Geology and Geoarchaeology of the Black Sea Region: Beyond the Flood Hypothesis (GSA Special Paper 473)

Geology and Geoarchaeology of the Black Sea Region: Beyond the Flood Hypothesis edited by Ilya V. Buynevich Department of Earth and Environmental Science Temple University 1901 N. 13th Street Philadelphia, Pennsylvania 19122, USA Valentina Yanko-Hombach Avalon Institute of Applied Science 976 Elgin Avenue Winnipeg, Manitoba R3E 1B4, Canada and Department of Physical and Marine Geology Odessa National I.I. Mechnikov University 2 Shampansky per. Odessa, 65058, Ukraine Allan S. Gilbert Department of Sociology and Anthropology Dealy Hall 401 Fordham University Bronx, New York 10458, USA Ronald E. Martin Department of Geological Sciences University of Delaware 103 Penny Hall Newark, Delaware 19716-2544, USA

Special Paper 473 3300 Penrose Place, P.O. Box 9140

Boulder, Colorado 80301-9140 USA

2011

Copyright © 2011, The Geological Society of America (GSA), Inc. All rights reserved. GSA grants permission to individual scientists to make unlimited photocopies of one or more items from this volume for noncommercial purposes advancing science or education, including classroom use. For permission to make photocopies of any item in this volume for other noncommercial, nonprofit purposes, contact The Geological Society of America. Written permission is required from GSA for all other forms of capture or reproduction of any item in the volume including, but not limited to, all types of electronic or digital scanning or other digital or manual transformation of articles or any portion thereof, such as abstracts, into computer-readable and/or transmittable form for personal or corporate use, either noncommercial or commercial, for-profit or otherwise. Send permission requests to GSA Copyright Permissions, 3300 Penrose Place, P.O. Box 9140, Boulder, Colorado 80301-9140, USA. GSA provides this and other forums for the presentation of diverse opinions and positions by scientists worldwide, regardless of their race, citizenship, gender, religion, or political viewpoint. Opinions presented in this publication do not reflect official positions of the Society. Copyright is not claimed on any material prepared wholly by government employees within the scope of their employment. Published by The Geological Society of America, Inc. 3300 Penrose Place, P.O. Box 9140, Boulder, Colorado 80301-9140, USA www.geosociety.org Printed in U.S.A. GSA Books Science Editors: Marion E. Bickford and Donald I. Siegel Library of Congress Cataloging-in-Publication Data Geology and geoarchaeology of the Black Sea Region : beyond the flood hypothesis / edited by Ilya V. Buynevich ... [et al.]. p. cm. -- (Special paper ; 473) Includes bibliographical references. ISBN 978-0-8137-2473-7 (pbk.) 1. Geology--Black Sea Region. 2. Archaeological geology--Black Sea Region. 3. Paleoclimatology-Holocene. I. Buynevich, Ilya V. (Ilya Val) QE350.22.B55G465 2011 554.9--dc22 2010046616 Cover: Satellite image of the Black Sea. NASA image courtesy of the MODIS Rapid Response Team (http://earthobservatory.nasa.gov/IOTD/view.php?id=8817). Inset, left: Eroding cliffs of Berezan Island, Ukraine, an important archaeological site along the northern Black Sea coast. Photo by I. Buynevich. Inset, center: ROV Hercules over a shipwreck with amphorae on the bottom of the Black Sea. Photo ©IFE/COE. Inset, right: Remnants of Tauric Chersonesos, an important Greek colony and port on the Crimean Peninsula. Photo by I. Buynevich.

10 9 8 7 6 5 4 3 2 1

Contents

Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . v V. Yanko-Hombach, A.S. Gilbert, I.V. Buynevich, and R.E. Martin 1. Surface runoff to the Black Sea from the East European Plain during Last Glaciation Maximum–Late Glacial time . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 A.Yu. Sidorchuk, A.V. Panin, and O.K. Borisova 2. Modeling extreme Black Sea and Caspian Sea levels of the past 21,000 years with general circulation models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27 A. Kislov and P. Toropov 3. Assessment of the Black Sea water-level fluctuations since the Last Glacial Maximum . . . . . . . 33 G. Lericolais, F. Guichard, C. Morigi, I. Popescu, C. Bulois, H. Gillet, and W.B.F. Ryan 4. Rapid Holocene sea-level and climate change in the Black Sea: An evaluation of the Balabanov sea-level curve . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51 R.E. Martin and V. Yanko-Hombach 5. Global climate change and sea-level fluctuations in the Black and Caspian Seas over the past 200 years . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59 E. Konikov and O. Likhodedova 6. Paleogeography of the Pontic Lowland and northwestern Black Sea shelf for the past 25 k.y. . . 71 E. Larchenkov and S. Kadurin 7. Nonpollen palynomorphs: Indicators of salinity and environmental change in the Caspian–Black Sea–Mediterranean corridor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89 P.J. Mudie, S.A.G. Leroy, F. Marret, N.P. Gerasimenko, S.E.A. Kholeif, T. Sapelko, and M. Filipova-Marinova 8. Climatic and environmental oscillations in southeastern Ukraine from 30 to 10 ka, inferred from pollen and lithopedology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117 N.P. Gerasimenko 9. Late Pleistocene and Holocene paleoenvironments of Crimea: Pollen, soils, geomorphology, and geoarchaeology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133 C.E. Cordova, N.P. Gerasimenko, P.H. Lehman, and A.A. Kliukin

iii

iv

Contents

10. Bedforms, coastal-trapped waves, and scour process observations from the continental shelf of the northern Black Sea . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165 A. Trembanis, S. Nebel, A. Skarke, D.F. Coleman, R.D. Ballard, A. Yankovsky, I.V. Buynevich, and S. Voronov 11. Archaeological oceanography and environmental characterization of shipwrecks in the Black Sea . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 179 M.L. Brennan, R.D. Ballard, K.L. Croff Bell, and D. Piechota 12. Pontic-Baltic pathways for invasive aquatic species: Geoarchaeological implications. . . . . . . . 189 I.V. Buynevich, A. Damušytė, A. Bitinas, S. Olenin, J. Mažeika, and R. Petrošius

Preface These opening words convey only a few essential matters. The present volume is one of a growing number of works focusing on the Black Sea, and as such, its relationship to previous research and its links to that of the near future beg some clarification, and thereby perspective. In addition, no interdisciplinary publication is ever achieved without help from a wide range of contributors whose part in the process deserves a public statement of deep appreciation. The Black Sea is the largest anoxic basin in the world, encompassing a total area of 423,000 km2. The basin is surrounded by Alpide fold belts and was formed in the Mesozoic as a backarc structure above the northward-subducting Tethyan oceanic lithosphere. The Black Sea consists of two large subbasins on the west and east that are separated by the NW-SE–trending Mid–Black Sea ridge. The western subbasin is floored by oceanic crust over which lie thick sediment units probably of Cretaceous and younger age; the eastern subbasin has a thinned continental or oceanic crust with a sediment cover less than 10 km thick. As a marginal basin, the Black Sea acts as a paleoenvironmental amplifier, recording climatic events in great detail. In response to sea-level changes driven by climatic cycles and/or regional tectonics, its connections with adjacent basins (the Marmara, Mediterranean, and Caspian Seas) have periodically been altered, leading to coastline migration and drastic modifications in environmental conditions (i.e., salinity, oxygen regime, basin morphology, hydrology), with dramatic consequences for the sedimentary, geochemical, and ecological systems, as well as human adaptive strategies. RENEWED SCIENTIFIC INTEREST IN THE BLACK SEA REGION Lately, this basin has witnessed a tremendous surge in interest due to (1) the Great Flood hypotheses that tied the Biblical Flood to the Black Sea (Ryan et al., 1997, 2003; Chepalyga 2003, 2007), (2) the presence of huge methane reserves contained within gas hydrates beneath the seafloor that may be exploitable as new nontraditional energy sources (Shnyukov and Ziborov, 2004), (3) the growing tangle of underwater infrastructure (e.g., gas pipelines and communication cables) laid across the Black Sea floor that is increasingly subject to geohazards from landslides, tectonics, and other dynamic forces, and (4) the presence of vast amounts of raw materials (e.g., sapropels) that have economic applications in agriculture (Shnyukov et al., 1999). This new outlook on the Black Sea has fostered a series of meetings, symposia, and workshops targeting issues in the geology, climatology, geochemistry, and archaeology of the Pontic basin. Three of them, held in 2003—(1) NATO Advanced Research Workshop “Climate Change and Coastline Migration,” 1–5 October 2003, in Bucharest, Romania; (2) international conference “The Black Sea Flood: Archaeological and Geological Evidence” sponsored by the Columbia University Seminar on the Ancient Near East, 18–20 October 2003, in New York, USA; and (3) Geological Society of America (GSA) Topical Session “‘Noah’s Flood’ and the Late Quaternary Geological and Archaeological History of the Black Sea and Adjacent Basins” presented at the Geological Society of America Annual Meeting on 4 November 2003, in Seattle, USA—led to the publication of a 1000-page volume entitled The Black Sea Flood Question: Changes in Coastline, Climate, and Human Settlement, which appeared under the Springer imprint (Yanko-Hombach et al., 2007a). This volume included 35 papers dealing with the geological, hydrological, climatological, archaeological, and linguistic aspects of the Black Sea flood hypotheses. Although no final answer to the Black Sea flood question appeared there, the book made great strides in enabling expanded dialogue between western and eastern scientists, encouraging new collaborations, and familiarizing western researchers with the extensive amount of information obtained by eastern scientists, data that had previously been inaccessible owing to the local languages in which they had originally been published. Subsequently, east-west collaboration continued to grow in the research programs of individual scientists as well as in international multidisciplinary projects, such as International Geological Correlation Programme (IGCP) 521 “The Black Sea–Mediterranean Corridor during the last 30 k.y.: Sea-level change and human adaptation” and International Union for Quaternary Research (INQUA) 501 “The Caspian–Black v

vi

Preface

Lithuania Denmark Russia Canada USA

UK Ireland France Romania Bulgaria

Ukraine

Egypt

Black Sea

Figure 1. International Geological Correlation Programme (IGCP) 521 and International Union for Quaternary Research (INQUA) project 501 participating countries (highlighted on the map): Algeria, Australia, Austria, Azerbaijan, Belgium, Bulgaria, Canada, Croatia, Egypt, Germany, Finland, France, FYR of Macedonia, Georgia, Greece, Ireland, Israel, Italy, Kazakhstan, Latvia, Lithuania, Moldova, Romania, Russian Federation, Spain, Switzerland, The Netherlands, Turkey, Ukraine, UK, United States of America (countries with contributors to this volume are listed on the map). Note the wide geographic distribution of scientists carrying out research in the Black Sea region.

Sea–Mediterranean Corridor during the last 30 k.y.: Sea-level change and human adaptive strategies” (www .avalon-institute.org/IGCP). Today, these projects involve the work of ~400 scientists, not only from the Black Sea region, but from around the world (Fig. 1). After the first three conferences in 2003, five plenary meetings were conducted under the framework of the IGCP 521–INQUA 501 projects from 2005 to 2009, and numerous topical sessions were presented at leading geological forums, such as the Annual Assembly of the European Geological Union in Vienna, Austria (2005, 2006); the Annual Meeting of the Geological Society of America in Denver, USA (2007); the XIIth INQUA Congress in Cairns, Australia (2007); and the 33rd International Geological Congress in Oslo, Norway (2008). In addition, many other smaller meetings examined in further detail the flood hypotheses but also addressed issues of regional climate, tectonics, coastline migration, human adaptive strategies, economic resources, and the future environmental stability of the region. More than 1000 authors have made contributions to IGCP 521–INQUA 501 meetings by presenting ~500 papers (Yanko-Hombach et al., 2005, 2006, 2007b; Gilbert and Yanko-Hombach, 2008, 2009). Many of these papers have been published or will be published in five IGCP 521–INQUA 501 thematic volumes of Quaternary International. THE DENVER CONFERENCE AND COMPILATION OF THE VOLUME The papers contained within this special GSA volume are the outgrowth of a successful technical session at the 2007 Geological Society of America Annual Meeting in Denver, Colorado. A large number of participants from Eastern Europe, funded by the GSA International Division, had the opportunity to present their recent findings, and their contributions are an integral part of the volume. The twelve papers were written by contributors from twelve countries (Fig. 1), and they address a range of topics, including climatic and hydrologic modeling, paleogeographic reconstruction of late Quaternary landscapes, palynology and paleoclimate reconstruction, and geoarchaeological studies, both onshore and offshore. We hope that the volume will serve as a timely reference for continuing research in a region harboring a number of newly independent states that are now faced with population pressure and a variety of environmental issues.

Preface

Each paper in the present book underwent a lengthy review process (three reviewers as a rule per paper) and both language and graphics editing. Acknowledgment must first be given for the financial assistance that made the conferences and book possible. We thank the International Union of Geological Sciences (IUGS), IGCP, INQUA, United Nations Educational, Scientific, and Cultural Organization (UNESCO), and GSA, which provided grant sponsorship to support many presenters at IGCP 521– INQUA 501 meetings. All transliterations of cited sources in Cyrillic follow Library of Congress style for both consistency and compatibility with the Online Computer Library Center’s World Catalogue, to maximize ease of location for the references in question. Grateful acknowledgment is offered for the thoughtful efforts of many external reviewers: Patrick Conaghan, Australia; Veselin Peychev, Bulgaria; John McAndrews, Petra Mudie (internal editorial help), Canada; K. Petersen, Denmark; Goran Georgievski, Jürgen Herget, Jens Matthiessen, Germany; Eliso Kvavadze, Georgia; Michel Fontugne, France; Daniella Basso, Italy; Tomasz Kalicki, Poland; Oya Algan, Mustafa Ergin, Namık Çağatay, Erdinç Yiğitbaş, Turkey; Doug Levin, Antonio Rodriguez, Shelley Wachsmann, USA, and a number of anonymous reviewers. Lastly, we thank the GSA book editor’s staff for, above all, their patience in awaiting the delivery of the finished manuscript. Valentina Yanko-Hombach Co-Leader of IGCP 521 and Leader of INQUA 501 Allan S. Gilbert Ilya V. Buynevich Ronald E. Martin

REFERENCES CITED Chepalyga, A.L., 2003, Late Glacial Great Flood in the Black Sea and Caspian Sea: Geological Society of America Abstracts with Programs, v. 35, no. 6, p. 460. Chepalyga, A.L., 2007, The Late Glacial Great Flood in the Ponto-Caspian basin, in Yanko-Hombach, V., Gilbert, A.S., Panin, N., and Dolukhanov, P., eds., The Black Sea Flood Question: Changes in Coastline, Climate and Human Settlement: Dordrecht, Springer, p. 119–148. Gilbert, A., and Yanko-Hombach, V., eds., 2008, Extended Abstracts of the Fourth Plenary Meeting and Field Trip of IGCP 521–INQUA 501 Project “Black Sea–Mediterranean Corridor during the Last 30 k.y.: Sea Level Change and Human Adaptation,” 4–16 October 200: Bucharest, Romania, National Institute of Marine Geology and Geoecology (GeoEcoMar), and Varna, Bulgaria, Department of Natural History of the Regional Historical Museum, 215 p., ISBN 978-973-0-06271-7. Gilbert, A., and Yanko-Hombach, V., eds., 2009, Extended Abstracts of the Fourth Plenary Meeting and Field Trip of IGCP 521–INQUA 501 Project “Black Sea–Mediterranean Corridor during the Last 30 k.y.: Sea Level Change and Human Adaptation,” 22–31 August 2009: Izmir, Turkey, Kadir Has University, Dokuz Eylül University, and Çanakkale, Turkey, Çanakkale Onsekiz Mart University, 213 p., ISBN 978-975-441-265-9. Ryan, W.B.F., Pitman, W.C., III, Major, C.O., Shimkus, K., Maskalenko, V., Jones, G.A., Dimitrov, P., Görür, N., Sakinç, M., and Yüce, H., 1997, An abrupt drowning of the Black Sea shelf: Marine Geology, v. 138, p. 119–126, doi: 10.1016/S0025-3227(97)00007-8. Ryan, W.B.F., Major, C.O., Lericolais, G., and Goldstein, S.L., 2003, Catastrophic flooding of the Black Sea: Annual Review of Earth and Planetary Sciences, v. 31, p. 525– 554, doi: 10.1146/annurev.earth.31.100901.141249.

Shnyukov, E., and Ziborov, A., 2004, Mineral’nie bogatstva Chernogo moria [Mineral Riches of the Black Sea]: Kiev, Department of Marine Geology and Mineral Resources of the Ukrainian Academy of Sciences. Shnyukov, E.F., Kleschenko, S.A., and Kukovskaya, T.S., 1999, Sapropelevie ili Chernogo moria—Novii vid mineral’nogo siriia [Sapropels of the Black Sea—New kind of raw materials]. Geologiia i poleznie iskopaemie Chernogo moria [Geology and Mineral Resources of the Black Sea]: Kiev, p. 399–411. Yanko-Hombach, V., Buynevich, I., Chivas, A., Gilbert, A., Martin, R., and Mudie, P., eds., 2005, Extended Abstracts of the First Plenary Meeting and Field Trip of IGCP 521 Project “Black Sea–Mediterranean Corridor during the Last 30 k.y.: Sea level Change and Human Adaptation,” 8–15 October 2005: Istanbul, Turkey, Kadir Has University, 226 p. Yanko-Hombach, V., Buynevich, I., Chivas, A., Gilbert, A., Martin, R., and Mudie, P., eds., 2006, Extended Abstracts of the Second Plenary Meeting and Field Trip of IGCP 521 Project “Black Sea–Mediterranean Corridor during the Last 30 k.y.: Sea Level Change and Human Adaptation,” 20–28 August 2006: Odessa, Ukraine, Odessa National University, 188 p., ISBN 966-318-554-6. Yanko-Hombach, V., Gilbert, A.S., Panin, N., and Dolukhanov, P.M., eds., 2007a, The Black Sea Flood Question: Changes in Coastline, Climate and Human Settlement: Dordrecht, The Netherlands, Springer, 971 p. Yanko-Hombach, V., Buynevich, I., Dolukhanov, P., Gilbert, A., Martin, R., McGann, M., and Mudie, P., eds., 2007b, Extended Abstracts of the Joint Plenary Meeting and Field Trip of IGCP 521 “Black Sea–Mediterranean Corridor during the Last 30 k.y.: Sea Level Change and Human Adaptation,” and IGCP 481 “Dating Caspian Sea Level Change,” 8–17 September 2007: Gelendzhik (Russia)-Kerch (Ukraine), Southern Branch of the Institute of Oceanology, Russian Academy of Sciences and Demetra Beneficent Foundation, 178 p., ISBN 978-5-85941-151-0, IGCP 521-481.

vii

The Geological Society of America Special Paper 473 2011

Surface runoff to the Black Sea from the East European Plain during Last Glacial Maximum–Late Glacial time Aleksey Yu. Sidorchuk Andrey V. Panin Geographical Faculty, Moscow State University, Vorob’evy Gory, Moscow 119991, Russia Olga K. Borisova Institute of Geography, Russian Academy of Sciences, Staromonetny per., Moscow 119017, Russia

ABSTRACT Hydromorphological and hydroclimatic methods were used to reconstruct the former surface runoff from the East European part of the Black Sea drainage basin. Data on the shape and dynamics of the last Fennoscandian ice sheet were used to calculate meltwater supply to the headwaters of the Dnieper River. The channel width and meander wavelength of well-preserved fragments of large paleochannels were measured at 51 locations in the Dnieper and Don River basins (East European Plain), which allowed reconstruction of the former surface runoff of the ancient rivers, as well as the total volume of flow into the Black Sea, using transform functions. Studies of the composition of fossil floras derived from radiocarbon-dated sediments of various origins and ages make it possible to locate their modern region analogues. These analogues provide climatic and hydrological indexes for the Late Pleniglacial and Late Glacial landscapes. Morphological, geological, geochronological, and palynological studies show that the landscape, climatic, and hydrologic history of the region included: (1) a cold and dry interval close to the Last Glacial Maximum characterized by high meteoritic surface runoff supplemented by meltwater flow from ice-dam lakes; (2) a warmer humid interval at the end of the Late Pleniglacial with very high surface runoff and formation of extremely large meandering channels, combined with a short event of substantial inflow from the Caspian Sea; and (3) a period from the Oldest Dryas to the Preboreal of nonsteady surface runoff decrease, and transformation of large meandering channels into smaller ones against the background of climate warming.

Sidorchuk, A.Yu., Panin, A.V., and Borisova, O.K., 2011, Surface runoff to the Black Sea from the East European Plain during Last Glacial Maximum–Late Glacial time, in Buynevich, I.V., Yanko-Hombach, V., Gilbert, A.S., and Martin, R.E., eds., Geology and Geoarchaeology of the Black Sea Region: Beyond the Flood Hypothesis: Geological Society of America Special Paper 473, p. 1–25, doi: 10.1130/2011.2473(01). For permission to copy, contact [email protected]. © 2011 The Geological Society of America. All rights reserved.

1

2

Sidorchuk et al.

INTRODUCTION General information on the variability of climate and water budget in Europe for the period from the Last Glacial Maximum (LGM), through Late Glacial time (LGT), and into the beginning of the Holocene (ca. 18–10 radiocarbon ka) is controversial. According to vegetation reconstructions based on palynological data (Grichuk, 1982), the climate is believed to have been both cold and dry (the so-called cryoxerotic stage of the glaciation). The climate of the southern part of the East European Plain drives changes in the water balance of southern sea and lake basins, and, thus, most workers correlate the last major drop in the level of the Caspian Sea and Black Sea to the LGM (Varuschenko et al., 1987; Winguth et al., 2000; Dolukhanov et al., 2008; and others). To explain such a dramatic drop in sea level, a substantial decrease in river runoff into the seas has been suggested. Varuschenko et al. (1987) estimated annual river runoff into the Caspian Sea during the LGM at only 20%–28% of its present value. Estimates of river runoff into the Caspian and Black Seas during the LGM based on atmospheric general circulation models comprise 55% and 59% of the modern values, respectively (Kislov and Toropov, 2006). On the other hand, extensive dating of the Caspian deposits undertaken in the last decade have revealed a Late Glacial age for the highest stage of the Khvalynian transgression (Svitoch and Yanina, 1997; Leonov et al., 2002; Chepalyga et al., 2008; and others), which suggests large surface runoff, at least from the Volga River basin. Kalinin et al. (1966) estimated river runoff into the Caspian Sea during the maximum stage of the Khvalynian transgression at 517 km3 per year, a figure that is 1.5 times higher than the present value. Similarly controversial reconstructions exist for the Black Sea drainage basin: some researchers propose a relatively high runoff and a continuous outflow from the Black Sea (Lane-Serff et al., 1997), while others believe that the runoff into the Black Sea was low due to the dry climate, and as a result, sea level dropped to –110 m (Ostrovsky, 1982; Aksu et al., 2002) or even to –150 m (Winguth et al., 2000) because of negative water budget during the LGM. It has also been suggested that a massive inflow of meltwater from the Fennoscandian ice sheet into the Black Sea took place after the LGM (Kvasov, 1979; Kroonenberg et al., 1997), both directly through the Dnieper River valley and as an outflow from the Caspian Sea through the Manych Straight. A series of meltwater pulses is suggested by isotopic depletion of the Black Sea waters between 18 and 15.5 ka (Bahr et al., 2006). Hydromorphological and hydroclimatic methods of paleogeographic reconstructions allow quantitative estimation of the former surface runoff originating from melting glaciers and from precipitation over a river basin. The ice volume of the Quaternary ice sheets can be reconstructed from their area using transform functions derived from recent glaciological information and theoretical considerations about ice rheology (Markov and Suetova, 1964; Khodakov, 1982; Peltier, 1994). Therefore, with information on the age of the boundaries of the last Fennoscandian ice

sheet, changes in its volume and meltwater supply into adjacent rivers can be estimated (Kalinin et al., 1966). Valuable information about past hydrological river regimes can be derived from the morphology of the former river channels, especially if the former topography differs significantly from that of the present. Morphological data on the large Late Glacial paleorivers, which give distinct evidence of high surface runoff, were first investigated by Dury (1964, 1965) in Western Europe and North America, and by Volkov (1960, 1963) in northern Kazakhstan and western Siberia. Similar results were obtained for several rivers in the Black Sea basin: for the Seim (Borisova et al., 2006) and Khoper (Sidorchuk and Borisova, 2000) Rivers in the Dnieper and Don basins; for the basin of the Danube River in Hungary (Borsy and Felegyhazi, 1983; Kasse et al., 2000), and in Romania (Howard et al., 2004). Morphological, textural, palynological, and geochronological studies have shown that the LGT in Europe was a period of very large, widely spread river channels, which presumably were formed by high and powerful surface runoff. Paleobotany plays an important role in providing data for paleoclimatic reconstructions. Late Glacial climatic events and the chronology of vegetation development in Europe have been derived mainly from palynological data later dated through radiocarbon and correlated with the isotopic “events” in the Greenland ice-core record (e.g., Walker et al., 1999). To reconstruct the hydroclimatic conditions that existed at various stages of the LGM and LGT, palynological studies of dated alluvium, lake, and peat sediments can be applied. The use of paleobotanic data for paleoclimatic and paleolandscape reconstruction implies that flora and vegetation are strongly influenced by changes in the natural environment and by the climate in particular (e.g., Iversen, 1944; Grichuk, 1969). This paper is aimed at (1) analysis of the paleoenvironmental conditions and causes for the paleohydrological changes in the East European part of the Black Sea basin during the LGM and LGT, (2) paleohydrological reconstruction of the surface runoff there since the Late Pleniglacial, and (3) discussion of the landscape and surface runoff changes in the remaining part of the Black Sea basin during the last ~20,000 yr. METHODS Paleofloristic Method of Paleolandscape and Hydroclimatic Reconstruction Climate reconstructions usually rely on either the comparison of fossil and modern pollen assemblages and their associated modern climate, or selected indicator plant species with specific climatic requirements. Paleobotanical data used for such reconstructions are of two main kinds: plant macrofossils (seeds, fruits, leaves, etc.), and pollen and spores. Macrofossils have the advantage of usually being identifiable to the species level. On the other hand, the occurrence of macrofossils is relatively restricted, and they usually belong to aquatic and subaquatic

Surface runoff to the Black Sea from the East European Plain plants. Pollen diagrams provide more detailed information on the history of specific plant taxa as well as vegetation on the whole. Because plants sensitive to climatic conditions are mainly medium to low pollen producers, sufficiently detailed pollen data are essential. In the process of pollen identification, the highest possible taxonomic resolution should be achieved to obtain the most complete results. This is possible if well-preserved pollen of arboreal plants, as well as pollen and spores of certain groups of herbaceous plants (e.g., Thalictrum, Lycopodium, Equisetum), can be identified to genus or even species levels. Iversen (1944) was the first to use the pollen of certain plant species to estimate paleotemperatures. This author established the relationship between present occurrences of Ilex and Hedera and summer as well as winter temperatures. Such relationships can be established by comparing the present boundaries of the plant’s geographical range (its “area”) and climatic data, i.e., the coincidence of “area” limits and certain isotherms. For example, Iversen showed that present-day ranges of Ilex and Hedera are limited by the 0 °C and –2 °C January isotherms, respectively. A well-known example of this kind is the coincidence of the tree line with the 10 °C July isotherm in many lowland and mountain areas. This approach has subsequently been extended by increasing the number of indicator species and climatic indexes considered (e.g., Hintikka, 1963; Zagwijn, 1996). Grichuk (1969) further developed Iversen’s approach into the mutual climatic range method, in which as many species as possible are used to obtain paleoclimatic estimates for a specific site. Their geographical distributions are first converted to climatic ranges, expressed as climatic range diagrams (climatograms), and then their mutual climatic field is determined. Climatograms have the advantage that even plant species without present-day geographical overlap can be used for paleoclimatic reconstructions. Moreover, since this approach is based on the requirements of individual species, finding modern analogues for the fossil pollen spectra or plant assemblages is not necessary. Grichuk (1969) suggested yet another method of climatic reconstruction based on paleofloristic data, elaborating on the idea of Szafer (1946–1947). This method consists of identifying a modern region where all the species of a fossil flora grow presently. The mutual area of all individual species of a certain fossil flora is found by overlapping their present-day “areas.” The “areas” of as many plants as possible should be used. Geographical analysis of the modern spatial distribution of all the plants of a certain fossil flora (compilation of a so-called arealogram) allows one to find the closest modern floristic analogue to the past vegetation at the site. By identifying this modern regional analogue, it is possible to determine the closest modern landscape and hydroclimatic environment to that of the fossil flora under study. Accuracy in these reconstructions depends on (1) the accuracy of the paleofloristic definitions based on detailed pollen analysis, (2) the richness of the resulting fossil floras, (3) the accuracy of the data on present-day geographical ranges of plants that represent components of the fossil floras, (4) the sizes of the regional analogues, and (5) the variability of the hydroclimatic

3

characteristics within these analogues. Furthermore, the accuracy of palynological analysis and the richness of a paleoflora depend not only on the palynologist’s personal experience, but also on the type of sediment, vegetation type, and the degree of pollen preservation. Late Glacial floras are usually relatively poor, so that not all analyzed samples provide sufficient data for paleofloristic reconstruction. Usually, conditions that are suitable for all the species of a given fossil flora can be found within a comparatively small area. The present-day features of plant communities and the main hydroclimatic indices of such a regional analogue would be close, in most cases, if not identical, to those that existed at the sampled site in the past. For example, fossil flora from the cultural layer at the Yudinovo Early Man site in the Desna River basin consists of 19 plant species identified using pollen analysis. At present, all of them grow within a small area in the Biya River basin, downstream from Teletskoye Lake (Altai Mountains). Therefore, the current climatic characteristics of this region should be similar to those of the Desna River basin at the time when the studied paleoflora existed (ca. 14–15 14C k.y. B.P.). Method of Estimating Ancient Continental Ice-Sheet Volume Analysis of recent continental and mountain glaciers shows that their shapes can be approximated by ellipses in both the horizontal and vertical planes (Kapitsa, 1958):

x2 y2 + = 1. b2 a 2

(1)

Here, a is the half-length of the short axis of the ellipse in a horizontal plane (in a vertical plane, it corresponds to the maximum height Hm of the glacier), and b is the half-length of a long axis (in a vertical plane, it is the distance from the glacier’s center to its border Lm along a given transect), while x and y are corresponding running coordinates from the ellipse’s center. The unit volume Vu of a glacier’s vertical transect (for the unit width of a glacier) is:

Vu =

π H m Lm ± ε . 4

(2)

Here, ε stands for the volume of initial relief of the glacier’s base, which is neglected in the following equations. An empirical relationship exists between the maximum height Hm (in km) of modern sheet glaciers and the length of a short-axis transect Lm in km (Khodakov, 1982):

H m = ks L m .

(3)

According to the recommendations of Khodakov (1982), the shape coefficient ks in this formula was 0.094 (with the scatter

4

Sidorchuk et al.

from 0.075 to 0.12) for the last Fennoscandian glacier at the “cold” stage of its advance. For the “warm” stage of glacier retreat, he recommends using a coefficient ks equal to 0.061 (0.048–0.077), which yields a flatter glacier. For the stage of glacier retreat, the “dead-ice” glacier model also can be used, with “active” glacier in the center and a marginal zone of melting “inactive” ice; this model follows from geomorphologic considerations (Chebotareva and Faustova, 1982; Faustova, 1984). Therefore, in a short-axis vertical transect for two dated positions of a glacier’s border with a lag Δt (yr) and with distances Lm1 and Lm2 (km), the unit rate of change in “cold” or “warm” glacier ice volume (in km2 yr–1) is:

ΔVu π (ks2 Lm2 − ks1Lm1 ) . = Δt 4Δt 32

32

(4)

For the “dead-ice” glacier model, this unit rate of volume change is:

Methods of River Paleodischarge Estimation The two main ways of calculating paleodischarge from river channel morphology and texture are by either hydraulic or regime equations, both first used by Dury (1965). The advantages and disadvantages of these methods are discussed in Sidorchuk et al. (2008). In the regime equations approach (used here), the relationships between channel morphology and flow hydrology must be known. Our investigation of the rivers on the Russian Plain and in western Siberia (Sidorchuk et al., 2001, 2008) shows that it is important to use a broad range of river sizes and river basin landscapes to work out the empirical formula that suits the purposes of paleohydrological reconstructions. We used ~450 sections of rivers in northern Eurasia with mean annual discharge Qa from 1 to 13,000 m3 s–1 and channel bankfull width Wb from 15 up to 3000 m, and the drainage basins were situated in a variety of landscapes from steppe to tundra. Based on these data, the relationship took the form: Qa = 0.012y0.73Wb1.36.

ΔVu = Δt

π (k L

32 s2 am2

−k L

4Δt

32 s1 am1

)

+ Hd

Ldm2 − Ldm1 . Δt

(5)

Here, Hd is “dead-ice” thickness, and indices “a” and “d” refer to “active” and “dead” (inactive) parts of a glacier. In the case of decrease in glacier volume, this unit rate would be negative and equal to the sum of the positive snow accumulation rate I on the glacier’s surface and the negative rate of meltwater drainage ΔWu/Δt from the glacier plus snowice evaporation rate E from the glacier’s surface. The unit rate of water drainage for a unit width of the glacier (meltwater runoff in km2 yr–1) would therefore be (here ice volumes are recalculated into water volumes with the coefficient 0.9):

ΔWu 0.9ΔVu = + I −E. Δt Δt

(6)

The last two terms in Equation 6 are difficult to estimate for Quaternary glaciers. Unit (for a unit width) snow accumulation rate I (in km2 yr–1) for recent sheet and mountain glaciers (Khodakov, 1982) is related to their size Lm (km):

I = ki L2m3 ,

(7)

⎛ L 2 3 + Lm2 2 3 ⎞ ΔWu 0.9ΔVu = + ki ⎜ m1 ⎟−E. Δt Δt 2 ⎝ ⎠

(8)

and

Coefficient ki for recent glaciers varies in a broad range from 0.001 to 0.005, and therefore its mean value of 0.00224 is recommended for use (Khodakov, 1982).

(9)

Parameter y is inversely related to the seasonal flow variability and represents the ratio between the mean annual discharge Qa and the mean maximum discharge Qmax: y = 100(Qa/Qmax).

(10)

The range of parameter y is from 4 to 5 for rivers with high seasonal flow variability to more than 20 for those with low seasonal flow variability, and up to 100 for rivers with stable flow, such as those draining large lakes. An increase in the flow variability (a decrease in y) generally causes an increase in floodplain height and flow concentration in a single channel with larger bankfull width. Flow variability depends on the basin area F (km2): y = aF 0 .125.

(11)

Parameter a in Equation 11 reflects the geographical distribution of climatic flow variability independent of river basin size. It can be calculated from measured mean annual discharge Qa, mean maximum discharge Qmax, and basin area F. Parameter a typically varies between 1.5 for river basins with high seasonal flow variability to over 4 for river basins with low variability. For paleolandscapes, parameter a for each paleochannel is estimated using recent fluvial analogues. It is then possible to calculate y for paleolandscapes with Equation 11, the mean annual discharge Qa from the paleochannel width with Equation 9, and the mean maximum discharge Qmax with Equation 10. All variables in Equations 9–11 can be obtained from maps and space images, as well as from Hydrological Service measurements. Estimation of coefficient a in Equation 11 for the past requires knowledge of this relationship for the former landscape or its modern hydroclimatic analogue. Geographic influences on river flow bring about similar hydrological regimes for rivers in

Surface runoff to the Black Sea from the East European Plain similar landscapes (Evstigneev, 1990). Geographic controls over river flow and their applications to paleohydrology lead to the principle of paleogeographical analogy (Sidorchuk and Borisova, 2000), which states that the hydrological regime of a paleoriver within a paleolandscape must have been similar to that of a present-day river within the same type of landscape. Therefore, the hydrological regimes of modern rivers in a certain type of landscape can be used to estimate the paleohydrological regime in the same type of paleolandscape. STUDY AREA General Characteristics of the Black Sea Drainage Basin within the East European Plain The Black Sea drainage basin covers the area of ~1,240,000 km2 and occupies ~40% of the East European Plain (Fig. 1). The basin is

5

drained by typical lowland rivers: 75% of the territory is situated at the elevations less than 250 m above sea level (asl), and only 2% of the territory is at elevations above 500 m (in the south and southeast, near the Crimean and Caucasian Mountains). The total length of the permanent drainage net of 72,500 rivers in the East European part of the Black Sea basin is 344,258 km (Domanitskiy et al., 1971). The main part of the basin (74.7%) belongs to the Dnieper River (504,000 km2) and the Don River (422,000 km2). All other rivers drain ~23.3% of the basin: the Dniester River (72,000 km2), the Southern Bug River (63,700 km2), the Kuban River (58,000 km2), and other smaller rivers (together 120,000 km2). The large basin size, extending from ~57°N to 44°N, and from 23°E to 41°E, incorporates substantial climate and landscape variability. Mean annual temperature increases from 3 to 9 °C north to south, while annual precipitation decreases from 600 to 300 mm in the same direction. Surface runoff decreases from 200 to 10–20 mm from the north to the south, and then

Figure 1. The East European part of the Black Sea drainage basin. Key: (1) river basin boundaries; (2) region boundaries (see the text for region descriptions); (3) data sites with large paleochannel remnants.

6

Sidorchuk et al.

increases again in the southern mountains. The main part of the basin, including the Dniester River basin, the northern parts of the Dnieper and Don basins, and the Kuban River basin, would be covered with broad-leaved forest under natural conditions. The lower Dnieper and Don basins would be covered by forest-steppe and steppe vegetation under natural conditions. Now, ~70% of the basin is plowed and used for agriculture.

The recent annual flow of all the rivers to the Black Sea from the East European Plain is ~110 km3 (83 km3 from the Dnieper and Don Rivers) or ~30% of the runoff from the total drainage area of the sea. The water regime of the large rivers has been substantially changed by the construction of a system of hydroelectric dams with large reservoirs. Small rivers are often regulated by chains of ponds.

Figure 2. The region with the late Valdai (late Weichselian) glacial and periglacial features. Key: (1) deposits of proglacial lakes; (2) sandy fluvioglacial deposits; (3) meltwater blow-out channels; (4) present-day direction of flow; (5) boundaries between ice sectors. Boundaries of the Last Glaciation stages and keys 1 and 2 are after Faustova and Chebotareva (1969). Key 3 is after Kvasov (1979) with corrections based on space images.

Surface runoff to the Black Sea from the East European Plain

The main geomorphologic features used in hydromorphological reconstruction belong to (1) glacial and periglacial topography (as well as glacial and periglacial deposits), which allows one to trace the boundaries of the former sheet glaciers on the East European Plain, and (2) fluvial topography (with fluvial deposits), which can be used for former runoff estimates. These features, dated to the LGM and LGT, can be found within three regions of the Black Sea basin (see Fig. 1). Region I with the Late Valdai (Late Weichselian) Glacial and Glaciofluvial Features Glacial and periglacial topography and deposits of the Fennoscandian ice sheet in the Dnieper River basin were investigated mainly in 1960s and 1970s, when several large monographs were published (Gerasimov, 1969; Chebotareva and Makarycheva, 1974). The position of the southern boundary of the Fennoscandian ice sheet here during the LGM has been generally confirmed by recent works (Velichko et al., 2004; Svendsen et al., 2004). It is firmly established that during its maximum extent, the last ice sheet covered only the northernmost part of the Upper Dnieper River basin (region I in Figs. 1 and 2). Glacial topography is represented by a system of moraine hills and ridges often clearly bordering the former ice lobes. Three main bands of such moraine morphology, dated with 14C and pollen analysis of under-moraine deposits, show the position of the glacier boundary (Fig. 2) during the Bologoye/Brandenburg stage (ca. 18 14C k.y. B.P.), the Edrovo/Frankfurt stage (ca. 17 14C k.y. B.P.), and the Vepsovo/Pomeranian stage (ca. 15.5 14C k.y. B.P.)—all the dates are after Chebotareva and Makarycheva (1974), the last being entirely beyond the Black Sea basin. Taking the center of the southern half-ellipse of the Fennoscandian glacier at the center of the northern part of the Gulf of Bothnia, the distance (Lm) to the glacier border (at the headwaters of the Berezina River) was ~1230 km at the Bologoye stage, ~1180 km at the Edrovo stage, and ~1100 km at the Vepsovo stage (Table 1). Each stage was characterized by an “active” phase of the glacier advance and by an “inactive” phase of the glacier retreat, when a marginal zone of “dead-ice” could exist. Another important morphological and depositional feature is represented by the fluvioglacial plains formed by fluvioglacial streams, terraces, and deposits of ice-dam lakes and river terraces related to them. These features are also dated with 14C and by their correspondence to the glacial topography. According to Faustova and Chebotareva (1969) and Chebotareva and Makarycheva (1974), the “glacial” part of the Black Sea drainage basin achieved its maximum area during the Bologoye (Brandenburg) stage ca. 18 14C k.y. B.P. At that time, a chain of ice-dam lakes (Kvasov 1979), or a series of short-lived glacial lakes (Mangerud et al., 2004), formed a broad band between the glacier front and the main pre–last glaciation water divide of the East European Plain. The chain of lakes stretched from the headwaters of the Dnieper River in the east to the upper part of the Neman River

TABLE 1. MELTWATER SUPPLY TO THE DNIEPER RIVER HEADWATERS DURING THE EARLY STAGES OF THE FENNOSCANDIAN ICE-SHEET RETREAT F Hm E Sector of drainage Ice-sheet stage Lm I tcal Lf 0.9 ΔVu/Δt Δt ΔWu/Δt 3 1 3 1 3 1 (km yr– ) (km yr– ) (km3 yr–1) (km) (km) from ice sheet (yr B.P.) (km) (km) (km yr– ) (yr) “Cold” glacier scenario Bologoye 1230 670 410,000 3.3 21,500 86 25 Edrovo 1180 670 390,000 3.2 20,300 83 24 A+B+C 1200 54 115 Edrovo 1180 440 260,000 3.2 20,300 55 15 B+C 1800 26 65 Vepsovo 1100 440 245,000 3.1 18,500 53 15 Edrovo 1180 270 160,000 3.2 20,300 34 10 Vepsovo 1100 270 150,000 3.1 18,500 33 9 C 1800 16 40 “Warm” glacier scenario Bologoye 1230 670 410,000 2.1 21,500 86 25 A+B+C 1200 35 96 Edrovo 1180 670 390,000 2.1 20,300 83 24 Edrovo 1180 440 260,000 2.1 20,300 55 15 Vepsovo 1100 440 245,000 2.0 18,500 53 15 B+C 1800 17 55 Edrovo 1180 270 160,000 2.1 20,300 34 10 Vepsovo 1100 270 150,000 2.0 18,500 33 9 C 1800 10 34 “Dead-ice” glacier scenario Bologoye 1230 670 410,000 0.5 21,500 86 25 A+B+C Edrovo 1180 670 390,000 0.5 20,300 1200 83 24 7 68 Edrovo 1180 440 260,000 0.5 20,300 55 15 B+C 1800 3 42 Vepsovo 1100 440 245,000 0.5 18,500 53 15 Edrovo 1180 270 160,000 0.5 20,300 34 10 Vepsovo 1100 270 150,000 0.5 18,500 33 9 C 1800 2 26 Note: See explanations of the indexes in the text. Note that all unit rates are multiplied on the sector half-width.

Main Geomorphologic Regions in the Black Sea Basin

7

8

Sidorchuk et al.

basin in the west (Fig. 2). The lakes were connected by channels and formed a pool-step system with a general slope from east to west. Mangerud et al. (2004) did not assume any southward meltwater drainage from these lakes. Nevertheless, there is clear geomorphic evidence of meltwater blow-outs through a number of river valleys. During the Bologoye stage, when this lake system was separated from the marginal valleys in the Polish and German lowlands, the meltwaters drained to the south through the lowest parts of the main water divide. At different times, the flows used different passes, so that a complicated pattern of blow-out valleys was formed at the headwaters of the Dnieper, Dvina, and Neman Rivers. Kalicki (1995) even suggested a new “Dnieper-type” of paleochannel with multiple old valleys. On the whole, there were three main routes of meltwater drainage: through upper Dnieper valley, through the upper Berezina River, and through the Pripyat River, each corresponding to one of three sectors of the ice sheet (A, B, C in Fig. 2). At that time (ca. 17 14C k.y. B.P.), the first terrace in the Upper Dnieper basin was formed (Kalicki and San’ko, 1992). This terrace is ~1800 m wide in the Upper Dnieper valley, ~1200 m wide in the Upper Berezina valley, and ~1600 m wide in the valley of the Shara River—a tributary of the Neman River that connected a paleolake at the Upper Neman basin with the Pripyat River valley (Fig. 2). The sizes of large meanders in these paleochannels correspond well to their widths. With the last Fennoscandian glacial retreat from the boundaries of the maximum (Bologoye) stage, the area of the additional “glacial” part of the Black Sea basin (F) decreased from 410,000 km2 (sectors A + B + C) to 260,000 km2 (sectors B + C), as the width of the ice-sheet front (Lf) was reduced from 670 to 440 km (see Table 1). It happened when the paleo–Upper Neman River became connected with the marginal valleys in the Polish and German lowlands through the Neman paleovalley, and meltwater flow into the Pripyat River stopped. The meltwater fluvioglacial streams and ice-dam lakes continued to drain into the Upper Dnieper and Upper Berezina only in the eastern part of the system. During the Edrovo stage, the area of the additional “glacial” part decreased again, from 260,000 km2 to 150,000 km2 (sector C), when meltwater drainage into the Upper Berezina River stopped. After the Dvina River valley formed during the Vepsovo stage, the entire system of meltwater drainage shifted into the Baltic Lake and into the western marginal valleys, so that the Black Sea lost its connection with the meltwater source. Region II with the Late Glacial Large Alluvial Paleochannels Well-preserved fragments of large meandering paleochannels can be distinguished on large-scale maps and space images. In our former investigations, 16 such fragments over rivers >200 km long were described in the Black Sea basin (see table 1 in Sidorchuk et al., 2001). The use of Landsat-7 images with 15 m resolution allowed us to find an additional 35 fragments within river valleys with basins greater than 5000 km2 (Table 2 in this paper). Although a future increase in image resolution will potentially increase the number of large paleochannels traced, currently available data on the Black Sea drainage basin

provide the possibility of analyzing the general distribution of such features. Remnants of the large alluvial paleochannels are wellpreserved in the basins of the tributaries of the Dnieper and Don rivers (region II in Fig. 1). These remnants are mainly situated at the level of the modern floodplain. In the north, the necks of paleomeanders form fragments of the first terrace, while the point bars and filled paleochannels constitute the floodplains of the recent valleys. In several river valleys, the paleochannel fragments are partly situated at the low terraces and partly included within the recent floodplain. In the southern part of the region, all elements of paleochannels are included within the recent floodplains. Therefore, rivers in this region are characterized by very high ratios of floodplain to channel widths. The relationships between river channel plan geometry and flow discharge are of prime importance for paleohydrological reconstruction. The large paleochannels (in terms of their width and meander wavelength) were up to 15 times larger than the recent channels in the same river basin. Such sizes indicate large surface runoff at the time of paleochannel formation. Two key sites were investigated in this region: one in the Dnieper River basin (Seim and Svapa Rivers), and the other in the Don River basin (the Khoper River). Paleochannels in the Seim and Svapa River Valleys (Dnieper River Basin) The floodplain of the Seim River and its main tributaries is characterized by a sequence of arcs with their radii of curvature exceeding that of the modern river channel bends by an order of amplitude (Figs. 3 and 4). Systems of natural levees and large abandoned oxbows are well defined on aerial photographs. Systems of levees and hollows with relative relief of 0.5–1.5 m reflect steeply curved meandering paleochannels with wavelengths of 3.3–6.5 km and widths of 350–700 m (for the recent Seim River, these are 0.2–1.0 km and 20–100 m, respectively). The floodplains with remains of large paleochannels lie largely at 2.0– 2.5 m above low water level (LWL) and usually become inundated during floods. Only the tops of the highest levees are as high as 3–3.5 m and usually remain above flood level. The investigated fragment of the large paleochannel of the Seim River near Kudintsevo village is the highly curved meander with a wavelength of 6 km. The texture of infilling of this large paleochannel was investigated by coring along the profile A′–A″ in the upper part of the bend (Fig. 3). The paleochannel cross section has an asymmetrical triangular shape with the deepest part (~10 m below LWL) near the steep concave bank of the paleomeander. The paleochannel fill includes three main units. The lowermost unit is fine silty sand 5–7 m thick, which belongs to the initial stage of the paleochannel infill. This sand is overlain by gray clay and silt accumulated largely in the oxbow lake on the floodplain. Deposits at the base of this unit were radiocarbon dated to 12,630 ± 70 and 13,800 ± 85 yr B.P., based on bulk organic matter (samples Ki-6985 and Ki-6984). We assume that the latter date indicates the time shortly after the macromeander

Surface runoff to the Black Sea from the East European Plain TABLE 2. PALEORUNOFF RECONSTRUCTIONS BASED ON THE SIZES OF LARGE 14 ALLUVIAL PALEOCHANNELS IN THE BLACK SEA DRAINAGE BASIN CA. 14–15 C k.y. B.P. W Q Latitude Longitude F λ River 2 3 (m) (m /s) (°N) (°E) (km ) (m) Aidar 49.01 3 8 .9 6 7015 845 1 74 48 Berezina 54.02 28.87 9557 991 260 74 Bitiug 51.03 40.06 7695 1177 211 69 Buzuluk (Khoper) 50.64 42.79 6390 1206 182 62 Desna 51.13 31.03 84,183 2145 385 194 Desna 51.44 31.95 71,145 1806 278 136 Desna 52.45 33.56 24,989 1513 297 115 Dnieper 54.97 32.98 7340 1997 136 Egorlyk 45.99 41.30 1 1 ,2 2 0 938 50 Egorlyk 45.87 41.40 8 707 845 171 48 Iput' 52.47 31.39 9400 809 41 Kalitva 48.42 40 . 96 10 , 14 0 80 5 133 40 Khoper 51.27 42.33 19,329 1906 328 140 Khoper 52.46 43.67 8981 1358 184 71 50.19 43.73 29,612 1810 213 107 Medveditsa (Don) Medveditsa (Don) 51.44 44.86 7582 973 184 55 Nerussa 52.44 33.90 5346 890 209 56 Orel' 48.97 34 . 4 9 91 58 1 26 5 187 68 Orel' 48.93 34 . 5 9 91 58 9 73 203 60 Orel' 49.16 34 . 9 3 74 76 9 12 152 46 Orel' 49.13 35 . 1 2 56 53 7 92 153 41 Psel 49.23 33.68 22,1 58 12 03 152 62 Psel 49.74 33.78 14,6 92 841 162 49 Psel 50.18 33.97 11,7 35 712 122 35 Psel 50.56 34.44 94 17 67 1 1 21 33 Psel 51.04 35.22 65 87 67 0 30 Ros' 49.49 31.48 11,021 3701 458 267 Sal 47.33 41.34 20,523 1392 288 105 1936 286 130 Samara (Dnieper) 48.64 35.37 19,903 Seim 51.39 33.41 27,070 1938 234 118 Seim 51.28 33.88 2 2,257 1602 220 97 Seim 51.48 34.78 19,678 1673 312 124 Seim 51.69 35.23 11,418 1924 349 140 18,706 742 152 44 Severskiy Donets 49.33 36.84 Snoc 51.66 31.64 8281 2311 167 Sozh 52.35 30.95 39,336 1693 380 154 Sozh 53.86 31.80 6630 225 77 Styr' 50.77 25.34 7201 1762 114 Sula 49.65 32.71 1 8 , 00 9 1 5 95 26 3 106 Sula 49.62 32.88 1 8 , 00 9 1 6 62 27 2 112 Sula 49.81 32.90 1 5 , 29 5 1 7 27 19 0 91 Sula 50.22 33.32 60 9 8 1 26 0 238 77 Tersa 50.87 44.11 6589 963 171 51 Udaj 50.21 3 2 .7 6 65 79 1 2 04 2 10 69 Volchia 48.11 3 6. 0 7 9627 77 1 13 9 40 Vorona 51.78 4 2.37 12,746 1 434 252 93 Vorskla 49.00 34 .16 14,433 12 93 256 89 Vorskla 49.04 34 .33 14,433 12 16 238 81 Vorskla 49.49 34 .58 11,124 83 8 1 15 38 Vorskla 49.73 34 .63 9 129 1259 207 72 Vorskla 50.14 34 .74 6 220 75 9 133 37 Note: F—catchment area, λ—mean paleomeander step (half-wavelength), W—mean paleochannel width, Q—mean annual discharge, X—mean annual runoff depth.

was abandoned. By the beginning of the Holocene, the oxbow lake had been transformed into a fen, and mineral deposition was followed by peat accumulation. The thickness of the peat layer is up to 2 m. In the Svapa River valley, the paleochannel formed steeply curved meanders with a mean wavelength of 2.8 km and a mean width of 300 m (for the recent Svapa River, these are 0.14–

9

X (mm) 216 245 282 307 72 60 145 583 142 175 136 124 228 248 114 229 328 233 207 194 229 89 104 94 110 146 763 161 205 138 137 199 387 74 637 124 364 500 186 195 188 396 245 329 130 230 194 177 107 249 185

0.6 km and 15–60 m, respectively). The paleochannel fragment near Semenovka village is clearly expressed in the modern topography (Fig. 4). Its surface lies only 4 m above the modern LWL, so that it is submerged during high floods. Coring reveals that the channel trough assumes a box-shaped profile. The top of the lower layer of fine- and medium-grained channel alluvial sands lies 1.5–2.5 m below the modern LWL. Silt and clay deposits

10

Sidorchuk et al.

with lenses of clayey sand fill the trough. Accumulation of finegrained sediments began because of the abandonment of the paleochannel in the Oldest Dryas (14,030 ± 70 yr B.P., Ki-6997), and continued during the Bølling and the Allerød (12,360 ± 110 yr B.P., Ki-6999 and 11,755 ± 80 yr B.P., Ki-6996). At the end of the Late Glacial to beginning of the Holocene, the paleochannel almost entirely dried up, and the rate of deposition became very low. Peat formation started in the late Preboreal (9120 ± 70 yr B.P., Ki-6995; 9300 ± 120 yr B.P., GIN-11951). Paleochannels in the Khoper River Valley (Don River Basin) One of the best preserved systems of large paleochannels with bankfull channel width of 0.8–1.4 km, maximum depth of 9 m, and mean wavelength L of 5 km, is found on the ancient floodplain of the Khoper River near Povorino (Fig. 5). At present, the Khoper River near Povorino has a channel width of 60 m, maximum depth of 4 m, and a meander length up to 360 m. Radiothermoluminescence dating (Vlasov and Kulikov, 1988) of the bottom deposits (17 ± 4 ka, RTL-808) shows that the large paleochannels were formed during the Late Glacial. The

thickness of their subsequent infilling varies at different parts of the paleochannel. The system of channels that follows the right bank of the valley has been rarely flooded after it was abandoned, so that the former bottom of the paleochannel is locally exposed. The maximum thickness of the channel-filling deposits there does not exceed 1.5–3.0 m (Fig. 5). There are large eolian dunes on the terrace at the macromeander neck. The system of channels along the left bank of the valley is situated nearer to the present river. The pools of the large paleochannel with maximum depth of 9–11 m were completely filled in by fine-grained alluvium, beginning from 11.3 to 10.8 14C k.y. B.P. A system of smaller paleochannels (although still larger than the recent Khoper River channel) was preserved at the outspread of the floodplain in the valley bend (see Fig. 5, cross-section C–C′). The channel width was ~200 m, and the meander wavelength was ~1200 m. These channels were active before Boreal time, presumably during the Younger Dryas. Deposition of floodplain sediments during the Holocene was concentrated in a narrow belt, 1 km wide, along the river channel, where sandy natural levees up to 3.5 m above LWL were formed.

Figure 3. Geological section and space image across the large meandering channel of the paleo–Seim River near Kudintsevo village.

Surface runoff to the Black Sea from the East European Plain Region III with Large Incised Paleochannels Region III (see Fig. 1) occupies the high western part of the East European Plain (Volyno-Podolsk Upland), the near–Black Sea plain, and pre-Caucasus highlands. Rivers are often incised here, presumably because of tectonic uplift. The channels of these rivers form large bends with systems of Quaternary terraces at their necks. The ages and origins of these large bends are mostly unknown, and their paleohydrological signal is unclear. These features are not used in further runoff estimates. RESULTS OF THE ESTIMATES OF PALEOGEOGRAPHIC CONDITIONS Hydroclimatic Parameters of River Development since the Late Pleniglacial To reconstruct the main climatic indexes in the Black Sea basin, composition of fossil floras derived from palynological data (the method of “arealograms”) was analyzed. Fossil floras used for paleoclimatic reconstructions were derived from five sites

11

(Fig. 6). At three of them, palynological studies were conducted on fluvial sediments that filled segments of large paleochannels found on the floodplains of the Seim, Svapa, and Khoper Rivers (Sidorchuk and Borisova, 2000; Borisova et al., 2006). Another of the localities is the Early Man site of Yudinovo, where loamy sediments containing a Late Paleolithic cultural layer were subjected to detailed pollen analyses (Borisova and Novenko, 1999). One more fossil flora includes both pollen and plant macrofossils identified in the so-called “Usvyacha” deposits exposed in several outcrops within the Dvina River valley near the villages of Sloboda and Drichaluki (Velichkevich, 1982; San’ko, 1987). In the course of pollen analysis, an attempt was made to achieve the highest possible taxonomic resolution. Pollen identifications have been made to species or genus levels for arboreal plants as well as certain groups of herbaceous plants. When identifications were possible only to the family level, the finds in question were not included in the resulting lists of fossil floras (Table 3). On the whole, these paleobotanical data proved to be sufficient to locate modern geographical analogues to 12 fossil floras and, therefore, to estimate climatic changes that occurred within the

Figure 4. Geological section and space image across the large meandering channel of the paleo–Svapa River near Semenovka village.

Figure 5. Geological sections and space image across the large meandering channels of the paleo–Khoper River near Povorino. RTL—radiothermoluminescence.

12 Sidorchuk et al.

Surface runoff to the Black Sea from the East European Plain

13

Figure 6. Location of the modern region analogues to the Last Glacial Maximum and Late Glacial time fossil floras. Key: (1) key sites (Sm— Seim, Sv—Svapa, Kh—Khoper, Yu—Yudinovo, Sl—Sloboda/Drichaluki); (2) modern region analogues of the fossil floras (number of a fossil flora and its 14C age in k.y. B.P.); (3) corresponding hydrological region analogues in the lowland areas (A—Bol’shezemel’skaya Tundra, B— Lena Plateau).

East European part of the Black Sea drainage basin since ca. 18 14 C k.y. B.P. The accuracy of the match between a modern geographical analogue and a fossil flora depends not only on the richness of the latter, but also on the knowledge of the modern geographical distribution of each plant genus or species. In some cases, the area of a region analogue remains rather large and includes different types of vegetation associations. Therefore, in this study, all such regions were located on the map (see Fig. 6) and additionally checked afterward against the types of plant communities, reconstructed from the pollen assemblage of a given sample. Sometimes, such comparisons enabled us to reduce the area of the region analogue, but even then, the resulting areas remained large enough to show a substantial variability of hydroclimatic characteristics within them. The reconstructed ranges of climatic parameters are shown in Figure 7 as shaded boxes, their vertical size corresponding to the time intervals characterized by each fossil flora. The earliest fossil flora (flora 1 in Table 3 and in Figs. 6 and 7)—derived from the palynological and plant macrofossil studies of the Sloboda and Drichaluki sections—belongs to the relatively cold stage of the Late Pleniglacial, dated by radiocarbon to ca. 17–18 ka (Velichkevich, 1982; San’ko, 1987). The flora includes several of the typical Arctic and Arctic-mountain species,

which at present grow in various European and west Siberian tundra and forest-tundra communities, along with some xerophytes tolerant to low winter temperatures, such as Ephedra distachya. It also includes trees and shrubs growing presently in regions with cold and highly continental climate, mainly in Siberia (Larix sp., Alnaster fruticosus). Such a complexity of flora is highly typical of the glacial floras in northern Eurasia (Grichuk, 1969). The region currently inhabited by the species of fossil flora 1 lies south of the East Sayan Mountains, in the upper part of the Oka River basin. The area has a cold climate with mean January air temperature of –21 to –22 °C and mean July air temperature of 8– 10 °C, which are characteristic of mountain tundra in this region. Because of permafrost, the runoff coefficient should be very high (more than 0.8). The mean annual precipitation is 400–600 mm, and the calculated runoff depth is 350–500 mm (Table 4). The Early Man site of Yudinovo is located on the first terrace of the Sudost’ River, a tributary of the Desna River in the Dnieper basin. Both the geomorphological position of the site and a series of radiocarbon dates based on the materials from the cultural layer indicate that the site was inhabited approximately from 14.5 to 14 ka (Svezhentsev, 1993). According to palynological data, fossil flora of this interval consists mainly of forest and meadow plants. It includes Boreal and Arctic-Boreal trees and shrubs, forest club-mosses (Lycopodium clavatum, L. selago, and

14

Sidorchuk et al. TABLE 3. THE LATE GLACIAL AND THE HOLOCENE PALEOFLORAS FOUND IN THE BLACK SEA DRAINAGE BASIN

Site names:

Sloboda/ Drichaluki Fl 1

Yudinovo

Seim

Seim

Numbers of fossil floras: Fl 2 Fl 3 Fl 4 Radiocarbon ages of the (17–18) (14–14.5) 13.8 12.6 fossil floras (k.y. B.P.)#: I. Trees and shrubs (a) Micro- and mesothermal species of the continental regions Abies sibirica * Alnaster fruticosus *† * * * § Larix sp. ** * Pinus sibirica * * (b) Species with broad ecological tolerances Betula alba * * * * Betula humilis * * * Picea abies * * * * Pinus sylvestris * * * * (c) Relatively thermophile mesophytes Acer tataricum Alnus glutinosa A. incana * * * Corylus avellana Quercus robur Tilia cordata Ulmus sp. U. campestris Viburnum opulus II. Arctic-Alpine microthermal mesophyte species Betula nana ** * Dryas octopetala ** Gastrolychnis apetala ** Lycopodium pungens * Polygonum viviparum ** Potentilla cf. nivea ** Salix herbacea ** Salix cf. polaris ** Selaginella selaginoides ** * III. Herbaceous plants associated with various forest communities Botrychium ramosum * Humulus lupulus Lycopodium annotinum L. clavatum * * * L. complanatum * L. selago * L. tristachyum * Thalictrum minus T. simplex * Pteridium aquilinum * IV. Xerohalophytes Atriplex cana * A. pedunculata A. verrucifera Chenopodium glaucum * C. acuminatum C. chenopodioides Kochia prostrata * * Plantago cornuti Salsola sp. * S. soda V. Xerophytes Helianthenum sp. * * * Eurotia ceratoides Ephedra (non-distachya) * * * * Ephedra distachya Kochia scoparia * VI. Psammophytes and plants growing on eroded soil Amaranthus sp. Atriplex tatarica Centaurea cyanus Chenopodium album C. botrys Linaria sp. Spergula sp.

Svapa

Seim

Khoper

Seim

Svapa

Svapa

Svapa

Svapa

Fl 5

Fl 6

Fl 7

Fl 8

Fl 9

Fl 10

Fl 11

Fl 12

12.4

12.2

11.9

11.5

(10.5)

9.8

7.5

4.9

*

* * * *

* *

* * * * * * * *

*

* * * *

* * * *

* *

* * * * *

* *

* * *

* * * *

*

* *

*

*

*

* * * *

* * * * * * * * *

*

*

* * *

*

* * *

* * *

* * *

* * *

* * * * *

* * *

*

*

* *

* *

* *

* * * * *

* * *

* * * *

* * *

* *

* (continued)

Surface runoff to the Black Sea from the East European Plain

15

TABLE 3. THE LATE GLACIAL AND THE HOLOCENE PALEOFLORAS FOUND IN THE BLACK SEA DRAINAGE BASIN (continued) Site names:

Sloboda/ Drichaluki Fl 1

Yudinovo

Seim

Seim

Numbers of fossil floras: Fl 2 Fl 3 Fl 4 Radiocarbon ages of the (17–18) (14–14.5) 13.8 12.6 fossil floras (k.y. B.P.)#: VII. Steppe and meadow plants Cannabis sp. Fagopyrum sp. * Plantago ramosa Rumex acetosella Botrychium lanceolatum * B. simplex Bupleurum sp. Papaver nudicaule ** Plantago lanceolata * Sanguisorba officinalis Scabiosa sp. * Thalictrum foetidum Valeriana sp. * VIII. Plants of wet meadows, water margins, and mires (helophytes) Alisma gramineum A. plantago-aquatica * Calystegia sp. Equisetum palustre E. scirpoides E. variegatum Filipendula ulmaria Menyanthes trifoliata Myosoton aquaticum ** Polygonum amphybium Ranunculus reptans ** Sparganium sp. * * S. hyperboreum ** Sagittaria sagittifolia Thalictrum angustifolium T. flavum Typha angustifolia T. latifolia Urtica sp. IX. Aquatic plants Batrachium sp. ** Lemna sp. Myriophyllum sp. * M. spicatum * M. verticillatum Nymphaea alba Nymphaea candida Potamogeton filiformis ** P. natans P. perfoliatus ** P. vaginatus ** # Age estimations based on interpolation are shown in parentheses. † Plants identified by their pollen or spores. § Plants identified by macrofossils.

others), Pteridium aquilinum, and other mesophile plants (see Table 3). Such a floral composition suggests relatively humid conditions during the interval in question. A region analogue for fossil flora 2 lies in the Altai Mountains, in the middle reaches of the Biya River, on the east-facing slopes (Fig. 6). In this region, pine and birch forests and spruce–fir–Siberian pine mountain taiga forests occur along with wet meadows. The area is characterized by milder and wetter climatic conditions compared to the region analogue 1: the mean January air temperature there is –16 °C, and the mean July temperature is 16–17 °C. The mean annual precipitation is 700–800 mm, and the runoff depth is 500– 550 mm (see Table 4).

Svapa

Seim

Khoper

Seim

Svapa

Svapa

Svapa

Svapa

Fl 5

Fl 6

Fl 7

Fl 8

Fl 9

Fl 10

Fl 11

Fl 12

12.4

12.2

11.9

11.5

(10.5)

9.8

7.5

4.9

* * *

*

* * * * *

*

*

*

*

* *

*

* *

* * *

*

* * *

* *

*

* * *

*

* *

*

*

*

*

*

* *

*

*

*

* * *

* *

* *

* *

* *

* * *

* * *

* * * *

Sediments of the initial stage of filling of the large paleochannel of the Seim River (core S-4 in Fig. 3; 13,800 ± 85 yr B.P., Ki-6984) generally correspond to the beginning of the Oldest Dryas. Fossil flora 3, derived from these sediments, combines cryophile (Alnaster fruticosus, Selaginella selaginoides) and xerophile (Ephedra sp.) plants, inhabitants of the boreal forest, steppe, meadow, and riverine communities. The closest modern floristic analogue for this assemblage can be found in the forest steppe in the middle reaches of the Irkut River basin, west of Lake Baikal (see Fig. 6). Within this small area, larch and pine forest grow next to southern Siberian meadow steppes, with patches of spruce forest in the river valleys. The area is characterized by a

Figure 7. Estimations of the main climatic indexes (deviations from modern values) based on the composition of fossil floras (numbers of the floras are as in Tables 3 and 4 and Fig. 6). PB—Preboreal; DR3—Younger Dryas; AL—Allerød; BØ—Bølling; DR1—Oldest Dryas.

16 Sidorchuk et al.

2.0 3.5 3.0 2.6 3.4 1.6 1.6 1.8 2.0 1.2 1.4 0.8 350 to 400 225 to 275 175 to 225 240 to 290 25 to 125 40 to 90 75 to 125 40 to 190 –10 to 45 20 to 70 –50 to 15

140 to 290 350 to 500

500 to 550 350 to 400 300 to 350 350 to 400 150 to 250 150 to 200 200 to 250 150 to 300 100 to 155 130 to 180 65 to 125 125 to 225 –125 to –75 –175 to –125 –125 to –75 50 to 150 40 to 140 125 to 175 –150 to –50 –50 to 50 100 to 150 75 to 125

–200 to 0 400 to 600

700 to 800 425 to 475 375 to 425 425 to 475 600 to 700 500 to 600 675 to 725 400 to 500 500 to 600 650 to 700 625 to 675 –1.5 to –2.5 –3.5 to –2.5 –2 to 0 –4 to 0 –3 to –1 –3 to –4 –0.5 to –1.5 –3.5 to –4.5 –0.5 to 0.5 0 to –1 0 to –1

–7 to –9 8 to 10

16 to 17 15.5 to 16.5 17 to 19 15 to 19 16 to 18 16 to 17 17.5 to 18.5 14.5 to 15.5 18.5 to 19.5 18 to 19 18 to 19 –5.5 to –9.5 –14 to –18 –15 to –19 –13 to –15 –7.5 to –8.5 –6 to –8 –7.5 to –8.5 –9 to –11 –7 to –8 –7 to –8 1 to 2

–13 to –14 –21 to –22

–14 to –18 –22 to –26 –23 to –27 –21 to –25 –15.5 to –16.5 –16 to –18 –15.5 to –16.5 –17 to –19 –15 to –16 –15 to –16 –6 to –7

(17–18)

(14–14.5) 13.8 12.6 12.4 12.2 11.9 11.5 (10.5) 9.8 7.5 4.9

1

2 3 4 5 6 7 8 9 10 11 12

Sloboda/ Drichaluki Yudinovo Seim Seim Svapa Seim Khoper Seim Svapa Svapa Svapa Svapa

ΔRunoff depth (mm/yr) Runoff depth (mm/yr) ΔPrecipitation (mm/yr) Precipitation (mm/yr) ΔTemp. July (°C) Temp. July (°C) ΔTemp. January (°C) Temp. January (°C) No. of fossil flora

Site names Radiocarbon ages of the fossil floras (k.y. B.P.)

TABLE 4. MAIN CLIMATIC INDEXES IN THE BLACK SEA DRAINAGE BASIN IN THE LATE GLACIAL AND THE HOLOCENE

Runoff depth ratio past/recent

Surface runoff to the Black Sea from the East European Plain

17

cold semiarid and extremely continental climate. The mean January temperature is –24 °C, and that of July is ~16 °C. The mean annual precipitation varies between 425 and 475 mm. The region is situated near the boundary of the permafrost zone with a high annual runoff coefficient. The mean annual surface runoff depth is 350–400 mm (see Table 4). Another sample of fluvial deposit fill from the large paleochannel of the Seim River was obtained from the basal part of core S-7 taken in the same profile as core S-4 (Fig. 3). According to the radiocarbon date (12,630 ± 70 yr B.P., Ki-6985), it corresponds to the final part of the Oldest Dryas. The flora of this sample (flora 4 in Table 3) is distinctive for the diversity of its xerophytes and xerohalophytes, which suggest a rapid drying of the topsoil in the watershed areas during relatively warm summers. Similar conditions occur at the present time in southern Siberia. A region analogue for fossil flora 4 lies at the headwaters of the Yenisei River, within an intermountain depression downstream from the confluence of the Biy-Khem and Ka-Khem rivers (Fig. 6). In this area, southern Siberian dry grassy steppes are found next to a mountain forest of Pinus sibirica and Larix sibirica. The area is characterized by extremely cold winters with mean January air temperature from –23 °C to –27 °C, while the summer is warm with mean July temperature being ~18 °C. The annual magnitude of air temperature changes is ~43 °C, and that is ~15 °C greater than at the study site at the present time. The mean annual precipitation is ~400 mm, and calculated runoff depth is ~325 mm. Fluvial deposits filling the paleochannel of the Svapa River were cored near Semenovka village (see Fig. 4). According to the 14 C date (12,360 ± 110 yr B.P., Ki-6999), fossil flora 5, deriving from core SV-1–8, belongs to the beginning of the Bølling interstadial. It includes species of dark coniferous taiga forest (Picea abies, Abies sibirica, Pinus sibirica), and light coniferous and mixed boreal forest (Pinus sylvestris, Pteridium aquilinum, Betula alba), along with species of riverine shrub associations, meadows, psammophytes (e.g., Spergula), and xerophytes. A region analogue for fossil flora 5 lies at the headwaters of the Yenisei River. It is shifted to the north with respect to region analogue 4 (see Fig. 6). In this area, southern Siberian dry grassy steppes are also the main vegetation type, though the role of the dark coniferous mountain forest of Picea, Pinus sibirica, and Abies sibirica is slightly greater here compared to region analogue 4. The area is characterized by milder climatic conditions, with mean January air temperature of –23 °C and mean July air temperature of ~17 °C, which are similar to the reconstruction based on fossil flora 3. Therefore, the runoff coefficient could be the same as at the region analogue of flora 3. The mean annual precipitation is ~450 mm, and the calculated runoff depth is ~375 mm. Palynological data for the fluvial deposits infilling the paleochannel of the second generation in the Seim River valley were obtained from borehole S-11 (see Fig. 3). Flora 6, derived from a sample dated to 12,250 ± 70 yr B.P. (Ki-6987), includes broad-leaved temperate tree species (Quercus, Ulmus, and Tilia cordata), as well as Alnus glutinosa, tree alder, growing on the

18

Sidorchuk et al.

waterlogged ground. Its composition indicates that in the late Bølling, the area was covered by a complex vegetation of forest steppe type, with birch and pine copses and minor participation by broad-leaved trees. The presence of broad-leaved trees in the region implies a complete degradation of permafrost at the end of the Bølling. This is confirmed by the generally mesophile character of the flora and by the presence of pollen of relatively heatdemanding aquatic plants, such as Nymphaea alba (see Table 3). A diversity of the hygro- and hydrophytes and an absence of xerohalophytes in this flora suggest an increase in rainfall. A region analogue for fossil flora 6 lies at the headwaters of the Ufa River in the Southern Ural Mountains (Fig. 6). In this region, meadow steppes come into close contact with the southern Urals pine and birch forests and spruce-fir subtaiga forests, with a minor presence of broad-leaved trees. The area is characterized by significantly milder climatic conditions compared to the region analogue of the early Bølling flora 5: the mean January air temperature there is –16 °C, and the mean July temperature is ~17 °C. The mean annual precipitation is 650 mm, and the runoff depth is ~200 mm, with 70% of the flow passing during the spring flood. The annual runoff coefficient is 0.31, and that for the flood period is ~0.6. The alluvium deposits infilling the meandering Khoper paleochannel were studied near the town of Povorino (see Fig. 5). At the base of core B, a radiocarbon date was obtained for the layer enriched with organic matter: 11,900 ± 120 yr B.P. (Ki-5305). The flora of this horizon includes some arboreal species with broad ecological tolerances, such as Scots pine (Pinus sylvestris) and tree and shrub birch (Betula alba, Betula humilis), but it also includes Siberian pine (P. sibirica), growing in the regions with continental climate, xerohalophytes, and xerophytes (Atriplex pedunculata, Atriplex verrucifera, and others; see Table 3). A region analogue of fossil flora 7 is located in northeastern Kazakhstan, in the Bukhtarma River basin (Irtysh River basin) at the boundary between dry steppe and semidesert and close to the region where communities of shrub birch and open dark coniferous forest are spread on the western slopes of the Altai Mountains (Fig. 6). The mean January air temperature in this region is –17 °C; mean July temperature is ~16.5 °C. The mean annual precipitation varies between 500 and 600 mm. The annual runoff depth is 150–200 mm. The region analogue is situated beyond the permafrost zone but close to its boundary. Fossil flora 8 of the dated sample from section S-11 (see Fig. 3; 11,450 ± 60 yr B.P., Ki-6986) includes many helophytes—plants growing on wet meadows, near the water’s margin, and in shallow water (Sparganium spp., Menyanthes trifoliata, Sagittaria sagittifolia, and others; see Table 3). Composition of this flora indicates that the Allerød warm interval was favorable for the expansion of dark coniferous trees (Picea abies, Pinus sibirica, and Abies) in the Seim River basin. Forest communities were dominated by birch and Scots pine, and larch participated in the pine forest at this period. Of the relatively thermophile trees and shrubs, flora of the Allerød included Ulmus, Tilia, Corylus, and Viburnum. A modern analogue for fossil flora 8 is located at the headwaters

of the Belaya River near the Southern Urals (Fig. 6). The region is currently occupied by forest steppe, where meadow steppes are associated with birch and aspen woods and grow next to (1) larch-pine open forests with steppe elements in the herbaceous cover, and (2) broad-leaved (oak-linden) forests of Southern Ural type. This area is characterized by the mildest climatic conditions for the entire Late Glacial, with mean January air temperature at –16 °C and July air temperature at 18 °C. Mean annual precipitation there reaches 700 mm, and runoff depth is ~225 mm, with 60% of the flow passing during the spring flood. Annual runoff coefficient is 0.32, and that for the flood period is ~0.56. Fossil flora 9 is derived from a loam and clay sediment layer in core SV-1–8, which is correlated with the Younger Dryas on the basis of pollen composition and radiocarbon dating. This horizon is distinguished by high nonarboreal pollen content, as well as a distinct maximum of Artemisia pollen and a smaller peak of Chenopodiaceae. Among trees, Pinus sylvestris and Betula alba were predominant. Of the dark coniferous trees, Picea abies and Pinus sibirica are registered in this flora. Of the cold-tolerant shrubs, Betula humilis and Alnaster are present. Relatively thermophilic trees once again disappear from the local flora (see Table 3). Changes in the flora are indicative of both colder and more arid climatic conditions compared to the previous time interval, with probable reestablishment of permafrost in the region. A region analogue for fossil flora 9 is situated in the middle reaches of the Biya River in the Altai Mountains (Fig. 6). In the vegetation cover of this area, open larch woodlands with steppe elements and patches of steppes are distributed next to fir and Siberian pine forests on the mountain slopes. Birch and Scots pine communities occur on sandy soil. The area is characterized by a cold, continental, and relatively dry climate with mean annual precipitation of 450 mm and runoff depth of 150–300 mm. The mean January air temperature is –18 °C, and that of July is 15 °C. Floras 10–12, deriving from palynological data of peat in core SV-1–8 from the Svapa River valley, indicate a conspicuous change of landscape and climatic conditions at the transition from the Late Glacial to the Holocene. These floras are dominated by forest and meadow plants (see Table 3). Their closest present-day analogues are located within the boreal forest zone and, for the Atlantic period of the Holocene, broad-leaved forest zone near the boundary of the steppe zone (see Fig. 6). These changes suggest a considerable rise in temperature and precipitation, decrease in the runoff coefficient because of degradation of permafrost, and, therefore, a substantial decrease in surface runoff (see Table 4). On the whole, changes in geographical position and hydroclimatic characteristics of the paleofloristic region analogues reflect a complexity of climate change during the Late Glacial against the background tendency toward warmer and wetter climate (see Fig. 7). Secondary oscillations can be seen on this general trend. Air temperature was the lowest in the LGM, in the late part of the Oldest Dryas, and in the Younger Dryas. Relatively warm intervals precede the Oldest Dryas and correspond to the Bølling and the Allerød. Extremely low winter temperatures

Surface runoff to the Black Sea from the East European Plain

19

indicate the existence of permafrost during the Late Pleniglacial, the Oldest Dryas to early Bølling, and in the Younger Dryas. Precipitation changes in the LGM and LGT generally followed temperature changes, the cold stages being relatively dry, and the warm stages relatively humid, with the maximum precipitation achieved in the pre–Oldest Dryas warming (ca. 14.5–14 14C k.y. B.P.), and in the Bølling-Allerød interstade (ca. 12.5–11 14C k.y. B.P.). Surface runoff is not determined entirely by annual precipitation, since water losses depend on air temperature. The runoff coefficient is inversely correlated with temperature changes, showing a strong overall decrease during the LGM-LGT. Generally following changes in runoff coefficient, the surface runoff also decreased considerably during this period (Fig. 7). Secondary oscillations of precipitation and water losses led to significant variations in surface runoff. The greatest runoff was characteristic of the pre–Oldest Dryas warming, because of relatively high precipitation, low losses, and preserved permafrost. The second highest runoff maximum was achieved at the beginning of the Bølling interval, when the temperatures still remained low while precipitation began to rise. During the cold and dry Younger Dryas, an increase in runoff was caused by low water losses to evaporation. During the warm and relatively humid Bølling and Allerød interstadials, higher losses to evaporation caused a relative decrease in surface runoff. Nevertheless, the entire LGM-LGT interval was characterized by high surface runoff compared to the present-day level, which was achieved at the beginning of the Holocene (see Fig. 7).

was assumed to be constant and equal to 60 mm yr–1, according to measurements on recent glaciers. The results are more qualitative than quantitative because of the great uncertainty about the structure and coefficients of Equations 1–8. Nevertheless, these estimates show a very large meltwater supply to the headwaters of the Dnieper River and its tributaries (Berezina and Pripyat Rivers), which at the early stages of deglaciation was comparable with recent annual surface runoff from the Dnieper River basin: 54 km3. The sizes of the blow-out channels coeval with this large meltwater supply (see Fig. 2) yield an additional possibility for estimating its volume. Using Equation 9 along with the width W and meander half-wavelength λ of these paleochannels (Table 5), the mean annual discharge Q and annual yield V were estimated. Presumably, the seasonality of runoff was high, so index of runoff variability y = 5. Paleochannel effective width W* was calculated as

Estimation of the Meltwater Supply in the Dnieper River Basin during the Early Stages of the Ice-Sheet Retreat

Estimation of Surface Runoff in the Dnieper and Don River Basins during the Period of Large Paleochannel Formation

Morphometrical parameters of the Fennoscandian ice sheet at the LGM (Bologoye)–Vepsovo stages (see Fig. 2) along with Equations 1–8 provide an estimate of change in glacier volume, snow accumulation, and final mass budget (Table 1). Such estimates were made for three main scenarios: a “cold” glacier with a shape similar to those of the Antarctic and Greenland glaciers; a “warm” glacier with a flatter hypothetical shape, derived from some assumptions about Fennoscandian glacier dynamics (Khodakov 1982); and a “dead-ice” glacier with a marginal zone of “inactive” ice with 500 m thickness, which follows from geomorphologic considerations (Chebotareva and Faustova, 1982; Faustova, 1984). Mean annual evaporation from the glacier surface

Calculations with Equations 9–11 require knowledge of paleochannel morphology and a modern hydrological analogue of the ancient river. Paleochannel effective width W* was estimated with Equation 12 with the mean relative error of 15% (see Table 2). The climatic region analogues for Late Glacial time were estimated from paleofloristic analysis (see Fig. 6). For relatively mild periods of the Late Glacial (pre–Oldest Dryas warming, Bølling and Allerød interstadials), analogues are situated in the basins of the Biya River (the Altai Mountains) and of the Belaya River (southeastern Ural Mountains). For the cold periods (the LGM, the Oldest, Older, and Younger Dryas), the modern region analogues are located in the southeastern sub-Baikal

W* =

W + (λ / kλ ) . 2

(12)

Here, kλ is λ/W; the mean value of the ratio is equal to 5.6 for the paleochannels in the plains of northern Eurasia (Sidorchuk et al., 2008). This estimate of annual meltwater supply (from 60 km3 annually) is close to the “dead-ice” scenario and in general confirms a very large volume of meltwater flowing to the Black Sea basin in the early stages of the ice-sheet retreat.

TABLE 5. MELTWATER FLOW THROUGH THE MAIN PALEOCHANNELS FROM THE ICE-DAM LAKES IN THE UPPER DNIEPER RIVER BASIN Meander Channel Effective Mean annual width, W width, W* discharge, Qa Paleochannel half-wavelength, λ 3 (m /s) (m) (m) (m) Paleochannel from the Upper Neman ice-dam lake to the Pripyat' River through the Schara River 5250 1600 1270 650 Paleo–Berezina River near Borisov 3540 1200 920 400 Paleo–Dnieper River near Rogachev 8600 1800 1670 950

Annual water yield volume, V 3 (km )

20 13 30

20

Sidorchuk et al.

region and at the headwaters of the Yenisei River (the Sayan Mountains). These territories can be used as hydrological analogues only partly (for example, for the estimation of runoff coefficient) because of the mountainous relief, high inclinations, and small areas of the river basins. For the estimation of such characteristics as discharge variability, lowland territories with similar climatic indexes (mean temperature of January and July, annual precipitation) and landscapes, close to those of the LGT (sparse open vegetation, widespread permafrost), must be selected. The lowland with climatic characteristics similar to those in the Altai and Ural Mountains, is the Bol’shezemel’skaya Tundra in the northeastern part of European Russia. The lowland with climatic characteristics similar to the Yenisei River headwaters is located within the Lena Plateau in central Yakutia (see Fig. 6). Hydrological regimes of large rivers in these two territories are quite similar: spring flood is high and sharp, and most of the annual flow passes during the flood. Low water period during the summer, autumn, and winter is long, but some years only 10%–12% of the annual runoff passes during this period. Coefficient a in Equation 11, which depends on hydrological regime, is therefore very similar for these two territories: it is 2.25 for the Bol’shezemel’skaya Tundra and 2.23 for the Lena Plateau. The sequence of paleohydrological calculations is as follows. The coefficient of within-year flow variability y for the past hydrological regime was estimated with Equation 11 for each river basin in Table 2, using coefficient a = 2.25. Basin area was assumed to be constant since the LGT. The mean annual paleodischarge Qa m3 s–1 was calculated with Equation 9, using y and effective width W*. The mean maximum paleodischarge Qmax

was estimated with Equation 10, using y and Qa. Surface runoff depth X (mm yr–1) from basin area F (km2) was calculated as

X = 31,536

Qa . F

(13)

The structure of Equation 9 leads to an increase in the relative error of discharge estimation compared to the relative error of channel width estimation. Therefore, the errors of paleodischarge and runoff depth estimates for the Dnieper and Don River basins are close to ±20%. Although we obtained only a few 14C dates from paleochannel deposits in the Dnieper and Don River basins (see descriptions of the key sites), these dates support a hypothesis of synchronous activity of the large rivers. As the oldest large paleochannels were abandoned 15–13 14C k.y. B.P., the large rivers were active in pre–Oldest Dryas time. Therefore, information from Table 2 can be used to reconstruct the spatial distribution of surface runoff within the Dnieper and Don River basins for this period (Fig. 8), and to estimate the volume of water flowing into the Black Sea from these basins. The general pattern of the pre–Oldest Dryas runoff is quite similar to the modern longitudinal decrease in runoff depth from north to south in the Dnieper and Don River basins (Figs. 8A and 8B). Maximum runoff was reconstructed for the area adjacent to the ice sheet, though none of the rivers used in the calculations was fed by glacier meltwater in the pre–Oldest Dryas. There was some lateral differentiation in the runoff: the Dnieper River basin was drier than the Don River basin. The same effect can be

Figure 8. Surface runoff in the East European part of the Black Sea drainage basin: (A) recent (measured) (after Evstigneev, 1990); and (B) calculated for the period of activity of large meandering rivers (pre–Oldest Dryas time). Key: (1) annual depth of surface runoff (mm); (2) data sites.

Surface runoff to the Black Sea from the East European Plain

River Don Dnieper Total

TABLE 6. VOLUME OF ANNUAL SURFACE RUNOFF FROM THE DNIEPER AND DON RIVER BASINS Recent Late Pleniglacial Annual runoff volume, VR Basin area Annual runoff volume, VP Basin area 2 3 2 3 (km ) (km ) (km ) (km ) 422,000 29 422,000 110 504,000 54 504,000 166 926,000 83 926,000 276

distinguished in the recent runoff pattern (Fig. 8A). The runoff during the pre–Oldest Dryas warming was much greater than the modern value. Runoff depth reached 600 mm in the basins of the Upper Dnieper and Don. It was also ~600 mm in the western part of the Dnieper River basin. In the eastern part of the Dnieper River basin, runoff depth decreased rapidly in the northsouth direction. It was ~400–600 mm at the headwaters and did not exceed 100–200 mm in the middle and lower Dnieper River basin. Runoff was more than 200 mm in the upper and middle Don River basin, and only at the lower reaches did it decrease to 100 mm. The excess of annual water flow above the modern value can be explained by greater precipitation and lower losses (greater runoff coefficient values). The rate of north-to-south decrease in runoff was lower in the past than at present. In the Dnieper River basin, the modern ratio between the runoff depth at the headwaters (200 mm) and lower reaches (20 mm) is 10; during the pre–Oldest Dryas warming, it was 6 (600 mm/100 mm). In the Don River basin, these ratios are closer to one another (150 mm/20 mm and 600 mm/100 mm) because of a smaller north-south extent of the basin. The map in Figure 8B allows one to calculate an annual volume of surface runoff for the period of large river activity and to compare it with recent characteristics (Table 6). During the pre– Oldest Dryas warming, the Don River supplied about four times the present-day water volume, and the Dnieper River supplied about three times the volume. If we apply these proportions to other rivers, the total annual surface runoff from the East European part of the Black Sea basin can be estimated at ~365 km3. Even if we keep recent values for the runoff from region III, the estimate will be more than 300 km3, i.e., almost three times greater than the modern runoff from this area and even greater than the modern runoff from the entire Black Sea drainage basin.

21

VP/VR 3.8 3.1 3.3

that the meltwater supply was at least equal to recent runoff from the Dnieper River basin, and possibly in excess. This estimate is confirmed by the morphology of the blow-out channels, through which the meltwater passed. Paleoclimatic reconstructions for the LGM are based on data from the Drichaluki site, which characterize the Upper Dnieper basin. They reveal a cold and relatively dry environment with precipitation lower than at present (see Table 4). Low winter and summer air temperatures and the existence of deep permafrost led to low precipitation losses from infiltration and evaporation. Therefore, surface runoff was higher than recently, in spite of the lower precipitation. During the Late Pleniglacial, landscape and climatic conditions over the Russian Plain were less differentiated than at present, and the so-called periglacial hyperzone was formed (Velichko, 1973). At the LGM, permafrost in the Russian Plain reached 45°N (Velichko et al., 1982), thus including the major part of the Dnieper River basin. Taking this into consideration, annual meteoritic water runoff from the Dnieper basin on the whole was estimated at 100 km3, which is about two times greater than recent totals. With the additional 60 km3 of meltwater runoff, the cumulative runoff to the Black Sea from the Dnieper River basin at the LGM was significantly larger than recent runoff from the entire East European part of the Black Sea drainage basin. Extraordinary morphological features—large meandering paleochannels—mark the next hydrological period at the end

DISCUSSION Morphological, geological, geochronological, and palynological information on the East European part of the Black Sea drainage basin enables a reconstruction of the landscape, climate, and hydrological history of this region (Fig. 9). The amount of available paleogeographical data on different time intervals varies, influencing the reliability of the suggested interpretations. The weakest evidence is related to the hydrological regimen at the LGM. The great uncertainty in the paleoglaciological reconstruction of the shape and budget of the Fennoscandian ice sheet makes the estimate of meltwater supply to the Dnieper River basin more qualitative than quantitative. It is possible to say

Figure 9. Reconstructed surface runoff in the East European part of the Black Sea drainage basin during the Last Glacial Maximum (LGM) and Late Glacial time (LGT). PB—Preboreal; DR3—Younger Dryas; AL—Allerød; BØ—Bølling; DR1—Oldest Dryas.

22

Sidorchuk et al.

of the Pleniglacial (pre–Oldest Dryas warming). Existing dates (see Table 1 in Sidorchuk et al., 2008) show that this period had begun before 15 ka (the Moskva River paleochannel). In the Dnieper and Don River basins, widths and meander wavelengths of the large paleochannels are ten to fifteen times greater than those of the modern channels with the same drainage areas. The closest modern hydrological analogue for this period is an open tundra landscape with permafrost and long, severe winters with considerable snowfall (see Fig. 6). In such landscapes, short and high spring floods with large discharges formed large meandering channels. During the rest of the year, these channels were nearly empty. Using the large paleochannel morphology, we can roughly calculate the annual surface runoff from the East European part of the Black Sea basin during the pre–Oldest Dryas warming at ~300–365 km3. Because of low differentiation of landscapes within the periglacial hyperzone, we can assume that past/recent runoff ratios, estimated in the region analogues, can be extended to the whole East European part of the Black Sea basin. This assumption is supported by the results of our reconstructions. Thus, in the region analogue located for fossil flora 2 (see Fig. 6), the ratio between the past and recent runoffs is equal to 3.5 (see Table 4). A very similar ratio (3.2) was reconstructed for the Late Pleniglacial in the Moskva River basin near Moscow (Sidorchuk et al., 2009). Therefore, using the reconstructed values of the main hydroclimatic indexes for various intervals of the LGT (see Table 4), we calculated water discharge from the East European part of the Black Sea drainage (Fig. 9). Following the maximum reached at the end of the Pleniglacial, the discharge generally decreased, although it remained greater than the modern value until the beginning of the Holocene. The decrease in water discharge was interrupted by two events of high runoff. A secondary system of large paleochannels in the Khoper River basin (see Fig. 5) was formed during the second of these events, which corresponds to the Younger Dryas. The East European part of the Black Sea basin makes up about one third of its recent basin. Of the rest of the Black Sea basin, the Danube River basin is the largest part. When the Caspian Sea was connected to the Black Sea through the Manych Strait, the Black Sea drainage basin increased dramatically. Next, we shall discuss briefly the water supply from the Danube River and from the Caspian Sea during the LGM and LGT. The Danube River basin can be divided into two main morphological regions: (1) the Alpine and subalpine region of glacial morphology, which can be used for paleoglaciological reconstruction, and (2) the sub-Carpathian region, which shows clear traces of large river activity. Within the latter, the morphology of the large paleochannels can be used for paleohydrological reconstructions. The Last Glaciation in the eastern Alps, where most of the Danube tributaries have their origin, was described in detail by van Huzen (2004). Application of Equations 1–8 to van Huzen’s estimates of the position and age of various stages of the Last Glaciation gives the following results: during the period between ca. 21 and ca. 17 14C k.y. B.P., meltwater supply to the Upper Danube was very low; it was much higher (~35 km3 yr–1) during

the short period when glaciers retreated very rapidly between ca. 17 and ca. 16.5 14C k.y. B.P., and afterward, it gradually decreased until the beginning of the Holocene. The morphology and deposits of the large paleorivers were investigated by Borsy and Felegyhazi (1983), by Kasse et al. (2000) in Hungary, and by Howard et al. (2004) in Romania. These investigations showed that large rivers in the Tisza River basin were active at the end of the Pleniglacial, i.e., at the beginning of the LGT (Dr. K. Kasse, 2008, personal commun.). Large rivers in Romania were active before 12 14C k.y. B.P. (Howard et al., 2004). These estimates correspond in general to the dates obtained in the East European Plain (Sidorchuk et al., 2008). The widths and meander lengths of radiocarbon-dated large paleochannels were measured on Landsat-7 images. Calculations using Equations 9–11 show that, during the LGT, the Carpathian tributaries of the Danube supplied three times their present-day discharge. Presumably, the runoff from the Danube River basin changed in similar ways to that in the East European part of the Black Sea basin (see Fig. 9), with its maximum shifted toward the beginning of the LGT. The history of the last connection between the Black Sea and the Caspian Sea through the Manych Strait has been analyzed by many researchers (e.g., Goretskiy, 1957; Popov, 1983; Menabde and Svitoch, 1990), but still many questions remain unsolved, mostly regarding the chronology of events and the water budget. This connection occurred at some time during the early phase of the Khvalynian transgression of the Caspian Sea, which is dated to 20–7 14C k.y. B.P. (Svitoch, 1991). The maximum level of the first stage of this transgression was about +50 m asl or +78 m above the recent level of the Caspian Sea. This stage is marked by +50 m marine terraces with the deposits containing the Khvalynian malacofauna, and by the river terraces of the tributaries of the Lower Volga River. Alluvium that built up these terraces corresponds to marine and estuarine deposits of the transgression (Obedientova, 1977). The remnants of the large paleochannels on these river terraces (Sidorchuk et al., 2009) show that their formation took place within the stage of large river activity at the LGT, which corresponds to the latest radiocarbon dates associated with the maximum phase of the Khvalynian transgression: 14– 11 ka (Svitoch and Yanina, 1997), 16–11 ka (Leonov et al., 2002), 16–13 ka (Chepalyga et al., 2008). Our calculations for that stage of the Caspian Sea water budget (Panin et al., 2005; Sidorchuk et al., 2009) show that runoff from the large rivers in the Volga River basin (~500 km3 yr–1) was sufficient to build up the highest level of the Khvalynian transgression. However, this runoff was not large enough to maintain the long-term flow of Caspian water into the Black Sea through the Manych Strait. The mean sizes of fluvial features in the Manych Straight (channel width ~8 km, length of alternating bars ~20 km, height 15–20 m, and meander half-wavelength ~40 km) support the hypothesis of a very high discharge during their formation. The mean annual discharge of ~65,000 m3 s–1 calculated for a stream with such geometry using Equation 9 and y = 100 (stable runoff from a large lake, when Qmax is close to Qa) is greater than the flood discharge at the Lower Mississippi River. Annual volume of runoff through the Manych

Surface runoff to the Black Sea from the East European Plain Strait was ~2000 km3 yr–1. Note, that Chepalyga (2007) used a different method of discharge estimation: the product of channel cross-section area and flow velocity. For the Zunda-Tolga profile, cored by Popov (1983), the cross-section area was estimated as 250,000 m2 (width of 10,000 m on a mean depth of 25 m). Flow velocity was estimated at 0.2 m s–1 according to sediment particle size. Calculated discharge (Chepalyga, 2007) was 50,000 m3 s–1. That is rather close to our estimate, taking into account the completely different approach to figuring the result. The most probable hypothesis that can explain these facts is that a high threshold with its top at about +50 m asl existed in the middle of the Manych Strait. This natural dam separated the high-standing Caspian Sea from the relatively low-standing Black Sea (Menabde and Svitoch, 1990). At some moment, the dam was eroded to a level of about +22 m asl. As a result, the Caspian Sea water from between +50 and +22 m asl (~23,000 km3) was emptied into the Black Sea during a period of only 20–30 yr. This event happened within the large river stage. This assumption is supported by the existence of remnants of large paleochannels on the terraces of the tributaries of the Volga River, which were formed both before and after incision of those tributaries due to the Caspian Sea level drop (Sidorchuk et al., 2009). Some radiocarbon dates indicate that the flow via the Manych Strait took place between the beginning of the late Valdai/late Weichselian deglaciation and 12 ka (Arslanov and Yanina, 2008). Our reconstruction of high river runoff into the Black Sea at the LGM and LGT (2–2.5 times greater than the present; see Fig. 9) does not support the “flood hypothesis” proposed by Ryan et al. (1997). This hypothesis holds that the –120 to –150 m stage of the Black Sea lasted until ca. 7 14C k.y. B.P. The estimated high river runoff does not agree with a low level and isolated-lake status of the Black Sea throughout the LGM and LGT. Today, the Black Sea has a positive water balance with some 300 km3 of excess water removed annually via the Bosphorus (Özsoy et al., 1995). To maintain a low stage at the LGM and LGT, the values of effective evaporation should have been much higher than today. Reconstructions of Tarasov et al. (1999) for the LGM provide quite the opposite result, suggesting a rate of evaporation that was lower than present. Low evaporation aided by high water supply from the drainage basin make it more probable that the Black Sea was filled up to the Bosphorus sill (now ~35 m below sea level) in the LGT, with large amounts of excess freshwater flowing via this channel straight into the Marmara Sea. This is supported by recent findings of Giosan et al. (2009) that the subaerial deltaic plain of the Danube River, which is indicative of the sea-level position, was located around 30 m below sea level directly before the Black Sea to world ocean reconnection was established at ca. 8.4 ka. CONCLUSIONS Three time intervals can be distinguished within the general sequence of hydrological events in the East European part of the Black Sea drainage basin during the LGM and LGT.

23

1. A cold and dry interval with precipitation lower than today and a high runoff coefficient caused by low evaporation losses and continuous spread of permafrost correlates with the LGM and the beginning of deglaciation, 18–15 14C k.y. B.P. A relatively high surface runoff (210 km3 yr–1) caused by precipitation was then supplemented by meltwater influx (60 km3 yr–1) from proglacial lakes. 2. A warmer, humid interval with precipitation exceeding that of today and a high runoff coefficient caused by relatively low temperatures correlates with the pre–Oldest Dryas time, 15–14 14C k.y. B.P. This was a period of very high surface runoff (300–365 km3 yr–1), when large meandering river channels were formed. A short event occurred within this period when a large amount of Caspian water (23,000 km3) was supplied to the Black Sea through the Manych Strait. 3. An interval with short-term climatic oscillations and nonsteady decrease of surface runoff, from ca. 14 to 10 14C k.y. B.P., corresponds to the LGT. Secondary climatic oscillations during the LGT were expressed against the overall trend toward warming, primarily the rise of winter temperatures. Phases with the lowest air temperatures (the Oldest and Younger Dryas) were favorable for permafrost development and to low losses, providing high runoff coefficients and, therefore, increasing the runoff (up to 355 km3 yr–1) despite a decrease in precipitation. Phases with relatively high air temperatures (Bølling and Allerød interstadials) were favorable for permafrost degradation and higher losses to evapotranspiration, providing low runoff coefficients and therefore decreasing the runoff (to 165 km3 yr–1) despite an increase in precipitation. By the end of this time interval, large meandering paleochannels were transformed into smaller channels. The entire LGM-LGT interval was characterized by high surface runoff from the East European part of the Black Sea drainage basin compared to that of the present day (110 km3 yr–1), which was achieved at the beginning of the Holocene. ACKNOWLEDGMENTS Financial support for this work was received from the Russian Foundation of Basic Research (RFBR), project no. 97-05-64708, 00-05-64021, and 09-05-00340. Valuable suggestions by reviewers J. Herget and T. Kalicki were incorporated into the text. REFERENCES CITED Aksu, A.E., Hiscott, R.N., Yaşar, D., Işler, F.I., and Marsh, S., 2002, Seismic stratigraphy of late Quaternary deposits from the southwestern Black Sea shelf: Evidence for non-catastrophic variations in sea-level during the last ~10,000 yr: Marine Geology, v. 190, p. 61–94, doi: 10.1016/S0025 -3227(02)00343-2. Arslanov, Kh., and Yanina, T., 2008, Radiocarbon age of the Khvalynian Manych passage, in Gilbert, A.S., and Yanko-Hombach, V., eds., Extended Abstracts of the Fourth Plenary Meeting and Field Trip of IGCP 521 “Black Sea–Mediterranean Corridor during the Last 30 k.y.: Sea Level Change and Human Adaptation,” and INQUA 0501 “Caspian–Black Sea– Mediterranean Corridor during the Last 30 k.y.: Sea Level Change and Human Adaptive Strategies” (4–16 October 2008; Bucharest, Romania, and Varna, Bulgaria): Bucharest, GeoEcoMar, p. 15–18.

24

Sidorchuk et al.

Bahr, A., Arz, H.W., Lamy, F., and Wefer, G., 2006, Late Glacial to Holocene paleoenvironmental evolution of the Black Sea, reconstructed with stable oxygen isotope records obtained on ostracod shells: Earth and Planetary Science Letters, v. 241, p. 863–875, doi: 10.1016/j.epsl.2005.10.036. Borisova, O.K., and Novenko, E.Yu., 1999, Sreda obitaniya pozdnepaleoliticheskogo cheloveka na stoyanke Yudinovo (po palinologicheskim dannym) [Environmental conditions at the Late Paleolithic site Yudinovo (based on palynological data)], in Bolikhovskaya, N.S., and Rovnina, L.V., eds., Aktual’nye problemy palinologii na rubezhe tret’ego tysyacheletiya [Topical Problems in Palynology at the Beginning of the Third Millennium]: Moscow, IGIRGI Press, p. 54–60 (in Russian). Borisova, O.K., Sidorchuk, A.Yu., and Panin, A.V., 2006, Palaeohydrology of the Seim River basin, Mid-Russian Upland, based on palaeochannel morphology and palynological data: Catena, v. 66, p. 53–73, doi: 10.1016/ j.catena.2005.07.010. Borsy, Z., and Felegyhazi, E., 1983, Evolution of the network of water courses in the north-eastern part of the Great Hungarian Plain from the end of the Pleistocene to our days: Quaternary Studies in Poland, v. 4, p. 115–124. Chebotareva, N.S., and Faustova, M.A., 1982, Struktura i dinamika poslednego Evropeiskogo lednikovogo pokrova [Structure and dynamics of the last European ice-sheet], in Gerasimov, I.P., and Velichko, A.A., eds., Paleogeografiya Evropy za poslednie sto tysyach let (Atlas-monografiya) [Palaeogeography of Europe during the Last 100,000 Years (AtlasMonograph)]: Moscow, Nauka, p. 36–48 (in Russian). Chebotareva, N.S., and Makarycheva, I.A., 1974, Poslednee oledenenie Evropy i ego geokhronologiya [Last Glaciation in Europe and Its Geochronology]: Moscow, Nauka, 216 p. (in Russian). Chepalyga, A.L., 2007, The Late Glacial Great Flood in the Ponto-Caspian basin, in Yanko-Hombach, V., Gilbert, A.S., Panin, N., and Dolukhanov, P.M., eds., The Black Sea Flood Question: Changes in Coastline, Climate, and Human Settlement: Dordrecht, Springer, p. 119–148. Chepalyga, A.L., Arslanov, Kh., and Svetlitskaya, T., 2008, Chronology of the Khvalynian sea-level oscillations: New data and approach, in Gilbert, A.S., and Yanko-Hombach, V., eds., Extended Abstracts of the Fourth Plenary Meeting and Field Trip of IGCP 521 “Black Sea–Mediterranean Corridor during the Last 30 k.y.: Sea Level Change and Human Adaptation,” and INQUA 0501 “Caspian–Black Sea–Mediterranean Corridor during the Last 30 k.y.: Sea Level Change and Human Adaptive Strategies” (4–16 October 2008; Bucharest, Romania, and Varna, Bulgaria): Bucharest, GeoEcoMar, p. 32–34. Dolukhanov, P.M., Chepalyga, A.L., Lavrentiev, N.V., Kadurin, S.V., Larchenkov, E.P., and Shkatova, V.K., 2008, Black Sea and Caspian Basins in late Pleistocene: sea-level changes, climate and early human settlement, in Gilbert, A.S., and Yanko-Hombach, V., eds., Extended Abstracts of the Fourth Plenary Meeting and Field Trip of IGCP 521 “Black Sea–Mediterranean Corridor during the Last 30 k.y.: Sea Level Change and Human Adaptation,” and INQUA 0501 “Caspian–Black Sea–Mediterranean Corridor during the Last 30 k.y.: Sea Level Change and Human Adaptive Strategies” (4–16 October 2008; Bucharest, Romania, and Varna, Bulgaria): Bucharest, GeoEcoMar, p. 44–47. Domanitskiy, A.P., Dubrovina, R.G., and Isayeva, A.I., 1971, Reki i ozera Sovetskogo Soyuza [Rivers and Lakes of the Soviet Union]: Leningrad, Gidrometeoizdat, 104 p. (in Russian). Dury, G.H., 1964, Principles of Underfit Streams: U.S. Geological Survey Professional Paper 452-A, 67 p. Dury, G.H., 1965, Theoretical Implications of Underfit Streams: U.S. Geological Survey Professional Paper 452-C, 43 p. Evstigneev, V.M., 1990, Rechnoy stok i gydrologicheskiye raschety [River Flow and Hydrological Prognosis]: Moscow, Moscow University Press, 304 p. (in Russian). Faustova, M.A., 1984, Late Pleistocene glaciation of European USSR, in Velichko, A.A., ed., Late Quaternary Environments of the Soviet Union: Minneapolis, University of Minnesota Press, p. 3–12. Faustova, M.A., and Chebotareva, N.S., 1969, Degradatsiya valdaiskogo oledeneniya, Belorusskaya SSR i Smolenskaya oblast’ [Degradation of the Valdai glaciation, Belorussian SSR and Smolensk region], in Gerasimov, I.P., ed., Posledniy lednikovyi pokrov na severo-zapade Evropeiskoi chasti SSSR [The Last Ice Sheet in the Northwestern European Part of the USSR]: Moscow, Nauka, p. 151–176 (in Russian). Gerasimov, I.P., ed., 1969, Posledniy lednikovyi pokrov na severo-zapade Evropeiskoi chasti SSSR [The Last Ice Sheet in the Northwestern European Part of the USSR]: Moscow, Nauka, 322 p. (in Russian).

Giosan, L., Filip, F., and Constatinescu, S., 2009, Was the Black Sea catastrophically flooded in the early Holocene?: Quaternary Science Reviews, v. 28, p. 1–6, doi: 10.1016/j.quascirev.2008.10.012. Goretskiy, G.I., 1957, O sootnoshenii morskikh i kontinental’nykh osadkov Priazoviya, Primanychya i Nizhnego Pridonya [On the correlation of marine and continental deposits in the Azov, Manych and lower Don basins]: Trudy Komissii po Izucheniyu Chetvertichnogo Perioda [Proceedings of the Commission on Quaternary Studies], v. 13, p. 36–54 (in Russian). Grichuk, V.P., 1969, Glyatsial’nye flory i ikh klassifikatsiya [Glacial floras and their classification], in Gerasimov, I.P., ed., Posledniy lednikovyi pokrov na severo-zapade Evropeiskoi chasti SSSR [The Last Ice Sheet in the Northwestern European Part of the USSR]: Moscow, Nauka, p. 57–70 (in Russian). Grichuk, V.P., 1982, Rastitel’nost’ Evropy v pozdnem pleistotsene [Vegetation of Europe during the late Pleistocene], in Gerasimov, I.P., and Velichko, A.A., eds., Paleogeografiya Evropy za poslednie sto tysyach let (Atlasmonografiya) [Palaeogeography of Europe during the Last 100,000 Years (Atlas-Monograph)]: Moscow, Nauka, p. 92–109 (in Russian). Hintikka, V., 1963, Über das Großklima einiger Pflanzenareale in zwei Klimakoordinatensystemen dargestellt: Annales Botanici Societatis Zoologicae, Botanicae Fennicae ‘Vanamo,’ v. 34, p. 1–64. Howard, A.J., Macklin, M.G., Bailey, D.W., Mills, S., and Andreescu, R., 2004, Late-Glacial and Holocene river development in the Teleorman Valley on the southern Romanian Plain: Journal of Quaternary Science, v. 19, no. 3, p. 271–280, doi: 10.1002/jqs.805. Iversen, J., 1944, Viscum, Hedera, and Ilex as climate indicators. A contribution to the study of the post-glacial temperature climate: Geologiske Förenigens i Stockholm Förhandlingan, v. 66, p. 463–483. Kalicki, T., 1995, Late Glacial and Holocene evolution of some river valleys in Byelorussia: Paläoklimaforschung, v. 14, p. 89–100. Kalicki, T., and San’ko, A., 1992, Genesis and age of the terraces of the Dnieper River between Orsha and Shklow, Byelorussia: Geographia Polonica, v. 60, p. 151–174. Kalinin, G.P., Markov, K.K., and Suetova, I.A., 1966, Kolebaniya urovnya vodoemov Zemli v nedavnem geologicheskom proshlom: Okeanologiya [Oceanology], v. 6, p. 737–746. Kapitsa, A.P., 1958, O zavisimosti formy lednikovogo kupola Vostochnoi Antarktidy ot reliefa podlednogo lozha i kharaktera rastekaniya l’da [On the dependence of the East Antarctic ice-sheet shape on the underlying topography and ice rheology]: Informatsionnyi Bulleten’ Sovetskoi Antarkticheskoi Ekspeditsii [Information Bulletin of the Soviet Antarctic Expedition], v. 1, p. 1–17. Kasse, C., Vandenberghe, J., Beets, C., and Van Huissteden, J., 2000, Changing fluvial styles during the last glacial-interglacial transition in the Tisza valley, Hungary, in Georgiadi, A., ed., Conference Papers and Abstracts Proceedings of the Fourth International Meeting on Global Continental Palaeohydrology (GLOCOPH 2000), Hydrological Consequences of Global Climate Changes: Moscow, Institute of Geography of Russian Academy of Sciences, p. 34–36. Khodakov, V.G., 1982, Aktualisticheskaya model’ Evropeiskogo pokrovnogo lednika [An actualistic model of the European ice-sheet], in Gerasimov, I.P., and Velichko, A.A., eds., Paleogeografiya Evropy za poslednie sto tysyach let (Atlas-monografiya) [Palaeogeography of Europe during the Last 100,000 Years (Atlas-Monograph)]: Moscow, Nauka, p. 48–62 (in Russian). Kislov, A.V., and Toropov, P.A., 2006, Modeling of variations in river runoff on the East European Plain under different climates of the past: Water Resources, v. 33, p. 471–482, doi: 10.1134/S0097807806050010. Kroonenberg, S.B., Rusakov, G.V., and Svitoch, A.A., 1997, The wandering of the Volga delta: A response to rapid Caspian sea-level change: Sedimentary Geology, v. 107, p. 189–209, doi: 10.1016/S0037-0738(96)00028-0. Kvasov, D.D., 1979, The late Quaternary history of large lakes and inland seas of Eastern Europe: Annales Academiae Scientarium Fennicae, ser. A III, v. 127, p. 1–71. Lane-Serff, G.F., Rohling, E.J., Bryden, H.L., and Charnock, H., 1997, Postglacial connection of the Black Sea to the Mediterranean and its relation to the timing of sapropel formation: Paleoceanography, v. 12, p. 169–174, doi: 10.1029/96PA03934. Leonov, Yu.G., Lavrushin, Yu.A., Antipov, M.P., Spiridonova, Ye.A., Kuzmin, Ya.V., Jull, E.J.T., Burr, S., Jelinowska, A., and Chalie, F., 2002, New age data on sediments of the transgressive phase of the early Khvalyn transgression of the Caspian Sea: Doklady Earth Sciences, v. 386, p. 748–751. Mangerud, J., Jakobsson, M., Alexanderson, H., Astakhov, V., Clarke, G., Henriksen, M., Hjort, C., Krinner, G., Lunkka, J.-P., Möller, P., Murray, A.,

Surface runoff to the Black Sea from the East European Plain Nikolskaya, O., Saarnisto, M., and Svendsen, J.I., 2004, Ice-dammed lakes and rerouting of the drainage of northern Eurasia during the last glaciation: Quaternary Science Reviews, v. 23, p. 1313–1332, doi: 10.1016/ j.quascirev.2003.12.009. Markov, K.K., and Suetova, I.A., 1964, Evstaticheskiye kolebaniya urovnya okeana [Eustatic changes of the ocean level], in Sovremennye Problemy Geografii [Current Problems in Geography]: Moscow, Nauka, p. 24–34 (in Russian). Menabde, I.V., and Svitoch, A.A., 1990, O kharaktere soedineniya Kaspiyskogo i Chernogo morei v pozdnem pleistotsene [On the character of the connection between the Caspian and Black Sea in the late Pleistocene], in Lebedev, L.I., and Maev E.G., eds., Kaspiyskoye More: Voprosy Geologii i Geomorfologii [Caspian Sea: Questions in Geology and Geomorphology]: Moscow, Nauka, p. 34–41 (in Russian). Obedientova, G.V., 1977, Erozionnye Tsikly i Formirovanie Doliny Volgi [Erosion Cycles and Formation of the Volga River Valley]: Moscow, Nauka, 239 p. (in Russian). Ostrovsky, A.B., 1982, Paleogeograficheskiye criterii prichin noveyshikh izmeneniy paleogidrologicheskogo rezhima vnutrikontinental’nykh morei [Palaeogeographical causes of the changes in the palaeohydrological regime of the intra-continental seas], in Kaplin, P.A., Klige, R.K., and Chepalyga, A.L., eds., Kolebaniya urovnya morei i okeanov za 15000 let [Level Changes of the Seas and Oceans during the Last 15,000 Years]: Moscow, Nauka, p. 102–112 (in Russian). Özsoy, E., Latif, M.A., Turul, S., and Ünlüata, Ü., 1995, Exchanges with the Mediterranean, fluxes and boundary mixing processes in the Black Sea, in Briand, F., ed., Mediterranean Tributary Seas: Bulletin de l’Institut Océanographique, Monaco, Special No 15, CIESME (International Commission for the Scientific Exploration of the Mediterranean Sea) Science Series 1, p. 1–25. Panin, A.V., Sidorchuk, A.Yu., and Borisova, O.K., 2005, Flyuvial’nye protsessy i rechnoi stok na Russkoi ravnine v kontse pozdnevaldaiskoi epokhi [Fluvial processes and river runoff in the Russian Plain in the end of the late Valdai glacial epoch], in Yanina, T.A., ed., Gorizonty Geografii [Horizons of Geography]: Moscow, Moscow University Press, p. 114– 127 (in Russian). Peltier, W.R., 1994, Ice age paleotopography: Science, v. 265, p. 195–201, doi: 10.1126/science.265.5169.195. Popov, G.I., 1983, Pleistotsen Chernomorsko-Kaspiyskikh Prolivov [The Pleistocene of Black Sea–Caspian Straits]: Moscow, Nauka, 216 p. (in Russian). Ryan, W.B.F., Pitman, W.C., III, Major, C.O., Shimkus, K., Moskalenko, V., Jones, G.A., Dimitrov, P., Görür, N., Sakınç, M., and Yüce, H., 1997, An abrupt drowning of the Black Sea shelf: Marine Geology, v. 138, p. 119– 126, doi: 10.1016/S0025-3227(97)00007-8. San’ko, A.F., 1987, Neopleistotsen Severo-vostochnoi Belorussii i Smezhnykh Raionov RSFSR [The Neopleistocene of North-East Byelorussia and Neighboring Districts of the RSFSR]: Minsk, Nauka i Tekhnika, 178 p. (in Russian). Sidorchuk, A.Yu., and Borisova, O.K., 2000, Method of palaeogeographical analogues in palaeohydrological reconstructions: Quaternary International, v. 72, p. 95–106, doi: 10.1016/S1040-6182(00)00025-2. Sidorchuk, A.Yu., Borisova, O.K., and Panin, A.V., 2001, Fluvial response to the late Valdai/Holocene environmental change on the East European Plain: Global and Planetary Change, v. 28, p. 303–318, doi: 10.1016/ S0921-8181(00)00081-3. Sidorchuk, A.Yu., Panin, A.V., and Borisova, O.K., 2008, Climate-induced changes in river runoff on the north-Eurasian plains during the Late Glacial and Holocene: Water Resources, v. 35, p. 386–396, doi: 10.1134/ S0097807808040027. Sidorchuk, A.Yu., Panin, A.V., and Borisova, O.K., 2009, Morphology of river channels and surface runoff in the Volga River basin (East European Plain) during the Late Glacial period: Geomorphology, v. 113, p. 137–157, doi: 10.1016/j.geomorph.2009.03.007. Svendsen, J., Alexanderson, H., Astakhov, V., Demidov, I., Dowdeswell, J., Funder, S., Gataullin, V., Henriksen, M., Hjort, C., Houmark-Nielsen, M., Hubberten, H., Ingólfson, O., Jakobsson, M., Kjær, K., Larsen, E., Lokrantz, H., Lunkka, J., Lyså, A., Mangerud, J., Matiouchkov, A., Murray, A., Möller, P., Niessen, F., Nikolskaya, O., Polyak, P., Saarnisto, M., Siegert, C., Siegert, M., Spielhagen, R., and Stein, R., 2004, Late Quaternary ice sheet history of northern Eurasia: Quaternary Science Reviews, v. 23, p. 1229–1271, doi: 10.1016/j.quascirev.2003.12.008.

25

Svezhentsev, Y.S., 1993, Radiocarbon chronology for the Upper Paleolithic sites on the East European Plain, in Soffer, O., and Praslov, N.D., eds., From Kostenki to Clovis: New York, Plenum Press, p. 23–30. Svitoch, A.A., 1991, Kolebaniya urovnya Kaspiyskogo morya v pleistotsene [Fluctuations of the Caspian Sea level in the Pleistocene], in Scherbakov, F.A., and Svitoch, A.A., eds., Paleogeografiya i Geomorfologiya Kaspiyskogo Regiona v Pleistotsene [Palaeogeography and Geomorphology of the Caspian Region in the Pleistocene]: Moscow, Nauka, p. 5–100 (in Russian). Svitoch, A.A., and Yanina, T.A., 1997, Chetvertichnye otlozhenia poberezhiy Kaspiyskogo morya [Quaternary Sediments on the Caspian Sea Coasts]: Moscow, Rosselkhozacademia Press, 270 p. (in Russian). Szafer, W., 1946–1947, Flora plioceńska z Krościenka nad Dunajcem (The Pliocene flora of Kroscienko on Dunajec River): Rozprawy Wydzialu matemat.-przyrodn: Polska Akad. Umiejet., Krakow, 72 B, 1, 162 p.; 2, 213 p. (in Polish). Tarasov, P.E., Peyron, O., Guiot, J., Brewer, S.V., Volkova, V.S., Bezusko, L.G., Dorofeyuk, N.I., Kvavadze, E.V., Osipova, I.M., and Panova, N.K., 1999, Last Glacial Maximum climate of the former Soviet Union and Mongolia reconstructed from pollen and plant macrofossil data: Climate Dynamics, v. 15, p. 227–240, doi: 10.1007/s003820050278. van Huzen, D., 2004, Quaternary glaciations in Austria, in Ehlers, J., and Gibbard, P.L., eds., Quaternary Glaciations—Extent and Chronology: Amsterdam, Elsevier, p. 1–13. Varuschenko, S.I., Varuschenko, A.N., and Klige, R.K., 1987, Izmeneniya rezhima Kaspiyskogo morya i besstochnykh vodoyemov v paleovremeni [Changes of the Caspian Sea and Other Closed Lake Regimes in PalaeoTime]: Moscow, Nauka, 240 p. (in Russian). Velichkevich, F.Yu., 1982, Pleistotsenovye Flory Lednikovykh Oblastei Vostochno-Evropeiskoi Ravniny [Pleistocene Floras of the Glaciated Districts of the East European Plain]: Minsk, Nauka i Tekhnika, 239 p. (in Russian). Velichko, A.A., 1973, Prirodnyi protsess v pleistotsene [Natural Process in the Pleistocene]: Moscow, Nauka, 256 p. (in Russian). Velichko, A.A., Berdnikov, V.V., and Nechaev, V.P., 1982, Rekonstruktsiya zony mnogoletnei merzloty i etapov eyo razvitiya [Reconstruction of the permafrost zone and stages of its development], in Gerasimov, I.P., and Velichko, A.A., eds., Paleogeografiya Evropy za poslednie sto tysyach let (Atlas-monografiya) [Palaeogeography of Europe during the Last 100,000 Years (Atlas-Monograph)]: Moscow, Nauka, p. 74–81 (in Russian). Velichko, A.A., Faustova, M.A., Gribchenko, Yu.N., Pisareva, V.V., and Sudakova, N.G., 2004, Glaciations of the East European Plain—Distribution and chronology, in Ehlers, J., and Gibbard, P.L., eds., Quaternary Glaciations—Extent and Chronology: Amsterdam, Elsevier, p. 337–354. Vlasov, V.K., and Kulikov, O.A., 1988, Radiotermolyuminestsentnyi metod datirovaniya rykhlykh otlozheniy [Radio-thermo-luminescence method of Non-Consolidated Sediment Dating]: Moscow, Moscow State University Press, 72 p. (in Russian). Volkov, I.A., 1960, O nedavnem proshlom rek Ishim i Nura [On the recent past of the rivers Ishim and Nura]: Trudy Laboratorii Aerometodov AN SSSR [Proceedings of the Laboratory of Aeromethods, USSR Academy of Sciences], v. 9, p. 15–19 (in Russian). Volkov, I.A., 1963, Sledy moschnogo stoka v dolinakh yuga Zapadnoi Sibiri [An evidence of the powerful flow in the valleys of west Siberian rivers]: Doklady AN SSSR [Reports of the USSR Academy of Sciences], v. 151, p. 648–651 (in Russian). Walker, M.J.C., Björck, S., Lowe, J.J., Cwynar, L.C., Knudsen, K.-L., Wohlfarth, B., and INTIMATE Group I, 1999, Isotopic ‘events’ in the GRIP ice core: A stratotype for the late Pleistocene: Quaternary Science Reviews, v. 18, p. 1143–1150. Winguth, C., Wong, H.K., Panin, N., Dinu, C., Georgescu, P., Ungureanu, G., Krugliakov, V.V., and Podshuveit, V., 2000, Upper Quaternary water level history and sedimentation in the northwestern Black Sea: Marine Geology, v. 167, p. 127–146, doi: 10.1016/S0025-3227(00)00024-4. Zagwijn, W.H., 1996, An analysis of Eemian climate in Western and Central Europe: Quaternary Science Reviews, v. 15, p. 451–469, doi: 10.1016/0277-3791(96)00011-X.

MANUSCRIPT ACCEPTED BY THE SOCIETY 22 JUNE 2010 Printed in the USA

The Geological Society of America Special Paper 473 2011

Modeling extreme Black Sea and Caspian Sea levels of the past 21,000 years with general circulation models Alexander Kislov Pavel Toropov Department of Meteorology and Climatology, Faculty of Geography, M.V. Lomonosov Moscow State University, Leninskiye Gory, Moscow 119992, Russia

ABSTRACT This paper describes the relationship between sea levels and climate based on the links between sea-level variations and river runoff. During the final late Pleistocene and postglacial periods, the Caspian Sea fluctuated between regression and transgression stages. The Black Sea experienced fluctuations as well, but these were mainly controlled by the world ocean due to water exchange through the Bosporus Strait. Sometimes, the Caspian Sea overflowed into the Black Sea through the Manych Strait, and they periodically coalesced. Change in the level of both seas could be interpreted as responses to the regional-scale water budget (the balance between inflow and outflow components). These components can be calculated from atmospheric general circulation models. This approach uses climate modeling data to reproduce river runoff changes, and, consequently, variations in seawater and sea level under contrasting climate conditions. In response to glacial conditions of the last cold Pleistocene event, the lowering levels of the Black Sea (post-Karangatian regression stage) and the Caspian Sea (Atelian regression stage) are simulated simultaneously. This lends credence to the idea of the connection between deep regression states of the Caspian and Black Seas and mature stages of the late Quaternary glacial/cooling/drying planetary events. Analysis of observed information allows us to conclude—taking into account the uncertainties of reconstructed data—that at least two regression stages occurred simultaneously with late Quaternary glacial planetary events. The simulation of transgression stages (their onset and duration) remains a very difficult problem. Results of modeling have shown that during the warm periods (taking as examples the mid-Holocene and Allerød events), simulated river runoff did not increase to the extent needed for a strong transgression and overflow of the Caspian Sea into the Black Sea through the Manych Strait. Thus, there is no clear understanding about the source of “additional” water volume necessary to elevate the level of the Caspian Sea to a point that would permit overflow into the Black Sea.

Kislov, A., and Toropov, P., 2011, Modeling extreme Black Sea and Caspian Sea levels of the past 21,000 years with general circulation models, in Buynevich, I.V., Yanko-Hombach, V., Gilbert, A.S., and Martin, R.E., eds., Geology and Geoarchaeology of the Black Sea Region: Beyond the Flood Hypothesis: Geological Society of America Special Paper 473, p. 27–32, doi: 10.1130/2011.2473(02). For permission to copy, contact [email protected]. © 2011 The Geological Society of America. All rights reserved.

27

28

Kislov and Toropov

INTRODUCTION The targets of our exploration are the Caspian Sea and the Black Sea (including the Sea of Azov). This paper describes the relationship between sea level and climate based on links between variations in sea level and the runoff from rivers. During the final late Pleistocene and postglacial periods, the Caspian Sea fluctuated between regressive and transgressive stages. The Black Sea experienced fluctuations as well, but these were mainly controlled by the world ocean due to water exchange through the Bosporus Strait. Sometimes, the Caspian Sea overflowed into the Black Sea through the Manych Strait (Chepalyga, 2007), and they periodically coalesced. The reasons for these sea-level changes could be different, and such a problem can be considered from the perspective of responses to the regional-scale water budget based upon planetary climate changes. The variability of seawater mass is a function of the balance between inflow and outflow components, and these factors are functions of the climate regime. Therefore, they could be calculated from atmospheric general circulation models (GCMs). This approach would allow the use of climate modeling data to reproduce river runoff changes, and, consequently, variations in seawater and sea levels under contrasting climate conditions. This paper is organized as follows: first, previous studies are reviewed; second, global climate changes during the Quaternary and their causes are presented; third, a description of the models is presented. The discussion then considers the closed-basin model that addresses the Caspian and Black Seas. The next sections present results from a seawater balance model that addresses the effects of drainage basin runoff changes and sea levels. The last section examines the paradox of “additional water.”

GLOBAL CLIMATE CHANGES DURING THE QUATERNARY The Quaternary has been characterized by both cold and warm phases. One candidate for a forcing agent that could produce such pronounced global climate variations is the Milankovitch mechanism (Berger, 1988). According to this theory, Earth’s orbital parameters change due to the influence of the Moon, Sun, and planets. Over 100,000 and 400,000 yr periods, eccentricity slowly varies, inducing small changes in the mean annual total insolation received by Earth. Obliquity oscillates from ~22° to ~25° over a 41,000 yr period, and the position of the equinoxes precesses relative to the perihelion with 19,000 and 23,000 yr periodicities. Obliquity and precession do not lead to global changes in mean annual energy but strongly modulate the seasonal pattern of insolation. The evidence for Milankovitch forcing of climate changes may be questioned for several reasons. First, there is no evidence that the climate cycles are periodic rather than aperiodic. Core records, both ice and deep sea, suggest that the dominant character is that of a random red-noise process. Second, much of the energy in low-frequency climate change occurs at periods around 100 ka, where the insolation forcing is very weak. The contribution of the Milankovitch periodicities (41,000, 19,000, and 23,000 yr periods) to climate change provides only a small fraction (15%–20%) of total climate variance. At times when orbital agents work synchronously, however, climatic response to variation in solar insolation can be distinguished from such noise. This effect led to the transition from the cold late Pleistocene to the warm Holocene, a transition that was not gradual but instead complicated by short-term events—e.g., the Allerød (AL)–Younger Dryas (YD) cycle. Many authors link the origin of these cycles to the behavior of the Atlantic thermohaline circulation.

PREVIOUS STUDIES RESULTS Various conceptual models have been used to link climate change to lake level (h), lake surface area (f), and catchment area (F) (e.g., Kalinin, 1968; Street-Perrot and Harrison, 1985; Benson and Paillet, 1989). Kalinin (1968) demonstrated that h asymptotically comes to an equilibrium level that is determined by the steady-state water-budget condition. Street-Perrot and Harrison (1985) classified closed lakes using functions relating f and F based on precipitation onto and evaporation from a lake surface and catchment (runoff) processes. Benson and Paillet (1989) argued that topographic constraints in lakes with more than one subbasin mean that lake area is the most appropriate measure of lake response to hydrologic balance. Few studies have been able to quantify the lake-climate relationship with precision for individual lakes. Functions of h, f, or precipitation and evaporation relationships have been used to evaluate climate model simulations (e.g., Kislov and Sourkova, 1998; Qin et al., 1998) and diagnose past climate (e.g., Harrison et al., 1993, 1996; Jones et al., 2001).

We studied the history of sea-level variation as influenced by climate conditions based on the results of climate modeling. Numerical experiments were used to understand how climatic parameters important to the Black and Caspian Sea water budget might change due to external forces. A modeling initiative, the Paleoclimate Modeling Intercomparison Project (PMIP), has focused on two slices of the past: (1) the mid-Holocene (6 ka calendar yr B.P. or ca. 5.3 ka radiocarbon yr B.P.), and (2) the last cold event of the late Quaternary (21 ka calendar yr B.P. or ca. 18 ka radiocarbon yr B.P.) because climatic conditions were remarkably different at those times, and abundant data are available that describe their environmental properties. In addition, both the Allerød (ca. 14.5 ka calendar yr B.P.) and the Younger Dryas (ca. 12 ka calendar yr B.P.) were studied as examples of short-scale variability. These slices (together with the modern state) are used in this paper to assess the linkage between climate variability and hydrological regime.

Modeling extreme Black Sea and Caspian Sea levels of the past 21,000 years A GCM (general circulation model) consists of an atmospheric model interactively coupled to submodels of the ocean, sea ice, land surface, and soil. Models have a typical horizontal resolution ~2°–4° latitude × longitude. Designations of the PMIP GCMs are presented in Kislov and Toropov (2007). Table 1 specifies all model boundary conditions and parameters. Closed-Basin Water-Budget Model and Data Quality We calculated changes in sea-surface area for each climatic event assuming that the closed sea was in hydrologic equilibrium with climate conditions. This is a reasonable assumption when one considers the impact of gradual climatic change on a sea with a short hydrologic response time compared to the typical rate of change for external forces. The steady-state equation for the annual water budget for a closed basin has the form: ef = YF,

(1)

where e = E – P, P is on-sea precipitation (m/yr), E is the evaporation (m/yr) per unit of lake area, Y is the runoff (m/yr) per unit of basin area, F is the drainage basin area, and f is the sea-surface area. Equation 1 assumes that the net groundwater flux into or out of the sea was probably minimal. Variation in lake area relative to the present status (denoted by index “0”) may be expressed in the form:

Δf ΔY ΔF Δe . = + − f0 Y0 F0 e0

(2)

This allows for the evaluation of the contribution of different factors toward the change in level (h) using information about lake size, bathymetry, and the surrounding topography as: Δh = (Δh)Y + (Δh)F + (Δh)e.

(3)

Evaluation of these factors can be undertaken using different approaches. Information about the change in catchment area

29

(Δh)F can be extracted from paleogeographical data, or it can be calculated based on simulation of Earth’s surface paleotopography. Level change (Δh)Y + (Δh)e due to both runoff change and change in e is calculated based on data acquired through numerical simulations of GCMs. Under current climate conditions, the volume of water annually delivered by the rivers of the East European Plain to the Black Sea is 312 km3 (total runoff to the Black Sea is 338 km3). The largest rivers are the Danube, Dneiper, Southern Bug, and Don. Their contribution is ~90% of the total mean volume of river runoff. The river flow is determined by the balance between precipitation and evaporation; hence, river runoff responds immediately to climate changes. It is important that both precipitation and evaporation be calculated using a GCM. Moreover, on a large plain such as the East European Plain, GCM data reflect the state of the climate better than areas with a mosaic of complex surface conditions. The contribution of runoff represents a large fraction (~50%) of inflow volume. Contributions from sea-surface precipitation and inflow from the Sea of Marmara represent 30% and 20%, respectively. Sea evaporation is 430 km3 (~53% of outflow volume), and release of water to the Sea of Marmara comprises ~47% of the outflow volume. The Caspian Sea (a vast inland lake) is fed by several rivers. Their annual runoff from the East European Plain is 274 km3. The greatest contribution (more than 80% of the mean total volume of runoff) derives from the Volga River. Other principal components of the annual water budget are precipitation over the sea (which adds 76 km3) and evaporation from the sea surface (which removes 362 km3). A special concern for climate modeling is the quality of the information. The data presented in Kislov and Toropov (2007) indicate how well PMIP GCMs simulate today’s river runoff within the basins of the Caspian Sea and Black Sea. Results were considered “successful” only if the errors of modeled runoff volume for each basin lay within 20%, because this variability does not fall outside the limits of natural variability. According to this classification, the most “successful” GCMs were chosen. Data from these models were examined more closely during modeling of other climatic regimes.

TABLE 1. PALEOCLIMATE MODELING INTERCOMPARISON PROJECT BOUNDARY CONDITIONS AND PARAMETERS Time Time Boundary conditions and parameters Control experiment (6 ka B.P.) (21 ka B.P.) Sea-surface temperature and sea ice Modern Modern Calculated or prescribed by CLIMAP Continental ice sheets Modern M odern Prescribed (Peltier, 1994) Vegetation and land-surface characteristics Modern Modern Modern (besides areas covered by ice) Aerosol optical depth Modern M odern M odern –2 –2 –2 1365 Wm 1365 Wm Solar constant 1365 Wm Ecc = 0.018994 Ecc = 0.016724 Ecc = 0.018682 Orbital parameters ε = 23.446° ε = 24.105° ε = 22.949° λ = 102.04° λ = 0.87° λ = 114.42° 28 0 p p m 280 p pm 2 00 p pm CO2 Note: CLIMAP—Climate: Long-range Investigation, Mapping, and Prediction. Ecc—eccentricity; ε—obliquity; λ—longitude of the perihelion.

30

Kislov and Toropov

Simulation of River Runoff and Sea-Level Changes during the Mid-Holocene Warm Event The Holocene has been marked by a relatively stable, warm climate that experiences weak global changes on average. By the early Holocene (9 ka), CO2 was up to its preindustrial level (Raynaud et al., 1993), the Scandinavian ice sheet was almost gone, and the Laurentide ice sheet had shrunk considerably (Dyke and Prest, 1987; Svensson, 1991). Sea-surface temperature (SST) was not significantly different from today. In the mid-Holocene, at 6 ka, insolation anomalies (compared to modern ones) were +5% in summer and −5% in winter, less than those at the beginning of the Holocene. PMIP simulations of the 6 ka climate were thus designed as pure sensitivity experiments to gauge changes of insolation forcing (see Table 1). Global model results were evaluated (Joussaume, 1999; Kohfeld and Harrison, 2000) against terrestrial proxy data (lake levels, pollen- and plant macrofossil–based reconstructions of 6 ka vegetation). Our calculations indicate that at 6 ka, there was no significant change in the East European Plain river runoff (see Table 2). The Volga River’s contribution to Caspian Sea inflow was slightly increased (93%) compared to today (88% based on PMIP model simulation and 84% based on observation). The value of (Δh)e over the Caspian Sea was estimated—based on simplified regional climate modeling (Kislov and Sourkova, 1998)—as a small value relative to the first term in Equation 2. Therefore, there was no large change in sea-surface area or level (this assessment is not true for the Black Sea because, at this time, it was not closed, and changes in its level were controlled by water exchange with the Mediterranean Sea). Simulation of River Runoff and Sea-Level Changes during the Last Cold Pleistocene Event The last cold event of the Pleistocene involved substantial changes in surface boundary conditions, and within the PMIP, a specific set was prepared (see Table 1). At 21 ka, the latitudinal distribution of insolation and its relative seasonal strength were similar to those of today (for example, in the northern summer, the solar energy deficit received by Earth was −2–4 Wm2). The ice-sheet extent and height were provided by Peltier (1994). CO2 concentration was estimated at 200 ppm as inferred from Antarctic ice cores (Raynaud et al., 1993). Over the oceans, two sets of PMIP experiments were defined: (1) prescribing

TABLE 2. CHANGE IN RIVER RUNOFF (Y – Y0)/Y0 (%) BELONGING TO THE BLACK AND CASPIAN SEAS AT 6 ka B.P. BASED ON PMIP DATA Black Sea Caspian Sea Ensemble of PMIP models +9 +14 –5 +5 Successful PMIP models Note: PMIP—Paleoclimate Modeling Intercomparison Project.

SST changes from estimates given by CLIMAP (1981), and (2) SST computed using coupled atmosphere–mixed-layer ocean models. Other kinds of intrinsic variability (e.g., due to ocean circulation changes) or other kinds of natural forces (e.g., solar irradiance and volcanic forces) were not incorporated into these experiments. Note that the boundary conditions used by the PMIP contain some uncertainties. For example, there is an underestimation of CLIMAP SST anomalies in the tropics (Anderson and Webb, 1994; Guilderson et al., 1994; Hostetler and Mix, 1999) and a problem with location and seasonal behavior of sea-ice cover (Weinelt et al., 1996). Another example touches upon the location of Quaternary ice sheets. Model sensitivity to these scenarios has been investigated previously (Kislov et al., 2002). It was shown that most differences (2–6 °C) between the results occurred within regions where the position of the ice sheet has changed. Thus, the difference in boundary condition provides only a regional effect. Moreover, taking into account the intermodel deviations in simulation data, these uncertainties are not crucial for global climate modeling. We evaluated the global model results against terrestrial proxy data (Joussaume, 1999; Kohfeld and Harrison, 2000; Kislov et al., 2002). These results clearly demonstrate that temperatures in the PMIP experiment reproduce the main peculiarities of reconstructed land temperature fields, but over the tropics, the simulations with prescribed CLIMAP SSTs produce too weak a cooling effect over land. All models produce drying in the extratropical zone, although the extent and location of the regions experiencing increased aridity vary between models. Consider the results of calculation of annual river runoff volumes for the Caspian Sea and the Black Sea (Table 3). At 21 ka, the total river runoff to the Caspian Sea (calculated by “successful” models) was substantially lower (–50%) compared to today. The relative contribution of Volga River runoff has a value of 72%. These facts are in accordance with the observational data. Taking into account that the second term in Equation 3 is equal to zero, and the value Δe/e0 can be estimated as small relative to the first term in Equation 2, we can estimate the relative decrease in Caspian Sea area as 50%, which means a substantial drop in level (~50 m). Calculated river runoff into the Black Sea decreased substantially as well (–45%), and the drop in level has been estimated at –200 m. During this period, there was no water exchange through the modern Bosporus Strait due to dropping of both the Black Sea level and the world ocean level.

TABLE 3. CHANGE IN RIVER RUNOFF (Y – Y0)/Y0 (%) INTO THE BLACK AND CASPIAN SEAS AT 21 ka B.P. BASED ON PMIP DATA Black Sea Caspian Sea Ensemble of PMIP models –22 –40 Successful PMIP models –45 –56 Note: PMIP—Paleoclimate Modeling Intercomparison Project.

Modeling extreme Black Sea and Caspian Sea levels of the past 21,000 years Simulation of River Runoff and Sea-Level Changes during the Cold Younger Dryas and the Warm Allerød Events At 12–14 ka, insolation anomalies were +5% in summer and −5% in winter (compared to mid-Holocene levels). During the YD, SST over the North Atlantic Ocean was significantly less than today, but during the Allerød, there were no large differences. For the climate simulation, GCM T42L15 was used with prescribed SSTs. Therefore, experimentally, the Allerød was practically similar to that discussed for the mid-Holocene experiment but with more active radiation forcing. The YD experiment yielded both strong radiation forcing and an SST anomaly in the North Atlantic Ocean. During the Allerød, annual river runoff volume into the Caspian and Black Seas was slightly increased (6%), and during the YD, it was lower (–12%) compared to today. Therefore, it could not cause serious sea-level changes. DISCUSSION The results of climate simulation presented here are important in light of the problem of chronologically correlating paleogeographical events that belong to different regional scales. In response to glacial conditions of the last Pleistocene cold event, the declining levels of the Black Sea (post-Karangatian regression stage) and the Caspian Sea (Atelian regression stage) are simulated simultaneously. Hence, these changes in sea level reflect climate forces at the planetary scale. This lends credence to the idea of a connection between the deep regression states of the Caspian and Black Seas and the mature stages of the late Quaternary glacial/cooling/drying planetary events. The next question is whether all deep regression stages exhibited by these seas have had a similar origin. Is the conclusion assigned to one snapshot at 21 ka similarly applicable to other events? This idea is probably true when sea-level curves denoting time-behavior of the Black and Caspian Seas are compared to the curve depicting global climate changes (Fig. 1). Taking into account uncertainties in the reconstructed data, it is possible to conclude that at least two final regression stages occurred simultaneously with late Quaternary glacial planetary events. As far as the transgression stages are concerned, the simulation of their onset and duration remains a very difficult problem. The aforementioned modeling results have shown that during the warm periods (taking as an example the mid-Holocene and Allerød events), the simulated river runoff did not increase sufficiently to create a strong Caspian Sea transgression leading to overflow into the Black Sea through the Manych Strait. In short, there is no evidence to connect the large, well-documented Khvalynian transgression stage of the Caspian Sea with the warm Allerød event. Thus, there is no clear understanding about the source of “additional” water volume capable of establishing a high enough Caspian Sea level that would permit overflow from the Caspian into the Black Sea. There are several speculative hypotheses that

31

lie beyond the paradigm of climate modeling that attempt to solve this paradox. One idea is that Siberian rivers such as the Ob’, Yenisei, Lena, and others were bordered by ice sheets along the Arctic coast, and water from huge dammed lakes overflowed into the Caspian Sea through different spillways (Grosswald, 1998). A second hypothesis posits that the source of additional water may have been connected to an increase in precipitation over the Tian Shan Mountains due to penetration of the Indian monsoon during warm periods (Kislov and Toropov, 2007). Subsequently, meltwater outflow was directed from the Amu Daria toward the Uzboy Valley and the Caspian Sea throughout the intermediate Sarykamish Lake. However, there is no reliable evidence to support these hypotheses. The volume of river runoff would be effectively increased if the runoff coefficient (i.e., the ratio of runoff to precipitation) were to change. This coefficient increases, for example, if water from precipitation is not absorbed into the soil due to the presence of permafrost. There is some evidence that permafrost conditions existed within the East European Plain during the postglacial, and it is thought that the effect of such a scenario would have been to produce changes as large as 30% in runoff volume. ACKNOWLEDGMENTS Financial support for this work was provided by the Russian Fund for Basic Research. This research is a contribution to the PMIP (Paleoclimate Modeling Intercomparison Project) and IGCP (International Geological Correlation Programme).

Figure 1. Global climate change (marine isotope data, after Imbrie et al., 1984) and Black Sea and Caspian Sea level variations (after Shuisky [2007] and Svitoch [2003]).

32

Kislov and Toropov

REFERENCES CITED Anderson, D.M., and Webb, R.S., 1994, Ice-age tropics revisited: Nature, v. 367, p. 23–24, doi: 10.1038/367023a0. Benson, L.V., and Paillet, F.L., 1989, The use of total lake-surface area as an indicator of climate change: Examples from the Lahontan Basin: Quaternary Research, v. 32, p. 262–275, doi: 10.1016/0033-5894(89)90093-8. Berger, A., 1988, Milankovitch theory and climate: Reviews of Geophysics, v. 26, p. 624–657, doi: 10.1029/RG026i004p00624. Chepalyga, A., 2007, The Late Glacial Great Flood in the Ponto-Caspian basin, in Yanko-Hombach, V., Gilbert, A.S., Panin, N., and Dolukhanov, P.M., eds., The Black Sea Flood Question: Changes in Coastline, Climate, and Human Settlement: Dordrecht, The Netherlands, Springer, p. 119–148. CLIMAP, 1981, Seasonal Reconstructions of the Earth’s Surface at the Last Glacial Maximum: Geological Society of America Map Series MC-36. Dyke, A.S., and Prest, V.K., 1987, Late Wisconsinan and Holocene history of the Laurentide ice sheet: Géographie Physique et Quaternaire, v. 41, p. 237–263. Grosswald, M.G., 1998, New approach to the Ice Age paleohydrology of northern Eurasia, in Benito, G., Baker, V.R., and Gregory, K.J., eds., Paleohydrology and Environmental Change: Chichester and New York, John Wiley & Sons, p. 199–214. Guilderson, T.P., Fairbanks, R.G., and Rubenstone, J.L., 1994, Tropical temperature variations since 20,000 years ago: Modulating interhemispheric climate change: Science, v. 263, p. 663–665, doi: 10.1126/ science.263.5147.663. Harrison, S.P., Prentice, I.C., and Guiot, J., 1993, Climatic controls on Holocene lake-level changes in Europe: Climate Dynamics, v. 8, p. 189–200, doi: 10.1007/BF00207965. Harrison, S.P., Yu, G., and Tarasov, P.E., 1996, Late Quaternary lake-level record from northern Eurasia: Quaternary Research, v. 45, p. 138–159, doi: 10.1006/qres.1996.0016. Hostetler, S.W., and Mix, A.C., 1999, Reassessment of ice-age cooling of the tropical ocean and atmosphere: Nature, v. 399, p. 673–676, doi: 10.1038/21401. Imbrie, J., Hays, J.D., Martinson, D.G., McIntyre, A., Mix, A.C., Morley, J.J., Pisias, N.G., Prell, W.L., and Shackleton, N.J., 1984, The orbital theory of Pleistocene climate: Support from a revised chronology of the marine δ18O record, in Berger, A., Imbrie, J., Hays, H., Kukla, G., and Saltzman, B., eds., Milankovitch and Climate: Understanding the Response to Astronomical Forcing, Proceedings of the NATO Advanced Research Workshop (held 30 November–4 December 1982, in Palisades, New York): Dordrecht, Netherlands, D. Reidel Publishing, p. 269–305. Jones, R.N., McMahon, T.A., and Bowlers, J.M., 2001, Modelling historical lake levels and recent climate change at three closed lakes, Western Victoria, Australia (c. 1840–1990): Journal of Hydrology (Amsterdam), v. 246, p. 159–180, doi: 10.1016/S0022-1694(01)00369-9. Joussaume, S., 1999, Modeling extreme climates of the past 20,000 years with general circulation models, in Holland, W.R., Joussaume, S., and David,

F., eds., Modeling the Earth’s Climate and its Variability: Amsterdam, Elsevier, p. 527–565. Kalinin, G.P., 1968, Problemy Globalnoi Gidrologii (Problems of Global Hydrology): Leningrad, Gidrometeoizdat, 376 p. (in Russian). Kislov, A.V., and Sourkova, G.V., 1998, Simulation of the Caspian Sea level changes during last 20000 years, in Benito, G., Baker, V.R., and Gregory, K.J., eds., Palaeohydrology and Environmental Change: Chichester, John Wiley & Sons, p. 235–246. Kislov, A.V., and Toropov, P.A., 2007, Climate modeling results for the circumPontic region from the late Pleistocene to the mid-Holocene, in YankoHombach, V., Gilbert, A.S., Panin, N., and Dolukhanov, P.M., eds., The Black Sea Flood Question: Changes in Coastline, Climate, and Human Settlement: Dordrecht, The Netherlands, Springer, p. 47–62. Kislov, A.V., Tarasov, P.E., and Sourkova, G.V., 2002, Pollen and other proxybased reconstructions and PMIP simulations of the Last Glacial Maximum mean annual temperature: An attempt to harmonize the data-model comparison procedure: Acta Palaeontologica Sinica, v. 41, p. 539–545. Kohfeld, K.E., and Harrison, S.P., 2000, How well can we simulate past climates? Evaluating the models using global palaeoenvironmental datasets: Quaternary Science Reviews, v. 19, p. 321–346, doi: 10.1016/S0277 -3791(99)00068-2. Peltier, W.R., 1994, Ice age paleotopography: Science, v. 265, p. 195–201, doi: 10.1126/science.265.5169.195. Qin, B., Harrison, S.P., and Kutzbach, J.E., 1998, Evaluation of modelled regional water balance using lake status data: A comparison of 6 ka simulations with the NCAR CCM: Quaternary Science Reviews, v. 17, p. 535–548, doi: 10.1016/S0277-3791(98)00011-0. Raynaud, D., Jouzel, J., Barnola, J.-M., Chappellaz, J., Delmas, R.J., and Lorius, C., 1993, The ice record of greenhouse gases: Science, v. 259, p. 926–934. Shuisky, Y., 2007, Climate dynamics, sea-level change, and shoreline migration in the Ukrainian sector of the circum-Pontic region, in Yanko-Hombach, V., Gilbert, A.S., Panin, N., and Dolukhanov, P.M., eds., The Black Sea Flood Question: Changes in Coastline, Climate, and Human Settlement: Dordrecht, The Netherlands, Springer, p. 251–277. Street-Perrot, F.A., and Harrison, S.P., 1985, Lake levels and climate reconstruction, in Hecht, A.D., ed., Paleoclimatic Analysis and Modeling: New York, John Wiley and Sons, p. 291–340. Svensson, N.O., 1991, Postglacial land uplift patterns of south Sweden and the Baltic Sea region: Terra Nova, v. 3, p. 369–378, doi: 10.1111/j.1365 -3121.1991.tb00165.x. Svitoch, A., 2003, Morskoi Pleistotsen Poberezhii Rossii (Marine Pleistocene of the Russian Coasts): Moscow, GEOS, 362 p. Weinelt, M., Sarnthein, M., Pflaumann, U., Schulz, H., Jung, S.J.A., and Erlenkeuser, H., 1996, Ice-free Nordic seas during the Last Glacial Maximum? Potential sites of deepwater formation: Paleoclimates, v. 1, p. 283–309.

MANUSCRIPT ACCEPTED BY THE SOCIETY 22 JUNE 2010

Printed in the USA

The Geological Society of America Special Paper 473 2011

Assessment of Black Sea water-level fluctuations since the Last Glacial Maximum G. Lericolais Institut Français de Recherche pour l’Exploitation de la Mer (IFREMER), Centre de BREST, BP 70, F29200 Plouzané cedex, France F. Guichard Laboratoire des Sciences du Climat et de l’Environnement (LSCE), CNRS-CEA, Avenue de la Terrasse, BP 1, 91198 Gif-sur-Yvette cedex, France C. Morigi Geological Survey of Denmark and Greenland (GEUS), Department of Stratigraphy, Øster Voldgade 10, 1350 Copenhagen, Denmark I. Popescu Institutul National de Cercetare-Dezvoltare pentru Geologie si Geoecologie Marina (GeoEcoMar), 23-25 Dimitrie Onciul Str, BP 34-51, Bucuresti, Romania C. Bulois School of Geological Sciences, University College Dublin, Belfield, Dublin 4, Ireland H. Gillet Unité Mixte de Recherche (UMR) 5805, Environnements et Paléoenvironnements Océaniques (EPOC), Université Bordeaux 1, Avenue des Facultés, F33405 Talence, France W.B.F. Ryan Lamont-Doherty Earth Observatory, Columbia University, 61 Route 9w, Palisades, New York 10964, USA

ABSTRACT This paper presents geophysical and core data obtained from several marine geology surveys carried out in the western Black Sea. These data provide a solid record of water-level fluctuation during the Last Glacial Maximum in the Black Sea. A Last Glacial Maximum lowstand wedge evidenced at the shelf edge in Romania, Bulgaria, and Turkey represents the starting point of this record. Then, a first transgressive system is identified as the Danube prodelta built under ~40 m of water depth. The related rise in water level is interpreted to have been caused by an increase in water provided to the Black Sea by the melting of the ice after 18,000 yr B.P., drained by the largest European rivers (Danube, Dnieper, Dniester). Subsequently, the Black Sea Lericolais, G., Guichard, F., Morigi, C., Popescu, I., Bulois, C., Gillet, H., and Ryan, W.B.F., 2011, Assessment of Black Sea water-level fluctuations since the Last Glacial Maximum, in Buynevich, I.V., Yanko-Hombach, V., Gilbert, A.S., and Martin, R.E., eds., Geology and Geoarchaeology of the Black Sea Region: Beyond the Flood Hypothesis: Geological Society of America Special Paper 473, p. 33–50, doi: 10.1130/2011.2473(03). For permission to copy, contact editing@geosociety .org. © 2011 The Geological Society of America. All rights reserved.

33

34

Lericolais et al. lacustrine shelf deposits formed a significant basinward-prograding wedge system, interpreted as forced regression system tracts. On top of these prograding sequences, there is a set of sand dunes that delineates a wave-cut terrace-like feature around the isobath −100 m. The upper part of the last prograding sequence is incised by anastomosed channels that end in the Danube (Viteaz) canyon, which are also built on the lacustrine prograding wedge. Overlying this succession, there is a shelfwide unconformity visible in very high-resolution seismic-reflection profiles and present all over the shelf. A uniform drape of marine sediment above the unconformity is present all over the continental shelf with practically the same thickness over nearby elevations and depressions. This mud drape represents the last stage of the Black Sea water-level fluctuation and is set after the reconnection of this basin with the Mediterranean Sea.

INTRODUCTION The location of the Black Sea, between Europe and Asia, makes its water level dependent on Eurasian climatic fluctuations. This inland sea is a perfect present-day example of a marginal basin where connection changes dramatically with sea level (Ross, 1971, 1978; Ross et al., 1970; Ross and Degens, 1974; Ryan et al., 2003, 1997). The Black Sea is at present the world’s largest anoxic basin, making it an important modern analogue for past anoxic conditions, while during the last glacial period, it was a low-salinity oxygenated lake, isolated from the Mediterranean (Deuser, 1972, 1974; Lericolais et al., 2006b; Wall and Dale, 1974). The marine and lacustrine deposits at the Black Sea represent a valuable archive for the study of past climate changes. During the Quaternary glacial periods, a northern ice cap prevented major East European rivers flowing from north as they do today. During ensuing interglacial periods, these rivers were diverted to the south in the direction of the Black Sea and Caspian Sea receiving basins and consequently have increased the size of these drainage basins (Arkhipov et al., 1995). Therefore, unique conditions specific to the Black Sea were established while this water body became isolated from the global ocean. This isolation results in a sedimentary record without the hysteresis effect, which is the latent period needed by the global ocean to respond to the consequences of ice melting. During these isolation phases, the Black Sea was more sensitive to climate changes than the Caspian Sea is today. Arkhipov et al. (1995) and Chepalyga (1984) interpreted the Caspian Sea fluctuations opposed to those of the global ocean to have caused the possible connection between the Black Sea and the Caspian Sea through the Manych Strait (Fig. 1). When the Black Sea was isolated, both the lack of saltwater input and the increase of freshwater runoff from the rivers led to reduced salinity levels in the Black Sea. This process during the glacial periods, linked to water-level fluctuation, is measured in the fauna succession, which shows an abrupt change from saltwater to freshwater or brackish-water species. The initial hypothesis of a rapid saltwater flooding of the freshwater lake that was the Black Sea in the Late Glacial Maximum (LGM) was proposed in 1996 by Ryan et al. (1996, 1997). The flood hypothesis raised controversy and initiated refutation (Aksu et al., 2002a, 2002b, 1999; Görür

et al., 2001; Hiscott and Aksu, 2002; Hiscott et al., 2008, 2002; Yanko-Hombach et al., 2006), but recently also received support (Algan et al., 2007; Eriş et al., 2007, 2008; Gökaşan et al., 2005; Lericolais et al., 2007b, 2007c; Siddall et al., 2004). Nevertheless, most of each opposing view is supported by only a small amount of data in the Black Sea, and not all of the 420,000 km2 have been surveyed using modern scientific equipment and interpretation in light of modern ideas. Recently, the European Project ASSEMBLAGE (EVK3CT-2002-00090) provided geophysical and sedimentary data collected in the northwestern part of the Black Sea from the continental shelf and slope down to the deep-sea zone. This project focused on applying sequence stratigraphic models to seismic data recorded on the northwestern Black Sea shelf, in order to correlate the sequences interpreted using seismic stratigraphy methods to sea-level fluctuations. To achieve the project’s objectives, very high-resolution seismic data were acquired during the BlaSON cruises (1998 and 2002) using the research vessel Le Suroît and during the ASSEMBLAGE 1 (2004) cruise of the research vessel Le Marion Dufresne. During the first two cruises, paleoshorelines and sand ridges were identified, and a set of seismic data was acquired on these targets to support pseudo–three-dimensional (3-D) analyses. This, coupled with a multiproxy approach, emphasizes that the Black Sea water level is dependent on Eurasian climatic fluctuations. This sequence stratigraphy study was validated by dated samples obtained from long cores (up to 50 m long) providing a firm calibration of Black Sea water-level fluctuation since the LGM. These data show that the Black Sea experienced a contemporary rise in water level with the melting of the Fennoscandian ice sheet, followed by a drop of the water level from the Younger Dryas to the Preboreal. This recent lowstand is confirmed by the presence of the forced regression sequences, the wave-cut terrace, and the coastal dunes still preserved on the shelf, even after the Black Sea was rapidly invaded by Mediterranean/Marmara marine waters. PREVIOUS STUDIES Already in the seventies, Kuprin et al. (1974) and Shcherbakov et al. (1978) documented the lowstand shorelines of the Black

Assessment of Black Sea water-level fluctuations Sea. Numerous Soviet, Romanian, Bulgarian, and Turkish coring and echo-sounding surveys conducted in the western part of the Black Sea had previously identified a littoral zone near the shelf edge. Several cores cut by these studies penetrated an erosional surface. From the 1990s Romanian data, Popescu et al. (2004) identified the presence of ancient river valleys entrenching the shelf, especially in front of major canyons. Other workers confirmed the shoreline position with the recovery of sand, gravel, and freshwater molluscs typical of the coastal zone (Major et al., 2002b; Ryan et al., 2003). Ostrovskiy et al. (1977) published results on the stratigraphy and geochronology of Pleistocene marine terraces of the Black Sea, where extensive down-cutting of coastal river valleys was recognized as evidence of a major water-level drop of the ice-age Black Sea on the order of −110 m. A key limitation of this previous research is that no seismic-reflection profiles were published to document their findings, even though one can read that the former eastern country researchers have documented the exposed margin of the Black Sea lake, and that numerous piston and drill cores also confirmed the existence of an ancient coast. The first interna-

35

tional publication related to the work done by the Soviet block on this topic was issued from the U.S.-Russia-Turkey expedition of 1993 led by professor Shimkus (Shimkus et al., 1997) with the objective to examine the impact of the Chernobyl contamination in the Black Sea. The results obtained from this joint survey allowed the mapping of the Dnieper River valleys in more detail with reflection profiling methods and explored the coastal deltas on the Ukrainian shelf. Later reflection profiling gave evidence of the same shelfwide erosion surface at different Black Sea locations, i.e., on the Romanian shelf (Lericolais et al., 2007b; Popescu et al., 2004), on the Bulgarian shelf (Dimitrov and Peychev, 2005; Dimitrov, 1982; Khrischev and Georgiev, 1991), and on the Turkish margin (Aksu et al., 2002b; Algan et al., 2002, 2007; Demirbag et al., 1999; Okyar and Ediger, 1999; Okyar et al., 1994). The general assumption about the Black Sea before the Ryan et al. (1997) hypothesis was that the lake’s surface had risen correlatively with the global sea level. This required a relatively early connection through the Bosporus Strait. Based on hydrologic considerations, Chepalyga (1984) and Kvasov and

Figure 1. General location of the studied area.

36

Lericolais et al.

Blazhchishin (1978) stipulated that outflow from the Black Sea through this strait had always been continuous, even at maximum lowstand conditions. For this to be the case, the outlet level of the Black Sea during glacial times would have to have been in concordance with the Black Sea lowstand shorelines. In opposition to the evidence discovered about the depth of the Bosporus sill, Chepalyga (1995) used an old suggestion to place the former Black Sea–Marmara connection in the Sakarya River valley. This idea was consistent with the work of Elmas (2003) who demonstrated that the late Cenozoic tectonics and stratigraphy of northwestern Anatolia allowed a connection between the Sakarya outlet to the eastern arm of the Gulf of Izmit. If such theory is viable, the Black Sea sill controlling its lake level would therefore have been the Dardanelles bedrock sill that is at −85 m (Ryan, 2007). Recently, a number of researchers have rejected the hypothesis of a deep Black Sea outlet (Bahr et al., 2005, 2006; Major et al., 2002b; Myers et al., 2003). Previously, Lane-Serff et al. (1997) had proposed that a deep outlet would have permitted a vigorous outflow of semibrackish waters from the Black Sea strong enough to keep out Mediterranean saltwater from entering the enclosed basin. However, their hydraulic models only prevented Mediterranean inflow until sea level rose 5 m above the sill. It is now admitted that the increase in salinity in the Sea of Marmara occurred at least 12,000 yr ago, as determined from the mollusc assemblage and stable isotopes (Çagatay et al., 2000; Sperling et al., 2003) and could have been even earlier (Popescu, 2003, 2004). Authors who are against a late connection of the Black Sea to the Mediterranean sea are now publishing evidence of a late salinization of the Black Sea obtained from their studies conducted on cores recovered on the Black Sea Turkish shelf, e.g., Hiscott et al. (2007) explains that the Ostracoda of Caspian affinity indicate ~5‰ salinity until ca. 7500 yr B.P. Dinocysts and foraminifera confirm a low but rising salinity no later than ca. 8600 yr B.P., and a first major pulse of marine waters was recorded at around 8460 yr B.P. by Marret et al. (2009). Hence, they confirm the previous observations published by Ryan et al. (1997, 2003) and Major et al. (2002a, 2002b, 2006). These results propose that the first marine signal in the Black Sea is recorded between 9000–8000 yr B.P. At that time, the Mediterranean sea level was around −30 m (Lambeck and Bard, 2000; Lambeck et al., 2002) or even less, as the model predictions from the northern coast of Israel indicate sea level at about –13.5 ± 1 m for ca. 8000 yr B.P., whereas observation places it between –14.5 m and –16.5 m (Lambeck et al., 2007; Sivan, 2003). If the Black Sea outflow through a deep connection was truly so vigorous and persistent, it remains to be explained how this outflow could have permitted the early and sustained salinization of the Sea of Marmara at the downstream end of the water cascade. On the other hand, since the observation of post-LGM lowstand shorelines characterized by wave-cut terraces in different areas of the Black Sea, i.e., at –110 m off Ukraine (Ryan et al., 1997), −100 m on the Romanian shelf (Lericolais et al., 2003, 2007b, 2007c), −122 m for the Bulgarian shelf (Dimitrov, 1982),

and −155 m off Sinop where the shelf is really narrow (Ballard et al., 2000), it is necessary to consider a shallow outlet for the Bosporus with interrupted outflow. Even if the subsidence effect since the LGM is negligible (Wong et al., 2005), some consideration has to be taken for the tectonic effect, especially at the foot of major mountain chains such as the Carpathians, the Balkans, the Caucasus, or Anatolia. This effect explains why some 20 m of difference exists for the topset location of the LGM lowstand wedges in different Black Sea areas. Major et al. (2006), quoting their former work on strontium isotopes published in 2002 (Major et al., 2002b), established that the lake level would have been controlled mainly by the balance of evaporation versus inflow from rivers and rainfall, even though intervals of enclosure of the Black Sea may have been of relatively short duration. These authors also confirm that the Black Sea was an enclosed semibrackish lake during these periods. Lake-level fluctuations might also account for the observed repetition of “cut and fill” in the sediments of the river valleys that cross the shelf (Heller et al., 2001; Koss et al., 1994; Lericolais et al., 2001; Newell, 2001; Popescu et al., 2004; Ryan et al., 2003; Talling, 1998; Zaitlin et al., 1994), as well as the presence of wave-cut terraces on the edges of the shelf (Shimkus et al., 1980). The presence of authigenic aragonite layers correlative with the onset of the sapropel deposit (Giunta et al., 2007) can be correlated to a response to climatic change (Lamb, 2001) despite the hydrothermal influence and calcite precipitation (Peckmann, 2001). A detrital/biogenic source has also been interpreted by Reitz and de Lange (2006) as a possible mechanism for the major part of the aragonite enrichments found in sapropel sediments. Possibly, offshore-directed surface-water flows related to wind stress and/ or enhanced runoff (consistent with Mediterranean flooding and enhanced precipitation) during sapropel deposition may have assisted in the transport of near-coastal aragonitic organisms to more coast-remote areas. Recent studies carried out in the Black Sea confirm that authigenic calcite precipitation of calcareous mud appears following the deglacial meltwater delivery (Bahr et al., 2005, 2006; Major et al., 2002b; Ryan et al., 2003) and can be interpreted as a result of water evaporation (Giunta et al., 2007). SYNTHESIS OF RESULTS OBTAINED IN THE FRAME OF THE ASSEMBLAGE PROJECT Recently, an assessment of the northwestern part of the Black Sea sedimentary systems from the continental shelf and slope down to the deep-sea zone was provided by the ASSEMBLAGE European Project. Here, we summarize the results of this project, obtained from geophysical data and core analyses. These results provide a solid record of the Black Sea Last Glacial Maximum (LGM) water-level fluctuations and shed new light on the controversy concerning the Black Sea water-level fluctuation since the Last Glacial Maximum. The ASSEMBLAGE project attempted to assess the last sealevel rise in the Black Sea and provide scenarios quantifying the processes governing the transition of the Black Sea system from

Assessment of Black Sea water-level fluctuations a low-salinity lake to a marine state while addressing the variability in this system. Six major observations are used to reconstruct the Black Sea water-level fluctuations since the LGM. 1. The first observation is the existence of a LGM lowstand wedge at the shelf edge offshore Romania, Bulgaria, and Turkey. This observation is completed with the evidence of a second small lowstand wedge dated from 11,000 yr B.P. to 8000 yr B.P. from −100 to −120 m of water depth identified during ASSEMBLAGE cruises on the outer shelf of Romania and Bulgaria and described on the Turkish shelf by Algan et al. (2002). This wedge is associated with the recovery of strata immediately below an observed unconformity consisting of dense, low-water-content mud containing desiccation cracks, plant roots, and sand lenses rich in freshwater molluscs (Dreissena rostriformis) with both valves still together. 2. The second observation is deduced from results providing information on the construction of the Danube delta/prodelta, showing that a former prodelta was built up at −40 m after the post-LGM meltwater pulses. 3. The third observation comes from mapping of meandering river channels capped by a regional unconformity and extending seaward across the Romanian shelf to the vicinity of the –100 m isobath. 4. The fourth observation is the presence of submerged shorelines with wave-cut terraces and coastal dunes, or delta mouth bars at depths between –80 to –100 m, below the Holocene Bosporus and Dardanelles Strait outlet sill to the global ocean. 5. The fifth observation to be underlined is that, on the western part of the Black Sea continental shelf, a shelfwide ravinement surface is visible in very high-resolution seismic-reflection profiles. 6. The sixth observation useable for the understanding of the last water-level fluctuation of the Black Sea is the presence of a uniform drape of sediment beginning at the same time above the unconformity with practically the same thickness over nearby elevations and depressions and with no visible indication of coastal-directed onlap across the outer and middle shelf, except in the vicinity of the Danube Delta, where this mud drape is overlapped by recent Danube sediments.

37

digitally recorded. The navigation profiles presented here are displayed in Figure 2. The very high-resolution seismic sources were hull-mounted Chirp sonar systems with frequencies ranging between 1.5 and 7 kHz. Their vertical resolution is less than 1 m, with penetration reaching 500 ms in some deep areas where the sediment cover is constituted by soft sediment. On the shelf, the presence of gasbearing sediments masking the information beneath ~20 m depth below seafloor (bsf) decreases the amount of usable data. We also present results obtained from cores. The sediment cores were recovered using a Kullenberg piston corer; a conventional one used during BlaSON surveys, and Calypso type one developed for IPEV (French Institute for South Polar Seas) with a long tube (up to 60 m) system. Each of the core sections recovered were cut horizontally into two pieces and scanned to get an image before analyzing the samples. The general properties of the sediments were measured by the Multi System Track (MST) to get P-wave velocity and amplitude, density, impedance, and magnetic susceptibility values. All cores were packaged at 4 °C, and sampling was done at the IFREMER Brest laboratory. Dating was conducted on samples at various distances and various depths from the coast to reveal any possible bias in ages due to coastal or current influence. The Poznań Radiocarbon Laboratory in Poland performed 14C dating.

Methods of Data Acquisition The data were acquired during surveys realized in the framework of two main projects; (1) BlaSON: a French-Romanian bilateral project for which two surveys coordinated by IFREMER (Institut Francais de Recherche pour l’Exploitation de la Mer) were carried out on board the French RV Le Suroît in 1998 and 2002, and (2) ASSEMBLAGE: an FP5 European project for which two surveys were carried out on board the French RV Le Marion Dufresne in 2004 and the Romanian RV Mare Nigrum in 2005. For all these surveys, a differential global positioning system (GPS) was deployed for accurate (~1 m) positioning, and every vessel was equipped with swath bathymetry systems. Very high-resolution seismic lines were shot simultaneously using a Chirp sonar system. All data acquisition was synchronized and

Figure 2. Bathymetry of the semi-enclosed Black Sea basin and BlaSON and ASSEMBLAGE survey route locations.

38

Lericolais et al.

Before the evidence of wave-cut terraces in the Black Sea, no reliable sea-level markers were described to allow a good sea-level reconstruction (Giosan et al., 2006; Pirazzoli, 1991). Moreover, the lack of radiocarbon ages on in situ materials and the difficulty in calibrating radiocarbon ages in a setting with variable reservoir ages (Giosan, 2007) led to ongoing discussion about the Black Sea water-level fluctuations. At least for the more recent period, Siani et al. (2000) have proposed a reservoir age of 415 ± 90 yr B.P. for the Black Sea, based on six samples from the Black Sea, the Sea of Marmara, and in the vicinity of the Bosporus. One of the reasons for this is that a reservoir age of ~1280 yr was deduced from the occurrence of the Santorini Minoan ash in the Unit II of Jones and Gagnon (1994) from several south Black Sea cores recovered between 400 and 700 m below present sea level (Guichard et al., 1993). If we include documentation from Jones and Gagnon (1994), Bahr et al. (2005), and Kwiecien et al. (2008), then reservoir ages extending from 0 to 1280 yr. For the lacustrine period, no measurement has been proposed for the past. If no terrestrial influence exists, then age and residence time of deep waters will be the main factors. Östlund (1974) calculated ages of deep waters between 1470 yr (at ~600 m) and more than 2000 yr (when deeper than 1400 m). If a low stratification occurs in poorly salted water, then the residence time would be equal or less than 935 yr. The question of which reservoir age should be given to old water of the “Black Lake,” depending on depth in the water column, is still a matter of debate, and it is the reason why we still use uncalibrated and uncorrected 14C ages throughout this study for our obtained dates. The seismic profile and core locations are displayed on Figure 3. Core location, length, and water depth where samples were recovered are presented in Table 1. First Observation: The Lowstand Wedges (LGM and Preboreal) The seismic line Chirp B2CH96 (AB on Fig. 3) was shot during the BlaSON2 survey off Romania in front of the Danube delta (Fig. 4). At around 150 m water depth, this line displays prograding parallel but undulating reflectors characterizing a seismic unit LSW1 (Fig. 5). These reflectors toplap at the top of LSW1. Above the erosional truncation, unit LSW2 is located on the slope part of this dip line. This unit presents reflectors beveling the LSW1 slope slightly to the northwest. Throughout the seismic line, there is a thin unit that corresponds to the mud drape known to be present all over the western Black Sea (Lericolais et al., 2007b; Major et al., 2002b; Ryan et al., 1997, 2003). Age control of these two lowstand wedges is given by the dates obtained from core MD04-2771 and presented in Table 2. Dating of the seismic units interpreted as the LSW1 and LSW2 was possible, and older dates reach back to 29,450 ± 320 14C yr B.P. for unit LSW1 at 11.90 m on core MD04-2771. This date was obtained on organic matter, but because a Dreissena shell sampled at 2.18 m

in the same core returned a date of 24,980 ± 160 14C yr B.P., we can be confident in attributing the deposition of LSW1 to the Last Glacial period. A second lowstand is evidenced by seismic sequence LSW2. This lowstand wedge has a shape characteristic of a low-energy wedge. Core MD04-2771 confirms that this lowstand wedge started to be deposited around 12,180 ± 60 14C yr B.P. On the southwestern part of the Black Sea, another Chirp profile (B2CH56) displays more precisely the two successive lowstand wedges LSW1 and LSW2 (Fig. 5). Age control of these lowstand deposits was obtained from core MD04-2752 (Table 3; Fig. 6) dating Dreissena shells. Here again, the LSW1 is correlative to LGM time. It is very clear that LSW2 is a lowstand wedge deposited between 12,010 ± 50 14C yr B.P. and 8130 ± 50 14C yr B.P., showing that the Black Sea encountered a second lowstand after the LGM lowstand. Second Observation: The Post-LGM (Ante–Younger Dryas) Danube Prodelta The Danube prodelta is located at the coastal part of the Danube delta and can be seen on line B2CH96, section AB (Fig. 3). On the Chirp seismic profile, an erosion surface interpreted as a ravinement surface R1 is identified (Fig. 7). Above this, a prograding wedge U.S.2 is well delimited on the Chirp seismic data; this wedge corresponds to a former prodelta lobe. Another prodelta lobe corresponding to seismic unit U.S.3 presents reflectors onlapping on the previous unit U.S.2. The geometric relationship between these prodelta lobes would have been best imaged on shore-parallel profiles showing in detail the lap-out patterns of seismic reflectors, such as those shown for the Po River prodelta by Correggiari et al. (2005a, 2005b). Units U.S.2 and U.S.3 (Fig. 7) represent the sites of deposition and progradation at the distal part of the previous Danube River outlets and channels. These prodelta lobes represent part of the main depocenters that extend offshore, being a considerable portion of the prodelta deposit (Correggiari et al., 2005a, 2005b). As seen on Figure 4, these units are the highest part of the dip seismic line and are restricted to the prodelta area. Our data set is not dense enough to be able to decipher prodelta autocyclic processes from external forcing. Nevertheless, their position and nature are in accordance with our interpretation obtained from the comparison of the seismic data interpreted all along the profile and the core results. Above these prodelta lobes, seismic unit U.S.4 (Fig. 7) is prograding also, but in a more gentle shape. This unit can also be interpreted as a prodelta lobe, but ages obtained on core MD042774 return an average age of ca. 9500 14C yr B.P. (see Table 4). U.S.4 is contemporaneous to the onset of the Preboreal regression responsible of the deposit of LSW2 at the shelf edge as presented in the previous paragraph. Recent studies (Giosan et al., 2009) confirm that the Danube was building a ramp delta lobe at 8860 ± 45 14C yr B.P. (ages obtained from Dreissena polymorpha). From the morphology of

Assessment of Black Sea water-level fluctuations

39

Figure 3. Western Black Sea shelf presenting the location map of the seismic line B2CH96 (A–B) and B2CH56 (C– D) and of the cores MD04-2752, MD042771, MD04-2772, MD04-2773, and MD04-2774, along with geomorphologic interpretation issued from previous work (Lericolais et al., 2007c; Popescu et al., 2001, 2004). DA—dune area, DD—Danube delta, PDR1 and PDR2—paleo–Danube River 1 and 2, PCL—paleocoastline, VC—Viteaz Canyon, DSF—Danube deep-sea fan, BSF—Bosporus shallow fan delta.

TABLE 1. CORES PRESENTED IN THIS STUDY WITH THEIR POSITION, LENGTH OF RECOVERY, AND WATER DEPTH AT LOCATION Latitude (°N) Longitude (°E) Core Core length Water depth (m) (m) MD04-2752 41°56.76 28°36.56 24.50 169 MD04-2771 44°16.32 30°54.24 12.38 168 MD04-2772 44°18.07 30°51.56 7.51 106 MD04-2773 44°37.96 30°20.61 3.63 68 MD04-2774 44°57.47 29°50.12 7.30 30

40

Lericolais et al.

Figure 4. Line B2CH96 with core location (vertical exaggeration is ~200). DPD—Danube prodelta, PCL—paleocoastline, LSW—lowstand wedge, f—faults; twt—two-way traveltime.

Figure 5. Chirp profile B2CH96: Distal part of Figure 4. Lowstand wedge 1 (LSW1) and LSW2 can be distinguished. twt—two-way traveltime.

Assessment of Black Sea water-level fluctuations

41

TABLE 2. DATES OBTAINED FOR THE LOWSTAND WEDGES FROM CORE MD04-2771 AND TYPE OF SAMPLES DATED (MOLLUSC OR ORGANIC MATTER) Calibrated age* Water depth Core length Depth in core Age Sample Unit Core 14 (yr) (m) (m) (m) ( C yr B.P.) 0.34 12,180 ± 60 13,650 ± 120 Dreissena LSW2 MD04-2771 168 12.38 2.18 24,980 ± 160 28,840 ± 180 Dreissena LSW1 – 11.90 29,450 ± 320 Organic matter LSW1 *Calibrated ages are here as indicator and were obtained using the Radiocarbon Calibration Program Calib5 (Stuiver et al., 1998) with 400 yr for reservoir correction.

TABLE 3. DATES OBTAINED FOR THE LOWSTAND WEDGES FROM CORE MD04-2752 AND TYPE OF SAMPLES DATED (MOLLUSC OR ORGANIC MATTER) Calibrated age* Water depth Core length Depth in core Age Sample Unit Core 14 (yr) (m) (m) (m) ( C yr B.P.) 12.20 8130 ± 50 8605 ± 110 Organic matter LSW2 MD04-2752 169 24.50 12.30 12,010 ± 50 13,430 ± 100 Dreissena LSW2 19.85 25,020 ± 180 28,730 ± 300 Dreissena LSW1 *Calibrated ages are here as indicator and were obtained using the Radiocarbon Calibration Program Calib5 (Stuiver et al., 1998) with 400 yr for reservoir correction and Cariacco data for age >24,000 yr.

the lacustrine-marine contact, Giosan et al. (2009) supposed that the Black Sea lake level at that time was around 30 mbsl. Third Observation: Meandering River Channels Preserved on the Black Sea Shelf The third observation is deduced from previous Romanian surveys carried out by the GeoEcoMar Institute, where several

recent paleoriver channels incising the continental shelf down to −90 m water depth (Popescu et al., 2004) were identified (PDR1 and PDR2 on Fig. 3). These paleochannels are completely filled by sediments and are no longer visible in the bathymetry. These erosive features reach 400–1500 m in width and 20–30 m in depth. They present conventional asymmetry on some cross sections (Fig. 8) and seem to have been beveled by a subsequent phase of erosion. Here also these paleochannels are sealed by the

Figure 6. Chirp profile B2CH56: Distal part of segment line C–D on Figure 3 with core MD04-2752 location. Lowstand wedge 1 (LSW1) and LSW2 can be distinguished. Dated core samples are a = 8130 ± 50 14C yr B.P., b = 12,010 ± 50 14C yr B.P., c = 25,020 ± 180 14C yr B.P. (cf. Table 3). twt—two-way traveltime.

42

Lericolais et al.

Figure 7. Prodelta section of line B2CH96 with location of core MD04-2774: R1 is a ravinement surface; U.S.2 is a prograding wedge interpreted as a former prodelta lobe; U.S.3 is a retrograding sequence on U.S.2 and can be a second prodelta lobe; U.S.4 is a prograding sequence to be correlated to the forced regression described at −100 m by Lericolais et al. (2006a); and U.S.8 is the present-day prodelta deposit.

mud drape described earlier (Lericolais et al., 2007a; Major et al., 2002b; Popescu et al., 2004; Ryan et al., 2003). For Popescu et al. (2004), the stratigraphic position of these incisions lying directly under the discontinuity at the base of the Holocene strongly suggested that they formed during the last lowstand. The cartography of these buried channels shows that they are concentrated around two main directions. This distribution leads to their interpretation as anastomosed fluvial systems corresponding to two distinct drainage systems (Fig. 3). These would correspond to former paleo–Danube River flooding on the shelf to the outer shelf, where they apparently split into several arms, similar to a fluvial deltaic structure comparable in size to the modern Danube

delta, that lie close to the Danube Canyon (Popescu et al., 2004), also named Viteaz Canyon (VC on Fig. 3). The channels extend right to the paleoshoreline and pass under the belt of coastal sand ridges and depressions. Consequently, the regression that exposed the shelf surface into which the river channels were cut was followed by a transgression that led to the filling of the channels and then to another regression that deflated the channel fills and reexposed the entire region to coastal dune and pan development. The argumentation about the origin of the coastal features at ~−100 m has been presented in Lericolais et al. (2007b). Core MD04-2773 was recovered at one incised valley section (Fig. 8). The core (Fig. 9) got through the marine drape and

TABLE 4. DATES OBTAINED ON THE CORE MD04-2774 AND TYPE OF SAMPLES DATED (MOLLUSC OR ORGANIC MATTER) Calibrated age* Sample Unit Core Water depth Core length Depth in core Age 14 (yr) (m) (m) (m) ( C yr B.P.) 5.43 9030 ± 50 9720 ± 140 Pisidium US4 MD04-2774 30 7.3 6.91 9570 ± 50 10,440 ± 90 Pisidium US4 *Calibrated ages are here as indicator and were obtained using the Radiocarbon Calibration Program Calib5 (Stuiver et al., 1998) with 400 yr for reservoir correction.

Assessment of Black Sea water-level fluctuations

43

Figure 8. Incised valley section of line B2CH96 with location of core MD04-2773. TWT—two-way traveltime.

passed the Dreissena hash layer (Major et al., 2002b) described later herein. Even if the mollusc Pisidium sampled at 103 cm in the core (Table 5) is at the limit of the hash layer, it shows that this mollusc is of the same genus as the one collected in core MD042774. It is a genus of very small or minute freshwater clams known as pea clams, aquatic bivalve molluscs in the family Sphaeriidae. This confirms that the infilling of the valleys was still active at more than 60 m before the onset of the marine drape. Fourth Observation: Presence of Submerged Shorelines The submerged shorelines characterized by the presence of a wave-cut terrace at depths between –80 and –100 m are the key elements of the fourth observation. At the top of this coastal feature recognized on the Romanian shelf, there is a set of coastal dunes or delta mouth bars described by Lericolais et al. (2007a, 2007b). Analysis of the very high-resolution seismic data in pseudo–3-D mode (Lericolais et al., 2009) demonstrates that the lacustrine shelf deposits form an important basinward-prograding wedge system interpreted as a forced regression system tract eroded at the distal part by a wave-cut terrace (see figs. 4 and 5 in Lericolais et al., 2009). On line B2CH96, located north of the dune

field studied area, the wave-cut terrace is also visible (Fig. 10). On top of the prograding units (FR on Fig. 10), there is a set of sand dunes that delineates a berm-like feature around the −100 m isobath (WCT on Fig. 10), similar to the ones described by Ryan et al. (2003), Popescu et al. (2004), and Lericolais et al. (2007b). Analyses of cores retrieved from the dune field area demonstrate that the prograding wedges are lacustrine in origin and document a low water level characterized by forced regression– like reflectors mapped from the pseudo–3-D seismic data (Lericolais et al., 2009). Here, too, the hinge point corresponds to the wave-erosion surface mapped around the −100 m isobath. The ages returned by the core analysis range between 11,000 and 8000 14C yr B.P., with the formation of dunes being around 8500 14C yr B.P. The prograding reflectors deepen seaward and are truncated by an erosional surface described as the wave-cut terrace. On the Chirp profile, it is clearly seen that all the area is covered by a drape of less of 1 m thick (see “Sixth Observation: Uniform Drape above the Unconformity”), confirming that the dune system is not active any more. Everywhere across the midand outer shelf, the ridges, mounds, and depressions are draped by this thin layer of sediment with a remarkably uniform thickness of no more than a meter (Fig. 11).

44

Lericolais et al. where the present-day washed wave zone is marked by debris of whitened mollusc shells. Such facts are in favor of a rapid transgression in the Black Sea and are in agreement with the previous works published by Khrischev and Georgiev (1991) and Lericolais et al. (2004, 2007b). Actually, Khrischev and Georgiev (1991) attributed “fast rising” water level to the transition from lacustrine to marine conditions. For them, this change corresponds to a stratigraphic break (“washout”) in the cores that interrupts the lacustrine calcite precipitation and is followed by terrigenous mud with marine molluscs. They reported this “washout” in more than 100 cores. This same transition was described for the BlaSON and ASSEMBLAGE cores, where the transition was interpreted as either a ravinement surface or an erosion surface (Lericolais et al., 2007b, 2007c; Major et al., 2002b). Algan et al. (2007, p. 621) also made the observation of dense, dry mud below the erosional unconformity on the Thrace margin in cores from the shelf edge. These authors noted “a marked contact” between a 2-cm-thick shell-enriched layer and a “stiff clay deposit with low water content at the base of these cores.” The 14C age of the shells (Dreissena sp.) is 8590 ± 145 yr B.P., comparable to the age of the shell material constituting the “hash layer.” Algan et al. (2007, p. 623) considered that “the lithological characteristic of this core indicates that the deposition starts with high-energy condition over the stiff eroded substrate at about −100m, and continued with low-energy, suggesting a rapid deepening of a shallow environment.” Sixth Observation: Uniform Drape above the Unconformity

Figure 9. Photo of the core MD04-2773 sections. I—Modiulus ecozone; II—Mytillus ecozone, III—Dreissena ecozone. I and II are marine indicators, while III is from semibrackish state (Giunta et al., 2007). a—Pisidium at 103 cm in the core aged 7890 ± 50 14C yr B.P.

Fifth Observation: Ravinement Surface On the western part of the Black Sea continental shelf, a shelfwide ravinement surface is always present and can be recognized both on very high-resolution seismic-reflection profiles and in all the collected cores. In the cores, this surface corresponds to the described “hash layer” of Major et al. (Major et al., 2002b). This “hash layer” is composed of debris of whitened Dreissena. This corresponds to the surf zone as it is shown on Figure 12

Along the Black Sea margin, Wong et al. (2005), Algan et al. (2002), Ryan (2007), Ryan et al. (2003), Major et al. (2002b), and Lericolais et al. (2007b, 2007c) already described the presence of a uniform mud drape deposited above the unconformity. This mud drape layer was sampled during BlaSON and ASSEMBLAGE and corresponds in cores to the layer of terrigenous mud containing marine molluscs such as Mytilus galloprovincialis and Mytilus edulis, Cerastoderma edule, and Cardium edule (Giunta et al., 2007). This lithologic and biostratigraphic interval on the shelf corresponds to units 1 and 2 in basin sediments as defined by Ross et al. (1970). This uniform mud drape is clearly seen on the highresolution seismic profiles obtained by the Chirp sonar system and is displayed for instance on Figures 8 and 11. Its thickness, when calculated from acoustic travel time to meters, corresponds

TABLE 5. DATES OBTAINED ON CORE MD04-2773 AND TYPE OF SAMPLES DATED (MOLLUSC OR ORGANIC MATTER) Calibrated age* Sample Unit Water depth Core length Depth in core Age 14 (yr) (m) (m) (m) ( C yr B.P.) MD04-2773 68 3.63 1.03 7890 ± 50 8350 ± 60 Pisidium Incised valley *Calibrated age are here as indicator and were obtained using the Radiocarbon Calibration Program Calib5 (Stuiver et al., 1998) with 400 yr for reservoir correction. Core

Assessment of Black Sea water-level fluctuations

45

Figure 10. Forced regression (FR) progradation limited basinward by the wave-cut terrace (WCT) visible on section of line B2CH96 with location of core MD04-2773 (note the prograding reflectors inside the FR wedge). TWT—two-way traveltime.

in cores to the layer of terrigenous mud containing marine molluscs. Similar to the Romanian continental shelf, this layer has also been found above the unconformity on other Black Sea margins (Algan et al., 2002, 2007). Initial deposition of this uniform drape of sediment started at the same time above the unconformity and has practically the same thickness over nearby elevations and depressions, and it presents no visible indication of coastal-directed onlap across the outer and middle shelf. Such a layer deposited over the “hash

layer” ravinement surface, composed of in situ mussel molluscs at the bottom of this infra-meter layer, is characteristic of a rapid change. The size and disposition of the Mytilus edulis found in the cores are in accordance with the natural biotope of such a species. While the highest biomass is in general recorded at water depths ranging between 5 and 30 m, being lower at deeper depths, and living in niche beyond 40 m (Stea et al., 1994; Westerbom et al., 2002), this is not the case in our cores, where they are abundant everywhere. Such an observation is also an argument in favor of

Figure 11. Close-up of Figure 10 Chirp profile B2CH96 showing the seismic signal of the mud drape. TWT— two-way traveltime.

46

Lericolais et al.

a rapid sea-level rise, making it hard for these mussels to survive the transgression. SYNTHESIS: WATER-LEVEL FLUCTUATION OF THE BLACK SEA SINCE THE LAST GLACIAL EXTREME The synthesis presented here corresponds to an essay on the assessment of the last sea-level rise in the Black Sea and supports a scenario quantifying the processes governing the transition of Black Sea system from a semi-freshwater lake to a marine state. This work addresses the important postglacial variability in the Black Sea system, as the transition of this system from a semifreshwater lake to a marine environment was perhaps one of the most dramatic late Quaternary environmental events in the world. Back to the Last Glacial Maximum, 21,000 yr ago, the Black Sea was probably a giant freshwater to semi–brackish-water lake, as proposed by Arkhangelskiy and Strakhov (1938), or at least a brackish enclosed basin. Its water level stood more than 120 m below than today’s level. During ASSEMBLAGE project, analysis of high-resolution seismic-reflection profiles, Chirp and side-scan data together with piston core analyses from surveys taken on the Danube fan, and on the Black Sea shelf, provided new insights into the recent sedimentation processes in the deep northwestern Black Sea. The deep-sea fan studies (Popescu et al., 2001) demonstrate that the last channel-levee system on the Danube fan developed during the Neoeuxinian lowstand (stage 2) in a semi-freshwater basin with a water level ~120 m lower than today. Sediments supplied by the Danube were transported to the deep basin through the Viteaz canyon (Popescu et al., 2004). Functioning of the deep-sea fan is a good indicator of lowstand periods (Popescu et al., 2001; Winguth et al., 2000; Wong et al., 1997).

Figure 12. The present-day wave action zone showing debris of coquinas at the berm. This hash layer is the modern equivalent of the one described in the Black Sea core and having an average age of 8600 14C yr B.P.

Because the Black Sea was in a very close vicinity to the Scandinavian-Russian ice cap, the supply of the melting water from the glaciers into the Black Sea through the major drainage system constituted by big European rivers (Danube, Dnieper, Dniester, and Bug) was recorded by a brownish layers described in cores (Bahr et al., 2005; Major et al., 2002b). The water volume brought to the Black Sea after the meltwater pulse 1A (MWP1A) at ca. 12,500 14C yr B.P. (14,500 yr cal. B.P.) (Bard et al., 1990) was sufficient to raise the water level between −40 m and −20 m, where the Dreissena layers were deposited. The −40 m upper limit is interpreted from our records and especially deduced from the construction of the Danube prodelta (Lericolais et al., 2009), which are not exhaustive, and the −20 m limit is certified by Yanko (1990). This last value for the transgression upper limit would have brought the level of the Black Sea even higher to the Bosporus sill, and possible inflow of marine species like Mediterranean dynoflagellate populations can be envisaged (Popescu, 2004). Nevertheless, the rise in the Black Sea water level, which stayed between freshwater to brackish conditions, stopped the deep-sea fan sedimentation. Palynological studies conducted on BlaSON cores (Popescu, 2004) show that from the Bølling-Allerød to the Younger Dryas, a cool and drier climate prevailed. Northeastern rivers converged to the North Sea and to the Baltic Ice Lake (Jensen et al., 1999), providing reduced river input to the Black Sea and resulting in a receding shoreline. These observations are consistent with some evaporative drawdown of the Black Sea and are correlated to the evidence of an authigenic aragonite layer present in all the cores studied (Giunta et al., 2007; Strechie et al., 2002). This drawdown is also confirmed by the determination of the forced regression– like reflectors recognized either on the dune field mosaics (Lericolais et al., 2009) or on the B2CH96 transect profile and dated to this period. This lowered sea level in the Black Sea persisted afterward. The post–Bølling-Allerød climatic event favored the lowering of the Black Sea water level, and the presence of the coastal sand dunes and wave-cut terraces confirms this lowstand. This had already been observed by several Russian authors who considered a sea-level lowstand at about −90 m depth. Their observations were based on the location of offshore sand ridges described at the shelf edge south of Crimea. The anastomosed buried fluvial channels described by Popescu et al. (2004) that suddenly disappear below −90 m depth and a unique wave-cut terrace on the outer shelf, with an upper surface varying between −95 and −100 m, are therefore consistent with a major lowstand level situated somewhere around −100 m depth. Around the Viteaz Canyon, the paleocoastline was forming a wide gulf into which two rivers flowed (Fig. 3). Previous studies have already proposed a depth of −105 m for this lowstand, according to a regional erosional truncation recognized on the southern coast of the Black Sea (Demirbag et al., 1999; Görür et al., 2001), but also based on a terrace on the northern shelf edge (Major et al., 2002b). On the Romanian shelf, preservation of the sand dunes and buried small incised valleys are to be linked with a rapid transgression where the ravinement processes related to the

Assessment of Black Sea water-level fluctuations water-level rise had no time to erode sufficiently the sea bottom (Benan and Kocurek, 2000; Lericolais et al., 2004). Circa 7500 14C yr B.P., the surface waters of the Black Sea suddenly attained present-day conditions owing to an abrupt flooding of the Black Sea by Mediterranean waters, as shown by dinoflagellate cyst records (Popescu, 2004). This can also be related with the beginning of widespread and synchronous sapropel deposition across slope and basin floor. At 7160 14C yr B.P., Popescu (2004) demonstrated a sudden (100‰ and a history of strong pollution over the past 1000 yr (Aladin and Potts, 1992), and in Lake Issyk-Kul, a slightly brackish (6‰), oligotrophic lake in Kyrgyzstan (Giralt et al., 2004). Modern prasinophytes are regarded as primarily marine, but they also live in brackish and freshwater, and their fossils are associated with lagoonal or deltaic environs. They are most abundant in cold waters, with high nutrients being more important than temperature or salinity, according to Batten (1996). The data for the Black Sea corridor show that Pterosperma (Cymatiosphaera) dominates only

Figure 5. Life cycles of algal nonpollen palynomorphs: (A) Pterosperma (from Martin, 1993); and (B) Pediastrum boryanum (from Batten, 1996). Stages: 1—coenobium; 2–5—asexual reproduction stages: 2, 3—release of vesicle with zooids, 4—new coenobium forming within vesicle, 5—release of coenobium, 5f–5k—sexual reproduction stages.

104

Mudie et al.

in the lower-salinity eutrophic waters, and it is absent in the oligotrophic marine–hypersaline Aegean waters. In contrast, Pterospermella (Plate 1, fig. 7) is only reported for the hypersaline Red Sea, and the small spiny acritarch Micrhystridium (Plate 2, figs. 3, 4, and 8) was found only in the Red Sea, the hypersaline Crimean liman-lagoons, and, as Micrhystridium braunii–type, in the Aral Sea. The acritarch Polyasterias problematica is found in the Caspian and Aral Seas (as Hexasterias problematica; Sorrel et al., 2006); it is usually considered as a brackish to marine organism, but appears to be euryhaline (Matthiessen et al., 2000; Sorrel et al., 2006). The acritarch Radiosperma corbiferum was present mostly in the brackish intervals of the Aral Sea (rare and possibly reworked in the hypersaline intervals) and in the modern sediments of the freshwater Lake Sapanca east of the Marmara Sea (Leroy and Albay, 2010). The biological affinity of this nonpollen palynomorphs is unknown; a morphologically similar organism in the plankton of the Gulf of St. Lawrence, Canada (Bérard-Therriault et al., 1999, their plate 149b), is listed as Protiste sp. 2, in the group Protiste incertains (Uncertain Protists). Brenner (2001) reported that an apparently similar morphotype Radiosperma cf. corbiferum has a widespread distribution ranging from low-salinity Arctic estuaries to upwelling marine waters off Peru. Colonial Algae The colonial algae are unicellular Chlorophyceans that form colonies (coenobia) from a number of cells linked together and arranged in a specific pattern. In Pediastrum (Plate 2, figs. 17 and 18), the colonies have a fixed number of cells, the coenobia are flattened, and reproduction is by means of biflagellated cells, either aggregated within a vesicle that is liberated from the parent colony cell, or during sexual reproduction, and they are briefly free-swimming (Fig. 5B). In Botryococcus, the colonies range from small (~30 µm) subspherical clumps of cells (Plate 2, fig. 19), embedded in an oil-rich matrix, to relatively large (~140 µm) grape- or mulberry-like (botryoidal) groupings held together by gelatinous fibers or membranes (Batten and Grenfell, 1996). In the Black Sea corridor, the distribution of the colonial algae Pediastrum and Botryococcus shows the same trend as for Pterosperma (Cymatiosphaera). These nonpollen palynomorphs are abundant in parts of the Black Sea, sparse in the Marmara Sea, and almost absent in the Aegean and Levantine Seas, but they reappear in the Nile Delta and on the Egyptian shelf. These colonial algae are commonly considered to be indicators of riverwater inflow (e.g., Head, 1992, and references therein). However, they also appear in sediments of the Aral Sea and the hypersaline Kara-Bogaz Gol of the Caspian Sea, and Botryococcus braunii ha been found in plankton surveys of the Aral Sea (Piontkovski and Elmuratov, 2008). Detailed studies of surface sediment samples in the Baltic Sea show that Pediastrum boryanum and P. kawraisky are dominant in salinities from 6‰ to 8‰ (full range is 5‰–9‰), while P. simplex and P. duplex occur in salinities of less than 3‰–5‰ (Matthiessen and Brenner, 1996). The Baltic

study also shows that Botryococcus cf. braunii tolerates salinities up to 8‰. Zalessky (1926) reported that large botryoidal colonies of B. braunii occurred in the freshwater Russian Lake Beloë, while those in the brackish water (4‰) of Lake Balkash (Kyrgyzstan) were closely packed and globular. Growth of Botryococcus may be favored by seasonally cold, oligohaline conditions (Batten and Grenfell, 1996), in either eutrophic shallow or deep oligotrophic lakes (Chmura et al., 2006); in contrast, blooms of Pediastrum in the Canadian Great Lakes are triggered by excess phosphate loading (Nicholls, 1997). Zygnematales The Zygnematales are charophycean green algae that reproduce by conjugation to produce resting spores or zygotes with a sporopollenin-like cell wall. This order includes unbranched filamentous green algae of the families Zygnemataceae and singlecelled Desmidiaceae (sometimes classified in a separate order Desmidiales). Modern Zygnemataceae typically live in shallow, stagnant freshwater lakes, ponds, or in wet soil (Van Geel, 2001) and produce spores in the spring when conditions are warm. In late Holocene sediments of the Black Sea corridor, zygospores known to be formed by Zygnematacean algae have been found only in low-salinity lakes or freshwater lakes. These include Spirogyra, which was found in Lake Durankulak (Marinova and Atanossova, 2006), the montane inland Lake Izzyk-Kul, Lake Sapanca, and Lake Manzala of the Nile Delta. Zygospores of Spirogyra and the Zygnematacean genera Mougetia and Debarya were also found in sediments of the Nile Delta (Leroy, 1992). Zygnema spores (Plate 2, fig. 20) are common in the polluted, freshwater Lake Ulubat south of the Marmara Sea (Fig. 1). Pseudoschizaea rubina (Plate 2, fig. 9) may be a zygospore of the zygnematacean alga Debarya (Grenfell, 1995), although this relationship has not been confirmed by laboratory cultures, and it is often simply grouped with the acritarchs, as in Tables 1 and 2. This sphaeromorphitic acritarch, with distinctive concentric markings on both hemispheres of its dorsoventrally flattened vesicle, was first described as Sporites circulus in Pliocene brown coals, and then as Concentricystes rubina in marine sediments off Israel. Pseudoschizaea rubina is distinguished from the similar species Pseudoschizaea circula by an irregular, maze-like polar complex up to one quarter of the vesicle diameter (Christopher, 1976). Concentricystes s.l. is usually considered to be a freshwater alga because of its association with wadi or river terrace deposits (Christopher, 1976), and some species, including P. circula, have only been recorded from terrestrial or fluvial environments. However, both P. circula and P. rubina are occasionally present in the late Holocene marine sediments of the Black Sea but are absent from the early Holocene brackish-water sediments. Pseudoschizaea circula is rare in the hypersaline Lake Saki, and unspecified Pseudoschizaea species have been found in the Nile Delta, Red Sea, and Aral Sea. Desmids are most common in oligotrophic freshwater lakes and ponds, but some species e.g., Closterium aciculare,

Nonpollen palynomorphs mark eutrophic conditions (Graham and Wilcox, 2000). Desmid zygotes Coelastrum and Mougeotia occur in Lake Durankulak (Marinova and Atanassova, 2006), but in the Black Sea, Mougeotia and Closterium have been reported only for mid-Holocene sediments (Mudie et al., 2010), although they are markers of river transport to modern sediments in the Beaufort Sea (Matthiessen et al., 2000). Zygotes of Tetraedon and Coelastrum are sporadically abundant in the freshwater Lake Sapanca. Cyanobacteria Fossil Cyanobacteria (blue-green algae) are rare in the Black Sea corridor sediments, although the marine unicellular Synechococcus cynobacteria occurs in both eutrophic and oligotrophic waters of the Black Sea corridor (Uysal, 2006). Gloeotrichia is rare to common in the modern sediments of Lake Sapanca (Table 1; Plate 2, fig. 24), but in the Black Sea, Gloeotrichia-type sheaths were found only in the early Holocene sediments (Plate 2, fig. 23). Van Geel (2001) noted that Gloeotrichia marks nutrient-poor conditions in late glacial lakes because it is a nitrogen-fixing alga that subsequently makes conditions suitable for other aquatic plants. The planktonic filamentous heterocyst and akinete-producing alga Anabaena is present in both IssykKul and the Caspian Sea (Plate 2, fig. 16). Fungi and Animal Remains Other important nonpollen palynomorphs in the Black Sea are derived from fungi and various types of planktonic or benthic zooplankton. Plate 3 illustrates various fungal spores, conidia (Plate 3, figs. 1–5), ascomata (Plate 3, fig. 6), or germlings from Black Sea core 45. There are very few marine fungi, and most of the spores and other fungal remains must have been transported to the inland seas by air or soil erosion. Some dung fungi, such as Sporormiella (Plate 3, fig. 29), mark nutrient enrichment from domestic animals and are a reliable proxy for faunal biomass (van Geel and Aptroot, 2006). Several spore types (e.g., Tilletia, Ustilago) are produced by parasites of specific native plants and domestic crops. In archaeological middens, Glomus-type fungal spores (Plate 3, figs. 24 and 28) are extremely resistant to fire and biological degradation; hence, they persist in shell middens and soils after almost all the pollen has disappeared from degradation (Bryant and Holloway, 1983; Leroy et al., 2009), and they may survive very long-distance transport by river water. Glomus-type spores are present on the Nile cone and the Red Sea, in addition to the Black Sea and almost all lakes, unless the catchment is very small and the sample is far from the shore. Various zooplankton remains have been reported for the Black Sea corridor, including the chitinous skeletal remains of a juvenile cladoceran (Plate 3, fig. 21), copepod egg capsules (Plate 3, fig. 19) and their eggs (Plate 3, fig. 20), the organic linings of benthic microforaminifera (Plate 3, figs. 7, 13, 14), polychaete worm jaws and pincers (known to geologists as scolecodonts; Plate 3, fig. 10), and various morphologically similar

105

palynomorphs that include arthropod sclerites (Plate 3, fig. 15). The jaws of unknown ostracods were identified in Lakes Anzaleh and Sapanca (Plate 3, fig. 26) and in early Holocene sediments of the Black Sea (Plate 3, fig. 22). The lining of a very small ostracod (Plate 3, fig. 18; probably Leptocythere according to David Horne, 2007, personal commun.) is found in the Black Sea, and ostracod linings were also present in Lake Saki and on the Egyptian shelf. Eggs of the rotiferan Filinia longiseta were found only in the freshwater Lake Sapanca. Several types of tintinnid loricas are reported for the Aral Sea (Sorrel et al., 2006) and the Black Sea (Plate 3, fig. 8), and a Tintinnopsis is present in Lake Sapanca (Plate 3, fig. 9). The marine palynomorph Palaeostomocystis (Plate 3, fig. 11) also resembles tintinnid loricas found in arctic regions (Matthiessen et al, 2000); it is rare in the Marmara Sea. Other relatively large (~60–80 µm), brown, vase-shaped palynomorphs (Plate 3, figs. 16 and 17) occur in the Marmara Sea and on the Egyptian shelf. These nonpollen palynomorphs resemble the resting eggs (oocytes) or egg capsules/cocoons of microturbellarian flatworms (Platyhelminthes, Order Neorhabdocoela), which are mostly freshwater organisms (Haas, 1996); however, other turbellarians are common in coastal marine environments, and predatory marine flatworms parasitize mussels in the Black Sea (Murina and Grintsov, 1998). According to Ole Bennicke (2008, personal commun.), the egg walls of the marine flatworms tend to be thicker than those of the freshwater taxa. Morphologically similar nonpollen palynomorphs that occur in deep-sea sediments of the Banda Sea have been referred to the chitinous loricas of marine tintinnids (van Waveren, 1994). Clearly, more research on this group of palynomorphs is required before these nonpollen palynomorphs can be used reliably as environmental indicators when recovered from brackish-water or marine sediments. There is also further need to confirm the biological link between the brackish-water nonpollen palynomorphs referred to the tintinnids: several illustrations of vase- or urn-shaped testate amoebae living in peatland moss (Swindles and Roe, 2007) are very similar in size and morphology to some of the tintinnid nonpollen palynomorphs found in the Black Sea corridor. The organic-walled palynomorph Halodinium (Plate 3, fig. 12) was first recorded and described as an acritarch of unknown affinity occurring in subarctic marine sediments of the Bering Sea, and it is widely distributed in the Arctic (Matthiessen et al., 2000), including ponds of the Mackenzie Delta. The shape of Halodinium is similar to that of the testate amoebas Cyclopyxis and Arcella (illustrated by Beyens and Meisterfeld, 2001), and it is possible that this palynomorph is the organic lining of a testate amoeba (thecamoebian). The testate amoebas are a polyphyletic group of protozoans, the largest group (75%) being the Arcellinida. The empty tests remain intact after death of the amoeba and can be recovered fully from anaerobic sediments by dispersion in water and gentle sieving, but they decompose within a few weeks under aerobic conditions (Beyens and Meisterfeld, 2001). Extraction by palynological processing with acids and/or alkalis, however, gives variable recovery and does not produce a reliable picture of the thanatocoenoses (Swindles and Roe, 2007).

106

Mudie et al.

Nonpollen palynomorphs Halodinium is rare in the Black Sea and present in the Nile Delta and Red Sea. Unspecified thecamoebians have been reported for the calcareous mire near Nowshahr in northern Iran and are present in some freshwater lakes of the region. The distribution of these nonpollen palynomorphs in the surface samples of the Black Sea corridor (Table 1) shows that the fungi are most abundant in the low-salinity waters of the inland lakes, and the Black and Marmara Seas, they disappear in the Aegean, and then they return off the Nile Delta, being clearly linked to the proximity of terrestrial environments where they originate. In contrast, microforaminiferal linings continue to be present in the marine and high-salinity waters of the northern Aegean Sea (Aksu et al., 1999; Kotthof et al., 2008) and the Egyptian shelf (Kholeif, 2010), but they are absent from both the anoxic deep basins of the southern Black Sea and from late Holocene marine sediment of the Nile cone in the hyperoligotrophic water of the Levantine basin. Microforaminiferal linings might be expected to be good markers of sustained marine connection with the Mediterranean, but they also occur in the Kara-Bogaz Gol, which, during most of the late Holocene, was only connected to the global oceans by a canal linking the Volga River and the Azov Sea. They are also present in the hypersaline water of Lake Dzharylgach, which has been isolated from the Black Sea during historical time. Kotthoff et al. (2008) reported

Plate 3. Light microscope photographs of fungi and animal remains from Black Sea cores M02-45T, 30 cm, Marmara Sea core M98-12, Caspian Sea core 31, Lake Sapanca samples (SAK), Lake Dzharylgach, Egypt (NC), and the Red Sea. Scale bar 10 µm. Letters in parentheses are the initials of contributing authors other than P.J. Mudie. Figures 1–5. Fungal spores from SW Black Sea. (1, 2) Tilletia-type teliospores, M45T; (3) Chaetomium ascospore, M45T; (4) Valsaria sp. Ascospore; (5) Coniochaeta ligniaria ascospore (MM-F). Figure 6. Fungal germling similar to hyphopodium of Gaeumannomyces sp., B7, 70 cm. Figure 7. Caspian microforaminiferal lining CS31-3/2 (S.A.G.L.). Figure 8. Pacillina-type tintinnid lorica. Figure 9. Tintinnid lorica, with embedded charcoal fragments, SA0361 (S.A.G.L). Figure 10a. Simple scolecodont, core NC-1, Egyptian shelf (S.E.A.K.). Figure 10b. Bifurcated scolecodont; Red Sea (S.E.A.K.). Figure 11. Paleostomocytis sp.; Figure 12. Thecamoeban cf. Halodinium minor. Figures 13, 14, Microforaminiferal linings; (13) trochospiral form, NC-1, Egyptian shelf (S.E.A.K.); (14) Black Sea Mar02-45 planispiral, open lining. Figure 15. Arthropod sclerite, Lake Ulubat, AK104_24.5_1 (S.A.G.L.). Figures 16, 17. Cf. tintinnid loricae/turbellarian egg capsules from core M98-12, 30 cm; (16) lorica type 1, with short apiculate base; (17) lorica type 2, with rounded base. Figure 18. Organic lining of brackish water ostracod, cf. Leptocythere, M02-45, 790 cm. Figures 19, 20. Copepod eggs; (19) copepod eggs; (12) ?copepod egg capsule with fibrous outer wall; M45. Figure 20. Spiny copepod egg. Figure 21. Exoskeletal parts (probably the first thoracic antenna, 150 μm long) of a juvenile cladoceran, M45P, 440 cm. Figure 22. Mouth part of a small ostracod. Figure 23. Sorosporium-type spore mass; Red Sea (S.E.A.K.). Figure 24. Glomus-type fungal spore, NC core 2 (S.E.A.K.). Figure 25. M45T Desmid or Tardigrade egg. Figure 26. Ostracod jaw CP14, 45 cm (S.A.G.L.). Figure 27. Unknown amphipod from Lake Dzharylgach, Dz70, 160–155 cm (T.S.). Figure 28. Glomus group SA03K71-13 (S.A.G.L.). Figure 29. Sporormiella spore SAK71-1 (S.A.G.L.).

107

that microforaminiferal linings are among the most oxidationsensitive palynomorphs. These palynomorphs are abundant and well-preserved at ~1000 m water depth in the Mt. Athos Basin of the northwest Aegean; therefore, their absence in the anoxic ~2-km-deep basins of the southeastern Black Sea strongly suggests that few benthic foraminifera live under conditions of low surface salinity (200 µm) arthropod parts from lake sediments include a sclerite from an aquatic Hemipteran (Plate 3, fig. 15), and it is possible that the smaller (100 µm) sclerites of arthropods are frequently found in freshwater lakes and lagoons, e.g., Lakes Ulubat and Sapanca, and in the inland brackish Caspian Sea. The size and shape of these nonpollen palynomorphs differ considerably from the sclerotic mouthparts of copepods and polychaetes found in continental shelf sediments off California (Mudie, 2009, personal observation). Ostracoda, Cladocera, and Chironomidae are arthropod groups that have received a lot of attention recently in Holocene studies owing to their potential to reconstruct past water temperatures and salinities (Smol et al., 2001). Ostracod mandibles (Fig. 5) and shells (Plate 3, fig. 18) and Cladoceran (Plate 3, fig. 21) remains are complex structures, and their identification is outside of the normal specialization of a palynologist. The jaws of ostracods (Plate 3, figs. 22 and 26) can be distinguished from the similar but more complicated mouthparts of chironomid midge larvae (Fig. 5; Walker, 2001; Eggermont et al., 2008). Chironomids have a joined symmetrical left and right part with identical rows of teeth, and the palynomorphs are usually found with cojoined central and ventral jaw parts. In contrast, the left and right teeth of ostracods are not joined, and the palynomorphs are found as single jaws with toothed ends. Anthropogenic influences in the Black Sea corridor are most clearly marked by changes in forest pollen influxes (Mudie et al., 2007) that correspond to deforestation after ca. 7.5 ka and by the appearance of cereal pollen around 5 ka, followed by various horticultural taxa, e.g., olives and vine grapes. However, nonpollen palynomorphs are unique palynomorphs for tracing the history of livestock production from marine sediment records. For example, the occurrence of herbivore dung fungal spores of Sporormiella at coastal sites off Varna clearly shows that livestock agriculture was practiced by the Bronze Age, but there is no evidence of earlier animal farms. In Lake Durankulak, dung-fungus spores of Chaetomium, Coniochaeta, Podospora, and Sordaria indicate extensive local stockbreeding and grazing during the Early Bronze Age (Marinova and Atanassova, 2006). Peaks of Neurospora spores, a fire indicator, correspond to maxima in charcoal particles and Glomustype spores that indicate soil erosion (Marinova and Atanassova, 2006). The smut fungus Sorosporium parasitizes mainly grasses, including important crops like sorghum, maize, and millet, and the presence of Sorosporium-type spores in the Red Sea marks long-distance transport from highland regions where the crops are mainly grown (Kholeif, 2004). In Issyk-Kul, increased Botryococcus and fungal spore percentages are taken as signs of greater erosion and heavier grazing in the catchment after A.D. 1560 (Giralt et al., 2004). Dinocyst records are also important for understanding the history of toxic red-tide blooms in the Black Sea corridor. In MAR02-45, it is clear that blooms of the toxic species L. polyedrum are endemic to the Black Sea waters. In contrast, the late

Nonpollen palynomorphs arrivals of the toxic red-tide species Gymnodinium catenatum and Alexandrium-cyst species around 2400 and 500 yr B.P., respectively, appear to be related to recent introductions via ship-ballast discharge, which today persists as a serious environmental problem in the Black Sea (e.g., Moncheva and Kamburska, 2002). Recent toxic cyst–forming species include Alexandrium monolatum (first seen in 1991), Gymnodinium uberrimum (1994), G. fuscum (1970–1980), and Gyrodinium cf. aureolatum or G. mikkimotoi (1970–1980). These cysts have not yet been recovered as fossils in the surface sediments, however, so it is not yet clear if they survive and reproduce in the Black Sea.

111

CONCLUSIONS There is a large range of nonpollen palynomorphs in the Black Sea corridor that remain to be fully exploited by palynologists. Only a small proportion has been identified and related to living organisms or, failing that, have been given a name or a type number. Nonetheless, the change from use of geological fossil names (e.g., Cymatiosphaera) to biological names (e.g., Pterosperma) by Quaternary paleoecologists illustrates the great progress made in the study of Holocene micro- and macrofossil groups during the last decades. Within much of the old geological

Figure 7. Diagrams of characteristic features of nonpollen palynomorphs with possible environmental importance in the Black Sea–Mediterranean corridor. (A-1) Ostracod mandible, showing last mandible coxa and teeth (from Horne et al., 2002); scale bar = 50 µm. (A-2) Chironomid mouthparts (from Walker, 2001). m—mentum; v—ventromentum plates. (B) Microforaminiferal shapes (after Stancliffe, 1996): 1—single chamber; 2—uniserial;3—biserial type ii; 4—coiled biserial; 5—coiled uniserial; 6—planispiral; 7—spiral uniserial; 8—spiral coiled; 9—trochospiral. (C) Pseudoschizaea species (after Christopher, 1976).

112

Mudie et al.

information, the new biological identifications have led to new, more in-depth interpretations of environmental change. These powerful new paleoenvironmental tools are likely to spread fairly quickly to the whole of the Pleistocene and further down the Cenozoic. Depending on the awareness of the various palynologists, and the use of acetolysis versus hydrofluoric acid for sediment processing, only a small portion of the nonpollen palynomorphs in marine/brackish water lakes in the Black Sea corridor has been regularly counted and published. However, our overview of the available, mostly qualitative data clearly shows the value of extracting the full complement of nonpollen palynomorphs, particularly thin-walled peridinioid dinocysts and fungal spores, because of the potential for detailed interpretation of paleoenvironmental conditions and human activity (e.g., livestock husbandry, agriculture, and burning). The most important initial results are summarized as follows. 1. The surface distributions of the nonpollen palynomorphs dinocysts, Pediastrum and Botryococcus, zygnematacean algae, and zooplankton remains show that during the early Holocene, the Black Sea was a brackish sea like the modern Caspian Sea or the outer Baltic Sea: It was not a freshwater lake. The presence of Pediastrum may indicate relatively high phosphorus levels from river inflow, but there are no nonpollen palynomorph indices of fire, soil erosion, or human settlement. 2. Core-top samples from the wide range of salinity in the Black Sea corridor show that the prasinophyte Pterosperma (fossil name Cymatiosphaera) and other small unicellular acritarchs are usually more abundant in low-salinity environments and they include freshwater species; hence, Traverse’s marine influence index cannot be used as a reliable marker of detailed sea-level change. 3. Peaks in the fossil Cymatiosphaera globosa (= Prasinophyte Pterosperma), colonial algae, and some dinocysts, e.g., Lingulodinium machaerophorum (= L. polyedrum), mainly reflect nutrient levels and stratification of the water column, and are not reliable markers of sea-level change. 4. Fungal spores appear to be the best index of terrigenous input from soil erosion, and they are important markers of Bronze Age farming practices. 5. Laminated sediments in the hypersaline liman-lagoons of the Crimean Peninsula are characterized by high concentrations of pollen regardless of salinity, by an absence of dinocysts, and variable amounts of low-diversity nonpollen palynomorph assemblages. Ponds with NaCl salt concentrations greater than 100‰ have few nonpollen palynomorphs compared to ponds with a salinity of 40‰–80‰. High concentrations of the acritarch Micrhystridium may characterize these environments. 6. In palynological preparations for arthropod and polychaete groups, only partial information is left that does not allow detailed paleoenvironmental reconstruction compared to the algal nonpollen palynomorphs groups, e.g., Chlorophyceae and Cyanobacteria, where identification to generic and species level is possible and leads to significant enhancement of

palynological spectra interpretation. More research on the biological links between sclerotized microfaunal remains and morphologically similar nonpollen palynomorphs in brackish and marine environments is required before precise environmental interpretations can be made from this group of Pleistocene– Holocene palynomorphs. ACKNOWLEDGMENTS The project International Geological Correlation Programme (IGCP) 521 “Black Sea–Mediterranean corridor during the last 30 k.y.: Sea-level change and human adaptation” allowed the co-authors to meet and initiate the idea of this manuscript at the 2007 joint meeting with IGCP 481 “Dating Caspian sea level change.” The senior author acknowledges field-work support from various scientists. Black Sea, Marmara, and Aegean core samples were made available by A.E. Aksu from the archives at Memorial University of Newfoundland (MUN), providing the foundation for this study, and the technical assistance of Helen Gillespie, MUN, for sample processing is gratefully acknowledged. We thank David Horne (Queen Mary University London), Ian Walker (Okanagan University College), Ole Bennike (Geological Survey Denmark and Greenland), and Dirk Verschuren (Ghent University) for help with identifications of animal remains. The Gloeotrichia filament photo (Plate 2, fig. 15) was taken and permission granted for use by Peter A. Siver, at Chrysophytes LLC. Bas van Geel (University of Amsterdam) provided much helpful advice, and we thank J.H. McAndrews (University of Toronto) and J. Matthiessen (Alfred Wegener Institute) for their critical reviews. REFERENCES CITED Aksu, A.E., Yasar, D., and Mudie, P.J., 1995, Paleoclimatic and paleoceanographic conditions leading to development of sapropel layer S1 in the Aegean Sea: Micropaleontological and stable isotope evidence: Palaeogeography, Palaeoclimatology, Palaeoecology, v. 116, p. 71–101, doi: 10.1016/0031-0182(94)00092-M. Aksu, A.E., Abrajano, T., Mudie, P.J., and Yaşar, D., 1999, Organic geochemical and palynological evidence for terrigenous origin of the organic matter in Aegean Sea sapropel S1: Marine Geology, v. 153, p. 303–318, doi: 10.1016/S0025-3227(98)00077-2. Aladin, N.V., and Potts, W.T.W., 1992, Changes in the Aral Sea ecosystem during the period 1960–1990: Hydrobiologia, v. 237, p. 67–79, doi: 10.1007/ BF00016032. Atanassova, J., 2005, Paleoecological setting of the western Black Sea during the past 15000 years: The Holocene, v. 15, no. 4, p. 576–584, doi: 10.1191/0959683605hl832rp. Batten, D.J., 1996, Chapter 26B. Palynofacies and palaeoenvironmental interpretation, in Jansonius, J., and McGregor, D.C., eds., Palynology: Principles and Applications: American Association of Stratigraphic Palynologists Foundation, v. 3, p. 1065–1084. Batten, D.J., and Grenfell, H.R., 1996, Botryococcus, in Jansonius, J., and McGregor, D.C., eds., Palynology: Principles and Applications: American Association of Stratigraphic Palynologists Foundation, v. 1, p. 205–214. Bérard-Therriault, L., Poulin, M., and Bossé, L., 1999, Guide d’Identification du Phytoplankton Marin de l’Estuaire et du Golfe du Saint-Laurent Incuant également Certains Protozaires: Publication Spéciale Canadienne des Sciences Haliétiques et Aquatiques 128, 387 p. Beyens, L., and Meisterfeld, R., 2001, Protozoa: Testate amoebae, in Smol, J.P., Birks, H.J.B, and Last, W.M., eds., Tracking Environmental Change Using

Nonpollen palynomorphs Lake Sediments. Volume 3: Terrestrial, Algal and Siliceous Indicators: Dordrecht, The Netherlands, Kluwer Academic Publishers, p. 121–153. Brenner, W.W., 2001, Organic-walled microfossils from the central Baltic Sea, indicators of environmental change and base for ecostratigraphic correlation: Baltica, v. 14, p. 40–51. Brenner, W.W., 2005, Holocene environmental history of the Gotland basin (Baltic Sea)—A micropalaeontological model: Palaeogeography, Palaeoclimatology, Palaeoecology, v. 220, p. 227–241, doi: 10.1016/j.palaeo.2004.12.010. Brenner, W.W., 2006, Organische Mikrofossilien der holozänen Ostsee: Meyniana, v. 58, p. 33–55. Bryant, V.M., Jr., and Holloway, R.G., 1983, The role of palynology in archaeology: Advances in Archaeology Method and Theory, v. 6, p. 191–224. Bryant, V.M., and Wrenn, J.H., eds., 1998, New Developments in Palynomorph Sampling, Extraction, and Analysis: American Association of Stratigraphic Palynologists Contributions Series 33, 155 p. Chmura, G.L., Stone, P.A., and Ross, M.S., 2006, Non-pollen microfossils in Everglades sediments: Review of Palaeobotany and Palynology, v. 141, p. 103–119, doi: 10.1016/j.revpalbo.2006.03.019. Christopher, R.A., 1976, Morphology and taxonomic status of Pseudoschizaea Thiergart and Frantz ex R. Potonié emend: Micropalaeontology, v. 22, p. 143–150, doi: 10.2307/1485396. Dale, B., 1976, Cyst formation, sedimentation, and preservation: Factors affecting dinoflagellate assemblages in recent sediments from Trodheimsfjord, Norway: Review of Palaeobotany and Palynology, v. 22, p. 39–60, doi: 10.1016/0034-6667(76)90010-5. Dale, B., 1996, Dinoflagellate cyst ecology: Modeling and geological applications, in Jansonius, J., and McGregor, D.C., eds., Palynology: Principles and Applications: American Association of Stratigraphic Palynologists Foundation, v. 3, p. 1249–1275. Dawes, C.J., 1998, Marine Botany (2nd ed.): New York, John Wiley & Sons, 480 p. Denys, L., and de Wolf, H., 1999, Diatoms as indicators of coastal paleoenvironments and relative sea-level change, in Stoermer, E.F., and Smol, J.P., eds., The Diatoms: Applications for the Environmental and Earth Sciences: Cambridge, UK, Cambridge University Press, p. 277–297. Djamali, M., de Beaulieu, J.-L., Shah-Hosseini, M., Andrieu-Ponel, V., Ponel, P., Amini, A., Akhani, H., Leroy, S.A.G., Stevens, L., Lahijani, H., and Brewer, S., 2008, A late Pleistocene long pollen record from Lake Urmia, N.W. Iran: Quaternary Research, v. 69, p. 413–420. Eggermont, H., Verschuren, D., Fagot, M., Rumes, B., Van Bocxlaer, B. and Kröpelin, S., 2008, Aquatic community response in a groundwater-fed desert lake to Holocene desiccation of the Sahara: Quaternary Science Reviews, v. 27, p. 2411–2425. Fensome, R.A., and Williams, G.L., 2004, The Lentin and Williams Index of Fossil Dinoflagellates: American Association of Stratigraphic Palynologists Contributions Series 42, 909 p. Fensome, R.A., Williams, G.I., Barss, S., Freeman, J.M., and Hill, J.M., 1990, Acritarchs and Fossil Prasinophytes: An Index to Genera, Species and Infraspecific Taxa: American Association of Stratigraphic Palynology Foundation Contributions Series 25, 771 p. Filipova-Marinova, M., and Christova, R., 2004, Sea-level fluctuations in the western part of the Black Sea during the Holocene: Journal of Environmental Micropalaeontology, Microbiology and Meiobenthology, v. 1, p. 122–135. Gerasimenko, N., and Subetto, D., 2007, Vegetational and climatic changes in the southwestern Crimea during the middle-late Holocene, based on pollen study of the Lake Saki, in Yanko-Hombach, V., Buinevich, I., Dolukhanov, P., Gilbert, A., Martin, R., McGann, M., and Mudie, P., eds., Extended Abstracts of the Joint Plenary Meeting and Field Trip of IGCP 521 “Black Sea–Mediterranean corridor during the last 30 k.y.: Sea-level change and human adaptation,” and IGCP 481 “Dating Caspian Sea Level Change,” 8–17 September 2007: Gelendzhik (Russia)–Kerch (Ukraine), Southern Branch of the Institute of Oceanology, Russian Academy of Sciences and Demetra Beneficent Foundation, p. 69–71. Giralt, S., Julià, R., Klerkx, J., Riera, S., Leroy, S.A.G., Buchaca, T., Catalan, J., De Batist, M., Beck, C., Bobrov, V., Gavshin, V., Kalugin, I., Sukhorukov, F., Brennwald, M., Kipfer, R., Peeters, F., Lombardi, S., Matychenkov, V., Romanovsky, V., Podsetchine, V., and Voltattorni, N., 2004, 1,000 year environmental history of Lake Issyk-Kul, in Nihoul, J., Zavialov, P., and Micklin, P., eds., Dying and Dead Seas; Climatic versus Anthropic Causes: Dordrecht, Netherlands, Kluwer Academic publishers, NATO ASI Series, v. 36, p. 253–285. Graham, L.E., and Wilcox, L.W., 2000, Algae: Upper Saddle River, New Jersey, Prentice Hall, 640 p.

113

Grenfell, H.R., 1995, Probable fossil zygnematacean algal spore genera: Review of Palaeobotany and Palynology, v. 84, p. 201–220, doi: 10.1016/0034 -6667(94)00134-6. Guy-Olsen, D., 1996, Chapter 7. Green and blue-green algae. 7B. Prasinophycean algae, in Jansonius, J., and McGregor, D.C., eds., Palynology: Principles and Applications: American Association of Stratigraphic Palynologists Foundation, v. 1, p. 181–189. Haas, J.N., 1996, Neorhabdocoela oocytes—Palaeoecological indicators found in pollen preparations from Holocene freshwater lake sediments: Review of Palaeobotany and Palynology, v. 91, p. 371–382, doi: 10.1016/0034 -6667(95)00074-7. Head, M.J., 1992, Zygospores of the Zygnemataceae (Division Chlorophyta) and other freshwater algal spores from the uppermost Pliocene St. Erth Beds of Cornwall, southwestern England: Micropaleontology, v. 38, no. 3, p. 237–260, doi: 10.2307/1485790. Head, M.J., 1993, Dinoflagellates, sporomorphs and other palynomorphs from the upper Pliocene St. Erth Beds of Cornwall, southwestern England: Journal of Paleontology, v.67, no. III, p. 1–62. Heim, C., Nowaczyk, N., Negendank, J., Leroy, S., and Ben-Avraham, Z., 1997, Middle-East desertification: Evidence from the Dead Sea: Naturwissenschaften, v. 84, no. 9, p. 398–401, doi: 10.1007/s001140050416. Hiscott, R.N., Aksu, A.E., Mudie, P.J., Marret, F., Abrajano, T., Kaminski, M.A., Evans, J., Cakiroglu, I., and Yaşar, D., 2007, A gradual drowning of the southwestern Black Sea shelf: Evidence for a progressive rather than abrupt Holocene reconnection with the Eastern Mediterranean Sea through the Marmara Sea gateway: Quaternary International, v. 167-168, p. 19–34, DOI: 10.1016/j.quaint.2006.11.007. Horne, D.J., Cohen, A., and Martens, K., 2002, Taxonomy, morphology and biology of Quaternary and living Ostracoda, in Holmes, J., and Chivas, A., eds., The Ostracoda: Applications in Quaternary Research: American Geophysical Union Geophysical Monograph 131, p. 5–36. Ignatiades, L., 2005, Scaling the trophic status of the Aegean Sea, Eastern Mediterranean: Journal of Sea Research, v. 54, p. 51–57, doi: 10.1016/ j.seares.2005.02.010. Kalgutkar, R.M., and Jansonius, J., 2000, Synopsis of Fossil Fungal Spores, Mycelia and Fructifications: American Association Stratigraphic Palynologists Foundation Contributions Series 39, 423 p. Kazancı, N., Gulbabazadeh, T., Leroy, S., and Ileri O., 2004, Sedimentary and environmental characteristics of the Gilan-Mazenderan Plain, northern Iran: Influence of long- and short-term Caspian water level fluctuations on geomorphology: Journal of Marine Systems, v. 46, no. 1–4, p. 145–168. Kholeif, S.E.A., 2004, Palynology and palaeovegetation reconstruction in late Quaternary sediments of the southern Suez Isthmus: Egypt: Journal of African Earth Sciences, v. 22, no. 1, p. 93–112. Kholeif, S.E.A., 2007, Palynology of mangrove sediments in the Hamata area, Red Sea coast, Egypt: Vegetation and restoration overview, in Isermann, M., and Kiehl, K., eds., Restoration of Coastal Ecosystems: Coastline Reports, v. 7, p. 5–16. Kholeif, S.E., 2010, Holocene paleoenvironmental change in inner continental shelf sediments, Southeastern Mediterranean, Egypt: Journal of African Earth Sciences, v. 57, p. 143–153. Kholeif, S.E.A., and Mudie, P.J., 2009, Palynomorph and amorphous organic matter records of climate and oceanic conditions in late Pleistocene and Holocene sediments of the Nile cone, southeastern Mediterranean: Palynology, v. 32, p. 1–24. Kholeif, S.E.A., Ibrahim, M.I.A., and Dilcher, D., 2007, Pollen evidence for Holocene vegetation and anthropogenic influence in the Fayoum Depression, Egypt: Egyptian Journal of Paleontology, v. 7, p. 171–195. Korhola, A., and Rautio, M., 2002, Cladocera and other branchiopod crustaceans, in Smol, J.P., Birks, H.J.B, and Last, W.M., eds., Tracking Environmental Change Using Lake Sediments, Volume 4: Zoological Indicators: Dordrecht, The Netherlands, Kluwer Academic Publishers, p. 5–42. Kotthof, U., Pross, J., Muller, U.C., Peyron, O., Schmiedl, G., Schultz, H., and Bordon, A., 2008, Climate dynamics in the borderlands of the Aegean Sea during the formation of sapropel S1 deduced from a marine pollen record: Quaternary Science Reviews, v. 27, p. 832–845. Kouli, K., Brinkhuis, H., and Dale, B., 2001, Spiniferites cruciformis: A freshwater dinoflagellate: Review of Palaeobotany and Palynology, v. 113, p. 273–286, doi: 10.1016/S0034-6667(00)00064-6. Leroy, S.A.G., 1992, Palynological evidence of Azolla nilotica Dec. in Recent Holocene of eastern Nile Delta, and its environment: Vegetation History and Archaeobotany, v. 1, p. 43–52, doi: 10.1007/BF00190700.

114

Mudie et al.

Leroy, S.A.G., and Albay, M., 2010, Palynomorphs of brackish and marine species in cores from the freshwater Lake Sapanca, NW Turkey: Review of Palaeobotany and Palynology, v.160, no. 3, p. 181–188. Leroy, S.A.G., Marret, F., Giralt, S., and Bulatov, S.A., 2006, Natural and anthropogenic rapid changes in the Kara-Bogaz Gol over the last two centuries by palynological analyses: Quaternary International, v. 150, p. 52–70, doi: 10.1016/j.quaint.2006.01.007. Leroy, S.A.G., Marret, F., Gibert, E., Chalié, F., Reyss, J.-L., and Arpe, K., 2007, River inflow and salinity changes in the Caspian Sea during the last 5500 years: Quaternary Science Reviews, v. 26, p. 3359–3383, doi: 10.1016/j.quascirev.2007.09.012. Leroy, S.A.G., Boyraz, S., and Gürbüz, A., 2009, High-resolution palynological analysis in Lake Sapanca as a tool to detect earthquakes on the North Anatolian Fault over the last 55 years: Quaternary Science Reviews, v. 28, no. 25-26, p. 2616–2632. Lewis, J., and Hallett, R.I., 1997, Lingulodinium polyedrum (Gonyaulax polyedra) a blooming dinoflagellate, in Ansell, A., Barnes, M., and Gibson, R.N., eds., Oceanography and Marine Biology: An Annual Review, v. 35, p. 96–161. Limaye, R.B., Kumaran, K.P.N., Nair, K.M., and Padmalal, D., 2007, Nonpollen palynomorphs as potential palaeoenvironmental indicators in the late Quaternary sediments of the west coast of India: Current Science, v. 92, no. 10, p. 1370–1382. Marinova, E., and Atanassova, J., 2006, Anthropogenic impact on vegetation and environment during the Bronze Age in the area of Lake Durankulak, NE Bulgaria: Pollen, microscopic charcoal, non-pollen palynomorphs and plant macrofossils: Review of Palaeobotany and Palynology, v. 141, p. 165–178, doi:10.1016/j.revpalbo.2006.03.011. Marret, F., 1993, Les effets de l’acétolyse sur les assemblages des kystes de dinoflagellés: Palynosciences, v. 2, p. 267–272. Marret, F., and Zonneveld, K.A.F., 2003, Atlas of modern organic-walled dinoflagellate cyst distributions: Review of Palaeobotany and Palynology, v. 125, p. 1–200. Marret, F., Leroy, S.A.G., Chalié, F., and Gasse, F., 2004, New organicwalled dinoflagellate cysts from recent sediments of central Asian seas: Review of Palaeobotany and Palynology, v. 129, p. 1–20, doi: 10.1016/ j.revpalbo.2003.10.002. Marret, F., Mudie, P.J., and Kholeif, S., 2008, Dinocyst assemblages in the Black Sea–Mediterranean corridor, in Eighth Dinocyst Workshop Abstracts, 4–10 May 2008: Montreal, Quebec. Marret, F. Mudie, P., Aksu, A.E., and Hiscott, R.N., 2009, A Holocene dinocyst record of a two-step transformation of the Neoeuxinic brackish water lake into the Black Sea: Quaternary International, v. 197, no. 1-2, p. 72–86, doi:10.1016/j.quaint.2007.01.010. Martin, F., 1993, Acritarchs: A review: Biological Reviews of the Cambridge Philosophical Society, v. 6, p. 475–538. Mathison, S.W., and Chmura, G.L., 1995, Utility of microforaminiferal test linings in palynological preparations: Palynology, v. 19, p. 77–84. Matthiessen, J., and Brenner, W., 1996, Chlorococcalalgen und dinoflagellatenZysten in rezenten sedimenten des greifswalder boddens (südliche Ostesee): Senckenbergiana Maritima, v. 27, no. 1/2, p. 33–48. Matthiessen, J., Kunz-Pirrung, M., and Mudie, P., 2000, Freshwater chlorophycean algae in recent marine sediments of the Beaufort, Laptev and Kara Seas (Arctic Ocean) as indicators of river runoff: International Journal of Earth Sciences, v. 89, p. 470–485, doi: 10.1007/s005310000127. Mertens, K.N., Ribeiro, S., Bouimetarhan, I., Caner, H., Combourieu-Nebout, N., Dale, B., de Vernal, A., Ellegaard, M., Filipova, M., Godhe, A., Goubert, E., Grøsfjeld, K., Holzwarth, U., Kotthof, U., Leroy, S.A.G., Londeix, L., Marret, F., Matsuoka, K., Mudie, P.J., Naudts, L., PeñaManjarrez, J.L., Persson, A., Popescu, S.-M., Pospelova, V., Sangiorgi, F., Van Der Meer, M.T.J., Vink, A., Zonneveld, K.A.F., Vercauteren, D., Vlassenbroeck, J., and Louwye, S., 2009, Process length variation in cysts of Lingulodinium machaerophorum, in surface sediments; investigating its potential as salinity proxy: Marine Micropaleontology, v. 70, p. 54–69, doi: 10.1016/j.marmicro.2008.10.004. Moncheva, S.P., and Kamburska, L.T., 2002, Plankton stowaways in the Black Sea—Impacts on biodiversity and ecosystem health: Commission Internationale pour l’Exploration Scientifique de la mer Méditerranée Workshop Monograph 20, p. 47–52 (www.ciesm.org/publications/Istanbul02.pdf). Mudie, P.J., Aksu, A.E., and Yaşar, D., 2001, Late Quaternary dinoflagellate cysts from the Black, Marmara and Aegean Seas: Variations in assem-

blages, morphology and paleosalinity: Marine Micropaleontology, v. 43, p. 155–178, doi: 10.1016/S0377-8398(01)00006-8. Mudie, P.J., Rochon, A., Aksu, A.E., and Gillespie, H., 2002, Dinoflagellate cysts, freshwater algae and fungal spores as salinity indicators in late Quaternary cores from Marmara and Black Seas: Marine Geology, v. 190, p. 203–231, doi: 10.1016/S0025-3227(02)00348-1. Mudie, P.J., Rochon, A., Aksu, A.E., and Gillespie, H., 2004, Recent dinoflagellate cyst assemblages in the Aegean–Marmara–Black Sea corridor: Statistical analysis and re-interpretation of the early Holocene Noah’s Flood hypothesis: Review of Palaeobotany and Palynology, v. 128, p. 143–167, doi: 10.1016/S0034-6667(03)00117-9. Mudie, P. J., Marret, F., Aksu, A. E., Hiscott, R. N., and Gillespie, H., 2007, Palynological evidence for climatic change, anthropogenic activity and outflow of Black Sea water during the late Pleistocene and Holocene: Centennial- to decadal-scale records from the Black and Marmara Seas: Quaternary International, v. 167-168, p. 73–90, DOI: 10.1016/ j.quaint.2006.11.009. Mudie, P.J., Marret, F., Rochon, A., and Aksu, A.E., 2010, Non-pollen palynomorphs in the Black Sea corridor: Vegetation History and Archaeobotany, v. 334, DOI: 10.1007/s00334-010-0268-9. Mukhanov, V.S., Naidanova, O.G., Nickolay, V., Shadrin, V., and Kemp, R.B., 2004, The spring energy budget of the algal mat community in a Crimean hypersaline lake determined by microcalorimetry: Aquatic Ecology, v. 38, p. 375–385, doi: 10.1023/B:AECO.0000035169.08581.10. Murina, V.V., and Grintsov, V.A., 1998, The ecology of a predatory Black Sea turbellarian: Hydrobiological Journal, v. 34, p. 54–61. Nicholls, K.H., 1997, Planktonic green algae in western Lake Erie: The importance of temporal scale in the interpretation of change: Freshwater Biology, v. 38, p. 419–425, doi: 10.1046/j.1365-2427.1997.00247.x. Piontkovski, S., and Elmuratov, A. E., 2008, Abundance of Algae in the Aral Sea: A.O. Kovalevsky Institute of Biology of the Southern Seas, The National Academy of Sciences of Ukraine, doi: 10.1594/PANGAEA.695221. Playford, G., 2003, Acritarchs and Prasinophyte Phycomata: A Short Course: American Association of Stratigraphic Palynologists Foundation Contributions Series 41, 39 p. Ramezani, E., Mohdjer, M.R.M., Knapp, H.-D., Ahmadi, H., and Joosten, H., 2008, The late-Holocene vegetation history of the central Caspian (Hyrcanian) forests of northern Iran: The Holocene, v. 18, p. 307–321, doi: 10.1177/0959683607086768. Riding, J.B., and Kyffin-Hughes, J.E., 2006, Further testing of a non-acid palynological preparation procedure: Palynology, v. 30, p. 69–88, doi: 10.2113/gspalynol.30.1.69. Roman, S., 1974, Palynoplanktologic analysis of some Black Sea cores, in Degens, E.T., and Ross, D.A., eds., The Black Sea: Geology, Chemistry, and Biology: American Association of Petroleum Geologists Memoir 20, p. 396–410. Rossignol-Strick, M., 1995, Sea-land correlation of pollen records in the Eastern Mediterranean for the glacial-interglacial transition: Biostratigraphy versus radiometric time-scale: Quaternary Science Reviews, v. 14, p. 893–915, doi: 10.1016/0277-3791(95)00070-4. Saniawski, H., 1996, Chapter 12. Scolecodonts, in Jansonius J., and McGregor, D.C., eds., Palynology: Principles and Applications: American Association of Stratigraphic Palynologists Foundation, v. 1, p. 337–354. Sapelko, T.V., and Subetto, D., 2007, Vegetation change reconstruction during the Holocene in the north-western Crimea, based on pollen data, in Yanko-Hombach, V., Buynevich, I., Dolukhanov, P., Gilbert, A., Martin, R., McGann, M., and Mudie, P., eds., Extended Abstracts of the Joint Plenary Meeting and Field Trip of IGCP 521 “Black Sea–Mediterranean corridor during the last 30 k.y.: Sea-level change and human adaptation,” and IGCP 481 “Dating Caspian Sea Level Change,” 8–17 September 2007: Gelendzhik (Russia)–Kerch (Ukraine), Southern Branch of the Institute of Oceanology, Russian Academy of Sciences and Demetra Beneficent Foundation, p. 135–136. Smol, J.P., Birks, J.B., and Last, W.M., 2001, Zoological indicators in lake sediments: An introduction, in Smol, J.P., Birks, H.J.B, and Last W.M., eds., Tracking Environmental Change Using Lake Sediments. Volume 4: Zoological Indicators: Dordrecht, The Netherlands, Kluwer Academic Publishers, p. 1–4. Solomina, O., Davi, N., D’Arrigo, R., and Jacoby, G., 2005, Tree-ring reconstruction of Crimean drought and lake chronology correction: Geophysical Research Letters, v. 32, p. L19704, doi: 10.1029/2005GL023335.

Nonpollen palynomorphs Sorrel, P., Popescu, S.-M., Head, M.J., Suc, J.P., Klotz, S., and Oberhansli, H., 2006, Hydrographic development of the Aral Sea during the last 2000 years based on a quantitative analysis of dinoflagellate cysts: Palaeogeography, Palaeoclimatology, Palaeoecology, v. 234, p. 304–327, doi: 10.1016/j.palaeo.2005.10.012. Stancliffe, R.P.W., 1996, Microforaminiferal linings, in Jansonius, J., and McGregor, D.C., eds., Palynology: Principles and Applications: American Association of Stratigraphic Palynologists Foundation, v. 1, p. 373–379. Strother, P.K., 1996, Chapter 5: Acritarchs, in Jansonius, J., and McGregor, D.C., eds., Palynology: Principles and Applications: American Association of Stratigraphic Palynologists Foundation, v. 1, p. 81–106. Swindles, G.T., and Roe, H.M., 2007, Examining the dissolution characteristics of testate amoebae (Protozoa: Rhizopoda) in low pH conditions: Implications for peatland palaeoclimatic studies: Palaeogeography, Palaeoclimatology, Palaeoecology, v. 252, p. 486–496, doi: 10.1016/ j.palaeo.2007.05.004. Taylor, F.J.R., 1987, General and marine ecosystems, in Taylor, F.J.R., ed., Ecology of Dinoflagellates: Oxford, UK, Blackwell Science Publications, p. 398–502. Tomas, C.R., 1993, Marine Phytoplankton: San Diego, Academic Press, 263 p. Traverse, A., 1974, Palynological investigation of two Black Sea cores, in Degens, E.T., and Ross, D.A., eds., The Black Sea: Geology, Chemistry, and Biology: American Association of Petroleum Geologists Memoir 20, p. 381–388. Traverse, A., 1978, Palynological analysis of DSDP Leg 42B (1975) cores from the Black Sea, in Ross D.A., et al., Initial Reports of the Deep Sea Drilling Program, v. XLII, Part 2: Washington, D.C., Government Printing Office, p. 993–1015. Turney, C.S.M., and Brown, H., 2007, Catastrophic early Holocene sea level rise, human migration and the Neolithic transition in Europe: Quaternary Science Reviews, v. 26, p. 2036–2041. Uysal, Z., 2006, Vertical distribution of marine cyanobacteria Synechococcus spp. in the Black, Marmara, Aegean and eastern Mediterranean Seas: Deep-Sea Research. Part II, Topical Studies in Oceanography, v. 53, p. 1976–1987, doi: 10.1016/j.dsr2.2006.03.016. van der Meer, M.T.J., Sangiorgi, F., Baas, M., Brinkhuis, H., Sinninghe Damsté, J.S., and Schouten, S., 2008, Molecular isotopic and dinoflagellate evidence for late Holocene freshening of the Black Sea: Earth and Planetary Science Letters, v. 267, p. 426–434, doi: 10.1016/j.epsl.2007.12.001. van Geel, B., 2001, Non-pollen palynomorphs, in Smol, J.P., Birks, H.J.B, and Last W.M., eds., Tracking Environmental Change Using Lake Sediments, Volume 3: Terrestrial, Algal and Siliceous Indicators: Dordrecht, The Netherlands, Kluwer Academic Publishers, p. 99–119. van Geel, B., 2006, “Quaternary non-pollen palynomorphs” deserve our attention: Review of Palaeobotany and Palynology, v. 141, p. vii–viii, doi: 10.1016/j.revpalbo.2006.04.001.

115

van Geel, B., and Aptroot, A., 2006, Fossil ascomycetes in Quaternary deposits: Nova Hedwigia, v. 82, p. 313–329, doi: 10.1127/0029-5035/2006/0082-0313. van Waveren, I.M., 1994, Tintinnomorphs from deep-sea sediments of the Banda Sea (Indonesia): Scripta Geologica, v. 105, p. 27–51. Versteegh, G.J.M., and Blokker, P., 2004, Resistant macromolecules of extant and fossil microalgae: Phycological Research, v. 52, p. 325–339, doi: 10.1111/j.1440-1835.2004.tb00342.x. Walker, I.R., 2001, Midges: Chironomidae and related Diptera, in Smol, J.P., Birks, H.J.B., and Last, W.M., eds., Tracking Environmental Change Using Lake Sediments. Volume 4. Zoological Indicators: Dordrecht, Netherlands, Kluwer Academic Publishers, p. 43–66. Wall, D., and Dale, B., 1974, Dinoflagellates in late Quaternary deep-water sediments of Black Sea, in Degens, E.T., and Ross, D.A., eds., The Black Sea: Geology, Chemistry, and Biology: American Association of Petroleum Geologists Memoir 20, p. 364–380. Wall, D., Dale, B., and Harada, K., 1973, Description of new fossil dinoflagellates from the late Quaternary of the Black Sea: Micropaleontology, v. 19, no. 1, p. 18–31, doi: 10.2307/1484962. Wood, G.D., Gabriel, A.M., and Lawson, J.C., 1996, Palynological techniques—Processing and microscopy, in Jansonius, J., and McGregor, D.C., eds., Palynology: Principles and Applications: American Association of Stratigraphic Palynologists Foundation, v. 1, p. 29–50. Yanko-Hombach, V.V., 2007, Controversy over Noah’s Flood in the Black Sea: Geological and foraminiferal evidence from the shelf, in Yanko-Hombach, V., Gilbert, A.S., Panin, N., and Dolukhanov, P.M., eds., The Black Sea Flood Question: Changes in Coastline, Climate, and Human Settlement: Dordrecht, The Netherlands, Springer, p. 149–204. Zalessky, M.D., 1926, Sur les nouvelles algues découvertes dans les sapropélogène du Lac Beloë (Hauteurs de Valdai) et sur une algue sapropélogène, Botryococcus braunii Kützing: Revue Générale de Botanique, v. 38, p. 31–42. Zonneveld, K.A.F., Versteegh, G.J.M., and De Lange, G.J., 2001, Palaeoproductivity and post-depositional aerobic organic matter decay reflected by dinoflagellate cyst assemblages of the Eastern Mediterranean S1 sapropel: Marine Geology, v. 172, p. 181–195, doi: 10.1016/S0025-3227(00)00134-1. Zonneveld, K.A.F., Versteegh, G., and Kodrans-Nsiah, M., 2008, Preservation and organic chemistry of late Cenozoic organic-walled dinoflagellate cysts: A review: Marine Micropaleontology, v. 68, p. 179–197, doi: 10.1016/j.marmicro.2008.01.015.

MANUSCRIPT ACCEPTED BY THE SOCIETY 22 JUNE 2010

Printed in the USA

The Geological Society of America Special Paper 473 2011

Climatic and environmental oscillations in southeastern Ukraine from 30 to 10 ka, inferred from pollen and lithopedology Natalia P. Gerasimenko* Earth Sciences and Geomorphology Department, Taras Shevchenko National University of Kyiv, Glushkova 2, Kyiv, DSP 680, Ukraine

ABSTRACT Pollen and lithopedological data were obtained from Upper Paleolithic sites and Upper Pleistocene loess-soil sequences located between the Sea of Azov and the River Donets, and in the foothills of the Crimean Mountains. During the last Middle Pleniglacial interstadial (Upper Vytachiv soil, 30–27 ka), there existed boreal steppe (south-boreal forest-steppe in Crimea). During the Late Pleniglacial, two main phases of loess accumulation occurred, which were separated by the phase of initial pedogenesis. The loess accumulated under subperiglacial xeric steppe (particularly dry at 15–13 ka), and the incipient soils (Dofinivka unit, 18–15 ka) formed under boreal grassland. During the Late Glacial interstadials, there existed boreal and southboreal forest-steppe with a relatively wet climate (middle Prychernomorsk soil unit, the upper soil 11.8–11.4 ka). During the Younger Dryas, grassland reappeared under a dry and cool climate (10.9–10.5 ka). Paleoclimatic changes demonstrate the same pattern in both studied areas, and they correspond well with Black Sea transgressiveregressive cycles. Regional differences still existed—during all phases, the climate was the mildest in the western foothills of the Crimean Mountains, the coldest in the Donetsk Upland, and the driest near the Sea of Azov.

INTRODUCTION The Upper Pleistocene deposits of southern Ukraine have a distinct cyclical pattern of loess and paleosol alternation, which has been used in the Quaternary stratigraphic framework of Ukraine (Veklich, 1993). According to various studies (Shelkoplyas et al., 1986; Gerasimenko, 1999; Gozhik et al., 2000; Rousseau et al., 2001), the Vytachiv soil unit of the Ukrainian framework corresponds to the Middle Pleniglacial, marine iso-

topic stage 3, whereas the Bug and Prychernomorsk loess units, and the Dofinivka soil unit between them, are correlated with the Late Pleniglacial, MIS 2. In this paper, paleoenvironmental information is based on the results of pedostratigraphic and pollen investigation of the aforementioned late Pleistocene units and the Late Glacial deposits. The sections are located in two regions (Fig. 1): the area between the Sea of Azov and the River Donets, and the foothills of the Crimean Mountains. In the Azov-Donets area, several key

*[email protected]. Gerasimenko, N.P., 2011, Climatic and environmental oscillations in southeastern Ukraine from 30 to 10 ka, inferred from pollen and lithopedology, in Buynevich, I.V., Yanko-Hombach, V., Gilbert, A.S., and Martin, R.E., eds., Geology and Geoarchaeology of the Black Sea Region: Beyond the Flood Hypothesis: Geological Society of America Special Paper 473, p. 117–132, doi: 10.1130/2011.2473(08). For permission to copy, contact [email protected]. © 2011 The Geological Society of America. All rights reserved.

117

118

Gerasimenko

Pleistocene sites have previously been studied for lithopedology and palynology (Veklich et al., 1973; Artyushenko et al., 1973; Sirenko and Turlo, 1986; Gerasimenko and Pedanyuk, 1991), but, because of inefficient pollen retrieval techniques in previous times, percentage pollen diagrams were plotted only latterly. Data have also been obtained from Final Paleolithic sites (Gerasimenko, 1997a, 1997b; archaeology by Gorelik, 2001), and the

Upper Paleolithic site at Amvrosievka (Gerasimenko, 1997b; archaeology by A. Krotova, 1996). In the Crimean Mountains, after the first investigation of the Zaskal’naya sites (Velichko, 1988; Gubonina, 1985), a multidisciplinary paleoenvironmental study of the Crimean Paleolithic was done (Marks and Chabai, 1998; Chabai and Monigal, 1999; Chabai et al., 2004), and this included investigation

Figure 1. Location map of the sections, discussed in the paper. 1—Kabazi II, 2—Skalisty rock shelter, 3—Buran-Kaya III, 4—Novotroitsk, 5—Amvrosievka, 6—Novoraysk, 7—Rogalik II and XII.

Climatic and environmental oscillations in southeastern Ukraine of small mammals, malacofauna, large mammals (Burke et al., 2004; Markova, 2004; Patou-Mathis, 2004), palynology (Gerasimenko, 1999, 2004, 2007), and absolute dating (Rink et al., 1998; McKinney, 1998, Pettit, 1998). The Late Glacial stratigraphy and paleoenvironments have also been revealed from the Crimean Final Paleolithic sites (Cohen et al., 1996; Yanevich et al., 1996, Gerasimenko, 2004). ENVIRONMENTAL SETTING The Azov-Donets area includes the Pryazov and Donets Uplands (200–350 m above sea level [asl]), coastal plains of the Sea of Azov, and the Donets River alluvial plain (Fig. 1). The uplands are dissected by valleys and gullies to a depth of 100– 200 m. Slopes of the uplands and the plains have a continuous loess cover. The Novoraysk and Novotroitsk sites are sections of low plateaus (180–200 m asl), and Amvrosievka is located in the bottom of a paleogully. Rogalik II and XII are situated on a terrace slope at the lower level of the Donets alluvial plain (70–90 m asl). The Azov-Donets area belongs to the northern subzone of the Ukrainian steppe. From north to south, the average January and July temperatures are–7.5 °C, +22 °C (Rogalik), –7 °C, +21 °C (Novoraysk), –6.5 °C, +22 °C (Novotroitsk), and –6 °C, +22.5 °C (Amvrosievka). The annual precipitation in the Donetsk Upland (Novoraysk and Amvrosievka) is higher than in the Pryazov Upland and the Donets Plain (Novotroitsk and Rogalik): 500–550 and 450–470 mm, respectively. Chernozems (Mollisol), which dominate the soil cover, have a higher humus content and thicker humus layer in the northern part of the area. The steppe coenoses consist of grasses (Stipa, Festuca) and mesophytic herbs (Herbetum mixtum). Arboreal vegetation grows in gullies: oak-ash forest (Quercus robur and Fraxinus excelsior) in the Donets Upland, and scrub in the Pryazov Upland (Malus silvestris, Pyrus communis, Crataegus monogyna, Prunus spinosa, and Ulmus campestris). The highest part of the Donets Upland is regarded as a forest-steppe because oak-maple-lime forest on gray forest soils partially spreads on watersheds here. Pine forest occupies sands on the river terraces. The sections of the Crimean Mountains studied in this paper are located on the slopes of a cuesta, formed in CretaceousPaleogene limestones, at 240–315 m asl. The slope and the sites have a southern exposure. In western Crimea, the Middle Paleolithic open-area site of Kabazi II is situated in the Alma riverbank, 90 m above the water level, and the Late–Final Paleolithic site, Skalisty rock shelter, is 20 m above the water level in the Bodrak River bank. The thick sequence at Kabazi II was accumulated in a sedimentation trap behind a rock resting on the slope bench. In eastern Crimea, the Middle Paleolithic–Neolithic site Buran-Kaya III is a rock shelter in the Burulcha River bank, 10 m above the water level. All sites are located in the low-mountain forest-steppe. The plateau-like tops of the cuesta are covered by meadow-steppe on chernozems. The lower parts of the slopes are occupied by

119

arboreal vegetation, the main components of which are oak (Quercus pubescens), hornbeam (Carpinus orientalis), hazelnut (Corylus avellana), and bloody dogwood (Cornus sanguinea). The average January and July temperatures are, respectively, +0.3 °C and +21.5 °C (Kabazi II and Skalisty rock shelter) and –1 °C and +22 °C (Buran-Kaya III). Precipitation in eastern Crimea is much lower (400–450 mm) than in western Crimea (500–550 mm), so forests are thinner, and steppe is drier in the east. The lower forest belt of the Main Mountain Ridge consists of oak and hornbeam, whereas beech and hornbeam grow in the second belt. Pine forests occupy the highest parts of the mountain slopes. METHODS For lithopedological study, humus, CaCO3, and dry salt content, absorbing capacity, and results of bulk chemical and grainsize analyses are interpreted on the basis of the paleopedological approach of Veklich et al. (1979). In order to obtain sufficient pollen counts from subaerial deposits, the following technique was used for sample processing: treatment with 10% HCl, disintegration in a solution of 15% Na4P2O7, treatment with HCl and 10% KOH, cold treatment with HF, and double separation in a heavy solution of CdI2 and KI (specific gravity 2.0 and 2.2). The pollen counts varied between 100 and 500 grains per sample, and pollen was mostly well preserved, particularly in the sections from depressions. Pollen preservation did not depend on carbonate (or gypsum) content in deposits. No palynotypes of Neogene (or older) plants were traced, and in loesses, no pollen of broad-leaved species occurred. This may indicate the absence of pollen redeposition. The transfer functions of vegetation and palynospectra, based on surface samples from different ecosystems, including mountain ones (Grichuk and Zaklinskaya, 1948; Arap, 1976; Dinesman, 1977; Klopotovskaya, 1976; Bezus’ko et al., 1997), were used in the interpretation of pollen data. Pollen percentages were counted from the total sum of microfossils, with the exception of the Buran-Kaya III site (spores were strongly over-represented there, and the percentages of arboreal and nonarboreal taxa were plotted from the arboreal pollen (AP) and nonarboreal pollen (NAP) sum). Pollen of dicotyledonous herbs (with the exception of the Chenopodiaceae and Asteraceae families) is grouped as “Herbetum mixtum”—an indicator of mesophytic steppe type (Grichuk and Zaklinskaya, 1948). In the following sections, the identification of loess-soil units has been primarily made on the basis of lithopedostratigraphy described in the Quaternary framework of Ukraine (Veklich, 1993). The geochronology in loess-soil sections is rather poor and mainly based on 14C data using bones from the archaeological sites. The obtained thermoluminescence (TL) dates, archaeological data, and paleomagnetic age estimations have also been used for the correlation. The geochronological information is summarized in Table 1.

120

Gerasimenko TABLE 1. THE GEOCHRONOLOGICAL DATA FOR THE SITES STUDIED

Site

Stratigraphic Archaeological Material Method of Age unit horizon dating (yr B.P.) Buran-Kaya III pc3 4A Bone AMS 10,580 ± 60 Buran-Kaya III pc3 4A Bone AMS 10,920 ± 65 B1 Bone AMS 28,840 ± 460 Buran-Kaya III vt3 B1 Bone AMS 28,520 ± 460 Buran-Kaya III vt3 C Bone AMS 32,350 ± 700 Buran-Kaya III vt2 C Bone AMS 32,200 ± 650 Buran-Kaya III vt2 Buran-Kaya III vt2 C Bone AMS 36,700 ± 1,500 II/1A Tooth ESR 30,000 ± 2,000 Kabazi II vt3b II/1 Bone AMS 31,550 ± 600 Kabazi II vt2 II/2 Bone AMS 35,100 ± 850 Kabazi II vt2 II/4 Bone AMS 32,200 ± 900 Kabazi II vt2 II/5 Bone AMS 33,400 ± 1,000 Kabazi II vt2 II/7AB Tooth ESR 36,000 ± 3,000 Kabazi II vt1c Kabazi II vt1c II/7AB Tooth ESR 38,000 ± 4,000 Novotroitsk vt1b Paleosol TL 57,500 ± 3,000 Novoraysk ud Loess TL 75,000 ± 4,000 III/2 Bone AMS 11,680 ± 110 Rock shelter Skalisty pc2 III/3 Bone AMS 11,750 ± 120 Rock shelter Skalisty pc2 IV Bone AMS 14,570 ± 140 Rock shelter Skalisty df3 Rock shelter Skalisty df3 V Bone AMS 15,550 ± 310 VI Bone AMS 15,030 ± 150 Rock shelter Skalisty df1 Rock shelter Skalisty df1 VII Bone AMS 18,300 ± 220 14 I Bone C 11,400 ± 140 Rogalik XII pc2 I Loess TL 13,500 ± 2,000 Rogalik II pc2 I Paleosol TL 13,000 ± 2,000 Rogalik II pc2 Rogalik II pc2 I Paleosol TL 13,500 ± 1,500 Rogalik II df Paleosol TL 15,500 ± 3,000 Rogalik II df Paleosol TL 17,000 ± 3,000 Peredel’sk bg Loess TL 25,500 ± 2,000 Amvrosiivka df I Bone AMS 18,220 ± 200 Amvrosiivka df I Bone AMS 18,620 ± 220 Amvrosiivka df I Bone AMS 18,700 ± 240 Amvrosiivka df I Bone AMS 18,860 ± 220 Note: AMS—accelerator mass spectrometry; ESR—electron spin resonance; TL—thermoluminescence.

PEDOSTRATIGRAPHY AND PALYNOLOGY Vytachiv (vt) Unit Depending on paleorelief, the Vytachiv unit is represented either by polygenetic “welded” soil or by a pedocomplex of three soils, correlated with the three Middle Pleniglacial interstadials (Gerasimenko, 1999, 2001). The upper paleosol vt3, formed after 30 ka, is examined in this paper (Figs. 2 and 3). This soil is frequently separated from the lower Vytachiv soils by loesslike loam vt2, dated between 36.0 and 31.5 ka, which has low counts of AP (and no pollen of broad-leaved taxa) and high NAP values (Figs. 3A and 3B). The vt2 material is distinguished within the BCca horizon of vt3 soil by much lower contents of R2O3 (Al2O3 and Fe2O3), clay particles, and humus than in the whole Vytachiv pedocomplex (Figs. 2A and 2C). In the DonetsAzov area, vt3 soils are thin, brown-colored, not rich in humus, but they are strongly calcified. The content of humus decreases, whereas contents of CaCO3 and dry salts increase to the south. In the north, water-salt residue in the soil is not considerable and is dominated by Ca-Mg hydrocarbonates (Fig. 2A), whereas in the south, it increases strongly (0.49%–0.60%) and consists of

Reference Chabai et al. (2004) Chabai et al. (2004) Chabai et al. (2004) Chabai et al. (2004) Chabai et al. (2004) Chabai et al. (2004) Chabai et al. (2004) Chabai (2004) Chabai and Monigal (1999) Chabai and Monigal (1999) Chabai and Monigal (1999) Chabai and Monigal (1999) Chabai (2004) Chabai (2004) Gerasimenko and Pedanyuk (1991) Gerasimenko and Pedanyuk (1991) Cohen et al. (1996) Cohen et al. (1996) Cohen et al. (1996) Cohen et al. (1996) Cohen et al. (1996) Cohen et al. (1996) Gorelik (2001) Gerasimenko (1997b) Gerasimenko (1997b) Gerasimenko (1997b) Gerasimenko (1997b) Gerasimenko (1997b) Gerasimenko (1997b) Krotova (1996) Krotova (1996) Krotova (1996) Krotova (1996)

Na-Ca-Mg sulfates. Absorbed Na becomes significant (Fig. 2C), and small gypsum crystals appear. By morphology and properties, vt3 soils are similar to Haplic Kashtanozem in the north and Gypsic Kashtanozem in the south. In the Crimean Mountains, the vt3 unit formed with an admixture of limestone colluvium. It is a brown rendzina (Eutric Leptosol), consisting of a graybrown humic A1 horizon and a light-brown transitional B horizon (Fig. 3), or its pedosediment. Pollen spectra of vt3 soil are dominated by NAP. The AP percentages are higher (23%–44%) in the Donets area and in western Crimea (Figs. 2B and 3A), and lower (4%–13%) in the Pryazov area and in eastern Crimea (Figs. 2D and 3B). Pinus dominates

Figure 2. Sites of the Azov-Donets area (the upper parts of the profiles): (A) Novoraysky quarry, pedolithology; (B) Novoraysky quarry, pollen diagram; (C) Novotroitsk quarry, pedolithology; (D) Novotroitsk quarry, pollen diagram. Legend for Figures 2–7. 1—Chernozem or A1 (humus) soil horizon; 2—Bt soil horizon; 3—Eutric Leptosol (brown rendzina) and its derivatives; 4—Cambisol and its derivatives; 5—Dystric Cambisol; 6—Kashtanozem; 7—Haplic Calcisol; 8—incipient soil; 9—nonsoil loam; 10—loamy pedosediment; 11— sandy pedosediment; 12—loess; 13—loess-like loam; 14—ashy layer.

Climatic and environmental oscillations in southeastern Ukraine

121

A

B

C

D

Archaeological

B

Figure 3. Pollen diagrams of the Crimean sites: (A) Kabazi II, the upper part of the section (modified from Gerasimenko, 1999); and (B) Buran-Kaya III, the upper part of the section (modified from Gerasimenko, 2004).

A

122 Gerasimenko

Climatic and environmental oscillations in southeastern Ukraine the AP, with the exception of the Buran Kaya III riverine site, in which Alnus prevails. In western Crimea, pollen counts of broadleaved species (Quercus and Carpinus) reach 4%–14%, but no pollen of broad-leaved trees occurs in the Pryazov Plateau, and only a few grains of Quercus or Tilia appear in the other sites. A small amount of Betula pollen is traced in all sections, and single pollen grains of bushes (Corylus, Cornaceae, Elaeagnaceae) are detected in the Donets Upland. Herbetum mixtum dominates the NAP in Crimea, and it shares dominance with Chenopodiaceae (or Asteraceae) in the Donets-Pryazov area. Poaceae and Cyperaceae pollen occur in all sections (Cyperaceae counts are high in Crimea), whereas Artemisia and Chenopodiaceae values are very low (with the exception of Chenopodiaceae in Novoraysk). Ephedra pollen occurs in Crimea. Spore percentages (Bryales and Polypodiaceae) are the highest in the Buran-Kaya III rock shelter and the lowest in the Pryazov Plateau. The top layer of the soil (subunit vt3c) shows a drop in AP pollen percentages and an increase in pollen of xerophytes. Bug (bg) Unit

stone debris, in Crimea. Bug loess has a lower content of humus and clay and, thus, much lower absorbing capacity than the Vytachiv soil. On the contrary, the loess is richer in carbonates and poorer in dry salts, and has higher contents of SiO2 and lower contents of R2O3 than the Vytachiv soil (Figs. 2A and 2C). NAP strongly dominates in the Bug unit (Figs. 2B, 2D, 3A, 3B, and 4). AP incidence falls to 5%–7% (in western Crimea, not lower than 13%–14%) and mostly consists of Pinus. Betula pollen occurs in small numbers, and few grains of Alnus and Juniperus are present in Crimea (Elaeagnaceae and Rhamnaceae at Kabazi II). The NAP is dominated by xerophytes (Chenopodiaceae and Artemisia), and Asteraceae pollen is seen in all sites with the exception of western Crimea. There, pollen of Herbetum mixtum, Poaceae, and Cyperaceae are also significant (Fig. 3A). In the northern Novoraysk site and the Buran-Kaya III rock shelter, Lycopodiaceae and Botrychium boreale spores appear. Dofinivka (df) Unit This unit consists of two to three weakly developed soils, separated by thin loess beds, or of one “welded” soil on the plateau. In the Donets-Pryazov area, the latter (0.5–1.0 m thick) has light-brown color, low humus content, and a very high amount of CaCO3 (Figs. 2A and 2C), frequently in a form of soft carbonate

Archaeological horizon

The Bug unit is a typical loess (1–2 m thick) in the DonetskPryazov area, and a loess-like loam, with a lot of angular lime-

123

Figure 4. Pollen occurrence in the Skalisty rock shelter site.

124

Gerasimenko

nodules. At the soil’s lower limit, thin “tongues” indicate former desiccation fissures. The soil does not differ from loess on the basis of its clay contents, but is richer in R2O3, has a higher absorbing capacity (at the expense of absorbed sodium), and a slightly higher amount of dry salts (including sodium sulfates). In southern Ukraine, Dofinivka soil is justly assigned to a brown semidesert type (Veklich et al., 1973), i.e., Haplic Calcisol. In paleogullies, similar soils developed in sandy loams and, thus, are poor in clay and R2O3, but are still calcified (Fig. 5A). In the Crimean Mountains, the Dofinivka unit includes two to three thin beds of brown and gray pedosediments, separated by loess-like loams (Figs. 3B and 4). AP counts in the unit are the lowest (2%–13%) in the Pryazov Upland and eastern Crimea (Figs. 2D and 3B). In western Crimea, they are 25%–55% (Fig. 4), and, in the Donetsk area, increase from the plateau, 17%–31% (Fig. 2B), to paleogullies, 23%–62% (Figs. 2D, 5A, and 6). The lower subunit df1 and the lower brown horizon of the welded soil are richer in AP than the upper subunit df3 and the A1 horizon of the welded soil. Pinus dominates the AP, particularly in the plateau sites. In depressions, the AP is more diverse, especially in subunit df1. In the Donetsk area, Betula pollen is significant, and grains of Betula Nanae et Fruticosae are present in the paleogully (Fig. 6). A few Quercus, Tilia, Corylus, and Picea pollen grains also occur there in subunit df1. In eastern Crimea, pollen of Juniperus, Salix, Betula, and Alnus are represented, and in western Crimea, single grains of diverse broad-leaved taxa appear (only in subunit df1). The NAP composition differs between the sections of plateaus and depressions: in the former, pollen of Asteraceae and Artemisia (or Chenopodiaceae) dominate (Figs. 2B and 2D), and, in the latter, Herbetum mixtum strongly prevails over xerophytes, and Artemisia counts are very low. Spores are not abundant and include Bryales, Polypodiaceae, and, in depressions, Lycopodiaceae.

totype area near the Black Sea, the subunit includes two brown carbonate soils (Veklich, 1993). In the Donetsk area, pc2 soils are leached of CaCO3 (Gerasimenko, 1997a, 1997b). The upper soil is a Dystric Cambisol with dark-gray A1 horizon, brightbrown prismatic B horizon, low CaO content, and signs of clay and R2O3 translocation (Fig. 5A). AP counts are higher in this soil than in the lower one or the thin loess bed between them (Fig. 5B). Pinus and Betula sect. Albae dominate the AP, and Herbetum mixtum prevails in the NAP of the soils. Few Rhamnaceae and Elaeagnaceae pollen grains occur in the lower soil. In the upper one, Alnus, Corylus, Malaceae, and few Picea also appear, and single pollen grains of Quercus have been detected (Gerasimenko, 1997b). In the loess between these soils, the AP is dominated by Betula sect. Nanae et Fruticosae, and pollen counts of xerophytes increase (particularly of Artemisia and Ephedra). In western Crimea, the two beds of pc2 brown pedosediments are also separated by a thin light loam (Fig. 4). Pollen in the pedosediments is small in number, but rich in composition, and includes few grains of Quercus, Tilia, and Corylus in the lower bed, and also Carpinus, Fagus, and Ulmus in the upper one. The thin pc3 loess has pollen spectra dominated by NAP (Figs. 2B and 4). Betula sect. Nanae et Fruticosae prevails in the AP of the Donetsk area (Gerasimenko, 1997b), and no pollen of broad-leaved trees is present in Crimea (only a few Corylus and Rhamnaceae in western Crimea). In the pc3 subunit, pollen counts of xerophytes are much lower than in the pc1 loess and in the loess between the pc2 soils (the Donetsk sites). Pollen percentages of Poaceae and Herbetum mixtum become significant instead, and in Crimea, spores of Polypodiaceae increase in number, and Botrychium boreale almost disappear (Fig. 3B).

Prychernomorsk (pc) Unit

Vytachiv (vt) Unit

The unit consists of two loess subunits (pc1 and pc3), separated by the middle Prychernomorsk (pc2) incipient soils (Veklich, 1993). Because of Holocene erosion, frequently only the thin loess pc1 is traceable. It is somewhat enriched in humus (an impact of Holocene pedogenesis), but its clay content and absorbing capacity are lower than in the soils (Figs. 2A and 2C). The loess is rich in CaCO3 (13%–16%), but not in dry salts. The NAP domination in the pc1 subunit is large or even absolute (Figs. 2B, 2D, and 3B). In the Donets-Pryazov area, pollen of Betula sect. Nanae et Fruticosae is revealed in the loess (Fig. 6). In the Crimean foothills, few grains of Pinus, Juniperus, Betula, and Alnus occur only at the bottom of this subunit. The NAP is dominated by xerophytes in all sites (with the exception of paleogullies, Fig. 6), and an increase in Artemisia pollen is characteristic. Lycopodiaceae and Botrychium boreale spores appear in the Buran-Kaya III rock shelter. The pc2 subunit is dated to the Late Glacial interstadials (Gerasimenko, 1997a, 1997b; Gozhik et al., 2000). In the stra-

The upper soil of the Vytachiv pedocomplex (vt3), which yielded 14C dates 28,800 ± 500 and 28,500 ± 500 yr B.P., electron spin resonance date 30.0 ± 2.0 ka (Figs. 3A and 3B), is correlated with the Denekamp interstadial of Western Europe. The preceding interval, vt2, dated between 31,500 ± 600 and 36,700 ± 1500 yr B.P., had a dry and cold climate: loess-like deposits with low indices of clay weathering formed under grassland both in the eastern Ukraine and in the foothills of the Crimean Mountains. Xerophytic herbs were frequent on the steppe, and only boreal trees occurred in valleys (mainly pine; in the Crimean foothills, also alder and birch). Presently, birch does not grow in Crimea (only sporadically on the highest ridges). Spore plants of boreal climate (Lycopodiaceae and Botrychium boreale) grew in rock shelters of eastern Crimea. Such an environment can be compared with the stadial established for the corresponding time period in Western Europe (Van der Hammen, 1995), in central Ukraine (Gerasimenko, 2001), and in western Ukraine (Bolikhovskaya, 1986, 1995; Haesaerts et al., 2003).

PALEOENVIRONMENTS AND CORRELATION

Figure 5. The Rogalik II site (modified from Gerasimenko, 1997b): (A) pedolithology; and (B) pollen diagram.

B

A

Climatic and environmental oscillations in southeastern Ukraine 125

Archaeological age

Figure 6. Pollen diagram of the lower part of the Amvrosievka site (modified from Gerasimenko, 1997b).

126 Gerasimenko

Climatic and environmental oscillations in southeastern Ukraine During the interstadial vt3, the soils formed under a warmer climate (higher indices of clay weathering). In eastern Ukraine, the spread of Kashtanozems with carbonates high in the profile indicates that the climate still was dry, and its aridity increased to the south, where gypsum and salts intensely accumulated in the soils. Pollen of the Asteraceae family, dominating in the site of Pryazov Upland, includes many palynotypes of Erigeron acer and Helichrysum arenarium (species that can grow on salt soils). Dry grasslands with scattered Elaeagnus bushes occupied the Pryazov area. In the steppe of the Donets area, valleys and gullies included patches of pine forest with some birch, and, in betterprotected places, a few oak, lime, and hazel. In the foothills of the Crimean Mountains, brown rendzina soils (Eutric Leptosols) formed, and broad-leaved species existed in the woodlands, particularly in western Crimea. There, mixed forest, which included oak and hornbeam, expanded much more extensively than in eastern Crimea, where mesophytic steppes strongly dominated, and only a few Quercus pubescens grew in better-protected places. The retreat of boreal trees (particularly of Betula) from the forest and of boreal spore plants (Lycopodiaceae and Botrychium) from rock shelters occurred. Xerophytes almost disappeared from the herbal cover, particularly in western Crimea. This indicates significant climatic humidity. At the end of the interstadial, an extensive advance of steppe and intense suppression of broad-leaved trees marked a transition to the next cold stage. The 14C dates from the top soil of the Vytachiv alluvial suite of the Dnieper terrace are between 28 and 27 ka (Stepanchuk et al., 2004), and these fit with the end of the last Middle Pleniglacial interstadial. The environment in the lower Dnieper valley was cool and similar to the transitional phase detected from the top of soil vt3 at Kabazi II. In the vt3 soil, cultural horizons II/1A and B1-B (Figs. 3A and 3B) include Mousterian artifacts, whereas the underlying vt2 loess, horizon C (Fig. 3B), contains Upper Paleolithic material. It proves a coexistence of Middle and Upper Paleolithic cultures in Crimea and shows that the Crimean Middle Paleolithic lasted considerably longer than in many other areas (Chabai, 2004). Bug (bg) Unit In the study area, this unit yields a range of ages between 27,080 ± 400 and 18,300 ± 220 14C yr B.P. It also yielded a thermoluminescence (TL) date of 25.5 ± 2.0 ka in its lower part (Gerasimenko, 1997b). This enables correlation of the Bug unit with the first half of the Late Pleniglacial. The Bug unit was marked by the strongest loess accumulation of the Pleniglacial. Loess derivatives covered the lower slopes of the Crimean Mountains. The low contents of clay and wide SiO2:R2O2 ratios in the loess are evidence of a drastic decrease in clay weathering, which together with intense accumulation of carbonates and depletion of humus formation indicate a dry and cold climate. This is also confirmed by an increase in input of large angular debris on mountain slopes.

127

Pollen data indicate that southern Ukraine was covered by steppe. In the Pryazov Upland, grassland included a high proportion of xerophytic species of Asteraceae. In the Donets Upland and in Crimea, at the beginning of Bug times, the steppe communities included mesophytic herbs. Nevertheless, in Crimea, they were considerably drier than the meadow steppe, which existed at the end of Vytachiv time. During the second half of Bug time, depletion and decrease in the variety of both mesophytic and hygrophytic herbs occurred. Grassland with strong xerophytic components (including Artemisia) was established then, and trees completely abandoned eastern Crimea. This indicates a rising aridity through Bug times—a trend that is also detected in the other areas of Ukraine (Gerasimenko, 2006). Broad-leaved trees disappeared in Bug times, as did Polypodiaceae ferns in western Crimea. Cold-resistant Lycopodiaceae and Botrychium boreale became more abundant in mountain rock shelters and in gullies of the Donetsk area. This proves the existence of a cold (boreal to subperiglacial) climate. Some pine and birch grew in the dissected Donetsk Upland, but the Pryazov Upland was treeless. The Crimean rivers were framed by alder in places. Pine and juniper (in western Crimea also Elaeagnaceae and drought-resistant species of Rhamnaceae) occurred on slopes. The climate was wetter in the foothills of Crimea (and particularly in western Crimea) than in the plains of southern Ukraine. The strong climatic deterioration in Bug times corresponds closely to the harsh climate of the first half of the Late Pleniglacial. Dofinivka (df) Unit The Dofinivka unit has yielded 14C dates in the range 14.0–17.1 ka in its stratotype region (Gozhik et al., 2000). In the studied area, the Dofinivka soils were 14C dated at 14,570 ± 140, 15,550 ± 310, 15,030 ± 150, and 18,300 ± 220 yr B.P. (Cohen et al., 1996), 18,220–18,860 yr B.P. (Krotova, 1996), and TL dated at 15.5 ± 3.0 and 17.0 ± 3.0 ka (Gerasimenko, 1997b). The complexity of the Dofinivka unit shows that it was formed as pulses of incipient soil formation and loess accumulation alternated. In the Azov-Donets area, weak weathering indices and salt-carbonate accumulation in the Dofinivka soils indicate that the climate was strongly continental. The plateaus were covered by dry steppe. Nevertheless, in gullies, the conditions were wetter, and birch-pine woodland spread much wider than during loess accumulation. During df1 times, deep gullies provided a habitat for all mesophytic plants, including arctoboreal Betula sect. Nanae et Fruticosae and a few broad-leaved trees (oak, lime, and hazel). Ecologically different species have been shown to coexist during the late Pleistocene interstadials in other East European sites (Bolikhovskaya, 1986, 1995). During df3 times, the climate became harsher: broad-leaved species disappeared, the role of herbal xerophytes increased, and reduction of woodland occurred (though not as dramatically as during the units of loess accumulation).

128

Gerasimenko

In the foothills of the Crimean Mountains, the climate of Dofinivka times was much wetter than that of Bug times—foreststeppe was established in the west and meadow-steppe in the east of the area. In eastern Crimea, arboreal vegetation was limited to river valleys, where some alder, willow, birch, pine, and juniper grew. Incipient soils, formed under mesophytic steppe, were richer in humus than the soils of the drier Donets-Pryazov area. In western Crimea, during the climatic optimum df1, woodland included an admixture of diverse broad-leaved trees (beech, oak, elm, lime, and maple), which obviously started to spread from refugia. The presence of beech can be seen as evidence that mountain forest belts were in a lower position than at present. In modern surface samples from the Crimean Mountains (Artyushenko and Mishnev, 1978), pollen proportions of broad-leaved trees are much higher than in df1 soils. Thus, even during the optimum, the climate was cooler than present day, and an interstadial environment existed. After df1 times, significant reduction of forest happened and broad-leaved trees retreated. During df3 times, birch-pine woodland included alder and diverse shrubs (Corylus, Sambucus, Rhamnaceae, and Malaceae). During df2 times, pine strongly dominated (only a few birch, juniper, and Rhamnaceae occurred), and thin loess accumulated, indicating a climate similar to that of a stadial, which was colder and drier than the climates of the df3 and df1 soil subunits. Of the last two, df1 times were warmer and wetter, and that was also the case in the Azov-Donets area. A climatic amelioration during the Dofinivka interstadial is reflected in distinctive vegetational changes only in the wetter areas and localities (Crimea and paleogullies). Nevertheless, soil development during Dofinivka times indicates the stability of the sedimentary environment, controlled by a decrease in loess storms, and, thus, by a regional decrease in climatic aridity. The poorly developed Dofinivka soils might be analogues of the set of incipient soils in western Ukraine and Molodova that were 14 C dated between 19.4 and 17.2 ka (Haesaerts et al., 2003). The Dofinivka unit can also be correlated with the Plyussky interstadial of central Russia (16.5–15.0 ka). Cryophytic species of clubmosses and ferns completely disappeared then from the boreal plant communities (Bolikhovskaya, 1995). This was also the case for the rock shelter vegetation in Crimea during Dofinivka times. Prychernomorsk (pc) Unit Loess of the earliest Prychernomorsk subunit pc1 was C dated in the range between 14.0 and 11.9 ka in its stratotype region (Gozhik et al., 2000), and, in the studied area, it fits between 14C 15,500–14,570 and 13,500–11,400 yr B.P. The main phase of loess accumulation occurred during the early Prychernomorsk time. In the Donets-Pryazov area, xerophytic communities of Artemisia, Chenopodiaceae, and Asteraceae predominated, making the landscape similar to a semidesert. Dry ArtemisiaPoaceae steppe occupied eastern Crimea. Sparse arboreal vegetation was restricted to gullies and valleys, and none existed during some time slices. In western Crimea, only drought-resistant 14

pine and juniper formed woodland. All of this indicates an extradry climate. Absence of broad-leaved trees and appearance of arcto-boreal species of Lycopodiaceae and Botrychium boreale in rock shelters are evidence that it was cold even in Crimea. Shrub birches (elements of periglacial vegetation) started to grow abundantly in gullies of the Donetsk Upland. In central Russia, the driest spell of the late Valdai, marked by a strong increase in Artemisia pollen, is related to the Luzhsky stadial, 15–13 ka (Bolikhovskaya, 1995). The pc2 subunit is dated to the Alleröd in its stratotype region near Odessa (Gozhik et al., 2000) and in Crimea (Fig. 4), where it includes the Final Paleolithic cultural layer (Cohen et al., 1996). In many profiles of the Rogalik and Peredel’sk sites (the Donets area), the pc2 pedocomplex consists of two soils, separated by a thin loess (Gerasimenko, 1997b). The upper soil (14C 11,400 ± 140 yr B.P.) corresponds to the Alleröd, the lower soil is marked by a paleomagnetic event, assigned to 13–12 k.y. B.P. (Vigilyanskaya, 1999), the loess includes the Final Paleolithic layer of 10,000 to 9000 B.C. (Gorelik, 2001), and the lower part of the subunit yielded undifferentiated TL dates 13.5 ± 1.5 and 13.5 ± 2.0 ka. Subunit pc2 is correlated with the Late Glacial. During the lower soil formation, birch-pine forest alternated with mesophytic steppe. This marks a sharp environmental change from the dry Late Pleniglacial climate to the interstadial environment, which was cooler but wetter than present day. This interstadial is correlated with the Bölling. In western Ukraine, forest-steppe vegetation of the first Late Glacial interstadial, dated at 12,300 ± 140 to 11,900 ± 230 yr B.P., already included a few broad-leaved species (Ivanova, 1987). The overlying loess bed was formed under typical grassland with considerable share of xerophytes (Artemisia, Ephedra). Arboreal vegetation almost completely disappeared (only a few shrub birches occurred). This stadial, correlated with the Older Dryas, was much colder and drier than the Bölling, but less harsh than the Late Pleniglacial. During the Alleröd interstadial, forest-steppe was reestablished in the Donetsk area. Woody species spread more extensively, and steppe communities included more mesophytic herbs than during the Bölling. Arboreal Rosaceae and Rhamnaceae were abundant. The existence of a “bush steppe” might be indicated, which is typical for the wetter steppe areas. The other evidence of humidity is a small but frequent presence of spruce. Spread of Picea is a typical feature of the Alleröd interstadial in the central part of Eastern Europe (Spiridonova, 1991). Few oaks and hazelnuts appeared in pine-alder-birch forest at the interstadial optimum, after 11,400 ± 140 yr B.P. (Gerasimenko, 1997a). This indicates that the Alleröd was warmer and wetter than the Bölling. Low ridges of the Crimean Mountains were occupied by forest and meadows. At 11,750 ± 120 to 11,680 ± 110 yr B.P. (Cohen et al., 1996), pine, birch, alder, oak, lime, and hazelnut grew in the forest. Later on, its composition became richer and included also hornbeam, beech, and elm. For the whole area, there is evidence that the Alleröd was cooler and wetter than the present time. Boreal plants, including wetloving spruce and alder, penetrated much farther south than

Climatic and environmental oscillations in southeastern Ukraine present day, and the role of broad-leaved trees was lower (particularly in the east). The subunit pc3 corresponds to the Younger Dryas. It was the last interval characterized by the presence of arcto-boreal plants (Betula nana) within depressions of the Donets-Pryazov area. Accumulation of thin loess layers occurred there (Gerasimenko, 1997b). In eastern Crimea, the corresponding layer (14C 10,920 ± 65, 10,580 ± 60 yr B.P.) is marked by the Swiderian culture (Yanevich et al., 1996). Steppe zone occupied the studied area, with the exception of western Crimea. In the north, mesophytic herbs were rather significant, particularly in depressions. Arboreal associations included sparse pine, alder, and arboreal birch. Eastern Crimea was occupied by grassland with high participation of xerophytes, though in river valleys, sparse tree stands from boreal species occurred. In western Crimea, boreal forest-steppe existed (with few Corylus and Frangula). The Younger Dryas was wetter and less harsh than the Older Dryas. CONCLUSIONS In southeastern Ukraine, the time span between 30 and 10 ka generally corresponds to the main period of loess accumulation in the late Pleistocene. Nevertheless, paleopedological and pollen data indicate climatic and environmental oscillations, the alternation of stadials and interstadials, during this time. Its beginning corresponds to the last interstadial of the Vytachiv unit, dated to 30–28 ka, which is the last Middle Pleniglacial (middle Valdai) interstadial. The first half of late Pleniglacial corresponds to Bug times, during which the most loess accumulation occurred (28–18 ka). The second half of the Late Pleniglacial included the two weak Dofinivka interstadials (between 18 and 14.5 ka), separated by a short stadial, and the early Prychernomorsk stadial pc1 (14.5–13 ka). The middle Prychernomorsk time pc2 corresponds to the two Late Glacial interstadials, separated by a short stadial. The late Prychernomorsk stadial pc3 (10.9–10.5 ka) is correlated with the Younger Dryas. The interstadials are represented by paleosols that are less developed than the modern soils. Their pollen is marked by a larger proportion of arboreal taxa and mesophytic herbs than the sediments from stadials. A characteristic feature of interstadial deposits in southeastern Ukraine is the consistent occurrence of a small number of pollen from broad-leaved species in all the sections studied. This might show that broad-leaved trees existed in the gully refugia and increased in number during the interstadials, which had a southern-boreal climate. In the northern Donetsk area, the stadial deposits (mainly loesses) include palynotypes of arcto-boreal shrub (Betula sect. Nanae et Fruticosae), and boreal species of Lycopodiaceae and Botrychium boreale, alongside pollen of xerophytic herbs. This indicates a subperiglacial climate during the loess formation. A combination of dry steppe communities with abundant arcto-boreal plants is typical for the periglacial area of the East European Plain (Bolikhovskaya, 1986, 1995). In the rock shelters of Crimea, an increase in spores of boreal Lycopodiaceae and

129

Botrychium boreale, but absence of arboreal cryophytes in the stadial deposits, gives evidence of a northern-boreal climate. Climatic and environmental dynamics in the Donets-Azov area and the foothills of the Crimean Mountains are schematically reconstructed in Figure 7. In Crimea, the interstadials had a southern-boreal climate, with the exception of the interstadial df3, which had a boreal climate. Forest-steppe existed in western Crimea, and mesophytic steppe spread in eastern Crimea. In the Donets-Pryazov area, the interstadials had a boreal climate, though the role of broad-leaved trees evidently increased in refugia. The climate of the interstadial df3 was northern-boreal (broad-leaved trees retreated, and a few shrub birch appeared). During the Late Pleniglacial interstadials, grassland dominated the landscape, and woody plants were restricted to gullies. During the Late Glacial interstadials, forest-steppe occupied the whole area of southeastern Ukraine. The df3 interstadial was the coldest, and the Alleröd was the warmest of the interstadials. The last Vytachiv interstadial vt3 was also rather warm in Crimea, but its transitional phase to Bug times, vt3c, was cool in both areas (boreal and northern-boreal climates). Stadial environments included subperiglacial dry steppe in the Donetsk-Pryazov area and boreal grassland in the foothills of the Crimean Mountains. The latest Late Pleniglacial stadial pc1 was the driest—a semidesert in the Donets-Pryazov area and a dry steppe in eastern Crimea. The stadial pc3, correlated with the Younger Dryas, was wetter in the Donetsk area and warmer in Crimea. In both areas, the beginning of Bug times was wetter than its later half. During the period studied, western Crimea was wetter than eastern Crimea and provided more habitats for arboreal and broad-leaved flora. The ecosystems were generally warmer and wetter in Crimea than in the Donetsk-Pryazov area. This is obviously a reason why the Middle Paleolithic population survived here up to the very end of the Middle Pleniglacial (28 ka), and coexisted with the Upper Paleolithic cultures (Chabai, 2004). This was not a case for the plains area of Ukraine, where the Middle Pleniglacial deposits yield only Upper Paleolithic artifacts. In both the studied areas, paleoclimatic changes demonstrate the same pattern, and they correspond well with transgressiveregressive cycles of the Black Sea. The last Vytachiv interstadial vt3 is correlated with the end of the Tarkhankutian (Surozhian) transgression, which finished at around 31.3 ka (Arslanov et al., 2005), or 30.5 ka BP (Balabanov and Izmailov, 1988). The cold Bug times, when the thickest loess of southern Ukraine was formed, correspond to the Neoeuxinian regression. The latter started at 28.5 ka (Yanko-Hombach et al., 2002). A decrease in aridity during the Dofinivka interstadials can be correlated with the first rise in the Neoeuxinian basin water level. This early Enikalian transgressive phase is related to a period between 17 and 15 ka (Chepalyga, 2002; Murdmaa et al., 2006). The early Prychernomorsk stadial pc1 could correspond then to the deep “post-Enikalian” regression. The middle Prychernomorsk interstadials pc2 occurred at the same time, when the Neoeuxinian transgression became particularly apparent between 13.8 and

130

Gerasimenko

A

B Figure 7. Soil-vegetational dynamics during 30–10 ka: (A) southeastern Ukraine; and (B) foothills of the Crimean Mountains.

Climatic and environmental oscillations in southeastern Ukraine 12.0 ka (Arslanov et al., 2005). The late Prychernomorsk stadial, pc3, correlated with the Younger Dryas, corresponds to the next Neoeuxinian low stand (Chepalyga, 2002; Murdmaa et al., 2006). ACKNOWLEDGMENTS The author is thankful to V. Chabai, A. Yanevich, V. Cohen, A. Krotova (Archaeological Institute of Ukrainian National Academy of Sciences), and to A. Gorelik (Lugansk Pedagogical University) for the chance to participate in the multidisciplinary study of the Paleolithic sites, and also thanks G. Pedanyuk for his cooperation in the investigation of the Donets area. I thank E. Kvavadze and an anonymous reviewer for useful comments and constructive reviews of the manuscript. REFERENCES CITED Arap, R.Ya., 1976, Sporovo-pyltsevye issledovania poverhnostnyh prob pochv rastitel’nyh zon ravninnoy Ukrainy: Kiev, Institut Botaniki Akademii nauk Ukrainy, 25 p. (in Russian). Arslanov, Kh.A., Dolukhanov, P.M., and Gei, N.A., 2005, Climate, Black Sea levels and human settlements in Caucasus littoral 50,000–9,000 years before present, in Yanko-Hombach, V., Buynevich, I., Chivas, A., Gilbert, A, Martin, R., Mudie, P., eds., Extended Abstracts of the First Plenary Meeting and Field Trip of IGCP 521 Project “Black Sea–Mediterranean corridor during the last 30 k.y.: Sea-level change and human adaptation,” (8–15 October 2005, Kadir Has University, Istanbul, Turkey), Ankara, Tübitak, p. 10–11. Artyushenko, A.T., and Mishnev, V.G., 1978, Istoria rastitel’nosti Krymskikh yayl i ikh sklonov v golotsene: Kyiv, Naukova Dumka, 131 p. (in Russian). Artyushenko, A.T., Pashkevich, G.A., Parishkura, S.I., and Kareva, Ye.V., 1973, Paleobotanicheskaya kharakteristika opornykh razrezov chetvertichnykh otlozheniy sredney i yuzhnoy chasti Ukrainy: Kyiv, Naukova Dumka, 96 p. (in Russian). Balabanov, I.P., and Izmailov, Y.A., 1988, Izmenenie urovennogo i gidrokhimicheskogo rezhimov Chernogo i Azovskogo morey za poslednie 20 tysyach let: Vodnye Resursy, v. 6, p. 54–62 (in Russian). Bezus’ko, L.G., Kostylyov, O.V., and Popovich, S.Yu., 1997, Fitotsenologichna interpretatsia palinologichnikh danykh na prykladi Chornomors’kogo biosfernogo zapovidnyka: Ukrainsky Botanichny Zhurnal, v. 54, no. 1, p. 80–86 (in Ukrainian). Bolikhovskaya, N.S., 1986, Paleogeography and stratigraphy of Valdai (Wurm) loesses of the southwestern part of the East-European Plain by palynological data: Annales Universitatis Mariae Curie-Sklodowska (Lublin), Section B, v. XLI, no. 6, p. 111–124. Bolikhovskaya, N.S., 1995, Evolutsia lessovo-pochvennoy formatsii Severnoy Evrazii: Moscow, Moscow University Press, 269 p. (in Russian). Burke, A., Markova, A.K., Mikhailesku, C., and Patou-Mathis, M., 1999, The animal environment of Western Crimea, in Chabai, V., and Monigal, K., eds., The Middle Paleolithic of Western Crimea, vol. 2: Liège, ERAUL, 87, p. 143–151. Chabai, V.P., 2004, Sredny Paleolit Kryma: Simferopol, Shlyah, 323 p. (in Russian). Chabai, V., and Monigal, K., eds., 1999, The Paleolithic of Crimea: II. The Middle Paleolithic of Western Crimea, Volume 2: Etudes et Recherches Archéologiques de l’Université de Liége 87, 260 p. Chabai, V., Monigal, K., and Marks, A., eds., 2004, The Paleolithic of Crimea: III. The Middle Paleolithic and Early Upper Paleolithic of Eastern Crimea: Etudes et Recherches Archéologiques de l’Université de Liége 104, 479 p. Chepalyga, A.L., 2002, Chernoye more, in Velichko, A.A., ed., Dinamika landshaftnykh komponentov i vnutrennikh morskikh baseynov Severnoy Evrazii za poslednie 130 000 let: Moscow, GEOS, p. 170–182 (in Russian). Cohen, V., Gerasimenko, N., Rekovets, L., and Starkin, A., 1996, Chronostratigraphy of rockshelter Skalisty (Crimea): European Prehistory, v. 9, p. 325–358.

131

Dinesman, L.G., 1977, Biogeotsenozy stepey v golotsene: Moscow, Nauka, 159 p. (in Russian). Gerasimenko, N., 1997a, Environmental and climatic changes between 3 and 5 ka BP in southeastern Ukraine, in Dalfes, H.N., Kukla, G., and Weiss, H., eds., Third Millennium B.C. Climate Change and Old World Collapse: Berlin, Springer-Verlag, p. 371–399. Gerasimenko, N., 1997b, Prirodnaya sreda obitania cheloveka na yugo-vostoke Ukrainy v pozdnelednikovye i golotsene: Arkheologicheskiy Al’manakh, v. 6, p. 3–64 (in Russian). Gerasimenko, N., 1999, Late Pleistocene vegetational history of the Kabazi-II Paleolithic site, in Chabai, V., and Monigal, K., eds., The Paleolithic of Crimea: II. The Middle Paleolithic of Western Crimea, Volume 2: Etudes et Recherches Archéologiques de l’Université de Liége 87, p. 115–141. Gerasimenko, N., 2001, Late Pleistocene vegetation and soil evolution at the Kiev loess plain as recorded in the Stari Bezradychy section, Ukraine: Studia Quaternaria, v. 17, p. 19–28. Gerasimenko, N., 2004, Vegetational history of Buran-Kaya III, in Chabai, V., Monigal, K., and Marks, A., eds., The Paleolithic of Crimea: III. The Middle Paleolithic and Early Upper Paleolithic of Eastern Crimea: Etudes et Recherches Archéologiques de l’Université de Liége 104, p. 19–34. Gerasimenko, N., 2006, Upper Pleistocene loess-paleosol and vegetational successions in the Middle Dnieper Area, Ukraine: Quaternary International, v. 149, p. 55–66, doi: 10.1016/j.quaint.2005.11.018. Gerasimenko, N., 2007, Environmental changes in the Crimean mountains during the Last Interglacial–Middle Pleniglacial as recorded by pollen and lithopedology: Quaternary International, v. 164–165, p. 207–220, doi: 10.1016/j.quaint.2006.12.018. Gerasimenko, N.P., and Pedanyuk, G.I., 1991, Paleogeograficheskie etapy pliotsena i pleistotsena Zapadnogo Donbassa: Moscow, Viniti, 498 p. (in Russian). Gorelik, A.F., 2001, Pamyatniki Rogaliksko-Peredel’skogo rayona. Problemy final’nogo paleolita Yugo-Vostochnoy Ukrainy: Kyiv-Lugansk, RIO LIVD, 364 p. (in Russian). Gozhik, P.F., Shovkoplyas, V.M., Komar, M.S., Matviishina, Zh.M., and Peredriy, V.I., 2000, Korelyatsia lesiv i lyodovykovyh vidkladiv Pol’shchi i Ukrainy: Kyiv, Instytut Geologichnykh nauk Natsional’noi Akademii nauk Ukrainy, 34 p. (in Ukrainian). Grichuk, V.P., and Zaklinskaya, E.D., 1948, Analiz iskopaemykh pyl’tsy i spor i primenenie v paleogeografii: Moscow, OGIZ-Geografiz, 91 p. (in Russian). Gubonina, Z.P., 1985, Uslovia obitania paleoliticheskogo cheloveka v predgor’yah Kryma, in Palinologia Chetvertichnogo Perioda: Moscow, Nauka, p. 95–103. Haesaerts, P., Borziak, I., Chirica, V., Koulakovska, L., and Damblon, F., 2003, The East-Carpathian loess record: A reference for the Middle and Late Pleniglacial stratigraphy in Central Europe: Quaternaire, v. 14, no. 3, p. 163–188, doi: 10.3406/quate.2003.1740. Ivanova, I.K., 1987, Paleogeografia i paleoekologia sredy obitania lyudey kamennogo veka na Srednem Dnestre, Stoyanka Molodova V., in Mnogosloynaya paleoliticheskaya stoyanka Molodova V Lyudi kamennogo veka i okruzhayushchaya sreda: Moscow, Nauka, p. 94–123. Khotinsky, N.A., 1982, Golotsenovye hronosfery: diskussionnye problemy paleogeografii golotsena, in Razvitie prirody territorii SSSR v pozdnem pleistotsene i golotsene: Moscow, Nauka, p. 69–87. Klopotovskaya, N.V., 1976, K metodologii rekonstruktsii rastitel’nosti gornykh rayonov na osnove palinologicheskikh dannykh, in Palinologia v SSSR: Moscow, Nauka, p. 38–41. Krotova, A., 1996, Amvrosievka new AMS dates for a unique bison kill site in Ukraine: European Prehistory, v. 9, p. 357–363. Markova, A.K., 2004, Small mammal fauna from Buran-Kaya III, in Chabai, V., Monigal, K., and Marks, A., eds., The Middle Paleolithic and Early Upper Paleolithic of Eastern Crimea, vol. 3: Liège, ERAUL, 104, p. 75–98. Marks, A., and Chabai, V., eds., 1998, The Paleolithic of Crimea: I. The Middle Paleolithic of Western Crimea, Volume 1: Etudes et Recherches Archéologiques de l’Université de Liége 84, 382 p. McKinney, C., 1998, U-series dating of enamel, dentine, and bone from Kabazi II, Staroselye, Kabazi V, and GABO, in Marks, A.E., and Chabai, V., eds., The Middle Paleolithic of Western Crimea, vol. 1: Liège, ERAUL, 84, p. 341–354. Murdmaa, I., Ivanova, E., Chepalyga, A., Cronin, T., Levchenko, O., Howe, S., Manushkina, A., Pasechnik, I., and Platonova, E., 2006, Paleoenvironments on the North Caucasian Black Sea shelf since LGM, in YankoHombach, V., Buynevich, I., Gilbert, A., Martin, R., eds., Extended

132

Gerasimenko

Abstracts of the Second Plenary Meeting and Field Trip of IGCP 521 Project “Black Sea–Mediterranean corridor during the last 30 k.y.: Sealevel change and human adaptation,” (20–28 August 2006, I.I. Mechnikov Odessa National University, Odessa, Ukraine): Astroprint, Odessa, p. 127–129. Patou-Mathis, M., 2004, Archeozoological analyses of large mammal fauna from Buran-Kaya III layer B, in Chabai, V., Monigal, K., and Marks, A., eds., The Middle Paleolithic and Early Upper Paleolithic of Eastern Crimea, vol. 3: Liège, ERAUL, 104, p. 355–370. Pettit, P., 1998, Middle Paleolithic and early Upper Paleolithic Crimea: the radiocarbon chronology, in Otte, M., ed., Préhistoire d’Anatolie: Genèse des deux mondes, vol. 1: Liège, ERAUL, 85, p. 329–338. Rink, W.J., Lee, H.-K., Rees-Jones, J., and Goodger, K.A., 1998, Electron spin resonance (ESR) and mass spectrometric U-series (MSUS) dating of teeth in Crimean Paleolithic sites: Staroselye, Kabazi II, and Kabazi V, in Marks A.E., and Chabai, V., eds., The Middle Paleolithic of Western Crimea, vol. 1: Liège, ERAUL, 84, p. 323–340. Rousseau, D.D., Gerasimenko, N., Matviishina, Zh., and Kukla, G., 2001, Late Pleistocene environments of the central Ukraine: Quaternary Research, v. 56, p. 349–356, doi: 10.1006/qres.2001.2270. Shelkoplyas, V.N., Gozhik, P.F., Khristoforova, T.F., Matsuy, V.M., Chugunny, Yu.G., Palatnaya, N.N., Shevchenko, A.I., Morozov, G.V., and Lysenko, O.V., 1986, Chetvertichnye otlozhenia Ukrainy: Kiev, Naukova Dumka, 152 p. (in Russian). Sirenko, N.A., and Turlo, S.I., 1986, Razvitie pochv i rastitel’nosti Ukrainy v pliotsene i pleistotsene: Kyiv, Naukova Dumka, 186 p. (in Russian). Spiridonova, E.S., 1991, Evolyutsia rastitel’nogo pokrova basseyna Dona v verchnem pleistotsene i golotsene: Moscow, Nauka, 218 p. (in Russian). Stepanchuk, V.M., Cohen, V.Yu., Gerasimenko, N.P., Damblon, F., Haesaerts, P., Zhuravlev, O.P., Kovalyukh, M.M., Petrun’, V.F., Plicht, van der, J., Puchkov, P.V., Rekovets, L.I., and Turner H.G., 2004, Bagatosharova

stoyanka Mira na Serednyomu Dnipri: Osnovni rezul’taty rozkopok 200 roku: Kamyana Doba, v. 6, p. 62–98. Van der Hammen, T., 1995, The Dinkel valley revisited: Pleniglacial stratigraphy of the eastern Netherlands and global climatic change: Mededelingen Rijks Geologische Dienst, v. 52, p. 343–355. Veklich, M.F., ed., 1993, Stratigraficheskya shema chetvertichnykh otlozheniy Ukrainy: Kiev, Gosgeolkom Ukrainy, 48 p. (in Russian). Veklich, M.F., Sirenko, N.O., Dubnyak, V.A., Mays’ka, Zh.M., Mel’nichuk, I.V., and Paryshkura, S.I., 1973, Rozvytok gruntiv Ukrainy u piznyomu kaynozoi: Kyiv, Naukova Dumka, 210 p. (in Ukrainian). Veklich, M.F., Matviishina, Zh.N., Medvedev, V.V., Sirenko, N.A., and Fedorov, K.N., 1979, Metodika paleopedologicheskikh issledovaniy: Kyiv, Naukova Dumka, 272 p. (in Russian). Velichko, A.A., 1988, Geoecology of the Mousterian in east Europe and the adjacent area, in Otte, M., ed., Actes du Colloque International de Liège. L’homme de Neandertal: Volume 2. L’Environment: Etudes et Recherches Archéologiques de l’Université de Liège 29, p. 179–207. Vigilyanskaya, L.I., 1999, Paleomagnetic research of the Rogalik-Peredelsk archeological monument: Geophysical Journal, v. 18, p. 1031–1037. Yanevich, A.A., Stepanchuk, V.N., and Cohen, V.Yu., 1996, Buran-Kaya III and Skalisty Rockshelter: Two new dated late Pleistocene sites in Crimea: European Prehistory, v. 9, p. 315–325. Yanko-Hombach, V., Balabanov, I., Mitropolsky, A., and Glebov, A., 2002, Late Pleistocene–Holocene history of the Black Sea: Is there any room for Ryan-Pitman hypothesis “Noah’s flood”?, in Proceedings of the Third International Conference on Environmental Micropaleontology, Microbiology and Meiobenthology, 1–6 September 2002: Vienna, p. 203–206. MANUSCRIPT ACCEPTED BY THE SOCIETY 22 JUNE 2010

Printed in the USA

The Geological Society of America Special Paper 473 2011

Late Pleistocene and Holocene paleoenvironments of Crimea: Pollen, soils, geomorphology, and geoarchaeology Carlos E. Cordova* Department of Geography, Oklahoma State University, Stillwater, Oklahoma 74078, USA Natalia P. Gerasimenko Department of Earth Sciences and Geomorphology, Taras Shevchenko University of Kyiv, Glushkova 2, Kyiv, DSP 680, Ukraine Paul H. Lehman Department of Geography, University of Texas at Austin, Austin, Texas 78712, USA Alexander A. Kliukin Cathedra of Geography, Taurida Vernadsky National University, Simferopol, Crimea, Ukraine

ABSTRACT We discuss pollen, soil, geomorphologic, and archaeological records used for reconstructing climatic, biogeographic, and human-environment events in the Crimean Peninsula during the past 130 k.y. Warm and moist conditions conducive to forest growth prevailed during the Eemian Interglacial (marine isotope stage [MIS] 5e). Although sea levels were higher than at present, a review of the stratigraphic and geomorphic data suggests that the peninsula was not detached from the mainland. During the last glacial period (MIS 5d–MIS 2), conditions fluctuated between steppe and tree growth in warmer places during the stadials, and forest-steppe during the interstadials. The Pleistocene–Holocene transition involved forest growth during the Bølling-Allerød interstadials, steppe during the Younger Dryas, and a forest-steppe during the early Holocene. The establishment of the modern Black Sea ca. 7 ka and increasing temperatures led to the formation of the modern vegetation belts, ushering in optimal conditions for the establishment of Neolithic communities. A dry period peaked around 4–3.5 ka, followed by milder conditions that lasted until the colonization of Crimea by Greek farmers during the middle part of the first millennium A.D. Dry conditions at the end of the same millennium led to the abandonment of agriculture and settlement decline. Sea-level oscillations during the late Holocene had an important effect on shoreline configuration, lagoonal systems, coastal wetlands, and human settlements. Data used in this paper were drawn from a number of published papers, mostly in Russian and Ukrainian, as well as records produced by the authors’ research.

*[email protected] Cordova, C.E., Gerasimenko, N.P., Lehman, P.H., and Kliukin, A.A., 2011, Late Pleistocene and Holocene paleoenvironments of Crimea: Pollen, soils, geomorphology, and geoarchaeology, in Buynevich, I.V., Yanko-Hombach, V., Gilbert, A.S., and Martin, R.E., eds., Geology and Geoarchaeology of the Black Sea Region: Beyond the Flood Hypothesis: Geological Society of America Special Paper 473, p. 133–164, doi: 10.1130/2011.2473(09). For permission to copy, contact editing@ geosociety.org. © 2011 The Geological Society of America. All rights reserved.

133

134

Cordova et al.

INTRODUCTION The Crimean Peninsula occupies an area of ~26,000 km2 bounded on the west and south by the Black Sea, on the east and northeast by the Azov Sea, and on the north by the Sivash (a lagoonal system connected to the Azov Sea) (Fig. 1). At present, Crimea is connected to the Ukrainian mainland by the Perekop, a 25-km-wide strip of land with its highest elevation around 20 m above sea level. The Crimean Peninsula is surrounded by a relatively shallow continental shelf, and as a consequence, significant drops in sea level have exposed extensive areas of land, thereby modifying local climate and vegetation (Cordova, 2007). Conversely, a rise in sea level, such as the one experienced during the Eemian Interglacial (marine isotope stage [MIS] 5e), meant the inundation of low-lying areas, and a possible separation from the mainland (Lazukov et al., 1981; Chabai, 2007). The northern two thirds of the peninsula are composed of flatlands and rolling plains, while the southern extremity is occupied by mountains. The flat, steppe-like region has strong geologic, climatic, and biogeographic similarities with the Ukrainian plains (Berg, 1950; Didukh, 1992). The southern mountainous area shows geological affinities with the Caucasus Mountains and other orogenic belts of Eastern Europe (Мuratov, 1974), as well as biogeographic affinities with the Caucasus, the Balkans, and

Asia Minor (Didukh, 1992). The semi-isolation of the Crimean Mountains from other regional mountainous areas has resulted in a considerable number of endemic species (Biodiversity Support Program, 1999). The lack of glacial activity on its mountains suggests that temperatures in Crimea during the cold stages of the Ice Age were relatively milder than in other regions, creating favorable conditions for tree refugia (Comes and Kadereit, 1998; Gerasimenko, 1999; Morozova and Kozharinov, 2001; Cordova, 2007). Consequently, Crimea has been an important terrain for the study of plant and animal communities and hominin/human paleoecological change during the last interglacial and glacial stages. Several geoarchaeological and paleoecologic projects during the past decade have provided new pieces of information for reconstructing Pleistocene ecosystem change (Chabai et al., 1999, 2004; Gerasimenko, 1999, 2004, 2005a, 2005b, 2007) and Holocene climatic and human-related change (Cordova and Lehman, 2003, 2005, 2006; Cordova, 2007). The objective of this paper is to discuss some of the paleoenvironmental issues investigated through these earlier studies, as well as research developed in the course of the twentieth century. The main points that this paper addresses are: (1) the effects of late Pleistocene climatic change on vegetation, soils, and Paleolithic subsistence in the Crimean Mountains and Piedmont; (2) the climatic and vegetation changes during the Pleistocene-Holocene

45°

33°

34°

35°

Figure 1. The Crimean Peninsula, general topography and geological transects.

36°

Late Pleistocene and Holocene paleoenvironments of Crimea transition; (3) the development of the Holocene landscape in relation to climate fluctuations and human development; and (4) the effects of late Holocene sea-level oscillations on coastal morphology and human settlement.

135

Dates of climatic events in the text are often rounded to ka, unless specific dates are referred to, in which case the calibrated yr B.P. or B.C./A.D. is used. GEOMORPHIC AND GEOLOGIC CONTEXTS

APPROACHES AND METHODS The Crimean Peninsula consists of three main physiographic units: the plains, the piedmont, and the mountains (Fig. 1). Areas of rolling plains and low hills break the uniformity of the plains in the Tarkhankutsk and Kerch Peninsulas. The piedmont corresponds to the gentle slope of the Outer Ridge, and the mountains constitute the Inner and Main Ridges. The top of the Main Ridge has a plateau-like summit, known as the Yaila or Yailas (a Turkic term referring to summer pastures), consisting of predominantly calcareous terrain with caves, sinkholes, and uvalas, most of which contain Quaternary sediments. Tectonically, the Crimean Peninsula occupies two megastructures: the Scythian Platform and the Megaanticlinorium of the Crimean Mountains (Мuratov, 1974). The Scythian Platform constitutes the northern part of the peninsula, including the Tarkhankutsk Peninsula and the northern part of the Kerch Peninsula. Prominent structures of the Scythian Platform include the Novoselovsk Rise and the Tarkhankutsk anticline, the Simferopol Rise, the Al’ma Basin, and the Indolo-Kuban Depression. The latter occupies most of the northern and northwestern plains, including the northern part of the Kerch Peninsula. The Megaanticlinorium of the Crimean Mountains, as defined by Мuratov (1974), consists of a series of anticlines and synclines forming a larger upwarp of Mesozoic–Paleogene origin. The Crimean Mountains are the expression of the northern half of the Megaanticlinorium; the southern half is submerged beneath the sea after having slid down along faults (Мuratov, 1974).

The sources of information discussed here include published and unpublished research by the authors of this paper as well as published research papers of other researchers. The localities discussed in detail are those studied by the authors, and these include the mountains and piedmont (for the late Pleistocene vegetationclimate reconstruction), and southwestern Crimea, the Yaila, and the Kerch Peninsula (for the Holocene). Research by other authors was obtained mainly from papers in Russian and Ukrainian, most of which have limited distribution in the west. The topics include marine, alluvial, and loess sequences, as well as research on sealevel changes and the evolution of the Sivash lagoonal system. Lithostratigraphic and chronostratigraphic names are anglicized where possible (e.g., Karangatskii = Karangat). In other cases the Russian, Tatar, or Ukrainian names are retained. In many instances, spelling of names conformed to the most accepted spelling, or the spelling used in the chapters of the compilation by Yanko-Hombach et al. (2007), or from the most common transliteration of the Cyrillic spelling into English. Radiocarbon dates on sections in the figures are presented in radiocarbon yr B.P. (RCYBP). Holocene ages are calibrated to calendar yr B.P. or B.C./A.D., particularly when historical events are discussed. The calibration method used is the Cologne Radiocarbon Calibration and Paleoclimate Package (CalPal, available at http:www.calpal.de). Tables 1, 2, and 3 show the 68% probability range for all radiocarbon dates mentioned in this paper.

TABLE 1. ACCELERATOR MASS SPECTOMETRY DATES FOR THE HERAKLEAN PENINSULA AND THE TARKHANKUTSK PENINSULA Section

14

BBBP-3

Depth (cm) 60–80

Laboratory number Beta-156480

C age (yr B.P.) 3730 ± 40

68% age range (cal yr B.P.) 4014–4144

Bulk carbon

BBBP-2 " " "

50–90 90–100 90–135 590–600

Beta-156478 Beta-127551 Beta-156479 AA48298

5450 ± 40* 8070 ± 40 8550 ± 40 12,028 ± 90

6221–6289 8882–9036 9508–9544 13,793–14,272

Charcoal Bulk carbon Bulk carbon Bulk carbon

AA " " "

25–30 75–90 170–180 310–320

AA48202 Beta-127550 AA48293 AA48294

2907 ± 30 3060 ± 40 5233 ± 40 9499 ± 50

3000–3116 3233–3337 5947–6095 10,707–11,019

Ash Bulk carbon Bulk carbon Bulk carbon

MM2 " "

110–120 245–270 340–350

AA48296 Beta-127552 T-16421A

4637 ± 60 5730 ± 40 8342 ± 70

5315–5458 6479–6602 9259–9438

Bulk carbon Bulk carbon Bulk carbon

BB

185–190

Beta-137094

7000 ± 70

7754–7915

Charcoal

4090 ± 60

4523–4771

Bulk carbon

EF 130–140 Beta-137095 Note: From Cordova and Lehman (2005).

Material dated

136

Cordova et al. TABLE 2. ACCELERATOR MASS SPECTOMETRY DATES FOR THE CHËRNAYA RIVER SECTIONS, TARKHANKUTSKAYA BALKA, AND YALTINSKAYA YAILA Section NG2

SB-1

TKB-1 Yaila

14

Depth (cm) 62–74 139–152 230–240 261–274

Laboratory number Beta-127553 Beta-137097 AA48303 Beta-127554

C age (yr B.P.) 1550 ± 40 3530 ± 40 4258 ± 60 5380 ± 40

68% age range (cal yr B.P.) 1402–1504 3744–3868 4720–4863 6088–6259

Material dated

109–123 220–236 292–312 445–461 445–461

AA77565 AA77566 AA77567 AA77568 AA77568

2689 ± 38 4970 ± 64 5082 ± 96 4591 ± 41 5740 ± 50

2771–2839 5646–5839 5718–5922 5150–5430 6486–6648

Bulk carbon Bulk carbon Bulk carbon Bulk carbon Bulk carbon

275

AA48284

4477 ± 39

5028–5252

Charcoal

56–69 90–108 158–170 194–205

AA78717 AA78718 AA78719 AA48275

2034 ± 32 1618 ± 36 3172 ± 39 7823 ± 93

1951–2042 1440–1547 3373–3439 8524–8837

Bulk carbon Humic acids Humic acids Humic acids

Charcoal Bulk carbon Bulk carbon Bulk carbon

TABLE 3. ACCELERATOR MASS SPECTOMETRY AND STANDARD DATES FROM DEPOSITS IN LASPI BAY (LASP 1 SECTION) 14

Unit or Laboratory C age 68% are range layer number (yr B.P.) (cal yr B.P.) 2540 ± 40 2539–2725 Unit 9 -329* Layer A AA48290 2501 ± 35 2514–2725 Layer A AA48290 2781 ± 35 2835–2930 2940 ± 60 3012–3203 Layer A -331* *Reported by Firsov (1972); these are standard radiocarbon dates.

In general terms, the oldest lithological units are exposed along the southern coast, while the younger ones are exposed on the central and northern parts of the peninsula (Fig. 2). The Crimean Megaanticlinorium consists of rocks of Triassic and Jurassic age in its core, and Cretaceous and Tertiary rocks on its exterior (Fig. 2B). The Heraklean Peninsula is formed predominantly by Miocene marine deposits creating a cuesta-like structure, with a veil of terrestrial Pliocene and Pleistocene deposits on top (Fig. 2A). The Kerch Peninsula combines Paleocene– Oligocene rocks in the south with Miocene marine deposits on the north, all of which form a series of synclines and anticlines (Blagovolin, 1962) (Fig. 2C). The Pliocene–Pleistocene geology of Crimea presents several facies, showing first an alternation of loess (or loess-like clays) and paleosols (Blagovolin, 1962; Мuratov, 1974; Veklich and Sirenko, 1976). During the early Pleistocene, the precursors of the modern stream valleys developed, and during the middle and late Pleistocene, a series of fluvial terraces formed within them.

Calibrated calendar age 682 ± 43 BC 653 ± 89 BC 933 ± 47 BC 1158 ± 95 BC

Material dated Timber Hearth Charred wood Charred wood

ing from a higher elevation and orographic precipitation. The southern slopes of the mountains have a relatively warm, subMediterranean climate (Douguedroit and Zimina, 1987; Yena et al., 1996). The Heraklean Peninsula and most of the foothills of the mountains in the southwest possess climatic conditions that are moister than the semiarid steppes and somewhat cooler than the warm southern coast (Cordova and Lehman, 2005; Cordova, 2007). The main weather patterns contributing to the Crimean climatic conditions are the seasonal variation of the Westerly Winds, the Siberian High, and the common cold fronts associated with mid-latitude cyclones. Westerly Winds predominate most of the year, bringing moisture from the adjacent sea, although occasionally, winds from the steppe weaken the effect of the Westerly Winds, particularly north of the mountains. Accordingly, the western and southwestern coasts of Crimea enjoy a relatively wetter climate than the rest of the peninsula.

CLIMATIC AND BIOGEOGRAPHIC CONTEXTS While the northern plains, and the Tarkhankutsk and Kerch Peninsulas, experience a semiarid climate similar to that of the southern Ukrainian and Russian plains, the piedmont and the mountains benefit from a wetter and cooler climate result-

Figure 2. Geological transects and greater lithostratigraphic units. Transect locations are indicated on the map of Figure 1. Data have been assembled from a diverse number of sources (e.g., Blagovolin, 1962; L’vova, 1982; Мuratov, 1974; Podgorodetskiy, 1998; Polkanova and Varuschenko, 1969; Semenenko et al., 1979; Veklich and Sirenko, 1976; Slavin, 1975).

Late Pleistocene and Holocene paleoenvironments of Crimea

137

Figure 3. Biogeographic complexes and climate variables. Biogeographic data are based on Biodiversity Support Program (1999). Climatic data were assembled from public records.

138 Cordova et al.

Late Pleistocene and Holocene paleoenvironments of Crimea During the winter, the north and east are directly exposed to blasts of Arctic cold, dry air resulting from the expansion of the Siberian High. In contrast, the southern coast, still under the influence of the Westerly Winds, receives cyclonic waves from the southwestern Black Sea and the Mediterranean Sea (Podgorodetskiy, 1988). During the summer, frontal waves associated with cyclones bring rain to the plains and northern part of the mountains (e.g., Fig. 3, Simferopol climograph). During this time, the mountains create a rain shadow on the southern coast, where precipitation remains relatively low, except for a very short period centered in June (e.g., Fig. 3, Yalta climograph). In essence, the climate of the southern coast, with its the relatively mild winters and dry summers, resembles the typical Mediterranean type of climate (Douguedroit and Zimina, 1987), and this presents one reason why native and introduced Mediterranean vegetation thrives in this region. The northern plains, and the Kerch and Tarkhankutsk Peninsulas have steppe vegetation, grading from meadow steppes in the south to the semiarid and salinized steppes of the northwestern coast, the region around the Sivash, and on the southern coast of the Kerch Peninsula (Fig. 3). Although today, the meadow steppe has largely been replaced by crop fields, the original vegetation according to Podgorodetskiy (1988) had four variants: the Tarkhankutsk steppe, which was dominated by grasses and petrophytic herbs, particularly thyme; the central steppe, which was dominated by grasses and a variety of herbs (i.e., herbetum mixtum); the southern section, which was a meadow steppe with

139

shrubs; and the Kerch section of the steppe, which was dominated by patches of grasses and xerophytic herbs, particularly wormwood (Artemisia). The semiarid and salinized steppe is less appropriate for agriculture, and therefore, patches of the original vegetation are still found. The dominant vegetation of the salinized steppe consists of small shrubs such as wormwood and chenopods, as well as a variety of salt-tolerant grasses. Descriptions of vegetation for the rest of the region can be summarized from Didukh (1992) and ordered into a simplified north-south altitudinal transect (Fig. 4). The piedmont and the Inner and Outer Ridges, as well as the Heraklean Peninsula, present mainly forest-steppes, where the dominant tree species are pubescent oak (Quercus pubescens) and eastern hornbeam (Carpinus orientalis). The humid northern slopes of the mountains have dense forests of oak and eastern hornbeam at lower elevations, and hornbeam (Carpinus betulus) and beech (Fagus sylvatica) at higher elevations. The mountain summit, or Yaila, consists of meadow-steppe vegetation with alpine elements and numerous petrophytic herbs and grasses. The forests of the southern slope are dominated by pine (Pinus pallasiana and Pinus kochiana), although stands of beech (Fagus sylvatica) and oaks (mainly Quercus petrae) are not uncommon. The lowest 250 m of the southern slopes consist of pubescent oak and eastern hornbeam woodlands, and sub-Mediterranean elements grouped into communities known as shiblyak and phrygana. The shiblyak is a community of shrubs that includes Juniperus oxycedrus, J. excelsa, Pistacia mutica, Paliurus spina-christi, Arbutus

Figure 4. Simplified vegetation transect across the mountains of Crimea (after Cordova et al., 2001).

140

Cordova et al.

andrachne, Cotinus coggygria, and Ruscus ponticus, with scattered stands of eastern hornbeam, pubescent oak, and Stankevicz’s pine (Pinus pityusae). The phrygana is a community of scrub and herbs of the mint family (Lamiaceae), cistus family (Cistaceae), and Asphodeline spp., among other scrub typical of the Mediterranean region. The distribution of soil types in Crimea is closely associated with the vegetation types. The mountain forest reveals predominantly Rendzina soils on steep limestone and marl slopes, forest brown soils under dense forest on relatively gentle slopes, and meadow soils on the Yaila. The piedmont and plains are characterized by Chernozem (black earth) in the lower part of the piedmont, corresponding to the meadow steppes and open areas of the forest-steppe. The black earth variant of Crimea, which corresponds to the meadow-steppe and open areas of the foreststeppe, is known as the “southern Chernozem of the piedmont” (Krupskiy and Polupan, 1979). The northern steppes of Crimea, which are considerably drier, are characterized by salinized soils such as solonetz and solonchak. In areas around the Sivash, the combination of high water table, low precipitation, and high summer temperatures is the main agent of soil salinization (Podgorodetskiy, 1988; Krupskiy and Polupan, 1979). The southern and southwestern coastal areas tend to have cinnamonic soil, a variation of the terra-rossa identified by Gerasimov (1954) for the sub-Mediterranean parts of the USSR. While cinnamonic soils are common on relatively gentle slopes, stony Rendzina soils predominate on mountainsides. In numerous areas of the western end of the piedmont, however, relict cinnamonic soils exist on gentle slopes. LATE PLEISTOCENE CHRONOSTRATIGRAPHY Marine Deposits A chronology of marine deposits and terraces has been constructed based on the sequences of regressions and transgressions and their correlation with cold and warm stages, respectively. Marine transgression-regression cycles during the late Pleistocene were studied in the Strait of Kerch (Shcherbakov et al., 1977; Fedorov, 1978), the shelf off the southern coast (Shcherbakov et al., 1977), the mouth of the Salgir River (Semenenko et al., 1979), and the western and northwestern shelf (Shcherbakov et al., 1976; Gozhik et al., 1987; Gozhik and Novosel’skiy, 1989). Regressions are commonly marked by alluvial and eolian (mainly loess-like) sediments deposited on the exposed shelf. Transgressions are expressed as marine terraces above the present sea level or by marine sediments backing up stream valleys. Schemes of correlation between MIS (marine isotope stages) and transgressional and regressional events in the northern Black Sea region have also been described by Veklich et al. (1993), Gozhik et al. (2000), and Chepalyga (2007). In the general chronology of sea-level events, the Paleo-Euxinian represents the lowering of sea level accompanied by accumulations of eolian silts related to the Riss Glaciation (Fedorov, 1978), or MIS 6, 8,

and 10 (Chepalyga, 2007). The Karangat transgression has been assigned to the Eemian Interglacial, or MIS 5e (Fedorov, 1978; Veklich et al., 1993; Chepalyga, 2007). The early Neoeuxinian regression is expressed in terrestrial deposits, in limans, and on shelves, as evidence of the coldest stages (MIS 4 and 2), separated by the Surozh transgression (roughly MIS 3 and early MIS 2) (Chepalyga, 1984; Veklich et al., 1993). The Karangat transgression was presumably the highest sea level reached during the late Pleistocene (Fedorov, 1978; Chepalyga, 2007). This was the basis for a suggestion that during the height of the Eemian Interglacial, the Crimean Peninsula became isolated from the mainland by a shallow strait in the Sivash and Perekop areas (Chabai [2007] based on Lazukov et al. [1981]). Reports by Olenkovsky (2000) on prehistoric site survey and in the Sivash area have shown no Middle Paleolithic sites, most of which should correspond to the Karangat and Surozh transgressional stages. On the contrary, several Upper Paleolithic and Mesolithic sites have been found in the area, which suggest that the reduced soil salinity induced by the low sea levels of the Neoeuxinian regression may have made this area more attractive to a diverse assemblage of flora and fauna. The assumption of an island situation during the Karangat transgression is not supported by the stratigraphic record, however. Cores from the Sivash deposits described by Stashchuk et al. (1964) show a series of loess-like and paleosol sequences, but no characteristic marine deposits. Likewise, stratigraphic sections in the Perekop and Solenoye Ozero (Veklich and Sirenko, 1976) show no evidence of marine deposits, but loesspaleosol sequences. Topographic and tectonic factors must be considered in the postulated isolation of Crimea by the Karangat transgression. Accordingly, if the area had been tectonically stable since the time of the transgression, the 7–14 m rise above the present level (Fedorov, 1978; Chepalyga, 1984) would not have inundated the Perekop, where the highest elevations are around 20 m. The Perekop-Sivash area, however, is part of the IndoloKuban Depression, which is constantly sinking, a process that has resulted in the relatively recent inundation of the area and the formation of the Sivash. Therefore, during the Eemian Interglacial, the area should have been higher than it is today, and as a result, the hypothesis regarding the island of Crimea during the Karangat transgression lacks geological support. Sea-level highstands, however, should have created more salinization and a less attractive environment. Loess-Paleosol Sequences The areas with loess and the so-called loess-like deposits occur primarily on the Crimean Plains. Veklich and Sirenko (1976) proposed a chronological classification of Pliocene– Pleistocene silt deposits and paleosols of the western plains near the mouth of the Kacha, Al’ma, and Bulganyak Rivers—the Perekop area. Although the ages of the sequences in the Crimean Plains were not determined through absolute dating methods,

Late Pleistocene and Holocene paleoenvironments of Crimea these deposits and their associated paleosols have been correlated with broad time depositional phases and paleosol formation elsewhere in Ukraine. According to this scheme, the late Pleistocene consists of several pedocomplexes and loess (or loess-like) units that, in turn, are correlated with warm and cold stages, respectively. The Kaydaky soil unit is correlated with the Karangat transgression (Veklich et al., 1993). In the Crimean Plain, this unit is represented by a pedocomplex of two Chernozem soils. The next late Pleistocene soil unit—the Pryluky pedocomplex—includes the upper “chestnut” soil (a soil type of the present dry steppe) and the lower “braunerde.” The Vytachiv soil unit (coeval with the Surozh transgressive phase) is represented by specific “graybrown” soils of the dry steppe, frequently with gypsum and salts. The last late Pleistocene soil unit—the Dofinivka pedocomplex—includes incipient semidesert soils. The loess units that separate these pedocomplexes are generally no thicker than 2 m, with the exception of the Bug loess (coeval with the early Neoeuxinian regression), which reaches 10 m in thickness. Pollen data indicate cold dry steppe during the deposition of all loess units: meadow steppe, steppe, and forest-steppe of temperate climate for the Kaydaky and Pryluky units, and cool steppe climate for the Vytachiv and Dofinivka units (Veklich et al., 1993; Sirenko and Turlo, 1986). Correlation of units established by the Ukrainian stratigraphical framework with their respective marine isotope stages was attempted in later papers (Veklich, 1990; Veklich et al., 1993; Gozhik et al., 2000; Rousseau et al., 2001; Gerasimenko, 1999, 2001, 2006, 2007; Lindner et al., 2002). The scheme of Veklich et al. (1993) does not necessarily match those of other authors (Table 4), but it suits the late Pleistocene pedostratigraphy of the mountains and plains of Crimea. For this reason, this scheme is used in this paper. Alluvial Terraces The chronology of alluvial terraces has been a subject of debate since its preliminary classification by Andrusov (1912), whose work in the Sudak area resulted in a correlation between the drop of the sea in post-Karangat times and the formation of the late Pleistocene terraces. Later on, correlation of the Sudak

141

terraces with the river terraces north of the mountains led to a broader scheme of terrace sequences (Мuratov, 1974). This scheme is based on the numbering of terraces from the modern floodplain upward, beginning with terrace I, or the Garden Terrace (Sadovaya Terrassa), of terminal Pleistocene–early Holocene age. Novik (1990) attempted a chronology based on the weathering intensity of clasts in the sediments of the alluvial terraces of the Al’ma River Valley (Table 4). Years earlier, Kliukin and Shchepinskiy (1983) assigned relative ages to terraces II and III in the Al’ma River Valley using diagnostic Paleolithic artifacts. Terrace II has Upper Paleolithic material, while terrace III has late Mousterian (late Middle Paleolithic). Knowing that the boundary between these two lithic industries lies between 40 and 30 ka, the archaeological dates reported by Kliukin and Shchepinskiy (1983) match the dates estimated by Novik (1990) (Table 5). In addition to the lack of absolute dates, a series of problems haunts this chronology of alluvial terraces. These problems involve the local tectonic controls in each valley, which should be considered when cross-correlating terraces among river valleys. Other problems include gradients, catchment size, and elevation in relation to the changing sea levels. One example of the latter is the case of terraces along the Chërnaya River (a special section is presented herein), which presents only one clear terrace above the modern floodplain. It seems that the Chërnaya terraces potentially lie below its recent floodplain, which is barely above the present sea level. VEGETATION AND SOIL DEVELOPMENT Late Pleistocene Vegetation and Soil Development in the Mountains and Piedmont Quaternary stratigraphy and palynology in the mountains and the piedmont areas have been carried out in tandem with archaeological excavations of Paleolithic sites (Chabai et al., 1999, 2004; Gerasimenko, 1999, 2004, 2007). Although many sites are rich in faunal remains, few have pollen records or paleobotanical remains in general. The sites that have provided paleovegetation information are Kiik-Koba (M.N. Klapchuk in Stepanchuk, 2006), Zaskalnaya (Gubonina, 1985; Velichko, 1988), Grot Skalisty (Cohen et al., 1996), Kabazi II (Gerasimenko,

TABLE 4. CORRELATION OF LATE PLEISTOCENE SOIL SEQUENCES AND ISOTOPIC STAGES Unit (index) and equivalent marine Gozhik et al. (2001) Rousseau et al. (2001) Lindner et al. (2002) oxygen isotope stage Prychernomorsk loess (pc) 2 2 2 2 Dofinivka soil (df) 3 2 2 2 Bug loess (bg ) 4 2 2 2 Vytachiv soil (vt) 5a 3 3 3 Uday loess (ud) 5b 4 4 4 5a–c 5c–e 5a–c (or 5) 5 Pryluky soil (pl) Tyasmyn loess (ts) 6 5d (or 6 ) 5d 6 Kaydaky soil (kd) 7 5e (or 7 ) 5e 7 Dnieper loess/glacial unit (dn) 8 6 (or 8 ) 6 8 Potygaylivka soil (pt (zv3)) 9 7 (or 9 ) 7 9 Note: Based on Gerasimenko (2007). Stratigraphic framework is from Veklich et al. (1993)

142

Cordova et al. TABLE 5. TERRACE SEQUENCE IN THE AL’MA RIVER AT THE INNER RIDGE WITH WEATHERING-BASED RELATIVE DATES (NOVIK, 1990) AND ARCHAEOLOGICAL DATING (KLIUKIN AND SHCHEPINSKIY, 1983)

Terrace and incision phases Erosion phase High floodplain Erosional phase Terrace I (Garden Terrace) Erosional phase Terrace II Erosional phase Terrace III

Elevation from the channel (m) 0

Estimated age, weathering (ka)

1–2

6.0

1 Present 1

2–4

10.7 –10 .6

1

5–7

31.9

15–25

5 2. 3– 43.6

Erosional phase Terrace IV (Sudak Terrace) 35–45 82–78, 60 Erosional phase 240 , 209 , 151 –1 53 Terrace V (Manddzhil Terrace) Erosional phase Note: Terraces VI to VII are not included in table. MIS—marine oxygen isotope stage. *Broad lithic industry spanning from ca. 130 to 40 ka.

1999, 2005b), and Buran Kaya III (Gerasimenko, 2004). Pollen data in these sites show a conspicuous difference between west (e.g., Kabazi II and Grot Skalistiy) and east (e.g., Zaskalnaya and Buran Kaya) (Fig. 5). This eastward decrease in atmospheric moisture is also apparent in the modern distribution of temperature and precipitation (Fig. 3). Pollen and soil-sediment sequences from Kabazi II and Buran Kaya III provide a source for the reconstruction of vegetation change from the Eemian Interglacial to the end of the Pleistocene (Fig. 6). The Middle Paleolithic open-area site Kabazi II is located on the southern slopes of the Inner Ridge in western Crimea (44°50′N, 34°02′E, and 301 m elevation). The site is located 90 m above the bottom of the Al’ma River valley. The thick sequence of Pleistocene deposits occurs in a sedimentation trap, formed behind a rock slab that fell onto a slope bench. Buran Kaya III (Gerasimenko, 2004, 2007) is located on the northeastern part of the piedmont (45°00′N, 34°25′E; 250 m elevation) on the Burulcha River bank, 10 m above the present water level. It is a rock shelter with a southwestern exposure, and the valley was cut through limestone of the low external ridges of the mountains. Both sites are presently located in the mountain forest-steppe vegetation. The Quaternary Stratigraphical Framework of Ukraine has been used for stratigraphic subdivision in the Crimean Mountains. This framework was elaborated by a research team under the leadership of M.F. Veklich (Veklich et al., 1993). The suggested correlation of Ukrainian units with the chronological scheme of marine isotopic stages and the European paleoclimatic chronology is shown in Table 4 and Figure 6. The detected evolution of vegetation and climate in the western and eastern Crimean foothills, based on pollen and lithopedology, is described next. During the formation of the Kaydaky (kd) unit, a foreststeppe appeared under warm conditions in the western foothills

Estimated age, lithics (ka)

Malinovka I, middle Upper Paleolithic (33–28 ka) Mousterian* (Middle Paleolithic)

MIS

3

4–3

5a–4 7–6

of the Crimean Mountains. Carpinus-Quercus woodland alternated with meadow steppe (mesophytic type). The forest had a mixture of Ulmus (elm) and Tilia (linden) and well-developed shrub undergrowth. This vegetational type is similar to that presently found in the area. The wet lower part of slopes provided an environment for sedges. An abundance of sedges and ironreduction processes (evidenced in the greenish color of the sediments) indicate an excessive ground-moisture supply in the studied locality at Kabazi II. After that, broad-leaved forests declined to the point of total disappearance and were replaced by steppe. As this happened, the pine forest belt expanded, and birch appeared on the northern slope of the mountains. Xerophytic components appeared in the steppe assemblages, and clumps of shrub possibly existed apart from stands of trees. Reduction of broad-leaved species, appearance of birch, and expansion of pine indicate cooling. Presently, birch does not grow in Crimea (it sporadically occurs only in the highest mountain ridges). Later on, the forest-steppe was replaced by a sparse Pinus forest with admixture of broad-leaved trees and rich herb-fern ground cover. The climate was southernboreal, cooler than today. It is also evidenced by the appearance of Fagus, which presently grows in the higher forest belt of the mountains. Humus accumulation was weak under the light forest (A1B horizon of the Mollisol). Mollisols are generally formed under herb vegetation, or under a well-lit forest (with enough light to enable a rich herbal ground cover). Soil humification gradually increased to the end of the interval, along with reduction of the pine population and spread of Poaceae and Cyperaceae. At the end of the Kaydaky unit deposition, the forest-steppe vegetation type became established again. The woodland vegetation was of broad-leaved type (dominated by Quercus). The steppe vegetation was characteristic of the meadow steppe and meadow types, with abundant Cyperaceae and Ranunculaceae.

Figure 5. Late Quaternary stratigraphic sequences and localities mentioned in the text.

Late Pleistocene and Holocene paleoenvironments of Crimea 143

144

Cordova et al.

Figure 6. Pedopalynological and geoarchaeological sequence in the northern foothills of the Crimean Mountains: reproduced from Gerasimenko (2007) with modifications. Soils and sediments: (1) meadow soil (Mollisol), (2) Rendzina and brown soil (mollic and eutric leptosol), (3) Luvisol (slope derivative), (4) Cambisol (or pedosediments), (5) loess (light-yellow sediment), (6) light-gray loam, (7) coarse colluvial sediments. Paleoenvironmental dynamics: (8) western Crimea, (9) eastern Crimea.

Late Pleistocene and Holocene paleoenvironments of Crimea Humus accumulated much more strongly than before (A1 horizon of Mollisol). The climate became drier and slightly warmer than in the preceding phase. In sum, during the interval represented by the Kaydaky unit, a temperate stage of an interglacial was replaced by its post-temperate stage. The complete sequence of the Kaydaky unit is not recorded at Kabazi II—first, because the sedimentation started only when the trap was formed behind the fallen rock, and, second, because the upper soil limit is truncated. The first hominin occupation, presumably by Neanderthals, at the Kabazi II site happened during the formation of the A1B horizon of a Mollisol when sparse pine forest spread over the foothills. At the beginning of the Tyasmyn (ts) unit deposition, an intense erosional incision occurred at the site, followed by a thick colluvial accumulation. The colluvial deposits include brown clay beds, pedosediments that are absent in the underlying sedimentary sequence. This represents possible confirmation that the sedimentation break lasted for a long while, and so might correspond to a climatic shift. Phases of strong colluviation correspond with stadial conditions (Antoine et al., 1999; Haesaerts and Mestdagh, 2000). During the formation of the first Pryluky subunit (pl1b1), forest-steppe existed in the foothills, where the soil cover was Luvisols and their slope derivatives. The pollen counts of broadleaved species indicate relatively warm southern-boreal environments for the whole subunit, and the climate was temperate. Broad-leaved forests were dominated by Carpinus orientalis. The steppe associations were mesophytic. Subunit pl1b1 corresponds to a warm interstadial. Its final phase was marked by the appearance of Abies, which could indicate a downhill encroachment of mountain forest belts at the end of the interstadial. The next Pryluky subunit (pl2) was marked by a sharp decline of broad-leaved vegetation. The presence of birch and pine increased, as well as steppe vegetation. The boreal forest-steppe was established, and Artemisia began to appear, although in small numbers. Pedogenic processes were replaced by colluvial accumulation. This subunit corresponds to a stadial in which the climate was cooler and drier than during the preceding interstadial. During the formation of the pl3b2 subunit, humus accumulated under a southern-boreal forest-steppe. Birch and alder coexisted with other broad-leaved species (oak, hornbeam, and elm). Birch expansion indicates that the climate was cooler than today, and the subunit represents an interstadial. The transition to the next cold stage at the end of the interstadial (sharp reduction of arboreal and broad-leaved trees) was related to the final stage of soil formation—subunit pl3c. The characteristic feature of this subunit is the extensive distribution of hygrophytic vegetational components: alder and sedges. A decrease in evaporation, caused by the cooling, could possibly have led to excessive moisture supply at Kabazi II. The Mousterian cultures existed in Crimea during the formation interval of the Pryluky unit. During the deposition of the Uday (ud) unit, pedogenic processes ceased, and light colluvial material accumulated. This was a cold and relatively dry stadial revealing an absence of broad-

145

leaved trees and a prevalence of steppe. Boreal trees (pine, birch, and alder) formed groves. Xerophytization of the former meadow steppe is evidenced by the spread of Chenopodiaceae, Artemisia, Ephedra, Plumbaginaceae, and the vanishing of Polypodiaceae ferns in the western foothills of the Crimean mountains. During this same stadial, a xeric steppe occupied eastern Crimea (Velichko, 1988; Gubonina, 1985). At the beginning of the accumulation of the Vytachiv unit (subunit vt1b1), Cambisols and their slope derivatives developed under southern-boreal forest-steppe conditions in western Crimea. Although Pinus dominated the woods, Quercus and Carpinus were common. The appearance of Fagus might indicate that the higher mountain forest belt had moved lower. The climate was cooler than today, and the mesophytication of the herbal cover and reappearance of Polypodiaceae and Lycopodiaceae indicate an increase in precipitation. During the time of accumulation of subunits vt1b1-b2, light colluvial material accumulated instead of soil formation. The mountain foothills were still covered by forest-steppe, but herbal coenoses prevailed, and participation of xerophytes increased considerably (particularly Artemisia and Plumbaginaceae). This, as well as the disappearance of Polypodiaceae and Lycopodiaceae, indicates aridification, whereas the retreat of broad-leaved trees and reappearance of birch give evidence of cooling. Accordingly, subunits vt1b1-b2 corresponds to a stadial. In western Crimea, subunit vt1b2 is marked by a new expansion of broad-leaved trees and mesophytic herbs, and by the development of Cambisols and their derivatives. Pine and birch still prevailed in the forest, and the climate was southern-boreal, characteristic of an interstadial. The domination of woods over steppe, the great herbal diversity, and the spread of ferns indicate an increase in humidity by comparison with the preceding stadial (i.e., MIS 5b). The time of subunit vt1c was marked by a strong expansion of pine forest and a retreat of broad-leaved trees, some of which survived in refugia. Juniperus grew under the light pine stands, the ground cover consisted of mesophytic herbs, and Lycopodiaceae also appeared. The climate was cooler than during the preceding phase and evidently indicated a transition from the interstadial to the next stadial. In the eastern mountain foothills, the forest-steppe environments that existed during the formation of subunit vt1 were much drier than in western Crimea. Meadow steppes (herbetum mixtum and Cyperaceae) strongly dominated over woods, which consisted of boreal trees and a few Tilia. Rock shelters were occupied by diverse spore plants, particularly green mosses. During the formation of subunit vt2, typical grassland was established both in western and eastern Crimea. The woodland was drastically reduced, particularly at the end of the subunit. The western foothills witnessed the first appearance of a steppe vegetation type, and at the end of the interval, the tree population was represented only by alder in the river valleys. In the eastern foothills, trees seem to have disappeared completely. The extensive distribution of Poaceae, xerophytes (Chenopodiaceae, Plumbaginaceae, Artemisia, Ephedra), and Asteraceae occurred.

146

Cordova et al.

Hygrophytes and spore plants grew only in rock shelters and depressions. At the end of the subunit, xerophytes also expanded to the formerly well-wetted localities. Fern and sedge populations strongly declined. The absence of broad-leaved trees and the spread of Lycopodiaceae and Botrychium boreale indicate the boreal climate of this dry stadial, marked by strong colluviation and a complete suppression of pedogenic features. The steppe climate became colder and progressively drier from the beginning to the end of the interval. During the deposition of subunit vt3b, eutric Leptosols were formed, and broad-leaved species reappeared in the woods, particularly in western Crimea. There, a mixture of hornbeam and oak expanded much more extensively than in eastern Crimea, where mesophytic steppes strongly dominated and only a few Quercus pubescens grew in better protected places. The appearance of broad-leaved species and the retreat of boreal trees (particularly Betula) and of boreal spore plants (Lycopodiaceae and Botrychium) indicate an interstadial climate. Xerophytes almost did not exist in the herbal cover during the whole interstadial in western Crimea and at its beginning in eastern Crimea, a condition that indicates relatively high moisture. At the end of this interstadial (subunit vt3c), an extensive advance of steppes and an intense suppression of broad-leaved trees marked the transition to the next cold stage. Thus, during Vytachiv unit times, three interstadials occurred, separated by two stadials. During the first half of the Vytachiv unit, Middle Paleolithic industries occurred in the Crimean Mountains, and starting from the stadial vt2, they coexisted with the Upper Paleolithic occupations in the area (Chabai, 2005). During the formation of this unit, loess-like loams accumulated in the mountain foothills under typical grassland conditions with the substantial participation of xerophytes. Arboreal vegetation was drastically reduced. No broad-leaved species occurred in the valley woods. This indicates the dry boreal climate of a stadial. In western Crimea, Alnus, Betula, and Rhamnus grew in the river valleys, but ferns were completely absent. Drought-resistant and heliophytic plants of the Elaeagnaceae family appeared, indicating the existence of a vast open landscape. In eastern Crimea, at the beginning of the Bug (bg) unit, herbetum mixtum, especially Lamiaceae and Asteraceae, still prevailed in the herb cover, birch and alder trees framed the river, and Pteridae grew in rock shelters. In the second part of the stadial, the area was treeless. Spore plants were dominated by mosses and included boreal elements (club-mosses and grape-ferns). The vegetation assemblages for cold stages (e.g., MIS 4 and MIS 2) show a reduction of broad-leaved species in favor of open steppe, with the prevalence of forest-steppes in some areas (Fig. 6). The presence of boreal trees was evident, but not to the point of dominance. Conversely, the warm stages (e.g., MIS 5e, 5c, 5d, and 3) show the expansion of tree vegetation in the form of forest and forest-steppe communities. The scheme of soil and vegetation change just described refers to the foothills region. No data exist for the higher elevations of the mountains, where environmental conditions and veg-

etation communities during stadials and interstadials can only be inferred. The relict forms of periglacial processes suggest an environment influenced by low temperatures (Podgorodetskiy, 1988), yet the lack of cirque landforms and tills indicates that ice caps and glaciers did not form. Furthermore, some habitation sites are found on the Demerdzhii Yaila at high elevations (often above 800 m), suggesting that conditions may not have reached a full glacial environment. Given the relatively mild conditions of the Crimean Mountains during the coldest periods, ideas for the possible existence of glacial refugia of temperate and boreal trees in Crimea have been put forward based on the presence of haploid types in Fagus (Comes and Kadereit, 1998), stands of relict boreal species in the modern landscape (Morozova and Kozharinov, 2001; Cordova et al., 2001), and modeling and simulation of glacial temperatures (Leroy and Arpe, 2007). The Buran Kaya III and Kabazi II pollen diagrams (Gerasimenko, 2007) show basically an absence of temperate species on the northern slopes. Carpinus is present in small amounts in Kabazi III, but it is more likely to be the more cold tolerant Carpinus betulus. It is probable, however, that glacial refugia for temperate trees were on the warmer southern slope of the mountains. Unfortunately, there are no dated paleobotanical sites on the southern slopes, but the existence of refugia on south-facing slopes of valleys dissecting the northern slopes of the mountains is possible. A study of this aspect, however, requires the use of temperature modeling on slopes using data drawn from the existing pollen record. Faunal assemblages changed in step with vegetation and climate changes. During the coldest stages of the glaciation, the Crimean Mountains had a distinctive mammalian assemblage, distinct from other mountainous regions around the Black Sea (see Markova and Puzachenko, 2006). During the early Weichselian and Denekamp Interstadials, the mountains supported an assemblage represented by woolly mammoth, woolly rhinoceros, wild horse, saiga, red, roe, and giant deer, mountain sheep and goat, cave bear, cave hyena, and lemmings, among others. The Last Glacial Maximum was characterized by European ass, saiga, red and giant deer, northern mole-vole, and steppe and yellow lemmings (Markova and Puzachenko, 2006). During the same period, faunas in the steppes seem to have been dominated by boreal species, strongly represented by woolly mammoth, woolly rhinoceros, reindeer, primitive bison, saiga, arctic fox, cave hyena, and arctic hare. Overall, during these three periods, steppe and forest-steppe fauna dominated, coinciding with the vegetation communities demonstrated by pollen data (Fig. 6). Although boreal and even tundra species existed, they seem to have been of less importance, or perhaps animals associated with the coldest part of the mountains or seasonal immigrants from the boreal areas to the north (Burke et al., 1999). Pleistocene-Holocene Transition After the LGM in the western foothills of the Crimean Mountains, conditions began to favor trees over steppe, as evidenced by

Late Pleistocene and Holocene paleoenvironments of Crimea pollen records from Grot Skalisty (Cohen et al., 1996; Gerasimenko, 2005a). On the eastern foothills of the mountains, however, conditions remained steppic, as evidenced by pollen from zone X (level 5) in Buran Kaya III (Gerasimenko, 2004). Pollen records from the west and east show fluctuations to mesic conditions indicated by a peak in arboreal vegetation consisting of Carpinus (hornbeam), Tilia (lime), and Fagus (beech), suggesting mesic bioclimatic conditions during the Allerød (11,800– 10,900 radiocarbon yr B.P.; 13,900–12,800 cal. yr B.P.). For this interval, pollen records from the Heraklean Peninsula provide evidence for dense forests of Quercus (oak) with a significant presence of Ulmus (elm) and Corylus (hazel), which reflect mild and wet conditions (Cordova and Lehman, 2005). The onset of the Younger Dryas in the Heraklean Peninsula represented a decline in arboreal vegetation (Fig. 7). A transition from a brown forest soil to a meadow Chernozem in geoarchaeological sections parallels pollen evidence for steppic vegetation (Cordova and Lehman, 2005). Reduction of arboreal vegetation occurred also in the western foothills of the mountains, although Rhamnus and Corylus pollen still appeared in Grot Skalistiy (Gerasimenko, 2005a). Farther east, at Buran Kaya III, conditions indicate grass-steppe vegetation. After the Younger Dryas, conditions improved in the foothills of the mountains as the forest-steppe vegetation appeared, even around the eastern side of the foothills at Buran Kaya III (Gerasimenko, 2004). In the Heraklean Peninsula, however, conditions after the Younger Dryas never reverted to forest. Steppe conditions with a few shrubs persisted through the Preboreal and Boreal periods, until roughly 7.4 ka. This development was interpreted by Cordova (2007) as a result of a dry and continental climate maintained by the still low sea level, which continued to influence effective precipitation by an increased continentality. Although the sea level was low during the moist Allerød, lower temperatures, and consequently low evaporation rates, kept conditions mesic. Another explanation might be that the effects of the Westerlies were not strong enough in the southern part of Crimea during this time. It is not until 7.5–6 ka that arboreal pollen appears in larger frequencies in the Heraklean Peninsula (Fig. 7). This development might have been associated with the return of high water levels in the Black Sea, i.e., near modern levels, as well as the overall moister conditions achieved in the region during the Atlantic stage, as evidenced in other areas of southern Ukraine (Kremenetski, 1995; Gerasimenko, 1997). The Pleistocene-Holocene transition encompassed the Mesolithic period, which is strongly represented in practically all regions of the Crimean Peninsula (Bibikov et al., 1994; Olenkovsky, 2000). However, most of the paleoenvironmental and cultural information for the Mesolithic comes from sites in the mountains (e.g., Shan Koba, Fatma Koba, Laspi; Fig. 5), where many lithic and animal bone materials have been recovered (Bibikov et al., 1994; Cohen et al., 1996). Mesolithic sites in the mountains provide macroscopic floral and faunal remains that suggest forest-steppe and forest environments, quite opposite those of the predominantly steppe condi-

147

tions of the Upper (Final) Paleolithic (Bibikov et al., 1994). The main aspect in the faunal remains lies in the gradual disappearance of the saiga (Bibikov et al., 1994, p. 167). Unfortunately, beyond the Buran Kaya III and Grot Skalistiy pollen records, little is known of vegetation during the Mesolithic of Crimea. The number of archaeological sites in the Sivash area, as well as the lithic and faunal material recovered in them (Olenkovsky, 2000), suggests conditions more attractive than today’s semiarid climate and salinized soils and water. The early Neolithic cultures of Crimea developed out of the Mesolithic Shan-Koba and Murzak-Koba cultures and appear no later than 8 ka (Cohen, 1996). This emergence coincides with the Boreal period and with cool and dry—but gradually improving— conditions in the southwest. At section BBBP-2, Mesolithic Murzak-Koba lithics are associated with a Chernozem soil and continuing steppic conditions in the Crimean Peninsula (Fig. 8). The later Neolithic cultures, however, had the benefit of relatively warm and moist conditions during the Atlantic period (Fig. 7). The Neolithic meant the beginning of farming and pastoralism, which should be seen as the beginning of a deeper vegetation transformation. The increase of phrygana vegetation in pollen records between 8 and 7 ka (Fig. 7) suggests possible impacts of pastoral activities (Cordova and Lehman, 2003, 2005, 2006). However, the increase in Mediterranean elements in the vegetation assemblages and the possible shift to a dry summer may have played a role in this development. On the other hand, accumulation of sediments in the balkas of the Heraklean Peninsula increased as soils became redder (cinnamonic soils) (Fig. 8). Whether this was the result of climatic shifts or the impact of pastoralism, or a combination of both, is a matter that deserves further research. Middle-Late Holocene in Southwestern Crimea Southwestern Crimea Results obtained through geoarchaeological and palynological studies in southwestern Crimea offer a chronological model that links vegetation-soil development and climate change (Fig. 7). The rise of temperature and moisture that occurred between 7.4 and 6 ka brought about an increase in arboreal vegetation. The apparent reduction of tree cover and increase in nongrass plants suggest a dry phase that peaked between 4.5 and 4 ka. The strong presence of Lamiaceae (mint family) associated with this period suggests that pastoralism may also have been involved in the deterioration of vegetation. Although the pastoral influence is possible, records from the Ukrainian mainland show a drying trend during this period (Kremenetski, 1995; Gerasimenko, 1997). Moist conditions returned by 3.5 ka, and the trend seems to have persisted until around 2 ka (approximately the beginning of the first century A.D.). The forest expansion that took place due to this moist period shows evidence of an interruption: a reduction of tree vegetation coinciding with the establishment of ancient Greek farms (Fig. 7). Pollen records from the floodplain

Figure 7. Pedopalynological composite section for the Heraklean Peninsula and Lower Chërnaya (based on Cordova and Lehman 2005). See Table 6 for designations of Mediterranean, sub-Mediterranean shrub, and Phrygana taxa. PC—post-classic chernozem; MR—meadow rendzina; CC—calcic cinnamonic; BC—brown cinnamonic; Ch—chernozem; BF—alluvial broen forest soil; GS—gray alluvial soil.

148 Cordova et al.

Late Pleistocene and Holocene paleoenvironments of Crimea

149

TABLE 6. POLLEN TYPES IN THE GROUPS OF THE POLLEN DIAGRAMS Group Sum of the following pollen types and taxa Broad-leaf trees Quercus, Fagus, Carpinus, Corylus, Pistacia, Tilia, and Ulmus Aquatic and subaquatic Cyperaceae, Typha, Juncus, Sparganium, and Potamogeton Phrygana Cistaceae, Lamiaceae, Asphodeline, Sarcopoterium, Thymelea hirsute Shiblyak Quercus pubescens-type, Pistacia, Jasminum, Paliurus, Cotinus Mediterranean Pistacia, Jasminum, Thymelea hirsuta, Sarcopoterium, Asphodeline, and Cistaceae Sub-Mediterranean Quercus pubescens-type, Cotinus, Cornus, Paliurus, and Verbascum Note: From Cordova and Lehman (2005).

Figure 8. Holocene pedogeoarchaeological sections in ancient farming territories in southwestern Crimea and the Kerch Peninsula. Reproduced from Cordova and Lehman (2005) with modifications and additions. See paleosol types (e.g., MR, PC, CC, MC) on Table 7.

150

Cordova et al.

TABLE 7. PALEOSOL TYPES AND THEIR PEDOGENIC HORIZONS IN THE SECTIONS OF THE HERAKLEAN PENINSULA Soils and paleosols Abbreviation Key characteristics Approximate time span Meadow rendzina MR Organic horizon developed on colluvial material. 3 ka to present 10YR 3/3-4/3-5/2 Medieval to modern Post-Classic Chernozem

PC

Organic horizon with low to moderate carbonate development. 10YR 4/1-5/2

Roman to modern

Calcic cinnamonic

CC

Truncated AB horizon and Bwk (carbonate filaments) horizon with moderate blocky to columnar structure. 10YR 5/1-5/3-5/4

4.2–3.5 ka Bronze Age (on top of soil)

Meadow cinnamonic

MC

A horizon on silty loam with reduction marks. No structure. Dark gray-brown color 10 YR 3/3-4/3

5.5–4.2 ka

Brown cinnamonic

BC

Poorly developed A horizon. Poorly developed Bw (cambic) and Bk (calcic) horizon. Brown color. 7.5YR 4/2-4/4, 10YR 4/3

7.5–5.5 ka

Meadow Chernozem

Ch

Partially truncated dark A horizon on a Bk (carbonate horizon). Massive to weak blocky structure. 10YR 4/1-4/2

10.5–7.5 ka Mesolithic, Murzak Koba lithics (bottom of soil)

Alluvial brown forest

BF

Dark brown A horizon developed on rapidly aggrading silts. Moderate blocky structure. 10YR 4/2-5/3-5/4

12–10.5 ka

Alluvial gray soil

GS

Poorly developed soil on silts and sand. No structure. High gravel content. 2.5 Y 5/4-6/4

12 ka or older

Note: From Cordova and Lehman (2005).

wetlands in the lower Chërnaya Valley show the appearance of cultivated grasses and an increase in weeds and phrygana elements during the establishment of the Greek rural economy during the second half of the first century B.C. (Cordova and Lehman, 2003). Ironically, the impact of Greek farming had little or no effect on rapid sediment accumulation in the ravines of the Heraklean Peninsula and the Chernaya Valley. This contrasts significantly with the period of rapid accumulation presumably associated with the Neolithic. Archaeological records show a decline in farming during the last two centuries B.C. The abandonment of farms occurred during a dry phase that is not represented in the pollen record due to its short duration, but it appears in some geomorphological records discussed later herein. Nonetheless, pollen records show an increase in shrub and tree pollen during the first half of the first millennium A.D., suggesting a decrease in farmed areas (Cordova and Lehman, 2003). During the past 1500 yr, the pollen record gives poor resolution of any changes in vegetation; however, the trend during the last millennium indicates an increase in phrygana and sub-Mediterranean species (Fig. 7). The Yaila and the Mountains The presence of a steppic landscape surrounded by forests on the summit of the Crimean Mountains has prompted biogeographers to propose several hypotheses to explain the treeless nature of the Yaila. In an exhaustive bibliographic compilation

of these hypotheses, Artiushenko and Mishnev (1978) were able to single out at least two main groups of hypotheses. One group states that the Yaila was forested, but gradually lost its cover to pastoral nomads who eliminated the forest over time through burning to create summer pastures. The second group states that the Yaila was always a steppe resulting from local climatic conditions of cold and strong winds, and the hydrological characteristics of its karstic nature. These contrasting views about the origin of the Yaila vegetation motivated Artiushenko and Mishnev (1978) to investigate the case using pollen analysis from soil profiles in each of the Yaila subdivisions: Ai-Petrinskaya Yaila, Yaltinkskaya Yaila, Demerdzhi Yaila, Babugan Yaila, and Dolgorukaya Yaila. Depths of the soil profiles rarely reached 50 cm, and no radiocarbon dates were provided. However, the soil profiles were deemed to have spanned the entire Holocene. Артюшенко and Мишнев’s pollen data from the soil profiles showed that, for the most part, the frequencies of arboreal pollen were minimal, ranging from 10% to 40%. Considering the possibility of wind-transported tree pollen, the amounts of tree pollen, although low, were still an overrepresentation of the tree cover on the Yaila. Artiushenko and Mishnev (1978) eventually favored the climatic reasons for the lack of forests, concluding that the Yaila had been treeless for the entire Holocene. In fact, winds are stronger here than anywhere in Crimea (Artiushenko and Mishnev, 1978), and cryogenic soil processes are common (Вахрушев and Клюкин, 2001), both of which curb

Late Pleistocene and Holocene paleoenvironments of Crimea tree development. Overall, the conclusion points to the existence of an Alpine-type of meadow, as opposed to a deforested area turned into meadow. In reality, small pine trees are scattered in the area, as well as planted trees. Our 2 m core from a depression in the Yaltinskaya Yaila provides a pollen record that further contributes to the paleoecological information of the Yaila (Fig. 9). The bottom date is 7823 ± 93 radiocarbon yr B.P. (cal. 8524–8837 yr B.P.). Throughout the core, the frequency of arboreal pollen is never above 35%, except for two samples: 65% and 57%. A large percentage of the arboreal pollen sum is pine, a characteristic present also in the profiles by Artiushenko and Mishnev (1978), who assumed that most of it was blown in from the areas around the Yaila. Pine pollen is produced in abundance and usually carried very easily by the wind, to the point that large amounts of pine pollen appear even in modern samples of the Heraklean Peninsula, where there are no pine trees (Cordova and Lehman, 2003). In general, the results of our study support the idea of a steppic Yaila at least during the past 8000 yr. Peaks of arboreal vegetation are evident at 50 cm depth. Unfortunately, the dates of this part were reversed. Nonetheless, the peak occurs after 2000 yr B.P., suggesting that this is a climatic event of the subAtlantic that correlates with high arboreal pollen data from the Crimean Peninsula and the Chërnaya Valley during the early centuries of the first millennium A.D. (Cordova and Lehman, 2005). An alternative possibility could be the warm conditions attained during the Medieval Climatic Anomaly (i.e., the Medieval Warm Period), but no other record in the Crimean Peninsula attests to the development of this event. Unfortunately, neither the cores nor other pollen records for the past 2000 yr have the chronological resolution to ascertain events of centennial duration. The Yaltinskaya Yaila pollen diagram, however, shows interesting paleoenvironmental information pointing to possible human influences, namely pastoral groups. The first clue for such potential modifications lies in the accumulation rates and types of sediment. While low sedimentation rates occur between 8000 and ca. 4000 yr B.P., particularly in silt-dominated sediment, afterward this time, sedimentation rates increase, particularly with the inclusion of clay, sand, and small gravel. This change in sedimentation rates (ca. 3500 yr B.P.) coincides with an interval of favorable climate and an increase in herding, which may have led to overgrazing and degradation of meadows on the ridges. In order to explore this hypothesis, we are carrying out a series of tests, including magnetic susceptibility, organic matter, opal phytoliths, burnt grass phytoliths, and microscopic charcoal. THE DYNAMIC ENVIRONMENT OF THE HOLOCENE Soils and Geoarchaeology The geoarchaeological study of the Heraklean Peninsula has shown how soil genesis can impact geoarchaeological and paleoecological interpretation (Figs. 7 and 8). A decline in arboreal pollen (Fig. 7) and a shift from brown forest soil to Chernozem in

151

the BBBP sections (Fig. 8) mark the transition from the Allerød to the Younger Dryas. This Chernozem soil appears also in sections MM, AA, and BP. Steppe conditions persisted throughout the Boreal stage; however, the soil and pollen record lack sufficient resolution to reveal changes during the Younger Dryas, Preboreal, and Boreal periods, for which the record still shows semiarid, cool, continental conditions (Cordova and Lehman, 2005; Cordova, 2007). The transition from the dry and cool conditions represented by the Chernozem to a moist-warm marine sub-Mediterranean environment is marked by a reddish brown (cinnamonic) soil dated to the Atlantic Period. The cinnamonic soils in the Black Sea region of the former Soviet Union represent warm and dry summer conditions, as do the Mediterranean terra rossa soils (Gerasimov, 1954). Carbonate accumulations in cinnamonic soils increase toward the top of the sequence, indicating that a decrease in moisture occurred between 4.5 and 4 ka. Moist conditions returned before 3 ka, after which an organic-rich meadow soil and Rendzina appeared, persisting until the present (Cordova, 2007). The studied section at the Tarkhankutskaya Balka in the Kerch Peninsula shows somewhat different soil and geomorphological developments (Fig. 8). Its location in the east of Crimea, away from the moist conditions provided by the Westerly winds, is a clue to understanding climatic changes in the area. Here, modern vegetation and soils suggest persistent steppe conditions and soil salinization. Paleosols typical of brown forest and cinnamonic types are absent in this region. Despite the persistent dry conditions, periods of dry and wet fluctuations are evident in the stratigraphy of the Tarkhankutskaya Balka section. The series of alluvial sediments and soils, which can be associated with several occupations, are divided into four stratigraphic units (Fig. 8). Unit I accumulated at a very slow rate, contrary to what occurred in the Heraklean Peninsula during the past 3000 yr. Floodplain sedimentation increased between the tenth and fourth centuries B.C., evident in the accumulation of unit II. This unit represents a rapid though steady accumulation that had short periods of stability until finally stabilizing in the fourth–third centuries B.C. Unit III is represented by floodplain deposits capped by a dark soil, dated to the third century B.C. Unit IV is an alluvial fill deposited after a period of floodplain erosion. Afterward, a second floodplain erosional phase occurred, shaping the valley into the form we see today. Unit I is characterized by a clayey, organically rich deposit with large amounts of cultural material, altogether suggestive of wet and stable conditions. Unit II represents unstable geomorphic conditions with persistent floods, suggesting drastic wet-dry fluctuations. As evidenced by the regional pollen diagrams, the period of rapid accumulation of init II coincides with a transition from drier to wetter conditions, that is, the transition from the Subboreal to Early Subatlantic (Kremenetski, 1995, 1997, 2003; Gerasimenko, 1997; Cordova and Lehman, 2005). During this period, the site was occupied by a native group from the steppes, which was the first to enter into contact with the Greek colonizers

Figure 9. Yaltinskaya Yaila section with pollen frequencies. AP—arboreal pollen; NAP—nonarboreal pollen.

152 Cordova et al.

Late Pleistocene and Holocene paleoenvironments of Crimea of Panticapaion. The pottery found atop unit I can be assigned to the indigenous Late Bronze Age ceramics. The top of unit II is marked by stable and presumably moist conditions with a strong presence of fourth and third century B.C. Greek pottery. The moment coincides with the expansion of farmland within the Greek Bosporan Kingdom centered at Panticapaion (Fig. 9). The improved climatic conditions of this time are evident in the formation of meadow soils on the Heraklean Peninsula (Fig. 8). It is possible that the influence of these moist conditions may have reached the Kerch Peninsula. Records from the steppes of eastern Ukraine also show a wet climate during this period (Gerasimenko, 2007). Unit III contains water-reworked pottery of the third century B.C. This implies that pottery fragments were transported out of eroded sites and deposited by floodwater at this location, suggesting slope instability and/or perhaps site abandonment. Archaeological records show that at the end of the third century B.C., entire farming areas were abandoned in the city state territories of the Kerch Peninsula (Maslennikov, 1997) and on the western coast of Crimea (Shcheglov, 1978). Floodplain erosion ensued after the deposition of unit III. Pottery of the third century B.C. is embedded in the deposits of unit III, presumably as material removed from abandoned sites. A second incision occurred, eroding the deposits of units I, II, and III, and this event may have occurred soon after the occupation (third century B.C.), suggesting that this may have been a period of dry conditions that favored erosion and incision. An event of erosion and floodplain incision occurred again, followed by the accumulation of unit IV, which contains seventh–twelfth century A.D. pottery. The modern soil formed on this unit, and since then, no incision or accumulation has occurred in Tarkhankutskaya Balka. The geoarchaeological value of soils and sediments lies not only in the information they contain reflecting the conditions of soil genesis and their influence on human occupations, but also the sedimentation/erosion events that may have been related to nonclimatic events. Thus, rapid sedimentation seen in the sediments of the Crimean Peninsula, although related to climatic changes, may also be connected to the beginning of landscape transformation by farmers and herders of the Neolithic and later periods (Cordova and Lehman, 2003, 2006). The Bronze–Iron Age inhabitants of Crimea, or the Tauri of the Greek sources, practiced farming and livestock rearing (Sorochan et al., 2000). The Greek colonists who arrived in Crimea beginning in the seventh century B.C. introduced a specialized agricultural system adapted from their Mediterranean homeland. Beyond the Greek agricultural districts, the indigenous agricultural system continued to be practiced for centuries. It was not until the fourth and third centuries B.C. that agriculture expanded beyond the farming districts of the city states, or chora. Two main areas were eventually developed, one in the east led by the cities of Panticapaion, Nymphaios, and Theodosia, and one in the west led by the city of Chersonesos (Fig. 10). The collapse of agriculture in the western part of

153

Crimea seems to have occurred at the end of the third century and beginning of the second century B.C. Archaeologists argue that the main causes of decline were political and military, particularly in view of the strength of the Scythian seminomadic groups of the steppe (Shcheglov, 1978; Sorochan et al., 2000). Lower Chërnaya Valley Another area of focus regarding Holocene geomorphological change is the lower Chërnaya Valley (Fig. 11), which corresponds to the lower reaches of the river basin and the depression that forms Sevastopol Bay. The depression is formed by a fault system that cuts through sediments of Eocene to Miocene age (Fig. 2A). Marine terraces of unknown age border the northern flank of the valley, in the vicinity of Inkerman, properly northwest of the Kalamita Fortress (Fig. 11). At a lower elevation, an alluvial terrace with carbonate sediments borders the basin north and south of the SB-1 location. The degree of carbonate density is suggestive of a Pleistocene age; however, its position within the Pleistocene alluvial terrace scheme (Table 4) is difficult to determine due to a lack of dating. Two cores were taken from the middle and upper part of the floodplain (Fig. 11). Core NG-2 represents mainly overbank flood silts, in some cases associated with soils. A paleosol is clearly visible at the bottom of the sequence, and cumulic soils are visible in other areas. Pollen data from the core reveal a wetland environment, which became stagnant water in levels below the aforementioned clay deposit (Cordova and Lehman, 2003, 2005). The bottom of the sequence was dated at 5380 ± 40 radiocarbon yr B.P. (6088–6259 cal. yr B.P.). Core SB-1 is 460 m long, but at its bottom, a paleosol is dated at about the same age as the bottom paleosol in NG-2 (Fig. 11). Although the two cores seem to cover the same time span, their stratigraphy is somewhat different. The most conspicuous difference is a thick layer of horizontally bedded clay, which is thicker in SB-1, suggesting changes in the preexisting topography during the deposition of this unit. These two cores yield only limited insight into the basin’s Holocene stratigraphy; it is evident that the basin possesses a deeper column of Holocene sediments. Previous engineering cores in the Chërnaya have reported 11.5 m of sediment at Chernomorskoye, just west of SB-1, and 40 m from the mouth of the river. This also shows that sediment accumulation may be associated with sea-level change and tectonics, which is one reason why deeper cores are needed. It is not clear what the effects of late Holocene sea-level change were on the basin, but this is a topic for future study, particularly linking some of the late antique sites in the area. Paleosols, such as the one at the bottom of the sequence, may also suggest buried archaeological sites, which is another geoarchaeological topic to pursue in this area. The relation between sedimentation and late Quaternary sea-level change in the basin becomes complicated if periods of alluvial sedimentation are factored in. Therefore, a project of this

154

Cordova et al.

kind will require extensive coring and digging to study a longer Quaternary sedimentation history.

bottom of the sequence had a paleosol that provided a radiocarbon date of 5597 ± 40 yr B.P. (AA-48302). This finding suggests that terrace I has been partially buried by recent floodplain deposits.

Other Holocene Floodplains Geoarchaeology of Laspi Bay The floodplains of the two major rivers north of the Chërnaya Valley have different morphology, due to their geological and topographic conditions. Unlike the deep graben of the lower Chërnaya (Fig. 2A), these rivers flow through a tilting platform that has caused migrations. The Holocene floodplain sediments of the lower Bel’bek Valley are relatively shallow compared to the lower Chërnaya Valley. In some areas, channel incision has exposed modern sediments that lie barely above the bedrock. The lower Kacha floodplain seems to have a longer history of sedimentation, although no channel cut exposures exist. A trench across the floodplain of the Kacha River next to the Sevastopol-Yevpatoria highway exposed a 2 m sequence of floodplain deposits, however. The top of the sequence had overbank flood deposits associated with twentieth century material. The

Lack of soil formation in geoarchaeological profiles usually means rapid accumulation and landscape instability dominated by slope sediments and wave action. The section exposed on the cliffs of Laspi Bay on the southern coast shows an example of unstable slopes, human occupations, and sealevel changes (Fig. 12). The area was originally studied by Firsov (1972), who obtained the first radiocarbon dates. Although originally the deposits of this section were classified by Firsov into lower and upper terraces, a more detailed study has discovered four Holocene terraces. Because the term terrace in this environment may portray the idea of marine terraces, we use the term “surface” to denote the terrace surfaces originally proposed by L.V. Firsov.

Figure 10. Ancient Greek settlement and areas of farming. Figure is after Kryzhitskii (1997) with modifications by the authors.

155 Figure 11. The lower Chërnaya and sections NG-2 and SB-1.

Late Pleistocene and Holocene paleoenvironments of Crimea

Figure 12. Section at Laspi Bay.

156 Cordova et al.

Late Pleistocene and Holocene paleoenvironments of Crimea Surface 1 (S1) corresponds to units 5 and 6, which represent the middle Holocene deposits buried by a late Holocene landslide. The remains of human occupation are associated with the deposition of unit 6 (layer A), which lies on surface 2 (S2). Radiocarbon dates obtained from charred wood and ashes produced ages between the twelfth and the seventh centuries B.C. (Table 3). This dating supports the analysis of pottery associated with it, which is classified as Bronze Age (Firsov, 1972), or pottery that precedes the Greek occupation of southwestern Crimea. A series of three landslide deposits (units 7, 8, and 9) subsequently formed, overlying the occupation layers. A tree log embedded in unit 9 produced a radiocarbon date that placed the landslide deposit in the seventh century B.C. (Table 4). This indicates that the landslides occurred during the occupation or shortly after its abandonment. The surface created by a hiatus in landslide accumulation (unit 9) corresponds to surface 3. Later on, gully erosion dissected this surface. In one of the valleys, alluvial deposits of unit 10 accumulated to create surface 4, which was occupied during and after accumulation during the Middle Ages. One of the buried surfaces of this period corresponds to layer B. After the incision of surface 4 by ravines, accumulation resumed to form the modern alluvial deposits (unit 11). Besides the rapid geomorphic events associated with landslides, stream erosion, and accumulation, the sea-level change is recorded in surface 1, which was originally described by Firsov (1972). Because the occupation surface is barely a meter above the modern sea level, the occupation may have occurred during a regressional phase, most likely the Phanagorian regression. The units below the occupation, namely 5 and 6, consist of a mixture of colluvial and marine deposits, indicating most probably a higher sea-level stand than the preceding transgression. Sea-Level Fluctuations Reconstructions of coastal change have been developed for numerous areas of the Black Sea. The only one that takes data from the Crimean shores is the one published by Shilik (1997), which is supported by evidence collected only from archaeological sites on the coast. Visible shoreline changes in Crimea are not reported for periods prior to 6 ka, since most of the evidence is underwater. In this region, sea-level change information in the early Holocene comes from the boreholes on the northwestern shelf and limans (Shcherbakov et al., 1976; Gozhik et al., 1987), in the Kerch Strait (Fedorov, 1978), and the southern coastal shelf and the Kerch Strait (Shcherbakov et al., 1977). Coastal and alluvial deposits found at various depths show the low levels that preceded the Black Sea transgression prior to 6 ka. The fluctuations after 6 ka are better known through evidence in coastal sediments and erosional benches, where the position of archaeological sites has often been used as a chronological reference. These fluctuations have been arranged into a series of regressions and transgressions, the magnitude of which fluctuates within the range of 8 m. Thus, the general

157

scheme includes the Old Black Sea transgression, the Phanagorian regression, the Nymphaean transgression, and the Medieval regression. The main problem is that the constructed curves and phases present discrepancies in the timing and magnitude of the sea-level fluctuations (for example, the curves proposed by Fedorov [1978], Chepalyga [1984], and Shilik [1997]; Fig. 13). Федоров suggests fluctuations of no more than 2 m above and below modern sea level for the past 6 k.y, whereas Shilik shows up to 8 m of sea-level drop during the Phanagorian regression, and according to Chepalyga, the Phanagorian regression dipped to more than 10 m below the present level. Models created for the Kuban and Danube deltas by Kaplin and Selivanov (2004) and Giosan et al. (2006), respectively, show that during the past 5 k.y., the deltas have evolved under the influence of a relatively stable sea with fluctuations within +1.5 and –2.0 m. The curves by Fedorov (1978), Chepalyga (1984), and Shilik (1997) are much simpler than the curve of Balabanov (2007), which shows fluctuations within the major transgressions-regressions. Balabanov established stages (eustatic events on the order of millennia), and phases (fluctuations on the order of centuries) in the history of the Holocene sea-level change. Giosan et al. (2006) attributed these regional discrepancies in sea-level curves to the differences in tectonic activity and hydrostatic rebound following the Holocene infilling of the Black Sea basin. Accordingly, sea-level reconstructions should focus on data obtained locally in order to recognize properly the role of tectonics and local conditions in sea-level correlations. Fedorov’s (1978) and Shilik’s (1997) curves show parallel changes. When compared with Balabanov’s (2007) stage-phaseregression chronology, Shilik’s curve shows a correlation with the regression at the end of the Kalamitan phase (ending around 4000 B.C.) and with the regression following the Old Black Sea transgression. The Phanagorian regression seems to reach its lowest level around 500 B.C. in Balabanov’s scheme and around 100 B.C. in Shilik’s curve. The Nymphaean transgression in Balabanov’s Nymphaean stage is interrupted by a short regression sometime between 600 and 700 A.D., which is not recorded in Shilik’s curve. Balabanov’s scheme does not record the Medieval regression of Shilik’s curve. Three localities with geomorphic, historical, and archaeological landmarks are compared with the scheme of sea-level curves, the Chërnaya floodplain (previously discussed), Laspi Bay, and the Sivash (Fig. 13). The dates from the two cores in the Chernaya floodplain (Fig. 11) show a change from overbank flood silt deposition with cumulic soils to permanent impounding from ca. 5 ka to 3.5 ka, which is also indicated by the increase in aquatic vegetation in the basin (Cordova and Lehman, 2005). This could be the effect of a back-up of seawater from Sevastopol into the floodplain as a result of a rising sea level, probably the Old Black Sea transgression (Fig. 13). Overbank sedimentation with cumulic soils in the floodplain returned and remained until recent centuries. No evidence for a Phanagorian regression and Nymphaean transgression exists in the two cores.

Figure 13. Sea-level curves by Fedorov (1978), Shilik (1997), and periodization by Balabanov (2007) with modifications by the authors.

158 Cordova et al.

Late Pleistocene and Holocene paleoenvironments of Crimea The second kind of evidence comes from the stratigraphic section and archaeological settlement in Laspi Bay (Fig. 12). The chronology developed by Firsov (1972) was later modified on the basis of a recent stratigraphic analyses and more recent radiocarbon dates (Table 2). The third case presented here corresponds to the Sivash, which is more complicated and requires a full analysis of its stratigraphic history. Nonetheless, although these examples seem to agree with the regressions and transgressions, they seem to show no evidence for an 8 m Phanagorian regression as proposed by Shilik (1997). The Sivash Phenomenon The system of lagoons between the Perekop and the Azov Sea presents an interesting aspect of environmental change in which tectonics, sea-level changes, and human history are combined. Although numerous descriptions of sediment and soil stratigraphy have been presented for the region (Stashchuk et al., 1964; Veklich and Sirenko, 1976; Olenkovsky, 2000), no absolute dates for a master chronology exist. The origin and age of the Sivash are two issues that have haunted Quaternary scientists to this day. The Sivash consists of interconnected lagoons of shallow water occupying an area of 2540 km2. Differences in depth, salinity, and mineralogy are notable across the area, and, therefore, Stashchuk et al. (1964) divided the Sivash into western, central, eastern, and southern basins (Fig. 14). Depth varies from less than 0.5 m in the western basin to ~3 m in the southern Sivash. The central and eastern basins have an average maximum of 1 m and 2 m, respectively. The deepest of all (3 m) is reached in the southern part of the southern basin. The Sivash, referred to as the Putrid Lake (Gniloe Ozero, in Russian) is known for salt production, where high water salinity is the result of high evaporation rates. Precipitation in this area is the lowest in Crimea, reaching ~200 mm/yr, where potential evapotranspiration is five times this amount (L’vova, 1982). Inside the Sivash, salinity increases considerably from the mouth eastward and southward; the mean salinity varies from 90‰ to 200‰ in the southern basin, from 80‰ to 270‰ in the central basin, and from16‰ to 260‰ in the western basin (values calculated from percentages reported by Stashchuk et al., 1964, p. 12–15). Just for comparison, the salinity of the Azov Sea is 10‰–11‰ (Stashchuk et al., 1964), which is considerably lower than the mean salinity of the world ocean (35‰) due to the influx of freshwater from rivers. Sediment cores from the Sivash bed show first a series of loess-like silt and paleosols, covered by a relatively thin sequence of lagoonal sediments (Stashchuk et al., 1964). This sequence indicates that the Sivash is a relatively recent phenomenon, particularly in the western and central basins, and probably of Holocene origin (Podgorodetskiy, 1988). Archaeological survey of the Sivash area shows that the Upper Paleolithic (40–16 ka) and the Mesolithic (14–8 ka) are the most common settlements in the area (Olenkovsky, 2000). The middle and late Holocene sites are very rare,

159

suggesting that interest in the Sivash by human groups became minimal after the Mesolithic period (ending around 7.5 ka). Using historical maps and descriptions since Hellenistic and Roman times, Stashchuk et al. (1964) identified changes in the landscape for the past 2000 yr, particularly for the Arabatskaya Strelka and the western basin. The Arabatskaya Strelka does not appear on maps and historical accounts until the sixteenth century. Before then, the present location of the Arabatskaya Strelka was occupied by two islands in the early first millennium A.D., and then by a chain of islands during the late first millennium and early second millennium A.D. (Stashchuk et al., 1964). This development suggests that the Arabatskaya Strelka formed as an original accumulation of sand along a spit that eventually connected the small islands. Strabo’s descriptions and later Claudius Ptolemy’s map refer to Lake Biki, which occupied the present area of the western Sivash (Stashchuk et al., 1964, p. 160–161). Additionally, a series of shallow lakes, probably similar to the ones existing today south of the Perekop, were often reported in the area of the central basin. The eastern and southern basins were part of the Sea of Azov, until they were separated from it by the Arabatskaya Strelka. Historical observations of the coastal configuration of the Sivash region can be seen in the context of sea-level changes in the late Holocene (Fig. 13). The observations of Strabo and Pliny the Elder were made at the end of the Phanagorian regression, perhaps when large shallow areas had been exposed as islands. In the historical data collected by Stashchuk et al. (1964), there is no report of a sand spit for most of the first millennium A.D.— roughly coinciding with the Nymphaean transgression. The first report of the Arabatskaya Strelka occurs at the end of the seventeenth and more clearly during the eighteenth century, that is, at the beginning of the last regression. The other possible effects of sea-level change on the formation of the Arabatskaya Strelka are shown by Zenkovich (1958), who proved that the accumulation of so great a size and volume of sand and shells in the bar of Arabatskaya Strelka (estimated at more than 300 million m3) could have occurred only over a long period, with the area continuously being exposed and drowned. This scenario is based on the existence of a submerged sand bar parallel to the Arabatskaya Strelka (Fig. 14A). This bar is ~2 km wide, and its crest is 4 m below the sea surface (Zenkovich, 1958, p. 171). Sand along the bar is continuously being accumulated by currents on its west side, consequently producing a westward migration of the entire bar (Zenkovich, 1958, p. 171). Based on the development of the offshore bar, Зенкович came to the conclusion that the Arabatskaya Strelka migrated westward until it reached an elevation high enough to be exposed and become attached to the mainland to the south, thus creating the eastern and southern lagoonal basins of the Sivash. This scenario does not explain the flooding of the central and western basins, however. Therefore, a more complicated scenario involving tectonics, sea level, coastal accumulation processes, and even fluvial input of sediment should be considered.

160

Cordova et al.

Figure 14. Morphological changes in the Sivash region over the past two millennia according to Stashchuk et al. (1964).

Late Pleistocene and Holocene paleoenvironments of Crimea PATTERNS OF ENVIRONMENTAL CHANGE: A CONCLUSION This paper pursued the objective of identifying the most important paleoenvironmental events in the Quaternary of Crimea over the last 130 k.y. using data from research published in previous decades as well as recent work by the authors in order to explore several aspects of late Pleistocene environmental change in the Crimean Peninsula. These include changes in vegetation, soils, sedimentation, and coastlines in relation to climatic changes, and the relations among humans, vegetation, and climate, and the need for higher resolution in the chronology of terrestrial and marine deposits. Coastal changes related to fluctuations of the Black Sea and the Sea of Azov have been important in shaping landform patterns and sediment deposition directly by changing the base level of streams to expose large tracts of shelf or drowning modern valleys and inlets. Evidence for sea-level stands appears in the marine terraces, particularly the Karangat terraces, and in the filling of marine sediments of river valleys, as in the case of the Lower Salgir. In turn, conditions inland were warmer and wetter, in part because these changes occurred during interglacials and interstadials, and in part because the marine sources increased as the peninsula became surrounded by water. The question whether Crimea became an island during the Eemian Interglacial, or MIS 5e, as proposed by Chabai (2007), is still debatable, particularly given the lack of tangible stratigraphic evidence in the Perekop and Sivash area. Vegetation and soil development correlate with the sequence of stadials and interstadials during the Pleistocene. Pollen data point to the predominance of steppe vegetation during cold periods, and forest-steppe and forest during warm periods, a relation that is confirmed by faunal assemblages. The PleistoceneHolocene transition is characterized by the return of forest and forest-steppe communities, which was the dominant vegetation during the Eemian Interglacial. This transition was delayed by oscillations between cold and warm periods, however, and the attainment of modern forest communities in the Heraklean Peninsula and the foothills of the mountains occurred after 7.4 ka, at which time the Black Sea reached levels close to those of today. A dry period centered between 4.5 and 4 ka is apparent in most regional pollen records. The return of humid conditions after 3.5 ka seems to have coincided with the spread of Bronze and Iron Age communities and finally the Greek colonization between 3 and 2.5 ka. Records show little resolution for the past millennium. The influence of humans on the geomorphic and ecological processes of the peninsula can be assumed by coupling geoarchaeological and paleoecological records. During the middle and late Holocene, the Heraklean and Kerch Peninsulas underwent cycles of erosion. Although these cycles might have been climatically driven, they occurred at a time when farming and pastoralism arose in Crimea. In the pollen and soil records of the Heraklean Peninsula, the beginning of the Neolithic is represented by an increase in phrygana scrub. Although an increase

161

of phrygana is often related to pastoral activities, it may have resulted from climatic conditions. The effects of the establishment of Greek farming are more clearly seen in the records, particularly in the pollen data. Ironically, there is no evidence for increased erosion. It is not until the abandonment of farms that erosion seems to occur, particularly in the Tarkhankutskaya Balka in the Kerch Peninsula. Although coring and studies of sections have been undertaken for decades, a tighter chronology is needed. In particular, these studies are needed for marine deposits to determine whether they correspond to the Karangat or Surozh transgressions. Dated loess sequences in the plains of Crimea would complement paleoenvironmental data already obtained in the Pleistocene deposits of the mountains. The terraces of the large river valleys are another source of relevant information, particularly through dating and correlation with sea-level changes. Holocene deposits such as those of the Sivash, bays, and lakes are still in need of further study and better understanding. In many cases, data from pollen, diatoms, and other proxy records can also add interesting information to the climatic context in areas where the most recent cultures of the peninsula have evolved. ACKNOWLEDGMENTS This research was funded in part by a grant from the Packard Humanities Institute to the Institute of Classical Archaeology (ICA) of the University of Texas at Austin and small grants by the College of Arts and Sciences at Oklahoma State University. Field research was carried out during summer seasons of years 1998–2001. We thank the numerous members of the staff of the Natural Preserve of Tauric-Chersonesos (Sevastopol), who helped this project in many ways and for providing us with a base for field research. We are thankful to many members of the Faculty of Geography at Taurida Vernadsky National University in Simferopol and the Nikitsky Botanical Gardens in Yalta, and the Institute of Geography of the National Ukrainian Academy of Sciences in Kyiv. REFERENCES CITED (Russian and Ukrainian references are immediately followed by their translations.) Andrusov, N.I., 1912, Terraces of the Sudak Area, Zapiski Kievskogo Oshschestvo Estestvopyt, v. 22, no. 1, no paging (in Russian). Aндpycoв, H.И., 1912, Teppacы oкpecтнocтeй Cyдaкa, Зaпиcки Kиeвcкoгo Oбщecтвa Ecтecтвoиcпыт, v. 22, issue 2. Antoine, P., Rousseau, D.-D., Lautridou, J.-P., and Hatté, C., 1999, Last interglacial-glacial climatic cycle in loess-palaeosol successions of northwestern France: Boreas, v. 28, p. 551–563, doi: 10.1080/030094899422046. Artiushenko, А.Т., and Mishnev, В.Г., 1978, History of the vegetation on the Yailas and surrounding slopes during the Holocene: Kiev, Naukova Dumka, 140 p (in Russian). Apтюшeнкo, and Mишнeв, 1978, Иcтopия pacтитeльнocть Kpымcкиx Яйл и пpияйлинcкиx cклoнoв в Гoлoцeнe: Kiev, Hayкoвa Дyмкa, 140 p. Balabanov, I.P., 2007, Holocene sea-level changes of the Black Sea, in YankoHombach, V., Gilbert, A.S., Panin, N., and Dolukhanov, P.M., eds., The Black Sea Flood Question: Changes in Coastline, Climate, and Human Settlement: Dordrecht, Springer, p. 711–730.

162

Cordova et al.

Berg, L.S., 1950, Natural Regions of the USSR (translated from the Russian by O.A. Titelbaum): New York, The Macmillan Company, 436 p. Bibikov, S.N., Stanko, V.N., and Cohen, V.Yu., 1994, Final Paleolithic and Mesolithic of the Crimean Mountains: Odessa, Vest’, 238 p. (in Russian). Бибикoв, C.H., Cтaнкo, B.H., and Koeн, B.Ю., 1994, Финaльный Пaлeoлит и Meзoлит Гopнoгo Kpымa: Odessa, Becть, 238 p. Biodiversity Support Program, 1999, Priority-Setting in Conservation: New Approach for Crimea: Washington, D.C., Biodiversity Support Program and World Wildlife Fund, 257 p. Blagovolin, Н.С., 1962, Geomorphology of the Kerch-Taman District: Moscow, Izdatel’stvo Akademiy Nauk SSSR, 190 p. (in Russian). Блaгoвoлин, H.C., 1962, Гeoмoфoлoгия Kepчeнcкo-Taмaнcкoй Oблacти: Moscow, Издaтeльcтвo Aкaдeмий Hayк CCCP, 190 p. Burke, A., Markova, K., Mikhailesku, C., and Patou-Mathis, M., 1999, The animal environment of western Crimea, in Chabai, V.P., and Monigal, K. eds., The Middle Paleolithic of Western Crimea, Volume 2: Etudes et Recherches Archéologiques de l’Université de Liège v. 87, p. 143–151. Chabai, V.P., 2005, Kabazi II: Stratigraphy and archaeological sequence, in Chabai, V.P., Richter, J., and Uthmeier, T., eds., Kabazi II: Last Interglacial Occupation, Environmental Subsistence: Simferopol, National Academy of Sciences of Ukraine, Institute of Archaeology, Crimean Branch, and Cologne, University of Cologne, Institute of Prehistoric Archaeology, p. 1–24. Chabai, V.P., 2007, The Middle Paleolithic and Early Upper Paleolithic in the northern Black Sea region, in Yanko-Hombach, V., Gilbert, A.S., Panin, N., and Dolukhanov, P.M., eds., The Black Sea Flood Question: Changes in Coastline, Climate, and Human Settlement: Dordrecht, Springer, p. 279–296. Chabai, V.P., Marks, A.E., and Monigal, K., 1999, Western Crimea Middle Paleolithic paleoenvironment and paleoeconomy, in Chabai, V.P., and Monigal, K., eds., The Middle Paleolithic of Western Crimea, Volume 2: Etudes et Recherches Archéologiques de l’Université de Liège v. 84, p. 211–233. Chabai, V.P., Marks, A.E., and Monigal, K., 2004, Crimea in the context of the Eastern European Middle Paleolithic and Early Upper Paleolithic, in Chabai, V.P., Monigal, K., and Marks, A.E., eds., The Middle Paleolithic and Early Upper Paleolithic of Eastern Crimea: Etudes et Recherches Archéologiques de l’Université de Liège, v. 104, p. 419–460. Chepalyga, A.L., 1984, Inland sea basins, in Velichko, A.A., Wright, H.E., Jr., and Barnosky, C.W., eds., Late Quaternary Environments of the Soviet Union, English Edition: Minneapolis, University of Minnesota Press, p. 229–247. Chepalyga, A.L., 2007, The Late Glacial Great Flood in the Ponto-Caspian Basin, in Yanko-Hombach, V., Gilbert, A.S., Panin, N., and Dolukhanov, P.M., eds., The Black Sea Flood Question: Changes in Coastline, Climate, and Human Settlement: Dordrecht, Springer, p. 119–148. Cohen, V., 1996, Neolithization of the Crimean Mountains (current stage of investigations): Préhistoire Européenne, v. 9, p. 417–433. Cohen, V., Gerasimenko, N., Rekovetz, L., and Starkin, A., 1996, Chronostratigraphy of Rockshelter Skalistiy: Implications for the Late Glacial of the Crimea: Préhistoire Européenne, v. 9, p. 325–356. Comes, H.P., and Kadereit, J.W., 1998, The effect of Quaternary climatic changes on plant distribution: Trends in Plant Science, v. 3, p. 432–438, doi: 10.1016/S1360-1385(98)01327-2. Cordova, C.E., 2007, Holocene Mediterranization of the southern Crimean vegetation: Paleoecological records, regional climate change, and possible nonclimatic influences, in Yanko-Hombach, V., Gilbert, A.S., Panin, N., and Dolukhanov, P.M., eds., The Black Sea Flood Question: Changes in Coastline, Climate, and Human Settlement: Dordrecht, Springer, p. 319–344. Cordova, C.E., and Lehman, P.H., 2003, Achaeopalynology of synanthropic vegetation in the Chora of Chersonesos, Ukraine: Journal of Archaeological Science, v. 30, p. 1483–1501, doi: 10.1016/S0305-4403 (03)00044-X. Cordova, C.E., and Lehman, P.H., 2005, Holocene environmental change in southwestern Crimea (Ukraine) in pollen and soil records: The Holocene, v. 15, p. 263–277, doi: 10.1191/0959683605hl791rp. Cordova, C.E., and Lehman, P.H., 2006, Mediterranean agriculture in southwestern Crimea: Paleoenvironments and early adaptations, in Peterson, D.L., Popoval, L.M., and Smith, A.T., eds., Beyond the Steppe and the Sown. Proceedings of the 2002 University of Chicago Conference on Eurasian Archaeology: Colloquia Pontica, v. 13, p. 425–447. Cordova, C.E., Rybak, A., and Lehman, P.H., 2001, Vegetation patterns and conservation issues in southern Crimea: Post-Soviet Geography and Economics, v. 42, p. 49–72.

Didukh, Ya.P., 1992, The vegetation cover of the Crimean Mountains: Kiev, Naukova Dumka. 251 p. Дидyx, Я.П., 1992, Pacтитeльный Пoкpoв Гopнoгo Kpымa: Kiev, Hayкoвa Дyмкa, 251 p. Douguedroit, A., and Zimina, R.P., 1987, Le climat méditerranéen en France et en U.R.S.S.: Mediterranée, v. 61, p. 75–84. Fedorov, P.V., 1978, The Pleistocene of the Ponto-Caspian Region: Moscow, Nauka, 164 p. (in Russian). Фeдopoв, П.B., 1978, Плeйcтoцeн Пoнтo-Kacпия: Moscow, Hayкa, 164 p. Firsov, L.V., 1972, Archaeological radiocarbon dating of the Laspi Terrace in Crimea and its origin: Biulleten’ Komissii po Izucheniiu Chetvertichnogo Perioda, v. 38, p. 72–87. Фиpcoв, Л.B., 1972, Apxeoлoгичecкaя paдиoyглepoднaя дaтиpoвкa тeppac Лacпи в Kpымy и иx пpoиcxoждeниe: Бюллeтeнь Koмиccии пo Изyчeнию Чeтвëpтичнoгo Пepиoдa, v. 38, p. 72–87. Gerasimenko, N., 1997, Environmental and climatic changes between 3 and 5 ka BP in southeastern Ukraine, in Dalfes, H.N., Kukla, G., and Weiss, H., eds., Third Millennium BC Climate Change and Old World Collapse: Berlin, Springer-Verlag, p. 371–399. Gerasimenko, N., 1999, Late Pleistocene vegetational history of the Kabazi-II Paleolithic site, in Chabai, V.P., and Monigal, K., eds., The Middle Paleolithic of Western Crimea, Volume 2: Etudes et Recherches Archéologiques de l’Université de Liége, v. 87, p. 115–141. Gerasimenko, N., 2001, Late Pleistocene vegetation and soil evolution at the Kiev loess plain as recorded in the Stari Bezradychy section, Ukraine: Studia Quaternaria, v. 17, p. 19–28. Gerasimenko, N., 2004, Vegetational history of Buran Kaya III, in Chabai, V.P., Monigal, K., and Marks, A.E., eds., The Middle Paleolithic and Early Upper Paleolithic of Eastern Crimea: Etudes et Recherches Archéologiques de l’Université de Liége, v. 104, p. 19–34. Gerasimenko, N., 2005a, Climatic oscillations in the foothills of Crimean Mountains during 30–10 ky BP, in Extended Abstracts of the Plenary Meeting of Project IGCP 521 “Black Sea–Mediterranean Corridor during the Last 30 k.y.: Sea-Level Change and Human Adaptation”: Istanbul, Kadir Has University, p. 51–53. Gerasimenko, N., 2005b, Vegetational evolution of the Kabazi II site, in Chabai, V.P., Richter, J., and Uthmeier, T., eds., Kabazi II: Last Interglacial Occupation, Environment and Subsistence: Simferopol, National Academy of Sciences of Ukraine, Institute of Archaeology, Crimean Branch, and Cologne, University of Cologne, Institute of Prehistoric Archaeology, p. 25–49. Gerasimenko, N., 2006, Upper Pleistocene loess-paleosol and vegetational successions in the Middle Dnieper area, Ukraine: Quaternary International, v. 149, p. 55–66, doi: 10.1016/j.quaint.2005.11.018. Gerasimenko, N., 2007, Environmental changes in the Crimean mountains during the Last Interglacial–Middle Pleniglacial as recorded by pollen and lithopedology: Quaternary International, v. 164–165, p. 207–220, doi: 10.1016/j.quaint.2006.12.018. Gerasimov, I.P., 1954, Cinnamonic Soils of Mediterranean Regions (Contribution to the 5th International Congress of Pedologists): Moscow, Izdatel’stvo Akademii Nauk SSSR, 38 p. (in Russian). Giosan, L., Donnelly, J.P., Constantinescu, S., Filip, F., Ovejanu, I., Vespremeanu-Stroe, A., Vespremeanu, E., and Duller, G.A.T., 2006, Young Danube delta documents stable Black Sea level since middle Holocene: Morphodynamic, paleogeographic and archaeological implications: Geology, v. 34, p. 757–760, doi: 10.1130/G22587.1. Gozhik, P.F., and Novosel’skiy, F.A., 1989, Geological setting of the DnieperBug hydrological complex: Kiev, Naukova Dumka, 118 p. (in Russian). Гoжик, П.Ф., and Hoвoceльcкий, Ф.A., 1989, Гeoлoгичecкиe ycлoвия cтpoитeльcтвa Днeпpoвcкo-Бyгcкoгo гидpoyзлa: Kiev, Hayкoвa дyмкa, 118 p. Gozhik, P.F., Karpov, V.A., Ivanov, V.T., Sibirenko, M.G., and Grapov, A.K., 1987, The Holocene of the northwestern area of the Black Sea, sequences, radiocarbon dating, and correlation of sections: Kiev, Institute of Geological Sciences, Academy of Sciences of the Ukrainian SSR, 43 p. (in Russian). Гoжик, П.Ф., Kapпoв, B.A., Ивaнoв, B.T., Cибиpчeнкo, M.Г., and Гpaпoв, A.K., 1987, Гoлoцeн ceвepo-зaпaднoй чacти Чëpнoгo Mopя, pacчeнeниe, paдиoyглepoднoй дaтиpoвaниe, кoppeляция paзpeзoв: Kiev, Инcтитyт Гeoлoгичecкиx Hayк, AH YCCP, 43 p. Gozhik P.F., Shovkoplyas, V.M., Komar, M.S., Matviishina, Zh.M., and Perediy, V.I., 2000, Correlation of loess and glacial deposits of Poland

Late Pleistocene and Holocene paleoenvironments of Crimea and Ukraine. Guidebook of the 5th Polish-Ukrainian Seminar: Kiev, 34 p. (in Ukrainian). Гoжик П.Ф., Шoвкoпляc, B.M., Koмap, M.C., Maтвiëшинa, Ж.M., and Пepeдepiй, B.I., 2000, Kopeляцiя лeciв i льoдoвикoвиx вiдклaдiв Пoльщi i Yкpaëни, Пyтiвник X Пoльcькo-Yкpaëнcькoгo ceмiнapy: Kiev, 34 p. Gozhik, P., Matviishina, Zh., Gerasimenko, N., Rekovets, L., and Shelkoplyas, V., 2001, Quaternary stratigraphy, in The Quaternary of Ukraine Explored, Excursion Guide of the Subcommission on European Quaternary Stratigraphy 2001 Conference: Kiev, p. 8–11. Gubonina, Z.P., 1985, Environment of Paleolithic humans in the piedmont of Crimea, in Grichuk, V.P., and Zaklinskaya, Ye.D., eds., Palinologiya chetvërtichnogo perioda: Moscow, Nauka, p. 95–103 (in Russian). Гyбoнинa, З.П., 1985, Ycлoвия oбитaния пaлeoлитичecкoгo чeлoвeкa в пpeдгopиьяx Kpымa, in Гpичyк, B.П., and Зaклинcкaя, E.Д., eds., Пaлинoлoгия Чeтвepтичнoгo Пepиoдa: Moscow, Hayкa, p. 95–103. Haesaerts, P., and Mestdagh, H., 2000, Pedosedimentary evolution of the last interglacial and early glacial sequence in the European loess belt from Belgium to central Russia: Netherlands Journal of Geosciences, v. 79, p. 313–324. Kaplin, P.A., and Selivanov, A.O., 2004, Late Glacial and Holocene sea level changes in semi-enclosed seas of North Eurasia: Examples from the contrasting Black and White Seas: Palaeogeography, Palaeoclimatology, Palaeoecology, v. 209, p. 19–36, doi: 10.1016/j.palaeo.2004.02.016. Kliukin, A.A., and Shchepinskiy, A.A., 1983, New Paleolithic findings in the terrace deposits of the Al’ma River (Crimea): Biulleten’ Komissii po Izucheniiu Chetvertichnogo Perioda, v. 52, p. 156–164 (in Russian). Krupskiy, N.K., and Polupan, N.I., 1979, Atlas of soils of the Ukrainian SSR: Kiev, Urozhay, 64 p. (in Russian). Kpyпcкий, H.K., and Пoлyпaн, H.И., 1979, Aтлac Пoчв Yкpaинcкoй CCCP: Kiev, Ypoжaй, 64 p. Kremenetski, C.V., 1995, Holocene vegetation and climate history of southwestern Ukraine: Review of Palaeobotany and Palynology, v. 85, p. 289– 301, doi: 10.1016/0034-6667(94)00123-2. Kremenetski, K.V., 1997, Human impact on Holocene vegetation of South Russian Plain, in Chapman, J., and Dolukhanov, P., eds., Landscape in Flux: Central and Eastern Europe in Antiquity: Colloquia Pontica, v. 3, p. 275–287. Kremenetski, K.V., 2003, Steppe and forest-steppe belt of Eurasia: Holocene environmental history, in Levine, M., Renfrew, C., and Boyle, K., eds., Prehistoric Steppe Adaptation and the Horse: Cambridge, McDonald Institute, p. 11–27. Kryzhitskii, S.D., 1997, The landscape of the North Pontic city-states: A case study from Olbia, in Chapman, J., and Dolukhanov, P., eds., Landscapes in Flux: Central and Eastern Europe in Antiquity: Colloquia Pontica, v. 3, p. 101–114. Lazukov, G.I., Gvozdover, M.D., Roginskiy, Ya.Ya., Uryson, M.I., Kharitonov, V.M., and Yakimov, V.P., 1981, Nature and prehistoric people: Basic phases of development of prehistoric peoples and their culture in the territory of the USSR during the Pleistocene: Moscow, Mysl, 220 p. (in Russian). Лaзyкoв, Г.И., Гвoздoвep, M.Д., Poгинcкий, Я.Я., Ypыcoн, M.И., Xapитoнoв, B.M., and Якимoв, B.П., 1981, Пpиpoдa и дpeвныий чeлoвeк: Ocнoвныe этaпы paзвития пaлeoлитичecкoгo чeлoвeкa и eгo кyльтypы нa тeppитopий CCCP в Плeйcтoцeнe: Moscow, Mыcль, 220 p. Leroy, S.A.G., and Arpe, K., 2007, Glacial refugia for summer-green trees in Europe and south-west Asia as proposed by ECHAM3 time-slice atmospheric model simulations: Journal of Biogeography, v. 34, p. 2115–2128, doi: 10.1111/j.1365-2699.2007.01754.x. Lindner, L., Bogutsky, A., Gozhik, P., Marciniak, B., Marks, L., Łanczont, M., and Wojtanovich, M., 2002, Correlation of main climatic glacialinterglacial and loess-palaeosol cycles in the Pleistocene of Poland and Ukraine: Acta Geologica Polonica, v. 52, p. 459–469. L’vova, E.V., 1982, The Plains of Crimea: Simferopol, Tavriya, 79 p. (in Russian). Львoвa, E.B., 1982, Paвнины Kpымa, Cимфepoпoль, Издaтeльcтвo Taвpия, 79 p. Markova, A., and Puzachenko, A., 2006, Vertebrate records: Late Pleistocene of northern Asia, in Elias, S., ed., Encyclopedia of the Quaternary Period, Volume 4: Amsterdam, Elsevier Science, p. 3158–3175.

163

Maslennikov, А.А., 1997, Research on the rural territories of the European Bosphorus: results and perspectives at the end of the century: Rossiyskaya Arkheologiya, v. 3, p. 60–70 (in Russian). Macлeнникoв, A.A., 1997, Иccлeдoвaниe ceльcкoй тeppитopии eвpoпeйcкoгo Бocпopa: Итoги и пepпeктивы к кoнцy вeкa: Poccийcкaя Apxeoлoгия, v. 3, p. 60–70. Morozova, O.V., and Kozharinov, A.V., 2001, Influence of historical factors in the spread of floristic species richness in Eastern Europe: Izvestiya Akademiya Nauk RAN, Geograficheskaya Seriya, v. 5, p. 39–50 (in Russian). Mopoзoвa, O.B., and Koжapинoв, A.B., 2001, Bлияниe иcтopичecкиx фaктopoв нa pacпpeдeлeниe видoвoгo бoгaтcвa флopы вocтoчнoй Eвpoпы, Извecтия Aкaдeмия Hayк, PAH, Cepия Гeoгpaфичecкaя, v. 5, p. 39–50. Muratov, M.V. 1974, Crimea I Geological Description: Geologiya SSSR, Volume 8: Moscow: Ministry of Geology of the USSR, 547 p. (in Russian). Mypaтoв, M.B., 1974, Kpым I Гeoлoгичecкoe Oпиcaниe. Гeoлoгия CCCP., Volume 8: Moscow, Mиниcтepcтвo Гeoлoгии CCCP, 547 p. Novik, N.N., 1990, The possibility of determining the age of river terraces from the thickness of weathered layers in alluvial pebbles: Geologicheskiy Zhurnal, v. 5, p. 57–62 (in Russian). Hoвик, H.H., 1990, O вoзмoжнocти oпpeдeлeния вoзpacтa peчниx тeppac пo мoщнocти вывeтpeлoгo cлoя в гaлькax aллювия: Гeoлoгичecкий Жypнaл, v. 5, p. 57–62. Olenkovsky, M.P., 2000, The Paleolithic and Mesolithic of the Sivash Region, The issue of the Epigravettian of Ukraine: Kherson, Vydavnytstvo Prydniprov’ya, 170 p. (in Ukrainian). Oлeнкoвcкий, M.П., 2000, Пaлeoлiт тa Meзoлiт Пpиcивaшшя. Пpoблeми Eпiгpaвeтy Yкpaëни, Kherson: Bидaвництвo Пpиднiпpoв’я, 170 p. Podgorodetskiy, P.D. ,1988, Nature of Crimea: Simferopol, Tavriya, 192 p. (in Russian). Пoдгopoдeцкий, П.Д., 1988, Kpым: Пpиpoдa: Simferopol, Taвpия, 192 p. Polkanova, L.P., and Varushchenko, S.I., 1969, Structural-geomorphic analysis of the relief of the Tarkhankutsko-Novoselovskoye hilly plain of the Crimean steppes: Petroleum Geology, v. 8, p. 204–206. Rousseau, D.-D., Gerasimenko, N., Matviischina, Zh., and Kukla, G., 2001, Late Pleistocene environments of the central Ukraine: Quaternary Research, v. 56, p. 349–356, doi: 10.1006/qres.2001.2270. Semenenko, V.I., Luparenko, A.V., and Kovaliukh, N.N., 1979, On the correlation of marine and continental deposits of the Crimean plains in relation to the Karangat ingressions in the valley of the Salgir River: Biulleten’ Komissii po Izucheniiu Chetvertichnogo Perioda, v. 49, p. 17–25 (in Russian). Ceмeнeнкo, B.И., Лyпapeнкo, A.B., and Koвaлюx, H.H., 1979, O кoppeляции мopcкиx и кoнтиeнтaлныx плeйcтoцeнoвыx oтлoжeний paвниннoгo Kpымa в cвязи c Kapaнгaтcкoй ингpeccиeы в дoлинe p. Caлгиp: Бюллeтeнь Koмиccии пo Изyчeнию Чeтвëpтичнoгo Пepиoдa, v. 49, p. 17–25. Shcheglov, A.N., 1978, Northwest Crimea in antiquity: Leningrad, Nauka, Leningradskoye Otdeleniye, 156 p. (in Russian). Щeглoв, A.H., 1978, Ceвepo-зaпaдный Kpым в aнтичнyю эпoxy: St. Petersburg, Hayкa, Лeнингpaдcкoe Oтдeлeниe, 156 p. Shcherbakov, F.A., Kuprin, P.N., and Polyakov, A.S., 1976, The shelf of the northwestern part of the Black Sea in the Late Pleistocene and Early Holocene: Chetvërtichniy Period, v. 16, p. 141–152 (in Russian). Щepбaкoв, Ф.A., Kyпpин, П.H., and Пoлякoв, A.C., 1976, Шeлф, ceвepoзaпaднoй чacти Чepнoгo Mopя в пoзнeм плeйcтoцeнe-гoлoцeнe: Чeтвëpтичный Пepиoд, v. 16, p. 141–152. Shcherbakov, F.A., Kuprin, P.N., Zabelina, Ye. K., and Parunin, O.B., 1977, Paleogeography of the Azov and Black Seas in the Late Pleistocene and Holocene, in Kalinin, P.A., and Shcherbakov, F.А., eds., Paleogeography and Pleistocene deposits of the southern seas of the USSR: Moscow, Izdatel’stvo Nauka, 251 p. (in Russian). Щepбaкoв, Ф.A., Kypпин, П.H., Зaбeлинa, E.K., Cкибa, C.И., and Пapyнин, O.Б., 1977, Пaлeoгeoгpaфия Aзoвo-Чepнoмopья в пoзднeм Плeйcтoцeнe и Гoлoцeнe, in Kaплин, П.A., and Щepбaкoв, Ф.A., eds., Пaлeoгeoгpaфия и oтлoжeния Плeйcтoцeнa южныx мopeй CCCP: Moscow, Издaтeльcтвo Hayкa, 251 p. Shilik, K.K., 1997, Oscillations of the Black Sea and ancient landscapes, in Chapman, J., and Dolukhanov, P., eds., Landscapes in Flux: Central and Eastern Europe in Antiquity: Colloquia Pontica, v. 3, p. 115–129.

164

Cordova et al.

Sirenko, N.A., and Turlo, S.I., 1986, The development of soils and vegetation of Ukraine during the Pliocene and Pleistocene: Kiev, Naukova Dumka, 186 p. (in Russian). Cиpeнкo, H.A., and Typлo, C.И., 1986, Paзвитиe пoчв и pacтитeльнocть Yкpaины в плиoцeнe и плeйcтoзeнe: Kiev, Hayкoвa Дyмкa, 186 p. Slavin, V.I., 1975, Modern geological process in southwestern Crimea: Moscow, Izdatel’stvo Moskovskogo Universiteta, 194 p. Cлaвин, B.И., 1975, Coвpeмeнныe гeoлoгичecкиe пpoцeccы в югo-зaпaднoм Kpымy: Moscow, Издaтeльcтвo Mocкoвcкoгo Yнивepcитeтa, 194 p. Sorochan, S.B., Zubar’, V.M., and Marchenko, L.V., 2000, Life and decline of Chersonesos: Kharkov, Ukraine, Maydan, 826 p. (in Russian). Copoчaн, C.Б., Зyбapь, B.M., and Mapчeнкo, Л.B., 2000, Жизнь и гибeль Xepcoнeca: Kharkov, Maйдaн, 826 p. Stashchuk, M.F., Suprichev, V.A., and Khitraya, M.Ts., 1964, Mineralogy and geochemical conditions and the formation of the Sivash sediments: Kiev, Naukova Dumka, 173 p. (in Russian). Cтaщyк, M.Ф., Cyпpичeв, B.A., and Xитpaя, M.Ц., 1964, Mинepaлoгия, гeoxимия и ycлoвия фopмиpoвaния дoнныx oтлoжeний Cивaшa: Kiev, Hayкoвa Дyмкa, 173 p. Stepanchuk, V.N., 2006, Lower and Middle Paleolithic of Ukraine: Zelena Bukovina, 463 p. Cтeпaнчyк, B.H., 2006, Hижный и cpeдний пaлeoлит Yкpaины: Зeлeнa Бyкoвинa, 463 p. Vakhrushev, B.A., and Kliukin, A.A., 2001, Cryogenic processes of the in the Crimean Yailas, Geomorfologiya, v. 2001-2, p. 48–54 (in Russian). Baxpyшeв, Б.A., and Kлюкин, A.A., 2001, Kpиoгeнныe пpoцeccы кpымcкиx яйл, Гeoмopфoлoгия, v. 2001-2, p. 48–54. Veklich, M.F., 1990, Basics of paleolandscape research: Kiev, Naukova Dumka,190 p. (in Russian). Beклич, M.Ф., 1990, Ocнoвы пaлeoлaндшaфтoвeдeния: Kiev, Hayк. дyмкa., 190 p.

Veklich, M.F., and Sirenko, N.A. 1976, The Pliocene and Pleistocene of the left bank of the Dnieper and the plains of Crimea: Kiev, Naukova Dumka, 186 p. (in Russian). Beклич, M.Ф., and Cиpeнкo, H.A., 1976, Плиoцeн и Плeйcтoцeн лeвoбepeжья нижнeгo Днeпpa и paвниннoгo Kpымa: Kiev, Hayкoвa Дyмкa, 186 p. Veklich, M.F., Sirenko, N.A., Matviishina, Zh.N., et al. 1993, Stratigraphic scheme of the Pleistocene deposits of Crimea: Kiev: Goskomitet geologii Ukrainy, 40 p. (in Russian). Beклич M.Ф., Cиpeнкo, H.A., Maтвиишинa, Ж.H., et al., 1993, Cтpaтигpaфичecкaя cxeмa плeйcтoцeнoвыx oтлoжeний Yкpaины: Kiev, Гocкoмитeт гeoлoгии Yкpaины, 40 p. Velichko, A.A., 1988, Geoecology of the Mousterian in East Europe and adjacent areas, in Otte, M., ed., L’Homme de Neandertal, L’Environment, Volume 2: Etudes et Recherches Archéologiques de l’Université de Liège, v. 29, p. 181–206. Yena, V.G., Tverdokhlebov, I.T., and Shantyr’, S.P., 1996, The Southern Coast of Crimea: Simferopol, Biznes-Inform, 304 p. (in Russian). Eнa, B.Г., Tвepдoxлeбoв, И.T., and Шaнтыpь, C.П., 1996, Южный Бepeг Kpымa: Simferopol, Бизнec-Инфopм, 304 p. Yanko-Hombach, V., Gilbert, A.S., Panin, N., and Dolukhanov, P.M., eds., 2007, The Black Sea Flood Question: Changes in Coastline, Climate, and Human Settlement: Dordrecht, Springer, 971 p. Zenkovich, V.P., 1958, The coasts of the Black and Azov Seas: Moscow, State Publishing of Geographic Literature, 373 p. (in Russian). Зeнкoвич, B.П., 1958, Бepeгa Чëpнoгo и Aзoвcкoгo Mopeй: Mocквa, Гocyдapcтвeннoe Издaтeльcтвo Гeoгpaфичecкoй Литepaтypы, 373 p.

MANUSCRIPT ACCEPTED BY THE SOCIETY 22 JUNE 2010

Printed in the USA

The Geological Society of America Special Paper 473 2011

Bedforms, coastal-trapped waves, and scour process observations from the continental shelf of the northern Black Sea A. Trembanis* S. Nebel A. Skarke Department of Geological Sciences, University of Delaware, 109 Penny Hall, Newark, Delaware 19716, USA D.F. Coleman R.D. Ballard Graduate School of Oceanography, University of Rhode Island, South Ferry Road, Narragansett, Rhode Island 02882, USA A. Yankovsky Marine Science Program and Department of Geological Sciences, University of South Carolina, Columbia, South Carolina 29208, USA I.V. Buynevich* Geology & Geophysics Department, Woods Hole Oceanographic Institution, MS 22, Woods Hole, Massachusetts 02543, USA S. Voronov Department of Underwater Heritage, Institute for Archaeology, Academy of Sciences of Ukraine, Kyiv, Ukraine

ABSTRACT The Black Sea basin presents an ideal laboratory for investigations of morphodynamic interplay between response (morphology) and force (processes) associated with shelf sedimentation. Recent studies along the perimeter of the basin have documented the existence of a complex, heterogeneous seafloor varyingly composed of sand, gravel, silt, and clay. Side-scan sonar data are utilized to establish the spatial patterns of bedform types in the area. In addition, a benthic tripod, configured with an acoustic Doppler current profiler, a rotary fanbeam sonar, and a conductivity-temperature sensor was deployed to record seabed dynamics in response to changing forcing conditions. Together, the tripod and side-scan survey data sets provide a complementary basis for deciphering the processes responsible for the observed seafloor morphology. The side-scan sonar data allows for the determination of spatial patterns of bedform length and orientation. In total, 2376 individual large sand wave bedforms were

*E-mail, Trembanis—[email protected]; present address, Buynevich—Department of Earth and Environmental Science, Temple University, 1901 N. 13th Street, Philadelphia, Pennsylvania 19122, USA; [email protected]. Trembanis, A., Nebel, S., Skarke, A., Coleman, D.F., Ballard, R.D., Yankovsky, A., Buynevich, I.V., and Voronov, S., 2011, Bedforms, coastal-trapped waves, and scour process observations from the continental shelf of the northern Black Sea, in Buynevich, I.V., Yanko-Hombach, V., Gilbert, A.S., and Martin, R.E., eds., Geology and Geoarchaeology of the Black Sea Region: Beyond the Flood Hypothesis: Geological Society of America Special Paper 473, p. 165–178, doi: 10.1130/2011.2473(10). For permission to copy, contact [email protected]. ©2011 The Geological Society of America. All rights reserved.

165

166

Trembanis et al. digitized in geographic information systems with mean and modal wavelengths of 72.8 and 15.7 m respectively. The correlation of near-inertial waves (velocity amplitude 12–20 cm/s and period 12–16 h) and bedform geometry suggest that the extensive sand-wave patches imaged across the shelf are affected by active modern processes and may themselves be modern features or perhaps relict features that remain active presently. Progressive vector diagrams of the nearbed mean current flow indicate a component of cross-shelf directed flow, suggesting an enhanced potential for artifact preservation via cross-shelf advection of anoxic bottom waters by the near-inertial flows measured in this study.

BACKGROUND The Black Sea provides an ideal natural laboratory for testing the role of shelf transport processes on bedform and artifact interaction (Özsoy and Ünlüata, 1997; Neretin et al., 2001; Coleman and Ballard 2004). This unique setting presents opportunities to test concepts of artifact-related scour and transport associated with complex sorted bedform features (Murray and Thieler, 2004; Trembanis et al., 2004; Green et al., 2004) and shipwreck artifacts (McNinch et al., 2006). The widespread occurrence and previous documentation of large-scale bedforms on the continental shelf (Ryan et al., 1997; Coleman and Ballard 2004; Lericolais et al., 2006) raises the question of whether these bedforms are (1) strictly modern; (2) ancient relicts; or (3) palimpsest features (i.e., relict but reworked) features. The objective of this study is to examine how the hydrodynamics (mean currents) of the shelf interact with the seafloor morphology over spatial scales ranging from meters to kilometers in a morphodynamic context similar to that used in other shelf studies (e.g., Trembanis et al., 2004; McNinch et al., 2006). In addition to analyzing the distribution of bedforms (size and orientation), a secondary goal of this study is to relate the shelf hydrodynamic processes (internal waves) and seafloor morphology (bedforms) as important causative factors in shipwreck site and artifact preservation sensu (McNinch et al., 2006) with application to recent marine archaeological expeditions in the region (Coleman and Ballard, 2004). It has been hypothesized (Ryan et al., 1997; Coleman and Ballard, 2004) that interfacial internal waves assist in transporting anoxic waters onto and across the shelf providing a mechanism for enhanced artifact preservation above the normal oxycline. At the center of these hypotheses is the interplay between hydrodynamics and bed roughness. Previous observations (Ryan et al., 1997; Coleman and Ballard, 2004; Lericolais et al., 2006) suggest that there are strong analogs between the Black Sea shelf settings and the ubiquitous shelf sand body features originally termed “Rippled Scour Depressions” (Cacchione et al., 1984) that have been termed “Sorted Bedforms” in recent years (Murray and Thieler 2004; Trembanis et al., 2004). Of particular parallelism was the finding in New Zealand (Hume et al., 2003; Trembanis et al., 2004) that anoxic organic material in the sediment may have played a stabilizing role in controlling the lateral stability of the bedform features. It is possible that such deposits exist in the vicinity of

the wreck sites in the Black Sea and perhaps these shear resistant deposits play a similarly important role in artifact and site preservation, whereby reduced organic layers might cap and help preserve the artifacts. SHELF MORPHODYNAMICS Field observations and theoretical refinements by numerous investigators over the past several decades have significantly advanced our understanding of shelf sediment transport processes (Thieler et al., 1995; Wright, 1995; Grant and Madsen, 1986). The continental shelf is an important transition region for physical, biological, and geological processes—one that forms a critical link between the nearshore and the deep-sea basin. The shelf is a morphodynamic system influenced by coupled physical, geological, chemical, and biological processes (Wright, 1995). Processes and phenomena of the shelf exhibit strong spatial and temporal variability, making this a complex four-dimensional region of study (Wright, 1995). Within this relatively shallow setting, frictional forces are important in connecting hydrodynamics to the behavior of seabed forms of varying scale (Grant and Madsen, 1986). In a strongly bidirectional manner, the bottom boundary layer structure depends heavily on the morphology of the seabed that in turn is shaped by gradients in the hydrodynamics (Wright, 1995). Furthermore, numerous studies of shelf settings around the world (Trembanis et al., 2004; Schwab et al., 2000; Drake, 1999; Wright et al., 1999; Riggs et al., 1998; Thieler et al., 1995; Cacchione and Drake, 1990) have documented that complexity is more the norm than the exception. RIPPLES AND BEDFORMS Ripples and other large bedforms on the shelf (e.g., sandwaves and subaqueous dunes) are important sources of seabed roughness to waves and currents (Ardhuin et al., 2002; Grant and Madsen, 1986) and play a key role in the nature and magnitude of sediment resuspension (Traykovski et al., 1999; Li et al., 1996; Cacchione and Drake, 1990). Several field and laboratory studies have been conducted in attempts to develop empirical formulae between ripple geometry (e.g., height, length, steepness) and flow conditions (Wiberg and Harris, 1994; Wikramanayake, 1993; Clifton and Dingler, 1984; Grant and Madsen, 1986; Nielsen, 1981; Miller and Komar, 1980). Under the typically irregular

Bedforms, coastal-trapped waves, and scour process observations flow conditions encountered in the field, the ability of these models to accurately predict observed ripple geometry has been shown to be rather poor (Trembanis and Traykovski, 2005; Trembanis et al., 2004; Doucette, and O’Donoghue 2002; Traykovski et al., 1999; Li et al., 1996; and Osborne and Vincent, 1993). In part, the reason for the poor agreement between field data and model estimates is that ripple geometries encountered in the field are often partially relict products of forcings from past events and not solely products of instantaneous hydrodynamic conditions. Both Traykovski et al. (1999) and Li and Amos (1999) observed significant hysteresis in ripple development on the shelf. In addition to non-equilibrium evolution, spatially varying grain size is another important issue affecting ripple dynamics (Green et al., 2004; Trembanis et al., 2004). Grain size and ripple dimensions on the shelf often exhibit large variations over spatial domains both greater than 1 km (e.g., Green et al., 2004; Hume et al., 2000; Black and Oldman, 1999; Barnhardt et al., 1998; and Field and Roy, 1984) and less than 1 km (e.g., Trembanis et al., 2004; Ardhuin et al., 2002; Thieler et al., 1995; Hunter et al., 1988; and Schwab and Molnia, 1987), often in correlation with sorted bedforms (previously termed “rippled scour depressions”) that have wavelengths much longer than the ripples themselves (Green et al., 2004; Traykovski and Goff, 2003). According to Holland et al. (2003), heterogeneous patches of contrasting sediment grain size are frequently encountered along shelf environments around the world. SCOUR Scour is the morphodynamic response of the seabed as a result of the presence of an object or structure that disturbs the fluid flow (Soulsby, 1998). Scour is important for a variety of marine situations including bridge piers, dock pilings, breakwaters, oil platforms, offshore pipelines, marine artifacts, heterogeneous seabed bedforms, and naval mines (Whitehouse, 1998). The presence of an object on the seabed produces local flow acceleration due to continuity and thus drives a flux of local sediment and concomitant bed adjustments (Whitehouse, 1998). Another manifestation of scour is an increase in bed shear stress and turbulence as structured vortices are generated and released from around the object (Trembanis et al., 2007; McNinch et al., 2006). Scour can be classified both in terms of spatial extent and hydrodynamics. Three spatial classes of scour are defined as: “local scour” which is in the immediate vicinity of the object (on the order of meters), “global scour” composed of wide depressions around large or multiple objects (on the order of 10s of meters), and “overall seabed movement” associated with large scale (100s–1000s of meters) patterns of erosion, deposition, and bedform movement (Whitehouse, 1998). In terms of hydrodynamic intensity, scour is classified as either clear-water, when the ambient flow (bed shear stress) is below threshold velocity, or live-bed, when ambient flow is above threshold velocity and the entire bed is active. In the former, the amplification of flow

167

about the object induces transport locally but elsewhere the bed is immobile. In the latter case, sediment is being transported by flow everywhere, but especially near the object, where turbulence and bed shear stress are enhanced (Trembanis et al., 2007; McNinch et al., 2006). Once an object is exposed on the seabed, scour is initiated around the lateral ends of the object because of converging accelerated flow. This convergence leads to progressive erosion of the sediment from under the ends of the object, forming an expanding scour pit and shrinking the support pedestal. The object then will settle into its scour pit in a series of rocking and rolling motions until it is no longer protruding above the ambient seabed or until flow conditions subside and backfilling (deposition) ensues (see McNinch et al., 2006, and figures therein). In non-steady flows, periods of excavation (scour) will normally be interrupted by episodes of backfilling (deposition) (Trembanis et al., 2007; Richardson and Traykovski, 2002; Fredsoe, 1978). A possible sequence for scour around a free settling horizontal object, such as those examined in this study, is illustrated in McNinch et al. (2006; their fig. 4). HYDRODYNAMICS AND BEDFORMS Like no other large body of water in the world, the Black Sea (Fig. 1) has an upper oxygenated layer and a lower anoxic layer. The interface between the layers reaches down to 180 m depth along the coastal margins, and 500 m near the Bosporus (Özsoy and Ünlüata, 1997; Neretin et al., 2001). This interface appears to be unstable and at varying times, probably during severe weather conditions, creates internal waves that break upon the Black Sea’s continental shelf along the oxic/anoxic boundary. The surface layer in the Black Sea averages 18‰ salinity and 22 °C, and is highly oxygenated. The surface circulation features a cyclonic rim current about the entire basin. Two cyclonic gyres occur within the outer rim current, as well as eddies and intermittent convection to intermediate depths by surface cooling. The transition from the surface to the denser, trapped deep layer is marked by an oxycline with steep gradients in salinity, temperature, and chemical content. The deep-water averages 22‰ salinity and 8 °C and is completely anoxic while being rich in hydrogen sulfide (H2S) and ammonium (Özsoy and Ünlüata, 1997). The density contrast between these layers is large enough to support the propagation of internal waves. Internal waves with periods of six minutes and more, and with bottom velocities on the order of 30 cm/s, fast enough to exhume and carry suspended silt and fine sand (Hjulström, 1935; Prothero and Schwab, 2004), have been observed on the southern coast of Crimea (Filonov, 2000). When the crest of such a wave meets the continental shelf like that of the Danube delta in the northwest, the anoxic, H2S-rich water runs up and down the slope in a manner precisely analogous to the swash and backswash of a surface wave on a beach—whether in linear or turbulent fashion—and almost certainly has a large effect on the benthic environment. The anoxia and high H2S content of the wave water would kill most or all organisms it

168

Trembanis et al.

contacts. In certain areas bottom topography may contribute to a rapid enough backswash to cause sediment entrainment, erosion, and the development of bedforms (McPhee-Shaw, 2006). Internal waves have two important geoarchaeological effects. The first is the creation of a mixed layer between 85 and 185 m that periodically leads to anoxic water conditions. These anoxic water conditions are of interest to archaeologists since they lead to the long-term preservation of ancient wooden shipwrecks (Coleman and Ballard, 2004). The second is the creation of bedforms characterized by large sand waves that lay in a depth zone of 85–185 m. These bedforms are of interest because: • the existence of these shelf-sorted bedform features is poorly documented and not well understood; • the hydrodynamics responsible for these bedforms may help transport anoxic water across the shelf; and • the introduction of anoxic water onto the continental shelf by internal waves may preserve ancient wooden shipwrecks at far shallower depths than previously thought. FIELD SITE AND RESEARCH METHODS The field site was located off the southwest coast of the Crimean Peninsula of the Northern Black Sea (Fig. 1). In

2006, a geoacoustical survey was conducted of the southwestern Crimean shelf and slope (Fig. 2) along a suspected deepsea trade route between the Bosporus and the Crimea. A number of side-scan sonar targets were identified that were later inspected with a remotely operated vehicle. Several of the targets turned out to be modern shipwrecks and aircraft (mostly Russian from the Crimean War, World War I, and World War II eras), but one of the targets, located ~23 km off the coast, was a pile of ancient ceramic jars. Based on the typology of similar-looking jars from the Chersonesos site (Ryzhov and Sedikova, 1999) and elsewhere around the Black Sea region (e.g., Hayes, 1992), these jar types are estimated to date between the ninth and eleventh centuries C.E., placing the ship in the early Medieval Period. At a depth of ~150 m, this shipwreck lies above the normal anoxic interface but within the mixed layer of temporally varying dissolved oxygen content. It has thus escaped the ravages of marine borers and appears to be in an unusually good state of preservation. In general, this mixed layer is prevalent throughout the entire Black Sea basin along the shelf break between depths of ~80–180 m (Ballard et al., 2001). The side-scan sonar data collected in 2006 forms the basis of the large-scale bedform mapping presented in this paper (Figs. 2 and 3).

Scale: 300 km

Scale:

10 km

Figure 1. Study-site location map (star) off the SW coast of the Crimean peninsula at a local depth of ~135 m. Inset illustrating the side-scan sonar surveys in the vicinity of the bottom mooring associated with the shipwreck site known as Chersonesos A (star).

Bedforms, coastal-trapped waves, and scour process observations SENSOR CONFIGURATION AND SAMPLING REGIME In addition to the side-scan sonar surveys conducted in 2006 (Figs. 2 and 3), a set of bottom-mounted instruments (Fig. 4) were deployed in 2007 measuring currents, temperature, salinity, and seabed geometry. The bottom-mount (Fig. 4A) was located at a depth of 135 m and included an upwardlooking acoustic Doppler current profiler (ADCP), conductivitytemperature (CT) logger, and rotary fanbeam sonar. The CT sensor provides a time-series point measurement of the ambient salinity and temperature used to determine changes to the density of the surrounding water and the passage of thermocline or pycnocline oscillations (Fig. 6). The ADCP gathers vertical profiles of three-dimensional hydrodynamic flow (Fig. 7) recorded in earth-coordinates (i.e., east-west, north-south, up-down) based on an internal compass. The rotary fanbeam sonar (Fig. 4B) obtains a high-resolution planview image of the seabed surrounding the bottom mount to a range of 9 m thus providing a

A

169

time-lapse picture (Fig. 5) of the small-scale bedform geometry and dynamics in for comparison to the hydrodynamic measurements. Table 1 summarizes the sampling scheme settings for each of the bottom-mount instruments. On 15 August 2007, the ADCP/CT/Sonar bottom mount was deployed in the vicinity of the Chersonesos A wrecksite at a local depth of 135 m and was recovered after ~38 h. Upon recovery, data was downloaded and archived. The data analysis results are presented below. RESULTS AND DISCUSSION Salinity and Temperature Record Figure 6 illustrates the recorded time series of temperature (blue line) in °C and salinity (green line) in ‰ at a height of 0.75 m above the bed. The Sea-Bird SBE-37SM sensor recorded every 23 seconds, the fastest sampling scheme that would cover the expected deployment duration. The sharp drop in temperature

C

B

60 km

B 200 m

B C 10 km

C 4 km Figure 2. Map plot illustrating several hundred of the more than 2300 bedforms digitized in geographic information systems. The crest line of each visible bedform was digitized from the side-scan sonar survey data. Insets are as follows: (A) Cape Sarych and the surrounding vicinity; (B) Digitized bedform crests in a 10 km box surrounding the Chersonesos A wreck site; (C) Close-up portion of the survey site showing digitized bedforms (blue lines) over the side-scan sonar record. Star illustrates location of the acoustic Doppler current profiler/conductivity-temperature sensor/Rotary sonar bottom mooring.

170

Trembanis et al.

and rise in salinity at the beginning of the deployment represents the measurements as the instrument mount was descending through the water column during deployment. A reverse signal is seen at the end of the deployment when the bottom mount was recovered from the seabed. During the interval between descent and recovery, the CT sensor recorded essentially static values for salinity and temperature. The mean temperature was 8.3 °C with a standard deviation of 0.03 °C. Mean salinity was 20.6‰ with a standard deviation of 0.08‰. The slight changes in the values during the deployment and the sharp changes at the beginning and the end confirm that the sensor was in fact working and that the nearly flat line plots of temperature and salinity were in fact real. We initially expected sharp fluctuations in temperature and salinity coinciding with the interface of the pycnocline oscillating about the depths of the deployment site with a high-frequency interval, which might have come from internal waves breaking across the shelf. The absence of these high-frequency oscillations in our data does not imply that these waves do not exist but simply that we did not observe them during our short deployment, which

A

was conducted during fair-weather conditions rather than a storm period when internal waves would most likely occur. Salinity and temperature measurements illustrate the general background conditions for this site. If our measurements are well and truly above or below the pycnocline, it could be that the signature of the internal waves (if present) are not reflected in the temperature and salinity signals but rather in the mean current flow, examined in the next section. Current Structure and Velocity Time-Series The key critical observations of mean current structure and time-series behavior are illustrated in Figures 7 and 8. The 300 kHz ADCP used in this study was set to the smallest vertical bin size (2 m) and a rapid profile repeat rate (30 second interval) in order to maximize the vertical and temporal resolution of the measurements with hopes of encountering high-frequency internal waves. Each 30-second burst represents an ensemble average of 16 individual acoustic pings over a vertical range from 4.2 m

B

Figure 3. Side-by side comparison of side-scan sonar record showing raw georeference image (A) and geographic information system digitized bedforms (B). Star illustrates location of the acoustic Doppler current profiler/conductivity-temperature sensor /Rotary sonar bottom mooring.

Bedforms, coastal-trapped waves, and scour process observations above the bed to over 80 m above the bed, with horizontal and vertical measurements occurring every 2 m. A resulting color contour plot of the magnitude of the velocity (i.e., speed) of the horizontal current component, i.e., the Pythagorean addition of the east-west and north-south components of the flow, is illustrated in Figure 7. While a great deal of vertical and temporal structure exists in this figure, we will limit our discussion of the current velocity structure near to the bed (