speleothem dates to the marine chronology and terrestrial pollen sites. . There are many sites across Europe with pollen sequences of the Eemian type. Very few of these sites have been absolutely

613

dated with dates published. If there are other interglacials/interstadials with a pollen succession similar to the Eemian, correlation of records on biostratigraphical grounds can lead to misinterpretations. Thus, as many as possible interglacial pollen records should be dated by U/Th or advanced luminescence techniques to assure that the classical Eemian pollen sequence in central Europe is a pattern indeed typical only of MIS 5e. . Radiometric ages for past interglacial sediments have large analytical errors and cannot be precise enough to correlate any sediment section by absolute dating. Thus, we will always depend on biostratigraphy or event correlation. Event correlation is done at the moment mostly relative to North Atlantic cold events, but this works reliably only if the records are long enough and reach from MIS 5e at least into MIS5c and show the structure of the C21–C26 events clearly. We need independent marker layers, probably from tephra and dust, which allow to correlate terrestrial records independently from the marine and ice core chronology. . The time-transgressive development of vegetation succession and local/regional/overregional patterns across Europe complicate a detailed correlation of records by biostratigraphy and palynology only. The successful analysis of the terrestrial vegetation in marine records off Iberia was a big step forward for the land–ocean correlation in southern Europe. Comparable records to link the vegetation development of Central and Northern Europe to the North Atlantic SST are still missing. . There are at the moment only two varve counted records for the Eemian; these are from Northern Germany and the Eifel region. None of these records is completely varved and undisturbed. In Northern Germany, only about 4000 individual varve laminae have been counted, in the Eifel, 12 000 varves were counted. There is an urgent demand for cores totally

614

.

.

.

.

Frank Sirocko et al.

undisturbed and also varve counted from the late MIS 6 and well into MIS 5d. There is only one varve counted core for the Holsteinian (MIS 9 or MIS 11) and one varve counted core for the Rhume interglacial (MIS 11 or Cromerian). No varve counted record is available for MIS 7 or any older interglacial. U/Th dates from peat is a rather newly developed dating method, but distinguishing MIS9 (320 kyr) from MIS 11 (420 kyr) is at the limit of the applicability of this method, and it can be problematic to proove that a respective site is a closed system for uranium. There is general agreement that insolation is the prime forcing mechanism of past interglacials. The character and exact value of the threshold at which a glacial terminates and develops into an interglacial is, however, not determined yet. This is particularly difficult because the interaction between changes in insolation, icesheet stability, sea-level change, continental vegetation and albedo, atmospheric dust and greenhouse gas concentration is a complex process which lasts over several millennia with abrupt events intercalated. Final interpretation from data has not yet been possible, mainly because the records from different geoarchives (ice, land and ocean) cannot be plotted on a commonly agreed timescale for past interglacials. Accordingly, we have avoided to plot all results from this final synthesis into one big figure. At least the 4000-year offset between the marine and speleothem chronologies has to be explained before this can be tried with some faith for the last interglacial. The most reliable approach would probably be a detailed parallel study of cave speleothems and ocean terrace corals, both dated by the same U/Th technique, in comparison with Ar/Ar-dated and varve counted records from laminated lakes. The reconstruction of past sea level from corals is, however, hampered by the unknown isostatic changes of the

.

.

.

.

.

continents after the melting of the large glacial icesheets. This process was quantified for Scandinavia, but it is hardly understood for ocean islands, where the isostatic adjustment is caused by the changing sea level. Thus, glacio-hydro-isostatic modelling is an important target for linking the land and ocean records worldwide. Interpretation of palaeorecords by using climate models is still a problem, because climate models provide information on average over a large area, whereas most palaeorecords contain mainly local information. First attempts to bridge this gap in scale are underway, but it has not been feasible to apply these so-called downscaling or data nudging methods to long-term climate simulations. Most climate system models which include ice-sheet dynamics are able to reproduce a glacial inception. A termination of a glacial or the beginning of an interglacial has not yet been successfully simulated. Most climate system models show a strong sensitivity to shifts in vegetation zones, or biomes, in the sense that if biome dynamics is missing, a glacial inception does not occur or is greatly diminished in the model. The role of biome dynamics as an amplifier of glacial inception seems to be a robust result in climate system models, the role of ocean dynamics, in particular the meridional overturning circulation is less clear. Climate system models yield ambiguous results. Most climate system models do not yet include all components of the climate system. In particular, the biogeochemical cycles are often not explicitly simulated. Instead, concentration of greenhouse gases, mineral dust and other biogeochemical substances is prescribed from data in most models. Hence, the course of atmospheric CO2 concentrations during interglacials is still an unresolved riddle.

6.2. DATA APPENDIX (CD-ROM)

Index 100-kyr cycle, 29, 37, 42 100-kyr paradox, 30 40-kyr cycle, 530 23-kyr cycle, 42 14 C-method, 464 230 Th/U, 112–14, 392–5 Abies, 114, 218, 250, 261, 280, 281, 355, 360, 365, 368, 383, 391, 208, 407–409, 430, 432, 433, 435, 436, 437, 439, 605, 609 Abri, 178, 186 Abrupt climate change, 7, 277, 574 Acceleration scheme, 513, 530, 540, 542, 555 Acropora palmata, 122, 123, 130, 131, 132, 135, 136 Aeolian dust, 55–60, 599 Age model, 279, 327–30 Age determination, 163–8 Albedo, 17, 33–4, 37, 40, 41, 200, 201, 202, 283, 342, 548, 549, 551, 553, 554, 558, 564, 565, 567, 568, 570, 571, 573, 574, 585–7, 590–1, 608, 610, 614 Alces alces, 462, 466 Alkenones, 308–309 Allactaga jaculus – great jerboa, 175 Alluvial loess, 417, 418, 432 Alopex lagopus, 458, 465 Alpine chamois, 466 Alpine foreland, 281–4, 437, 438 Alpine ibex, 466 Alpine marmot, 465 Alpine stalagmite, 201 Alps, 471 Amargiers interstadial, 391, 438 Amersfoort, 197, 240, 241 Amino acid racemisation, 77, 85 Amphibia, 462 Amsterdam, 197, 241, 351 Ancestral Cave bear, 458 Annual irradiance, 18, 24, 25 Antarctic ice cores, 54, 56, 57, 63, 70, 362, 387, 599 Anthropogenic forcing, 9, 550 AOGCM, 33, 565 AO temperature signal, 522–3, 524 Apodemus cf. sylvaticus, 458 Apodemus flavicollis, 461 Apodemus sylvaticus, 461, 462 Arctic, 38, 236, 448, 471, 507, 508, 509, 512, 513, 518, 520, 528, 548, 555, 557, 566, 569, 578, 590, 607, 610 Arctic fox, 465 Arctic Oscillation, 68, 518, 528, 607 Argenterola, 85 Artefact, 5, 178, 180, 182, 183, 184, 185, 200, 365, 581, 601

Artemisia, 99, 257, 259, 261, 270, 279, 282, 283, 284, 360, 361, 404, 421, 430 Arvicola hunasensis, 461, 465 Arvicola terrestris, 458, 462 Arvicolides, 456 Asia, 21, 51, 53, 57, 324, 509, 528, 569 Assimilation mode, 496 Astronomical parameters, 13, 15, 16, 30, 200 Astronomical theory, 13, 29, 30, 541 Atlantic, 289 Atmospheric circulation, 53–6, 58, 59, 62, 66, 68, 324, 364, 507, 517, 528, 535, 540, 542, 607 Atmospheric dynamics, 527 Atmospheric teleconnections, 607 Aurochs, 182, 183 Australia, 81, 569, 571 Austria, 472 Aves, 461, 462 Badger, 466 Bahamas, 477 Balzi Rossi, 182 Bank vole, 465 Barbados, 119 Baume Moula-Guercy, 179, 186 Bear, 154, 179, 184, 186, 250, 399, 425, 456, 458, 460, 461 Bechstein’s bat, 465 Beech marten, 465 Belem, 364 Betula, 48, 99, 210, 218, 219, 223, 249, 256, 259, 270, 367, 391, 392, 409, 412, 413, 414, 438, 609 Betula-Pinus peak, 409, 410, 412, 413, 414 Bifurcation, 31, 585 Big game hunting, 173, 184, 187, 601 Bilshausen interglacial, 353, 434 Bilzingsleben, 437, 449 BIOME 4, 202 Biostratigraphy, 233, 613 Bis´nik Jaskinia, 464, 466 Bison priscus, 461, 462 Bispingen, 210, 245, 247, 250, 251 Bleaching, 48, 158, 159, 160, 167, 168, 600 Boreal, 21, 31, 99, 100, 103, 200, 240, 272, 316, 317, 354, 355, 360, 383, 414, 535, 571, 612 Boreal tree line, 553, 574, 586 Bos primigenius – aurochs, 425, 449, 450 Bossel, 109–11, 114 Boundary-value problem, 497 Bovinae, 458 Bro¨rup interstadial, 100, 112, 277, 282, 283, 342

616 Brown bear, 182 Brown long-eared bat, 465 Bru¨ckner, 13, 31, 387 Bubalus murrensis – water buffalo, 173, 449, 450 Burgtonna, 176, 181, 185, 449, 450 Burial customs, 96, 114, 185 Buxus sempervirens – box tree, 173 C21, 611, 613 C22, 283 C23, 208, 211, 215, 282 C24, 199, 200, 207, 211, 215, 223 C25, 200, 208, 221–3 C26, 216–21, 613 Caching, 184 Calendar definition, 588–9, 21–2 Calendar insolation, 19–21 Caloric insolation, 18 Caloric season, 18 Calcite, 146–7, 151, 479, 480, 482, 485, 486 Canadian Archipelago, 579, 590, 607 Canis lupus, 458, 459, 461, 462, 465 Canis mosbachensis, 461, 465 Caours, 186 Capra ibex, 461, 466 Capreolus capreolus – roe deer, 173, 425, 448, 450, 461, 466 Carbonates, 61, 94, 95, 160, 161, 175, 267, 268, 269, 271, 292, 296, 327, 393 Carpinus, 177, 178, 197, 218, 222, 248, 249, 250, 261 Carpinus betulus – European Hornbeam, 177, 261, 383, 556 Carpinus-phase, 245, 368, 430, 432, 507, 601 Cascais, 604 Castor fiber, 458, 462 Cave, 455, 471, 472–4 Cave bear, 456, 458, 460, 461 Cave hyena, 456 Cave lion, 456 Central Caribbean, 197 Central Europe, 93, 173, 209–10, 255, 267–70, 445, 455 Central Russia, 164–6 Cervus elaphus, 425, 450, 461, 462 Charcoal layer, 216 Cheremoshnik, 164–6 Chinese loess plateau, 57, 323, 324, 339, 341, 342, 571 Chiroptera, 465 Chlorines, 311 Circulation models, 499, 528 Clethrionomys glareolus, 461, 462 CLIMAP, 201, 500, 507 Climate change, 7–8, 609 Climate change trigger, 7–8 Climate gradients, 278 Climate modelling, 547, 548

Index Climate models CLIMBER, 574, 575, 576, 579 CLIMBER-GREMLINS, 573, 608 ECHAM, 528 ECHO-G, 501, 510, 518, 528, 530, 607 HOPE-G, 501 LLN 2-D NH, 547, 548 MoBidiC, 547, 548 SICOPOLIS, 608 Climate reconstructions, 47, 241–4 Climate system, 495 Climate, classical definition, 3, 589 Climate, wider definition, 3, 597 Climatic precession, 15, 18, 21, 22, 37, 558 CLIMBER, 574, 575, 579 CO2, 3, 6, 8, 41, 42, 48–49, 550 Coelodonta antiquitatis, 458, 462 Cold event, 207, 589–90 Collared lemming, 465 Colour logging, 309–10 Combe Grenal, 186 Common hamster, 465 Common mole, 465 Common shrew, 465 Common vole, 465 Composite map, 521, 524, 537 Conifers, 259, 610 Continental ice volume, 38, 549, 556 Conturines-Ho¨hle, 459 Cooling, 207, 227, 603, 607, 610 Coral, 119, 122–3 Corylus, 435, 436, 437, 507, 600, 609 Corylus avellana – hazelnut, 181 Crested porcupine, 465 Cricetus cricetus, 462, 465 Cricetus major, 461, 465 Croatia, 178 Crocuta crocuta spelaea, 456, 462, 466 Cuon alpinus, 461, 465 Cuon sp, 458, 465 13 C, 202, 233, 268, 271, 272, 297, 475, 476, 479, 485, 486, 601, 610 18 O, 40, 47, 132, 135, 197, 199, 201, 207, 210, 211, 293, 312, 361, 367 Dama dama – fallow deer, 173, 450 Dating, 45, 47, 93, 144–5, 157, 310–12 Deciduous Quercus, 364 Deductive model, 495 Deforestation, 6, 283, 284, 551 DEKLIM, xi, xii, xiii, 137, 163 Denmark, 70, 197, 209, 240, 590 Devils Hole, 199, 479 Dhole, 465 Diachroneity, 283 Diatom, 145–6, 147 Diatomaceous gyttja, 143, 145

Index Diatomite, 166–8 Dicrostonyx gulielmi, 458, 465 Divje Babe I, 461 DNA, 180 DO event, 211, 342 Dome C, 55, 58, 362, 382, 599 Do¨mnitz, 352 Dose rate, 158, 159, 160, 161, 163, 166, 600 Dosimetric age determination, 159 Do¨ttingen, 398 Drenthe, 175, 355, 418, 431, 432, 440, 449, 450 Dry maar, 151–3, 397 Du¨mpelmaar tephra, 215 Du¨rnten interstadial, 277, 284 Dust, 50, 53, 55–63 Early Neanderthals, 178–80, 185 Early Weichselian, 261–2, 438–9 Earth system, 6, 31, 33, 34, 81, 495, 500, 563, 564, 576, 598, 601 Earth system model, 33, 495, 500, 541, 563, 564, 576, 583 Eastern Europe, 255, 265, 266–73 Ebelmen, 30 Eccentricity, 14, 16, 18, 21, 22, 24, 30, 34, 37, 528, 530, 609 ECHAM, 528 ECHO-G, 501, 518, 528, 530, 607 Edible dormouse, 465 Eem – Eemian, 141, 151, 153, 163–4, 200–203, 231, 239, 240–1, 249–50, 255, 259, 265, 432, 438, 450–3 Eemian cooling, 201, 210 Eifel, 141, 219, 397 Eigelbach, 212, 217 Einhornho¨hle, 458, 464 El Atrun, 182 El Nin˜o, 305 Elea, 182 Elephantidae, 461, 466 Elk, 466 ELSA, 211–12 Elsterian ice age, 353, 393, 419 Eleuthera, 83 EMICs, 37 Empirical Orthogonal Function, 519 ENSO, 50, 89, 305, 306, 307, 316, 317, 318 ENSO model, 305, 316, 317, 318, 541, 603 Eolian sediment, 324, 335 EPICA, 30, 41, 42, 47, 48, 49, 54, 55, 70, 277, 599 Equilibrium line altitude, 472, 475 Equinox, 13, 15, 17, 18, 19, 21, 29, 30, 501, 589 Equus caballus, 458, 462, 466 Equus cf. germanicus, 462, 466 Equus sp. – horse, 450 ESR, 77, 87, 179, 394, 434, 437, 457 Estoril, 364 Eurasia, 38, 40, 201, 325, 521, 525 Eurasian beaver, 465

617 Europe, 351, 359, 445, 455 European lynx, 466 European suslik, 465 European transect, 273, 602 European wild cat, 466 Event detection, 220–226 External variability, 4 Fagus, 383, 411, 434, 437, 438 Fagus sylvatica – European Beech, 176 Fallow deer, 173, 175, 182, 183 Feedback, 3, 563 Feldhofer Grotte, 186 Feldspar, 158, 159, 327 Felis silvestris, 461, 462, 466 Fen peat, 599, 605 Fennoscandian ice sheet, 419, 584 Field vole, 465 Fire, 181, 421, 434 Fish, 182, 417 Flood, 98, 308, 315, 417, 446 Flowstone, 476–80 Fluvial discharge, 178, 326 Forced models, 31–3 Forced variability, 33, 43, 523, 541, 556 Forest, 259–63 Fossil fuel, 6, 597 Fraxinus, 259, 261, 280, 282, 283, 361, 404, 425 Free models, 30–1 Free variability, 33 French Massif Central, 200, 360, 391 Freshwater disturbance, 589–90 Fundamental laws, 495 Fu¨ramoos, 278, 285 Future climate, 239, 277 Future Glacial inception, 583 Ga´novce, 177, 178, 179 Gastropods, 333, 460 Gathering, 181 Gaxun Nur Basin, 325, 326, 327, 328, 339, 342, 343, 604 GCM, 21, 22, 500, 517, 518, 523, 564 Gelkenbach interstadials, 392 General Circulation Model, 499 Geochronology, 157 Germany, 212, 277, 408, 417 Giant deer, 466 Giant hamster, 465 GIS 25, 198, 201, 202 GISP2, 47 Glacial, 55–7, 323, 527, 573 Glacial inception, 499, 517, 527, 573, 584, 587, 589–90 Glacial–interglacial variability, 13, 78, 80 Glaciation, 13, 353, 547, 579 Glacier, 324, 339, 344, 486 Gobi desert, 323 Golden jackal, 465

618 Gossau, 98–9 Grain size, 213, 335–9 Gramineae, 283, 284 Grande Pile, 231 Greenland, 585–6 Greenland Ice, 573, 585 GREMLINS, 574, 575, 576, 579 Grey wolf, 465 GRIP, 47, 200, 201, 208, 215 Gro¨bern, 99, 101, 103, 244, 245, 246, 248, 249 Gro¨bern-Schmerz, 244, 245, 248, 251 Groß-Rohrheim, 108 Grotte Scladina, 457–8 Grotte Vaufrey, 182 GS25, 199 GS26, 200 Gulf Stream, 202, 290, 291 Gyttja, 145 Hamburg-Dockenhuden, 240, 354, 408, 419 Hamburg-Hummelsbu¨ttel, 409, 392 Hamburg-Isfeldstrae, 106–108 Hare, 465 Harting, 197, 240 Harz, 458 Hawaii, 467 Hearth, 183, 186 Helicigona (Drobacia) banatica, 177, 178 Heliophytes, 256, 282, 407 Herdengelho¨hle, 460 Hetendorf, 240, 244, 245, 246, 248, 252 Hippopotamus amphibius – hippopotamus, 108, 163, 446, 448, 450, 606 HL2, 211–12 Hoher List, 142, 210, 211 Holocene, 63–6, 141, 153–4, 438, 453 Holocene climate simulation, 76, 528, 530, 627 Holsteinian Interglacial, 388–91 Hominid, 176–8 Homo Erectus, 417, 421, 435 Homo neanderthalensis, 462, 466 HOPE-G, 501 Hoˆrka-Ondrej, 178 Horse, 177, 183, 425, 430 Hoxnian Interglacial, 354, 391, 394, 434 Human impact, 145, 326, 432, 439, 550 Hystrix cf. vinogradovi, 176, 461 Hystrix cristata, 458 Iberia, 379 Iberian margin, 375, 376 Ice volume, 552 Iceberg, 200, 223, 231, 232, 293, 299, 497, 548, 589, 603, 610, 611 Icelandic Low, 528, 531, 532, 535, 540, 542, 607 Indian dhole, 465 Indicator species method, 241

Index Infrared Radiofluorescence (IR-RF), 162, 163, 166, 600 Infrared stimulated luminescence (IRSL), 158 Initial-value problem, 497 Inland ice, 324, 510, 583, 584, 585, 586, 590, 591 Insolation, 13, 17, 19 Interglacial, 379, 387, 388, 410, 417, 434, 445, 448, 499, 527, 547, 551, 583, 597 Interglacial sediments, 157, 163–8 Interglacial variability, 289–301 Internal variability, 6–7 Interstadial, 343, 432, 433, 434–8 Inverse modeling, 233, 601, 610 IRD, 199, 208, 219, 221, 226, 231, 232, 236 Irradiance, 16–17 Irradiation, 18–19 Italy, 70, 85, 135, 142, 284, 611 Jagonas interstadial, 392, 438 Juniperus, 99, 282, 360, 404, 419 Ka¨rlich interglacial, 353, 392, 434, 448 Kebara Cave, 186 Kepler’s law, 18, 531 Kieselgur, 166–8 Klinge, 163–4, 270 Ko¨nigsaue, 174 Krapina, 175, 178, 179, 182, 185, 186, 462 Laachersee Tephra, 142 La Grande Pile, 199, 202, 233, 234, 245 Lac du Bouchet, 360, 365, 367, 368, 369 Lacustrine basin/lake basin, 174–5 Lagurus lagurus – steppe lemming, 175, 458, 465 Lake level, 141, 142, 151, 216, 323, 326, 332, 333, 335, 336, 337, 338, 407 Lake sediment, 141 Laminae, 147, 211, 613 Lance, 181, 182 Land use, 6, 154 Land-ocean correlation, 359–71 Landos Interglacial, 387–95 Land–sea correlation, 375–84 Landscape evolution, 255 Last Glacial Inception, 517, 573, 587, 589 Last glacial-interglacial cycle, 38 Last Interglacial, 284, 289, 527, 551, 563 Late Eemian aridity pulse, 141, 176, 316, 601 Late glacial, 256–9, 438–9 Laurentide ice sheet, 584 LEAP, 216–21 Lehringen, 177, 178, 181, 182, 184 Lemmus lemmus, 458, 465 Leopard, 466 Lepus sp, 461 Les Echets, 281, 282 Levallois flakes, 180 LIGA members, 197, 201

Index Lignite, 99, 103, 112, 268, 270, 417, 599 Limnic sediments, 164–6 Linear regression, 41 Lisboa, 604 Lithics, 314 LLN 2-D NH, 547, 548 Loess, 57, 61, 215, 601 Longitude of the perihelion, 14, 566 Lonicera arborea – honeysuckle, 173 Lower Palaeolithic, 417, 425 Lower Pleistocene, 456 Luminescence, 157 Luminescence dating methods, 112 Lunz am See, 460 Lusitanian, 197 Lynx lynx, 461 MD01-2447, 377, 378 Maar, 141, 151, 397 Mafra, 364 Mammoth, 456 Mammuthus primigenius, 458, 465 Marillac, 186 Marine cores, 279, 283, 351, 609 Marine Isotope Stages, 94, 294, 361, 387, 388 Marine palynology, 112, 267 Marks Tey, 393, 394 Marmota marmota, 459, 461 Martes Martes, 461, 465 Mass balance, 574, 578, 579, 584 Massif Central, 355, 360, 361, 365, 368, 369, 391, 613 Mathematical models, 162, 495, 496 Mazovian interglacial, 390, 437 MD01-2443, 361–4, 367, 368, 369, 382, 604 MD03-2697, 377, 378 MD95-2042, 199, 232, 278, 279, 282, 329, 379 MD99-2331, 199, 377, 378, 379 Mediterranean, 445, 450, 453, 477, 479, 483, 485, 487, 549 Megaloceros giganteus, 450, 461, 466 Meles meles, 461, 465 Me´lisey I, 235 Me´lisey II, 198 Merck’s rhinoceros, 173, 182, 183, 184 Microlithozones, 147, 150–2 Microparticle, 208 Microtus agrestis, 459, 465 Microtus arvalis, 458, 465 Microtus arvalis/agrestis, 458 Microtus gregalis, 458, 465 Microtus nivalis, 459 Microtus oeconomus, 458, 462, 465 Microtus sp, 458, 465 Middle Palaeolithic, 174, 181, 185, 186 Middle Pleistocene, 349, 351, 359, 446–8 Mid-month insolation, 19 Milankovitch, 547

619 MIS 1, 583 MIS 2, 60, 99, 313, 329 MIS 3, 98–9 MIS 4, 51, 99, 112, 284, 313, 381, 456 MIS 5, 51, 86, 195, 583 MIS 5a, 99–101, 283–4 MIS 5b, 283 MIS 5c, 101, 282–3 MIS 5d, 281–2 MIS 5e, 173, 279–81, 438–9 MIS 6, 369, 486–8 MIS 7, 84–6, 108–109, 351, 483, 583 MIS 8, 367, 455 MIS 9, 84, 86–7, 109, 365–7, 583 MIS 10, 42, 88, 434, 556 MIS 11, 84, 87, 88, 556, 583 MIS 15, 173 MIS 17, 88 MIS 19, 88, 351, 353, 612 Mißaue, 417, 432, 433, 440 Mittelsteirischer Karst, 460 MoBidiC, 38, 547, 548, 554 Modern analogue, 242, 250, 292 Moisture transport, 324, 341, 506, 511, 574, 579 Mollusc, 197, 341, 342, 394, 435 Monsoon, 324, 341, 342, 343, 344, 563, 564, 568, 569, 571, 604 Montaigu event, 282, 283 Monticchio, 284 Moose, 466 Moraine, 106, 175, 266, 408, 450, 472 Mortality profile, 182 Munster-Breloh, 240, 244, 245, 247, 248, 249, 251, 252, 354, 419, 612 Murphy, 13 Mustela cf. Eversmanni, 462 Mustela foina, 458 Mustela putorius, 458, 462, 465 Mutual climatic range method, 242 Myotis bechsteini, 461, 465 Myoxus glis, 462 NAO, 501, 518, 528 Nar Valley, 392 Narrative models, 495 Narrow-headed vole, 465 Neanderthal, 51 Neodymium (Nd) isotopes, 60 Neogloboquadrina pachyderma, 198, 199, 231, 292, 293, 296–8, 361, 363, 378, 379, 380, 381 Neumark-Nord, 174, 175, 177, 181, 183, 265–7, 268–70, 273, 450 NGRIP, 201, 242, 265–8, 574, 585, 589, 590, 603 Nordic Seas, 236, 284, 541, 589, 590 Norfolk, 114, 392, 599 North America, 573 North Atlantic, 208

620 North Atlantic Oscillation, 300, 500, 501, 507, 518, 527, 528 Northern Hemisphere, 573 Northern Pacific, 199 Northern water vole, 465 NorthGRIP, 207 Northwestern China, 324, 341 North-western Iberian margin, 197, 375, 377, 383, 604 North-western Iberian Peninsula, 223, 377, 382 Norway lemming, 465 OAGCM, 499–501 Oak-mixed forests, 280 Obliquity, 15, 554 Ocean circulation, 5, 8, 37, 57, 76, 503, 517, 530, 540, 555, 556, 564 Ocean-atmosphere general circulation model, 499, 607 Odderade interstadial, 100, 101, 267, 283, 342 ODP, 87, 202, 289, 290, 291, 306, 310, 312, 318, 381, 479 Ognon interstadial, 284 Optically stimulated luminescence, 157, 158, 162, 600 Orbital changes, 5, 285 Orbital forcing, 527, 528 Orbital parameter, 588 Orography, 587 OSL, 157, 600 Oxygen isotopes, 309 Pacific, 40, 54, 57, 80, 86, 199, 202, 306, 307, 311–12, 314, 315, 317, 318, 326, 503, 512, 520, 527, 528, 538, 540–2, 566, 574 PAGES, xii, xiii, 627–9 Palaeoloxodon antiquus – straight tusked elephant, 173, 183, 425 Palynology, 112, 267, 352, 397, 613 Palynostratigraphy, 240 Panthera leo spelaea, 446, 458, 459, 461, 466, 467 Panthera pardus, 458, 461, 462, 466 Pdf/pdfs, 202, 239, 240, 242–4, 249 Peat, 93 Pebbles, 184–5, 460 Penck, 13, 31, 387 Penultimate deglaciation, 199, 202, 203, 471, 488, 601 Penultimate Interglacial, 113, 182, 231, 234, 471, 474, 483, 486, 488, 606 Perennial snow cover, 500, 511, 512, 587, 607 Perihelion, 5, 13–16, 18, 22, 316, 318, 501–503, 529–30, 566, 583, 589, 603, 608 Peru, 305 Physical models, 495 Picea sp. spruce, 103, 108, 177, 216–18, 234, 245–6, 250, 257, 262, 272, 280–4, 360, 365, 367–8, 391–3, 404, 407, 408, 410–14, 419, 424, 425, 430, 432, 433, 438, 439, 609 Pine marten, 466 Pine vole, 465

Index Pinus, 48, 103, 114, 210, 218, 219, 222, 223, 262, 270, 272, 280–5, 360–1, 364–8, 370, 378, 388, 391–3, 397, 407, 410–14, 419, 421, 424–5, 430, 432–3, 435–6, 438–9 Pitymys subterraneus, 458 Plant food, 181, 187 Plant migration, 282 Plant refugia, 252, 279 Plecotus auritus, 461, 465 Pleistocene, 349, 351, 359, 445, 446, 448 Pljos, 170 Poland, 209, 256, 257, 261, 352, 354, 387, 389, 390, 434, 437, 462, 555, 605 Polar Front, 57, 201, 236 Pollen, 213, 378, 397, 417 Pollen assemblage zone, 240–1, 246–9, 252, 267, 354, 387–9, 404, 408, 414 Porcupine, 176, 465 Portugal, 50, 210, 232, 279, 289, 362, 365, 367–70, 479 Praclaux interglacial, 353, 387, 391–2, 395, 434, 440, 605, 613 Precession, 15, 554 Precipitation, 231 Probability density functions, 202, 239, 240, 242 Prognostic mode, 496, 564, 606 Proxy data, 5, 30, 209, 239, 242, 244, 247, 299, 305–306, 312–15, 318, 472, 496, 505, 517–18, 522, 523–7, 541–3, 571 Pseudo proxy, 518 Pterocarya, 354, 391, 393, 407, 408, 410, 411, 419, 433, 437 Putorius sp., 458, 461, 462, 465 Quantitative climate reconstruction, 239, 280, 556 Quasi-deductive model, 34 Quartz, 60, 158, 159, 161–2, 164, 175, 207, 212–13, 215, 293, 335, 344, 600 Quaternary, 53, 55 Queluz, 364 Quercus, 197, 218, 248, 259, 261, 280, 282, 283, 360, 361–2, 364–71, 376, 382–3, 388, 391–3, 404, 407, 411, 413–14, 430, 450, 609 Rabutz, 177, 178, 181, 184 Radioactive decay, 95, 96, 98, 157, 160 Radioactive disequilibria, 157–8, 160–3, 165, 168, 600 Radioactivity, 158, 161, 166, 600 Radiofluorescence, 157, 158, 272, 600 Radionuclides, 158, 160–5, 167 RADIUS, 213 Ramesch-Knochenho¨hle, 459 Random motion, 7 Rangifer tarandus, 446, 448, 458, 461, 462, 466 Rapid aridification, 8 Ras Aamer, 182 Red deer, 182, 186–7, 425 Red fox, 465

Index Red Sea, 182, 500, 528, 542 Rederstall, 100, 210, 212, 214, 217–18, 223, 261, 267, 601 Reforestation, 279, 282–5, 438 Regourdou Cave, 186 Regression maps, 518 Reindeer, 466 Reinsdorf Interglacial, 114, 418, 421, 424–8, 430, 434–40, 605 Repolustho¨hle, 460, 466 Rhumian interglacial, 391, 392, 440 Roe deer, 173, 466 Rupicapra rupicapra, 458, 461, 462, 466 Saalian glacial, 261, 266, 269, 418, 432, 450 Saccopastore, 174, 179 Sahara, 7, 344, 506, 563, 568, 569, 571 Saint Germain the Vaux/Port Racine, 182 Saint-Sauveur, 178 Salix sp. – willow, 419, 430 Samerberg site, 278–83, 434 Scandinavian, 236, 286 Schalkenmehrener Maar, 141, 143, 145, 149 Scho¨ningen, 418 Scho¨ningen Interglacial, 431 Schwabenreith-Ho¨hle, 460, 463–4 Sclayn, 457 Scotland–Iceland Ridge, 284 Sea ice, 375, 552 Sea level, 75, 78, 79, 81, 84, 88, 119, 133 Sea level pressure, 504, 506, 508, 518, 519, 528, 531, 534, 539 Sea surface temperatures, 21, 47, 200–203, 282, 289, 298, 306, 313, 359–60, 365–9, 371, 381, 477, 500, 540, 549, 552, 576, 586, 599 Seasonal cycle, 16, 18, 37, 499, 503, 548, 567, 573, 576, 581, 589, 608 Seasonal irradiance, 18, 19, 25 Secular variability, 68 Secular variations, 15, 68 Sediment accumulation rate, 313, 314, 329–30, 362 Sedimentation rate, 142, 160, 212, 216–17, 231, 245–6, 283, 291, 294, 307, 313, 315, 317, 365, 404, 408, 414 Selaginella, 258, 261, 284 Settlement history, 154, 176–7, 187, 383 SICOPOLIS, 608 Simulation, 21, 499, 501, 555, 563, 587 Snow accumulation, 499, 509, 511–13 Snow albedo feedback, 574, 578, 585, 590, 598 Snow vole, 465 Solar constant, 14, 566 Solar energy flux, 4, 5, 8 Solar forcing, 5, 78, 338, 523 Solid state dosimetry, 157 Solstice, 14–19, 22–3, 529–30 Sorex araneus, 459, 465 South America, 53, 55, 57, 61–3, 68, 509, 569

621 Southern Europe, 359 Southern Pacific, 202 Spannagel Cave, 50, 210–11, 215, 471–7, 479–81, 483, 486–8, 609 SPECMAP, 24, 38, 40–1, 47, 49, 135, 294, 312, 359, 387–8, 613 Spectrometry, 93, 94, 98, 100–101, 103, 108, 115, 123, 176, 327, 435, 457 Speleothem, 471, 475 Spermophilus cf. citellus, 461, 465 Spermophilus sp., 465 SST, 207, 282, 289, 313, 477, 500, 549, 586, 599 St. Germain 1 interstadial, 231, 279, 282 St. Germain 2 interstadial, 279, 283 Stable isotopes, 55, 57–8, 66, 94, 135, 261, 273, 289, 290, 387, 471, 475–8, 480–3, 485 Stadial, 433, 434 Stalagmite, 481 Statistical uncertainty, 3, 8, 518 Stephanorhinus kirchbergensis – Merck’s rhinoceros, 173, 182, 425, 449, 450, 453, 462, 466 Steppe biome, 277, 279, 282–5 Steppe lemming, 175, 465 Steppe polecat, 465 Steppe wisent, 466 Stone marten, 465 Strontium (Sr) isotopes, 53, 61, 70, 599 Stuttgart-Untertu¨rkheim, 175, 178, 184 Subgrid-scale, 495–7 Subsistence, 181 Subtropical gyre, 202, 289, 297–8, 300 Succession, 150, 383 Summer temperatures, 207, 217, 242, 262–3, 280–1, 298, 507, 508–10, 553–5, 574–5, 578 Sus scrofa – wild boar, 108, 173, 446, 448–50, 453, 458, 461, 462, 466 Suslik, 465 Synergistic effect, 586 Talpa europaea, 458, 459, 461, 465 Tata, 176, 178 Taubach, 178–9, 182, 184, 186, 447, 449, 450 Tautavel wolf, 465 Taxus, 106, 178, 248, 261, 280, 388–91, 407–12, 414 Taxus baccata – yew, 260–1, 263, 425, 602 Tectonic forcing, 5, 31 Temperature, 231, 239 Temperature teleconnections, 523, 524, 526 Tenaghi Philippon, 365, 367–70, 558 Termination II, 50, 58, 89, 299, 313, 315, 486–8, 609 Terrace stratigraphy, 175, 179 Terrestrial ecosystems, 278 Th/U, 392 Theory of Ice Ages, 30 Thermomer, 388 Thermophilous taxa, 247–50, 285, 404, 407, 413–14 Thin section, 144, 146–7, 210, 213, 327, 332, 399, 475

622 Threshold models, 32 Tilt, 15, 527, 529, 589 TIMS, 93, 94, 123, 435, 459 TL, 100, 158, 421 Totes Gebirge, 459 Tottenhill Quarry, 114–15, 392, 395, 599 Transfer functions, 239, 241–2, 293, 309 Transient model simulation, 584, 587 Travertine, 175 Tundra, 40, 41, 207, 217, 408, 554, 601 Tundra biome, 200, 207, 218, 223, 226, 227, 273, 277, 284, 432, 435, 554, 571, 586, 610 Tundra vole, 465 Tundra-steppe, 277, 285, 419 Tuning, 29, 34, 38, 135, 142, 207, 210–11, 219, 279, 311, 312, 359, 387, 477, 496, 604 Turbidite, 143, 145–6, 152, 154, 212–13, 307 Turon, 197, 377, 378 Tyrol, 472 U/Th dating, 47, 48, 50, 77–8, 85, 87, 89 Ulmus, 198, 218, 259, 260, 261, 280, 282, 283, 389, 392, 404, 411 Ulmener Maar Tephra, 145 Uncertainty, 7, 77, 112, 134–6, 166, 243, 247, 496, 499 Unicorn Cave, 458 Unified theory, 33 Unstable Eemian, 141, 151, 153, 163, 173, 200, 231, 239, 240, 249, 255, 259, 265, 432, 438, 450, 499, 517, 585 Upper Pleistocene, 108, 112, 351, 456, 461, 464, 466, 606 Upscaling models, 518 Uranium series, 93, 199, 599 Uranium redistribution, 119, 131–3, 135–6, 599–600 Ursus arctos – brown bear, 182, 458, 461, 465 Ursus arctos priscus, 462 Ursus deningeri, 461, 465 Ursus sp., 425, 458, 465 Ursus spelaeus – cave bear, 458–9, 460–2, 465 Varve chronology, 142, 144, 145 Varve thickness, 148–50, 152–3 Varve types, 141 Varves, 145–7, 151, 244

Index Valle di Castiglione, 360, 361, 367, 368, 369 Vegetation phase, 177 Vegetation, 197, 255, 261, 277, 279, 349, 351, 563 Vegetation gradients, 284–6 Vienna, 464 Vindija, 461, 464 Vivianite, 144, 150 Volcanic ash, 61, 605 Volcanic forcing, 5–6 Volcanism, 141, 399 Vole, 446, 449, 450 Vosges, 199, 233 VOSTOK ice core, 41, 48, 54, 61, 66, 67, 68, 70, 392, 440, 575 Vulpes vulpes, 458, 461, 462, 465 Wacken, 355, 431, 435, 436, 437, 611 Wadi Haula, 182 Wallertheim, 178, 182, 183 Warsaw, 452, 464 Warthe glaciation, 175 Water vapour, 7, 9, 34, 37, 38, 40, 41, 324, 326, 342, 553, 554, 598, 608 Water vole, 465 Weasel, 465 Weichselian glacial, 106, 266 Weinbergho¨hlen bei Mauern, 456 Westerlies, 57, 323, 324, 326, 342, 343, 507, 519, 532, 504 Western polecat, 465 Wild boar, 173, 466 Wild horse, 466 Winter temperatures, 151, 202, 218, 247, 248, 250, 262, 263, 281, 282, 298, 300, 507, 508, 509, 533 Wood mouse, 465 Woodlands, 251, 261, 270, 273, 280, 282, 284, 285, 285, 376, 379, 602 Woolly mammoth, 466 Woolly rhinoceros, 456 Yellow-necked mouse, 465 Zeifen interstadial, 198 Zell, 101–106, 112

Juneau Icefield Research Project (1949-1958), Volume 8 (Developments in Quaternary Sciences) (Developments in Quaternary Sciences)

The Quaternary Period in the United States, Volume 1 (Developments in Quaternary Sciences)

Persistent Organic Pollutants in Asia, Volume 7: Sources, Distributions, Transport and Fate (Developments in Environmental Science) (Developments in Environmental Science)

The Climate of Past Interglacials, Volume 7 (Developments in Quaternary Science)

Juneau Icefield Research Project (1949-1958), Volume 8 (Developments in Quaternary Sciences) (Developments in Quaternary Sciences)

The Quaternary Period in the United States, Volume 1 (Developments in Quaternary Sciences)

Quaternary Dating Methods (Developments in Palaeontology and Stratigraphy 7)

Encyclopedia of Quaternary Science, Four-Volume Set

Ice Microdynamics (Developments in Quaternary Sciences)

Quaternary Glaciations - Extent and Chronology: A closer look (Developments in Quaternary Science 15)

Persistent Organic Pollutants in Asia, Volume 7: Sources, Distributions, Transport and Fate (Developments in Environmental Science) (Developments in Environmental Science)

Developments in Russian Politics 7

The Inquisition of Climate Science

The Late Cenozoic of Patagonia and Tierra del Fuego (Developments in Quaternary Sciences, Volume 11)

ADVANCES IN CATALYSIS VOLUME 7, Volume 7

Earth's Climate: Past and Future

Developments in sedimentology, Volume 19

Developments in Earth & Environmental Sciences, 7

Advancing the Science of Climate Change (America's Climate Choices)

Developments in British Politics 7: Seventh Edition (Bk.7)

Perspectives for Agronomy (Developments in Crop Science, Volume 25)

Well Completion Design, Volume 56 (Developments in Petroleum Science)

Developments in British Politics 7: Seventh Edition (Bk. 7)

Developments in Marine Geology, Volume 5: Quaternary Coral Reef Systems (History, development processes and controlling factors)

Volcanoes in the Quaternary, Volume 5,Nummer 161

PROGRESS IN MEDICINAL CHEMISTRY 7, Volume 7

Advances in Quaternary Entomology

Flavour Science: Recent Developments

Advances in Business and Management Forecasting, Volume 7 volume 7

Environmental Toxicology: Current Developments (Environmental Topics Series, Volume 7)

Peatlands, Volume 9: Evolution and Records of Environmental and Climate Changes (Developments in Earth Surface Processes)

Advances in Computers, Volume 7

Advances in Agronomy, Volume 7

Palaeoecology of quaternary drylands

The Climate of Past Interglacials, Volume 7 (Developments in Quaternary Science)

Juneau Icefield Research Project (1949-1958), Volume 8 (Developments in Quaternary Sciences) (Developments in Quaternary Sciences)

The Quaternary Period in the United States, Volume 1 (Developments in Quaternary Sciences)

Quaternary Dating Methods (Developments in Palaeontology and Stratigraphy 7)

Encyclopedia of Quaternary Science, Four-Volume Set

Ice Microdynamics (Developments in Quaternary Sciences)

Quaternary Glaciations - Extent and Chronology: A closer look (Developments in Quaternary Science 15)

Persistent Organic Pollutants in Asia, Volume 7: Sources, Distributions, Transport and Fate (Developments in Environmental Science) (Developments in Environmental Science)

Developments in Russian Politics 7

The Inquisition of Climate Science

The Late Cenozoic of Patagonia and Tierra del Fuego (Developments in Quaternary Sciences, Volume 11)

ADVANCES IN CATALYSIS VOLUME 7, Volume 7

Earth's Climate: Past and Future

Developments in sedimentology, Volume 19

Developments in Earth & Environmental Sciences, 7

Advancing the Science of Climate Change (America's Climate Choices)

Developments in British Politics 7: Seventh Edition (Bk.7)

Perspectives for Agronomy (Developments in Crop Science, Volume 25)

Well Completion Design, Volume 56 (Developments in Petroleum Science)

Developments in British Politics 7: Seventh Edition (Bk. 7)

Developments in Marine Geology, Volume 5: Quaternary Coral Reef Systems (History, development processes and controlling factors)

Volcanoes in the Quaternary, Volume 5,Nummer 161

PROGRESS IN MEDICINAL CHEMISTRY 7, Volume 7

Advances in Quaternary Entomology

Flavour Science: Recent Developments

Advances in Business and Management Forecasting, Volume 7 volume 7

Environmental Toxicology: Current Developments (Environmental Topics Series, Volume 7)

Peatlands, Volume 9: Evolution and Records of Environmental and Climate Changes (Developments in Earth Surface Processes)

Advances in Computers, Volume 7

Advances in Agronomy, Volume 7

Palaeoecology of quaternary drylands

Recommend Documents