Holocene Palaeoenvironmental History of the Central Sahara: Palaeoecology of Africa Vol. 29, An International Yearbook of Landscape Evolution and Palaeoenvironments ... of Africa and the Surrounding Islands)

HOLOCENE PALAEOENVIRONMENTAL HISTORY OF THE CENTRAL SAHARA

Palaeoecology of Africa International Yearbook of Landscape Evolution and Palaeoenvironments Volume 29

Editor in Chief J. Runge, Frankfurt, Germany

Editorial board

G. Botha, Pietermaritzburg, South Africa K.W. Butzer, Austin, Texas, USA E. Cornellissen, Tervuren, Belgium F. Gasse, Aix-en-Provence, France P. Giresse, Perpignan, France S. Kröpelin, Köln, Germany T. Huffmann, Johannesburg, South Africa E. Latrubesse, La Plata, Argentina J. Maley, Montpellier, France J.-P. Mund, München, Germany D. Olago, Nairobi, Kenya F. Runge, Altendiez, Germany L. Scott, Bloemfontein, South Africa I. Stengel, Pretoria, South Africa F.A. Street-Perrott, Oxford, UK M.R. Talbot, Bergen, Norway

Holocene Palaeoenvironmental History of the Central Sahara

Editors

Roland Baumhauer Department of Geography, Physical Geograpy, Julius-Maximilians University, Würzburg, Germany

Jürgen Runge Centre for Interdisciplinary Research on Africa (CIRA/ZIAF), Johann Wolfgang Goethe University, Frankfurt am Main, Germany

Front cover: Dead tree at the Plateau du Mangueni, NE-Niger. It symbolizes the changes in climate and landscape in Central Sahara caused by aridisation since the Early Holocene. Photograph by Jan Krause, Department of Earth Sciences, Physical Geography, Freie Universität Berlin, Germany. Financially Supported by DFG (German Research Foundation)

CRC Press/Balkema is an imprint of the Taylor & Francis Group, an informa business © 2009 Taylor & Francis Group, London, UK Typeset by Vikatan Publishing Solutions (P) Ltd., Chennai, India Printed and bound in Great Britain by TJ International Ltd., Padstow, Cornwall All rights reserved. No part of this publication or the information contained herein may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, by photocopying, recording or otherwise, without written prior permission from the publisher. Although all care is taken to ensure integrity and the quality of this publication and the information herein, no responsibility is assumed by the publishers nor the author for any damage to the property or persons as a result of operation or use of this publication and/or the information contained herein. Published by:

CRC Press/Balkema P.O. Box 447, 2300 AK Leiden, The Netherlands e-mail: [email protected] www.crcpress.com – www.taylorandfrancis.co.uk – www.balkema.nl

Library of Congress Cataloging-in-Publication Data Holocene palaeoenvironmental history of the Central Sahara / editors, Roland Baumhauer, Jürgen Runge. p. cm. Includes bibliographical references and index. ISBN 978-0-415-48256-1 (hardcover : alk. paper) 1. Palaeoecology -- Sahara. 2. Palaeoecology -- Holocene. 3. Palaeoecology -- Pleistocene. 4. Desert ecology -- Sahara -- History. I. Baumhauer, Roland. II. Runge, Jürgen, 1962- III. Title. QE720.2.S24H65 2009 560’.1793--dc22 2008045244 ISBN: 978-0-415-48256-1 (Hbk) ISBN: 978-0-203-87489-9 (e-book)

Contents

FOREWORD —Jürgen Runge

vii

IN MEMORIAM JOHANNA ALIDA COETZEE (1921–2007) —Klaus Heine

ix

PREFACE AND INTRODUCTION —Roland Baumhauer

xvii

CONTRIBUTORS

xxi

CHAPTER 1 GEOMORPHOLOGICAL AND PALAEOENVIRONMENTAL RESEARCH IN THE SOUTH-CENTRAL SAHARA IN REVIEW —Roland Baumhauer, Peter Felix-Henningsen & Brigitta Schütt CHAPTER 2 COMPARISON OF PROXY-BASED PALAEOENVIRONMENTAL RECONSTRUCTIONS AND HINDCAST MODELLED ANNUAL PRECIPITATION—A REVIEW OF HOLOCENE PALAEOENVIRONMENTAL RESEARCH IN THE CENTRAL SAHARA —Brigitta Schütt & Jan Krause CHAPTER 3 HOLOCENE PALAEOENVIRONMENTAL CHANGES IN CENTRAL SAHARA INFERRED FROM SEGGEDIM SCARP FOOT DEPRESSION (NE-NIGER) —Roland Baumhauer, Jens Brauneck, Barbara Sponholz, Erhard Schulz, Oumarou Faran Maiga, Ibrahim Sani & Simon Pomel CHAPTER 4 THE DESERT IN THE SAHARA. TRANSITIONS AND BOUNDARIES —Erhard Schulz, Abdelhakim Abichou, Aboubacar Adamou, Aziz Ballouche & Issa Ousseïni CHAPTER 5 PALAEO-CLIMATIC EVIDENCE OF SOIL DEVELOPMENT ON SAHELIAN ANCIENT DUNES OF DIFFERENT AGE IN NIGER, CHAD AND MAURITANIA —Peter Felix-Henningsen, Peter Kornatz & Einar Eberhardt

1

23

39

63

91

vi

Contents

CHAPTER 6 ARE THERE VALUABLE PEDOLOGICAL PALAEOENVIRONMENTAL INDICATORS IN NORTHERN CHAD? —Ludger Herrmann, Mohamed Mounkaila & Frieder Graef

107

CHAPTER 7 NEW DISCOVERY OF ROCK FULGURITES IN THE CENTRAL SAHARA —Barbara Sponholz

127

CHAPTER 8 FLUVIAL GEOMORPHOLOGY AND PALAEOHYDROLOGY OF A SMALL TRIBUTARY OF THE PLATEAU DE MANGUENI, NE NIGER —Jan Krause & Brigitta Schütt

137

CHAPTER 9 PALAEOECOLOGY OF THE GIANT CATFISH (ARIUS GIGAS, ARIIDAE) IN HOLOCENE SAHARAN AND TROPICAL WEST AFRICAN WATERS —Hélène Jousse & Wim Van Neer

159

CHAPTER 10 NEOLITHIC DOMESTICATION AND PASTORALISM IN CENTRAL SAHARA: THE CATTLE NECROPOLIS OF MANKHOR (TADRART ALGÉRIENNE) —Michel Tauveron, Karl Heinz Striedter & Nadjib Ferhat

179

CHAPTER 11 THE MICROSTRATIGRAPHY AND MICROMORPHOLOGY OF A HOLOCENE PALAEOLAKE IN SOUTHERN TUNISIA —Abdelhakim Abichou

187

CHAPTER 12 DIFFERENT DIMENSIONS OF RECENT VEGETATION DYNAMICS OF NORTH AND WEST AFRICA —Brian Beckers & Brigitta Schütt

197

CHAPTER 13 LANDMINES, DRUGS AND JUSTICE. THE RECENT HISTORY OF TWO SAHARAN MOUNTAINS (ADRAR DES IFORAS/MALI AND AIR MTS./NIGER) —Issa Ousseïni, Aboubacar Adamou & Erhard Schulz

221

CHAPTER 14 THE SAHELIAN AND SAHARAN DUNE SYSTEMS OF NIGER. A COMPARISON OF THEIR GRANULOMETRIC CHARACTERISTICS —Ibrahim Mamane Sani & Issa Ousseïni

239

REGIONAL/LOCATION INDEX

247

SUBJECT INDEX

249

Foreword

Volume 29 of the series ‘Palaeoecology of Africa’ on the ‘Holocene Palaeoenvironmental History of the Central Sahara’ continues the publishing of interdisciplinary scientific papers on landscape evolution and on former environments of the African continent, e.g. changes in climate and in vegetation cover interconnected to all kinds of environmental dynamics from the Cainozoic up to the present, including since the Late Quaternary the growing influence of humans in many of the study areas. Since its re-edition in 2007/2008 with a modified editorial board—attended by the ‘Frankfurt Centre for Interdisciplinary research on Africa’ (CIRA/ZIAF, www.ziaf.de)— and a completely new and up-to-date layout, assisted by the publishers group Taylor & Francis, the series has already begun to be again recognized by the scientific community. However, Palaeoecology of Africa is not as yet been accredited by international citation indices, what is regarded as a certain drawback for the series. As it is essential to change this in the future, the editor is already working on this often quite long and tedious process in being considered by citation indices. Young academics and upcoming university professionals are especially encouraged and invited to consider this traditional journal—which has been on the “market of science journals” since 1966—as a media to communicate, and as an opportunity for meetings in minds to exchange global knowledge and ideas across scientific and cultural borders (cf. Heine, 2007 in PoA 28). In the past (1970s to 1990s) quite regularly the scientific contributions of the Biennial SASQUA (South African Society for Quaternary Research) conferences were published within Palaeoecology of Africa. As a vision for future volumes the editor likes to propose that some follow-up editions of the series could have again a focus on SASQUA activities (“back to the roots”). The up coming SASQUA meeting is scheduled for September 2009 in Knysna, South Africa. Fourteen papers are gathered together in this volume focusing on the most recent (Holocene) dynamics of the Central Sahara. They are mainly the outcome of the since 2005 DFG funded LIMNOSAHARA project (www.limnosahara.de) of the Universities of Würzburg, Giessen, FU Berlin (Germany) and Niamey (Niger). Many thanks go to all colleagues for submitting their papers to Palaeoecology of Africa. Special thanks go to Roland Baumhauer for supporting the editing process. Formatting of the papers to the PoA layout was again reliably done by Erik Hock to whom I am most grateful. Ursula Olbrich revised numerous figures and assisted by carrying out cartographic work on the book. The Taylor & Francis team in Leiden (The Netherlands) with senior editor Janjaap Bloom supported in a professional manner the editorial work. Finally, I like to thank the Deutsche Forschungsgemeinschaft (DFG) for assisting in printing this volume. Jürgen Runge Bangui and Frankfurt August 2008

In Memoriam

Johanna Alida Coetzee (1921–2007) Klaus Heine University of Regensburg, Institute of Geography, Regensburg, Germany It is with deep regret that I have to report that Professor Joey Coetzee died in Somerset West on April 28, 2007. Quaternary and Palynological Research has suffered a great loss. We mourn a colleague of national and international distinction, a brilliant and honest scientist, who advanced the field of palynology in Southern and Eastern Africa. She has published palaeoenvironmentally relevant groundbreaking and pioneering investigations on African pollen of Cainozoic age. Joey Coetzee was educated at Jeppe High School for Girls in Johannesburg and then University of Witwatersrand. She earned her master’s degree in 1946 and subsequently worked at the Universities of Wits and Natal for a short time. In August 1946, she accepted a position as assistant in the Department of Botany at the University of the Orange Free State. Inspired by Eduard van Zinderen Bakker’s pollen analytical investigations in East Africa, she embarked in testing the widely held hypothesis that pluvial phases correlate with glacial phases during the Last Glacial Maximum. Joey Coetzee devoted several years of intense work on these problems as a senior member of the research team of Eduard van Zinderen Bakker Sr. who headed the Department of Botany at the University of the Orange Free State and the Palynological Research Unit of the South African Council for Scientific and Industrial Research. For some time during 1953, Joey Coetzee worked in the Palynological Laboratory, Stockholm-Bromma, and also visited the colleagues in Bergen, Velp, Utrecht, and Cambridge. In the 1950s she worked on the pollen morphology of Southern African species and completed a collection and description of several thousand pollen grains and spores. It was the real base for any palynological study in Southern Africa. Furthermore, she analysed the air-borne pollen collected weekly for three years at seven different stations in South Africa and Namibia. The data of these pollen traps gave valuable information on the pollen spectra of the main vegetation types (e.g. Karoo) in their natural condition. In 1959 Joey Coetzee made a trip to East Africa, collected many pollen samples at the East African Herbarium at Nairobi and studied fossil pollen material from Uganda. In the following years she started her investigations of cores collected on Mount Kenya. All these and many other pioneering studies of Joey Coetzee marked the progress in the achievement of palynology and of understanding the past. In 1964 Joey Coetzee had to leave her study to concentrate on her East African pollen analytical research. The core from Sacred Lake was finished and showed remarkable results (Coetzee, 1964). Joey Coetzee’s findings led to the approval of the name Mount Kenya Hypothermal for the pollen zone which indicates a synchronous drop in temperature in many parts of Africa during the Upper Pleistocene. The general features of her diagram were a sound proof of the world-wide changes which occurred in the climate of the earth. Later, in her D.Sc.-Thesis (Coetzee, 1967) Joey Coetzee showed that the hypothesis ‘glacial phases = pluvial phases in tropical Africa’ was not true in Eastern and Southern Africa. This work, recently featured as a classic work in the journal Progress in Physical Geography, and probably the most cited palynological paper from the African continent, illustrates ‘the very essence of Quaternary palaeoenvironmental reconstruction and, indeed of physical

x

In Memoriam

geography’ (Meadows, 2007). Joey Coetzee rapidly acquired a reputation as an eminent researcher of palaeoenvironmental studies in Africa. She was among the first scientists to recognize the significance of changes in temperatures during the Last Glacial Maximum in tropical mountain areas of Africa. It was 1973 when I first met Joey Coetzee. The location was the INQUA conference in Christchurch, New Zealand, in the shadow of the glacially shaped Southern Alps. Together with Eduard van Zinderen Bakker Sr. she presented a paper on Global Temperature Changes and the African Quaternary Environment (van Zinderen Bakker and Coetzee, 1973). After their talk I discussed with Joey the results: During the Quaternary, Africa has been subjected to a wide range of variations in humidity which for a long time have been considered to have been in-phase over the entire continent. The pattern of rainfall distribution and the evaporation rate, however, depend on global and local temperature conditions. The longterm variations in the earth’s energy budget are therefore the primary cause of changes in the Quaternary environment. Correlations of radiometrically calibrated temperature curves are consequently of basic importance for the understanding of Quaternary chronology. Vegetation changes which occurred in the interior of Southern Africa during Late Glacial times show that detailed correlations existed between lower temperature and higher humidity versus higher temperature and lower humidity. This result indicates that climatic settings in this region differed fundamentally from that of tropical Africa (Coetzee, 1967) and that ‘pluvial conditions’ cannot be used for correlation purposes. In those days, my own chronological investigations of glacial deposits of the Mexican volcanoes showed that glacial climatic evidence is suitable for stratigraphic correlations neither (Heine, 1974) and thus my records from the New World tropics added to Joey Coetzee’s observations. Since that time, Joey and I were friends and colleagues. I met Joey many times in Southern Africa during conferences of the Southern African Society for the Quaternary (SASQUA), in her home at Bloemfontein and, after her retirement from the university, at Somerset West. Several times we spent weeks together with colleagues from South Africa and Germany in the field, especially in the Kalahari and the Namib Desert (Figure 1). And we met in Europe, in the Austrian Alps, and in my home. One of the most debated topics in Quaternary science which Joey and I discussed was methodological problems arising from the reconstruction of Late Quaternary palaeoenvironments by interpreting pollen data. The ‘palynological records’ of the Younger Dryas climatic oscillation, a millennium-long cooling event approx. 12.700–11.500 cal yr B.P. that interrupted the transition from the last glacial to present interglacial (Holocene) period, were subject to much debate in Southern Africa, Northern and Southern South America and New Zealand. Two conflicting hypotheses, both based on palynological sequences, have been proposed to document either this cooling event in the Southern Hemisphere (e.g. Heusser and Rabassa, 1987) or to question it (e.g. Singer et al., 1998). Joey’s fundamental knowledge of the methodology of pollen research advanced my understanding of Quaternary vegetation changes and the palaeoclimatic implications, not only in Eastern and Southern Africa, but also in other regions of the earth. Despite the fact that Joey Coetzee anticipated finding evidence of the Younger Dryas in East Africa and described and correlated particularly marked temperature fluctuations in South Africa (Aliwal North site) during the Late Glacial with the European older Dryas, Allerød and Younger Dryas, she listened to and respected my diverse views about the existence of the Younger Dryas cooling in the Southern Hemisphere. Another topic we argued about many times was the ‘Pluvial Theory’, established by Richard Forster Flint (1957) who suggested that the major periods of Northern Hemisphere ice advance were associated with phases of more humid climates in tropical Africa. Joey Coetzee’s reconstruction of the LGM temperature depression of between 5,1 and 8,8 °C and more arid conditions relative to the present day in tropical Africa (Coetzee, 1967;

In Memoriam

xi

Figure 1. Rest during field work at Sossus Vlei, Namib Desert, Namibia. From left to right: Joey Coetzee, Louis Scott (in the back), Eduard van Zinderen Bakker Sr., Almut Heine. Photo: K. Heine, June 9, 1981.

van Zinderen Bakker and Coetzee, 1972) questions the ‘Pluvial Theory’ and corroborates my observations from tropical Mexico. What came through during our discussions was Joey’s enthusiasm, and the exhilaration and intellectual stimulation of working on a collaborative interdisciplinary research field. In the 1970s and 1980s, Joey Coetzee began research work to explore links between the Tertiary pollen records and the unique fynbos vegetation in the Cape region (Coetzee, 1978a, b, 1993; Coetzee and Rogers, 1982; Coetzee and Muller, 1984; Coetzee and Praglowski, 1984, 1988). She found that fynbos replaced palm-dominated subtropical to tropical woodlands that alternated with conifer forests during the Neogene and linked the process to global cooling and Antarctic glacial history (Scott, 2007). Joey Coetzee edited Palaeoecology of Africa and of the surrounding islands and Antarctica, volume 10 to 17 (1978–1987) in collaboration with E.M. van Zinderen Bakker Sr. and volume 18 (1988). While volume 1 (1966) of Palaeoecology of Africa was devoted to reports which have been published during the years 1950–1963 under the title of ‘Palynology in Africa’ and deal with the initial research done on palynology in South Africa, Volume 2 (1967) showed that pollen analytical research could be linked with so many related disciplines that the entire field was better covered by the title ‘Palaeoecology’. The volumes 1 and 2 show the achievements in African palynology, Joey Coetzee’s very great importance and her contributions to the rapid development of palynological fundamentals during the years 1950–1965. Volume 3 (1967) was entirely devoted to the D.Sc.-Thesis by Joey Coetzee on pollen analysis in East and South Africa. It was this volume of Palaeoecology of Africa that made the series of the ‘small yellow books’ (Nicole

xii

In Memoriam

Petit-Maire, Durban Febr. 1, 1989) well-known among African palaeoenvironmentalists. In the 1970s and 1980s, Palaeoecology of Africa became a forum for research results of scientists from various countries studying the palynology, palaeontology, biogeography, glacial geology, volcanology, geomorphology, oceanography, limnology, (palaeo-) climatology, archaeology, anthropology as well as theories on ice ages. In the publication process, Joey Coetzee distinguished herself as a personality with high principles and an intense interest in scientific communication and interchange of ideas. Joey Coetzee served on many committees in connection with Quaternary science and palynology. She was one who always gave credit to others rather than to herself. Through her far-ranging interests and achievements, she has left an enduring imprint on African palynological science. A world-class scientist, a woman of great integrity and an inveterate friend of animals, she was above all a nice fellow who just enjoyed nature in all its facets wherever she stayed. REFERENCES Coetzee, J.A., 1964, Evidence for a considerable depression of the vegetation belts during the Upper Pleistocene on the East African mountains. Nature, 204, pp. 564–566. Coetzee, J.A., 1967, Pollen analytical studies in East and Southern Africa. Palaeoecology of Africa, 3, pp. 1–146. Coetzee, J.A., 1978a, Climatic and biological changes in South-Western Africa during the Late Cainozoic. Palaeoecology of Africa, 10, pp. 13–29. Coetzee, J.A., 1978b, Late Cainozoic palaeoenvironments of Southern Africa. In: E.M. van Zinderen Bakker (ed.), Antarctic glacial history and world palaeoenvironments. A.A. Balkema, Rotterdam, pp. 110–127. Coetzee, J.A., 1993, African flora since the Terminal Jurassic. In: P. Goldblatt (ed.), Biological Relationships between Africa and South America. Yale University Press, New Haven and London, pp. 37–61. Coetzee, J.A. and Muller, J., 1984, The phytogeographic significance of some extinct Gondwana pollen types from the Tertiary of the Southwestern Cape (South Africa). Annals of the Missouri Botanical Garden, 71, pp. 1088–1099. Coetzee, J.A. and Rogers, J., 1982, Palynology and lithological evidence for the Miocene palaeoenvironment in the Saldanha region (South Africa). Palaeogeography, Palaeoclimatology, Palaeoecology, 39, pp. 71–85. Coetzee, J.A. and Praglowski, J., 1984, Pollen evidence for the occurrence of casuarina and myrica in the Tertiary of South Africa. Grana, 23, pp. 23–41. Coetzee, J.A. and Praglowski, J., 1988, Winteraceae pollen from the Miocene of the Southwestern Cape (South Africa). Relationship to modern taxa and phytogeographical significance. Grana, 27, pp. 27–37. Flint, R.F., 1957, Glacial and Pleistocene Geology. John Wiley, New York, pp. 1–553. Heine, K., 1974, Bemerkungen zu neueren chronostratigraphischen Daten zum Verhältnis glazialer und pluvialer Klimabedingungen. Erdkunde, 28, pp. 303–312. Heusser, C.J. and Rabassa, J., 1987, Cold climatic episode of Younger Dryas age in Tierra del Fuego. Nature, 328, pp. 609–611. Meadows, M., 2007 and Coetzee, J.A., 1967, Pollen analytical studies in East and Southern Africa. Palaeoecology of Africa 3, pp. 1–146. (Classics in physical geography revisited). Progress in Physical Geography 31, pp. 313–317. Scott, L., 2007, Professor Joey Coetzee 1921–2007. Review of Palaeobotany and Palynology, 147, pp. 1–2.

In Memoriam

xiii

Singer, C., Shulmeister, J. and McLea, B., 1998, Evidence against a significant Younger Dryas cooling event in New Zealand. Science, 281, pp. 812–814. van Zinderen Bakker, E.M. and Coetzee, J.A., 1972, A re-appraisal of Late-Quaternary climatic evidence from Tropical Africa. Palaeoecology of Africa, 7, pp. 151–181. van Zinderen Bakker, E.M. and Coetzee, J.A., 1973, Global temperature changes and the African Quaternary environment. 9th INQUA Congress, Christchurch, New Zealand, 2–10 Dec. 1973, Abstracts, pp. 1–385.

Schematic vegetation map of Northern and Western Africa (for text see chapter 4, pp. 63–89).

Preface and Introduction

The environmental setting within the Central Sahara was subject to considerable changes during Late Quaternary, mainly driven by major global climate variations, although human impact increased constantly since Early Holocene. Such global events can be reconstructed with the help of reliefs, sediments and palaeosoils and their specific morphological, chemical and mineralogical properties. The 14 papers gathered together in this volume are mainly the outcome of the interdisciplinary German research project LIMNOSAHARA (www.limnosahara.de) financed by the German Research Foundation (DFG). The investigations were carried out by a collaboration of multiple disciplines, reaching from physical geography, palaeopedology and palaeolimnology to palynology and prehistory. The project’s focus is to ascertain new and established data on climate variations and associated palaeoenvironmental changes within the Central Sahara and to systematically collate and correlate them to results obtained from the Afro-Asian dry land belt and adjacent areas. The joint analysis of Late Quaternary landscape development and present environmental conditions in the Central Sahara will result in the modelling of Late Pleistocene and Holocene palaeoenvironments, emphasising various aspects. This will be achieved by transferring the highly localized information obtained from palaeolake sediments to the region by means of spatially high-resolution information about the morphodynamic processes currently shaping the landscape and the factors controlling them. The first—introductory—chapter gives an overview of the geomorphological and palaeoecological research in south Central Sahara on the basis of a literature review. Subsequently different proxy-data sources are introduced and discussed to conclude on the former dynamics of the palaeoenvironment of the central part of Sahara. The following chapter turns towards two aspects: on the one hand it summarizes and discusses more than 50 original publications on climate proxy- and on model-data from the Central Sahara; on the other hand, it compares the findings from the proxies with numerically modelled data by Kutzbach and Guetter’s CCM0 estimations for African precipitation over time slices at 3, 6 and 9 ka. The presented combination of “empirical” versus “modelled” data is interesting as nowadays the discussion often is dominated by the model approach exclusively. In chapter 3 extensive work on former lake/sebhka-sediments and fossil soils in the Seggedim region gave further evidence of Holocene palaeoenvironmental changes in an up to now not well explored region of NE Niger. In the paper a new 15 meter long core gained in 2005 is analysed and discussed by an interdisciplinary approach using geomorphological, sedimentological and geochemical investigations. A trend to more humid conditions at the beginning of the Holocene changing later on—while getting drier—into a sebkha environment with alternating salt and sand layers was evidenced by the findings. The fourth paper by Erhard Schulz and colleagues from Niger, Tunisia and France is dealing in the way of a review on different approaches of how to define and of how to characterize transitions, limits or boundaries of the Saharan desert. The authors mainly outline these by the vegetation and floristic content of the landscapes which are described in detail. The contribution illustrates the sensitivity of ecological margins due to former and future changes in climate.

xviii

Preface and Introduction

Figure 1. Location map of study sites.

The subsequent chapters five and six are concerned with (palaeo)pedological investigations. The paper of Peter Felix-Henningsen and members of his working group (Peter Kornatz, Einar Eberhardt) examines different soils developed on nowadays inactive dune systems in Mauretania, Niger and Chad as a proxy data source for palaeoclimatic interpretation. Besides conventional pedological analysis on the soil profiles also OSL dating was carried out. The overall results are mostly confirming the findings by other

Preface and Introduction

xix

papers dealing with the course of Holocene environmental conditions in the Sahara. The second pedological paper gives an overview of a pedologically directed transect in Northern Chad in general search for suitable palaeoenvironmental indicators. Based on extended earlier research on fulgurites that were formed by lightning strike to silicate rocks, the author introduces in chapter 7 two new observations on rock fulgurites from the Central Sahara. In contrast to dune sand fulgurites that indicate monsoonal thunderstorms up to 18°N in the Sahara during the Mid-Holocene, the observed rock fulgurites are not necessarily correlated to the palaeoenvironmental conditions as they could have been formed by lightning events in the course of ephemeric thunderstorm events. On the basis of former geomorphological field work undertaken in the 1980 by the Würzburg working group, the so-called “Seeterrassental”—a catchment with only 31,5 km2 in size—in the Mangueni Plateau of NE Niger has been revisited by Jan Krause and Britta Schütt (chapter 8) using modern differential GPS measurements and SRTM data to reconstruct Holocene flow velocity and discharge, also applying different hydraulic equations on three identified terrace levels. The palaeoecological and ichthyological paper of Hélène Jousse and Wim Van Neer (chapter 9) studies the recent and former distribution of the Giant Catfish (Arius gigas Ariidae) in Westafrica, especially in the Niger basin. On the basis of the examination of a rare skeleton specimen from the British Museum anatomy of this species is studied and documented. The Holocene and the recent decline of this fish population is discussed from different perspectives considering former climate changes and also the influence of humans (subsistance fishing) in the region. In chapter 10 (Michel Tauveron and Karl Heinz Striedter) numerous findings of cattle bones and one almost complete cattle skeleton which had not been clearly identified as Bos taurus or Bos primigenious in the South-Eastern Algerian Sahara are suggesting palaeoenvironmental interpretations for the Holocene. Abdelhakim Abichou shows in his contribution (chapter 11) micromorphological investigations and cartographic surveys. In connection with some radiocarbon data the results allow the reconstruction of Holocene palaeoenvironmental conditions in the sebka Erg el Makhzen in Southern Tunisia. Chapter 12 is not mainly palaeoenvironmentally adjusted. Brian Beckers and Britta Schütt examine by an actualistic approach vegetation dynamics by using NDVI data and GBCP rainfall estimates on a huge transect running from North to West Africa. It underlines the recent sensitivity of the Central Saharan landscape that might have been in a way similar to the palaeoenvironmental conditions. Aside of variations in rainfall also human induced effects on vegetation dynamics are considered. Finally, the authors of chapter 13 (Issa Ousseïni, Aboubacar Adamou and Erhard Schulz) describe and discuss recent geoecological and environmental modifications in the Adrar des Iforas (Mali) and the Air (Niger) mountains that took and still take place against the backdrop of a civil war (the so-called “rebellion of Tuareg”). Therefore, this contribution is a combination of socio-economic and geopolitical with environmental and landscape issues. It is tried to show how environmental factors and dynamics of the environment can be linked to economic and political trends. Several very complex graphs/sketch maps represent time slices from 1910 to 2007 that are driving at a better understanding of this interdisciplinary approach. Another contribution (chapter 14) from the colleagues of the University of Niamey (Ibrahim Mamane Sani and Issa Ousseïni) makes an approach to understand the granulometric characteristics of the dune systems of Niger. Many thanks go to all colleagues for submitting their papers. Thanks are also due to the German Research Foundation (DFG) for its financial support to realize this publication. Special thanks go to Jürgen Runge for publishing in PoA. Roland Baumhauer Würzburg July 2008

Contributors

Abdelhakim Abichou Département de Géographie, CGMED Tunis, Faculté des Lettres et des Sciences Humaines de Tunis, BP 1123, Tunis, Tunisie. Email: [email protected]. Aboubacar Adamou Département de Géographie, Faculté des Lettres et des Sciences Humaines, Université Abdou Moumouni, BP 435, Niamey, République du Niger. Email: [email protected]. Aziz Ballouche Laboratoire “Paysages & Biodiversité”, Université Angers, F-49045 Angers Cedex 1, France. Email: [email protected]. Roland Baumhauer Department of Geography, Physical Geography, Julius-Maximilians University, Am Hubland, D-97074 Würzburg, Germany. Email: [email protected]. Brian Beckers Department of Earth Sciences, Physical Geography, Freie Universität, Berlin, Malteserstr. 74-100, D-12249 Berlin, Germany. Email: [email protected]. Jens Brauneck Department of Geography, Physical Geography, Julius-Maximilians University, Am Hubland, D-97074 Würzburg, Germany. Email: [email protected]. Einar Eberhardt Bundesanstalt für Geowissenschaften und Rohstoffe, Stilleweg 2, D-30655 Hannover, Germany. Email: [email protected]. Peter Felix-Henningsen Institute of Soil Science and Soil Conservation, Justus Liebig University Giessen, Heinrich-Buff-Ring 26, D-35390 Giessen, Germany. Email: Peter.Felix-H@umwelt. uni-giessen.de. Frieder Graef Institute of Soil Science and Land Evaluation (310), University of Hohenheim, D-70593 Stuttgart, Germany. Email: [email protected]. Klaus Heine Institute of Geography, University of Regensburg, Universitätsstr. 31, D-93053 Regensburg, Germany. Email: [email protected]. Ludger Herrmann Institute of Soil Science and Land Evaluation (310), University of Hohenheim, D-70593 Stuttgart, Germany. Email: [email protected]. Hélène Jousse Naturhistorisches Museum Wien, Säugetiersammlung, Burgring 7, A-1010 Wien, Austria. Email: [email protected].

xxii Contributors

Peter Kornatz Institute of Soil Science and Soil Conservation, Justus Liebig University Giessen, Heinrich-Buff-Ring 26, D-35390 Giessen, Germany. Email: peter.kornatz@umwelt. uni-giessen.de. Jan Krause Department of Earth Sciences, Physical Geography, Freie Universität, Berlin, Malteserstr. 74-100, D-12249 Berlin, Germany. Email: [email protected]. Oumarou Faran Maiga Département de Géographie, Faculté des Lettres et des Sciences Humaines, Université Abdou Moumouni, BP 435, Niamey, République du Niger. Email: [email protected]. Mohamed Mounkaila Institute of Soil Science and land Evaluation (310), University of Hohenheim, D-70593 Stuttgart, Germany. Email: [email protected]. Issa Ousseïni Département de Géographie, Faculté des Lettres et des Sciences Humaines, Université Abdou Moumouni, BP 435, Niamey, République du Niger. Email: [email protected]. Simon Pomel DYMSET/CNRS, University of Bordeaux III, F-33405 Talence Cedex, France. Email: [email protected]. Ibrahim Mamane Sani Département de Géographie, Faculté des Lettres et des Sciences Humaines, Université Abdou Moumouni, BP 435, Niamey, République du Niger. Email: [email protected]. Erhard Schulz Department of Geography, Physical Geography, Julius-Maximilians University, Am Hubland, D-97074 Würzburg, Germany. Email: [email protected]. Brigitta Schütt Department of Earth Sciences, Physical Geography, Freie Universität, Berlin, Malteserstr. 74-100, D-12249 Berlin, Germany. Email: [email protected]. Barbara Sponholz Department of Geography, Physical Geography, Julius-Maximilians University, Am Hubland, D-97074 Würzburg, Germany. Email: [email protected]. Karl Heinz Striedter Frobenius Institute, Johann Wolfgang Goethe University, Grüneburgplatz 1, D-60323 Frankfurt, Germany. Email: [email protected]. Michel Tauveron Independent Archaeologist, F-19600 Lissac sur Couze, France. Wim Van Neer Royal Belgian Institute of Natural Sciences, Vautierstraat 29, B-1000 Brussels, Belgium and Katholieke Universiteit Leuven, Laboratory of Animal Biodiversity and Systematics, Ch. Deberiotstraat 32, B-3000 Leuven, Belgium. Email: willem.vanneer@ bio.kuleuven.be.

CHAPTER 1

Geomorphological and palaeoenvironmental research in the South-Central Sahara in review Roland Baumhauer Department of Geography, Physical Geography, Julius-Maximilians University, Würzburg, Germany Peter Felix-Henningsen Institute of Soil Science and Soil Conservation, Justus Liebig University, Giessen, Germany Brigitta Schütt Department of Earth Sciences, Physical Geography, Freie Universität, Berlin, Germany ABSTRACT: The main interest of the interdisciplinary “Limnosahara” research project of the German Research Foundation (DFG) composed of Physical Geographers, Palaeopedologists, Palaeobotanists and Archaeologists from the Universities of Würzburg, Berlin, Giessen, Frankfurt and Niamey, is to elucidate the Holocene palaeoenvironmental history of the Central Sahara. This introductory chapter gives an overview of the geomorphological and palaeoecological research in South-Central Sahara on the basis of a literature review. Subsequently different proxy-data sources are introduced and discussed to conclude on the former dynamics of the palaeoenvironment of the central part of Sahara.

1.1 THE SOUTH-CENTRAL SAHARA—GEOGRAPHICAL SETTING The Southern Central Sahara, between 17° and 23°N and 11° and 15°E, belongs to the northern part of the hydrological Chad Basin. It is a typical part of the plateau and plains landscapes of the Central Sahara, framed in the West and East by the Saharan mountain regions of Aïr and Tibesti. To the North the study region extends to the southern fringe of the Murzuq Basin; to the South it reaches as far as the Sahara-Sahel boundary, marked by the northern limit of immobile ancient sand dunes. The central and largest part of the study region is taken up by vast sand plains, the largest being the Ténéré, grading southwards into the sand seas Erg de Ténéré, Erg de Fachi-Bilma and the Grand Erg de Bilma, all characterized by closely spaced NE–SW oriented longitudinal dunes. The sand plains and ergs are interrupted by N–S-oriented scarplands, partly broken up into isolated massifs or plateau remnants, a common feature of their western forelands being elongated endorheïc depressions. They are like stepping stones between the most conspicuous landforms in the North of the region—the

2

Roland Baumhauer et al.

Figure 1. Location map of NE-Niger.

plateaus of Djado and Mangueni at the southern fringe of the Murzuq Basin—and the fossil dune region of Tin Toumma and In-Madama to the South, already belonging to the geomorphological Lake Chad Basin. The average elevation of the entire area is 400 to 500 m asl; the escarpments may rise above the forelands by heights ranging from a few tens of metres to several hundred metres (Figure 1). The meteorological and climatological data situation of the region is extremely poor in terms of space and time. There are only few weather stations, and their data are often not available for political reasons. The plains of the Central Sahara are particularly sparsely equipped with only ten weather stations, almost all of which are located along its northern limit. For the South-Central Sahara, the only more or less complete record is that from the station at Bilma (18°41' N, 12°55'E, 335 m asl) from 1922 onwards. For the military post of Madama to the North (21°51' N, 13°45' E, 546 m asl), rainfall data are only available from 1939 to 1943. Because of the lack of proxy data, meteorological publications on the present climate and climatic history of the region are largely model-based and often rely too much on the transfer of data from deep-sea cores to be a suitable basis for further studies in a

Geomorphological and palaeoenvironmental research in the South-Central Sahara in review

3

continental region such as this (e.g. Flohn and Nicholson, 1980; Kutzbach, 1980; Adams and Tetzlaff, 1984; Macayeal, 1993; Cubasch et al., 1997; Montoya et al., 1998; Stauffer et al., 1998; Crowly, 1999; von Storch et al., 2000). Ultimately, the factors governing the climate of the Central Sahara are its latitudinal position within the trade-wind belt and its extreme continentality. Although the region is affected by tropical monsoonal air from the Equatorial Atlantic and, in the North, by the Mediterranean circulation, neither influence modifies the extreme aridity of the region. At a low cloud cover throughout the year and summer maximum temperatures of above 40 °C, average annual precipitation is barely above 20 mm. 1.2 GEOLOGICAL SITUATION The post-Palaeozoic development of Eastern Niger to the south rim of the Plateau du Djado was described by Faure (1966); together with the work by Pirard (1964) on the hydrogeology of Eastern Niger, this study is still the foundation of all geological work in the region. The Southern Central Sahara consists of a wide-ranging system of tectonic basins and broad uplifts. Mainly in the transition areas, gently inclined Palaeozoic, Mesozoic and Tertiary sediments, unconformably overlying an etchplain cutting across the Precambrian metamorphic and crystalline basement, have been eroded to form the typical plateau and scarpland terrains of the Central Sahara. The core of the South-Central Sahara belongs to the geological basin of Bilma; to the North, the plateaus at the southern Murzuq rim are part of the geological Murzuk Basin, while to the South, the Massif d’Agadem is part of the geological Chad Basin (cf. Faure, 1966). The geological basin of Bilma occupies an area of 400 × 300 km in the centre of the Nigrian part of the hydrological Chad Basin. Its longer axis is delineated by the Bilma Escarpment. The western half of the basin, delimited by outcropping basement to the North, South and West, has been filled by up to 1.000 m of Cretaceous to Palaeocene continental and marine sediments, without any underlying Palaeozoic strata (Faure, 1966). Hardly any geological and geomorphological information has been published about the eastern parts of the Bilma Basin. To the West the Bilma Basin is separated from the Aïr Mountains by the grabens of Achegour and Adrar Madet, to the Southwest, by the Téfidet-Lake Chad graben system from the Termit Basin. To the North, there is just a gradual transition to the Murzuq Basin. Within the system of Central-Saharan basins and uplifts the geological Murzuq Basin, approximately 1.000 × 600 km large, has a central position, geologically and morphologically being a depression with elevated rims. The gently inclined strata form an almost uninterrupted fringe of outward-facing escarpments around the basin (Grunert, 1983). Because of the basin structure, the ages of the sediments exposed at the surface decrease from the rim to the centre (Klitzsch, 1970, 1971). Along the southern fringe, the surficial rocks are mainly marine sediments of the Upper Carboniferous and the continental series of the Mesozoic, up to the Lower Carboniferous (Plauchut et al., 1960). As mentioned, the southernmost parts of the Central Sahara belong to the geological Chad Basin. 1.3 GEOMORPHOLOGICAL AND QUATERNARY EXPLORATION More recent geological studies of the South-Central Sahara have mainly been undertaken by the “Würzburg Africa Research Group” headed by Horst Hagedorn. From 1977 to 1991, ten expeditions to the region were carried out, with varying participants. From 1992 to 2000 all of the Nigrian part of the South-Central Sahara was inaccessible because of

4

Roland Baumhauer et al.

Figure 2. Emi Fezzan, Butte of the northern Plateau de Tchigai.

civil-war-like conflicts. Except for a few selective studies on the prehistory of the region, carried out on the Djado Plateau from a base in Algeria (Striedter et al., 1995; Striedter, 1997, 1999), a whole decade passed without field work in this part of the Sahara. Only a small number of studies exist on the Late Quaternary evolution of the SouthCentral Sahara. Those by Faure (1966), Servant (1973), Servant-Vildary (1978) and Maley (1980) have their regional focus in the Chad Basin; the Saharan space is dealt with only marginally and very selectively. The north easternmost parts of Niger, in particular from east of the plateaus of Djado and Mangueni to the Chadian and Libyan borders, is still largely unknown from this perspective, except for some reports by French colonial officers (e.g. Capit. Freydenberg, 1907), and some geological reconnaissance work (Dalloni, 1948; Kilian, 1937; Kilian and Furon, 1934; Pirard, 1964). 1.3.1 Tertiary to Mid-Pleistocene landform history According to the stratigraphy described by Faure (1966) and Klitzsch (1970, 1971), postsedimentary landform development in the Central Sahara began after the deposition and diagenesis of the Messak/Nubian Sandstone of the rim of the Murzuq Basin and of the Bilma and Emi Bao formation of the Bilma Basin. The shaping of the present landscape should already have started in pre-Upper Cretaceous times. From the study of the major landforms of North-Eastern Fezzan and groundwater studies of the Central Sahara, Klitzsch (1974) and Klitzsch et al. (1976) place the Messak Sandstone in the Jurassic and also assume that an escarpment landscape had already developed in pre-Late Upper Cretaceous times. In contrast, Faure (1966) for the Bilma Basin and Busche (1982) for the western and southern rim of the Murzuq Basin and the northern parts of the Bilma Basin assume an Upper Cretaceous age of the Bilma and Emi Bao formation and the Messak Sandstone. Up to the onset of the Cainozoic, the study region was part of a depositional surface rimmed by the exposed crystalline uplifts of Gargaf-Hoggar to the North and Northwest, and the Tibesti-Syrte uplift to the East and Northeast (Busche, 1982). To the South, during the Palaeocene and Eocene, the sea reached Western Niger, Mali and North-Western Nigeria (Greigert and Pougnet, 1967). The marine transgression from the South did not go beyond the present foreland of the Tibesti Mountains (Klitzsch, 1970). For the continental realm between the seas, Erhard (1956) and Elouard (1959) identified a time of lateritic deep weathering under a humid-tropical climate, based on their study of the Continental Terminal, the southern correlative sediment of extended etchplanation to the North. According to Faure (1966), tectonic uplift in Eastern Niger since the Upper Eocene and the Oligocene caused the large-scale removal of the lateritic weathering mantle there. For the southern and western rim of the Murzuq Basin, Busche (1982), in his comprehensive study on the development of the region, together with the Djado Plateau and northern “Kaouar”

Geomorphological and palaeoenvironmental research in the South-Central Sahara in review

5

(summarized by Baumhauer, 1986; Baumhauer et al., 1989; Skworonek, 1988; Busche, 1998) proposed a downwearing by etchplanation of the Messak Sandstone by at least 100 m. To him, the humid to wet-and-dry tropical conditions favouring etchplanation—of the African Surface of King 1967, cf. also Gellert (1971)—persisted to the Miocene, as Eocene sediments have also been etchplanated and intensive chemical weathering can be traced to at least that time, with decreasing intensity even to the end of the Tertiary (Busche, 1982; 1998). In a persistingly humid environment intensive silicification, following the almost complete removal of the lateritic weathering mantle, affected the level landscape close to a high groundwater level. Busche (1982, 1998) regards such silicified surfaces as a special type of planation landscape preserved on the plateaus of the Central Sahara, and also as a time marker: the end of silcrete formation, i.e. its beginning dissection, coincided with a time of increasing crustal movements, expressed in the doming and onset of volcanism first in what was to become the Hoggar, and then also in the Tibesti Mountain regions. Following silcrete formation there began a time of intensive surficial and subsurface silicate karstification, it appears to have continued to the Early Pleistocene (cf., among others, Busche, 1982; Busche, 1998; Busche and Sponholz, 1992; Hagedorn and Sponholz, 1990; Sponholz, 1992). The tectonic movements increasing since the Miocene uplifted most of the Cretaceous continental sediments and also, South of 17°N, the parts of Eastern Niger blanketed by the Continental Terminal, whereas the Chad Basin to the East began to subside. This led to a general gentle tilting of the strata to the East, together with continued downwearing by planation of parts of the rising terrains by an average of 100 m, the grand creusement postérieur au Continental terminal, followed, during the Early Pleistocene, by another phase of erosion during which also the Pliocene fluvial sediments (quartz gravel beds incorporating fragments of Continental terminal) were dissected in the North, and erosion to the South increased from 50 to 150 m (Faure, 1966). At the same time, in the region of subsidence west and northwest of present Lake Chad, up to 1.000 m of sediment were deposited of the Groupe du Tchad (Pirard, 1967), and, in the centre of the geological Chad Basin, about 100 m of the Bahr el Gazal series (Servant, 1983). According to Busche (1982, 1998) uplift still under the conditions of tropical chemical weathering, and a continuation of silicate karstification. The combined processes, up to the end of the Tertiary, resulted in the formation of a landscape of erosion scarps rimming dissected plateaus and a lower level of intra-plateau basins and etchplains. Regional differences of uplift and thus of different efficiency of etchplanation caused scarps to become a few hundred metres high, like the Messak Mellet now forming the western geomorphological rim of the Murzuq Basin, only a few tens of metres, like the southernmost parts of the Dissilak escarpment forming the western rim of the Djado Plateau, or, intermediate, the Bilma Escarpment. The latter is further characterized by the fact that selective planation, following intensive sandstone karstification, could keep pace with the rate of uplift, so that much of the original surface was transformed into a pattern of plateaus and basins (Busche, 1982; Busche, 1998). The fluvial and aeolian transformation of the landforms took place under semi-humid to semi-arid climates of the Early Pleistocene: also the lowermost etchplain level became fluvially dissected, those parts of the Late Tertiary etchplain abutting against the escarpments and rims of intra-plateau basins became more steeply inclined pediments, escarpment and inselberg profiles became steepened, with the development of a free face, and the endorheïc scarpfoot depressions (see above) came into existence (Busche, 1982, 1998) towards the end of the Early Pleistocene or somewhat later, during an extremely pronounced arid phase (Busche and Stengel, 1993; Busche, 1998). As described by Grunert (1983), in a detailed study focusing on the ancient landslides formed along the heterolithic parts of escarpment slopes of the western and southern rim of the Murzuq Basin, the slides followed a period of fluvial dissection and slope

6

Roland Baumhauer et al.

oversteepening under semi-arid conditions in the Plio-/Pleistocene transition period, possibly triggered by a renewed rise in precipitation. In comparison to present-day landsliding in humid regions, Grunert (1983) assumes at least 400 mm of winter rains, at lower temperatures and thus reduced evapotranspiration compared to today, for their formation. A common phenomenon throughout the Sahara, fossil landslides also exist along the escarpment rims of the Bilma Basin, but have not yet been studied there. For the landform history outlined above, mainly derived from geomorphological and palaeogeographical observations, it is assumed that climates were perennially humid to seasonally humid tropical, with deep-reaching chemical weathering especially during the earlier parts of the post-sedimentary development of the region. This is largely contradicted by Skowronek (1988) in a broad study of soils as indicators of climatically governed land formation of the Central Sahara. From palaeopedological and sedimentological studies he concludes that no humid-tropical soil was formed since the Early Tertiary, and that, in consequence, there was no climate since then which could have led to deep-reaching and intensive chemical weathering and soil formation. For the rim of the Murzuq Basin he assumes times of a soil-forming environment much more humid than today, but hardly beyond 200 mm of episodic rainfalls. Similarly, Barth and Blume (1975), from their study of the escarpments of the western rim of the Murzuq Basin, concluded that only marginally more humid conditions than today had existed throughout the Pleistocene, envisaging an arid-morphodynamic erosional history. 1.3.2 Late Pleistocene and Holocene development The first radiocarbon dates for the upper Late Pleistocene and Holocene of the plateau landscapes west and southwest of the Murzuq Basin were published by Busche et al. (1979), Busche (1982) and Grunert (1983). Largely on the basis of its morphostratigraphic position, they equalled a rhizome-rich terrace body grading into a diatomitic lake sediment in the Enneri Achelouma area, below the south rim of the Plateau du Mangueni, with the Middle Terrace of the Tibesti Mountains (mainly Jäkel 1979), although the dates from the plateau itself suggest a later onset of sedimentation. Dates from lake sediments/slack water sediments deposited in another tributary valley of Enneri Achelouma, grading into a reddish-brown sandy terrace at its mouth, indicated uninterrupted sedimentation from 8 to 7 ka BP. Another lake sediment from the foreland of Col d’Anai, about 250 km NNW of Achelouma, at the base of a reddish-brown, sandy terrace, was dated to 15,8 ka BP. The similarities among these terraces, and also with other sedimentary bodies of the region, led the authors to conclude that, despite the age differences, they are all part of the same Middle Terrace. Following the palaeoclimatic interpretation by Jäkel (1979) from the Tibesti (Enneri Bardagué), they assumed that the region around the Plateau du Mangueni, during the last major humid phase, between 16 and 7,4 ka BP, had come under the influence of a winter west-wind regime (from 16 to 8 ka BP according to Jäkel), and had thus experienced wet conditions earlier than the region to the South, where inroads of rain-bringing tropical air were thought to have started not before 13 ka. In conclusion, the northern and southern parts of the region had not always been subjected to the same climate (Busche et al., 1979). For Busche (1982, 1998) the deposition of the mostly fine-grained Middle Terrace is the major geomorphological event of the Late Pleistocene and Holocene. The several millennia of sedimentation are thought to have followed a long phase of pluvial-time soil formation during the Upper Quaternary, with largely stable conditions on the plateaus and slopes. With decreasing humidity the soil cover supplied the fines for the deposition on the valley floors of the region from about 16 ka to 7 ka BP. Schulz (1980), in studying the

Geomorphological and palaeoenvironmental research in the South-Central Sahara in review

7

Holocene vegetation development of the region, concluded that during the time of Middle Terrace deposition it resembled the present vegetation, but covered much larger areas with a higher ground cover and a greater number of species. He assumes a gradual thinning during the Holocene, interrupted by climatic oscillations, but the pollen spectra studied do not allow a detailed temporal development to be read from the sediments. First results of prehistoric studies by Striedter (1992) and Ferhat and Striedter (1993) confirm the presence of Palaeolithic and Mesolithic traces, like those identified by Baumhauer et al. (1991) and Tillet (1983, 1989) for the Southern Djado Plateau, the Ténéré and its frame of plateaus (see below). The Neolithic finds from the central Djado region allow the identification of a culture group mainly defined by its pottery (Striedter, 1992, 1995), possibly closely related to the Ténéréen (cf. Roset, 1982, 1987). If confirmed, this would not only considerably enlarge the region inhabited by the Ténéréen culture group, but would also mean that the Djado region was inhabited by herders, as long as the climatic conditions permitted. Of particular interest are the petroglyphs of the central Djado Plateau. The engravings, in particular, compare well with those of the Ahaggar, Tassili n’Ajjer, Acacus and Fezzan (among others, Aumassip, 1993); they are characterized by a sequence of epochs and schools whose overlaps should permit the setup of a relative chronology (Striedter, 1992; Ferhat and Striedter, 1995; Striedter, 1997). Lake sediments Palaeolake sediments are a common occurrence in the South-Central Sahara of NorthEastern Niger. Their location and state of preservation are determined by the geomorphic structure of the region. Area-wise and in the shape of yardangs up to 10 m high, they are found in the dune depressions of the ergs framing the Ténéré to the South and East (cf. figure 3), as well as in the endorheïc depressions West of isolated plateaus or the plateau remnants of the eroded scarplands of the South-Central Sahara (Busche 1998). In the depressions at the foot of Emi Bao (Seggedim oasis) and along the Southern Bilma escarpment (Kawar), they had in places developed to a maximum depth of about 40 m. Within the sand plain of the Ténéré there are only thin deposits in shallow basins of various sizes. The few lake deposits studied so far occur in extremely favourable locations within the plateau landscapes of the western and southern fringe of the Murzuq Basin (Busche et al., 1979; Grunert, 1983; Striedter et al., 1995) and in the Southern Plateau du Tchigai (Busche, 1998). From the next intended study region East of the plateaus of Mangueni and Djado, there are first observations on widespread plains with a lacustrine cover and reports on the terminal basins of rivers as well as bedrock basins or playas filled with clays and diatomites (mainly Pirard, 1964; Capitaine Renaud, 1926; Striedter and Ferhat, personal communication). Publications on the stratigraphy and its palaeoclimatic and -environmental interpretation of the lacustrine deposits of the outgoing Pleistocene and Holocene of North-Eastern Niger were first presented by Servant (1973), who—on the basis of Faure (1965, 1966, 1969) and Faure et al. (1963)—tied his studies in with Cainozoic continental sedimentation and climate history. Servant-Vildary (1978) outlined the history of the lakes by means of their diatom flores, and Maley (1980) presented the climatic history of the region during the last 30.000 years based on palynological data. All these studies focus on the Chad Basin; of North-Eastern Niger, only the southern fringe—Southern Ténéré, Erg de Fachi-Bilma, Grand Erg de Bilma, or the region around Bilma—is dealt with selectively. Several field campaigns between 1981 and 1991 addressed the potential and problems of palaeoenvironmental reconstruction from palaeolakes of the Southern Central Sahara (Baumhauer, 1986, 1987, 1988, 1989, 1990; among others).

8

Roland Baumhauer et al.

Figure 3. Kafra scarpfoot depression in Central Ténéré with yardangs formed in diatomite.

Swamp-iron crusts In the former shallow water margins of Pleistocene and Holocene palaeolakes of the Ténéré, the Tchigai Plateau and the Erg de Bilma, gleyic palaeodune sands are generally overlain at the surface by low ridges or a horizontal mesh of swamp-iron rhizoconcretions. There are only few references to them from the Sahara of Eastern Niger (Maley, 1980; Baumhauer, 1993). Similar deposits from palaeolake margins of the Eastern Sahara were described by Kröpelin (1993). The origin and nature of these stratigraphically and palaeoclimatically significant crusts were studied by Felix-Henningsen (1997, 1998): the former roots and stems of swamp vegetation had been transformed into massive goethite casts, which had impregnated the palaeodune sand around them, often with an outer zone of lepidocrocite. Together with the iron oxide, there were concentrations of P, Ca, Mg and heavy metals. Their formation and morphological differentiation were linked to the gently inclined shorelines and shallows of the palaeolakes within the ancient dunes. Formerly increasing water depth shows in the transition from a massive layer of rhizoconcretions to indurated plant stems sticking out of the sand following deflation. Because of the gently rolling to often perfectly horizontal surface of the Ténéré, swamp iron concretions are generally frequent and may cover large continuous areas. Within the more accentuated relief of the Erg de Bilma there is a pattern of narrow shoreline fringes on the flanks of sand-dune depressions. The concretions originated from the oxidation and precipitation of reduced iron ions of the lake water and their concentration around the roots of reeds and macrophytes. High redox potentials can be excluded for the altogether shallow lake environments because of the low oxygen content of the warm water in the first place and further microbiotic

Geomorphological and palaeoenvironmental research in the South-Central Sahara in review

9

consumption for the decomposition of plant remains. Thus the formation of rhizoconcretions can only be due to the specific physiology of the swamp vegetation, namely the existence of an aerenchyme supplying the roots with oxygen from the air. The release of excess oxygen from the roots caused a high redox potential locally, leading to oxide precipitation around each of the roots. The movement of ions from the reduced sediments of the lake floor to the root space was mainly due to diffusion. Palaeoclimatically the existence of swamp ore deposits stands for prolonged periods of a constant lake level. Their presence at various heights above the lake bottoms, most evident on the flanks of the dune depressions is witness to the trend of shrinking palaeolakes with increasing aridity, interrupted by phases of lake-level stability (cf. Felix-Henningsen, 2000). Ancient sand dunes, swamp iron deposits, palaeolake deposits, as well as the palaeosoils dealt with in the following chapter, are sensitive indicators for changing environmental conditions of widespread occurrence. Their dating is expected to contribute significantly to our knowledge of the regional differentiation of the Central and South-Saharan past environments. Palaeosoils On the ancient dunes of Niger’s Central and Northern Sahel, an almost continuous cover of palaeosoils has been preserved, the youngest of which developed during the Early Holocene humid phase. For the Nigrian region, studies have been published on their nature and palaeoclimatic interpretation by Grunert (1988), Völkel (1988, 1989), Pfeiffer and Grunert (1989), Pfeiffer (1991), and Völkel and Grunert (1990). Studies from other parts of the Sahel have been presented by, among others, Felix-Henningsen (1983, 1984, 2000), who studied palaeosoils from the Southern Central Sahara to the Northern Sahel during an expedition in 1991. The quite thick palaeosoils remained generally well preserved after their formation in the Holocene; only parts of their upper horizons underwent either erosion or burial by colluvium and windblown sand. In the South and Central Sahara, too, palaeosoils and relict soils are still widespread on Quaternary fluvial sediments and on ancient Pleistocene and Holocene dunes. In deflation areas they may also be exposed at the surface and are overlain by windblown sand of varying thickness or by dunes. Völkel (1988, 1989) and Felix-Henningsen (1992, 2000) report that the palaeosoils have been preserved to different degrees in the various landscapes of the region, depending on their degree of development during the pluvials and geomorphic activity in the subsequent phases of aridity. In the Ténéré the original ancient-dune terrain was fairly levelled, and pluvial-time palaeosoils have been preserved over wider areas than in other regions of the South and Central Sahara. Neolithic tumuli on level ground indicate that the soil surface around them has been lowered by a few decimetres at most since their construction. In the more accentuated palaeodune landscape of the Erg de Bilma, deflation and surface wash have removed much of the soils from dune tops and steeper slopes. Preservation was better only in the lower parts of depressions. On the windward and leeward dunes and fluvio-aeolian sand ramps of the scarplands to the North and the Zouar uplands, the environment was more favourable to soil formation during the pluvials than in the adjacent plains. Less evapotranspiration in wind-protected positions allowed the earlier onset of soil formation at the beginning of a pluvial phase, and a prolongation at its end. Additional runoff from the rocky slopes above was probably also conducive to weathering and thus soil formation. Differences in thickness, colour and stability of soil profiles of palaeosoils of most likely the same age may thus be explained by their position either on the escarpment slopes or far away from them. Their protected leeward position was also conducive to their preservation during the subsequent

10

Roland Baumhauer et al.

arid phases. As yet no palaeopedological information is available for the region east of the plateaus of Djado and Mangueni. Studies to date suggest that palaeosoils exist from at least three pluvials in the study area: a Late Pleistocene pluvial (c. 40–20 ka BP), the Ghazalien (Servant, 1983), an Early Holocene pluvial (c. 14–7,5 ka BP), the Tchadien (Servant, 1983), and a Mid-Holocene pluvial (6,5–4 ka in the Western Sahara), the Nouakchottien (Michel, 1973). Each of the pluvials was preceded by an arid phase of dune-sand deposition and dune formation. During the oldest arid phase, at 40 ka BP, palaeodunes were deposited in the Erg Ancien, extending far into the Southern Sahel (Gavaud, 1965). The pedostratigraphic interpretation of palaeosoil relics spread over large areas is based on grouping those with a similar inventory of characteristics and a comparable stage of development. For those deviating in both respects, formation during another pluvial phase is inferred. As there have been repeated shifts of the monsoonal climatic regime to the North into the Sahara (cf. Warren, 1970), zonal differences in weathering intensity and profile development have to be expected for the soils of each pluvial. This should be taken into account when comparing palaeosoils from different meridional regions. Some statistical evidence, though based on only a small number of samples, was presented by Felix-Henningsen (2000). 1.4 DEVELOPMENT OF THE PALAEO-ENVIRONMENTS The palaeoenvironmental evidence obtained by Baumhauer from the Ténéré, the escarpment forelands and the ergs to the South is in good agreement on a millennium scale with the chronology of the Late Pleistocene and Holocene climate and landscape changes obtained for the Southern Sahara (cf. Baumhauer, 1984–2002; Gasse, 1988; Maley, 1980; Servant, 1973; Servant-Vildary, 1978). With regard to lake formation, however, there appear to have been sufficiently humid conditions from the end of the Pleistocene to the Mid-Holocene only, with a decreasing tendency to 6,5 to 5,5 ka BP and interrupted by several short arid phases, with the onset of fully arid conditions since 5 ka BP at the latest. Finely bedded lake sediments from the depression of Seggedim, at the foot of the northernmost part of the Bilma Escarpment (and thus the northernmost deposit studied in the South-Central Sahara so far), have yielded evidence that the decisive change in climate and landscape development to the present hyperarid conditions took place even earlier in the Holocene and ended by about 6,5 ka BP. By that time, the transition from the Early Holocene fresh-water environment to sebkha conditions had become definitive (Baumhauer et al., 2004). The very close contact between Sudanese and Saharan vegetation evident for the Early Holocene became non-existent by 6,5 ka BP. Palynological evidence since 7 ka BP indicates a contracted permanent vegetation with a species composition similar to the present one (Baumhauer et al., 2004). This suggests a significantly weaker climatic gradient up to the Mid-Holocene than today, due to the interference of the monsoonal and Atlantic (Mediterranean) cyclonic circulation elements, which gradually became weaker during the Mid-Holocene (Baumhauer et al., 2004). Support comes from the palaeopedological studies conducted by Felix-Henningsen (1992–2000), as well as from first, still highly selective palaeoecological studies in the Djado Plateau region referred to above (Striedter et al., 1995). Felix-Henningsen (1992, 2000) conducted detailed pedochemical, micromorphological and clay mineralogical studies on the origin of palaeosoils. He concluded that the palaeosoils formed on the ancient dunes of the region developed during semi-humid conditions. His macro- and micromorphological, soil-physical, geochemical and claymineralogical analyses allow a sequence of processes to be reconstructed. The formation

Geomorphological and palaeoenvironmental research in the South-Central Sahara in review

11

of browned to partly rubefied horizons is characteristic of Pleistocene and Holocene chromi-cambic arenosols; they originated from silicate weathering as part of a pedogenetic sequence, preceded by desalinization, decarbonatization and acidification due to leaching by percolating soil water in a more humid climate. Newly formed clay minerals—though quite immobile in the soil-water solution—were very uniformly distributed within the Bv horizons, suggesting an originally quite uniform distribution of the weatherable minerals, best explained by the deposition of dust simultaneously with the sands. Electrostatic piggybacking of dust particles is a common phenomenon in aeolian sand transport (cf. Stengel, 1992). Dust deposited during the pluvial phases was probably also spread throughout the soil by percolation and bioturbation. For a review of the significance of dust sedimentation in the Sahel and Sahara cf. Völkel (1989, 1991), Stahr et al. (1994), Hermann et al. (1994), and Hermann (1996). The decreasing haematite content in the palaeosoils of the Early Holocene from South to North and West to East (Felix-Henningsen, 1992, 2000) suggests decreasing humidity to the North and East during the Early Holocene pluvial. Similarly the decreasing thickness and weathering intensity to the North suggests a reduction of time available for the soilforming processes by the later onset and earlier termination of the Early Holocene pluvial phase in relation to the shifting monsoon regime. Rubification, clay-mineral formation and aggregate stability show the same regional trend. A very small-scale, slope-related variation of the Fe-oxide minerals exists in the palaeosoils in transition to the Northern Sahel. In the formerly quite humid lower slopes and valley floors between the ancient dunes, only goethite was formed in the developing soil. The input of organic matter was possibly higher in those positions due to their higher vegetation density, which blocked the aging of the original ferrihydrite to haematite and was conducive to the neoformation of goethite. Similar slope conditions in the more central parts of the Southern Sahara could also have caused the simultaneous formation of goethitic and rubefied palaeosoils next to each other under pluvial conditions. These finds show the need for studying the small-scale or catena variation of soil properties in palaeopedological studies of the region, to make sure that specific local conditions are not mistaken as evidence for sweeping pedogeographical and palaeoclimatological interpretations. Palaeoclimatic evidence from numerous studies in the Western Sahara point to another humid phase between 4,5 and 2,5 ka BP, followed by the onset of the hyperarid conditions still prevailing today (among others: Michel, 1973; Petit-Maire, 1989, 1991, 1994; Petit-Maire and Kröpelin, 1991; Petit-Maire and Riser, 1983; Reichelt et al., 1992; Völkel, 1988; Mauz and Felix-Henningsen, 2005; Felix-Henningsen and Mauz, 2005). In contrast to evidence from the Western Sahara, studies by the Berlin Collaborative Research Centre SFB 69 in the Eastern Sahara identified only an Early- to Mid-Holocene pluvial, beginning around 9,5 ka BP and persisting, with decreasing humidity and interrupted by several arid phases, to around 4 ka BP—thus longer than in the South-Central Sahara— before the abrupt change to hyperaridity took place (among others: Guo et al., 2000; Pachur et al., 1987, 1990, 1991, 1996; Kröpelin and Soulie-Märsche, 1991; Pachur and Hoelzmann, 1991; Hoelzmann, 1992; Kroepelin, 1993, 1999; Pachur, 1997, 1999; Holzmann et al., 2000; Pachur and Altmann, 2006). Workers from all parts of the Sahara agree that, beyond the general changes of climate and the environment during the Holocene, there also existed a palaeoclimatic zonality with latitudinal shifts of the environmental belts, for the Southern Sahara Servant (1983), Servant-Vildary (1978) and Maley (1980), for the Western and West-Central Sahara the working group around Petit-Maire (among others: Petit-Maire, 1987, 1991, 1994; PetitMaire and Kröpelin, 1991; Petit-Maire and Riser, 1983; Schulz, 1991), and for the Eastern Sahara the group around Pachur (among others: Guo et al., 2000; Pachur and Hoelzmann, 1991; Hoelzmann, 1992; Kröpelin, 1993, 1999; Pachur, 1997, 1999; Hoelzmann et al.,

12

Roland Baumhauer et al.

2000; Pachur and Altmann, 2006), as well as Haynes et al. (1989) and Ritchie and Haynes (1983). From their palaeoecological studies Hoelzmann (1992), Kröpelin (1999), Neumann (1989), Pachur (1999) or Kuper and Kröpelin (2006) have inferred that ecosystem structures in the Eastern Sahara during the Early Holocene pluvial optimum resembled those of the savannas around 13°N of today, related to a northward shift of the Sahel boundary in East Africa by about 600 km. This is in good agreement with the known northward occurrence of the EarlyHolocene pluvial soils in Eastern Niger (Felix-Henningsen 1992 to 2000). A direct comparison of the soils is problematic, though, as only a few studies on the palaeosoils on ancient dunes are available for the Eastern Sahara, though with supporting evidence from environmental history studies in the Goz region of the Sudanese Sahel (Warren, 1970; Felix-Henningsen, 1983, 1984; Gläser, 1987). Palaeosoils from several pluvials have been identified there on ancient dune fields of different ages, characteristic enough to serve as stratigraphic markers. Another piece of evidence comes from the palynological interpretation of the sediment core of Seggedim (Baumhauer et al., see chapter 3): up to at least 20°N there existed a Mid-Holocene (Sahelian) savanna vegetation, with an influx of Saharan flora elements as early as 7 ka BP. This early apparent aridification is also attributed to a persistent human impact along the boundary between the Sudanian and Sahelian vegetation since MidHolocene times. In conjunction with progressive climate deterioration, this interference led to the typical Sahelian savanna vegetation, possibly sooner than would have happened under purely natural conditions (Baumhauer et al., 2004; Schulz, 1991; Wasylikowa, 1992). The studies conducted by the Collaborative Research Centre SFB 268 “Cultural development and language history of the West African savanna” are only marginally concerned with the Late Quaternary landscape development and present morphodynamics of the South-Central Saharan region. The high-resolution pollen diagrams from the Manga grassland (Salzmann, 2000; Salzmann and Waller, 1998) yield no evidence of humaninduced development of the Sahel, thus contradicting the findings from the Central Sahara. The authors admit, however, that the problem of finding indicators of human impact has not yet been solved (Salzmann, 2000). At the start of the SFB 268 studies, a very early human influence was assumed, mainly owing to the high pollen percentages of a secondary shrub and pollen of certain plants that exist as weeds today (Ballouche and Neumann, 1995). More recent evidence suggests that human impact dates back only to the third millennium B.C. and originally consisted in a combination of nomadic animal husbandry, food gathering and some local cultivation without permanent settlements (Neumann et al., 2001). The transition to a sedentary economy did not occur until around 3 ka BP. For the Southern Chad Basin the model presented by Servant (1973) was largely accepted, assuming an arid phase from 20 to 12 ka BP, a rapid lake-level rise around 9 ka BP, a maximum level around 6,5 ka BP, a short arid phase around 3,5 ka BP, and ever since a general shrinking of the lake. Pedological findings from Bama Ridge—older dune sands with chromic arenosols and younger ones with cambic arenosols (Thiemeyer, 2002)—suggest more than one maximum lake level, the youngest one occurring around 6,5 ka BP. Climate oscillations at the beginning of the Holocene, around 11,4 to 10,3 ka BP, have been derived from the analysis of clay deposits in the dune depressions. The situation in the Southern Chad Basin resembled that of Burkina Faso today: permanent settlement began around 3 ka, the people living on a combination of hunting, gathering and fishing, animal husbandry and the cultivation of domesticated Pennisetum millet (Gronenborn, 1998). Earlier permanent settlement had been prevented by frequent floods. The first settlers are assumed to have come from the North and Northwest. No information on an early human impact in the region has been found so far. However, if millet cultivation was introduced to the Sahel from the Sahara, as assumed by Neumann

Geomorphological and palaeoenvironmental research in the South-Central Sahara in review

13

et al. (2001), the question arises whether the environment there during the third millennium B.C. was still Sahelian or already Saharan. To explore environmental changes in the northern part of the Chad Basin, P. FelixHenningsen and R. Baumhauer led a geo- and bioscientific expedition to the Central Sahara in March and April 1997, as part of the DFG-funded project “Late Quaternary landscape development and present sediment dynamics between Lake Chad and Tibesti.” During this expedition, the palaeopedology group from the University of Giessen studied the palaeosoils on ancient dunes and lake sediments along a transect from East of the lake to the oasis of Faya, just South of the Tibesti Mountains. An additional research focus was the alternation of palaeodune and -lake sediments in former terminal lakes and in the Bodélé Depression West of Faya. Field work and TL/OSL dating revealed that the sand dunes between the Northern Sahel and Faya are of Early to Mid-Holocene age, with progressively younger ages to the South, consistent with progressive aridity since the pluvial maximum. Stratigraphy and datings agreed with previous palaeolimnological results on the former extent and lake levels of Lake Chad. However, the expedition’s results showed that earlier studies had erroneously attributed the ancient dunes of the region to the Pleistocene. Soil formation mainly took place during the subsequent Mid-Holocene pluvial starting around 3,1 ka. It led to moderately developed cambic arenosols, overlain by humic colluvia on lower dune slopes. Soils that had developed in a terrestrial environment on low dune shields during the early pluvial became gleyic with the rising groundwater table and were eventually buried beneath lake sediments when flooded by the Palaeochad. Compared to the deeply weathered and rubefied chromi-cambic arenosols of the Pleistocene sand dunes of Eastern Niger, those in the Northern Chad region show a clearly lesser degree of soil development. This is because in Niger soil formation had already started with the onset of the Early Holocene pluvial. Additionally, due to the West-East gradient of continentality and thus humidity of the present climate, pedogenic processes are still at work in the soils of the Nigrian palaeodunes, whereas soil formation in the Northern Sahel zone of the Chad region to the East came to a complete standstill with the onset of Saharan hyperaridity. Deposition and illuviation of carbonates, salts and silica and their aeolian sand blanket show that the A-horizons and B-horizons are fossil. As well as the dune soils West of Faya and of the Bodélé buried by lake sediments, gleyic to wet gleyic soils with humic A-horizons were found intercalated with the lake sediments, also subaquatic soils of former shallow-water areas with carbonatic and oxidic rhizoconcretions. For lack of funds they have not yet been dated and thus cannot be placed in a stratigraphy. Even so, repeated alternations of dune sands with lake sediments and the presence of gleys and rhizoconcretions within the lake sediments point to abrupt oscillations between humid and arid phases. Whereas the presence of palaeosoils on the fossil dunes is evidence of increasing humidity, palaeosoils intercalated with the lake sediments reflect somewhat lower lake levels and thus decreasing humidity. Towards the end of the last lake phase of the region, gleys with all their typical characteristics developed on the largely diatomitic sediments. 1.5 PRESENT MORPHODYNAMICS The present morphodynamics of the Southern Central Sahara has been underrepresented in research studies, as the emphasis of most publications has been on the landform, environmental and, derived from it, the palaeoclimatic history of the region, except for a number of publications on the Tibesti Mountains by workers at the Bardai Research Station of Freie Universität Berlin in the late 1960s (e.g. Grunert, 1970; Jäkel, 1971, 1979; Molle, 1969, 1971; Pachur, 1966), or by Mainguet (e.g. 1968) for neighbouring areas.

14

Roland Baumhauer et al.

The publications by Busche et al. (1979), Busche (1982), to a larger extent Busche (1998) and Grunert (1983) concern the western and southern fringe of the Murzuq Basin. Insights into most recent, largely aeolian morphodynamics may be gained from a number of wide-range palaeoclimatological studies, for example by Mainguet (1995), Servant (1973), Sirocko et al. (1993), or the numerous studies of the PAGES—PEPIII: Past Climate Variability Through Europe and Africa project (for project-summary see Batterbee, et al., 2004); however, their respective models and superordinate conclusions have still to be verified. No specific studies have been conducted modern fluvial processes in the region, for the simple reason that during the rather short stays, at least of the Würzburg team, no rainfall events took place in the hyperarid environment. Observations have been made, though, on aeolian activity. On the corrosion side, there is much evidence of present processes, but most of the aeolian corrosion landforms date from older times of much higher wind speeds, or a higher availability of grinding sands, when, channelled around the Tibesti Mountains, complete sections of the scarplands of the region were worn down and their remains became streamlined. During most likely several such phases, one of them also with extremely corrosive winds from the South (-West), the region affected was extended much more to the North and South than the region presently subjected to the low-energy corrosion. In most parts of the region, the fossil nature of the corrosion landforms is testified to by their coating of decaying desert varnish. In soft saprolitic rocks and in wind-exposed positions. In the Ténéré region and on the plateaus of the Western Murzuk Basin, corrosion may nevertheless have worn down surfaces by a few centimetres to decimetres since the Neolithic (Hagedorn, 1979). Studies by Busche and Stengel (1993) have shown that present aeolian sand movement and dune-shaping processes of the region are less strong compared to the phases of aeolian deposition and dune formation of the Pleistocene past. REFERENCES Adams, L.J. and Tetzlaff, G., 1984, Did lake Chad exist around 18.000 yrs BP. Archives for Meteorology, Geophysics and Bioclimatology, 34, pp. 299–308. Aumassip, G., 1993, Chronologies de l’art rupestre saharien et nord africain. Gandini, Calvisson, p. 32. Ballouche, A. and Neumann, K., 1995, A new contribution to the Holocene vegetation history of the West African Sahel: pollen from Oursi, Burkina Faso and charcoal from three sites in Northeast Nigeria. Vegetation History and Archaeobotany, 4, pp. 31–39. Barth, H.K. and Blume, H., 1975, Die Schichtstufen in der Umrahmung des MurzukBeckens (libysche Zentralsahara). Zeitschrift für Geomorphologie N.F., Suppl. 23, pp. 118–129. Battarbee, R.W., Gasse, F. and Stickley, C.E., (Eds.), 2004, Past climate variability through Europe and Africa. Developments in Palaeoenvironmental Research, Suppl. 23, Springer Netherlands. Baumhauer, R., 1986, Zur jungquartären Seenentwicklung im Bereich der Stufe von Bilma (NE-Niger). Würzburger Geographische Arbeiten, 65, pp. 1–235. Baumhauer, R., 1987, Holozäne limnische Akkumulationen im Bereich der Stufen von Zoo Baba und Dibella (NE-Niger). Palaeoecology of Africa, 18, pp. 167–177. Baumhauer, R., 1987a, Holocene limic accumulations in the Great Erg of Bilma. In: Petit-Maire, N. and Vanbesien, C., (Eds.), Past and future evolution of deserts, IGCP 252, pp. 1–11. Baumhauer, R., 1987b, Das Kawar—holozäne Seen in einem Schichtstufenvorland. Verhandlungsbd. 46. Deutscher Geographentag, pp. 332–341.

Geomorphological and palaeoenvironmental research in the South-Central Sahara in review

15

Baumhauer, R., 1988, Holozäne limnische Akkumulationen im Großen Erg von Bilma, NE-Niger. Würzburger Geographische Arbeiten, 69, pp. 137–148. Baumhauer, R., 1988a, Radiocarbondaten aus NE-Niger. Würzburger Geographische Arbeiten, 69, pp. 53–70. Baumhauer, R., 1989, Palaeolakes in South Central Sahara—problems of palaeoclimatological interpretation. Geoöko plus, 1, pp. 22–23. Baumhauer, R., 1990, Zur holozänen Landschafts- und Klimaentwicklung in der zentralen Sahara am Beispiel von Fachi/Dogonboulo (NE-Niger). Berliner Geographische Studien, 30, pp. 35–48. Baumhauer, R., 1991, Palaeolakes of the South Central Sahara—problems of palaeoclimatological interpretation. Hydrobiologica, 214, pp. 347–357. Baumhauer, R., 1992, Radiocarbon- und Grundwasserisotopendaten aus NE-Niger. Würzburger Geographische Arbeiten, 84, pp. 90–112. Baumhauer, R., 1993, Probleme der paläoökolgischen Interpretation limnischer Akkumulationen im Ténéré, NE-Niger. Trierer Geographische Studien, 9, pp. 33–50. Baumhauer, R., 1997, Zur Grundwassersituation im Becken von Bilma, zentrale Sahara. Würzburger Geographische Arbeiten, 92, pp. 131–146. Baumhauer, R. and Schulz, E., 1984, The Holocene lake of Séguédine, Kaouar, Niger. Palaeoecology of Africa, 16, pp. 283–290. Baumhauer, R., Busche, D. and Sponholz, B., 1989, Reliefgeschichte und Paläoklima des saharischen Ost-Niger. Geographische Rundschau, 41, 9, pp. 493–499. Baumhauer, R. and Felix-Henningsen, P., 1997, Late Pleistocene and Holocene palaeoenvironmental records of Ténéré, Erg of Ténéré and Erg of Fachi-Bilma (Central Sahara): new implications from palaeolimnological and palaeopedological data. Geografica Fisica e dinamica quaternaria, Suppl. III/1, pp. 1–73. Baumhauer, R. and Hagedorn, H., 1989, Probleme der Grundwasserschließung im Kawar (Niger). Die Erde, 120, pp. 11–20. Baumhauer, R. and Hagedorn, H., 1990, Problems of ground water capture in the Kawar (Niger). Applied Geography and Development, 36, pp. 99–109. Baumhauer, R., Morel, A. and Tillet, T., 1991, Palaeoenvironments and prehistoric populations of the Sahara in the Upper Pleistocene: Air-Ténéré-Djado-Kawar.- IGCP 252, Past and future evolution of deserts, Solignac. pp. 151–186. Baumhauer, R., Morel, A. and Tillet, T., 1997, Southern Central Sahara: Air-Ténéré-DjadoKawar. In: Tillet T., (Ed.), Sahara—Palaeoenvironments and Prehistoric Populations in the upper Pleistocene, L’Harmattan, Paris, pp. 229–266. Baumhauer, R., Schulz, E. and Pomel, S., 2004, Environmental changes in the Central Sahara during the Holocene—the Mid-Holocene transition from freshwater lake into sebkha in the into sebkha in the Seggedim depression, NE-Niger. Lecture Notes in Earth Sciences, 102, pp. 33–47. Busche, D., 1982, Die geomorphologische Entwicklung des Westlichen Murzuk-Beckens, des Djado-Plateaus und des nördlichen Kaouar (Zentrale Sahara). Habil.-Schrift Univ. Würzburg, pp. 1–440. Busche, D., 1998, Die zentrale Sahara: Oberflächenformen im Wandel. Gotha, pp. 1–284. Busche, D., Grunert, J. and Hagedorn, H., 1979, Der Westliche Schichtstufenrand des Murzukbeckens (Zentral-Sahara) als Beispiel für das Gefügemuster des ariden Formenschatzes. Festschrift Deutscher Geographentag Göttingen, 1979, pp. 43–65. Busche, D. and Sponholz, B., 1998, Morphological and micromorphological aspects of the sandstone karst of Eastern Niger. Zeitschrift für Geomorphologie N.F., Suppl.-Bd. 85, pp. 1–18. Busche, D. and Stengel, I., 1993, Rezente und vorzeitliche äolische Abtragung in der Sahara von Ostniger. Petermann Geographische Mitteilungen, 137, pp. 195–218.

16

Roland Baumhauer et al.

Busche, D. and Heistermann, C., 1992, Wechselbeziehungen zwischen geomorphologischer und geomorphologischer Forschung in der Sahara von Ost-Niger. Würzburger Geographische Arbeiten, 84, pp. 169–200. Crowley, T.J., 1999, Correlating high-frequency climate variations. Palaeooceanography, 14, pp. 271–272. Cubasch, U., Voss, R., Hegerl, G.C., Waszkewitz, J. and Crowley, T.J., 1997, Simulation of the influence of solar radiation variations on the global climate with an ocean-atmosphere general circulation model. Climate Dynamics, 13, pp. 757–767. Dalloni, M., 1948, Matériaux pour l’étude du Sahara oriental. Région entre la Libye, le Tibesti et le Kaouar (Niger): Géologie et préhistoire. In: Mission scientifique du Fezzan (1944–1945) 6, 1, pp. 1–119. Elouard, P., 1959, Etude géologique et hydrogéologique des formations sedimentaires du Guebla mauritanien et de la vallée du Senegal. Thèse Université Paris, pp. 1–415. Erhart, H., 1953, La nature minéralogique et la genèse des sediments de la cuvette tchadienne. C.R. Acad. Sci., 237, pp. 401–403. Faure, H., 1965, Evolution des grands lacs sahariens à l’Holocène. Quaternaria, 16, pp. 167–175. Faure, H., 1966, Une importante periode humide du Quaternaire supérieur au Sahara. Bulletin ASEQUA, 10/11, pp. 1–13. Faure, H., 1966, Reconnaissance géologique des formations sedimentaires postpaléozoiques du Niger oriental. Mém. de B.R.G.M., 47, pp. 1–630. Faure, H., 1969, Lacs quaternaires du Sahara. Mitt. Intern. Ver. Limnol., 17, pp. 171–146. Faure, H., Manguin, E. and Nydal, R., 1963, Formations lacustres du Quaternaire supérieur du Niger oriental: Diatomites et ages absolus. Bulletin B.R.G.M., 3, pp. 41–63. Felix-Henningsen, P., 1983, Zur Genese und Vergesellschaftung von Böden auf den Altdünen der nördlichen Goz-Region im Sudan. Mitteilungen der Deutschen Bodenkundlichen Gesellschaft, 38, pp. 485–490. Felix-Henningsen, P., 1984, Zur Relief-und Bodenentwicklung der Goz-Zone Nordkordofans im Sudan. Zeitschrift für Geomorphologie N.F., 28, pp. 285–303. Felix-Henningsen, P., 1997, “Sumpferze” in der Sahara Ost-Nigers. Würzburger Geographische Arbeiten, 92, pp. 147–169. Felix-Henningsen, P., 1998, Genese und paläoökologische Indikation von fossilen Böden mit Oxidkrusten auf Altdünen der Sahara Ost-Nigers. Zentralblatt für Geologie und Paläontologie Teil I, H 1–2, pp. 59–76. Felix-Henningsen, P., 2000, Palaeosols on Pleistocene dunes as indicators of palaeomonsoon events in the Sahara of East-Niger. Catena, 41, pp. 43–60. Felix-Henningsen, P., 2004, Genesis and palaeo-ecological interpretation of swamp ore deposits at Sahara palaeolakes of East Niger. In: Smykatz-Kloss, W. and FelixHenningsen, P. (Eds.), Palaeoecology of Quaternary drylands. Lecture notes in Earth Sciences, 102, pp. 47–72. Felix-Henningsen, P. and Mauz, B., 2005, Palaeo-environmental significance of soils on ancient dunes of the Northern Sahel and Sahara of Chad. Die Erde, 135, pp. 321–340. Ferhat, N. and Striedter, K.H., 1993, Art rupestre et paléoenvironnements. Résultats preliminaries de recherches dans la region de Dao Timmi (NE du Niger). Mem. Soc. Ital. Sc. Nat. XXVI/II, pp. 209–216. Flohn, H. and Nicholson, S., 1980, Climatic fluctuations in the arid belt of the “old world” since the last glacial maximum; possible causes and future implications. Palaeoecology of Africa, 12, pp. 3–22. Freydenberg, C., 1907, Exploration dans le Bassin du Tchad. La Géographie, pp. 35–46. Gasse, F., 1988, Diatoms, palaeoenvironments and palaeohydrology in the Western Sahara and the Sahel. Würzburger Geographische Arbeiten, 69, pp. 233–254.

Geomorphological and palaeoenvironmental research in the South-Central Sahara in review

17

Gasse, F., Fontes, J.Ch., Plaziat, J.C., Carbonel, P., Kacsmarska, J., De Decker, P., Soulie-Märsche, I., Callot, Y. and Dupeuple P., 1987, Biological remains, geochemistry and stable isotopes for the reconstruction of environmental and hydrological changes in the Holocene lakes from Northern Sahara. Palaeohydrology, Palaeoclimatology, Palaeoecology, 66, pp. 1–66. Gavaud, M., 1965, Interprétation chronologique des systémes de sols de dunes fixées du Niger méridional. Actes 6e Congr Panafr. Préhist. et. Quatern., pp. 376–380. Gellert, J., 1971, Das System der Entstehung und Gestaltung der Rumpfflächen, Inselberge samt Pedimenten und Rumpftreppen in Afrika. Petermanns Geographische Mitteilungen, 115, pp. 173–182. Gläser, B., 1987, Altdünen und Limnite in der nördlichen Republik Sudan als morphogenetisch-paläoklimatische Anzeiger. Untersuchungen zur jungquartären morphogenetischen Sequenz eines Regionalkomplexes. Akademie der Wissenschaften in Göttingen, pp. 1–193. Greigert, J. and Pougnet, R., 1967, Essai de description des formations géologiques de la République du Niger. Min. de Trav. Publ. et Transport Niamey, Niger, 3, pp. 1–238. Gronenborn, D., 1998, Archaelogical and ethnohistorical investigations along the southern fringes of Lake Chad. African Archaelogical Review, 15, 4, pp. 225–259. Grunert, J., 1970, Erosion und Akkumulation von Wüstengebirgsflüssen—eine Auswertung eigener Feldarbeiten im Tibesti-Gebirge. Unveröff. Staatsarbeit, pp. 1–127. Grunert, J., 1983, Geomorphologie der Schichtstufen am Westrand des Murzuk-Beckens (Zentrale Sahara). Relief, Boden, Paläoklima, 2, pp. 1–271. Grunert, J., 1988, Klima- und Landschaftsentwicklung in Ost-Niger während des Jungpleistozäns und Holozäns. Würzburger Geographische Arbeiten, 69, pp. 289–304. Guo, Z., Petit-Maire, N. and Kröpelin, S., 2000, Holocene non-orbital climatic events in present-day arid areas of Northern Africa and China. Global and Planetary Change, 26, pp. 97–103. Hagedorn, H., 1968, Über äolische Abtragung und Formung in der Südost-Sahara. Erdkunde, 22, pp. 257–269. Hagedorn, H., 1979, Zum Problem der inneren Gliederung der Wüsten. Stuttgarter Geographische Studien, 93, pp. 47–52. Hagedorn, H., 1985, Formen und Bilder der Wüste am Beispiel der Sahara. Natur und Museum, 115, 7, pp. 210–230. Hagedorn, H., 1988, Äolische Abtragungsformen im Massiv von Termit (NE-Niger). Würzburger Geographische Arbeiten, 69, pp. 277–288. Hagedorn, H., 1990, Beobachtungen über Paläowindrichtungen bei Faya im BorkouBergland (Tschad). Berliner Geographische Studien, 30, pp. 235–246. Hagedorn, H. and Sponholz, B., 1990, Silikatkarst in Nord- und Ostniger—Einbindung der Verkarstung in die tertiäre und quartäre Landschaftsgeschichte. Mainzer Geographische Studien, 34, pp. 77–90. Haynes, C.V., Eyles, C.H., Pavlish, L.A., Ritchie, J.C. and Rybak, M., 1989, Holocene palaeoecology of the Eastern Sahara: Selima oasis. Quaternary Science Reviews, 8, pp. 109–136. Herrmann, L., Sponholz, B. and Stahr, K., 1994, Quellregionen für den Harmattan-Staub in Westafrika: Ein mineralogischer und geochemischer Ansatz. Mitteilungen der Deutschen Bodenkundlichen Gesellschaft, 74, pp. 367–370. Herrmann, L., 1996, Staubdeposition auf Böden Westafrikas. Hohenheimer Bodenkundliche Hefte, 36, pp. 1–239. Hoelzmann, P., 1992, Palaeoecology of Holocene lacustrine sediments within the West Nubian Basin, SE-Sahara. Würzburger Geographische Arbeiten, 84, pp. 59–71.

18

Roland Baumhauer et al.

Hoelzmann, P., Kruse, H.-J. and Rottinger, F., 2000, Precipitation estimates for the Eastern Sahara palaeomonsoon based on a water balance model of the West Nubian Palaeolake Basin. Global and Planetary Change, 26, pp. 105–120. Jäkel, D., 1971, Erosion und Akkumulation im Enneri Bardague- Araye des Tibesti-Gebirges (Zentrale Sahara) während des Pleistozäns und Holozäns. Berliner Geographische Abhandlungen, 10, pp. 1–55. Jäkel, D., 1979, Run-off and fluvial formation processes in the Tibesti mountains as indicators of climatic history in the Central Sahara during the Late Pleistocene and Holocene. Palaeoecology of Africa, 11, pp. 13–36. Kilian, C., 1937, Esquisse géologique du Sahara francais à l’Est du 6° degré de longitude. Chron. Min. Colon. Kilian, C. and Furon, R., 1934, Primaire et Crétacé entre le Tibesti et l’Air. C.R. Acad. Sciences, 198, 20, pp. 1798–1800. Klitzsch, E., 1970, Die Strukturgeschichte der Zentralsahara. Neue Erkenntnisse zum Bau und zur Paläogeographie eines Tafellandes. Geologische Rundschau, 59, 2, pp. 459–527. Klitzsch, E., 1971, The structural development of parts of North Africa since Cambrian Time. In: Gray, C. (Ed.), Symposium on the geology of Libya, University of Libya, pp. 253–262. Klitzsch, E., 1974, Bau und Genese der Grarets und Alter des Großreliefs im Nordostfezzan (Südlibyen). Zeitschrift für Geomorphologie N.F., 18, pp. 99–116. Klitzsch, E., Sonntag, C., Weistroffer, K. and Shazley, E.M., 1976, Grundwasser der Zentralsahara: Fossile Vorräte. Geologische Rundschau, 65, 1, pp. 264–287. Kröpelin, S., 1993, Zur Rekonstruktion der spätquartären Umwelt am unteren Wadi Howar (Südöstl. Sahara/Nordwest Sudan). Berliner Geographische Abhandlungen, 54, pp. 1–293. Kröpelin, S., 1993, Environmental change in the Southeastern Sahara and the proposal of a geo-biosphere reserve in the Wadi Howar area (NW-Sudan). In: Thorweihe, U. and Schandelmeier H. (Eds.): Geoscientific Research in Northeast Africa, pp. 561–568. Kröpelin, S., 1999, Terrestrische Paläoklimatologie heute arider Gebiete: Resultate aus dem unteren Wadi Howar (südöstliche Sahara/Nordwest Sudan). Nordost-Afrika: Strukturen und Ressourcen, pp. 446–506. Kröpelin, S. and Soulie-Märsche, I., 1991, Charophyte-remains from Wadi Howar as evidence for deep Mid-Holocene freshwater lakes in the Eastern Sahara of Northwest Sudan. Quaternary Research, 36, pp. 210–223. Kuper, R. and Kröpelin, S., 2006, Climate-controlled Holocene occupation in the Sahara: Motor of Africa’s Evolution. Science, 313, pp. 803–807. Kutzbach, J.E., 1980, Estimates of past climates at palaeolake Chad, North Africa, based on a hydrological and energy balance model. Quaternary Research, 14, pp. 210–223. Macayeal, D.R., 1993, A low-order model of the Heinrich Event Cycle. Palaeooceanography, 8, pp. 767–773. Mainguet, M., 1968, Le Borkou. Aspects d’un modelé éolien. Annales de Géographie, 412, pp. 296–322. Mainguet, M., 1995, The saharo-sahelian Global Wind Action System: one facet of wind erosion analysed at a synoptic scale. Proc. Intern. Sem. On «Wind Erosion in WestAfrica», pp. 25–28. Maley, J., 1980. Etudes palynologiques dans le Bassin du Tschad et paléoclimatologie de l’Afrique nord-tropical de 30.000 ans à l’époque actuelle. Thèse, Université de Montpellier, pp. 1–520. Mauz, B. and Felix-Henningsen, P., 2005, Palaeosols in Saharian and Sahelian dunes of Chad: Archives of Holocene North African climate changes. The Holocene, 15, pp. 453–458.

Geomorphological and palaeoenvironmental research in the South-Central Sahara in review

19

Michel, P., 1975, Les basins des fleuves Sénégal et Gambie. Etude géomorphologiques. Mémoires de l’O.R.S.T.O.M., 63, pp. 1–752. Molle, H.G., 1969, Terrassenuntersuchungen im Gebiet des Enneri Zoumri (TibestiGebirge). Berliner Geographische Abhandlungen, 8, pp. 23–31. Molle, H.G., 1971, Gliederung und Aufbau fluviatiler Terrassenakkumulationen im Gebiet des Enneri Zoumri (Tibesti-Gebirge). Berliner Geographische Abhandlungen, 13, pp. 1–53. Montoya, M., Crowley, T.J. and Von Storch, H., 1998, Temperatures at the last interglacial simulated by means of a coupled general circulation model. Palaeooceanography, 13, pp. 170–177. Neumann, K., 1989, Vegetationsgeschichte der Ostsahara im Holozän. In: Kuper, R. (Ed.), Forschungen zur Umweltgeschichte der Ostsahara, pp. 12–181. Neumann, K., Schoch, W.H., Détienne, P. and Schweingruber, F.H., 2001, Woods of the Sahara and the Sahel—an anatomical atlas. Bern, pp. 1–321. Pachur, H.J., 1966, Untersuchungen zur morphoskopischen Sandanalyse. Berliner Geographische Abhandlungen, 4, pp. 1–35. Pachur, H.-J., Kröpelin, S., Hoelzmann, P., Goschin, M. and Altmann, N., 1990, Late Quaternary fluvio-lacustrine environments of Western Nubia. Berliner Geowissenschaftliche Abhandlungen (A) 120, 1, pp. 203–260. Pachur, H.-J. and Hoelzmann, P., 1991, Palaeoclimatic implications of Late Quaternary lacustrine sediments in Western Nubia, Sudan. Quaternary Research, 36, pp. 257–276. Pachur, H.-J. and Kröpelin, S., 1987, Wadi Howar: Palaeoclimatic evidence from an extinct river system in the Southeastern Sahara. Science, 237, pp. 298–300. Pachur, H.-J. and Wünnemann, B., 1996, Reconstruction of the palaeoclimate along 30°E in the Eastern Sahara during the Pleistocene/Holocene transition. Palaeoecology of Africa, 24, pp. 1–32. Pachur, H.-J., 1997, Der Ptolemäus-See in Westnubien als Paläoklimaindikator. Petermanns Geographische Mitteilungen, 141, 4, pp. 227–250. Pachur, H.-J., 1999, Paläo-Environment und Drainagesysteme der Ostsahara im Spätpleistozän und Holozän. Nordost-Afrika: Strukturen und Ressourcen, pp. 366–445. Pachur, H.-J. and Altmann, N., 2006, Die Ostsahara im Spätquartär—Ökosystemwandel im größten hyperariden Raum der Erde. Berlin, pp. 1–662. Petit-Maire, N., 1987, Local responses to recent global climatic change: hyperarid Central Sahara and coastal Sahara. In: Matheis, G. and Schandelmeier, H. (Eds.), Current Research in African Earth Sciences, pp. 431–434. Petit-Maire, N., 1991, Recent Quaternary climatic change and man in the Sahara. Journal of African Earth Sciences, 12, 1–2, pp. 125–132. Petit-Maire, N., 1994, Natural variability of the Asian, Indian and African monsoons over the last 130 ka. In: Desbois, M. and Désalmand, P., Global precipitations and climate change. NATO ASI Series, 126, pp. 3–26. Petit-Maire, N. and Kröpelin, S., 1991, Les climats holocenes du Sahara le long du tropique du cancer. In. Petit-Maire, N. (Ed.). Palaeoenvironnements du Sahara. Lacs holocenes à Taoudenni (Mali). CNRS, pp. 205–210. Petit-Maire, N. and Riser, J., 1983, Sahara ou Sahel. CNRS, Marseille, pp. 1–473. Pfeiffer, L., 1991, Schwermineralanalysen an Dünensanden aus Trockengebieten mit Beispielen aus Südsahara, Sahel und Sudan sowie Namib und der Taklamatan. Bonner Geographische Abhandlungen, 83, pp. 1–216. Pfeiffer, L. and Grunert, J., 1989, Heavy mineral associations in dune sands and soils of the Sahel and the Sudan (Niger, Burkina Faso, Togo). Palaeoecology of Africa, 20, pp. 55–68. Plauchut, B., Faure, H. and Jaquemont, P., 1960, Note explicative sur la carte géologique du bassin du Djado (feuilles Djado et Toummo), 1:500.000. Rapport B.R.G.M. Dak-54-A13, pp. 1–38.

20

Roland Baumhauer et al.

Pirard, F., 1964, Reconnaissance hydrogéologique du Niger oriental. Mémoires B.R.G.M. A 11, pp. 1–109. Reichelt, R., Faure, H. and Maley, J., 1992, Die Entwicklung des Klimas im randtropischen Sahara-Sahelbereich während des Jungquartärs—ein Beitrag zur angewandten Klimakunde. Petermanns Geographische Mitteilungen, 136, pp. 69–79. Ritchie, J.C. and Haynes, C.V., 1983, Holocene vegetation zonation in the Eastern Sahara. Nature, 330, pp. 352–354. Roset, J.-P., 1982, New data on the South saharian neolithisation problem: Air and Ténéré in Niger. XI INQUA Congress, Moscow, II, pp. 1–268. Roset, J.-P., 1987, Palaeoclimate and cultural conditions of Neolithic development in the Early Holocene of Northern Niger (Air and Ténéré). In: Close, A. (Ed.), Prehistory of Arid North Africa, pp. 211–234. Salzmann, U., 2000, Are modern savannas degraded forests? A Holocene pollen record from the Sudanian vegetation zone of NE Nigeria. Veget. Hist. Archaeobot, 9, pp. 1–15. Salzmann, U. and Waller, M., 1998, The Holocene vegetational history of the Nigerian Sahel based on multiple pollen profiles. Palaeobot. Palynol, 100, pp. 39–72. Schulz, E., 1980, Zur Vegetation der östlichen zentralen Sahara und zu ihrer Entwicklung im Holozän. Würzburger Geographische Arbeiten, 51, pp. 1–194. Schulz, E., 1991, Holocene environments in the Central Sahara. Hydrobiologica, 214, pp. 359–365. Schulz, E., Joseph, A. and Baumhauer, R., 1990, Upper Pleistocene and Holocene history of the Bilma region (Kawar). XVe Coll. de Géologie Africaine, Nancy, pp. 281–284. Servant, M., 1973, Séquences continentals et variations climatiques: Evolution du Bassin du Tchad au Cénozoiques supérieur. Thèse d’Etat Universté de Paris, pp. 1–348. Servant-Vildary, S., 1978, Etudes des diatomées et paléolimnologie du bassin tchadien au Cénozoique supérieur. Trav. Doc. O.R.S.T.O.M., 84, pp. 1–346. Sirocko, F., Sarntheim, M., Erlenkeuser, H., Lange, H., Arnold, M. and Duplessy, J.C., 1993, Century scale events in monsoonal climate over the past 24.000 years. Nature, 364, pp. 322–324. Skowronek, A., 1988. Böden als Indikator klimagesteuerter Landformung in der zentralen Sahara. Relief, Boden, Paläoklima, 5, pp. 1–184. Sponholz, B., 1992, Karstverfüllungen in Ostniger. Würzburger Geographische Arbeiten, 84, pp. 131–168. Stahr, K., Herrmann, L. and Jahn, R., 1994, Long distance dust transport in the SudanoSahelian zone and the consequences for soil properties. In: Buerkert, B.., Allison, B.R. and Oppen, M. von (Eds.), Proceedings of the International Symposium “Wind erosion in West Africa”, pp. 23–33. Stauffer, B., Blunier, T., Dällenbach, A., Indermühle, A., Schwander, J., Stocker, T.F. and Tschumi, J., 1998, Atmospheric CO2 concentration and millenial-scale climate change during the last glacial period. Nature, 392, pp. 59–62. Stengel, I., 1992, Morphodynamic types of longitudinal dunes in the Ténéré desert and the Erg of Bilma (Rep. of Niger). Würzburger Geographische Arbeiten, 84, pp. 147–168. Striedter, K.H., 1992, Rock art research on the Djado Plateau, Niger. IGNCA Rock Art Series, 1, pp. 113–128. Striedter, K.H., 1997, Rock art and palaeoenvironment in the Djado Plateau, Central Sahara (North-Eastern Niger). Rock art research—Moving in the 21st century, Conference Proc., Pt 1, Okahandjy, pp. 94–102. Striedter, K.H., 1999, Gravures rupestres de Dadafuy (Niger nord-oriental). Sahara, 11, pp. 113–117. Striedter, K.H., Vernet, R., Ferhat, N., Oumarou, A. and Tauveron, M., 1995, Prähistorische und paläoökologische Forschungen im Djado-Plateau, Nordost-Niger. Beiträge zur allgemeinen und vergleichenden Archäologie 15, pp. 49–84.

Geomorphological and palaeoenvironmental research in the South-Central Sahara in review

21

Thiemeyer, H., 2002, Maiduguri’s natural environment. In: Kawka, R. (Ed.). From Bulamari to Yerwa to Metropolitan Maiduguri. Westafrikanische Studien, 24, pp. 67–78. Tillet, T., 1983, Le paléolithique du bassin tchadien septentrional (Niger/Tchad). C.N.R.S. Paris, pp. 1–379. Tillet, T., 1989, La fin de temps paléolithiques dans le Bassin Tchadien nord-occidental. Studies in African Archeology, 2, pp. 507–511. Von Storch, H., Cubasch, U. and Gonzales-Rouco, J.F., 2000, Combining palaeoclimatic evidence and GCMs by means of data assimilation through upscaling and nudging. Proc. 11th Symp. on global change studies, pp. 28–31. Völkel, J., 1988, Zum jungquartären Klimawechsel im saharischen und sahelischen Ostniger aus bodenkundlicher Sicht. Würzburger Geographische Arbeiten 69, pp. 255–276. Völkel, J., 1989, Geomorphologische und pedologische Untersuchungen zum jungquartären Klimawechsel in den Dünengebieten Ost-Nigers (Südsahara und Sahel). Bonner Geographische Abhandlungen, 79, pp. 1–258. Völkel, J., 1991, Staubsedimentation im nordafrikanischen Sahel—Herkunft und Auswirkung auf die Landschaftsökologie eines semiariden Großraumes. Zeitschrift für Geomorphologie N.F., 89, pp. 73–85. Völkel, J. and Grunert, J., 1990, The problem of dune formation and dune weathering during the Late Pleistocene and Holocene in the Southern Sahara and Sahel. Zeitschrift für Geomorphologie N.F., 34, pp. 1–17. Warren, A., 1970, Dune trends and their implications in the Central Sudan. Zeitschrift für Geomorphologie N.F., 10, pp. 154–180. Wasylikowa, K., 1992, Plant macrofossils from the archaeological sites Uan Muhuggiag and Ti-n-Torha/Two Caves, Tatdrad Acacus, SW-Libya. Proceed. Intern. Symp. “Environmental changes and human culture in the Nile Basin and Northern Africa”, pp. 24–27.

CHAPTER 2

Comparison of proxy-based palaeoenvironmental reconstructions and hindcast modelled annual precipitation—a review of Holocene palaeoenvironmental research in the Central Sahara Brigitta Schütt and Jan Krause Department of Earth Sciences, Physical Geography, Freie Universität, Berlin, Germany ABSTRACT: The development of Late Quaternary palaeoenvironmental conditions in Northern Africa west of Sudan is described and discussed on the basis of a literature review of proxybased palaeoenvironmental research. Reconstruction of these palaeoenvironments includes references to analyses of all types of proxies from primary and secondary literature, especially of palynological, lacustrine, pedological and archaeological on-site findings. Palaeoenvironmental information given by climate proxy data is set in relation to palaeoclimates derived from hindcast modelling (Flohn and Nicholson, 1980; Kutzbach, 1979), and the difficulties of comparing the two approaches are demonstrated. Differences in lag time of system reactions can also be reflected in the palaeoenvironmental proxies, whereas hindcast modelling points to the climatic impulses triggering these reactions. It is emphasised that a valid palaeoenvironmental reconstruction needs to differentiate clearly whether proxy data provide information on pre-, syn- or post-sedimentary conditions and whether they refer to the source area or to the depositional environment.

2.1 INTRODUCTION 2.1.1 Objectives The fragmentation of Holocene palaeoenvironmental conditions of Northern Hemisphere Africa is a major topic in scientific literature. On the basis of an evaluation of primary literature, the Late Quaternary palaeoenvironmental conditions of the Sahara are analysed and collated. The resulting diachronous overview of palaeoenvironmental reconstruction discloses regional differences. The aim of this paper is to explore whether palaeoenvironmental conditions as given by climate proxy can be directly related to palaeoclimate data as given by climate hindcast modelling. The analysis is based on data from selected primary and secondary literature documenting research on different types of archives such as lacustrine sediments, palaeosoils and archaeological findings, and based on different multi-proxy approaches (pollen, chemistry, minerals). In total, 52 original publications concerning Northern Africa west of Sudan were taken into account for this analysis. In fact, many more publications exist. But an intensive review shows that many results were reproduced and published more than once, frequently in combination with

24

Schütt and Krause

Figure 1. Location of research areas of the analysed literature. Numbering according to figure 2.

other data. Extractions of palaeoenvironmental information from the literature analysis are spatially referenced (Figure 1), temporally itemised (Figure 2) and diachronously displayed (Figure 3a–c). On this basis it is also intended to re-assess the spatially differentiated trend of Holocene aridisation of the Central Sahara. For comparison, the CCM0-modelled hindcast of annual precipitation is used. These data are available with a resolution of 7,5 * 4,4° (Kutzbach and Guetter, 1986 and Wright et al., 1993) (Figure 3d–f). 2.1.2 Study site Environmental conditions in the Central Sahara are strongly driven by climate. On a continental scale, it is evident that the present-day distribution of ecoregions is similar to the distribution of climates (White, 1983; Schultz, 2005). In the Central Sahara the spatial distribution of rainfall and its seasonality are determined by the general circulation pattern of Africa. In the Northern Hemisphere summer the ITCZ (Intertropical Convergence Zone) and an associated low pressure cell over the Sahara lie between the moist SW monsoon and the dry NE harmattan (trade wind). During the Northern Hemisphere winter months the ITCZ is located further South and a high pressure cell is established over the Sahara. At this time of the year the northern part of Northern Africa is under the influence of the rainbearing mid-latitude westerlies, whereas its southern part is influenced by the

A review of Holocene palaeoenvironmental research in the Central Sahara

25

Figure 2. Palaeoenvironmental character over the last 20 ka cal BP for selected study sites of Northern Africa, for localization compare figure 1. The chart shows the palaeoenvironment differentiated by studies and is divided into more arid (light grey), more humid (black) and transition periods (dark grey). Additionally light triangles mark arid intervals, while dark ones point to humid intervals. White triangles denote the end of more humid conditions. The boundaries of humidity and aridity phases are set based on the median probability of the lower and higher age range (given by CALIB 5.0.1).

dry NE harmattan. Hence, Northern Africa separates into a winter and a summer rainfall regime with the Sahara in between (Nicholson, 2001). Thus, the central part of the Sahara is presently under the influence of the subtropical high-pressure belt with little impact by either the westerlies or the ITCZ (Weischet and Endlicher, 2000). Rainfall is only erratic, and mean annual rainfall totals 50–100 mm with parts of the area receiving no rainfall for years. Perennial vegetation only occurs at locations where groundwater is near-surface such as oases and mountain areas with higher rainfall (Beuchelt, 1968). Ephemeral xerophytes are found after erratic rainfall events (Busche and Stengel, 1993).

26

Schütt and Krause

Figure 3. Left column: Spatial distributed palaeoenvironmental reconstruction for the time slices 9, 6, 3 ka cal BP referring to the literary analysis. Right column: Spatial distributed results for the CCM0 hindcasts for the time slices 9, 6, 3 ka BP (modified after Kutzbach and Guetter, 1986; Wright et al., 1993; (www.ncdc.noaa. gov/paleo/modelvis.html).

The area of interest can be clustered into five sub-regions (Figure 1): − the Mediterranean Maghreb in the climatic transition zone from a Mediterranean environment with winter rainfall to a desert environment, − the Western Sahara (including Western Sahara, Mauritania and Mali) with a dry climate due to the influence of the cold Canaries current (coastal desert) and the dry NE trade winds,

A review of Holocene palaeoenvironmental research in the Central Sahara

27

− the Central Sahara with locally hyperarid conditions caused by high continentality and NE trade winds nearly all year round, − the Sahelian Zone with a subarid to dry–subhumid climate, receiving its moisture primarily from the SW monsoon, − the Sudanian-Guinean-Zone with a subhumid climate, also receiving its moisture from the SW monsoon. 2.2 METHODS AND SOURCES This literature analysis is mainly based on primary literature focusing on clear age determination of the palaeoenvironmental proxies and thus on palaeoenvironmental reconstruction. A further requirement is that the study site can be localised, even if it is sometimes not possible to geocode the sites. Owing to the necessity of clear age determinations the literature analysed is confined to the period since the 1970s, when 14C dating was established for palaeoenvironmental research. To make dating results comparable, data are checked to determine whether they are calibrated; if necessary calibration was performed, applying the method developed by M. Stuiver and P. J. Reimer with their calibration program CALIB 5.0.1 (copyright 1986–2005) to convert into calendar years (calBP). The time-driven processes of radioactive decay of 14C are the basis of this absolute physical dating method (Geyh, 1971, 1983). For the non-marine radiocarbon samples in the Northern Hemisphere the calibration curve IntCal04 (Stuiver et al., 2005) is used. The calibration curve error σk is set with an accuracy of 2-sigma (95% certainty) of the conventional age (Wagner, 1995). If lab errors are given in the literature these values are adopted; if not, the lab error is set at 0. The lab error is set as multiplier. The standard deviation of the conventional aging (SD years/age uncertainty) was also adopted if known, otherwise set to 1. The median of the calibrated calendar years is used to display the climate data. 2.3 LITERATURE ANALYSIS Palaeoenvironmental research was largely carried out in the Central Sahara with most sites in Mali, Niger and Tchad, indicating systematic studies over decades. By contrast, palaeoenvironmental research in the adjoining area of the Southern Sahel seems to be more erratic. To the North, in the Mediterranean Maghreb, palaeoenvironmental research also has a long tradition; thus, many publications are available. But as the regional focus of this paper is on the Central Sahara, only selected papers are considered. Regarding Libya, publications on the Sahara are frequently published in the Italian language and in journals or reports that are not internationally accessible and thus could not be considered. The same applies to miscellaneous French-language publications on the Algerian Sahara. The fragmentation of the Holocene climate in Northern Hemisphere Africa in temporal and spatial distribution is a major topic in scientific literature. Most palaeoenvironmental reconstructions focus on moisture conditions, differentiating between ‘arid’ or ‘dry’, ‘humid’ or ‘wet’ and ‘dryer’ and ‘wetter’. The definition of the dry and humid periods and the evaluation of moisture change vary depending on the proxies and timescale used. Clear definitions—e.g. by P/E-ratios or annual rainfall amounts—are lacking. Therefore the often comprehensive results given in the literature were, if necessary, attributed to three main categories (more humid, transition period and more arid) to describe the palaeoenvironment of the past 20 ka cal BP. The attribution was carried out as shown in the following two examples: − palaeolake levels: high lake level—humid, low lake level—arid − pollen records: Saharan type—arid, Sahelian type—transition period, SudanianGuinean-type—humid.

28

Schütt and Krause

2.3.1 From the Late Glacial period to the Early Holocene wet phase During the Last Glacial Maximum (LGM), around 18 to 20 ka cal BP, arid conditions are known to have occurred at many locations in Northern Africa, including the Sahara and its northern and southern margins (e.g. Rognon and Williams, 1977; Lézine et al., 1990; Reichelt et al., 1992). Only Flohn and Nicholson (1980) state that conditions were humid or at least transitional at the northern flank of the desert. The ‘first’ dune generation was deposited during this dry period in Saharan and sub-Saharan Africa, (Völkel and Grunert, 1990); the best known are the longitudinal Ogolian dunes in today’s Sahelian and Sudanian zone (Michel, 1973; Elouard, 1962). The length of this arid period varies in time and place from author to author, but it is repeatedly attested that humidity gradually increased after 16 ka cal BP (e.g. Völkel and Grunert, 1990; Reichelt et al., 1992). However, all palaeoenvironmental research carried out in the Central Sahara confirms that the Last Glacial Maximum arid phase was succeeded by the onset of a humid phase during the Late Glacial already: Flohn and Nicholson (1980a, b) assume a fluctuating climate between 14 and 10 ka cal BP, with a semiarid climate in the Northern Sahara (Flohn and Nicholson, 1980a), and a humid Southern Sahara (Flohn and Nicholson, 1980b). On the basis of investigations in the Zoumri-Bardagué-Arayé river system of the Tibesti Mountains, Jäkel (1979) constructs a climate curve from 19 ka cal BP onwards. He hypothesises humid conditions in the Tibesti Mountains from 19–8,9 ka cal BP with most humid conditions from 10,7–9,5 ka cal BP, caused by an overlapping of rainfall-bearing air masses occurring throughout the year and originating from both monsoonal influence during summer and westerlies influence during winter. For the neighbouring lowlands of the Massif du Termit, Gasse and van Campo (1994) set the onset of humid environmental conditions at 13,8 ka cal BP, documented in palaeolake deposits. These data almost agree with the palaeoenvironmental reconstruction for the Massif du Termit given by Servant (1973, 1983). The latter places the change from extremely arid (Kanemian/Ogolian) to moderately humid conditions ((Nigero-)Tchadien I) at around 15,3 ka cal BP. Gasse (1990) defines the first arid-to-humid transition (AHT) for Northern Algeria at around 17,5 to 13,8 ka cal BP. Lézine and Casanova (1989) assign the change in the Sahelian zone to around 14,7 ka cal BP. Figure 2 shows a shift in the onset of the first AHT from the Sudanian zone (Gasse and van Campo, 1994) to the Western Saharan region (Petit-Maire, 1988; Hillaire-Marcel et al., 1982). Finally, after 13,8 ka cal BP, during the Nigéro-Tchadien I, environmental conditions were humid nearly all over Northern Africa, and by 11,5–8,9 ka cal BP (Tchadien) the regional water balance reached its optimum (corresponding to the Early Holocene wet period) (Völkel and Grunert, 1990; Flohn and Nicholson, 1980; Servant and ServantVildary, 1980). The Sahel stretched up to 23°N, whereas today it is located around 18°N (Claussen et al., 2002). In the most recent literature humid environmental conditions during the Early Holocene are subsumed to the African Humid Period (AHP) which began between 11,5 ka cal BP (e.g. Renssen, 2006) or 14,8 ka cal BP (deMenorcal et al., 2000). During this phase Western Saharan and Sahelian climate was affected by (relatively) high summer temperatures and enhanced precipitation. For the period 9 to 7,5 ka cal BP Renssen (2006) hindcasts a precipitation of 290 mm/a due to a strong land-sea thermal gradient which increases wet air mass transport. This results in an average vegetation cover of 70%. By contrast, Early Holocene annual precipitation at Dibella (Eastern Niger) is estimated to have been at least 300–400 mm/a (Grunert et al., 1991) or 250–400 mm/a (Flohn and Nicholson, 1980). It is assumed that perennial rivers were flowing during this period even in the hyperarid desert centre between Kufra Oasis and Tibesti Mts. (Flohn and Nicholson, 1980). The improved humidity was caused by the intensification of the African monsoon due to changes of the orbital parameters (de Menocal et al., 2000). Around 10,6 ka cal BP

A review of Holocene palaeoenvironmental research in the Central Sahara

29

the intensity of the monsoon reached its maximum and lake water levels in the Sahara peaked (Baumhauer, 1986, 1991; Gasse et al., 1990; Flohn and Nicholson, 1980). The beginning of the Holocene humid phases corresponded with the end of glacial conditions in Europe and the North Atlantic Ocean and with a high concentration of methane in the atmosphere. The July insolation at 10–11 ka cal BP was 8% higher than today. The consequence was higher precipitation, 40% higher over North Africa than it is today (de Menocal et al., 2000; Kutzbach and Guetter, 1986; Prell and Kutzbach, 1987). 2.3.2

From the Early Holocene wet phase to the present

The climatic evolution to the hyperarid conditions that determine the climate today started in the Holocene (Baumhauer et al., 2004). The Early Holocene wet phase varied in duration depending on the location of the research area and on the beginning and the progression of the successional aridisation. Gasse (1990) states another two arid-to-humid transitions (AHT). The second AHT is dated 10,5–9,4 ka cal BP, and the third AHT from 8,3 to 8 ka cal BP. This instable phase lasted locally until 6 ka cal BP (Damnati, 2000). During this phase the annual precipitation and the average vegetation cover decreased to 210 mm/a and 50%, attended by an increased climate variability (Renssen et al., 2006). During this time, ongoing or amplified aridisation finally triggered the movement of Neolithic peoples, who had settled all over Northern Africa during the Early Holocene. They headed to the margins of the developing Sahara following the shifting Sahelian zone or to oasis regions such as the Nile valley or the mountain areas of Tibesti, Air and Hoggar (Reichelt et al., 1992; Flohn and Nicholson, 1980). The second dune generation of the Sahara was deposited in this period, indicating arid conditions (Nouakchottien after Michel, 1973; Völkel and Grunert, 1990). From 5,7 ka cal BP onwards the Sahara expanded, and its fringe moved approximately 450 km to the South and less extensively to the North (Reichelt et al., 1992). However, in the literature it is stated that the Holocene aridisation included one possible and one definitive interruption, i.e. phases of humidisation or temporarily stopped aridisation. Around 7,4 ka cal BP the second (Middle) Holocene humid phase started. It extended over the whole Sahara-Sahel and ended around 5,7 ka cal BP (Flohn and Nicholson, 1980; Geyh and Jäkel, 1973; Servant and Servant-Vildary, 1980; Gasse and van Campo, 1994). Petit-Maire (1987, 1988) sets the end of the freshwater lakes at around 5,2 ka cal BP. It is assumed that around 2,5 ka cal BP the climate finally changed to presentday conditions. Völkel and Grunert (1990) date the most recent dune generation to this period. 2.4 COMPARISON OF PALAEOENVIRONMENTAL RECONSTRUCTIONS AND ANNUAL PRECIPITATION 2.4.1 Source criticism Evaluating the available literature sources on proxy-based palaeoenvironmental reconstruction is complicated by the authors’ sometimes imprecise use of the terms palaeoclimate and palaeoenvironment. Both depend on each other, but mostly the coring or sampling results indicate palaeoenvironmental conditions that are sometimes described or interpreted as palaeoclimate indicators without any further comment. In addition, owing to fluctuating climates some uncertainty attaches to the results displayed in figure 2. For example, several sources state humid conditions for a certain period, although the climate was not constantly humid within the time slice selected. These fluctuations are

30

Schütt and Krause

frequently mentioned but neither clearly documented nor dated. Therefore in some cases we generalised the palaeoclimatic descriptions. In the literature analysed, palaeoclimatic or palaeoenvironmental conditions for the Central and Western Sahara are reconstructed. Their temporal assignment depends on the dating provided in the respective publication and thus reflects the uncertainty of several authors cited in this text. The ages of the publications and the exact notations of sample ages, especially in terms of calibration of 14C-data, may yield inaccuracies, especially as the dates used and shown in figure 2 are rarely confined to one single site. It is not always transparent whether the dates posted in the respective publication are calibrated or uncalibrated 14C measurements or whether different dating methods may have been used. 2.4.2 Community Climate Model (CCM0) The palaeoprecipitation data displayed in figure 3 (d–f) were hindcasted by Kutzbach and Guetter (1996) and Wright et al. (1993), using the Community Climate Model (CCM0) of the National Center for Atmospheric Research. Owing to the spatial resolution of the data provided and the model setup, these data contain strong generalisations and uncertainties. CCM0 includes atmospheric dynamics, which is based on the equations of fluid motions. It utilises radiative and convective processes, condensation and evaporation. The horizontal resolution totals 4° latitude and 7,5° longitude. The input data required are insolation, atmospheric gas concentrations, land albedo and soil moisture. Additionally the model considers mountain and ice-sheet orography, sea-surface temperature, sea-ice limits and snow cover. CCM computes an array of surface and upper-air parameters. The atmospheric energy budgets are subdivided into kinetic and potential energy and into zonal, stationaryeddy and transient-eddy components. Sea-level pressure and geopotential height fields are calculated. Annual precipitation is one of the output data the model provides. Modelling results are available in 3.000-year steps up to Last Glacial Maximum (20 ka cal BP). 2.4.3 Comparison The three time slices considered in this comparison correspond to the time slices of the available CCM0 modelling results and the palaeoenvironmental age information based on proxy data. Consequently, the diachronous overview focuses on the 9 ka cal BP, 6 ka cal BP and 3 ka cal BP time slices (see figure 3). 9 ka cal BP. This time slice falls within the African Humid Period (e.g. Renssen, 2006; de Menocal et al., 2000). Proxy data explicitly point out humid conditions all over Northern Africa west of Sudan (Figure 3a). This situation is imaged for all regions examined: ranging from the Maghreb, whose present-day precipitation predominantly originates from the westerlies, over the Western and Central Sahara, which are both presently strongly affected by the trade winds, to the Sahelian zone and the Sudan-Guinean zone, where present-day precipitation is generated by the monsoon. A regional differentiation of the palaeoenvironmental conditions due to humidity-aridity cannot be derived from the proxy data. All authors agree that maximum humidity during the Holocene occurred between ∼11–8,8 ka cal BP (Early Holocene wet phase). By contrast, the precipitation data provided by CCM0 (Figure 3d) distinctly show a zonal distribution with minimum values in the Maghreb and Western Sahara. Increased precipitation is shown for the Sahelian and the Sudan-Guinean zones with a peak in the area of the Niger river course. 6 ka cal BP. A first short aridisation phase in the Early Holocene, around 7,4 ka cal BP, was followed by a humid period, the Mid-Holocene wet period. Especially the Chad

A review of Holocene palaeoenvironmental research in the Central Sahara

31

basin was still dominated by high lake levels and freshwater conditions (e.g. Servant, 1983; Servant and Servant-Vildary, 1980; Fontes and Gasse, 1990). However, the main feeders of the Chad basin discharge from the South, where the more humid conditions of the Sahelian-Saharan and Sudanian-Guinean zones occur. By contrast, the Western Sahara features more arid conditions (e.g. Lézine et al., 1990) (Figure 3b). Also the frequent lack of proxy data for this time slice indicates more arid conditions, the hiatuses probably being due to deflation or the lack of carbon-rich sediments appropriate for dating. In the Southern Sahelian and the Sudanian-Guinean zones proxies indicate predominantly humid conditions (e.g. Fontes and Gasse, 1990b; Lézine and Casanova, 1989b). The same applies to the Maghreb, especially in sites close to mountainous or coastal regions (e.g. Faust et al., 2004; Zielhofer et al., 2004; Lamb et al., 1995; Gasse et al., 1990; Fontes and Gasse, 1990d, e; Rognon and Williams, 1977b). This palaeoenvironmental image given by the proxy data is only partly reflected by the modelled palaeoprecipitation (Figure 3e). For the Western and Central Sahara as well as for the Maghreb annual precipitation is low, totalling 2–3 mm/d (after Kutzbach and Guetter, 1986; Wright et al., 1993). The corresponding annual precipitation of 110–200 mm depends on the number of rainfall days: an average of ∼55 annual rainfall days is assumed for the Sahelian stations N`Djamena, Zinder, Sokoto, Maiduguri (after Müller, 1996). By contrast, the Sahelian and the Sudanian-Guinean zones show annual precipitation with values around 4–5 mm/d. Assuming an annual average of 100 rainfall days according to the stations at Makurdi, Batouri, Enugu and Tamale (after Müller, 1996), this value corresponds to approx. 450–650 mm annual rainfall. Also in this time slice, the course of the Niger River is displayed by increased precipitation amounts (4–5 mm/d) and extends into the Sahara. 3 ka cal BP. The shift from the Mid-Holocene humid period to extremely arid conditions occurred after 5,5 ka cal BP according to Renssen (2006), Damnati (2000) and de Menocal (2000). In the Western and Central Sahara most of the palaeoenvironmental studies point to arid environmental conditions for the time slice selected (e.g. Flohn and Nicholson, 1980a, b; Hoelzmann et al., 2000; Hoelzmann, 1992; Servant, 1983; Lézine et al., 1990; Maley, 1977). An exception is made by Lézine and Casanova (1989) who worked on a site in the Chad basin. But again it needs to be considered that the Chad basin is fed by rivers draining from the South with their headwater areas in the more humid Sudanian-Guinean and Sahelian-Saharan zones. According to Völkel (1988) the fixation of the second dune generation by weathering and soil forming processes occurred during this phase. In the Mediterranean Maghreb, too, the palaeoenvironmental picture shown is incoherent and, including information from figure 2, indicates short-term palaeoenvironmental changes (e.g. Rognon, 1988; Faust et al., 2004; Lamb et al., 1995). South of the Sahara it becomes obvious that arid conditions shifted South and affected wide areas of the Sahelian zone (e.g. Waller and Salzmann, 1999; Rognon and Williams, 1977a). Humid conditions are confirmed at the sites of the Sudanian-Guinean zone (e.g. Gasse and van Campo, 1994e; Talbot and Delibrias, 1977, 1980; Lézine and Casanova, 1989a). Again the spatial distribution of palaeoenvironmental conditions given by the proxy data is only partly reflected by the modelled palaeoprecipitation (Figure 3f ). The image is similar to the situation at 6 ka cal BP, with increased annual precipitation only in the Sahelian and the Sudan-Guinean zones and along the Niger River course, varying in the reduced total amount of modelled annual precipitation. 2.4.4 Discussion There are several reasons for a major change from arid to humid and back to arid conditions in Northern Africa. One triggering criterion is the increase in the Northern Hemisphere

32

Schütt and Krause

summer solar radiation. This leads to the already mentioned enhanced land-sea thermal contrast and resulting increase of convective precipitation. These increasing rainfalls caused the AHP with all its local variations (cf. Renssen, 2006; Damnati, 2000; de Menocal, 2000). The shift from humid to recent arid conditions can partly be explained by the tilt of the Earth’s axis—which, at 10,2 ka cal BP, was stronger than today—with a resulting time of the perihelion at the end of July (Berger, 1978; Claussen, 1999). Yet the quite smooth variations in these orbital parameters in the Holocene scale are in contrast to climate and vegetation change which occurred abruptly in this timescale (Claussen, 1999). Input parameters for CCM0 model runs correspond to radiation attributes, modified by local conditions affecting insolation such as relief and albedo (Kutzbach and Guetter, 1986). Thus, imaging palaeoenvironmental conditions generated by hindcast climate models such as CCM0 focus on climate as a triggering factor and are visible at the zonal scale (Kutzbach and Guetter, 1986). Results from hindcast palaeoclimatical modelling reflect the effects of solar driven forces on the climate, modified by regional particularities such as the effect of expanded wetlands on the regional evaporation and water vapour content. By contrast, palaeoenvironmental proxies reflect the synsedimentary environmental conditions as controlled by climate forces, endogenic forces and humankind. Launched processes can be observed in matter flows such as surface runoff, erosion and sediment deposition (Schütt, 2004a, b). Palaeoenvironmental proxies represent reactions of the natural system to endogenic or exogenic triggers. Palaeoenvironmental reconstruction requires differentiation between the different types of triggers and the influence exerted by climate-driven impulses (Schütt, 2004a, b). Imaging palaeoenvironmental conditions as a result of proxy data analysis always points to regional system reactions on external or endogenic forcing—and thus is of local to regional scale. Consequently, regional palaeoclimatic conditions are not appropriate to explain the different behaviours of Saharan palaeolakes (Baumhauer, 1991). When analysing palaeoenvironmental proxies we need to consider, first, that processes launched by both exogenic and endogenic forces show short-term reactions and, second, that processes are initiated whose implications underlie longer time-lag effects. Groundwater recharge and groundwater flow belong in this process group. This is convincingly pointed out by Grunert et al. (1991), who show that two neighbouring Holocene palaeolakes of the Central Sahara—Dibella and Zoo Baba, located only 70 km away from each other in the Erg of Bilma—reacted differently, with different time lags and with different process intensities, to changing Holocene climate. During Early and Mid-Holocene wet phases, Dibella had brackish water with great fluctuations in salinity, whereas Zoo Baba was a freshwater lake even up to the Mid-Holocene. Despite the proximity of both systems, Dibella is fed by a local aquifer, whereas Zoo Baba is connected to the large Bilma/Kaouar aquifer. It is now well known that the former and present-day wetland areas of the Central Sahara, in particular the depressions along the foreland of the escarpments (cuestas), were predominantly groundwater fed. Thus, if they are part of larger groundwater systems longterm reactions to climate change need to be considered. With reference to Thorweihe et al. (1984), Baumhauer (1997) estimates that groundwater flow in the Bilma depression has a velocity of approximately 6 m/year. Because of the large extent of the Bilma/Kaouar aquifer with its headwater area in the Tibesti Mountains, a period of 50.000 years is required to bring naturally replenished groundwater from the Tibesti Mountains to the foreland depressions of the Kaouar. Thus, groundwater recharge of the lake depression is still active today, although regional environmental conditions are already dry. In view of the hydrological situation of the Chad basin with its receiving Lake Chad, another kind of teleconnection may have adulterated proxy-based regional palaeoenvironmental information: Lake Chad is mainly fed by river systems of the Yedseram, Chari, Logone and Erguig, perennial rivers with their headwater areas in the Mandara Mts. or the Massif de l’Adamoua—and thus located in

A review of Holocene palaeoenvironmental research in the Central Sahara

33

the Guinean zone with an average of 1.200–1.500 mm precipitation annually (cf. Müller, 1996). Thus, expected sensitive reactions of a desert margin area such as the Lake Chad basin to regional climate changes are masked by inflow from less sensitive areas. In consequence, it is difficult to compare palaeoenvironmental information derived from the application of both approaches because short-, medium- and long-term reactions are all reflected in the palaeoenvironmental proxies, whereas hindcast modelling points to the climatic impulses triggering these reactions. At any rate, an individual, site-specific analysis of proxies is required to assess the factors controlling the processes. 2.5 CONCLUSIONS Palaeoenvironmental reconstruction is traditionally based on proxy data obtained from sediment analysis. Depending on the sediment facies and the proxies used, different information on the palaeoenvironment is derived. For instance, the sediment’s pollen content and composition allows conclusions about the synsedimentary vegetation cover and, if transfer functions are used, even about palaeoclimatological information. In contrast, sediment structure gives information on the relief-forming processes, and clay mineral composition yields information about presedimentary weathering conditions. In addition, we need to differentiate between proxies showing − synsedimentary palaeoenvironmental conditions such as salinity of a playa lake (e.g. carbonate composition, diatom and ostracode assemblages), − presedimentary palaeoenvironmental conditions such as soil-forming processes (e.g. soil sediments in lake deposits), or − long-term reactions, for example by a lake water body in a groundwater-fed system. Auditing the different palaeoenvironmental proxies shows that the quality, accuracy and complexity of the palaeoenvironmental reconstruction strongly depends on the archives and proxies used and on the specific concepts and approaches applied for analysis. Lake archives, especially of endorheic lakes, are ranked as the most valuable sedimentary archives for reconstructing the palaeoenvironment, as they have relatively few hiatuses and are frequently even laminated, corresponding to seasonal or event layers. However, final deposition of sediments takes place in a lake, whereas temporary deposition occurs in the drainage basin. Thus, lake sediments mainly provide information on the synsedimentary palaeolimnic environments, most likely without temporal interruption, whereas the generally temporally discontinuous fluvial and colluvial deposits in drainage basin archives provide high resolution and spatially differentiated information on synsedimentary morphodynamics and flow dynamics. Integrating studies, combining palaeoenvironmental environment and process information, are carried out only rarely— but ultimately these are the basis for understanding a system. Furthermore, it is common practice for palaeoenvironmental information to be considered applicable to the entire region, while ignoring that it is generally of punctiform character. In summary, palaeoenvironmental information derived from proxy data corresponds to conclusions on system reactions to exogenic, endogenic or human induced forces. In contrast, palaeoenvironmental information derived from climate hindcast modelling corresponds to exogenic triggers forcing geomorphological system reactions. Consequently, an integrated analysis of palaeoenvironmental information provided by proxy data and derived from climate hindcast modelling shows regionally diverging information. To allow further application and utilisation of palaeoenvironmental information derived from proxy data, data analysis must include the disclosure and consideration of all system

34

Schütt and Krause

characters. Without this, it will be difficult to assess whether the palaeoenvironmental indicators relate to short- or long-term effects. Satisfactory results are likely to be obtained by differentiating between these effects and by comparing palaeoenvironmental reconstructions based on proxy data (indicating system reactions) with palaeoclimatic data based on hindcast modelling (showing system impulses triggering the system reactions). ACKNOWLEDGEMENTS We would like to thank the German Research Foundation (DFG) for its financial support of the ‘Limnosahara’ research project (Schu 949/8), funded since 2005. Thanks are also due to Nicole Marquardt and Katharina Ducke for their considerable support in preliminary work within this project. REFERENCES Baumhauer, R., 1986, Zur jungquartären Seenentwicklung im Bereich der Stufe von Bilma (NE-Niger). Würzburger Geographische Arbeiten, 65, pp. 1–235. Baumhauer, R., 1987, Holozäne limnische Akkumulationen im Bereich der Stufen von Zoo Baba und Dibella (NE-Niger). Palaeoecology of Africa, 18, pp. 167–177. Baumhauer, R., 1988, Das Kawar-Holozäne Seen in einem Schichtstufenland. In: 46. Deutscher Geographentag München, 46, pp. 332–341. Baumhauer, R., 1997, Zur Grundwassersituation im Becken von Bilma, zentrale Sahara. Würzburger Geographische Arbeiten Band, 92, pp. 131–146. Baumhauer, R., 1991, Palaeolakes of the South Central Sahara: problems of palaeoclimatological interpretation. Hydrobiologia, 214, pp. 347–357. Baumhauer, R. and Hagedorn, H., 1990, Probleme der Grundwassererschließung im Kawar (Niger), Die Erde, 120, pp. 11–20. Baumhauer, R. and Schulz, E., 1984, The Holocene lake of Seguedine, Kaouar, NE Niger. Palaeoecology of Africa, 16, pp. 283–290. Baumhauer, R., Schulz, E. and Pomel, S., 2004, Environmental changes in the Central Sahara during the Holocene—the Mid-Holocene transition from freshwater lake into sebkha in the Segedim depression, NE-Niger. Lecture Notes in Earth Sciences, 102, pp. 33–47. Berger, A., 1978, Long-term variations of daily insolation and quaternary climatic changes. Journal of Atmospheric Sciences, 35, pp. 2362–2367. Beuchelt, E., 1968, Niger. Deutsche Afrika-Gesellschaft e.V. Bonn, Die Länder Afrikas, 38, Kurt Schroeder, Bonn, pp. 1–143. Busche, D. and Stengel, I., 1993, Rezente und vorzeitliche äolische Abtragung in der Sahara von Ostniger. Petermanns Geographische Mitteilungen, 137, 4, pp. 195–218. Claussen, M., Brovkin, V. and Ganopolski., 2002, Africa: Greening of the Sahara Africa: a hot spot of non-linear atmosphere-vegetation interaction. In: Steffen, W., Jäger, J., Carson, D.J., and Bradshaw, C., Challenges of a changing earth. Proceedings of the Global Change Open Science Conference, Amsterdam, the Netherlands, 10.–13.7.2001, Springer Verlag, Berlin, Heidelberg, N.Y., pp. 125–128. Claussen, M. and Brovkin, V., et al., 1999, Simulation of an abrupt change in Saharan vegetation in the Mid-Holocene. Geophysical Research Letters, 26, 14, pp. 2037–2040. Damnati, B., 2000, Holocene lake record in the Northern Hemisphere of Africa. Journal of African Earth Sciences, 31, 2, pp. 253–262. De Menocal, P., Ortiz, J., Guilderson, T., Adkins, J., Sarntheim, M., Baker, L. and Yaruskinsky, M., 2000, Abrupt onset and termination of the African Humid Period:

A review of Holocene palaeoenvironmental research in the Central Sahara

35

Rapid climate response to gradual insolation forcing. Quaternary Science Reviews, 19, pp. 347–361. Elouard, P., 1962, Etude géologique et hydrogéologique des formations sédimentaires du Guebla mauritanien et de la vallée du Sénégal. Mémoires du Bureau de recherches géologiques et minières, Paris, 7, pp. 1–274. Fabre, J. and Petit-Maire, N., 1987, Holocene climatic evolution at 22–23°N from two palaeolakes in the Taoudenni area (Northern Mali). Palaeogeography Palaeoclimatology Palaeoecology, 65, pp. 133–148. Faust, D., Zielhofer, C., Escudero, R.B. and Diaz del Olmo, F., 2004, High-resolution fluvial record of Late Holocene geomorphic change in Northern Tunisia: climatic or human impact? Quaternary Science Reviews, 23, pp. 1757–1775. Flohn, H. and Nicholson, S., 1980, Climatic fluctuations in the arid belt of the “Old World” since the Last Glacial Maximum; Possible causes and future implications. Palaecology of Africa, 12, pp. 3–22. Fontes, J.C. and Gasse, F., 1990, PALHYDAF (Palaeohydrology in Africa) program. Objectives, methods, major results. Palaeogeography Palaeoclimatology Palaeoecology, 84, pp. 191–215. Gasse, F. and van Campo, E., 1994, Abrupt post-glacial climate events in West Asia and North Africa Monsoon domains. Earth Planetary Science Letter, 126, pp. 435–456. Gasse, F., Téhet, R., Durand, A., Gilbert, E. and Fontes, J.C., 1990, The arid-humid transition in the Sahara and the Sahel during the last deglaciation. Nature, 346, pp. 141–146. Gasse, F., Fontes, J.C., Plaziat, J.C., Carbonnel, P., Kaczmarska, P., de Deckker, P., Soulie-Märsche, I. and Callot, Y., 1987, Biological remains, geochemistry and stable isotopes for the reconstruction of environmental and hydrological changes in the Holocene lakes from North Sahara. Palaeogeography Palaeoclimatology Palaeoecology, 60, pp. 1–46. Geyh, M.A., 1983, Physikalische und chemische Datierungsmethoden in der QuartärForschung. Clausthaler Tektonische Hefte, 19, pp. 1–163. Geyh, M.A., 1971, Die Anwendung der 14C-Methode und anderer radiometrischer Datierungsverfahren. Clausthal-Zellerfeld, Verlag Ellen Pilger, pp. 1–118. Geyh, M.A. and Jäkel, D., 1973, Late Glacial and Holocene climatic history of the Sahara desert derived from a statistical assey of 14C dates. Palaeogeography Palaeoclimatology Palaeoecology, 15, pp. 205–208. Grunert, J., Baumhauer, R. and Völkel, J., 1991, Lacustrine sediments and Holocene climates in the Southern Sahara: The example of palaeolakes in the Grand Erg of Bilma (Zoo Baba and Dibella, Eastern Niger). Journal of African Earth Science, 12, 1–2, pp. 133–146. Hillaire-Marcel, C., Riser, C., Rognon, P., Petit-Maire, N., Rosso, J.C. and Soulie-Marche, I., 1982, Radiocarbon chronology of Holocene hydrologic changes in Northeastern Mali. Quaternary Research, 20 (1983), pp. 145–164. Hoelzmann, P., 1992, Palaeoecology of Holocene lacustrine sediments within the West Nubian Basin, SE-Sahara. Würzburger Geographische Arbeiten, 84, pp. 57–71. Hoelzmann, P., Keding, B., Berke, H., Kröpelin, S. and Kruse, H.J., 2000, Environmental change and archaeology: Lake evolution and human occupation in the Eastem Sahara during the Holocene. Palaeogeography Palaeoclimatology Palaeoecology, 169, pp. 193–217. Hoogchiemstra, H., Stalling, H., Agwu, C.O.C. and Dupont, L.M., 1992, Vegetational and climatic changes at the northern fringe of the Sahara 240.000–5.000 years BP: Evidence from 4 marine pollen records located between Portugal and the Canary Islands. Review of Palaeobotany and Palynology, 74, pp. 1–53.

36

Schütt and Krause

Jäkel, D., 1979, Run-Off and fluvial formation processes in the Tibesti Mountains as indicators of climatic history in the Central Sahara during the Late Pleistocene and Holocene. Palaeocology of Africa, 11, pp. 13–36. Kutzbach, J.E., 1979, Estimates of past climate at palaeolake Chad, North Africa, based on a hydrological and energy-balance model. Quaternary Research, 14, pp. 210–223. Kutzbach, J.E. and Ruddiman, W.F., 1993, Model description, external forcing and surface boundary conditions. In: Wright, H.E., J.E. Kutzbach, Webb, III, T., Ruddiman, W.F. Street-Perrott, F.A. and Bartlein, P.J. (Eds.), 1993, Global climates since the Last Glacial Maximum. University of Minnesota Press, Minneapolis, MN, pp. 12–23. Kutzbach, J.E. and Guetter, P.J., 1986, The influence of changing orbital parameters and surface boundary conditions on climate simulations for the past 18.000 years. Journal of the Atmospheric Sciences, 43, 16, pp. 1726–1759. Lamb, H.F., Gasse, F., Benkaddour, A., El Hamouti, N., van der Kaars, S., Perkins, W.T., Pearce, N.J. and Roberts, C.N., 1995, Relation between century-scale Holocene arid intervals in tropical and temperate zones. Nature, 373, pp. 134–137. Lézine, A.M., 1987, Late Quaternary vegetation and climate of the Sahel. Quaternary Research, 32, 2, pp. 317–334. Lézine, A.M. and Casanova, J., 1989, Pollen and hydrological evidence for the interpretation of past climates in tropical West Africa during the Holocene. Quaternary Science Reviews, 8, pp. 45–55. Lézine, A.M., Casanova, J. and Hillaire-Marcel, C., 1990, Across an Early Holocene humid phase in Western Sahara: Pollen and isotope stratigraphy. Geology, 18, pp. 264–265. Maley, J., 1977, Palaeoclimates of Central Sahara during the Early Holocene. Nature, 269, pp. 573–577. Michel, P., 1973, Les bassins des fleuves Sénégal et Gambie. Études géomorphologiques. Mémoires de l´O.R.S.T.O.M. 63, Paris, pp. 1–752. Müller, M.J., 1996, Handbuch ausgewählter Klimastationen der Erde. Forschungsstelle Bodenerosion, 5, Trier. Nicholson, S., 2001, Climatic and environmental change in Africa during the last two centuries. Climate Research, 17, pp. 123–144. Pachur, H.J., Röper, H.-P., Kröpelin, S. and Goschin, M., 1987, Late Quaternary hydrography of the Eastern Sahara. Berliner Geowissenschaftliche Abhandlungen, 75, pp. 331–384. Petit-Maire, N., 1987, Local responses to recent global climatic change: Hyperarid Central Sahara and Coastal Sahara. In: Current research in African earth sciences, edited by Matheis and Schandelmeier, Balkema, Rotterdam, pp. 431–433. Petit-Maire, N., 1988, Taoudenni Basin (Mali), Holocene palaeolimnology and environments. Würzburger Geographische Arbeiten, 69, pp. 45–52. Prell, W.L. and Kutzbach, J.E., 1987, Monsoon variability over the past 150.000 years. Journal of Geophysical Research, 92, pp. 8411–8425. Reichelt, R., Faure, H. and Maley, J., 1992, Die Entwicklung des Klimas im randtropischen Sahara-Sahelbereich während des Jungquartärs- ein Beitrag zur angewandten Klimakunde. Petermanns Geographische Mitteilungen, 136, 2–3, pp. 69–79. Renssen, H., Brovkin, V., Fichefet, T. and Goosse, H., 2006, Simulation of the Holocene climate evolution in Northern Africa: The termination of the African Humid Period. Quaternary International, 150, pp. 95–102. Rognon, P., 1986, Late Quaternary climatic reconstruction for the Maghreb (North Africa). Palaeogeography, Palaeoclimatology, Palaeoecology, 58, pp. 11–34. Rognon, P. and Williams, M.A.J., 1977, Late Quaternary climatic changes in Australia and North Africa: a preliminary interpretation. Palaeogeography Palaeoclimatology Palaeoecology, 21, pp. 285–327.

A review of Holocene palaeoenvironmental research in the Central Sahara

37

Schultz, J., 2005, The ecozones of the world. The ecological divisions of the geosphere. Springer, Berlin, Heidelberg, New York. Schütt, B., 2004a, Reconstruction of Holocene weathering conditions from chemical character of playa-lake-sediments—a case study from Central Spain. Lecture Notes in Earth Science, 102, pp. 5–30, Springer Verlag. Schütt, B., 2004b, Zum holozänen Klimawandel der zentralen Iberischen Halbinsel. Relief, Boden, Paläoklima, 20, Stuttgart. Servant, M., 1973, 1983, Sequences Continentales et Variations Climatiques: Evolution du Bassin du Tchad au Cenozoique Superieur. Travaux et Documents de l´O.R.S.T.O.M., 159, pp. 1–573. Servant, M. and Servant-Vildary, S., 1980, L´environment Quaternaire du Bassin du Tchad. In: Sahara and the Nile, edited by Williams, M.A. and Faure, H., pp. 133–163. Street, F.A. and Grove, A.T., 1976, Environmental and climatic implications of Late Quaternary lake-level fluctuations in Africa. Nature, 261, pp. 385–390. Stuiver, M. and Reimer, P.J., 1989–2005, CALIB 5.0.1 Calib Radiocarbon Calibration Program. Stuiver, M., Reimer, P.J. and Reimer, R., 2005, CALIB 5.0.1 Manual, http://www.calib.qub. ac.uk/crev50/manual/, 09/05/2005. Talbot, M.R., 1980, Holocene changes in tropical wind intensity and rainfall: Evidence from Southeast Ghana. Quaternary Research, 16 (1981), pp. 201–220. Talbot, M.R. and Delibrias, G., 1977, Holocene variations in the level of Lake Bosumtwi, Ghana. Nature, 268, pp. 722–724. Talbot, M.R. and Delibrias, G., 1979, A new Late Pleistocene–Holocene water-level curve for Lake Bosumtwi, Ghana. Earth and Planetary Science Letters, 47 (1980), pp. 336–344. Thorweihe, U., Schneider, M. and Sonntag, C., 1984, New aspects of hydrogeology in Southern Egypt, Berliner Geowissenschaftlichen Abhandlungen, Reihe A/50, pp. 209–216. Völkel, J., 1988, Zum jungquartären Klimawandel im saharischen und sahelischen Ost-Niger aus bodenkundlicher Sicht. Würzburger Geographische Arbeiten, 69, pp. 255–276. Völkel, J. and Grunert, J., 1990, To the problem of dune formation and dune weathering during the Late Pleistocene and Holocene in the Southern Sahara and the Sahel. Z. Geomorph.‚ N.F, 34, pp. 1–17. Wagner, G.A., 1995, Altersbestimmung von jungen Gesteinen und Artefakten. Ferdinand Enke Verlag. Stuttgart, pp. 1–277. Waller, M. and Salzmann, U., 1998, The Holocene vegetational history of the Nigerian Sahel based on multiple pollen profiles. Revue Palaeobotany Palynology, 100, pp. 39–72. Weischet, W. and Endlicher, W., 2000, Regionale Klimatologie. Teubner, Stuttgart, Leipzig. White, F., 1983, The vegetation of Africa. A descriptive memoir to accompany the UNESCO/AETFAT/UNSO vegetation map of Africa. UNESCO, Paris, France. Wright, H.E., Kutzbach, J.E., Webb III, T., Ruddiman, W.F., Street-Perrott, F.A. and Bartlein, P.J. (Eds.), 1993, Global climates since the Last Glacial Maximum. University of Minnesota Press, Minneapolis, MN, pp. 1–569. Zielhofer, C., Faust, D., Escudero, R.B., Diaz del Olmo, F., Kadereit, A., Moldenhauer, K.-M. and Porras, A., 2004, Centennial-scale Late Pleistocene to Mid-Holocene synthetic profile of the Medjerda Valley, Northern Tunisia. The Holocene, 14, 6, pp. 851–861.

CHAPTER 3

Holocene palaeoenvironmental changes in Central Sahara inferred from Seggedim scarp foot depression (NE-Niger) Roland Baumhauer, Jens Brauneck, Barbara Sponholz and Erhard Schulz Department of Geography, Physical Geography, Julius-Maximilians University, Würzburg, Germany Oumarou Faran Maiga and Ibrahim Sani Department of Geography, University Abdou Moumouni, Niamey, Niger Simon Pomel DYMSET/CNRS, University of Bordeaux III, Talence Cedex, France ABSTRACT: Initial research has shown that palaeolimnic sediments and palaeosoils situated in the northeastern parts of Niger are suitable for reconstructing parts of the Quaternary palaeoenvironment of the Central Sahara. Several corings in 1989, 1990 and 2005 in the sandstone depression of Seggedim revealed a composition of high-resolution sections suitable for palaeoenvironmental investigations. Stratigraphical, structural and geochemical investigations as well as the analysis of thin sections allow the characterisation of different environmental conditions. Radiocarbon dates set the beginning of an initial swamp environment at about 10,6 ka cal BP, with an exceptionally stable regime to 6,6 ka cal BP, when a major change in the sedimentation regime of the basin is recorded in the core. By mineralogical and geochemical evidence, a rapid transformation from the sapropel setting to a sequence of freshwater lake stages occurred, the latter merely lasting a few centuries. About 6 ka BP, a transition phase, determined by the accumulation of huge bedrock fragments, separated the pure lake sediments from a following sebkha environment. The subrecent and present environment is characterised by the sedimentation of a continental sebkha, where salt and sand show a rhythmic deposition mode.

3.1 INTRODUCTION As shown in chapter one, there is a general agreement between the findings on Late Quaternary landscape and climate development of the Central Sahara with those obtained for the Southern Sahara at the 10³ years time scale. However, sufficiently humid conditions for the formation of freshwater lakes only seem to have existed from the Late Upper Pleistocene to the Mid-Holocene, becoming successively less pronounced by 6,5 to 5,5 ka BP, with the intercalation of several short arid phases, until hyper-arid conditions came to prevail in the whole region since 5 ka BP at least. The evidence of, first, a continuous increase in aridity and then the rapid and quite early transition to the present hyper-arid conditions disagrees with findings from the Western Sahara, for which numerous studies indicate yet another

40

Roland Baumhauer et al.

wet phase around 4,5 ka BP prior to the shift to hyper-aridity. In contrast to those findings, in the Eastern Sahara again there is evidence of a single Early to Mid-Holocene wet phase only, beginning around 9,5 ka BP, interrupted by several arid phases, and ending by 4 ka BP, thus later than in the South-Central Sahara (see chapter 1). The various groups working in the different regional parts of the Sahara agree that, beyond the general Holocene climate and landscape change, there also existed a zonal palaeoclimatic pattern, together with a latitudinal shift of landscape belts (see chapter 1). At last Kuper and Kröpelin (2006), conclude from their palaeoecological studies that conditions in the Eastern Sahara, during the Early Holocene humidity optimum, were similar to those presently found at 12°N, indicating a northward shift of the Sahelian boundary in East Africa by about 600 km. This is in good agreement with the present evidence from the pattern of Early Holocene pluvial soils in the Southern Central Sahara of East Niger (FelixHenningsen, 2004). Similarly, Baumhauer et al. (2004), from the palynological analysis of a former Seggedim core, conclude that during the Late Early Holocene, at least to 20°N there existed a Saharan savanna vegetation (cf. Schulz et al., this issue) which, by 6,5 ka BP, experienced a change towards Saharan desert conditions. The aridification shown in the pollen spectrum, however, is partly attributed to anthropogenic interference (grazing, fire) in the contact region between Sudanian and Saharan vegetation with the beginning of the Neolithic. This, together with the climatic deterioration, led to the development of the Sahelian savanna vegetation further in the South, the human component possibly accelerating it (Baumhauer et al., 2004; Schulz, 2004). New analyses from a core from the Seggedim region of NE-Niger extend through the Holocene period. Using high-resolution, multi-proxy evidence, we are able to clarify palaeoenvironmental examinations of two cores collected in 1989 and 1990 as well as analyse new aspects of the landscape history in this part of the Central Sahara over the last 10.000 years. 3.2 REGIONAL SETTING The endorheic scarp foot depression of Seggedim lies at the northern margin of the Chad Basin in North-Eastern Niger (20°10' N, 12°47' E), between the Djado-Plateau and modern Lake Chad. It is located in the western scarp foot depression of a cuesta of karstified marine sand- and siltstones of Senonian age (Faure, 1966). The absolute height of the freeface escarpment is 640 m asl, the relative one around 230 m. The depression, about 10 km2 large, is partly taken up by a sebkha environment, with groundwater inflow in the centre and at the eastern margin. On average it is 2 km wide East to West, with a maximum width of 7 km. A smooth aeolian sand layer covers most of the depression around the sebkha. The adjacent plain to the West is mostly a serir, its gravel originating from some conglomerate beds within the sandstone of the cuesta. The hydrological situation of the region is largely unknown. Traces of surface runoff are restricted to parts of the escarpment foot. There is no weather station in the Seggedim area. The nearest one at Bilma, some 150 km to the South, is the only one for all of Saharan Northeastern Niger. In Bilma, the annual precipitation does not exceed 10 mm, with a potential evaporation of about 2.700 mm/yr. The rare rainfalls are very irregular, mostly connected to the interaction of the monsoon and polar front. The present plant cover of the scarp foot depression shows a concentric mosaic of belts dependent on the structure of the sediments. An outer zone bears tussocks of Panicum turgidum on coarse sand and gravel, whereas the inner dune-sand parts have stands of Imperata cylindrica, Desmostachya bipennata and Sporobolus spicatus. The inner part of the depression is colonised by trees such as Acacia raddiana, A. ehrenbergiana, A. nilotica

Holocene palaeoenvironmental changes in Central Sahara

41

Figure 1. Map showing NE-Niger. The point marks the coring location of Seggedim (LANDSAT image).

or Hyphaene thebaica. Phoenix dactylifera grows on dune sands. The sebkha surface itself is plantless, but the salines and waterpoints are surrounded by Tamarix canariensis and Juncus maritima (cf. Boudouresque and Schulz, 1981). 3.3 METHODS The cores sequences described below were taken in 2005, using a modified Cullenberg corer with a core diameter of 63 mm. The core sequences were spilt: one half of the core was used for subsampling, the other half for archiving. High-resolution images were produced with a modified flatbed scanner. The main lithostratigraphical features, Munsell soil colour data and major changes in sedimentation visible with the naked eye were logged. 1 cm thick slices were taken about every 10 cm in consideration of major sediment changes. These subsamples were dried and prepared for analysis in bulk chemistry, mineralogy, palynology and microfossils with the main focus on the non-sebkha sediment (below 5 m depth). The first millimetres of the remaining archive-half were cut out for making thin sections. Bulk chemistry was measured for selected elements in aqua regia digestions using an ICPOES (Inductively Coupled Plasma Optical Emission Spectrometry). For the investigations

42

Roland Baumhauer et al.

of silicon and aluminium dynamics, additional XRF (X-Ray Fluorescence) analyses on fuse tablets were performed on 30 samples. Mineralogy was measured by XRD (X-Ray Diffraction). Total inorganic carbon (TIC) and total carbon (TC) were determined by using a Woesthoff Carmhograph and a carbon-sulphur analyzer. The pH values were determined in CaCl2 solution following ISO 10390. Samples for radiocarbon dating were taken from highly organic parts of the sequence and from pieces of charcoal found in the bottom parts of the core. These dates were calibrated for 95% confidence intervals using the CalPal2007 software (Weninger et al., 2007) with the calibration data set CalPal-2007Hulu. 3.4 RESULTS FROM THE 2005 CORING 3.4.1 Stratigraphic units Seven distinctive units determine the stratigraphy of the core. These units are defined by major colour changes, texture and lamination properties. Sand The most recent stratigraphic unit of the sebkha consists of sand and evaporitic material such as thenardite and syngenite. It extends from 0 to 600 cm depth and shows several dynamic changes of sedimentation. A comprehensive description of the Seggedim sebkha sedimentation can be found in preliminary studies (Baumhauer, 1986; Baumhauer et al., 2004). Clay, unstratified This unit is located at 650 cm to 740 cm depth and shows no visible stratification of any kind. Several distinct colour changes and large rock fragments give evidence of a rather dynamic sedimentation process. The main colours in this section range from 5 YR 5/8 to 10 YR 6/2 and 7,5 Y 4/2. Another homogenous unit can be found at 923 cm to 995 cm depth. Again, some sandstone fragments can be found here and no stratification is visible. The main colour is 7,5 Y 4/1. Inferred from the rock fragments deposited and the homogenous structure, these units are supposed to have originated as turbidites or from mud streams. Clay, layered The main features of these units are the fine laminations that appear in regular and ground-parallel sets of bands but show variations in thickness. One major section is located at 750 cm to 780 cm depth. It consists of finely laminated clays and shows no disturbances of any kind. The main colour is 2,5 GY 4/1. Another set of fine laminations can be found at 1.019 cm to 1.055 cm. The colours range from 7,5 Y 4/1 to 3/1. One particular feature of this segment is the occurrence of fish bones at 1.032 cm that prove the existence of a ecosystem teeming with life. Another section between 1.100 cm and 1.140 cm is comparable, yet with occasional white laminae. A third, very small unit between 1.253 cm and 1.270 cm shows variations in thickness. Few shell fragments were found in all three sections. Clay, disturbed The units referred to as “Clay, disturbed” show laminations that are thinly bedded, heavily disturbed, undulated and partly tilted. Such a major section lies between 1.140 cm and 1.250 cm. The contact between the laminae is blurred, and there are several

Holocene palaeoenvironmental changes in Central Sahara

43

Figure 2. Main properties of the 2005 core (5–15 m depth).

distinct colour changes, from 10 Y 3/1 to 7.5 Y 5/1. The disturbed bedding may be due to bioturbation during deposition. A similar segment occurs between 1.270 and 1.300 cm depth, with several white layers more than 2 cm thick. Two more small segments of disturbed clay appear at 995 to 1.019 cm and 1.055 cm to 1.083 cm depth, the latter with some ferrous lenses. Their layers are also tilted, in contrast to the adjacent segments. Tilting therefore, most have taken place during sedimentation and not coring. Sapropel This is the main unit of the lowest part of the core, from 1.300 cm to 1.470 cm depth, with continuous lamination, interrupted by a few white bands. It consists of a mixture of organic material and clay distinct from the adjacent units. Aside from white bands, its colours range from 7,5 Y 2/2 to 4/3.

44

Roland Baumhauer et al.

Bedrock The small unit referred to as bedrock consists of a debris of mollusc shell fragments, heavily weathered pieces of sandstone and ferrous rocks. It comprises the section from 1.470 cm to 1.500 cm, marking the bottom of the core. Gaps and disturbances During coring, a major gap occurred at 600 cm to 650 cm depth due to the massive changing sediment properties. Several disturbances are spread over the core due to the coring technique, which holds several difficulties such as falling sediment debris inside the borehole. 3.4.2 Mineralogy Quartz (silicon dioxide) and clay minerals (aluminium phyllosilicates), most likely kaolinite, are the essential elements of the mineralogical composition and are ubiquitously present. The amounts of quartz and clay sum up to more than 60% (semi quantitative) in mean throughout the core. Other abounding minerals, in a lesser amount, are pyrite and halite. Carbonates are widely spread throughout the core. They show a very heterogeneous composition of calcite, dolomite, aragonite and low-Mg calcite (in detail: figure 12). This may indicate the simultaneous occurrence of allochthonous and autochthonous carbonates followed by biogenic alteration of the carbonates (cf. ch. 6.2.) The maximum concentration of carbonates can be found at the depths of 764 cm (38%) and 1.280 cm to 1.295 cm (>30%). Evaporites, such as the sulphate minerals glauberite, syngenite and thenardite, show maximum values in the upper part of the core, deposited after the present hyperarid climate had set in. Gypsum occurs in several segments, with a maximum peak in the upper part, but other strong peaks at 11 to 12 m depth. Sylvine (KCl), another evaporitic mineral, has only been deposited in the lowest segment, from 1.300 to about 1.450 cm depth. Traces of muscovite occur at various depths throughout the core. 3.4.3 Geochemistry Performing normative calculations on the ICP-OES results (Boyle, 2002), the ascertained elements add up to about 30% in mean. With the addition of total carbon (TC), the mean value ascends to about 34%, leaving about 60% unaccounted. The essential elements of the remaining part are silicium and aluminium, compared to mineralogical analyses (quartz and clay). The results of additional XRF analyses on fuse tablets confirm the calculations. Figure 3 shows the distribution of silicon and aluminium and derived ratios in detail. SiO2 (silicon dioxide or silica) is the major component with a mean value of 41,89% (n = 30) and a maximum of 66,34% in the sandy unit. Al2O3 (aluminium oxide) shows a maximum value of 25,6% at 1.005 cm, with a mean of 15,79% (n = 30). An increase of silica in lower parts of the core and associated depletion of Al2O3 (assigned to clay) is notable. The mole ratio of Si/Al affirms the result and shows maximum values below 1.300 cm as well as a very constant ratio of about 1,6. The mole ratio of Na + K/Al, also referred to as salinization ratio, shows a pattern similar to Si/Al ratio. Potassium (K) and sodium (Na), both parts of alkali metals, represent major evaporites in the core, halite (NaCl) and sylvine (KCl) as well as thenardite (Na2SO4). The ratio of Fe/Mn is a strong indicator of reducing environments (Davison, 1993) and shows several maxima in the lower parts of the core.

Holocene palaeoenvironmental changes in Central Sahara

45

The weight percentage (wt %) maximum of sulphur amounts 31,52% and can be found at the bottom of the core at 1.496 cm, with a mean assay of 14,84%. The maximum amount of iron is again found in the lowest segment, where some heavily weathered bedrock fragments have been found. The average amount of iron throughout the core is 5,7 wt %. Calcium is spread all over the core but with differing values. Remains of gastropod shells lead to the calcium maximum of 14,74 wt % at 1.483 cm depth, immediately followed by a decline to below 1% for the following 1,5 metres. Striking is the high correlation (0,92) of inorganic carbon (TIC) and calcium throughout the core due to the amount of carbonates, although both values show high coefficients of variation (cv %) from 80 to 100%. 3.4.4 Main lake stages Regarding the macroscopic sedimentological, mineralogical, and chemical properties of the Seggedim core, six main units (Figure 2) have been classified and statistically tested. Figure 4 shows some of the main features of five units in mean percentage with their standard deviations, compared to their total values. Unit I—Initial lake (swamp) phase (1.470–1.300 cm) This sequence is characterised by its high organic carbon content (mean 6,87, n = 17) and by its extremely low inorganic carbon content ( mean 0,04, n = 17). Here, TIC and calcium content are at their lowest (0,4 wt %) compared to the mean value of the core (4,21 wt %). The same applies to strontium, which is highly correlated (0,83) with calcium throughout the whole core. Sulphur content is at its maximum (18,4 wt %). pH-values are very low and range from 2 (1.443 cm) to 4,3 (1.302 cm). Pyrite content is constantly at a high level, as is halite. Carbonates do not occur in this unit. Unit II—Transition phase 1 (1.300–1.140 cm) This unit shows rather unstable conditions. With massively decreasing LOI-values down to 1%, the pH-values increase simultaneously. CaO and TIC show maximum values that correspond to changing carbonate peaks. Gypsum occurs in the mineralogical fraction, as well as several halite peaks. Laminated material only occurs at a small sequence from 1.270 to 1.253 cm. Contrary to the initial lake phase, LOI is low and TIC as well as CaO and carbonates show a maximum peak due to fragments of carbonate shells. In contrast to the other values, FeOOH content remains stable at an average of 5,08% (n = 18): the lowest average value of iron throughout the core. Numerous sediment-colour changes and undulated sediment layers support the assumption of dynamically changing lake conditions. Unit III—Lake phase 1 (1.140–1.020 cm) Compared to underlying units, this sequence appears to be rather stable, although sedimentation conditions must have changed, as appears from the stratigraphic record. FeOOH, K2O, MgO and Na2O show low standard deviations and therefore low coefficients of variation at about 10% (n = 10). Gypsum occurs only in the lower parts in the sequence and vanishes at about 11 m depth, having led to diminishing amounts of CaO and SO3 and a high positive correlation of them. Corresponding to the decrease of gypsum, pyrite shows minimum qualitative values, whereas halite shows maximum reflections. In comparison to other units, carbonates are present throughout. Fish bones at 1.032 cm depth are evidence of life in the lake. Magnesium shows maximum values (2,6 %) with low coefficients of variation of 6,17 %. According to this, carbonate values origin solely in dolomite. TIC and iron show a high negative correlation

46

Roland Baumhauer et al.

Figure 3. SiO2 + Al2O3 distribution, mole ratios.

Figure 4. Five of the main chemical values in mean percentage with associated standard deviations.

(-0,81), so that peaks of carbonate result in decreased iron and sulphur values and vice versa. This interrelation results in a negative correlation of CaO and FeOOH (-0,75). Unit IV—Lake phase II (1.000–750 cm) Sulphur and iron show stable values throughout the sequence, with very low coefficients of variation at about 6% (n = 17). The stable values result in stable signals of pyrite and therefore a constant Fe/S mole ratio at 0,39, very close to pyrites stoichiometric mole ratio

Holocene palaeoenvironmental changes in Central Sahara

47

of 0,5. This difference and the bad correlation of iron and sulphur of 0,59 are evidence of a deficit of iron respectively a surplus of sulphur, as based on the pyrite values. The surplus of sulphur increases constantly towards the top of the sequence at 750 cm. TIC and iron show a high negative correlation (-0,81), so peaks of carbonates result in decreased iron and sulphur values and vice versa. This interrelation results in a negative correlation of CaO and FeOOH (-0,75). Unit V—Transition phase 2 (750–650 cm) Sediments of this period represent the transition from a freshwater lake to a sebkha environment. Just like the other transition phase, this is a very unstable sequence. All chemical values showing high coefficients of variation, multiple distinct colour changes and the presence of large rock fragments of problematic origin (see ch. 6.1). All are evidence of major changes in the environment. No laminations appear throughout this unit, which also shows that there has not been a stable phase of calm and undisturbed sedimentation such as in Unit I. Unit VI—Sebkha (600–0 cm) Only three subsamples of this unit have been studied, and therefore there will be no statistical interpretation. Nevertheless, the stratigraphic summary (Figure 2) shows some of the major traits, such as the high content of quartz and evaporites. SO3 and CaO correspondingly show increasing values, due to the development of sulphates (gypsum, glauberite, thenardite, syngenite). Pyrite and carbonates only appear as traces as do the clay minerals. A comprehensive description of the sebkha sedimentation can be found in the preliminary studies by (Baumhauer, 1986; Baumhauer et al., 2004). 3.5 THE ARCHIVE-FUNCTION OF THE SEGGEDIM RECORD Investigations on the palaeoenvironment of a region largely depend on the nature of the exploited archives. A valuable natural archive is one that, depending on its structure, will yield information on the ecological prerequisites that led to its formation, must have been able to trap and preserve remains of microorganisms or inblown dust. This will allow the reconstruction of the former environment from the life conditions of the respective organisms. Fine layered or laminated lake and sebkha sediments proved their ability in two ways: first by their quiet and regular sedimentation, pointing to a non- or less-disturbed nature of the water body, which may be permanent or periodic (Berglund et al., 1986; Clark et al., 1989; Schulz et al., 2002). The second point is the close cover of different layers ensuring all dust and remains of microorganisms trapped with them, which can thus serve as indicators of the conditions of the respective water body and characteristics of the surrounding landscape. Due to this fine layering, they consequently allow a high resolution of investigation on environmental history of the region. Investigations of this type proved to be necessary in order to understand the deposition type and possible alteration as well as to evaluate the abilitiy of sediments to reflect different types of ecosystems (cf. Mees, 1996, 1998). Here we present new micropetrographical analyses of the Seggedim cores. First results were already presented by Pomel in Baumhauer et al. (2004). As described above the 2005 coring in the depression of Seggedim revealed a composition of several units (cf. figure 2). An initial lake (swamp) environment (unit I) was replaced by a sequence of fine layered lake sediments (lake phases I and II /units II–IV). A transition phase separated the pure lake sediments from a following sebkha environment (unit V). The sub recent and actual

48

Roland Baumhauer et al.

environment is characterised by sedimentation of a continental sebkha (VI), where salt and sand show a rhythmic deposition mode. The initial lake (swamp) phase (unit I) The sediment consists of the remains of swamp plants and fine-grained silica material with some greater sandstone fragments. The material is monophasic deposited under wet or even subaquatic conditions over a period of several hundreds of years (cf. Chapter 6.3) and, to as typical for such swamp sediments, it trapped and well preserved the dust blown in such as pollen grains. It is of special interest that charcoal and ashes are regularly present in large quantities exclusively originating from grasses. These evidences of fires are equally present in the following sequences (units II + III). The lake phase II (unit IV) The sediments of this period are characterised by a regular alternation of two distinct layers of about 200–300 microns thickness, forming a kind of doublet (Figure 5a, b, d). A basic layer consists of diatom frustules and their remains in different states of alteration and silt and clay particles, which are often arranged in cross like patters. Small pieces or flakes of amorphous silica are an important element. Mostly horizontally aligned, they provide a textile structure for it. Charcoal and ashes are regularly present in significant quantities. Very often they form typical ash and charcoal layers in this part of the section. Pyrite as small isolated particles or clusters is regularly present as well. An upper layer is characterised by thin flakes of sesquioxides, mostly iron oxides showing a variable dimension from about 30 to 200 microns. They tend to be very flat reminiscent of parts of jellies coming from bacterial and/or algae films. These flakes have been deposited in a matrix of silt and clay also containing diatoms and their fragment as well as charcoal particles. These two layers characterise in their alternation the sediments of this period. The explanation of this alternating sedimentation however provides some difficulties. In general a deep lake with stable hypolimnion has to be assumed, rarely disturbed by lake water circulation. This would explain both the regularity of sedimentation and thin layering. The different fragmentation of diatoms and their aligned deposition as well as of flakes of amorphous silica combined with the presence of pyrite all suggest anoxic and acid conditions in the hypolimnion. The second type of layers, however, that of iron oxides and jelly particles may not be explained by this way. It calls for an oxygen-rich environment where dissolved iron II oxides could be transformed into iron-III-oxides and then be trapped by algae or bacterial layers (see Hamilton-Taylor and Davison, 1995). A possible explanation may be derived from the model of swamp-ore formation presented by Felix-Henningsen (2004), which in view of the morphology of the Seggedim depression should also apply to the former near shore environment of this palaeolake. Looking to the morphology of the Seggedim depression the model presented for the formation of swamp-ores would give a good explanation. In this shallow and highly insolated water with its fringe of reeds, one has to assume two types of iron-III-oxide formation. First it is caused by emanation of oxygen by the roots of Typha or Phragmites after its transportation via the aerenchyme cells and secondly they are formed in the water itself and trapped by the algae and bacterial jellies. Afterwards these yellies are fragmentated and transferred by the waves of the shallow water body into the centre where they could sink and form a characteristic layer of the lakes bottom together with low quantities of diatoms and charcoals. As the next diatom-, silica- and charcoallayer covers them, they are preserved from reduction in the anoxic and acid milieu.

Holocene palaeoenvironmental changes in Central Sahara

49

Figure 5. Lake sediments from the unit IV showing biphasic sediments (E. Schulz). a. Biphasic sediment containing flakes of iron oxides, silt diatoms and their fragments as well as some charcoals in the upper part, the lower part consisting of silt, charcoal and diatoms (845 cm depth); b. Double layer containing diatoms, silt, charcoal and some pyrite in the lower part and iron oxide flakes, charcoals, diatoms and silt in the upper one (790 cm depth); c. Monophasic turbidites of silt with some charcoals and pyrites (786 cm depth); d. Biphasic lake sediments containing strata of tile-like and cross-like oriented silts and thicker layers with coarse silica grains (780 cm depth).

Uniform and unstructured silt layers containing a few charcoal fragments and pyrites interrupt this alternation of thin layers. They are up to 5–7 cm thick and represent turbidites having slid down the steep slopes to the lake bottom (Figure 5c). There, short time and monophasic sedimentation stands in contrast to the biphasic deposition of the thin layers of diatoms, charcoal and silt. Unlike these, they contain very hardly any particles of inblown dust and thus their archive function is a very limited. Figure 6d shows a more differentiated sequence in this period. Fine layers of totally aligned silt and clay particles are in a tile like deposition. Thicker layers overlie them with

50

Roland Baumhauer et al.

Figure 6. Landscape and sediment conditions during transition phase 1 and lake phase I (unit II–III) (E. Schulz). a. The Emi Bao escarpment and Seggedim lake; b. Shallow lake margin with burning reed; c. Phragmites-Typha reeds, the formation of iron oxide flakes around their stems and roots and their transport into the centre of the lake; d. The formation of the fine layered sediments in the hypolimnion of the lake; e. Turbidite formation in the centre of the lake; f–j. Mode of sedimentation of the fine layers and the turbidite in the hypolimnion of the lake.

some coarse quartz grains in mixed with the diatom remains and charcoals. Some beds again show the pattern of the cross-like layering indicating a quiet sedimentation milieu (cf. Pomel, 2004). The model of the deep lake at the foot of Emi Bao (Figure 6a–j) summarises the formation of this type of line-layered deposits. The slopes are covered by an open savanna of the Acacia-Maerua type. A dense stand of Phragmites-Typha reed fringes the lake (Figure 6a). Here, the iron oxide flakes were originate before being washed to the centre of the lake. Figure 6d–j again summarises the two modes of sedimentation in the deeper part of the lake and the regular formation of the fine layers as an alternation of diatom-siltcharcoal and iron oxide sedimentation occasionally interrupted by turbidites. It present the discussion remains open, whether we have to assume a regular deposition of doublets indicating seasonal differences or whether it was a continuous sedimentation of diatom, charcoals and silt regularly interrupted by the inwash of iron-III-oxide flakes at the end of summer season. Transition from lake to sebkha (unit V) This represents the transition phase from a freshwater lake to a sebkha environment. It is an unstable period. All chemical values show high coefficients of variations. The presence of large rock fragments evoke the question of their origin. Figure 7 shows two different modes of sedimentation. The first one is the continuation of the fine layered deposition

Holocene palaeoenvironmental changes in Central Sahara

51

Figure 7. Sediments from the Transition Phase 2 (unit V) (E. Schulz); a. Transition from lake sediment of diatoms, diatom shell, silt and charcoal as well as pyrites to a silica slurry of clay and silt with pyrite and some charcoal. Apparently, there is no hiatus in between these two sediments. In the upper part the slurries are horizontally arranged containing with pyrites and some iron oxide flakes as well as sharp-angled quartzes (683 cm); b. Irregular slurry deposit of sharp-angled quartzes, silt and clay and some charcoals (677 cm); c. Dense slurry of clay, silt and some sharp angled quartzes and charcoal (629 cm); d. Dense slurry of clay, silt, sharp angled quartzes some charcoal and iron oxides. The silts are cross-like layered indicative of a calm sedimentation environment. Presence of some post sedimentary gypsum crystals (606 cm); e. Irregularly deposited clay slurry with isolated sharp- angled quartzes some iron oxide flakes and charcoal. Apparently, the sediment contains material from eroded soils (585 cm). Dense clay slurry with sharp- angled quartz grains and some iron oxide flakes. Some quartzes already show an alteration around their margins (543 cm).

52

Roland Baumhauer et al.

in a stable hypolimnion as described above. In Figure 7b a definite change takes place. The lower part shows the lake sedimentation. This is replaced by a totally different type of deposit consisting of a silica slurry of clay and silt with some coarse quartz particles and pyrite which will prevail from this time on. Remarkably, diatoms are no longer present, but flakes of iron oxides are still common. This change stands for the change to a shallow water body, being filled with slurry or dust. Initially there was still an open water body most likely a shallow lake or some pools. Further sedimentation still was of the slurry type (Figure 7f, g) with charcoal and iron oxides particles still being present. Coarse grained and sharp-angled quartz particles became more and more characteristic of this sedimentation period. At times deposition took place in a very calm environment as shown by a cross-like arrangement of the silty particles (Figure 7d). In the uppermost part of this section rounded quartz grains as reworked material from the cuesta sand stones were found with surficial weathering indicating a post-sedimentation alteration (Figure 7c). Figure 8 depicts the landscape during this period. The former lake has been reduced to some open pools of open water swamps. The plant cover on the slopes of Emi Bao escarpment has changed to the desert type. In contrast to the preceding period linear wadi vegetation made of trees shrubs and grasses of the Acacia-Panicum type was dominant. The Phragmites-Typha reed belt was still present though less dense than in the previous periods. Thus, more and more sediments could be washed into the centre of the depression. The formation of iron oxides in the swamp and charcoal from burning of the reed appears to be as in the previous period. The sedimentation, however, has become a mixture of sub aquatic and windborn deposits as dust storm have been more and more important (d, e, f). The character of the archive function changed definitely too in this period. The lower part of the sequence still allows and analysis for pollen and diatoms. The change of sedimentation (from the biphasic anoxic to the monophasic oxic) goes along with the end of the preservation of pollen or diatoms mainly due to the permanent oxic nature of the sedimentation milieu. Sebkha (unit VI) The recent and present phases are those of continental sebkha conditions. Figure 9 shows samples taken along a west-east transect through the sebkha, sample c was taken near the coring site. The profiles were drawn from soil cuts in the field and from a short core (c). The transect starts at both and ends at the contact with the dune belt surrounding the sebkha basin (Figure 9a). Common to all sections are the alternating sand layers. There is a pellicular silt cover on top occasionally crusted. Another regular phenomenon is the coarse-grained salt, which in the core taken at the centre core also appears as nests. In the central part of the sebkha (b–d) the salt layers either lie on top of a sand layer or at the base of the next stratum. Flooding of all or parts of the sebkha floor seems to be of extremely rare occurrence. The people of Seggedim could not recall a single one, but there are traces of running water from the Emi Bao escarpment to the centre of the sebkha. Fluctuations of the groundwater table could be observed over the year 2005/06. This points to an extreme continental type of sebkha dynamics, where inundation by runoff or precipitation did not occur or where it was extremely rare. Thus the sedimentation is done with inblown material, which secondly is incrusted by salt crystallisation according to the changing water table. This is in contrast to sebkhas from semiarid areas, where regular or periodic inundations and subsequent short time lacustrine conditions lead to the formation of threefold sequence consisting of in washed material at the base, a layer of clay and algae/ bacteria and finally an upper salt crystallisation (Abichou, 2004; Schulz et al., 2002).

Holocene palaeoenvironmental changes in Central Sahara

53

Figure 8. Landscape and sediment conditions during the transition from freshwater lake to sebkha (E. Schulz). a. The Emi Bao escarpmet and the sebkha with its swamp lake and sebkha sediments; b. The lake margin with burning reed; c. Phragmites-Typha reed with the formation of iron oxide flakes and their transport into the centre of the lake; d–f. Subaquatic-subaeric sedimentation in the sebkha environment.

Figure 9. Subrecent and present sebkha sedimentation in the Seggedim depression (E. Schulz). a. Silt coated alternating sand and salt layers near the southwester margin of the sebkha; b. Silt coated alternating layers of coarse and fine sands with salt emanations; c. Sequence of sand and salt layers near the coring site. It is visible, that the salt emanations are formed in the sand layers either as nests or strata. A pellicular crust with a tendency to crack and to form again covers the sequence; d. Alternating layers of sand and salt in the southwestern part of the sebkha. There is a comparable rhythmicity to the centre (c) with a similar coat and crust formation; e. Scheme of sand-silt layers near the present salines. Inblown/inwashed silt with some fine sand dominates the layers. Some salt nests are visible. The fine layering may be due to changing near surface groundwater horizon.

54

Roland Baumhauer et al.

Figure 9 resumes the conditions of the subrecent and present sebkha sedimentation. Figure 9a shows the sebkha in front of the Emi Bao escarpment. In contrast to the preceding period no vegetation is present in the wadis and runnels, whereas a dune belt around the sebkha bears an oasis vegetation of various trees and the remnants of an outer grass belt (Schulz, 1994). The Seggedim depression is completely filled and shows the four different sediment periods and their respective deposits. Figure 10b, c depict a transect from the outer dune belt to the interior of the sebkha explaining the dust sedimentation and some rare situation of surface run off. Figure 10d explains the formation of the sand layers and their subsequent salt formation due to lateral or vertical groundwater supply. A superficial pellicular crust cracks and due to different humidity it wells up giving space to dust accumulation. Figure 10e–g describe the formation of the sand layers, the salt crystallisation and the final crust formation. Comparable to the sediments of the transition phase these deposits are again monophasic and due to their permanent oxygen influence, they do not preserve any pollen or diatoms (Baumhauer et al., 2004) in contrast to the sebkhas of Northern Africa or the Mediterranean, which by their fine layering represent an excellent archive. Finally, figure 11 summarises the steps of the landscape history from first lake phase (unit III) on. The change from savanna environments in unit III to the desert conditions in the following periods is visible.

Figure 10. Present sedimentation in the Seggedim depression (E. Schulz). a. The Emi Bao Escarpment and the Seggedim sebkha with the swamp lake and sebkha sediments of the cored sequence. Not the dune belt with an oasis vegetation around the sebkha; b/c. Blockdiagrams showing a transect form the dune belt to the sebkha. Main sedimentation is done by blowing sand and dust and by rare inwash from the mountainside; d. Blockdiagram showing the sedimentation in the upper part of the sebkha sequence. Lateral and vertical groundwater influence controls the emanation of salt. The pellicular crust falts up. Cracks are filled with dust and sand; e–f. Blockdiagrams showing the subaeric sedimentation mode of the sebkha including the crust formation.

Holocene palaeoenvironmental changes in Central Sahara

55

Figure 11. The landscape history of Seggedim (after Schulz and Merkt, 1996 modified).

3.6 INTERPRETATION 3.6.1 Conditions of the initial sapropel phase (unit I) The initial lake phase referred to as sapropel, unit I or swamp is different in many ways from the other/later ones (Figure 4). It shows high organic carbon and low pH-values as well as extremely low content of calcium and a total lack of carbonates. The “carbon pump”, which has been described by Dean (1999), is a suitable model for describing the sedimentary conditions during this initial lake phase. The decomposition of a high amount of organic detritus in a hypolimnion will lead to a high oxygen demand and thus

56

Roland Baumhauer et al.

to anoxic conditions. CO2 will be produced in high amounts due to this decomposition, in turn leading to decreasing pH-values. CO2 production and the decreasing pH-value will cause the destruction of carbonate remains. The fine, undisturbed laminae in this segment suggest a lifeless lake bottom sealed by a certain depth of water, as the preservation of the organic content can only have been possible under anoxic conditions. This is supported by the high ratio of Fe/Mn, a strong indicator of reducing environments (Davison, 1993), which shows several maximum peaks in this segment (Figure 3). The increase of SiO2 and the decrease of Al2O3 (clay) in the sapropel stage (Figure 3) originate either in a massive increase in aeolian deposited sand (quartz) or in high populations of diatoms. The texture characteristics of the sediment indicate the latter, but ongoing investigations on the grain size distribution and microfossils will show what causes this effect. The maximum values of the salinization ratio originate in the unique occurrence of KCl in this segment and the characteristically constant content of halite, which supports the assumption of a halocline. As no other evaporitic minerals appear at this depth, sylvine and halite may have originated by solution from underlying sandstone. It seems likely that this initial and very stable lake phase was enabled by groundwater discharge from a salt-containing marine aquifer within the Emi Bao formation (Baumhauer, 1986). The saltworks not far from the coring site and freshwater wells right next to it indicate how complicated—and yet not understood—the hydrogeology in this depression is. There is as yet no way to estimate the share of inflowing salty groundwater during the initial lake phase. 3.6.2 Carbonates—distribution and origin Figure 12 shows the heterogeneous distribution of carbonates throughout the core. So far it has been assumed that there are no carbonatic rocks in the Seggedim area (Faure, 1966). Consequently, and due to the different conditions under which aragonite and dolomite would have formed, their presence in the sequence may only be explained by aeolian input. The Mg/Ca mole ratio is an approved indicator of changes in carbonate distribution. It shows two large peaks in the deeper parts of the core, but only the one at 12,50 m coincides with aragonite or low-Mg-calcite. Due to the large amounts of calcite in the upper parts of the core, the ratio remains low. The mole ratio of Sr/Ca is commonly used as an indicator of palaeosalinity (Gasse et al., 1987; Goschin, 1988; Hoelzmann et al., 2000). The authors state that a Sr/Ca mole ratio above 0,003 points is indicative of increased salinity. This threshold value is not exceeded throughout the core, thus freshwater conditions can be assumed. Ongoing investigations on microfossils may confirm the statement. Comparing the amounts of total inorganic carbon and calcium, the values are quite constant. Only in the sapropel stage, where the calcium content has decreased due to the low pH, the mole ratio by far exceeds the ideal one of 1. Other fluctuations will be due to changing calcium content throughout the core (Figure 2). As to the origin of carbonates, several fragments from transition phase II create some problems. The largest one is a massive piece, about 7 cm long, of dark dolomitic carbonate rock, coated by brighter solid crust of most likely biogenic origin. This is the first evidence of any carbonate rock in this area, calling for further investigations. Faure (1966), who studied the geology of Emi Bao and Seggedim formations, did not report any carbonate strata from the catchment. A second piece from the core consists of calcite and aragonite with a calcium content of nearly 30%, and thus a CaCO3 content of more than 50%. This material seems to be of biogenic origin, like the coating of the larger piece. The unusual carbonate composition may best be explained by the transformation of allochthonous carbonate from

Holocene palaeoenvironmental changes in Central Sahara

57

Figure 12. Distribution of carbonates and mole ratios of determining chemical features.

the catchment area, probably dolomite, to aragonite by biogenic activity and the regular weathering of primary carbonates to calcite. This is supported by the simultaneous presence of low-Mg-dolomite and calcite (and aragonite as well at about 12,50 m). 3.6.3 Age-Depth-Model Figure 13 shows the distribution of calibrated AMS radiocarbon (14C) dates compared to two conventional uncalibrated dates from the previous coring. 21 radiocarbon dates were analyzed in the Poznan Radiocarbon Laboratory with accelerator mass spectrometry (AMS). The amount of dates was necessary to provide evidence of the complex sedimentation history of the record. For 17 samples radiocarbon dating was performed on bulk organic matter and revealed a rather chaotic pattern of the upper dates, interpreted as repeated contamination by translocated organic material. To receive dates from other sources than organic material, radiocarbon was additionally measured on carbonates. The mineralogical composition revealed singular signals of aragonite at the beginning and the end of the lake phase which are interpreted as remnants of gastropod shells. Detailed studies on gastropods revealed the reliability of various gastropod taxa for radiocarbon data (Fritz and Poplawski, 1974) especially those terrestrial snails incorporating aragonite in their shells (Brennan and Quade, 1997; Pigati et al., 2004; Zhou et al., 1999). When analysing biogenic carbonates by radiocarbon methods, shifts of dates have to be considered, either caused by the contamination of carbon containing groundwater (hardwater effect) or by carbonate bedrock (carbon reservoir effect) (Fontes and Gasse, 1989; Snyder et al., 1994; Zhou et al., 1999). The results show that no nameable hardwater effect contaminated the carbonates. Therefore, it can be assumed that

58

Roland Baumhauer et al.

Figure 13. Calibrated radiocarbon dates of the 2005 coring compared to the two dates from 1984 (squares).

adjacent radiocarbon dates from organic material can be considered to show the true age of the sediment. The lowermost dated sample was taken from the bottom of the core, close to the bedrock. It places the onset of continuous sedimentation at about 10.410 a cal BP ± 180. Radiocarbon dates from within the first stage set the beginnings of the initial lake phase at about 10.050 a cal BP ± 200. Several dates from higher up are evidence of a very low accumulation rate in unit I. Only about 0,6 mm/a were deposited for over 3.000 years, at a very constant rate. Probably this phase has to be extended to even more than 4.000 years, if the date at 12,61 m depth is included (6,6 ka cal BP). This may be justified by the fact that the accumulation rate of approximately 0,6 mm/a remained the same, even though there had been a change in sedimentary conditions (cf. figure 2). The low accumulation rate is evidence of a very low minerogenic input, which might be interpreted by very low surface wash in the catchment for several millennia, due to the savanna-type vegetation cover inferred from the pollen spectra. The high organic content of the sediment suggests a supporting argument, namely that a closed belt of reed vegetation fringing the lake, as discussed above (Chapter 5), filtered the surface wash from the adjacent slopes. This could have led to accumulation of sediment and development of sediment banks. The following sequences (boxed) show rather confusing variations. Starting at 7.080 a cal BP ± 140 at 1.184 cm depth, all the radiocarbon dates lie inside a small range from 7.400 a cal BP to 6.250 a cal BP and, unlike those from the lower sequence, show no linear distribution. It may be assumed that these variations originate from translocated sediment, originally deposited higher up on the sloping lake bottom during the high-organic sapropel stage. With increasing surface wash and increased sediment input, turbidite movements are likely to have been triggered.

Holocene palaeoenvironmental changes in Central Sahara

59

The two dates from 6 and 8 meters depth (squares) are uncalibrated conventional radiocarbon dates from a preliminary coring in 1984 by Baumhauer (1986). In a preliminary study of another core from Seggedim (Baumhauer et al., 2004), thin-section analysis revealed the presence of large amounts of charcoal. Such fragments can be preserved unaltered in the sediment, whereas other freshly deposited organic matter will be decomposed almost immediately under oxic conditions. Most likely, such surviving charcoal flakes from burnt grass are the source of improper dates from 1984 and 2005 coring. In addition, the time of frequently discussed fires can be localized in a relatively short period from 7.500 to 6.200 a cal BP (boxed dates), but can be extended to the beginning of Initial Phase due to charcoal presence at the lowermost parts of the core. In conclusion, the five dates from the lowermost part of the core appear to be reliable. The majority of those from the upper section, as well as the earlier ones published (Baumhauer, 1986; Baumhauer et al., 2004), seem to be the result of mixing by translocation of sediments of various ages. Additional radiocarbon dating of biogenic carbonates set up a framework for validating the results obtained from bulk organic matter. 3.7 CONCLUSION The coring of 2005 revealed some new evidence and helped to elaborate on some known facts concerning the landscape history of Seggedim. An initial swamp environment came into existence in the Early Holocene and existed for approximately 4.000 years. No evidence was found of the formerly assumed Pleistocene lake stages. A high organic content, with organic fragments partially preserved due to extremely reducing conditions, and very low accumulation rates within the sediment are characteristic of the sapropel/swamp environment. The botanical record reveals the presence of Saharan savanna vegetation in the area, which may have been the main cause of diminished surface wash, but, in addition, a belt of reed vegetation may additionally have filtered any surface wash. Radiocarbon dates set the beginning of the stage at about 10,6 ka cal BP, with an exceptionally stable regime to 6,6 ka cal BP (at 12,6 m depth), when a major change in the sedimentation regime of the basin is recorded in the core. By mineralogical and geochemical evidence, there was a transition from the stable sapropel stage to a sequence of freshwater lake stages showing huge variations in all attributes investigated (cf. figure 2 and figure 4). The distinct increase in Al2O3 (Figure 3) indicates an ascending detritus input due to enhanced surface wash or a reduced belt of reed vegetation. Variations in the thickness and appearance of layers confirm variations in sedimentation most likely reflecting changes of precipitation. The occasional occurrence of gypsum suggests phases of desiccation. The discrepancies in the sequence of radiocarbon dates may best be explained by underwater sliding processes within the deposits. Variations among the freshwater lake stages may reflect variations in amount and seasonality of precipitation. The duration of lake stages has not yet been determined because of irregularities in the sequence of radiocarbon dates. From an earlier coring, the transition from freshwater to sebkha conditions had been assumed to lie further back in time. The new and more reliable dating, however, now indicates a somewhat longer existence of the Early Holocene freshwater lake(s), to at least about 5.310 ± 40 B.P. (uncalibrated, 728 cm depth; corr. 6.098 cal a B.P.), as compared to 6.850 ± 345 B.P (600 cm depth; 1984 bulk date). Additional radiocarbon ages from biogenic carbonates validated some of the dating results obtained from bulk organic matter and set up a framework for the reconstruction of palaeoenvironmental conditions within the Seggedim basin (Figure 14). Due to the lack of dateable material in the upper core section the termination of the lake stage and the onset of the subsequent sebkha stage cannot be determined precisely but

60

Roland Baumhauer et al.

Figure 14. Palaeoenvironmental succession scheme of the Seggedim basin.

can be narrowed to a period around 6 ka BP. Ongoing investigations will hopefully permit to determine more precisely the transition to the following sebkha stage, likely to indicate the onset of the present hyperarid phase. By isolating any aquifer impact, the results obtained from the core should then be linked with those from terrestrial and lacustrine sediments from outside the depression, situated a few hundred kilometres further to the North. Schütt and Krause (see chapter 2) have reviewed the literature on palaeoenvironmental research in the central Sahara. Several authors assume humid conditions from about 11 to 6,5 ka cal BP, with the maximum intensity of the monsoon activity around 10,6 ka cal BP. This is exactly when sedimentation appears to have started in the Seggedim core. Whether this was due to an increase in local precipitation or the rise of a replenished aquifer is still open to discussion, as the role of groundwater recharge in the hydrological balance of the depression is still unknown. The results of the radiocarbon dating show that no noticeable hardwater effect contaminated the carbonates and the adjacent radiocarbon dates from the organic material can be considered to show the true age of the sediment. However, as the aragonite is considered to be of terrestrial origin (gastropods), conclusions concerning groundwater charge during the initial stage cannot be drawn. The mineralogical and geochemical study of rock fragments of the core revealed the presence of carbonatic rocks in the Seggedim area. So far, the nearest known carbonatic outcrops are those of the Carboniferous Dembaba limestone of the Djado basin, at least 100 km northwest and at Ezerza North of Achegour in the Southwest. Evidence of frequent fires raises the question as to their causes. Despite the irregularities in the radiocarbon record, the dates from the translocated and redeposited charcoal pieces can be narrowed down to a tight time slice of 1.200 cal years, during which the regular burnings should have occurred. Two models are discussed for the causes of the fires: natural fires caused by lightning (Dolidon, 2005; Sponholz, 2004), versus intentional seasonal burning of the belt of reeds, which is still a regular procedure among traditional Sahelian herdsmen for providing access to the water for their animals (cf. Schulz, 2004, Figure 2c, d; Schulz and Merkt, 1996). ACKNOWLEDGEMENTS We would like to thank the German Research Foundation (DFG) for its financial support of the ‘Limnosahara’ research project (Ba 1000/21-1,2), funded since 2005. REFERENCES Abichou, A., 2002, Les changements des paysages du basin versant de l´oued Tataouine Fessi (Sudest Tunisie). Etude multiscalaire et micromorphologie des remplisages des sebkhas et études des états des surfaces. These Université Michel de Montaigne, Bordeaux, pp. 1–1207.

Holocene palaeoenvironmental changes in Central Sahara

61

Baumhauer, R., 1986, Zur jungquartären Seenentwicklung im Bereich der Stufe von Bilma (NE-Niger). Upper Pleistocene and Holocene evolution of lakes in the Bilma region, Northeastern Niger. Würzburger Geographische Arbeiten, 65, pp. 1–235. Baumhauer, R., Schulz, E. and Pomel, S., 2004, Environmental changes in the Central Sahara during the Holocene. The Mid-Holocene transition from freshwater lake into sebkha in the Seggedim depression, NE-Niger. In: Smykatz-Kloss, W., Felix-Henningsen, R. (Eds.), Palaeoecology of quaternary drylands. Lecture notes in earth sciences, 102, Springer, Berlin, pp. 31–45. Berglund, B.E. and Ralka-Jasiewiczowa, M. (1986), Handbook of Holocene Palaeoecology and Palaeohydrology. John Wiley and Sons, Chicester. Boudouresque, E. and Schulz, E., 1981, The flora and vegetation of NE-Niger (Djado, Kaouar and Ténéré). Willdenovia, 11, pp. 363–394. Boyle, J., 2002, Inorganic Geochemical Methods in Palaeolimnology. Tracking Environmental Change Using Lake Sediments, pp. 83–141. Brennan, R. and Quade, J., 1997, Reliable Late-Pleistocene stratigraphic ages and shorter groundwater travel times from 14C in fossil snails from the Southern Great Basin. Quaternary Research, 47, 3, pp. 329–336. Clark, J.S., Merkt, J. and Mueller, H., 1989, Post-glacial fire, vegetation and human history on the Alpine forelands, Southwestern Germany. Journal of Ecology, 77, pp. 997–925 Davison, W., 1993, Iron and manganese in lakes. Earth-Science Reviews, 34, pp. 119–163 Dean, W.E., 1999, The carbon cycle and biogeochemical dynamics in lake sediments. Journal of Palaeolimnology, 21, pp. 375–393. Dolidon, H., 2005, L´espace des feux en Afrique de l´Ouest. Thèse de l´Université de Caen, pp. 1–345. Faure, H., 1966, Reconnaissance geologique des formations sedimentaires post-paléozoiques du Niger Oriental. Direction des Mines et de la Géologie, Paris, pp. 1–587. Felix-Henningsen, P., 2004, Genesis and palaeo-ecological interpretation of swamp ore deposits at Sahara palaeo-lakes of East Niger. In: Smykatz-Kloss, W., Felix-Henningsen, R. (Eds.), Palaeoecology of Quaternary drylands. Lecture notes in earth sciences, 102, Springer, Berlin, pp. 47–72. Fontes, J.C. and Gasse, F., 1989, On the ages of humid Holocene and Late Pleistocene phases in North Africa—remarks on “Late Quaternary climatic reconstruction for the Maghreb (North Africa)” by P. Rognon. Palaeogeography, Palaeoclimatology, Palaeoecology, 70, 4, pp. 393–398. Fritz, P. and Poplawski, S., 1974, 18O and 13C in the shells of freshwater molluscs and their environments. Earth and Planetary Science Letters, 24, 1, pp. 91–98. Gasse, F., Fontes, J.C., Plaziat, J.C., Carbonel, P., Kaczmarska, I., de Deckker, P., SoulieMarsche, I., Callot, Y. and Dupeuble, P.A., 1987, Biological remains, geochemistry and stable isotopes for the reconstruction of environmental and hydrological changes in the Holocene lakes from North Sahara. Palaeogeography, Palaeoclimatology, Palaeoecology, 60, pp. 1–46. Goschin, M., 1988, El Atrun (Nubien)—Ein frühholozänzeitlicher See. Freie Univ. Berlin, pp. 1–221. Hamilton-Taylor, J. and Davison, W., 1995, Redox-driven cycling of trace elements in lakes. In: Lerman, A., Imboden, D. (Eds.), Physics and chemistry of lakes. Springer, Berlin, pp. 217–263. Hoelzmann, P., Kruse, H.J. and Rottinger F., 2000, Precipitation estimates for the Eastern Saharan palaeomonsoon based on a water balance model of the West Nubian Palaeolake Basin. Global and Planetary Change, 26, pp. 105–120. Kuper, R. and Kroepelin, S., 2006, Climate-controlled Holocene occupation in the Sahara: Motor of Africa’s evolution. Science, 313, pp. 803–807.

62

Roland Baumhauer et al.

Mees, F., 1996, Petrological studies of saline deposits of a perennial saline lake and a dry lake basis, and of calcareous deposits of small freshwater basin. Ph. D. thesis, University of Ghent, pp. 1–318. Mees, F., 1998, The alteration of glauberite in lacustrine depositsd of the Toudenni-Agargott basin, Northern Mali. Sedimentary Geology, 117, pp. 193–205. Pigati, J.S., Quade, J., Shahanan, T.M. and Haynes, C.V., 2004, Radiocarbon dating of minute gastropods and new constraints on the timing of Late Quaternary springdischarge deposits in Southern Arizona, USA. Palaeogeography, Palaeoclimatology, Palaeoecology, 204, 1–2, pp. 33–45. Schulz, E., 1994, The southern limit of the Mediterranean vegetation in the Sahara during the Holocene. Hist. Biol., 9, pp. 137–156. Schulz, E., 2004, Landscape and diseases. The Middle Holocene Sahara: A development of cultural landscape as reusult of a risk oriented behaviour. Riscure si catastrofe, 1/2, pp. 1–24. Schulz, E. and Merkt, J., 1996, Transsahara. Die Überwindung der Wüste. Würzburger Geographische Manuskripte, 38, pp. 1–116. Schulz, E., Abichou, A., Hachicha, T., Pomel, S., Salzmann, U. and Zouari, K., 2002, Sebkhas as ecological archives and the vegetation and landscape history of Southeastern Tunisia during the last two millenia. Journal of African Earth Sciences, 34, pp. 223–229. Snyder, J.A., Miller, G.H., Werner, A., Jull, A.J.T. and Stafford, T.W., 1994, AMSradiocarbon dating of organic-poor lake sediment, an example from Linnevatnet, Spitsbergen, Svalbard. The Holocene, 4, 4, pp. 413–421. Sponholz, B., 2004, Fulgurites as paleoclimatic indicators- the proof of fulgurite fragments in sand samples. In: Smykatz-Kloss, W., Felix-Henningsen, R. (Eds.), Palaeoecology of quaternary drylands. Lecture notes in earth sciences, 102, Springer, Berlin, pp. 73–78. Weninger, B., Jöris, O. and Danzeglocke, U., 2007, CalPal-2007. Cologne Radiocarbon Calibration and Palaeoclimate Research Package. http://www.calpal.de/, accessed 200711-14. (calibration data set is CalPal-2007Hulu ). Weninger, B. and Jöris, O., 2007 (in press), Towards an absolute chronology at the Middle to Upper Palaeolithic transition in Western Eurasia: a new GreenlandHulu time-scale based on U/Th ages. Journal of Human Evolution. Zhou, W., Head, M., Wang, F., Donahue, D. and Jull, A., 1999, The reliability of AMS radiocarbon dating of shells from China. Radiocarbon, 41, 1, pp. 17–24.

CHAPTER 4

The desert in the Sahara. Transitions and boundaries Erhard Schulz Department of Geography, Physical Geography, Julius-Maximilians University, Würzburg, Germany Abdelhakim Abichou Département de Géographie, Université de Tunis, Tunisia Aboubacar Adamou Département de Géographie, Université Abdou Moumouni de Niamey, Niger Aziz Ballouche and Issa Ousseïni Département de Géographie, Université d’Angers, France ABSTRACT: The Sahara is a system of three main landscape types. It comprises the semidesert in its northern part, the desert in the centre and the savanna in the South. The semidesert is a sparse but diffuse shrub vegetation, which is fed by winter rainfall and contains several Mediterranean affiliations in its floristic composition. A sharp boundary separates it from the desert, which is to define by the dominance of linear or contracted permanent vegetation. Aleatoric resources such as rainfall are exploited by achabs, short time vegetation plots of therophytes, which may cover large surfaces. A savanna covers the southern part of the Sahara. It is also separated from the desert by a sharp boundary. This savanna is alimented by monsoon summer rain. The floristic composition remains Saharan. The boundaries of the desert are climatic dependent because the annual rainfall is no longer sufficient enough to allow a diffuse plant cover. They can be followed over the whole continent. Even in the climatically favoured region of the Oceanic Sahara in Mauritania these boundaries are clearly identifiable. In general the vegetation shows two different strategies to cope with the severe physical conditions of the Sahara. The first is to equip the individual organisms with protections against drought and frost but also to adjust their number to the amount of water available. The second is to exploit aleatoric resources by short time organisms in a great number without special protection mechanisms. A great number of pioneer elements in the vegetation describes its general vulnerability but also a great resilience. Especially the repeated development of extended achabs is a sign of rapid reaction to slight modifications of the rainfall conditions without changing the general ecosystem. A vegetation map is presented as a modern model for the interpretation of former environments in the Sahara and adjacent areas.

64

Erhard Schulz et al.

4.1 INTRODUCTION The Arabic term “Sahara” means desert or steppe (Wehr, 1968) and describes a waste land being permanently poor in water and hostile for human existence. These estimations are easily to understand if one realises the limited number of water points along the main caravan routes of the Sahara regardless if they crossed the diffuse pastures of the North or the dune fields of the Central Sahara. The use of the term “desert” is wide. The only common feature in this variety of landscapes is that they are arid regions and poorly suitable for human existence. The permanent lack of water demands a highly elaborated organisation of life. Finally, both “Sahara” and “desert” have a more or less emotional dimension. Difficulties arise in that moment when one tries to transfer them into scientific terms. Both terms are largely employed but there is hardly any consensus neither on the nature of Sahara nor on its limits in space. Moreover the use of the terms “steppe”, “semidesert” and “savanna” is often contradictory if one compares the descriptions of Barry et al. (1976), Le Houerou (1989) or White (1983). Anyway, it is necessary to find common criteria in order to differentiate the waste areas especially for the monotony of plant cover (Barry et al., 1976). Regarding the wide discussion on an expanding Sahara or desert (cf. Stebbing, 1935, Tucker et al., 1991) one should choose phenomena which in principal could change in order to decide, whether there might be an extension of an ecosystem on the cost of another. Several attempts were made to define these regions. Describing the limits of ecosystems is a necessary way to understand their internal nature to define indicators of past changes in landscape to be read from various archives. 4.2 SAHARA AND DESERT. ATTEMPTS TO DIFFERENTIATE THE UNKNOWN AREAS The first attempt is to look for visible “real” limits in the landscape. This is the way voyagers proceeded along the main commercial routes during the last and the first half of this century (Denham, 1822–25; Barth, 1857–1858; Nachtigal, 1879; Vischer, 1910). They reported clear boundaries in the landscapes of the Southern Sahara, but in general these indications were not accepted later on to define or limit these regions of the desert itself. These landscape descriptions as well as those from the Western Sahara (Monod, 1954) and from Southwestern Africa (Walter, 1964) had one thing in common. They reported an important change within short distances either from forest to steppe or woodland to grassland. They also coincided in describing these changes in the same regions during the last century where they can be recognized still today (Schulz and Hagedorn, 1994). Monod (1954) and Walter (1964) described the general transition from outside into the desert in explaining it by the gradual diminishing of annual precipitation to a threshold, where the basic needs of plants could only be satisfied by a combination of runoff, groundwater and precipitation in wadis or depressions. These descriptions provided a suitable tool to classify the immense complex of Sahara with help of criteria, which easily could be followed and detected in the landscape. In this sense the desert is a region and an ecosystem where permanent life is only possible under special conditions e.g. the combination of run off, precipitation and of groundwater in depressions and wadis. This is valid both for vegetation and wildlife. It also has the advantage to deal with permanently present phenomena and not only with the areas of certain organisms Uromastix acanthinurus the Bell’s dab lizart, Cerastes cerastes, the Horned viper or the Dorcas gazelle Gazella dorcas (Le Berre, 1989, 1990) for the whole Sahara or the change from the Had (Cornulaca monacantha) to the Cram Cram (Cenchrus biflorus) for the southern limit of the Sahara (Capot-Rey, 1953).

The desert in the Sahara. Transitions and boundaries

65

In a second way one tries to impose an arrangement from predefined limits and their values. These were as well isohyets, aridity indices or percentages in the floristic composition (Klaus and Frankenberg, 1980). From the time Dubief (1959, 1963) presented the isohyetmaps of Northern Africa these maps were taken as suitable tools to define landscapes. There were parallels between landscape/vegetation boundaries and isohyets and inverse and consequently to limit the Sahara and desert at once by the isohyets of 100, 50 mm in the North and 150 mm in the South (Giessner, 1988; Hagedorn, 1985). However, the restricted number of meteorological stations and the great variability of rainfall set a principal limit for the use of these definitions. Anyway, a general consensus on the climatic conditions of the Sahara comprises: (1) The general diminishing of the annual amount of precipitation from the external regions to the centre. (2) The characterisation of the Northern Sahara by the Atlantic–Mediterranean cyclones and of the Southern Sahara by the monsoon and their general interaction in the Central Sahara. (3) The increasing irregularity of rainfall and the rising part of aleatoric precipitation towards the centre. Especially the last point makes it very difficult to use isohyets for a limitation of the desert and the Sahara (Barry et al., 1972, 1973). Anyway, the variability of precipitation and consequently of the annual isohyets often lead to the ideas of moving frontiers of the desert still lasting into most recent times (Tucker et al., 1991; Reichelt et al., 1992). In the end one took the most variable and fluctuating phenomenon to limit the main landscape units. If one compares these two ways to describe and to limit the landscape units, the characterisation of the desert by the repartition of vegetation has several advantages. It frees from the emotional use of the term “desert” as a barren hostile land but also from the exclusive connection to variable criteria as the amount of precipitation or the presence of single organisms. The vegetation summarises both climatic and soil conditions and also provides information of the rainfall regime in regions where records are not available. Moreover it represents phenomena, which could also be detected in the remnants of former landscapes as sediments or palaeosoils. In that way the vegetation provides chances to reconstruct the past and compare it to the present based on the same material. 4.3 THE LANDSCAPE UNITS IN A NORTH–SOUTH TRANSECT THROUGH THE SAHARA In a North–South-transect we will follow the landscape formations and their changes in the Sahara in order to detect, whether and where there are boundaries between ecosystems and in which way they can be recognized. A vegetation map of Northern and Western Africa (Figure 1) was conceived and compiled as a modern model to be compared with the Holocene vegetation. It refers to former regional maps (Pomel et al., 1994; Schulz and Lueke, 1994) and tries to correct the too simplistic model given by White (1983)—at least for the Saharan part. It combines own investigations in the central part of this region and a compilation of the recent vegetation mapping projects. An extensive bibliography is given in Schulz et al. (2000). A special attention is given to the dynamic of the southern part of the Sahara, which is widely discussed as an indicator of “Desertification” (see Tucker et al., 1991). The different formations incorporate remnants of formerly denser vegetation types and the whole spectra of the derivations and transformations by man. In this way it does not deal with “natural”, “potential-natural” or “near to nature” formations or their reconstructions (cf. Anhuf and Frankenberg, 1991). This concept follows the way already indicated by Trochain (1940), who pointed out that most of the vegetation types of these regions could only be characterised as “pseudo climaxes” (depending on soil- or groundwater types)

66

Erhard Schulz et al.

Figure 1. Schematic vegetation map of Northern and Western Africa (from Schulz 2000, modified) (for colour map and legend see pages xiv–xv in the preliminary section).

The desert in the Sahara. Transitions and boundaries

67

68

Erhard Schulz et al.

or even more “peni climax formations” being created by man. Difficulties based on differences of regional vegetation mapping projects are evident in the border region of the former English or French dominions. The limits of the Sudanian region differ in Cameroon (Letouzey, 1985) from those in Nigeria (Barbour et al., 1982; Keay, 1959). The map we present in this article refers to the system of Letouzey (1985) because it seemed to be more coherent and also comparable to own observations in Cameroon and Togo. Moreover, Ballouche and Dolidon (2000) and Dolidon (2005) could confirm the northern limit of the Sudanian zone by its frequency of annual fires. In addition there is a series of block diagrams (Figure 2), which may explain the physiognomy of the main vegetation units including their altitudinal zonation. To understand the Saharan system it is necessary to know about the adjacent areas too. Thus, the Mediterranean not only shows a sharp borderline to the Sahara it also interfingers with it. The southern part of the Mediterranean is characterised by the transition of forest and shrub communities to those of open grasslands (the steppe units). These plant communities are steppe units in the original sense as winter cold continental tussock grass communities of the genus Stipa (cf. Walter, 1964). Floristically and also climatically they still range into the Mediterranean system. A. The Mediterranean macchia and oak forest (from the coast to the Ain Draham Mts./Northern Tunisia Quercus faginea, Qu. suber, or the macchia units of Qu. coccifera, Qu.ilex, Pistacia therebinthus or Olea europea). B. The Djebel Chaambi near Kasserine (Central Tunisia) represents the Stipa steppe and the transition to the Pinus halepensis, Rosmarinus communis, Juniperus oxycedrus forests and on the top (about 1.500 m) the southernmost stands of Quercus ilex. 4.3.1 Northern Sahara and the semidesert. Two strategies of life Parallel to the so-called “Sahara line” south of the Atlas chain from Southwestern Morocco to Southern Tunisia the shrub communities of the Gymnocarpus-Rhanterium-Atractylisunits—rarely exceeding a soil cover of 30%—replace the tussock grass communities of the Stipa-steppe. This transition represents the change to the semidesert (Bornkamm und Kehl, 1985). The transition is characterised by two different physiognomic vegetation units but also by a change in the floristic composition. Whereas the Stipa tenacissima-units are characterised by Mediterranean floristic elements, the semidesert is dominated by Saharan elements. But we have to admit that both vegetation types include units characterised by Artemisia: A. campestris in the steppe units and by A. campestris and A. herba-alba in the semidesert. Stokker (1962, 1976) explained the characteristic composition of the semidesert as a treeless shrub vegetation type caused by the double stress against drought and frost, which impedes a noteworthy tree growth. Stokker also stated that the tussock grass Stipagrostis pungens physiologically behaves as a small shrub in colonising dune sands. However, the questions of a former content of trees evoked by Le Houerou (1969) remain open and may be solved only by records of vegetation history covering the complete Holocene. Perhaps one time this will explain the presence of Acacia raddiana-tree communities in the northern part of the semidesert area. Some principal problems limit the understanding of this borderline. The first is a simple case of nomenclature since the schools around Quezel (1965), Barry (Barry et al., 1973, 1974, 1981, 1985, 1986) and Gaussen and Vernet (1958) regroup these two

The desert in the Sahara. Transitions and boundaries

69

Figure 2. Block diagrams showing the main vegetation types along a North–South transect between the Mediterranean and the Northern Sudan. 1. Pinus pinea, Pinus halepensis, 2. Juniperus phoenicea, 3. Quercus ilex, 4. Quercus faginea, 5. Quercus suber, 6. Quercus cocciferea, 7. Pistacia lentiscus, P. terebinthus, 8. Olea europaea, 9. Phillyrea europaea, 10. Acacia ehrenbergiana, 11. Acacia raddiana, 12. Acacia nilotica, 13. Acacia laeta, 14. Aacia albida, 15. Maerua crassifolia, 16. Balanites aegyptiaca, 17. Salvadora persica, 18. Commiphora africana, 19. Hyphaene thebaica, 20. Guiera sengalensis, Combretum micranthum, C. nigricans, 21. Adansonia digitata, 22. Khaya senegalensis, 23. Parkia biglobosa, 24. Annona senegalensis, 25. Stipagrostis pungens, 26. Panicum turgidum, 27. Artemisia herba alba, 28. Salsola baryosma, Cornulaca monacantha, 29. Stipagrostis plumosa, 30. Aristida sp. 31. Andropogon sp., 32. Mediterranean forest and steppe, 33/34. Sahara, desert, 35. savanna. a. Macchia, cultivated land and oak forests near Ain Draham (N-Tunisia, Quercus coccifera, Qu. ilex, Qu. suber, Qu. faginea, Pistacia terebinthus, Olea europaea). b. Steppe, cultivated land and Pinus-Quercus forest at Djebel Chaambi as altitudinal change of vegetation (Central Tunisia, Stipa tenacissima, Rosmarinus officinalis, Juniperus phoenicea, Pinus halepensis, Quercus ilex). c. Semidesert at Dj. Dahar and Jeffara (S-Tunisia, Artemisia herba alba, Gymnocarpus decander, Rhanterium suaveolens, Olea europaea). d. Southern limit of semidesert and transition to desert south of El Golea (Central-Algeria, Ephedra alata Chenopodiacea, Retama raetam, Acacia raddiana, Panicum turgidum). e. Altitudinal change from desert to semidesert at Atakor/Ahaggar (Southern Algeria, Acacia raddiana, Panicum turdidum, Artemisia herba alba, Chenopodiaceae, Pistacia atlantica). f. Altitudinal change from desert to savanna at Bagzan Mts. (Northern Niger, Acacia raddiana, A. laeta, Panicum turgidum, Commiphora africana). g. Change from desert to savanna and the southern part of Sahara at Tigidit (Northern Niger, Acacia raddiana, A. ehrenbergiana, Maerua crassifolia). h. North-Sahelian savanna south of Tigidit (Northern Niger, Acacia raddiana, Maerua crassifolia, Commiphora africana, Aristida mutabilis). i. Mid-Sahelian savanna and cultivated land north of Tahoua (Central Niger, Maerua crassifolia, Balanites aegyptiaca, Acacia laeta, Aristida mutabilis). j. South Sahelian savanna near Dosso (Southern Niger, Combretaceae savanna- fallow bush, parks, alluvial vegetation, Acacia albida, A. raddiana, Combretum micranthum, Hyphaene thebaica, Adansonia digitata). k. North Sudanian savanna and parks near Gaya (Southern Niger, Khaya senegalensis, Parkia biglobosa, Adansonia digitata, Combretum micranthum).

vegetation types as “steppe”-units (grass-steppe and shrub-steppe). Differences between them are explained as secondary phenomena. The main point is presented by Le Houerou (1969). He describes both vegetation types—as steppe phenomena—in a degradation series starting from Pinus halepensis-Juniperus-(Quercus)-forests into the Stipa tenacissima-steppes and finally into the shrub formations of Gymnocarpus-RhanteriumAtractylis. Artemisia-communities are regarded as degradation stages of both principal units. It is obvious, that use and overuse of these landscapes are omnipresent and may go back to the last ten thousand years (Gabriel, 1984). Consequently pioneer elements, as Artemisia will dominate several communities. Moreover, they will make “steppe” and “semidesert” units similar in the first stages of recovering. However, in a time

70

Erhard Schulz et al.

of about twenty years they may be overtaken by Stipa or Gymanocarpus respectively (Abichou, 1988). One may characterise the semidesert as a landscape dominated by dwarf and small shrub communities. They range from Rhanterium-Gymnocarpus-Atractylis-units on medium to coarse-grained substratum to Calligonum-Retama-Ephedra-units colonising fine-grained material or Stipagrostis pungens on loose dune sands. The first units often are dominated by Artemisia sp., the Mediterranean pioneer shrubs. Silts and clay are normally colonised by Chenopodiaceae (Suaeda, Salsola or Arthrocnemum), which also may be explained by the salt accumulations in this material by evaporation of the soilwater. The semidesert is dominated by Saharan floristic elements, either Saharo-Sindic or with a Mediterranean affiliation (see Barry et al., 1973ff.). The vegetation remains diffuse with a soil cover of about 30% in the North. The distances between the individual shrubs are directed by the extended root system, which exploits all kind of humidity ranging from rainfall to fog and dew. There still is another strategy of life in the Sahara. Whereas these shrubs are equipped with a complicated physiological system in coping with drought and frost events in their life span, therophyte seeds answer immediately to the variable rainfalls in creating achabs of different extension. These achabs or short time rain floras have only a short life span and they have to fulfil their life cycle in the time the amount of water allows. They take a regular part in the semidesert system. Nomads know them and exploit them systematically. The achabs consist of grasses and herbs with a floristic composition belonging to the Sahara as well as to the Mediterranean (Ammodaucus leucotrichus, Convolvulus arvensis, Diplotaxix harra, Eragrostis papposa a.o). The semidesert is also represented in figure 2 by the unit C, which describes the situation of the Dahar Mts. in Southeastern Tunisia. On the plateau the southernmost steppe units are bound to sand cover and fog. The area is characterised by the intensive Olea-cultivation either in small gardens (jesur) or in plantations similar to those around Sfax in the Tunisian Sahel. The southern limit of the semidesert is astonishingly clear. Around 30°N in the Central Sahara the diffuse shrub vegetation comes to an end within a distance of a few km. This boundary is sharp, it traverses a variety of substrata, and it can be recognised over the continent. Near the boundary the semidesert is represented by the Ephedra- Arthrophytumunits on coarse material and Retama-Calligonum-units on loose sand. In the Western Sahara it is dominated by Anabasis- or Fredolia-units. In the Libyan or western desert of Egypt it is reported to find an end already by 31°N (Stahr et al., 1985), whereas in the Atlantic region it follows the Atlas chain to the Southwest and continues as a parallel vegetation band parallel to the coast down to 19°N (Quezel, 1965; see also below). The sharpness of the semidesert boundary and the fact, that it transverses several substrata, characterises it as a clear climatic boundary, where the sum of regular precipitation, runoff and dew is no longer sufficient enough to allow a diffuse permanent soil cover. Unit D in figure 2 explains this situation south of El Golea in Central Algeria. 4.3.2 The desert South of that sharp boundary a change in vegetation and landscape takes place. The vegetation is contracted and dominated by trees, tussock grasses and shrubs. The term “contracted” means that permanent plant life does only exist in special places where the basic needs of water are fulfilled by the combination of precipitation, run off, dew or groundwater. These are wadis or depressions (Figure 3). Physiognomically the vegetation type belongs to the so called “desert-savanna-units” (Quezel, 1965) and phytosociologically to the Acacio-Panicion alliance (Barry, 1982).

The desert in the Sahara. Transitions and boundaries

71

Figure 3. The contracted desert vegetation with Acacia raddiana, Fagonia aegyptiaca, Cassia italica and Colocynthis vulgaris in the Enneri Achlouma/Northeast Niger.

Two main types can be separated. Wadis and depressions with coarse sand are colonised by Acacia (mostly A. raddiana) and Panicum turgidum whereas areas with fine sand are settled by Tamarix aphylla or T. saharae and Stipagrostis pungens. The general vegetation map (Figure 1) shows that the contracted Aciacia Panicum- or Tamarix-Stipagrostis-vegetation is mainly restricted to the mountain areas and the wadis radiating from them into the plains. The main components of these vegetation units are trees like Acacia raddiana, A. ehrenbergiana, A. nilotica, Maerua crassifolia and Balanites aegyptiaca, Capparis decidua, Tamarix aphylla together with shrubs like Salvadora persica, Aerva javanica, Zilla spinosa or Leptadenia pyrotechnica. Tussock grasses are represented by Panicum turgidum Stipagrostis pungens or Cymbopogon schoenanthus. From the floristic composition also a lot of pioneer elements from the Sahelo-Sudanian regions are present like Balanites aegyptiaca, Aerva javanica or Solenostemma oleifolium and Cassia italica. Aerva javanica and Balanites aegyptiaca reach the Ahaggar and the Arak gorge north of Ahaggar respectively. The dominance of trees can be explained by the presence of the dualistic root system of the Acacias. A taproot may reach to ground water horizons as deep as 55 m (Abadie, 1927) whereas horizontal roots beneath the soil surface exploit precipitation and run off. In that way trees may become rapidly independent from the actual climatic conditions when they reach a ground water horizon (Ullmann, 1985). The intensive root system of the tussock grasses will act in the same way. Thus, acacias may react in two different ways to the climatic constraints. As often visible saplings from horizontal roots will form genetically identical groups whereas in favorable years a mass of seedlings will occur as one could observe in Northern Niger in 2006. Only when these seedlings will reach the first groundwater lenses or horizons they can continue to grow. Outside the area of the contracted vegetation the achab represent the second strategy of plant life. As already shown for the semidesert the presence of therophytes or even biennial grasses and herbs is directly connected to aleatoric rainfalls. Cyperus conglomeratus, Aristida acutifolia, Stipagrostis plumosa, Schouwia purpurea, Indigofera coerulaea, Aerva javanica etc. are the main components of these vegetation plots, which may extend from

72

Erhard Schulz et al.

some square meters to several tens of square kilometres. Observations in the Ténéré (Niger) showed a regular sprouting of grasses and Cyperaceae from their seed bank in layers of fine sand covered by coarse sand (Figure 4). In this way the humidity of rainfalls can be stored and used for the germination. Kehl and Bornkamm (1993) pointed out, that these aleatoric rainfalls also may lead to an installation of permanent plant communities for several years if they manage to reach the first ground water level. In the Southern and Southeastern Sahara these achabs are known as “gizu” (Murat, 1936; Wilson, 1978). Even if they depend on the aleatoric rainfalls and they are not present every year they characterise large regions in Northern Sudan, Chad and Niger between 14°30'N to about 16°N (Murat, 1936; Schulz and Adamou, 1988; Schulz et al., 1999). Wilson (1978) described their appearance as extended grass areas in the Southern Sudan and as limited and isolated plots of herbs and bushes in the Northern Sudan. These “gizu” pastures are regularly exploited by camel herders and regarded as an important part of their pasture area. Out of their economic value the appearance of “gizu” pastures serve as an important step to the stabilisation of soil surfaces and moreover as an indicator of the interaction of Monsoon and Mediterranean air masses. The “gizu” rainfalls last from the end of summer just to the end of winter. Observation in Northern Chad during spring 1997 evidenced repeated rainfalls in March and proved their existence from December 1996 on. This was based on the age estimation of “gizu” pastures. In March 1997 these rainfalls were connected to the interaction of cold air masses coming from the

Figure 4. Achab vegetation in the eastern Ténéré, Eastern Niger. Stipagrostis generally roots in fine sand layers, which store the humidity filtered down through the coarse sand layers.

The desert in the Sahara. Transitions and boundaries

73

North and intrusions of Monsoon air masses (Schulz, 1999). Southwest of Adrar des Iforas, Air Mts. and Tibesti these achabs regularly evolve into large areas providing the image of savannas combined to the extended wadi vegetation (see figure 1). Thus, the permanent vegetation characterises and defines the desert as a region, where permanent life is restricted to depressions or wadis where beside of rainfall, dew and run off groundwater is available. Even smaller changes in substratum or relief are followed by a change of vegetation. Sandy patches in rocky areas normally are colonised as the first places by achabs or small bushes. Units D and G in figure 2 describe these changes in the North between semidesert and desert and in the South between desert and savanna. 4.3.3 The change with altitude There are two modes of change with altitude. The first is represented by the “Mediterranean type”, visible in the Ahaggar, Tibesti and in parts of the Tassili-n-Adjjer. From about 2.000 m altitude (Ahaggar) or 2.900 m (Tibesti) the contracted vegetation of the Acacia-Panicumtype is replaced by diffuse shrub communities of the Artemisia-Ephedra type representing the change to semidesert environments (Quezel, 1958, 1964, 1965, 1968) In addition the Ahaggar shows relict tree- and shrub units as Olea laperrini and Pistacia atlantica, whereas the Tassilli-n-Ajjer is known for the presence of Cupressus duprezziana-stands (Leredde, 1957). Erica arborea is described for the summit area of the Tibesti Mountains (Quezel, 1958). Climatically these environmental changes depend on the regular presence of Mediterranean winter rainfall in the higher altitudes. The second type of altitudinal change is manifested in the Air Mts./Northern Niger. Isolated High Mountain massives expose the transition to a diffuse tree and shrub vegetation over 1.500 m asl (Schulz and Adamou, 1988; Schulz et al., 1999) consisting of Acacia raddiana, A. laeta, Commiphora africana and Rhus tripartita. These savannas depend on elevated summer rainfall and this model represents the “Sahel type” of altitudinal change. A variation of this type is the appearance of Acacia-Indigofera-savanna on the high plateaus of the Ennedi in Northern Chad (Gillet, 1968). The Adrar des Iforas in Northern Mali however does not reach the necessary altitude for a change of vegetation and remains a Saharan desert mountain area (Schulz et al., 2001; Voss and Krall, 1994a, b). The units E and F in figure 2 describe the two types of altitudinal change.

4.3.4 The southern boundary of both Sahara and desert. The Saharan savanna The southern boundary of the desert is the most disputed limit between landscape zones or over regional ecosystems. It caused the misunderstanding of degradation, desertification and extension of desert reaching back to the alarm cries from Stebbing (1935), even it was corrected very soon (Aubreville et al., 1937/1973). However, information on this boundary is still very scarce. For long times the Sahel was regarded as an poorly defined transition zone between the Sahara and the (Sudanian) savannas proper, forming more or less an ecotone (Barry et al., 1986). It took a long time until the Sahelian savannas were accepted as consistent vegetation units (cf. Monod, 1986; Le Houerou, 1989). The Maerua-savannas as the step of vegetational adapatation to rising humidity in the Southern Sahara In the lowlands of the Southern Sahara we could observe a change from the contracted desert vegetation to a diffuse tree formation in Northern and Eastern Niger. The figures

74

Erhard Schulz et al.

Figure 5. Block diagram showing the depression and oasis of Agadem, Eastern Niger. Note the Maerua-savanna on the slopes and on the plateau.

Figure 6. The Tigidit escarpment, Northern Niger. The linear wadi vegetation contrasts to the sparse Maerua-savanna on the sandstone plateau.

5 and 6 show the examples of the sandstone cuesta of Agadem and Tigidit in the East and North of Niger. The depression and oasis of Agadem represent the southernmost parts of the cuesta system of Bilma in Eastern Niger. The oasis of Agadem as well as the plateaus of Homodji and Tcheni Tchadi South of it consist of deeply fissured sandstone, which, is iron crusted in its upper part. In the oasis of Agadem (Figure 5) the centre is sand covered and it bears a group of Acacia ehrenbergiana, A. raddiana, Phoenix dactylifera and Hyphaene thebaica trees with some Salvadora persica bushes. The sandy plain supports some Capparis decidua and Maerua crassifolia trees, Panicum turgidum tussocks and some Salvadora bushes. Tussock grasses like Stipagrostis vulnerans or therophytes like St. adscensionis or Cyperus conglomeratus colonise the dune covered lower slopes. Trees like Maerua crassifolia, Aacia raddiana or Capparis decidua grow in the wadis accompanied

The desert in the Sahara. Transitions and boundaries

75

by Leptadenia pyrotechnica-bushes. The fissured sandstone of the cuesta as well as the block accumulations of the slopes expose a sparse but diffuse tree vegetation composed of Maerua crassifolia. In 1984 we observed, that the trees on the slopes were all living, including great and rather old exemplars. On the plateau however, trees were small and about the half of them were dead. The plateau may have been colonised in the 1960ies, but during colonial times there were many old trees on the slopes (A. Kelle, personal communication). The plateaus of Homodji and Tcheni-Tchadi Southeast of Agadem showed a similar picture. The sandstone is fissured and covered by an iron crust. The plateaus supported at that time a diffuse tree vegetation consisting of Maerua crassifolia accompanied by a few Acacia ehrenbergiana and Fagonia-bushes and grasses of Panicum turgidum, Stipagrostis vulnerans, St. adscensionis. Here, all Maerua and Acacia trees were alive. In the southern foreland of the plateaus some extended achabs of Stipagrostis adscensionis were present as well as isolated Capparis deciduas-trees in some dune depressions. Another example is the Tigidit escarpment in the Southeast of the Air Mts., which rises about 80 m above the foreland. The cuesta consists of sandstone of Cretaceous age. It is highly dissected and the sandstone shows a mosaic of fissures and has a thin sand cover. Figure 6 shows the different types of vegetation. The wadi in front of the cuesta follows the scheme of the contracted vegetation of the desert with an arrangement of large Acacia and Balanites trees and Salvadora bushes. Tussok grasses (Panicum) form an understorey. On the fine-grained outer parts grows an extended population of Schouwia purpurea (Cruciferae) providing an important pasture. The plateau bears a sparse tree vegetation of Maerua and Acacia (ehrenbergiana). In spring 2005 the Maerua-trees were all living and we estimated an age of 20 years. This sparse but diffuse tree formation with a variable grass component is to classify as a savanna. It is also present in North-Mali (see map, figure 1). These Maearua-savannas colonise fissured sandstone environments, where finegrained material and humidity is collected in the cracks and exploited by the flat root system of Maerua crassifolia. During long-lasting drought periods survival is difficult. Thus populations of the same age might be explained. These Maerua-savannas are well restricted to a substratum providing a cistern situation and consequently a diffuse colonisation in a desert environment, where else permanent life is still restricted to wadis and depressions. The Acacia-Panicum-savannas as the definite change from desert to savanna If one proceeds at about 16°30'N in Niger from the North to the South both in the Tigidit area southwest of Agadez and in the Tintoumma in the East, South of Bilma, one can observe a densification of the tree lines and within a distance of about one km a change to the diffuse repartition of the Acacia-Panicum-vegetation. This represents the transition from the desert into the savanna, a diffuse tree and grass-vegetation dependent on tropical summer rain. The floristic elements of the vegetation still remain Saharan. A similar change is performed in Northern Mali (Barry, 1982) and in Northern Mauritania (Barry et al., 1987) Observations in Northern Chad (Schulz, 1999) showed a modification by substratum. The large inundation plains are devoid of plants but the sand ridges between them bear a treevegetation and at about 16°N a diffuse tree and tussock grass savanna occurs in the sandy plains. A comparable transition belt characterised by substratum changes is announced for the Northern Sudan by Akthar-Schuster (1995). Contrary to the Maerua-savannas described above this transition takes place on the same substratum and it is no longer totally dependent on the underground conditions. Several kilometres South of this boundary another transition takes place. The plant cover remains as a savanna but the floristic components are different. Beside Acacia Commiphora-trees become characteristic. The tussock grass Panicum turgidum disappears and annual grasses like Aristida mutabilis or Chenchrus biflorus are dominant. This is the

76

Erhard Schulz et al.

Figure 7. Block diagram showing the southern margin of the Sahara in Niger. Above the change from linear desert to the diffuse savanna vegetation at about 150 km south of Agadez, below in the Erg of Bilma (Schulz, 1999, modified).

transition to the Sahelian savanna system, which for its northern part is phytosociologically characterised by Barry (1982) as the Acacio-Aristidion alliance in contrast to the Saharan Acacio-Panicion. Other tree elements in these savannas are Maerua crassifolia, Balanites aegyptiaca and shrubs like Grewia tenax, Ziziphus mauritiana or Boscia senegalensis. As the general map (Figure 1) shows, the Sahelian savannas change southwards to Acacia-LeptadeniaBalanites-Piliostigma-units and finally to Combretaceae-savannas. The Units G to K in figure 2 describe the southern boundary of the desert and the succession of different savanna types of the Sahel and the northernmost Sudan. It is necessary to mention again that the Sahelian and Sudanian savannas represent cultural landscapes in different intensity and history. 4.4 THE SPECIAL CASE OF THE ATLANTIC SAHARA Travelling in Southern and Western Mauritania along the road from Aleg, Boutilimit, Nouakchott and Nouadibou in April 2005 we observed the transition in the Southern Sahara where humidity is largely provided by fog. Hitherto only little information is available on this region. Quezel (1965) characterised the oceanic Sahara by shrub formations of Nitraria, Zygophyllum, Gymnocarpus, in the littoral zone also with Tamarix and Artrocmenom trees and bushes. On dune areas he noted a mixed tree and shrub formation of Balanites, Acacia (raddiana) or Salvadoraand Cornulaca-bushes. He described the presence of an Acacia raddiana, Euphorbia balsamifera, Lycium intricatum-tree and shrub land for these hills. Barry (1989) distinguishes in his floristic analysis between the alliances of Acacio Panicion and the

The desert in the Sahara. Transitions and boundaries

77

Figure 8. The degraded Acacia-Leptadenia savanna near Aleg, Southern Mauritania.

Figure 9. The semidesert vegetation around Nouakchott/Mauritania.

Aervo-Fagonion both of Saharan character. Diak (1977) presented a map showing a change from the Commiphora-Acacia-savannas of the Sahel to Acacia (senegal)-savannas and to the coastal Tamarix-Salsola-Salicornia formations giving place to the North to tussock grass (Stipagrostis pungens) and Acacia raddiana formations of the dune areas. Culmsee (2002a, b) described the savannas and semideserts around Nouakchott and Akjoujt and Tabia et al. (2005) analysed the recent vegetation changes in the area of Lake Aleg. In 2005 the dune areas between Aleg and Nouakchott were monotonously covered by a loose and degraded savanna of Acacia senegal, A. ehrenbergiana, Balanites and Leptadenia pyrotechnica (Figure 8). Some Panicum turgidum-tussocks remained in the dune depressions. The whole region was totally overgrazed, which explains the presence of this pioneer vegetation, where trees are generally replaced by Leptadenia-bushes. A diffuse scrubland of Salsola and Zygophyllum presents the image of an extended semidesert on the disturbed sand and marls north of Nouakchott (Figure 9), which is also described by Culmsee (2002). It certainly depends on the intensive overgrazing of the region. About 160 km north of Nouakchott, a diffuse tree and grass vegetation is present in the area of disturbed dunes (Figure 10). Some Zygophyllum and Salsola bushes, mainly along the roadside accompany these savannas.

78

Erhard Schulz et al.

Figure 10. The Acacia-Panicum savanna about 160 km north of Nouakchott, Western Mauritania.

Figure 11. The Maerua savanna about 180 km north of Nouakchott/Western Mauritania.

At about 180 km north of Nouakchott the type of the savanna changes again (Figure 11). The open Maerua crassifolia-savanna appears with an understorey of Panicum turgidum tussocks may reach to about 20% of soil cover. Zygophyllum or Salsola bushes are also presen (Figure 11). 90 km to the North the dunes and extended wadis bear some achabs of the Stipagrostis adscensionis-type together with some Chenopodiaceae bushes. Trees are only present as Acacia (raddiana) in lines representing the change to a contracted vegetation. This desert formation dominates over 150 km with a varying density but without changing its main character (Figure 12). From about 400 km north of Nouakchott on (Figure 13) the dune area is again colonised by a diffuse vegetation of small bushes of Zygophyllum, Salsola- and Linariabushes, some annual grasses and sparse Acacia (raddiana)-trees. Hyoscyamus stands are present. The maximum soil cover of this vegetation is 20%. It represents the change to the semidesert, which characterises the westernmost Sahara from this region on to the North.

The desert in the Sahara. Transitions and boundaries

79

Figure 12. The linear desert vegetation with some achabs about 270 km north of Nouakchott/Northwestern Mauritania.

Figure 13. The semidesert vegetation about 350 km north of Nouakchott/Northwestern Mauritania.

Thus the general mosaic of vegetation types in the Southern Sahara manifested by the change from diffuse to contracted vegetation can be found also in this region. But the semidesert shrub vegetation widely interfingers with the plant cover of desert and savanna. This may be due to the special climatic conditions of the Oceanic Western Sahara. However, it is also evident that an extreme overgrazing has influenced both the floristic composition and the density of the different vegetation types (cf. Tabia et al., 2005). From the inland part of Northern Mauritania Quezel (1965) described a savanna vegetation on the inselbergs of the Zemour region. A diffuse tree-shrub cover mainly consisting of Maerua crassifolia, Rhus tripartita, Ephedra rollandi and Lycium intricatum colonises isolated granitic massives. This points to the special ecological conditions of isolated mountains providing cisterns for humidity and fin grained material as well as it indicates an interfingering with the northsaharan plant units as it is visible in the Oceanic part of the Sahara. It is certainly necessary to invest a lot of fieldwork to verify these observations. However, it will probably be impossible for certain areas because the region was a theatre of war in the past and still it is mined over large distances.

80

Erhard Schulz et al.

4.5 THE CHANGE OF VEGETATION IN TIME AND SPACE As explained above the double strategy of life in the Sahara allows a twofold reaction to severe or favourable conditions. Short time fluctuations are answered by the therophytes. Achabs may rapidly cover large area out of their seed bank. Trees react in a double way too: vegetative with root saplings or generative with seeds. In good years a mass production of seeds may lead to uniformous groups of the same age. Trees equipped with tap-roots may survive long and even severe periods, if they can rely on a groundwater horizon. This causes a certain vulnerability and resilience of these Saharan ecosystems. Vulnerability, because periods of decreasing amount of water and/or accelerated exploitation may easily lead to a decline of individuals or greater groups, especially when they have a similar age. Resilience, because they will recover from their own resources (vegetative way) or because pioneers will replace them for the time of one generation or for a longer period. It could provoke degradation-regeneration systems like the replacement of Acacia nilotica, A. raddiana, A. ehrenbergiana, Leptadenia pyrotechnica and inverse or the disappearance of Commiphora from the Southern Sahara and the Northern Sahel for a certain period. The latter is caused by the unique lateral root system of the Commiphora trees making it sensitive to drought as well as the intensive grazing. During the last years the permanent vegetation recovered in many regions such as in the Mangueni in NE-Niger. This region showed the general background of regeneration in the Sahara and Sahel. Periods of renovelation and regeneration are connected to climate as well as to human interference. Man exploits intensively both Sahara and Sahel (Adamou, 1979; Spittler, 1989; Schulz et al., 1999). Pastoralism, either as a full- or as a semi-nomadism, still is the most adapted exploitation system for the desert and the North Sahelian savannas (Ousseïni, 1996). Agrior horticulture is necessarily restricted to favoured places following the oasis structure of these ecosystems. The intensive human exploitation reaches back to the Neolithic and the question remains open whether the whole savanna system of the Sahel did not evolve from the contact of Sudanian and Saharan savanna-systems during the Middle Holocene due to human impact and pressure (Pomel et al., 1994). In periods of human absence the vegetation may recover to the level the respective climatic conditions will allow. It is the case for the Mangueni-Achelouma region (cf. figure 3), which is part of the traditional pasture area of the Tubu. However, since the last decade after the last rebellion the local Tubu clan did not exploit these pastures (A. Kelle, pers. comm.). Olson et al. (2005) showed that the long war in Central Sudan caused a regeneration of vegetation. It goes parallel to the fact that the renovelation periods of acacias in the Air Mts. in Northern Niger coincided with the evacuation of the population during the Kaoucen upheaval (Spittler, 1989) or during the present civil war. Thus, the question of stability–lability and vulnerability–resilience cannot only be evaluated from the physical conditions of these regions. 4.6 DESERTIFICATION AND REGENERATION As the southern boundary of the Sahara is still connected to the discussion on “Desertification”, this subject merits the same remarks as for the savannas (Cole, 1978): most discussed and least understood. From the 1930ies there was a steady warning against the spreading of the Sahara and /or desert ranging from the first cries of Stebbing (1935) to the complete misunderstanding of the vegetation dynamic by Tucker et al. (1991). Annual changes of the plant cover in Sahara and Sahel as detected by satellite images

The desert in the Sahara. Transitions and boundaries

81

were interpreted as a spreading of the Sahara to the South. However, these ideas are still active. The NASA (2007) states that expansion and shrinking of the Sahara resume to a net stability during the period of 1984 to 1994. For this area we have valuable informations on the landscape development during the last 200 years. The descriptions given by various voyagers for the Southern Sahara during the last centuries (s.a.) had one thing in common. About 100 km north of Lake Chad in the region of Ngourti they reported an important change within short distances either from forest to steppe or woodland to grassland. They all pointed to a stability of the borders between the major ecosystems. These are the same regions where the principal changes can still be observed today (Schulz and Hagedorn, 1996). The present debate on the “greening” of the Sahel resumes the whole debate on “Desertification” of the last decade (Anyambo and Tucker, 2005; Hellden, 1991; Hutchinson et al., 2005; Herrmann and Hutchinson, 2005; Mainguet, 1991; Mortimore and Turner, 2005; Nicholson, 2005; Olson et al., 2005). Herrmann and Hutchinson (2005) characterised one fact as very problematic, that of the low resolution of the satellite images. However one states a general improvement of vegetation since 1982 and the discussion is, whether it represents a return to past humid periods or to the development of a new equilibrium. There is a general change in understanding of climatic variability, of social processes or on the question whether the humid periods of the last centuries were more an exception than normality in the regional climatic system. After all this reasoning and disinvestments it became evident, that combat this degradation is only possible when the human impact is modified and combined with a general acceptance of the people concerned. The Southwestern Niger furnished good examples (Bender and Ousseïni, 2000; Ousseïni, 2002). Figure 14 may explain these interdependencies. After an intensive overuse of the Combretaceae savanna system and

Figure 14. Block diagrams showing the soil erosion risk in the Acacia albida-parks of Liptako/Southwestern Niger. The necessity of an integral antierosional intervention is evident (after Ousseïni, 2002, modified).

82

Erhard Schulz et al.

of the Acacia albida-parks (1, 2) it was obvious that an interference of anti-erosion measurements, which was only restricted to slopes (3) could be dangerous and also could provoke retrogressive and deep cutting erosion. Only a complete and long time assured interference may be successful and be maintained by the local population (4). The regeneration potential of the Saharan and Sahelo-Sudanian ecosystems is still present (Schulz and Hagedorn, 1994). This is based on the experience of fence and observation experiments (such as Maas, 1991) and observations on the regeneration of the South Saharan and North Sahelian savannas in Northern Chad (Schulz, 1999). Also the behaviour of soil surfaces shows the regeneration potential of these regions (Casenave and Valentin, 1989, Malam Issa et al., 1999). 4.7 DISCUSSION These observations were done during the last 30 years on various expeditions, which mostly were for coring of sediments in order to reconstruct palaeoenvironments. Out of logistical reasons these expeditions were conducted in the dry period (February to April) with the exception of one project in the Air Mts., which allowed investigations during the rainy season (cf. Schulz and Adamou, 1988). Thus, it only allowed a restricted picture of the herb vegetation but the observations are comparable for the permanent vegetation. However, the detection of the continental wide distribution of these formations allows a discussion of the nature of these ecosystem changes. The general reasoning on boundaries between ecosystems focussed on “ecotones” as regions of interactions between the respective units (Di Castri and Hanssen, 1992), as contact zones of two different ecosystems with marginal effects or as continua between them (Duvigneau, 1984), as transit areas (Odum, 1991) or simply as border biotopes between biozones of a certain extension (Frey and Lösch, 1998). This also involves the classification of the Sahelian steppes/savannas as transition units between the Sahara and the real (Sudanian) savannas (Barry et al., 1976,) or even as a “zono-ecotone” (Walter and Breckle, 1984). However, Di Castri and Hansen (1992) admit, that there are sudden changes in these interaction zones. These classifications all come from regions of a permanent and diffuse plant cover. Question arises, whether one can transfer them to areas, where long living organisms come to their definite limits of existence and where biocoenosises are not only determined by concurrence between them. Thus, the observed clear boundaries between two basic distribution types of permanent live—as diffuse or contracted—show that the whole system of the Sahara has a basic threshold-ecology. These thresholds are hydric, based on regular annual rainfall or on the combination of rainfall or fog, run off or humidity storage in soil or on the availability of groundwater. These are primary thresholds and they determine the basic chances of life in a certain area. The discussion of ecotones or transit formation is not adapted for these regions. Even if one admits, that achabs may connect desert and savanna for short periods, as animals do it too, the distribution of permanent life does not follow a gradient but is directed by these principal thresholds. The continental-wide presence of these boundaries on various substrata characterises them as climatic boundaries. Semidesert and savanna as northern and southern part of the Sahara landscape system are principally different and should not be combined to one unit (cf. White, 1981). However, it is unknown whether the present shape of these formations represents their full nature or only a type of basic survival. Our observations are done in the period of recent droughts and only the investigations of Leredde (1957), Monod (1954) or Quezel (1965) refer to the last humid period of these areas. Comparing their descriptions there is a basic agreement on the contracted mode of the permanent vegetation in the desert and of the clear boundaries with

The desert in the Sahara. Transitions and boundaries

83

diffuse formations of savanna and semidesert even other terms are often employed for them such as “steppe” for the semidesert. On the other side some historical descriptions point out, that the Maerua-savannas had a much wider distribution in the Central Southern Sahara at the time Barth (1854), Nachtigal (1879) or Tilho (1911) travelled on the Borno–Tripolis road (Schulz and Hagedorn, 1994). These authors announced them for the dune areas south of 16° N. It seems, that what we could observe in Mauritania will represent a model, which we may transfer to the central part of Southern Sahara too. Certainly, the intensive overgrazing during the last decades will have modified the shape and the floristic and faunistic inventory of these formations. It may be comparable for the semidesert too, where the presence of trees south of the Atlas chain in Morocco and in Tunisia raises the question of the principal treeless nature of these formations. It is certain, that all these questions demand much more fieldwork—also in the rainy season. However, the present situation of the Sahara will not allow it. Wide areas are a theatre of war and the distribution of landmines will exclude many regions from travelling and investigations.

4.8 CONCLUSIONS The tripartition of the Sahara in semidesert, desert and savanna illustrates the strategies of living organisms in coping with severe conditions mainly the lack of water. In the interaction space of Mediterranean/Atlantic cyclones and the monsoon the semidesert as well as the Saharan savanna receive enough precipitation to maintain a diffuse cover of permanent shrubs or trees and tussock grasses respectively. The desert is characterised as a region where permanent life is restricted to favoured places where additional water is provided by runoff and groundwater. Thus the plant cover of mainly trees and tussock grasses appears as linear or contracted. In all these ecosystems there is a second type of exploiting aleatoric resources such as irregular rainfalls. Achabs or gizu appear as short time but diffuse plots of therophytes, which have to fulfil their life cycles according to these irregular rainfalls. The boundaries of the desert are sharp as for the North between semidesert and desert and for the South between desert and savanna. These can be followed over the continent on several substrata and thus they are climatic boundaries. Substratum, relief or exploitation cause changes and variations within these limits, as it is visible in the appearance of an intermediate savanna formation of Maerua, Capparis and Acacia on the sandstone plateaus of the Southern Sahara. There is a distance between the Maerua- and the Acacia-Panicum-savannas in the continental Sahara, whereas in the Atlantic province they have a direct contact. These ecosystems have an elevated vulnerability but also resilience is high both in Sahara as in the desert resulting into a general tendency of regeneration of vegetation and soil. Regeneration of vegetation is often restricted to the vegetative mode and only in favoured period of precipitation—or protection—a generative renovelation is visible. Thus the desert remains stable—in a secular scale—within these boundaries and the Sahara seems to be stable too. The sharp boundaries of desert and Sahara and the shape of the Saharan and Sahelian savannas explain that one has to revoke ideas of only transition between the desert and the savannas—often understood as Sudanian savannas. The Sahel too stands as a consistent savanna system and not as an ecotone interfingering Sahara and Sudan. The general limited resources force all living organisms to adjust to this situation. Neglect, overuse or mining of non-renewable resources may easily lead to long-lasting devastations.

84

Erhard Schulz et al.

As the vegetation map is conceived as a modern model to interpret palaeo-environments, question arises, to which extension the actualism may work. We have to accept, that we dispose on only very limited descriptions and investigations for these waste areas. However, several historical reports confirm the basic features of clear and sharp boundaries between a diffuse and contracted mode of permanent life in the Sahara. The present dynamic of vegetation points to the function of threshold. Short time fluctuations of rainfall are immediately answered by the achabs or gizu, which may cover large areas and also stabilise them for a certain time. A repeated formation of them will reduce deflation. A favourable combination of rainfall may cause a formation of tree formations in a vegetative or generative way. If their roots reach to a groundwater level, they will survive even if the rainfall diminishes again. It means that a small change in the precipitation regime will cause important changes in the landscape without changing its principal nature. The comparison of the continental distribution of the South Saharan savannas in Niger with the oceanic one in Mauritania already points to this fact. A next step would follow a longer period of slightly increased precipitation. This would be an extension of the Saharan savannas, which could conquer large areas. This still would remain in the Saharan dynamic and also include the sharp boundaries of savanna areas to a desert environment. During these periods plants—trees, shrubs or herbs—may invade and settle on favourable places like valleys, niches in cuestas or around springs and lakes. Remnants of these elements may be found in suitable sediments. They will allow a reconstruction of the former landscape type but the knowledge of the present dynamic of the vegetation will point to the narrow limits to take them as indicators for a different climate in a greater area. Landscapes may react very fast on slight changes in the precipitation regime, but several features of them will remain, even if the climatic modifications are already terminated. The present observations and the description from the last two hundred years all point to the stability of the Saharan ecosystems. They are vulnerable but they have a great recovering potential even if we still know very little on their dynamic in time. This also includes the various modes of human interference in the desert as in the savanna. ACKNOWLEDGEMENTS We are grateful to I. Ullmann for valuable suggestions and to K. Wepler for help in cartography. REFERENCES Abadie, M., 1927, La colonie du Niger, (Paris: Societé d’ Edition Geographique Maritime et Coloniale). Abichou, A., 1988, El Bahira (Sud tunisien) deux géofacies dans un milieu aride. (analyse et cartographie). (Caen; Mémoire de maitrise de Géographie). Adamou, A., 1979, Agadez et sa region. Etudes Nigriennes, 44, pp. 1–358. Akhtar-Schuster, M., 1995, Degradationsprozesse und Desertifikation im semiariden randtropischen Gebiet der Butana/Rep. Sudan, Göttinger Beiträge Land- und Forstwirtschaft in den Tropen und Subtropen, 105, pp. 1–166. Anhuf, D. and Frankenberg, P., 1991, Die naturnahen Vegetationszonen Westafrikas. Die Erde, 122, pp. 243–265. Anyambo, A. and Tucker, C.J., 2005, Analysis of Sahelian vegetation dynamics using NOAA-AVHRR NDVI data from 1981–2003. Journal of Arid Environments, 63, pp. 596–614.

The desert in the Sahara. Transitions and boundaries

85

Aubreville, A., Patterson, J.R., Collier, F.S., Brynmor, J., Dundas, J., Mathey, J. and Bachelier, R., 1973, Rapport de la mission forestière anglo-française Nigéria-Niger (décembre 1936–février 1937). In Revue Bois et Forêts des Tropiques, 148, pp. 3–26. Ballouche, A. and Dolidon, H., 2003, Forêts claires et savanes ouest-africains: Dynamiques et évolution de systèmes complexes à l’interface nature/societé. In: La foret: enjeux compares des formes d’appropriation, de gestion et d’exploitation dans les politiques environmentales et de contexte d’urbanisation generalisée edcited by M. Taabni, Université de Poitiers), pp. 56–70. Barbour, K.M., Oguntoyinbo, J.S., Onyemelukwe, J. and Nwafor, J.C., 1982, Nigeria in maps, (London). Barry, J.P., 1982, La frontière méridionale du Sahara entre l’Adrar des Iforas et Tombouctou. Ecologia Méditerranea, 7, 3, pp. 99–124. Barry, J.-P., 1989, Le problème des divisions bioclimatiques et floristiques au Sahara. Note VII: Les confines saharo-sahelien en Mauretanie, Bulletin de la Societe Botanique de France Actual Botanique, 136, 3–4, pp. 67–84. Barry, J.-P., 1991, La végétation des massifs sahariens: Hoggar, Tassilis, Adrar des Iforas et Adrar Tmar de Mauretanie. Revue de Geographie Alpine, 1, pp. 55–70. Barry, J.-P. and Celles, J.-C., 1972–73, Le problème des divisions bioclimatiques et floristiques au Sahara algerien. Naturalia monspeliensia, serie Botanique, 23–24, pp. 5–48. Barry, J.P., Celles, J.-C. and Faurel, L., 1974, Carte Internationale. du tapis végétal et des conditions écologiques 1:1 mio. Alger, Algiers. Barry, J.P., Celles, J.-C. and Manière, R., 1976, Le problème des divisions bioclimatique et floristique au Sahara algerien, Note II. Le Sahara central et le Sahara méridional. Naturalia monspeliensia, serie Botanique, 26, pp. 211–242. Barry, J.P., Celles, J.C. and Manière, R., 1981, Le problème des divisisons bioclimatiques et floristiques au Sahara algerien. III l’analyse de la végétation de la région d’In Salah et de Tamanrasset. Naturalia monspeliensia Série. Botanique, 44, pp. 1–8. Barry, J.P., Celles, J.C. and Musso, J., 1985, Le problème de divisions bioclimatiques et floristiques au Sahara algerien, Note IV, le plateau de Tademait et ses alentours (carte Ouargla). Ecologia Mediterranea, XI, 2/3, pp. 23–181. Barry, J.P., Celles, J.C. and Musso, J., 1986, Le problème des divisions bioclimatiques et floristiques du Sahara. Note V: du Sahara au Sahel, Un essai de définition de cette marche africaine aux alentours. Ecologia Méditerranea, XII, pp. 187–235. Barry, J.P. and Faurel L. without date. Carte de la végétation de l ’Algerie. 1: 2500 000. (Algiers; Institut. de cartographie. de la végétation. de l’universite. d’Alger). Barry, J.P., Jaquen, X., Musso, J. and Riser, J., 1987, Le problème des divisions bioclimatiques au Sahara. Note VI: entre Sahel et Sahara. L’ Adrar mauritanien. Approches biogéographiques et géomorphologiques. Ecologia. Méditerranea, XIII, 1/2, pp. 131–142. Barth, H., 1857–1858, Reisen und Entdeckungen in Nord-und Central-Afrika. 5. (Gotha). Bornkamm, R. and Kehl, H., 1985, Pflanzengeographische Zonen in der Marmarika. Flora, 176, pp. 141–151. Capot-Rey, R., 1953, Le sahara francais. (Paris; Presse Universitaire de France), pp. 1–564. Casenave, A. and Valentin, Chr., 1989, Les états de surface de la zone sahélienne. (Paris; Editions ORSTOM), pp. 1–229. Cole, M., 1978, The savannas. Biogeography and geobotany. (London: Academic Press). Culmsee, H., 2002a, The habitat functions of vegetation to the behaviour of the desert locust Schistocerca gregaria (Forskal) (Acridae: Orthoptera)—a study in Mauretania (West Africa). Phytocoenologica, 32, 4, pp. 645–664. Culmsee, H., 2002 b., Saharische und sudanische, Wüsten- und Savannenelemente in der Halbwüstenregion Mauretaniens. Berichte. der Reinhard-Tüxen-Gesellschaft, 14, pp. 69–79.

86

Erhard Schulz et al.

Diak, M.S., 1977, Végétation et faune. In Atlas de la Répubioque Islamique de Mauritanie (edited by Toupet, Ch., Laclavère, G., Monod, Th. (Paris: Eds. JA.). Di Castri, E. and Hansen, A.J., 1992, The environmental and development crisis as determinants of landscape dynamics. In Landscape boundaries, edited by Hansen, A.J. and Di Castri, F. Ecological studies, 92, (Berlin, Springer), pp. 3–18. Denham, D., 1822–25, Mission to the Niger. The Bornu mission, (Cambridge, Reprint 1966). Dolidon, H., 2005, L’espace des feux en Afrique del’Ouest. (Caen; Thèse l’Université de Caen), pp. 1–345. Dubief, J., 1959, 1963, Le climat du Sahara, Travaux de l’Institut des Recherches Sahariennes, 1, 2, 312, pp. 1–275. Duvigneau, P., 1984, La synthèse écologique, (Paris: Doin), pp. 1–380. Ern, H., 1979, Die Vegetation Togos, Gliederung, Gefährdung, Erhaltung. Willdenowia, 9, pp. 295–312. Frey, W. and Lösch, R., 1998, Lehrbuch der Geobotanik, (Stuttgart: G . Fischer), pp. 1–436. Gabriel, B., 1984, Vorgeschichte und Klimaentwicklung. In: Tunesien, edited by Schliephake, K., (Stuttgart: Thienemann), pp. 75–91. Gaston, A., 1979, Synthese agropastorale du bassin du lac Tchad, (Maisons Alfort; IEMVT). Gaussen, Ch. and Vernet, J., 1958, Carte internationale du tapis végétal au 1.000.000, Tunis- Sfax. Bulletin du Service de la Carte Phytogéographique Série, A, 3, 2. Giessner, K., 1988, Sahara—“Die grosse Wüste” als Forschungsobjekt der Physiogeographie. In Eichstätter Hochschulreden, 50, pp. 1–70. Gillet, H., 1968, Le peuplement végétal du massif de l’Ennedi (Tchad). Mémoire. du Mus. d’Histoire Naturelle, B, XVII. Hagedorn, H., 1985, Formen und Bilder der Wüste am Beispiel der Sahara. Natur und Museum, 115, pp. 210–230. Haaland, R., 1992, Fish, pots and grain: early to mid-Holocene adaptions in the Central Sudan. French Archaeological Revue, 10, pp. 43–64. Hanote, O., Bradley, D.E., Ochieng, J.W., Verjee,Y., Hill, E. and Rege, E.O., 2002, African Pastoralism: Genetic imprints of origins and migrations. Science, 296, pp. 336–339. Hellden, U., 1991, Desertification—Time for a reassessment, Ambio, 20, pp. 372–383. Herrmann, S.M. and Hutchinson, C.F., 2005, The changing context of the desertification debate. In Journal of Arid Environments, 63, pp. 538–555. Hooghiemstra, H., 1996, Aspects of Neogene–Quaternary environmental and climatic change in equatorial and Saharan Africa. Palaeoecology of Africa, 24, pp. 115–132. Hutchinson, C.F., Herrmann, S.M., Mankonen, T. and Weber, J., 2005, Introduction: The greening of the Sahel. Journal of Arid Environments, 63, pp. 535–537. Keay, R.W.J., 1948–1959, An outline of Nigerian vegetation, (Lagos), pp. 1–52. Kehl, H. and Bornkamm, R., 1993, Landscape ecology and vegetation units of the Western Desert of Egypt. Catena Supplement, 26, pp. 155–178. Klaus, D. and Frankenberg, P., 1980, Pflanzengeographische Grenzen der Sahara und ihre Beeinflussung durch Desertifikationsprozesse. Geomethodica, 5, pp. 109–137. Knapp, R., 1973, Die Vegetation von Afrika. (Stuttgart; G. Fischer) pp. 1–626. Le Berre, M., 1989, 1990, Faune du Sahara vol I, II., (Paris; Lechevalier-Chabaud). Le Houerou, H.N., 1969, La végétation de la Tunisie steppique. Annales Institut Natural Recherches Agronomiques, 142, (Tunis). Le Houerou, H.N., 1989, The grazing land ecosystem of the African Sahel, Ecological Studies, 75, pp. 1–282. Leredde, Cl., 1957, Etude écologique et phytogéographique du Tassili des Ajjer. Travaux. Institut des Recherches Sahariennes, Alger, série du Tassili, 2, pp. 1–455.

The desert in the Sahara. Transitions and boundaries

87

Letouzey, R., 1985, Notice de la carte phytogeographique du Cameroon au 1:50.0000. Institut de la carte internationale de la végétation, (Toulouse), 5. Maas, I., 1991, Weiderotation in Timia/Niger. Bericht über fünf Jahre Erfahrung. (Stuttgart), pp. 1–57. Mainguet, M., 1991, Desertification: Natural background and human mismanagement, (Berlin: Springer), pp. 1–235. Malam Issa, O., Trichet, J., Defarge, C., Coute, A. and Valentin, C., 1999, Morphology and microstructure of microbiotic soil crusts on a tiger bush sequence (Niger, Sahel). Catena, 37, pp. 175–196. Miehe, S., 1988, Vegetation ecology of the Jebel Marra Massif in the semiarid Sudan, Dissertationes Botanicae, 113, pp. 1–171. Monod, Th., 1954, Modes “contracté et diffus” de végétation saharienne. In: Biology of desert, edited by Cloudsley Thompson, (London), pp. 35–44. Monod, Th., 1986, The Sahel zone north of the equator. In: Hot deserts and arid shrublands, edited by Evenary, M., Nov-Meir, J., (Oxford), pp. 203–243. Mortimore, M. and Turner, B., 2005, Does the sahelian smallholders management of woodland, farmtree, rangeland support the hypothesis of human induced desertification, Journal of Arid Environments, 63, pp. 567–595. Murat, M., 1936, Végétation de la zone prédésertique en Afrique Centrale. Bulletin de la Societe. d’Histoire Naturelle.de l’Afrique du Nord, XXVIII, pp. 19–83. Nachtigal, G., 1879, Sahara und Sudan, (Berlin), 2 volumes. NASA, 2007, African vegetation: Comparing August 1984 and 1994. http://svs.gsfc.nasa. gov/vis/a000000/a000500/a000590/index.html Neumann, K., 1991, Vegetationsgeschichte der Ostsahara im Holozän. Holzkohlen aus prähistorischen Fundstellen (Mit einem Exkurs über die Holzkohlen von Fachi Dogonboulo/Niger), Forschungen zur Umweltgeschichte der Ostsahara. Acta Praehistorica, 2, pp. 13–181. Nicholson, S., 2005, On the question of the “recovery” of the rains in the West African Sahel, Journal of Arid Environments, 63, pp. 615–641. Odum, E., 1991, Prinzipien der Ökologie, (Heidelberg: Spektrum), pp. 1–305. Olson, L., Eklund, L. and Ardö, J., 2005, A recent greening of the Sahel-trend, patterns and potential causes. Journal of Arid Environments, 63, pp. 556–566. Ousseini, I., 1996, Les secheresses au Niger: cliches et realites, In: L’environnement au Niger, edited by Soumana, I. Resadep, (Niamey: Institute.Panos), pp. 17–29. Ousseini, I., 2002, Erosion et conservation des sols dans la vallée du Moyen Niger (République du Niger): interpretations préliminaires de mesure en parcelles et d’observations à l’echelle des petits et moyens bassins versants. In: Human response to drastic change of environment in Africa II, edited by Hori, N. (Tokyo: Metropolitan University), pp. 39–70. Pomel, S., Pomel-Rigeaud, I. and Schulz, E., 1994, Les indicateurs anthropogènes de la végétation et des sols de quelques savanes de l’Afrique de l’Quest. In: Enregistrateurs et indicateurs de l’evolution de l’environnement en zone tropicale, edited by Maire, R., Pomel, S. and Salomon, H.N., (Bordeaux; Presse univ. de Bordeaux), pp. 173–200. Quezel, P., 1958, La mission botanique au Tibesti, (Algiers) Institute des Recherches Sahariennes, Memoires, 4. Quezel, P., 1964, Carte internattionale du tapis végétal. 1:1 Mio. Feuille Largeau, (Toulouse). Quezel, P., 1965, La végétation du Sahara du Tchad à la Mauritanie. Geobotanica selecta, (Stuttgart: G. Fischer) 2, pp. 1–333. Quezel, P., 1968, Carte internationale du tapis végétal. 1:1 Mio. Feuille Djado, (Toulouse).

88

Erhard Schulz et al.

Reichelt, R., Faure, H. and Maley, J., 1992, Die Entwicklung des Klimas im randtropischen Sahara-Sahelbereich während des Jungquartärs—ein Beitrag zur angewandten Klimakunde. Petermanns Geographische Mittleilungen, 136, 2–3, pp. 69–79. Republique du Senegal, 1985, Carte du couvert végétal 1:500.000, (Dakar). Schliephake, K., 1995, Libyens Großer Künstlicher Fluß—Lösung der agrarischen Versorgungsprobleme? Mitteilungen des Deutschen Orientinstituts, 52, pp. 165–180. Schulz, E., 1994, The southern limit of the Mediterranean vegetation in the Sahara during the Holocene, Historical Biology, 9, pp. 1347–156. Schulz, E., 1999, The southern margin of the Sahara in the Republic of Chad. Vegetation, soil and present pollen rain. Zentralblatt Geologie. Paläontologie, I, pp. 483–496. Schulz, E., 2000, Present vegetation. In: The Holocene landscape and vegetation history of Northern and Western Africa. A palaeoecological atlas, edited by Schulz, E., Agwu, Ch., Dupont. L., Jahns, S. and Salzmann, U., www.uni-wuerzburg.de/geographie/fachi/pal_ atlas_afrika/index_atlas.htm Schulz, E. and Adamou, A., 1988, Die Vegetation des Aïr-Gebirges in Nord-Niger und ihre traditionelle Nutzung. Giessener Beiträge zur Entwicklungsforschung, I, 17, pp. 75–86. Schulz, E., Adamou, A. and Ousseïni, I., 2001, Air et Adrar des Iforas: une comparaison de deux montagnes du Sud du Sahara, et leur evolution actuelle, Espaces Tropicaux, 16, pp. 219–232. Schulz, E. and Hagedorn, H., 1994, Die Wüste, wächst sie denn wirklich? Die Geowissenschaften, 12, pp. 204–210. Schulz, E. and Himsel, H., 1993, Sahara-Sahel-Sudan. Honige als Spiegel der Vegetation des Niger, Würzburger Geographische Arbeiten, 87, pp. 373–388. Schulz, E. and Lueke, M., 1994, A two years pollen calendar for traditionally produced honey types from Gaya, Southern Niger. Grana, 33, pp. 254–259. Spittler, G., 1989, Dürren, Krieg und Hungerkrisen bei den Kel Ewey (1900–1985) (Wiesbaden: Steiner), pp. 1–199. Stahr, K., Bornkamm, R., Gauer, J. and Kehl, H., 1985, Veränderungen von Böden und Vegetation am Übergang von Halbwüste zur Vollwüste zwischen Mittelmeer und Quattara Depression in Ägypten, Geoökodynamik, 6, pp. 99–120. Stebbing, E.P., 1935, The advance of the desert. Geographical Journal, 91, pp. 356–359. Stokker, O., 1962, Steppe-Wüste-Savanne. Veröffentlichungen des Geobotanischen Institutes ETH Zürich (Rübel), 37, pp. 234–243. Stokker, O., 1976, The water-photosynthesis syndrome and the geographical plant distribution in the Saharan deserts. In: Lange, O., Kappen, L., Schultze, E.D. (Eds.), Ecological studies, analysis and synthesis, 11, pp. 506–521. Tabia, A.N., Gassanib, J., Elghadi, A.V., Ballouche, A. and Moguedet, G., 2005, Diagnostic de la dynamique des resources hydriques de surface et des processus de “desertification” du Lac d’Aleg et son basin basin versant (Brakna, Mauretanie) par teledetection multidate. Teledetection, 5, 1–3, pp. 123–137. Trochain, J., 1940, Contribution à l’étude de la végétation du Sénégal, Mémoires de l’IFAN, 2, pp. 1–433. Tukker, C.J., Dregne, H. and Newcomb, W.W., 1991, Expansion and contraction of the Sahara desert from 1980 to 1990, Science, 253, pp. 299–301. Ullmann, I., 1985, Tagesgänge von Transpiration und stomatärer Leitfähigkeit Sahelischer und Saharischer Akazien in der Trockenzeit, Flora, 176, pp. 383–409. UNESCO/FAO, 1968. Carte de la végétation de la région méditerranéenne 1:5000.000, (Paris). United Nations, 1992, Convention on desertification. (New York), pp. 1–44. Vischer, H., 1910, Across the Sahara from Tripoli to Bornu, (London), pp. 1–308.

The desert in the Sahara. Transitions and boundaries

89

Voss, F. and Krall, P., 1994a, Principaux biotopes du criquet pelerin dans le Nord dur Tilemsi (Mali), (Berlin, Eschborn). Voss, F. and Krall, P., 1994b, Principaux biotopes du criquet pelerin dans l’Adrar des Iforas (Mali), (Berlin, Eschborn). Walter, H., 1964, Die Vegetation der Erde. In ökophysiologischer Betrachtung, (Jena: VEB, G. Fischer, 2 vols). Walter, H. and Breckle, S., 1984, Ökologie der Erde, 2, (Stuttgart, G. Fischer), pp. 1–461. Wehr, H., 1968, Arabisches Wörterbuch, (Wiesbaden). White, F., 1983, The vegetation of Africa. Natural Resources Research (Paris: UNESCO), pp. 1–356. Wilson, R.T., 1978, The “gizu” grazing in the South Libyan desert. Journal of Arid Environments, 1, pp. 327–344.

CHAPTER 5

Palaeo-climatic evidence of soil development on Sahelian ancient dunes of different age in Niger, Chad and Mauritania Peter Felix-Henningsen and Peter Kornatz Institute of Soil Science and Soil Conservation, Justus Liebig Universität Giessen, Germany Einar Eberhardt Bundesanstalt für Geowissenschaften und Rohstoffe, Hannover, Germany ABSTRACT: Ancient dunes of the Sahel reflect lengthy arid climatic phases in which the desert margin of the Sahara extended southwards over a distance of several hundred kilometers from its present position. During the following semi-humid to humid climatic phases the dunes and sediments were fixed by vegetation and soil formation. The extent and duration of the humid phases determined depth and intensity of weathering and soil formation. In order to show the relationship between degree of soil development and age of the ancient dunes, soil profiles in the Sahel of Mauretania, Niger and Chad were investigated pedologically and by OSL dating. Due to the limited number of investigated sites and dated samples, general conclusions concerning the stratigraphy of ancient Sahel dunes are not possible. However, the examples described in this paper show that a relationship exists between dune age and depth and intensity of soil formation. The sedimentation of the Upper Pleistocene dunes of the Sahel in Niger started at least around 30 ka and ceased at ∼10 ka with the transition from the Upper Pleistocene to the Holocene, while the Middle Holocene dunes in Chad and Mauritania were sedimented between ∼5 and ∼3,5 ka. Due to the accumulated processes in the humid periods of the Early Holocene and the Mid-Holocene during soil formation on Upper Pleistocene dunes, the rubefication, accumulation of fines and leaching depth of soluble salts is much more pronounced than on Middle Holocene dunes, which were only affected by weathering from the Mid-Holocene until present. The alkaline pH values of the soils reveal that under recent climatic conditions in the Northern and Central Sahel, with annual precipitation up to 400–500 mm a−1 and deposition of dust containing carbonates and soluble salts, the progress of silicate weathering is inhibited. Only under higher precipitation in the southern Sahel weathering of silicates proceeds due to stronger soil acidity. Therefore the soils on ancient dunes of the Northern and Central Sahel are today relict soils, which mainly weathered in periods with conditions of higher precipitation than ∼500 mm a−1 during the Early and Middle Holocene humid periods.

5.1 INTRODUCTION In Sahelian countries ancient dunes form a several hundred kilometer broad belt south of the present southern desert margin of the Sahara, at approximately 16° northern latitude and stretching from East to West Africa. These morphologically more or less denudated dune fields, lee dunes at escarpments or inselbergs, eolian sand sheets and valley fillings

92

Peter Felix-Henningsen et al.

indicate lengthy arid climatic phases, in which the desert margin of the Sahara extended southwards over a distance of several hundred kilometers from their present position. During subsequent periods with semi-humid to humid climatic conditions, depending on the geographical position, the eolian sediments were covered by vegetation and fixed by soil formation. The latter is characterized by root channels, animal burrows, a brown to reddish brown colour and a cementation by silica and soluble salts, which indurates or stabilizes the normally loose eolian sand. In valleys the dune relief was dissected by water erosion due to the incision of wadis. Apart from archaeological findings or organic material, which allows radiocarbon dating, soils and palaeosols are also valid proxies for the stratigraphical and palaeo-climatical reconstruction of humid periods. With regard to soil characteristics, e.g. thickness, weathering intensity, depth of leaching or accumulation of soluble constituents, the influence of climate and biosphere accumulate over the time of soil development. Therefore soils in arid regions always represent combined information about the duration and intensity of humidity as the main factor. First investigations of soils on ancient dunes in the Goz-region of Sudan by Warren (1970) and Felix-Henningsen (1983, 1984) showed that two ancient dune generations existed, which can be subdivided by the intensity of soil development. On the older dunes of the “Low Goz”, which according to Warren (1970) are of Upper Pleistocene age, a thick rubefied palaeosol is developed, overlain by longitudinal dunes of the “High Goz”, on which shallow, weakly weathered soils have developed. The Holocene age of these dunes (Warren, 1970) is discussed by Gläser (1987). In Northern and Central Sahel of Niger and Burkina Faso, two different ancient dune generations also exist, which can be subdivided by the relief forms and the degree of soil development (Grunert, 1988; Völkel, 1988, 1989; Pfeiffer and Grunert, 1989; Pfeiffer, 1991; Völkel and Grunert, 1990; Felix-Henningsen, 2000). In addition to radiocarbon datings, which established an Early Holocene and a Mid Holocene humid period, the stratigraphy (supporting the previous results of French geologists) from three humid periods during the Upper Quaternary must be considered in the Southern Sahara of East Niger: (a) an Upper Pleistocene (Ghazalien) humid period between 40.000 and 20.000 BP (Servant, 1983), (b) a Late Pleistocene to Early Holocene (Tchadien) humid period between 14.000–7.500 BP with a maximum humid period between 10.000 and 7.500 BP (Servant, 1983), and (c) a Middle Holocene (Nouakchottien) humid period (Young Neolithic) between 4.500–3.000 BP (Michel, 1973). In arid periods prior to the humid climatic phases, ancient dunes and sand sheets were sedimented. Ancient dunes of the Sahel extend to the Southern (Völkel, 1988, 1989) and Central Sahara (FelixHenningsen, 2000) of East Niger. Relict and fossil palaeosols indicate the northward transgression of the Southern Sahara margin during the Early Holocene humid period over a distance of at least 600 km. In order to complete the stratigraphy of palaeo-climatic changes at the southern margin of the Sahara, OSL datings of ancient dunes of the Sahara and Sahel in Chad were carried out. The eastern Manga dunes, east of Lake Chad in the Northern Sahel, as well as the ancient dunes with palaeosols in the Sahara of Chad up to the oasis of Faya in the North, exhibited Mid Holocene ages between 4,7 and 3,1 ka and a development of rather weak and shallow soils (Mauz and Felix-Henningsen, 2005; Felix-Henningsen and Mauz, 2005). In the Sahel of Niger, west of Lake Chad, ancient dunes with a thick reddish brown soil, rubefied due to the formation of hematite, are distributed. A first OSL age of ∼9,5 ka (Mauz and Felix-Henningsen, 2005) indicates an end to dune sedimentation and consequently the beginning of soil formation in the following humid period of the Early Holocene. In order to show the relationship between degree of soil development and age of the ancient dunes, further investigations were carried out during the Sahara expeditions in 2005 and 2006 at one site in the Central Sahel of Mauritania, and at two sites in the

Palaeo-climatic evidence of soil development on Sahelian ancient dunes

93

Southern and Northern Sahel of Niger. Although general conclusions regarding the ancient dune stratigraphy are not possible due to the limited number of investigated sites and dated samples, the examples described in this paper reveal that the soils of the Sahel are partly relictic and, due to the duration and intensity of the weathering in previous humid periods, a relationship between the age of the ancient dunes and the degree of soil development does exists. Apart from palaeoenvironmental aspects, the results may have practical implications for land-use planning and combating desertification. Soil fertility decreases while soil stability against deflation, a common process of desertification as a consequence of overgrazing, increases with the duration, depth and intensity of soil development. 5.2 METHODS 5.2.1 Pedology At selected sites (Figure 1) soils were mapped by digging. Soil characteristics were described (Munsell colour, texture, structure and bioturbation, hardening and cementation, carbonatization and gleyification) and sampled for further pedochemical, physical, mineralogical and micromorphological investigation. The following physical, chemical and mineralogical methods were used: The Redness Ratio was calculated from the values of the Munsell moist colours RR = (10-Hue) × Chroma/Value according to Schwertmann et al. (1983). Texture: Amounts of sand >63 μm and fines <63 μm were determined by sieving without sample pretreatment. Organic carbon was measured using a C/N/S-Analyzer. Carbonates were dissolved with

Figure 1. Location of the study sites in the Sahel of Mauritania, Niger and Chad.

94

Peter Felix-Henningsen et al.

10% HCl and calculated from the CO2 gas volume according to Schlichting et al. (1995). Pedogenic Fe- and Mn-oxides were extracted by citrate-bicarbonate-dithionite (Schlichting et al., 1995) and the metal ions were analyzed by AAS (Perkin Elmer). Easily soluble (amorphous) Si- and Al-compounds were extracted with 0,5 M NaOH (Schlichting et al., 1995) and the elements analyzed by AAS (Perkin Elmer). pH values were measured in 1: 2,5 soil water suspensions (actual acidity). The total amounts of dissolvable salts (TDS) and the amounts of single salt species of different solubility were calculated from the concentrations of cations (Ca2+, Mg2+, K+, Na+, measured by AAS) and anions (SO42−, Cl−, NO3− , measured by a Dionex ion chromatograph, HCO3− calculated as the difference between summarized cations and anions) on the basis of the solubility products of salt species occurring in soils according to the method of Smettan and Blume (1987). 5.2.2 Optical dating The OSL dating of the samples from Profiles 1, 2 and 4 was carried out in the luminescence dating laboratory of the Department of Geography, University of Cologne. The basic methodological details and different protocols used for luminescence dating have been most recently described e.g. by Boetter-Jensen et al. (2003), Aitken (1998), Wintle (1997) and Murray and Wintle (2000). The sample preparation was carried out under subdued red light. All samples were treated with hydrochloric acid, sodium oxalate and hydrogen peroxide in order to remove carbonates, clay and organic material. To separate the quartz fraction for coarse-grain dating (100–200 μm), solutions of sodium polytungstate (2,62 and 2,68 g cm–3) were used. To purify and concentrate the quartz fraction and to remove the outer α-irradiated layer, the samples were etched in 40% HF for 40 minutes. Finally the grains were mounted on sample discs with the central 1, 2 or 8 mm covered with grains (100–200, 200–400 or >1.500 grains), in the following referred to as “aliquots”. Per sample about 12 up to 50 aliquots were measured to obtain equivalent dose (De) values, being an estimate of the amount of radiation dose that the sample accumulated within its crystal lattice since it was shielded from sunlight by subsequently deposited sediments. The equivalent dose values were estimated following the single-aliquot regenerativedose protocol (SAR) as proposed by Murray and Wintle (2000) that enables De determination for each individual sub-sample. Measurement of a range of aliquots allows a more detailed investigation of the dose distribution within the sediment, which is further improved by a reduction of the aliquot size or number of grains per sub-sample, respectively. All luminescence measurements were carried out on automated risø TL/OSL readers TL-DA12 or -15, equipped with 90Sr/90Y beta sources for irradiation and EMI 9235 photomultiplier tubes for luminescence detection. The BLSL signals (blue light stimulated luminescence) were measured for 50 s at constant temperature (125 °C) and the UV emission of quartz was detected in the in wavelength window 330 ± 40 nm generated by a Hoya U340 filter. The preheat temperature for all measurement cycles was set to 220 or 240 °C, respectively, according to preheat plateau tests and held for 10 s. The signal intensities were calculated from the luminescence signals integrated over the first second of stimulation with the signal measured during the last 5 s subtracted as background. To calculate the annual dose derived from the decay of lithogenic radionuclides in the sediment the concentration of uranium, thorium, and potassium was determined by laboratory gamma-spectrometry and/ or inductively coupled plasma–mass spectrometry (ICP-MS, Preusser and Kasper, 2001). A water content variation of 0,5 to 3,5 weight-percent was assumed. The cosmic dose contribution was calculated according to the sampling depth. It was assumed, that the thickness of the overburden sediments which serve as a shielding did not change throughout the burial time.

Palaeo-climatic evidence of soil development on Sahelian ancient dunes

95

The dating of the samples from profiles 3, 5 and 6 is described in detail in Mauz and Felix-Henningsen (2005). 5.3 STUDY SITES, DUNE AGES AND MORPHOLOGICAL SOIL CHARACTERISTICS 5.3.1 Northern Sahel, Mauretania (Site 1) In the Northern Sahel of Mauritania flat, shield-shaped ancient dunes are widely distributed in valleys and form ramps, stretching towards the valleys mainly on the lee-side of cuestas and inselbergs, which expose Carboniferous sandstones. In this area, intensive grazing under recent climatic conditions has generated an extensive mobilization of the ancient dunes. A site without such characteristics of desertification, situated about 70 km east of the village Kiffa (16°30'21,0" N, 10°46'20,8"W, 206 m asl), was investigated in a middle slope position of a weakly inclined ancient dune ramp. The investigated soil profile represents a shallow (Chromi-)Cambic Arenosol, which developed from the Upper Middle Holocene until today. The OSL age of 3,74 ± 0,31 ka of a sample from 60–70 cm depth indicates an end of dune deposition at about 3,5 ka BP. Since that time a 15 cm thick Ah horizon developed, followed by a reddish brown (7,5 YR 4/4) Bw horizon down to 80 cm depth. 5.3.2 Northern Sahel, Niger (Site 2) In the Northern Sahel of Niger flat shield-shaped ancient dune ridges are wide-spread. They also form valley fillings and dune ramps at cuestas. In contrast to Mauritania and Chad, the soils on those ancient dunes are thick and intensively red-brown coloured ChromiCambic Arenosols. Due to intensive grazing, exposed sites in the crest of dune ridges are especially subject to re-mobilization of the dune sand. A site was investigated 5 km SW of the village Abalak, situated between Tahoua and Agadez (15°25'54,4"N, 6°14'32,4"E, 408 m asl). Here, ancient dune plateaus and ridges cover the bottom of broad valleys between shallow cuestas. The profile of the ChromiCambic Arenosol was exposed at the cut bank of a wadi. Shallow erosion channels at the surface in the surrounding of the profile site indicate that, originally, the soil was thicker and stronger, weathered soil horizons were removed. Soil erosion under dryer climatic conditions led to a flattening and dissection of the dune relief. A sample from 240 cm depth shows an OSL age of 10,48 ± 0,73 ka and indicates an Upper Pleistocene age of the dune and an end of sedimentation in the transition period between Pleistocene and Holocene. 5.3.3 Central Sahel, Niger (Site 3) In the Central Sahel of Niger, north and northeast of the capital Niamey, extensive ancient eolian sand sheets and shallow dune ridges cover red sandstones and gravel beds with ferricretes of Tertiary age (Continental Terminal). The dunes are the parent material of Chromi-Luvic Arenosols. The higher precipitation in this region causes land use of agricultural fields mainly for the cultivation of sorghum. A soil profile of a Chromi-Luvic Arenosol was investigated 10 km north of Niamey (13°36'N, 02°05'E, 235 m asl). The soil is characterized by a > 1 m thick red-brown (5 YR 6/8) and weakly indurated Bwt horizon. Within this horizon thin dark redbrown clay bands occur as a consequence

96

Peter Felix-Henningsen et al.

of lessivation. An OSL sample, taken at 150 cm depth displays an age of the top layer of 9,9 ± 1,2 ka and indicates that the sedimentation ended in the transition period between Pleistocene and Holocene. 5.3.4 Southern Sahel, Niger (Site 4) Southwest of Niamey the relief is dominated by dissected plateaus and ridges of sedimentary rocks with lateritic crusts of the Continental Terminal (Tertiary). The bottoms of wide trough valleys are partly covered with ancient dunes, especially in lee-positions. In the deeper positions the valleys are dissected by wadis. The occurrence of the dunes at a lesser density as compared to the region north of Niamey, indicates the southern margin of the ancient dune distribution. While grazing mainly occurs on the poor and shallow soils of the rocky plateaus, the ancient dunes of the valley bottoms are frequently used by agriculture. A profile was investigated of a thick, red brown Chromi-Luvic Arenosol, exposed at the cut bank of a wadi, about 50 km southwest of Niamey near the village Kobadi (13°12'39,4"N; 01°51'34,9"E, 223 m asl). Three OSL dates were obtained from this profile. A sample from 280–290 cm depth shows an age of 29,31 ± 4,93 ka, which indicates the beginning of the eolian sand deposition. A sample from 170–190 cm depth shows an age of 17,34 ± 1,28 ka, while a sample from 110–115 cm depth has an age of 11,87 ± 1,42 ka, indicating that the eolian sedimentation terminated at the end of the Upper Pleistocene. With such a low number of samples it cannot be proven that eolian sedimentation occurred continuously under an arid climate between 30 and 10 ka BP. In this profile, however, as well as in other investigated profiles of that region, no interruption in sedimentation caused by a phase of fluvial erosion or formation of a palaeosol was detected. 5.3.5 Northern Sahel, Chad (Sites 5 and 6) The landscape is characterized by ancient longitudinal dunes forming plateau-like flattened ridges of more than 40 meters relative altitude and of a width from 1–2 km. The plateaus and rather steep slopes of the dunes are sparsely vegetated by grass and trees, mainly Acacia species. The soil surface is covered by a layer of young, weakly-weathered, mobile sand. A strew of stone artefacts and fragments of pottery indicates that the dunes were settled in former times. The interdunal valleys are up to some hundred meters wide and filled with dark grey colluvial flood sediments, rich in fines (about 60 mass-%) and organic matter (about 0,5–1 mass-%). At two sites soil catenas, each with several profiles, were studied and dated by OSL (sites 5 and 6 in Mauz and Felix-Henningsen, 2005). In order to compare the Sahelian soil formation on ancient dunes of different ages, the two sites from Chad are included in this study. Profile 5, situated 22 km north of Moussoro (13°50'31"N; 16°28' 43"E, 298 m asl) and profile 6, 30 km west of Moussoro (13°37'56"N; 16°11'52"E, 289 m asl), are both situated in a flat plateau position on top of ancient dune ridges. At two sites soil catenas, each with several profiles were studied and dated by OSL (sites 5 and 6 in Mauz and Felix-Henningsen, 2005). OSL dating of dune sand from a weakly developed Cambic Arenosols, situated at both sites in a flat plateau position on top of the dune ridges, reveals Middle Holocene ages. A sample from 90 cm depth in profile 5 displayed an age of 3,4 ± 0,2 ka while a sample from 80 cm depth in profile 6 had an age of 3,65 ± 0,74 ka. The dark grey (10 YR 3/3, moist) topsoil of a weakly developed Cambic Arenosol is covered by a thin layer of loose sand from redistribution of aolian sand in the course of desertification. A dark brown (7,5 YR-10 YR 4/3, moist) Bw horizon changes gradually at

Palaeo-climatic evidence of soil development on Sahelian ancient dunes

97

0,8 to 0,9 m depth into a yellow C horizon. Krotowinas, filled with humic topsoil material, and root channels are frequently developed within the Bw horizon. The Ah, which also contains fragments of pottery, and the upper part of the Bw horizons are strongly indurated, due to accumulation of salts and amorphous silica, while a weaker cementation of the lower part of the Bw horizon gradually decreases towards the loose C horizon. The upper soil horizons in profile 6 were disturbed by a prehistoric pit, refilled with sand containing fragments of pottery and bones. As the dating of two samples from the refilled material displayed ages of 900 and 400 years, respectively, the younger age may reflect the minimum age of the soil disturbance. 5.4 PHYSICAL AND CHEMICAL SOIL CHARACTERISTICS The morphological differences between the soils on ancient dunes of different ages are also reflected by their physical and chemical properties. The analytical data of the physical and chemical soil properties are presented in table 1. Correlation coefficients in the text refer to sample numbers of n = 20 for the soils on Pleistocene dunes and n = 24 for soils on Middle Holocene dunes. 5.4.1 Redness Ratio According to Schwertmann et al. (1983) the Redness Ratio is significantly correlated with the proportion of hematite in the total amount of pedogenic oxides. The formation of hematite is an indicator of high weathering intensity due to a relatively high rate of dissolution of silicates, which leads to a fast release and oxidation of iron followed by the formation of ferrihydrite as the precursor of hematite. A highly significant but rather loose correlation (r = 0,58**) exists between redness ratio and the content of pedogenic iron oxides (FeDCB). Therefore the depth and intensity of rubefication (RR values) of the Chromi-Cambic-Arenosols of the Northern Sahel is weaker developed than in the Central or Southern Sahel (Table 1). 5.4.2 Amounts of fines and pedogenic oxides The source of the fines is mainly dust from the Sahara, which sedimented together with the dune sand in arid periods. Furthermore, dust deposition in the Sahel is very effective during the dry season and supplies the ecosystems with bases and nutrients (Stahr et al., 1994; Herrmann, 1996). Apart from silicates the dust contains carbonates and soluble salts (Herrmann et al., 1994). After deposition of dusts on the soil surface or on the vegetation, which acts as a filter, the fines are washed into the soil horizons with the infiltrating rain water. While the amount of clay <2 μm is very low, the fines < 63 μm mainly consist of the silt fraction richer in primary silicates with iron and aluminium, which weathered to secondary clay minerals and pedogenic oxides. The amount of fines correlates significantly and rather closely with the amount of pedogenic Fe and Al oxides (fines/FeDCB r = 0,72**; fines/AlNaOH r = 0,72**). The Arenosols on ancient dunes of Middle Holocene age in the Northern Sahel have lower amounts of fines and pedogenic oxides than the soils on Pleistocene dunes. This reflects a shorter period of dust deposition and a lower intensity of weathering, due to a shorter humid period available for weathering since the Middle Holocene. Lacustrine sediments of palaeo-lakes, which cover large areas of the Sahara of East Niger (Baumhauer et al., 1989), are an important source for Sahelian dust

depth (cm)

RR

pH H2O

Fitnes <63μm mas.%

TDS (mg kg−1)

CaSO4 (mg kg−1)

3 4 3 3 2 1

5 5 6 5

Profile 2: Chromi-Cambic Arenosol—Northern Sahel, Abalac (Niger), 385 mm * a−1 Aa 23 2,08 8,01 14 169 96 10 Bw1 65 3,33 7,70 15 200 110 10 Bw2 110 2,50 7,77 14 170 77 11 Bw3 153 1,67 7,73 14 170 60 6 BwC 194 1,67 7,83 7 162 77 13 Cg > 250 0 7,71 8 106 21 9

Profile 3: Chromi-Luvic Arenosol—Central Sahel, north of Niamey (Niger), 480 mm * a−1 Ah1 22 5,00 7,76 8 25 4 0 Bwt1 53 3,75 7,28 12 11 2 3 Bwt2 86 3,75 6,48 14 17 8 2 Bw > 140 3,13 6,21 14 29 9 4

FeDCB (g kg−1)

2 3 3 2

MgCl2 (mg kg−1)

19 13 12 12

Site 1: Cambic Arenosol—Northern Sahel, Kiffa (Mauritania), 270 mm * a−1 Ah 15 0 7,14 16 207 108 BwAh 35 2,50 7,31 11 115 21 Bw 80 1,67 7,49 11 47 14 BwC 100 0 7,44 10 54 10

Horizon

9 13 9 9

7 6 6 6 7 4

6 7 7 6

SiNaOH (g kg−1)

5 7 6 6

4 3 3 3 3 3

2 3 4 4

AlNaoH (g kg−1)

19 25 21 20

14 12 12 12 12 8

10 13 14 12

Sum. ped. Ox. (g kg−1)

Table 1. Contents of fines and soil chemical characteristics of the representative soil profiles of the Sahel in Mauritania, Niger and Chad; RR: Redness ratio; TDS: total dissolvable salts; FeDCB: pedogenic iron oxides; SiNaOH, AlNaOH: amorphous Silica and Al-oxides.

98 Peter Felix-Henningsen et al.

4 4 4 3 4

5 4 4 4 3

1 1 1 1 1

Profile 5: Cambic Arenosol—Northern Sahel, 22 km north of Moussoro (Chad), 300 mm * a−1 Ah1 7 0 8,45 14 680 149 8 Ah2 15 0 8,16 4 172 25 5 Bw1 50 3,75 8,46 6 148 4 1 Bw2 70 3,13 8,25 6 35 3 1 C 100 0 8,02 3 33 5 0

Profile 6: Cambic Arenosol—Northern Sahel, 30 km west of Moussoro (Chad), 320 mm * a−1 Ah 20 0 8,16 6 303 70 7 1 Bw1 60 0 7,97 4 100 8 3 1 Bw2 80 4,17 8,02 5 72 4 1 1 Bw3 100 3,13 8,30 3 57 6 2 <1 C 120 0 8,01 3 34 7 2 <1

3 6 7 7 6 4 3

9 12 12 12 12 10 9

Profile 4: Chromi-Cambic Arenosol—Southern Sahel, Kobadi (Niger), 551 mm * a−1 Ah 20 2,08 6,37 9 77 7 14 Bw1 65 4,17 4,79 19 546 6 93 Bw2 110 4,17 5,30 21 171 7 0 Bw3 150 3,13 5,08 20 140 2 0 BwC1 200 3,13 5,53 21 345 2 0 BwC2 240 3,13 6,32 20 4157 1481 0 C 300 3,13 6,37 17 2483 832 0

1 1 1 2 1

1 1 1 1 1

3 7 8 9 8 5 4

7 6 6 6 4

6 6 6 5 6

15 25 27 28 26 19 16 Palaeo-climatic evidence of soil development on Sahelian ancient dunes 99

100

Peter Felix-Henningsen et al.

(Herrmann, 1996). Because the lake sediments are rich in diatomite, amorphous silica is also a constituent of eolian dust and was deposited on the soils of the Sahel. Under conditions of high temperatures and high humidity, silica was dissolved and leached. Therefore in the strongly leached Arenosols on Pleistocene dunes, no correlations exist between amorphous silica and other soil constituents. In the weaker leached soils on Middle Holocene dunes, however, highly significant and rather narrow correlations occur between amorphous silica and the total amounts of soluble salts (r = 0,79) and fines (r = 0,77). The weathering of silicates was obviously not a main source for the release of amorphous silica. This can be explained by the high proportions of amorphous silica on the sum of pedogenic oxides in the weakly developed soils and the missing correlation between amorphous silica and pedogenic iron oxides. 5.4.3 Soluble salts Apart from weatherable silicates, the dust contains carbonates and soluble salts. All soils presented in this study are free of carbonates, but they contain salts of different solubilities. A part of the salts, as sulphates and chlorides, results from the deposition of more or less saline dust from the Sahara, while a part of the nitrates was formed in situ due to mineralization of organic matter. Therefore the proportion of nitrates on the bulk of soluble salts increases from the Northern to the Southern Sahel. The salts are dissolved, leached or redistributed within the soil profile according to their solubility product and the annual precipitation. The correlation between the amount of fines and the concentrations of sulphates (r = 0,69**), chlorides (r = 0,66**) and nitrates (r = 0,73**) is highly significant in the weakly developed Arenosols of the Middle Holocene dunes of the Northern Sahel. In contrast these soil compounds show no correlation in the deeper weathered and strongly leached Arenosols of the Pleistocene dunes. Under higher humidity, salts of different solubilities are affected by leaching and redistribution within the soil. Therefore sulphates, chlorides and nitrates display, only in deeply weathered Arenosols on Pleistocene dunes, a very narrow and highly significant inter-correlation (r > 0,9**), whereas they show only weak inter-correlations (r < 0,5**) in Arenosols on Middle Holocene dunes. This is confirmed by the depth distribution of salt forms (Figure 2). The Arenosols of the Northern and Central Sahel are enriched with bases and display weak to moderate alkaline conditions. While leaching of bases under humid climatic conditions leads to acidification of soil horizons, with the consequence of enhanced weathering intensity of silicates, impeded leaching under drier climatic conditions leads to an enrichment of bases and neutral or alkaline pH values. Due to the labile nature of the soil pH in the Sahelian environment, the pH values are mainly independent from the dune and soil age and indicate the balance between enrichment and leaching of bases. The Arenosols on Middle Holocene dunes of the Northern Sahel in Chad especially show a strong increase in pH values, due to portions of highly alkaline Na2CO3 or NaHCO3 from palaeo-lake sediments. The Arenosol of the Central Sahel shows weakly alkaline conditions only in the upper horizons, while the Arenosol in the Southern Sahel is acidic throughout the profile. Because weathering of silicates is only possible and effective under acid conditions, the alkaline pH values of the soils in the Northern and Central Sahel show that the weathering of silicates must have occurred under more humid climatic conditions, when the concentration of bases was very low, due to strong leaching of salts. Obviously weathering of silicates does not continue at present. We conclude that the processes of fossilization by the deposition of sand and dust, as well as by the cementation of the upper soil horizons by amorphous silica and soluble salts, occurred at a rather fast rate and can be attributed to the recent semi-arid climate. But the

Palaeo-climatic evidence of soil development on Sahelian ancient dunes

101

Figure 2. Depth distribution of salts of different solubilities (left column: 2–200 g L−1, middle column: 200–500 g L−1, right column: >500 g L−1), profile 1: Cambic Arenosol, Northern Sahel, Mauritania; profile 2: Chromi-Cambic Arenosol, Northern Sahel, Niger; profile 4: Chromi-Cambic Arenosol, Southern Sahel, Niger. Location of the profiles see figure 1.

relative high amounts of amorphous silica and soluble salts in the refilled soil sediment (bABw horizon, Figure 3) indicate that the accumulation of these elements in the topsoil horizons may have previously existed. This would mean that the underlying brown 3bBw horizon must have developed, as in profile 15, during a long period with more humid conditions than today, when salts and carbonates were completely leached and the pH was below 7.

102

Peter Felix-Henningsen et al.

5.5 PALAEO-CLIMATIC INTERPRETATION The thicknesses, morphological, physical and chemical characteristics of the soils on ancient dunes are related to the humidity of the climate, and the duration of the humid period suitable for soil formation. The OSL ages of the ancient dunes give references of the maximum age of the soils, because sustainable soil development started with the end of eolian sedimentation. The real date of the end of sedimentation can only be approximated because the sampling depth was below the Bw horizons, which had been influenced by mixing due to bioturbation and rooting. Therefore it is not possible to use the youngest sediment layers for dating. Although the number of OSL data of ancient dunes of the Sahel in Mauritania, Niger and Chad is very limited and cannot be interpreted with respect to a general ancient dune stratigraphy for the entire Sahel, nevertheless they reflect evidence that only two ancient dune generations existed, which can be clearly subdivided by the thickness and weathering intensity of the soils. The sedimentation of the Upper Pleistocene dunes of the Sahel in Niger started at least around 30 ka and ceased at ∼10 ka with the transition from the Upper Pleistocene to the Holocene, while the Middle Holocene dunes in Chad and Mauritania were sedimented between ∼5 and ∼3,5 ka. The existence of two dune generations in the Sahel is in concordance with previous results of investigations in the Sahel in Sudan (Warren, 1970), the Sahara and Northern Sahel in Chad (Felix-Henningsen and Mauz, 2005), and the Northern Sahel and Southern Sahara in Niger (Völkel, 1988, 1989; Völkel and Grunert, 1990; Felix-Henningsen, 2000). With respect to older stratigraphical conclusions of French researchers, the data show that an Upper Pleistocene humid period (Ghazalien), which was assumed between 40.000 and 20.000 BP by Servant (1983), must have ceased before ∼30 ka if indeed such a humid period occurred. From the Sahara and the Sahel, Gasse et al. (1990) report an abrupt transition from arid to humid conditions at > 12 ka and at ∼9,3 ka, with a reversal (from humid to arid) at ∼10,5–10 ka. The climate changes took place within a few centuries or even decades. The latest eolian activity is confirmed by the OSL data of the ancient dunes in Niger. Servant (1983) dates the Late Pleistocene to Early Holocene humid period (Tchadien) between 14.000–7.500 BP, with a maximum of humidity between 10.000 and 7.500 BP. In view of the OSL data of the ancient dunes, this would mean that between 14.000 and ∼10.000 BP the climate of the Sahel in East Niger was only semi-arid and the low intensity of soil development still allowed sedimentation or local redistribution of eolian sand. The ages of the optically dated palaeo-dunes of Chad and Mauritania indicate that the uppermost sand layers were deposited in the Mid Holocene between ∼5 ka and ∼3 ka. The southernmost ancient dunes at Site 5 show the youngest age of 3,1 ka BP. If the ages of the “Nouakchottien” humid period (between 4.500 and 3.000 BP) provided by Michel (1973) are reliable, then the OSL data of the Middle Holocene ancient dunes in Chad, Mauritania and the Central Sahara (Felix-Henningsen, 2000) show that sufficient humidity existed in the Sahel after ∼3.500–3.000 BP, which led to fixation of dunes by vegetation and soil formation. The fact that palaeosols of this humid period occur in the Southern to the Central Sahara (Völkel, 1988, 1989; Felix-Henningsen, 2000) proves that the climatic conditions since the Middle Holocene humid period must have also changed in the Sahel to more arid conditions. The depth and intensity of weathering of the soils seem to be associated with the stratigraphical unit of the dune generation. Due to the accumulation of the processes of soil formation in the Early and the Mid Holocene humid periods until present, on Upper Pleistocene dunes the rubefication, accumulation of fines and leaching depth of soluble salts is much more pronounced than on Middle Holocene dunes, which were only affected

Palaeo-climatic evidence of soil development on Sahelian ancient dunes

103

by weathering from the Mid Holocene until present. In soils of both dune generations the latest climatic changes are expressed in the physical and chemical soil characteristics. The weathering of silicates indicates that acid conditions must have prevailed for a sufficient time. Especially in the rubefied soils on Upper Pleistocene dunes, the intensity of weathering was high due to intensive leaching and acid conditions, because the rate of iron release by weathering of silicates is one important factor in the formation of hematite (Schwertmann, 1985). Under recent climatic conditions, leaching of bases, which are deposited with dust, is limited. Therefore the pH values of soils in the Northern Sahel are neutral to alkaline and in the Central Sahel neutral in the uppermost part. Consequently, under recent climatic conditions with annual precipitation up to <400–500 mm * a−1 and deposition of dust containing carbonates and soluble salts, the progress of weathering of silicates is inhibited. Only under higher precipitation conditions in the Southern Sahel the weathering of silicates proceeds due to acid soil conditions. Therefore the soils on ancient dunes of the Northern and Central Sahel can be assumed to be relict soils, which mainly weathered in periods with conditions of precipitation greater than ∼500 mm * a−1 during the Early and Middle Holocene humid periods. 5.6 CONCLUSIONS From initial OSL datings of ancient dunes of Pleistocene and Holocene age, it can be assumed that soil properties and the degree of soil development contain stratigraphical information related to the age of the dunes due to the duration of humid periods. Weathering of silicates is a relic feature in soils on ancient dunes of the Northern and Central Sahel. Therefore soils on ancient dunes in these eco-zones are mainly palaeosols. Correlating relict soil properties with recent annual precipitation data for a climatic gradient from the Northern to the Southern Sahel, shows that humid periods with an annual precipitation of more than 500 mm * a−1 are necessary to generate acid conditions with weathering of silicates. Apart from stratigraphical and palaeoenvironmental aspects, the results may have practical implications for land-use planning and for combating desertification. A common process of desertification as a consequence of overgrazing is that soil fertility decreases, whereas soil stability against deflation improves with the duration, depth and intensity of soil development. ACKNOWLEDGEMENTS Fieldwork in the Sahara was only made possible by the professional organisation and preparation of the expedition by R. Baumhauer and B. Sponholz (University of Würzburg). We are grateful for their active support in the field and for many fruitful interdisciplinary discussions. The generous funding of the project by the German Research Society (DFG) is gratefully acknowledged. REFERENCES Aitken, M.J., 1998, An introduction to optical dating—the dating of Quaternary sediments by the use of photon-stimulated luminescence. Oxford University Press; pp. 1–267. Baumhauer, R., Busche, D. and Sponholz, B., 1989, Reliefgeschichte und Paläoklima des saharischen Ost-Niger. Geographische Rundschau, 41, pp. 493–499.

104

Peter Felix-Henningsen et al.

Boetter-Jensen, L., McKeever, S.W.S. and Wintle, A.G., 2003, Optically stimulated luminescence dosimetry, (Elsevier), pp. 1–404. Felix-Henningsen, P., 1983, Zur Genese und Vergesellschaftung von Böden auf den Altdünen der nördlichen Goz-Region im Sudan. Mitteilungen der Deutschen Bodenkundlichen Gesellschaft, 38, pp. 485–490. Felix-Henningsen, P., 1984, Zur Relief- und Bodenentwicklung der Goz-Zone Nordkordofans im Sudan. Zeitschrift für Geomorphologie, N.F., 28, pp. 285–303. Felix-Henningsen, P., 2000, Palaeosols on Pleistocene dunes as indicators of palaeomonsoon events in the Sahara of East Niger. Catena, 41, pp. 43–60. Felix-Henningsen, P. and Mauz, B., 2005, Palaeoenvironmental significance of soils on ancient dunes of the Northern Sahel and Sahara of Chad. Die Erde, 135, pp. 321–340. Gasse, F., Téhet, R., Durand, A., Gilbert, E. and Fontes, J.C., 1990, The arid-humid transition in the Sahara and the Sahel during the last deglaciation. Nature, 346, pp. 141–146. Gläser, B., 1987, Altdünen und Limnite in der nördlichen Republik Sudan als morphogenetisch-paläoklimatische Anzeiger. Untersuchungen zur jungquartären morphogenetischen Sequenz eines Regionalkomplexes. (Hamburg: Akademie der Wissenschaften in Göttingen), pp. 1–193. Grunert, J., 1988, Verwitterung und Bodenbildung in der Süd-Sahara, im Sahel und im Nord-Sudan. Mit Beispielen aus Niger, Burkina Faso und Nord-Togo. In: Hagedorn, J., Mensching, H.G., (Eds.), Aktuelle Morphodynamik und Morphogenese in den semiariden Randtropen und Subtropen. Abh. Akad. Wissensch. in Göttingen, 41, (Göttingen: Vandenhoek & Ruprecht), pp. 22–43. Herrmann, L., 1996, Staubdeposition auf Böden Westafrikas. Hohenheimer Bodenkundliche Hefte, 36, pp. 1–239. Herrmann, L., Sponholz, B. and Stahr, K., 1994, Quellregionen für den Harmattan-Staub in Westafrika: Ein mineralogischer und geochemischer Ansatz. Mitteilungen der Deutschen Bodenkundlichen Gesellschaft, 74, pp. 367–370. Krbetschek, M.R., Götze, J., Dietrich, A. and Trautmann, T., 1997, Spectral information from minerals relevant for luminescence dating. Radiation Measurements, 27, 5–6, pp. 695–748. Mauz, B. and Felix-Henningsen, P., 2005, Palaesols in Saharian and Sahelian Dunes of Chad: Archives of Holocene North African climate changes. The Holocene, 15, pp. 453–458. Michel, P., 1973, Les bassins des fleuves Sénégal et Gambie. Etude géomorphologiques. Mém. de l´O.R.S.T.O.M. 63, (Paris), pp. 1–752. Murray, A.S. and Wintle, A.G., 2000, Luminescence dating of quartz using an improved single-aliquot regenerative-dose protocol. Radiation Measurements, 32, pp. 57–73. Pfeiffer, L., 1991, Schwermineralanalysen an Dünensanden aus Trockengebieten mit Beispielen aus Südsahara, Sahel und Sudan sowie Namib und der Taklamatan. Bonner Geographische Abhandlungen, 83, pp. 1–216. Pfeiffer, L. and Grunert, J., 1989, Heavy mineral associations in dune sands and soils of the Sahel and the Sudan (Niger, Burkina faso, Togo). Palaeoecology of Africa, 20, pp. 55–68. Preusser, F. and Kasper, H.U., 2001, Comparison of dose rate determination using highresolution gamma spectrometry and inductively coupled plasma-mass spectrometry. Ancient TL, 19, 1, pp. 19–23. Schlichting, E., Blume, H.P. and Stahr, K., 1995, Bodenkundliches Praktikum. Pareys Studientexte, 81 (Berlin: Blackwell). Schwertmann, U., 1985, The effect of pedogenic environments on iron oxides minerals. J. Soil Sci., 1, pp. 172–200.

Palaeo-climatic evidence of soil development on Sahelian ancient dunes

105

Schwertmann, U., Fechter, H., Torrent, J. and Alferez, F., 1983, Quantitative relationships between soil colour and hematite content. J. Soil Sci., 136, pp. 354–358. Servant, M., 1983, Séquences continentales et variations climatiques: Évolution du Bassin du Tchad au Cénozoique supérieur. Trav. et Doc. de l´O.R.S.T.O.M. 159, (Paris), pp. 1–573. Smettan, U. and Blume, H.-P., 1987, Salts in sandy desert soils in South-Western Egypt. Catena, 14, pp. 333–343. Stahr, K., Herrmann, L. and Jahn, R., 1994, Long distance dust transport in the SudanoSahelian zone and the consequences for soil properties. In: Buerkert, B., Allison, B.R. and Oppen, M. von (Eds.): Proceedings of the International Symposium “Wind-erosion in West Africa”, pp. 23–33. Völkel, J., 1988, Zum jungquartären Klimawechsel im saharischen und sahelischen OstNiger aus bodenkundlicher Sicht. Würzburger Geographische Arbeiten, 69, pp. 255–276. Völkel, J., 1989, Geomorphologische und pedologische Untersuchungen zum jungquartären Klimawechsel in den Dünengebieten Ost-Nigers (Südsahara und Sahel). Bonner Geographische Abhandlungen, 79, pp. 1–258. Völkel, J. and Grunert, J., 1990, The problem of dune formation and dune weathering during the Late Pleistocene and Holocene in the Southern Sahara and Sahel. Zeitschrift für Geomorphologie, N.F., 34, pp. 1–17. Warren, A., 1970, Dune trends and their implications in the Central Sudan. Zeitschrift für Geomorphologie, N.F., 10, pp. 154–180. Wintle, A.G., 1997, Luminescence dating: laboratory procedures and protocols. Radiation Measurements, 27, 5–6, pp. 769–817.

CHAPTER 6

Are there valuable pedological palaeoenvironmental indicators in Northern Chad? Ludger Herrmann, Mohamed Mounkaila and Frieder Graef Institute of Soil Science and Land Evaluation, University of Hohenheim, Stuttgart, Germany ABSTRACT: The paper reports on field observations and analytical results from an expedition to Northern Chad. The objective is to evaluate on the basis of these data, which landscapes or surface types potentially contain palaeopedological information, and which proxies/approaches can be used for their exante detection in order to more efficiently direct field research. In the first part, in a regional approach, 56 samples are evaluated, which were chosen based on remote sensing data (surface type distinction). The underlying hypothesis is, that zonal soil formation gradients should exist, and that samples exhibiting extreme values contain additional (palaeo?) information. In the second part potential information of lacustrine and dune areas is discussed. Finally, a palaeosoil detected by remote sensing information is presented in detail. The following conclusions can be drawn from the study: (1) Fe-components can be helpful for the exante remote sensing detection of palaeopedologically relevant sites in the visited region. (2) The probability to find exploitable dune records is highest in the Sahelian domain. (3) The area between the Angamma escarpment and Ain Galaka is assumed to bear a high potential for palaeopedological information at the landscape scale.

6.1 INTRODUCTION There is only cumbersome recent scientific information on Northern Chad emerging during the last decades. This is not due to a lack of interesting scientific phenomena but depending on political instability, war and its remnants, which make the area between Lake Chad and Tibesti hardly accessible. Only few expeditions covered the area since the mid-seventies, under which was a joint one with a multidisciplinary team from Würzburg, Trier, Giessen and Hohenheim Universities in 1997. The expedition followed the Bahr el Ghazal northward, heading to the oasis Faya-Largeau, then turned to the west bound for the Angamma escarpment and finally entered into the Bodele Depression southward. One major aim of the expedition was to discover Holocene palaeo-climatic records like continuous lake sediments, fulgurites as lake level indicators etc. One major finding was, that the area visited is at present dominated by wind erosion and deflation as documented i.e. by frequent yardangs, which make complete sedimentary palaeo-records improbable. However, the question remained, whether there is other exploitable palaeoenvironmental information. In the first part this article tries to find indicators for palaeopedological information based on a rather statistical approach. The hypotheses behind are the following. In Northern Chad zonal soil formation gradients should be detectable due to increasing aridity towards the North. On the other hand, at present this region is an immense deflation area, where surface properties tend to be homogenised by repeated eolian entrainment and deposition.

108

Ludger Herrmann et al.

Due to representative sampling based on remote sensing information, these trends should be discernable. On the other hand, extreme outliers could contain valuable additional, i.e. palaeo-information. The second part reports about a palaeosoil formed on a river dam north of the Angamma escarpment (18°N), which was discovered based on remote sensing information. 6.2 SHORT DESCRIPTION OF THE VISITED ENVIRONMENT The region visited (Figure 1) stretches from the Sahel close to Lake Chad (ca. 13,5°N) to the desertic footslopes of the Tibesti (ca. 18°N). The East–West extension is determined by the accessible dust roads followed from N’Djamena over the Bahr el Ghazal to Faya Largeau, and over Ain Galaka to the Angamma escarpment (16–18,5°E). The N–S transect corresponds to a climatic gradient with approximately 300 mm annual rainfall in the South (unimodal) down to < 50 mm (erratic) in the North at above 26 °C average annual temperature (bimodal in the South, unimodal in the North). The relief energy is low with elevations of 200–400 m a.s.l. However, relief features like dunes or escarpments can reach several decametres. Petrography is dominated by eolian sands, lacustrine sediments (including diatomite), and (coarse) sandstones, the former of Holocene, the latter of Tertiary age. With respect to surfaces, three major types can be distinguished: dune, lacustrine, and serir surfaces. For description of the vegetation refer to Schulz (2000). At the southern end of the transect vast plains with grey to dark grey silty to clayey sediments are prominent. These represent deposits of Lake Chad and give an indication of its former size. At present they undergo pedogenesis, leading to vertic properties and organic matter accumulation. Partially, thin sheets of floating sand cover the surfaces. Between approximately 13,5–14,5°N longitudinal dune fields with an NW–SE orientation occur. In between the dunes again dark grey clayey sediments occur which are due to their mineralogical properties attributed to lacustrine sediments that were fluvially redistributed and partially mixed with dune sands. Further north, along the Bahr el Ghazal, in fact an overflow channel of Lake Chad, which drained towards the Bodele Depression in former times, levelled plains partly covered with achabs are present. Extended spots of whitish lacustrine sediments occur there. Locally yardangs can be detected. Farther north (ca. 16–17°N) lies the Erg de Djourab predominantly characterised by dunes of barchan type and sand sheets. However, even between the latter diatomites and swamp orees can be detected, which hint towards former lake presence. The same is true for the vast plain lying northwards, which is fringed by the Bodele Depression in the West. There, surfaces of serir-type dominate. They tend to get coarser towards the North. Sand sheets and even sandstone outcrops occur also. Approaching Faya Largeau sandstone escarpments appear to the West limiting the Bodele Depression at its eastern fringes. Serir surfaces in this area have often a dark red, dark brown or even black appearance, which we assume to be related to the (Fe-rich?) sandstone outcrops. The oasis Faya Largeau lies in a small depression in front of a sandstone escarpment. Variable sediments characterize the depression itself, dunes occur as well as lacustrine sediments and thin white carbonate crusts embedded in weakly consolidated siltstone could be observed. At one point lacustrine sediments overlying orange-brown dune sands with an abrupt boundary could be detected. Whether this phenomenon is an indicator of a climate change or only a result of local geomorphologic activity could not be clarified. Between Faya Largeau and the small oasis Ain Galaka the surfaces are very variable with a dominance of lacustrine sediments (spotwise reddish) and sandstone outcrops, which partly interfinger. Connected to the sandstone thin brown iron oxide crusts occur, which partly cover as remnants greater surfaces as serir even on lacustrine sediments. The latter

Are there valuable pedological palaeoenvironmental indicators in Northern Chad?

109

indicates a still high redistribution of material in the landscape, as do yardangs for erosion processes. Since the lacustrine sediments occur north of the escarpment that is thought to limit the Bodele Depression, it is in question whether they developed at extreme high lake levels (overflow) or are local phenomena depending on the geomorphological setting. From Ain Galaka westwards, serir surfaces of different pebble size, composition and colour become dominant. They can be sorted, well rounded, whitish, fine with quartz dominance or coarse, dark with dominance of only locally transported only slightly rounded Fe-sandstone components, or mixtures of all kind of components and sizes. The local physical setting (petrography, geomorphology), fluvial transport and sheet wash seem to determine the serir type present. In the terrain a clear morphological boundary

Figure 1. Sketch of the research area and expedition route in Northern Chad (Samples 1–32 north and south of the Angamma escarpment, 33–39 between 18°E and Faya Largeau, 40–44 between Faya Largeau and Koro Toro, 45–49 between Koro Toro and Nedeley, 50–54 between Nedeley and Moussoro, 55–56 between Moussoro and Massakori).

110

Ludger Herrmann et al.

can be observed at 17°80'N, 17°74'E in form of a river dam which separates a fluvial domain in the West from the terrestrial one in the East. Following the fluvial system to the South the Angamma escarpment can be reached, which separates the fluvial domain from the lacustrine domain of the Bodele Depression in the South. Channels filled with sandy weakly consolidated material are a hint that this fluvial domain represents a river delta, which is supported by recent findings (Drake and Bristow, 2006). The Bodele Depression itself is dominated by vast occurrences of lacustrine sediments. Yardang fields, with single yardangs more than 2 m high, and towards the South a pronounced undulating micro relief, which makes passage with vehicles impossible, mark the ongoing deflation. Since the lacustrine sediments are consolidated, particle entrainment is depending on saltating sand grains delivered by passing barchan dune fields and local sand sheets. The grey skirts made up by lacustrine sediments on the leeward side impressingly show the active role of the barchans. Where sandy material at the surface is absent, sandy fillings in desiccation cracks witness their former occurrence. Partly along the desiccation cracks gley phenomena (Fe-impregnation) can be observed as signs for shallow water to semi-terrestrial conditions. Locally Fe-accumulation is so accentuated that vertically oriented Fe-sheets weather out of the sediment, sharp as knifes and ready to cut vehicle tires. Strong Fe-translocation in the palaeoenvironment is also locally documented at the Angamma escarpment where silty sediments meet Fe-sandstone. There, large (up to 1 m diameter) Fe-impregnated sculptures weather out of the sediment. Also carbonate dynamics (rhizoconcretions) could be detected at several places. South of the Angamma escarpment the fluvial inputs are well established through sediments up to gravel size, which cease towards the South. Very localized, and in state of erosion, weakly developed soil profiles with gley properties (mottling, Mn-concretions) and very shallow B-horizons (aggregation, Fe-oxide formation) document the transition from lacustrine to semi-terrestrial conditions. All the latter findings are indicators for transition zones between aquatic and terrestrial domains. The observations are too erratic to decide whether they witness several shorelines related to discrete lake levels or small islands in a discontinuous lake surface. Both options are proposed by Servant (1983). 6.3 SAMPLING APPROACH AND METHODS Except for the Bodele Depression and the area north of the Angamma escarpment sampling was restricted to sites close to the flagged dirt roads (Figure 1) due to security reasons (mines). The sampling scheme was designed to differentiate as many surfaces as possible (including lacustrine sediments), and was based on spectral properties, as collected by Landsat TM satellite images. In order to create the false colour images the channel combination 7-4-3 (RGB) was used, which had shown to give good results for geomorphologic mapping in semi-arid areas of West Africa (Graef, 1999). Apart from surfaces discernable by spectral properties, local geomorphologic or pedological curiosities were sampled. In total 56 sites (partly catenawise) were sampled and described (FAO, 1990, 1994) with respect to landscape and strata properties. Since surface properties were in focus of the remote sensing study, in general only the first three strata were sampled as mass samples, which resulted in a mean sampling depth of 0,2–0,6 m. We are aware, that this is not the normal pedological approach. However, the high sample number, the representativity due to the satellite-based sampling, plus the information of the lowermost strata should allow for some significant conclusions. A detailed description of the sampling sites and their geographical location can be found in Mounkaila (2006). In the laboratory, the air-dry samples were sieved (2 mm mesh) in order to separate skeleton (> 2 mm) and fine soil. Fine soil was subject to further analysis according to Herrmann (2005).

Are there valuable pedological palaeoenvironmental indicators in Northern Chad?

111

6.4 GRADIENTS OF SURFACE SAMPLE PROPERTIES 6.4.1 Texture With respect to texture no latitudinal gradients could be detected. Sandy and clayey surfaces were present in all landscapes. Figure 2 shows the overall dominance of sandy surface textures (ca. 70%). Sands are omnipresent, even in the lacustrine domain of the Bodele Depression either as dunes, sand sheets or only thin veils on serir-type surfaces. This is a clear sign for the active redistribution of materials under the given environmental conditions. Main cause is eolian activity, which in its intensity leads to the fact that the areas south of the Tibesti are net exporters of fines (mineral dust), which are transported with the Harmattan in south-westerly directions (Herrmann et al., 1999). About 10% of the samples are very fine and have a maximum either in the clay or fine silt fraction. These are in their origin lacustrine samples. They were detected on the whole transect, transformed to soils in the Sahel and as deflated surfaces in the Sahara with vast occurrence in the Bodele Depression. Intermediate texture is either typical for moving sand, which has incorporated lacustrine material or mineral dust (coarse to medium silt dominance) accumulated under serir cover. However, the texture distribution as presented in figure 2 cannot be interpreted as representing coverage of textural surface types, especially since skeleton—significant for the serir—is not presented, and sampling was not based on spatial representativity. Typical for the great majority of sampled sites is strong layering as indicated by the comparison of grain size fraction ratios (i.e. medium/fine sand). Exceptions occur only at Sahelian sites on stable sand dunes (Cambic Arenosols) and levelled pedogenically transformed lacustrine sediments (Vertic Gleysols), indicating relative geomorphologic stability. With exception for the lacustrine sites, the second stratum sampled is always enriched in clay and fine silt. This is again a hint towards deflation as dominant recent process impoverishing the uppermost stratum. In conclusion, the granulometric data together with field observation indicate that complete pedological records as in situ developed profiles from rather uniform or

Figure 2. Frequency of sand concentration in surface samples on the visited Sahel-Sahara transect.

112

Ludger Herrmann et al.

reconstructable parent materials are to be expected mainly in the Sahelian environments, especially with dunes. 6.4.2 Kaolinite as a weathering indicator Though kaolinite can also be formed under initial weathering i.e. from feldspars, it is a good indicator for weathering processes if present in higher concentrations. It could be shown, that under Sahelian conditions (560 mm annual rainfall) in SW-Niger freshly deposited smectite (from mineral dust) was quantitatively transformed to kaolinite under the given leaching conditions in Arenosols (Herrmann et al., 2002). On the first look, kaolinite concentrations in the clay fraction are quite scattered over the whole N–S-transect (Figure 3). Looking closer, it can be remarked, that the minimum concentration stays the same in all areas (ca. 15%). On the other hand, maximum concentration increases with an inclination of about 7,5% per degree northern latitude. It remains the question, whether these trends are due to weathering or due to inheritance of the parent rock. Comparing the sites with minimal and maximal values and crosschecking with other data reveals in both groups lacustrine as well as terrestrial sediments and all texture classes occur. Low values occur more often where also carbonates (rhizo-concretions) were detected at the surface. In conclusion, kaolinite cannot be used ex ante as indicator for Holocene palaeoweathering, but must be interpreted in its local context here, i.e. very high kaolinite concentrations (60%) where Fe-sandstone outcrops appear. This does not undermine the interpretative value of kaolinite if the parent material for soil formation is known. 6.4.3 Fe-fractions Since frequent features of Fe-redistribution via the solute pathway on the landscape scale were observed (swamp ores, rhizo-concretions etc.), also Fe-fractions were investigated. The hypothesis behind is, that where high absolute iron accumulation (total iron) or

Figure 3. Kaolinite concentrations (estimated semi-quantitatively from XRD-diffractograms of sedimented samples) in the clay fraction of the top-stratum in dependence of geographical location (latitude N) in N-Chad.

Are there valuable pedological palaeoenvironmental indicators in Northern Chad?

113

Figure 4. Fe-compounds (Fet: rhomb, Fed/t: bullet) of top strata on a latitudinal (N) gradient in N-Chad.

high Fe-transformation (Fed/t ratio) is detectable, pedogenic (palaeo-) processes could be responsible. The results obtained yield a similar picture as already for kaolinite. Low Fe-concentrations (<0,5% Fet) can be detected everywhere (Figure 4), but the maximum values clearly increase towards the north, being around 3% at 13,5°N and rising up to 6% Fet towards the northern ranges (ca. 17,5°N). Highest values are always found in absolute (North and South) or relative (central) low landscape positions. All samples are either pure (North and South) or mixed lacustrine materials (sand with lacustrine flakes in the central part of the transect). The explanation for this fact is that landscape lows receive iron via particle (Fe3+ in clay fraction in the South and in the North partly also as Fe2+ in primary minerals) as well as through solute transport (Fe2+). Higher absolute values in the North could be due to petrographic differences, since the geologically relative young, partly basic magmatites of the Tibesti (Kusnir, 1995) and the detected Fe-sandstones should contain a higher share of Fe-rich minerals than the sandstones of the Ennedi in the East and the old weathered mantles in the South. However, highest concentrations were found in sediments without visible primary mineral contributions. So, although the general background might be higher due to regional petropgraphy, field observations indicate highest values to be related to local sedimentary conditions including absolute enrichment via the solute pathway. The Fed/t ratio is generally low (<0,5, average 0,29) with only few exceptions at the northern end of the transect. Samples with Fed/t > 0,5 occur only in the area between the Angamma escarpment and the oasis Ain Galaka. This is where developed B-horizons, partly with reddish colours, could be detected under serir surfaces. In this context, Fed/t seems to be a valuable indicator of potential palaeosoil occurrence. According to Mounkaila (2006), algorithms exist for the detection of Fet as well Fed via Landsat TM band combinations in this environment. If so, and if the detection has sufficient accuracy, areas with higher potential of palaeosoil occurrence could ex ante be detected based on remote sensing. 6.4.4 pH values The majority of samples lie in the neutral to alkaline range (Figure 5). The northern group (north of 17°N) has a centre between pH 8–9,5 which represents the transition between the

114

Ludger Herrmann et al.

Figure 5. pH (H2O) of top strata on a latitudinal (N) gradient in N-Chad.

carbonate and bicarbonate buffer range. Only few samples are outliers with pH either >10 (bicarbonate buffer) or lower 6,5 (silicate buffer) and concern (mixed) lacustrine materials. The same is true for the low values (pH < 6,5) around 16° northern latitude along the Bahr el Ghazal. Samples with pH around 5,5 represent mixtures of eolian sands with lacustrine sediment flakes, extreme low values with pH < 4 represent pure lacustrine surfaces. South of 15,5°N, entering the Sahelian domain the pH is depending on landscape position, which at the same time means petrography. Here, several catenas have been sampled from the dune top positions (pure sands) towards the interdunal inundation planes (lacustrine sediments, partly fluvially redistributed). Dune sites with Arenosols show neutral pH around 7,5. Landscape lows (Vertic Gleysols to Gleyic Vertisols) are characterized by pHvalues around 9. In conclusion, average pH-values of free draining sites are increasing towards the North with increasing aridity from 7,5 with ca. 350 mm to 9 with <50 mm annual rainfall. Arenosols on dunes in Chad have a one unit higher pH-value than comparable profiles at the same latitude in SW-Niger (compare Herrmann, 1996). This is explained by the fact that rainfall isohyets are shifting towards the South in Chad. Higher pH-values in Sahelian low landscape positions are explained by the endorheic situation and low water conductivity of the sediments. So soluble components either from atmospheric deposition or dune outwash are accumulated here. The extreme values of some lacustrine sites in the Saharan domain need further explanation. Overall, pH in this sample set has low indicative value with respect to potential palaeopedological information. 6.5 INFORMATION FROM LACUSTRINE/FLUVIAL SEDIMENTS It does not need explanation, that lacustrine sediments as such contain information on the recent and/or palaeoenvironment. The mere existence is an indication for water surplus on the landscape scale. Furthermore, the amount and type of biogenic remnants (i.e. radiolarians, diatoms) is indicative for specific environmental conditions (i.e. Baumhauer, 1991). Here, however, we want to concentrate on potential information, which hint towards pedogenic processes, which might have occurred under shallow water or semi-terrestrial conditions at

Are there valuable pedological palaeoenvironmental indicators in Northern Chad?

115

the transition from the aquatic to the terrestrial domain. Therefore, the samples that contain lacustrine material and exhibit extreme (geo-) chemical or mineralogical characteristics are further investigated. Already the first view on the data (i.e. pH in table 1) reveals the local similarity but regional disparity of the sample properties. Samples 12 and 13 are typical for the pure fine bright lacustrine sediments in the Bodele Depression south of the Angamma escarpment. Samples 39a and c are located north of the Angamma escarpment towards the oasis Ain Galaka. 39a does not represent a lacustrine sediment in the narrow sense but a filling in a local depression. Samples 45a and 47b where sampled along the Bahr el Ghazal and are potentially fluvially redistributed and samples 54c (interdunal depression) and 56 (vast plain) were collected in the Sahelian domain. The Bodele samples (12, 13) are characterized by slightly acidic pH, slightly elevated EC, high total Al and Fe, low kaolinite (ca. 20%) and moderate to high (42–61%) smectite concentration. Gypsum—determined through Rietveld analysis from powder XRDdiagram—yields 3–5%. The fine particle share is high (silt + clay = 90%). Diatoms were detected via SEM (see photographs p. 87 in Mounkaila, 2006). With exception for the lower than expected pH-values, these samples can be seen as a kind of standard without important post-sedimentary changes (i.e. gleying, salt accumulation etc.). They represent a lacustrine environment with reducing (grey sediment colour) and neutral to slightly alkaline (Mg-rich smectite, gypsum) conditions. Striking are the differences comparing these with samples from north of the Angamma escarpment (39a and c). Though of different petrography (39a dune sands over sandstone; 39c lacustrine sediments) the latter bear a number of common properties. These are a high pH (>10, bicarbonate buffer), elevated EC and dominance of water soluble Na, high total Na and the highest kaolinite shares in the clay fraction (>40%). In addition, Mn- and Fe-carbonate (rhodochrosite and siderite) could be detected in the lowermost Bwk-horizon of site 39a. Since the sites lie north of the main borderline of the Bodele Depression, the characteristics show strong local influences. The iron-impregnated sandstone occurring to the West should influence the high kaolinite share. The lateral influences on the landscape scale are also documented through the Mn- and Fe-carbonates, which indicate the former existence of Bg-horizons in the former terrestrial surroundings (Fe-/Mn-mobilization and solute transport to the depression). That high pH occurs also in the eolian sand cover of 39a is taken as a hint, that alkalinisation is a post-sedimentary process, since it is out of the zonal pH-range (see figure 5). Here palaeoenvironmental information concerning an aridisation phase might be available, but should be investigated in more detail at the landscape scale. Again, easily distinguishable from the samples before by the extreme low pH (< 4, Al-buffer) are those collected along the Bahr el Ghazal (45a, 47b). Very fine textured, high total Si (39–44%) and LOI (8%) argue for a higher share of diatoms in the sample. The presence of gypsum and texture are comparable to those of samples 12 and 13. So the question is, why are these samples that acid? A hint comes from the chemical characteristics: higher than average water soluble Ca and Mg concentration was detected. An explanation could be, that the lacustrine sediments contained reduced Fe-compounds, which oxidized consecutively and produced sulphuric acid (as known from acid sulphate soils) that was then buffered by dissolution of Ca- and Mg-carbonates. At least, this process could explain low pH as well as high water soluble Ca- and Mg-values at the same time. Also over-average free iron (Fed) values in all these samples hint towards this process. Unfortunately sulphur analytics, which could have helped to support the given hypothesis, was not possible in this framework. Why this process is especially intense in this area is not understood, but might be in relation to geological strata (Pliocene/Pleistocene) with high sulphur concentration (gypsum, Kusnir, 1995). If the given hypothesis is accepted, low pH-values are an effect of post-sedimentary processes during a desiccation phase.

116

Ludger Herrmann et al. Table 1. Characteristics of (mixed) lacustrine materials in N-Chad.

Surface Munsell

pH H 2O

EC μS/cm

Ca (H2O) mval kg−1

Mg

Na

17,45 17,38 17,95 17,95 17,95 17,93 16,02 15,89 13,84 13,84 13,84 13,62

2.5 Y8.2 2.5 Y8.2 7.5 YR7/6

10 YR6/1

5,7 5,2 10,5 10,7 10,3 10,8 3,7 3,9 9,3 8,2 8,5 9,3

512 1015 99 116 8780 2275 502 467 144 2500 800 183

17 119 17 19 2037 530 20 29 28 28 589 35

49 40 <1 <1 7 1 44 25 <1 10 1 <1

28 33 <1 <1 <1 <1 15 11 <1 <1 <1 <1

Depth cm

Nat %

Cat %

Mgt %

Sit %

Alt %

Fet %

Fed/t

12 13 39a1 39a2 39a3 39c 45a 47b 54c1 54c2 54c3 56

>30 >30 −0,5 −7 >7 >4 >30 −10 −3 −12 −40 −1

0,9 1,2 0,3 0,4 4,8 2,4 0,3 0,1 0,4 2,1 1,2 0,2

1,1 0,9 0,1 0,2 0,3 4,6 0,2 0,1 2,7 0,6 0,7 0,8

0,7 0,7 0,1 0,2 0,2 1,1 0,2 0,1 0,5 0,6 0,6 0,6

29,9 30,7 41,1 43,1 37,0 28,2 39,2 44,3 33,9 26,9 27,7 30,0

7,9 7,5 0,6 1,0 1,2 5,1 4,5 0,5 5,5 10,3 9,8 9,3

4,8 4,3 0,4 0,5 0,8 2,7 2,1 0,3 2,4 4,4 5,1 3,2

0,19 0,26 0,50 0,19 0,13 0,04 0,48 0,33 0,08 0,07 0,08 0,03

Sample No.

Depth Kaolinite cm %

Smectite Gypsum Calcite Sand % % % %

Silt %

Clay %

12 13 39a1 39a2 39a3 39c 45a 47b 54c1 54c2 54c3 56

>30 >30 −0,5 −7 >7 >4 >30 −10 −3 −12 −40 −1

57 62 4 4 9 15 37 39 7 15 47 14

42 35 2 2 4 45 63 61 2 34 43 61

Sample No.

Depth Latitude cm N°

12 13 39a1 39a2 39a3 39c 45a 47b 54c1 54c2 54c3 56

>30 >30 −0,5 −7 >7 >4 >30 −10 −3 −12 −40 −1

Sample No.

18 24 42 – – 62 28 35 – – – 17

10 YR8/1 7.5 YR7/2 7.5 YR7.2 10 YR6/1

61 42 32 – – 9 45 36 – – – 47

5 3 0 – – 0 2 1 0 – – 0

0 0 0 – – 17 0 0 4 – – 0

1 3 94 94 87 40 0 0 91 51 10 26

Are there valuable pedological palaeoenvironmental indicators in Northern Chad?

117

The Sahelian samples take an intermediate position, with pH in the carbonate buffer (8–9), low EC in the topsoil increasing with depth, elevated total Al and Fe and an expected kaolinite to smectite ratio. Exceptional are only the low Fed/t ratios. All in all the determined values are in the expected range taking the given climate and landscape position into account. Since the sites are under present soil formation (development of vertic and gleyic properties), pedological palaeoenvironmental information can hardly be expected. 6.6 ARE THERE VALUABLE DUNE RECORDS? Since this question is treated by long and large in other sections of this volume, only some general remarks are made here. Soils on dunes, which do not seem to match with the recent climate, can successfully be used as indicators of palaeo-climates as shown in the bordering Sudan and E-Niger by various authors (i.e. Felix-Henningsen, 1984; Völkel, 1989). However, there are some terrain characteristics and processes present, which should be carefully considered before drawing conclusions from soils on dunes in the visited region. First of all, complete terrestrial soil profiles could only be detected in the Sahelian domain. Due to anthropogenic disturbance even there eolian redistribution of materials is a common phenomenon. Apart from these recent phenomena, layering is very common in the described soils as indicated by texture data (Felix-Henningsen, 1999; Mounkaila, 2006). In order to derive conclusions about the weathering intensity in the sense of quantifiable characteristics we need to reconstruct the parent material. In order to do so, we have to consider three processes. The first is incorporation of fines into the dunes during dune formation, i.e. by transgressing lacustrine sediments, as it could be often observed in the Saharan domain. The second is recent dust deposition as a continuous ubiquitous process with a gradient depending on aridity and distance to sources (Herrmann, 1996). The third one is local disturbances of unknown origin (i.e. fires) or short term arid phases (several years) which enable re-translocation of fines from the finer textured inter-dune depressions. All these processes can influence the presence of fines (clay) and free iron (Fed) which are normally used to identify in situ weathering intensity. In order to separate these influences detailed analytical work, especially concerning mineralogy and geochemistry is indispensable but rarely executed. In the Saharan domain, apart from the recently moving sands, suspected older dune sand formations where only discovered in protected situations, i.e. under lacustrine sediments. Occurrence is erratic and state and setting allow only very limited conclusions. Normally, C-material could not be sampled or is not present, which means that no reliable reference to establish the weathering intensity exists. The locally different colour of barchans in the Saharan domain impressively shows, that potential C-material can be very different, especially with respect to iron minerals. Consequently, in situ weathering intensity and extent cannot be determined. So pedological information is scarce, it rests dating as a possibility to determine arid events. However, pre-requisite for interpretable TL and OSL dating is complete bleaching before sedimentation, which needs a certain transport distance. This information is normally derived from landscape context and sedimentary information but can hardly be collected under the given circumstances for the very localized “older” dune materials. In conclusion, Northern Saharan Chad with its extreme deflation activity (negative regional mass balance) has only low potential for valuable pedological information from dune palaeosoils. In the Sahelian domain exploitable sites are present. In future the sites in major eolian transport direction, i.e. SW of the Bodele depression where materials begin to accumulate should be of interest in order to find the connection between findings of this expedition (east of Lake Chad and Bodele Depression) and those established in E-Niger.

118

Ludger Herrmann et al.

6.7 A PALAEOSOIL IN THE SAHARAN REGION UNDER SERIR A rather unusual finding was a nearly complete palaeosoil (Figure 6a) several kilometres north of the Angamma escarpment. The site was visited due to a very sharp linear boundary visible on a Landsat image (Figure 6b). The profile is situated under a serir surface on the banks of a palaeo-riverbed. The river drained the Tibesti towards the Bodele Depression. The visit in the terrain showed a fluvial morphology. A natural N–S oriented dam is separating the fluvial domain with relatively unconsolidated greenish fresh and mica rich sandy (3% clay) sediments—establishing the magmatic Tibesti influence—and the terrestrial domain with significant signs of terrestrial weathering (BC-horizons present). The latter materials are brownish to reddish, finer (ca. 20% clay) and consolidated. From the morphological point of view the profile can be separated into three parts. The uppermost is characterised by a serir surface and a yermic phase. The central part contains coarse components and has a reddish colour. Finally, the finer textured subsoil is characterised by brown to greenish colours, impregnated along coarse vertically oriented pores with material from the reddish overlying horizons.

Figure 6a. Palaeosoil profile (Aridi-gleyic Luvisol) in a river dam north of the Angamma escarpment.

Are there valuable pedological palaeoenvironmental indicators in Northern Chad?

119

Figure 6b. Landsat 5 TM image showing an erosion event in the Bodele Depression mainly on lacustrine sediments (bright area), and the location of Angamma escarpment and palaeosoil sampling site.

The granulometry (Table 2a) indicates—with exception for the two lowermost horizons—strong layering with an increase of the skeleton content towards the topsoil, which is rather unusual for riverbank profiles developing under constant environmental conditions. A distinct downward increase of the clay content between 0,4–0,7 m depths indicates beside clay cutans clay illuviation, though the whole profile contains small

120

Ludger Herrmann et al. Table 2a–d. Analytical data of a palaeosoil profile (Aridi-Gleyic Luvisol) north of the Angamma escarpment in Northern Chad.

Horizon

Depth cm

Skeleton wt.%

cS %

mS %

fS %

Sand %

Silt %

Clay %

Corg %

pH H2O

Serir C E Btg B(t)kg B(t)gr

0–3 −7 −40 −70 −120 −140

100 63 61 3 2 2

10 35 12 5 6

18 20 23 17 25

24 7 13 16 18

53 62 47 38 48

24 12 21 33 27

23 26 32 30 25

0,10 0,02 0,03 0,02 0,02

9,1 8,1 7,9 8,1 8,1

Horizon

KAKpot

Exchangeable cations Ca

Mg

K

Na

−1

Horizon

256 207 343 470 407 EC 1/20 μS cm−1

222 148 192 329 265

BS

Ca

18 25 58 75 65

12 6 4 4 3

3 3 3 5 2

Mg

K

P

−1

%

mmol c+ kg C E Btg B(t)kg B(t)gr

Weatherable elements (HCl)

mmol kg 100 88 75 88 82

161 74 90 208 133

115 91 129 166 146

96 71 57 55 49

11 4 2 2 2

Water soluble ions Na

K

Ca

Mg

NO3

SO4

Cl

PO4

0 19 12 3 4

<0,2 1 8 16 12

<0,6 <0,6 <0,6 <0,6 <0,6

mmol kg–1 C E Btg B(t)kg B(t)gr Horizon

C E Btg B(t)kg B(t)gr

79 259 272 274 243

7 20 22 21 17

CaCO3 %

Feo/d

0,8 0,2 0,3 1,3 0,4

0,03 0,02 0,01 0,02 0,02

3 3 2 2 2

5 18 17 19 18 Fed

1 3 6 5 5 Mnd

0 1 9 12 10 Kaolinite

mg kg−1 11.216 10.351 11.514 8.308 8.694

Illite

Smectite

% 429 115 361 339 640

30 34 10 6 6

2 3 3 2 2

68 63 87 92 92 (Continued)

Are there valuable pedological palaeoenvironmental indicators in Northern Chad?

121

Table 2a–d. (Continued ).

Horizon

K H2O

Exch.

Weath.

Clayt

mmol kg−1 C E Btg B(t)kg B(t)gr

3 3 2 2 2

12 6 4 4 3

Horizon

% 96 71 57 55 49

1,17 0,85 0,66 0,61 0,65

0,76 0,59 0,84 0,77 0,87

Clayt

Total

Mg H2O

Exch.

Weath.

mmol kg−1 C E Btg B(t)kg B(t)gr

1 3 6 5 5

18 25 58 75 65

Horizon

% 115 91 129 166 146

1,21 0,86 1,23 1,48 1,51

0,5 0,39 0,74 0,62 0,85

Clayt

Total

Ca H2O

Exch.

Weath.

mmol kg−1 C E Btg B(t)kg B(t)gr

Total

5 18 17 19 18

222 148 192 329 265

% 161 74 90 208 133

1,57 1,01 1,18 1,51 1,54

0,95 0,49 0,91 0,61 1,28

amounts of carbonates. Taking into account the latter, the base saturation in the soil profile should reach 100%. This is not the case, which can only be explained by the fact that the soil was once desaturated and that recalcification occurred along major pores but not in the whole matrix. Organic matter content (Table 2d) is extremely low for all horizons. From this point of view, as well as from colour or aggregation, no A-horizon could be differentiated. However, magnitude of E- vs. Bt-horizons and difference in clay concentration could be an argument that the former A-horizon is included in the E and only organic matter was decomposed with time. Unfortunately the layering does not allow for a calculation of the clay balance, which could answer the question. The potential cation exchange capacity (Table 2b, 2d) is over proportion increasing in relation to the clay content towards the subsoil, indicating a shift in clay mineral assemblage. In fact, the clay mineral composition shows a clear gradient with downward decreasing

122

Ludger Herrmann et al.

kaolinite and increasing smectite (Figure 7). Illite is present only in traces. With respect to mineral composition the two uppermost and the two lowermost horizons are nearly identical, whereas the Btg represents a transition. The peaks in the XRD-diffractogramme at 14,8, 4,9 and 4,5 Å indicate a Mg-rich smectite, which is supported by the downward increasing Mg-fractions (H2O, exchangeable, weatherable, total elements in clay fraction). The high intensity and the symmetrical form of the peak at 14,8 Å are arguments for in situ neo-formation. We interpret these data with respect to the palaeoenvironment as follows. Texture and geomorphologic position in the landscape emphasize fluvial deposition as major sedimentary process. The low skeleton content in the subsoil indicates calm and uniform sedimentation processes during a more humid phase when the lake level was high and the affluent rather constant. If one accepts the “Mega Chad”-hypothesis, then the climate must have been between semi-arid to sub-humid conditions in order to feed the lake. The subsoil colour is an expression of prolonged groundwater influence. However, the clay mineral composition with dominance of Mg-rich clay minerals indicates a seasonal climate with a distinct dry season, which allows for concentration of solution and neoformation of such minerals (Jasmund and Lagaly, 1993). In fact, Carmouze (1976) and Gac (1980) described the same neoformation in the present Lake Chad. There the

Figure 7. XRD-diffractogramme (Cu-Ka, background corrected) of the clay fraction (<2 μm) of three horizons of a palaeosoil profile near the Angamma escarpment in Northern Chad.

Are there valuable pedological palaeoenvironmental indicators in Northern Chad?

123

chemical structure of the “montmorillonite magnésienne” is given with: Ca0.145(Si4.0) (Al1.32Fe0.46Mg0.18) O10(OH)2. The overlying gravel-rich material indicates a deposition with catastrophic events, as they occur with more distinct and drier seasonal climates and higher rainfall intensities as we find them at present under Sahelian conditions. Also clay illuviation and reddening of the upper profile part fit into this picture. A further support comes from the interfingering of reddish material into the greenish subsoil, which should have occurred along desiccation cracks, which could develop into the lowermost horizons due to a fall of the groundwater level. Probably already at that time the river became ephemeral. The weathering intensity is rather low as is clay illuviation. Though kaolinite occurs in the E horizon with 34%, this value should not be overemphasized, since we do not know the clay mineral composition of the parent material and the semi-quantitative analysis did not use weighting factors for different cristallinity of the clay minerals. Though not supported by granulometric data, a certain share of kaolinite might be due to incorporation of eolian material (mineral dust) deposited on top, which contain in these environments at least a share of 20% kaolinite as determined in the adjacent samples from the fluvial and terrestrial domain. However, leaching must have occurred, so that the soil profile was carbonate free, chemical weathering proceeded and clay illuviation took place. Under present semi-arid conditions in the Sahel (approx. 560 mm annual rainfall) leaching losses reach up to 200 mm in Luvic Arenosols (10–15% clay) as documented by Bley (1990) and Fechter (1993). In more clayey (subsoil-) profiles, as in this case, leaching is concentrated along the desiccation cracks at the beginning of the rainy season and leads to similar losses. Taking this information and the high skeleton content of the E-horizon into account, several hundred years of soil formation (after a potential decalcification if carbonates were present) could have been enough to produce the described characteristics. Finally, the Serir and C-material on top are witness for the present day climate. That erratic rains still transport solutes to the subsoil is shown by the depth distribution of conservative water-soluble ions (i.e. Na+, NO3− Cl−), increasing with depth. This is also true for the carbonates found in the profile, which are originating from the deposited mineral dust on top of the profile, which regularly contains carbonate in small amounts (ca. 1%, Herrmann, 1996; Moreno et al., 2006). Though the carbonate concentrations are low, the present arid phase must already last quite long to produce the subsoil carbonate accumulation, given the low solubility and the low number of rains capable to moisturize even the subsoil. A hint that the carbonates are not an effect of capillary rise is that also soluble ions tend to have a maximum in the B(t)kg-horizon, which indicates a change in the downward conductivity. Re-evaluation of the sampled sites shows, that in the visited Saharan region it is only in this restricted area, north of the Angamma escarpment and west of Ain Galaka under Serir surfaces, that soil horizons characterized by chemical weathering (brownish to reddish B and BC-horizons) could regularly be detected. Due to the given reasons, these sites were unfortunately not described to a greater depth. While the area farther east is under the influence of strong winds due to channelling effects around the flanks of the Tibesti and shows everywhere signs of deflation, north of the Angamma escarpment the Tibesti massive seems to function as a windshield, reducing the wind erosion forces. With more detailed investigation it should be possible to find further palaeosoil remnants in a landscape context, which would allow for much more in depth interpretation of landscape and palaeoenvironmental processes. This is also indicated by a number of relic soil profiles (mainly Gleysols, partly with fAh-horizon) described by Felix-Henningsen (1999).

124

Ludger Herrmann et al.

6.8 CONCLUSIONS The three major surface types bear a strongly different potential for palaeoenvironmental information. With the exception of the southern transect end, dunes are very mobile and only isolated occurrences of older than present formations could be detected. Though dating of these remnants seems to give some reasonable results (Felix-Henningsen and Mauz, 2004), the necessary interpretation in the landscape context (Herrmann et al., 2007) is hardly achievable. More in depth insight through interpretation of palaeosoils on Pleistocene dunes, as they could be revealed in the neighbouring country Niger (FelixHenningsen, 2000) is not to be expected from the northern areas visited. Of more interest should be dunes in leeward direction (SW) of the Bodele Depression to connect results from Chad with those established in E-Niger. Also complete lacustrine sedimentary series are not to be expected, because those crossed are all in a state of strong erosion. Passing dune fields and sand sheets abrade the consolidated lacustrine sediments as indicated by frequent eolian features like yardangs. However, frequently detected semi-terrestrial features, which are more resistant to wind erosion, like swamp ores and iron oxide fillings in cracks (in general: gley phenomena) bear a certain potential (compare Felix-Henningsen, 2003). They indicate conditions near the lakeshore, which could be spatially mapped in order to learn more about the history of the “Mega-Chad”. With further progress in remote sensing and availability of higher resolution images, these areas might be distinguishable on the basis of their spectral properties (Mounkaila, 2006). Analysis of Fe-isotopes in these features might help to reconstruct the major directions of affluents, simply because petrography around the Bodele depression is so different that it should be expressed by the isotopic signature. In addition, some lacustrine sediments show extreme features (i.e. pH 3–4) that are normally associated to swamp conditions (acid sulphate soils). These can also be used to map shallow water conditions. A high potential to conserve palaeoenvironmental information have serir surfaces in certain areas. Though initially created by erosion processes, once established, these surfaces are quite stable and even tend to accumulate fine particulate matter as could be seen by frequent shallow loess-like material accumulation underneath the pebbles on top (yermic phase). Especially the region around the Angamma escarpment is of interest. Here, a nearly complete palaeosoil indicating several climatic phases could be found. In addition, brownish to reddish (sub-?) soil material indicating terrestrial weathering under at least semi-arid conditions could be detected. So, with more thorough investigation a whole palaeo-landscape might be discovered which would allow for an in depth interpretation of the palaeoenvironment. The reason that this information is conserved at this place seems to be that it is apart from the strongest wind pathways caused by channelling effects around the Tibesti, but is more situated in the wind shadow of these mountains. This should be the area of more detailed research if pedological information is requested. The localized soil remnants found in the Bodele-Depression itself shall soon be gone due to wind erosion. REFERENCES Baumhauer, R., 1991, Palaeolakes of the South Central Sahara—problems of palaeoclimatological interpretation. Hydrobiologica, 214, pp. 347–357. Bley, J., 1990, Experimentelle und modellanalytische Untersuchungen zum Wasser- und Nährstoffhaushalt von Perlhirse (Pennisetum americanum L.) im Südwest-Niger. Dissertation. (Stuttgart: Institut für Bodenkunde und Standortslehre. Universität Hohenheim).

Are there valuable pedological palaeoenvironmental indicators in Northern Chad?

125

Carmouze, J.P., 1976, La regulation hydrogéochimique du lac Tchad. Travaux et Documents de l’ORSTOM, 58. Drake, N. and Bristow, C., 2006, Shorelines in the Sahara: geomorphological evidence for an enhanced monsoon from palaeolake Megachad. The Holocene, 16, 6, pp. 901–911. FAO, 1990, Guidelines for soil description. (Rome: FAO). FAO-UNESCO, 1994, Soil map of the world. Revised Legend with corrections. (Wageningen: ISRIC). Fechter, J., 1993, The simulation of pearl millet (Pennisetum glaucum L.) growth under the environmental conditions of Southwest Niger, West Africa. Hohenheimer Bodenkundliche Hefte, 10. Felix-Henningsen, P., 1999, Jungquartäre Landschaftsentwicklung und aktuelle Sedimentdynamik zwischen Tschadsee and Tibesti. Research Report to Deutsche Forschungsgemeinschaft. (Giessen: Institut für Bodenkunde und Bodenerhaltung. Justus Liebig Universität). Felix-Henningsen, P., 1984, Zur Relief- und Bodenentwicklung der Goz-Zone Nordkordofans im Sudan. Zeitschrift für Geomorphologie, N.F., 28, pp. 285–303. elix-Henningsen, P., 2000, Palaeosols on Pleistocene dunes as indicators of palaeo-monsoon events in the Sahara of East Niger. Catena, 41, pp. 43–60. Felix-Henningsen, P., 2003, Genesis and palaeoecological interpretation of swamp ore deposits at Sahara palaeo-lakes of East Niger. Lecture Notes in Earth Sciences, 102, pp. 47–72. Felix-Henningsen, P. and Mauz, B., 2004, Palaeoenvironmental significance of soils on ancient dunes of the Northern Sahel and Southern Sahara of Chad. Die Erde, 135, pp. 321–340. Gac, J.Y., 1980, Geochimie du bassin du lac Tchad. Travaux et Documents de l’ORSTOM, 123. Graef, F., 1999, Evaluation of agricultural potentials in semi-arid SW-Niger—A soil and terrain (NiSOTER) study. Hohenheimer Bodenkundliche Hefte 54. (Stuttgart: Institut für Bodenkunde und Standortslehre. Universität Hohenheim). Herrmann, L., 1996, Staubdeposition auf Böden West-Afrikas—Eigenschaften und Herkunftsgebiete der Stäube und ihr Einfluß auf Boden- und Standortseigenschaften. Hohenheimer Bodenkundliche Hefte, 36. Herrmann, L., ed., 2005, Das kleine Bodenkochbuch. Herrmann, L., Stahr, K. and Jahn, R., 1999, Identification and properties of source regions for soilborn dust. Case study: Dust sources in eastern West Africa. Contributions to Atmospheric Physics, 72, 2, pp. 141–150. Herrmann, L., Zarei, M., and Stahr, K., 2002, Temporal and spatial trends of mineral associations from the Sahara (N-Chad) to the Sahel (SW-Niger). Berichte der Deutschen Ton- und Tonmineralgruppe e.V. (DTTG) 9, pp. 47–56. Herrmann, L., Sommer, M. and Bleich, K.E., 2007, Paläoklimatische Interpretation von Paläoböden—Theoretische Überlegungen. Mitteilungen der Deutschen Bodenkundlichen Gesellschaft, 110, 2, pp. 475–476. Jasmund, K. and Lagaly, G., 1993, Tonminerale und Tone. (Darmstadt: Steinkopf). Kusnir, I., 1995, Géologie, ressources minerales et ressources en eau du Tchad. Travaux et Documents Scientifiques du Tchad. Connaissance du Tchad I, 2ème Edition actualisée et augmentée. (N’Djamena: Centre National d’Appui à la Recherche). Mauz, B. and Felix-Henningsen, P., 2005, Palaeosols in Saharan and Sahelian dunes of Chad: archives of Holocene North African climate changes. The Holocene, 15, 3, pp. 453–458. Moreno, T., Querol, X., Castillo, S., Alstuey, A., Cuevas, E., Herrmann, L., Mounkaila, M., Elvira, J. and Gibbons, W., 2006, Geochemical variations in mineral aerosol from the Sahara-Sahel dust corridor. Chemosphere, 65, pp. 261–270.

126

Ludger Herrmann et al.

Mounkaila, M., 2006, Spectral and mineralogical properties of potential dust sources on a transect from the Bodele Depression (Central Sahara) to the Lake Chad in the Sahel. Hohenheimer Bodenkundliche Hefte, 78. Schulz, E., 2000, The southern margin of the Sahara in the Republic of Chad. Vegetation, soil, and present pollen rain. Zentralblatt für Geologie und Paläontologie, I, 5, 6, pp. 483–496. Servant, M., 1983, Sequences continentals et variations climatiques: Evolution du Bassin du Tchad cenozoique superieur. Travaux et documents de l’ORSTOM, 159. Völkel, J., 1989, Geomorphologische und pedologische Untersuchungen zum jungquartären Klimawandel in den Dünengebieten Ost-Nigers (Südsahara und Sahel). Bonner Geographische Abhandlungen, 79.

CHAPTER 7

New discovery of rock fulgurites in the Central Sahara Barbara Sponholz Department of Geography, Physical Geography, Julius-Maximilians-University, Würzburg, Germany ABSTRACT: Rock fulgurites—formed by lightning strike to silicate rocks—are described from two sites in the Central Sahara/Republic of Niger. The fulgurites from the Emi Bao cuesta near Seggedim and from Achelouma are characterized by their content in lechatelierite, but also in microcrystalline quartz (Emi Bao) and by their change in colour from light beige to dark violet following the lightning path through the rock. Compared to dune sand fulgurites the rock fulgurites do not require any special hydrological conditions. They correspond to the lightning strike risk of exposed topographic points. Lightning strike may have occurred under humid as well as under arid climatic conditions with just ephemeric thunderstorm events. In consequence, the two analysed rock fulgurites fit with the palaeoclimatic interpretation of dune sand fulgurites in the Central Sahara published by Sponholz et al. (1993), but are not necessarily correlated to it.

7.1 INTRODUCTION Fulgurites—from “fulgur” (lat.) = lightning—are formed by lightning strike into quartz sands or silicate rocks. The lightning strike causes short-term heating of the substratum up to 3.000 °C and brings some minerals to melt, such as quartz, while others are vaporized or remain unaffected. Therefore, the most important mineralogical characteristic of fulgurites is the presence of lechatelierite, amorphous SiO2 that forms during the rapid cooling of the melt. Since the first reported description of fulgurites in 1706 by Hermann (after Rakov, 1999) and early scientific analyses (Julien, 1901; Lacroix, 1915, 1931/1932) they often have been described from recent dune areas, e.g. from the Saharan dune fields, but also from Palaeozoic and Mesozoic sedimentary rocks, where they are fossilized mainly in sandstones. They also occur in recent loose sediments all over the world, where they are present as singular features. Apart from those “sand fulgurites”, “rock fulgurites” are also known. They form when solid rock is struck by lightning. Rock fulgurites appear as glassy coatings on the rock’s surface or as lechatelierite veins following fractures within the host rock. In most publications fulgurites are considered as “exotic” mineral formations, and the main focus is put on their mineralogical characteristics. Only few scientific investigations however deal with the palaeoclimatic implication of fulgurites (Sponholz et al., 1993; Rakov, 1999; Navarro-Gonzales et al., 2007) and even less with rock fulgurites (Norin, 1986; Ege, 2005). This article focuses on the description of rock fulgurites in the Central Sahara and will compare those rock fulgurites with dune sand fulgurites for palaeoclimatic reconstruction.

128

Barbara Sponholz

7.1.1 Identification of fulgurites Any kind of fulgurite is easy to identify by thin section analyses: If the mineral lechatelierite—amorphous SiO2, formed by melting of quartz and immediate cooling of the melt (Lacroix, 1915)—is present and the confusion with pseudotachylyte may be excluded by the geological context, it proves the sample to be a fulgurite. Macroscopically, sand fulgurites figure as irregular, tubular shaped formations vertically sticking inside the dune sands (or lying on the sand surface in small fragments) (Figure 1). They form by melting processes along the lightning path through the ground. Sometimes they seem to follow plant roots and therefore are branched from the surface downwards. Their diameter varies from about 3 mm up to several centimetres. The central tubular void is surrounded by lechatelierite, locally in a foamy structure. On the outside wall of the fulgurite, quartz grains may be embedded in the lechatelierite matrix (Figure 1). The colour of dune sand fulgurites is white, even transparent, to blackish, depending on the composition of the melt. They are so characteristic in form and structure that they are easy to distinguish even in the field. At least in well sorted, uniform sands the fulgurite tubes are prominent features, but even in coarser fluvial pebbles they are evident (pebble fulgurites are exhibited in the Staatl. Museum für Mineralogie und Geologie, Dresden/Germany). At the surface, they are quickly reworked by fluvial or aeolian action because of their weak mechanical resistance. But even in sand samples, the presence of lechatelierite in the sand size fraction indicates fulgurite fragmentation at or near the same site (Sponholz, 2004). Rock fulgurites however have been rarely described in scientific literature. They are much more difficult to find in the field, because they do not show the tubular shape that is

Figure 1. Dune sand fulgurites from the Grand Erg de Bilma (Niger). The fulgurite fragment above shows the cortex with adherent quartz grains on the lechatelierite matrix. Below: lechatelierite coating of the central tubular void inside the fulgurite. The shiny, glassy amorphous structure is visible (longitudinal section).

New discovery of rock fulgurites in the Central Sahara

129

typical of fulgurites in loose materials. Most frequently, rock fulgurites are glassy coatings or crusts on the rock surface, mainly found on (or near) the top of mountain summits (Ege, 2005). Lightning strike on solid rock, however, may also cause the rock’s fragmentation into clasts (Norin, 1986), with their size depending on the rock’s chemistry, porosity/fracture density and content of pore water. Furthermore, confining pressure may stabilize the rock and prevent it from fragmenting. Those differences between rock fulgurites and fulgurites in loose materials are explained by the fact, that the sudden heating causes the melting or vaporization of minerals and other substratum compounds, such as organic matter, along the discharge path of the lightning strike. In solid rock, normally the porosity is much lower than in loose material, as is the content of pore water and volatiles. Therefore the explosion effect of pore air and pore water and of other volatiles is reduced and mainly limited to fractures inside the rock. This causes fragmentation of clasts and/or the glassy coatings on fracture surfaces. Alternatively, the discharge takes place on the rock’s surface and creates superficial glassy coatings without any fragmentation. 7.2 ROCK FULGURITES FROM THE CENTRAL SAHARA In 2005 and 2006, rock fulgurites were found in the Central Sahara, in the northeastern part of the Republic of Niger, near the oasis of Seggedim and Achelouma. Both rock fulgurites formed in iron-bearing quartz sandstone of Palaeozoic and Early Mesozoic age. The rock fulgurite sites as well as the main distribution area of dune sand fulgurites are shown in figure 2. 7.2.1 The rock fulgurite of Emi Bao/Seggedim Topographic situation: 20°13'15'' N, 12°58'58'' E, 588 m asl. Stratigraphical situation: Formation de Seggedim/Emi Bao (Mesozoic, after Faure, 1966; Upper Senonian, after BRGM, 1966). The fulgurite occurs near the top position of the Emi Bao cuesta, about 180 m above the foreland depression where the village and the sebkha of Seggedim are situated. The fulgurite site is positioned on a flat surface in iron-bearing quartz sandstone, and it figures as a cracked—not completely fragmented—sandstone layer that is still part of the solid rock. The center of the fulgurite that marks the entering point of the lightning strike figures as a very small pit in yellow-reddish sandstone. Cracks radiate from this center. In 10–20 cm distance from the central pit, there is an irregular dark reddish-violet, even blackish rim that describes, together with the radial cracks, a polygonal structure. A thin section of the near central sample of this fulgurite shows mainly undisturbed quartz sandstone. However, in the left part of figure 3, the quartz grains as well as the silica cement have been melted and transformed to amorphous and microcrystalline silica. Lechatelierite matrix is present in this part of the sample, as well as quartz grains that were partially melted and deformed or even transformed into lechatelierite and cristobalite (Figure 4). The melting path that passes through this part of the rock represents the way of the electrical discharge. As known from dune sand fulgurites, the lightning causes shortterm heating up to 3.000 °C along this path through the ground (Feldmann, 1988). As the heating is of short duration, it remains limited to a small area beneath this path. During the sudden heating, all pore water explodes and other volatiles substances are released, too. The explosion pressure and the evacuation of steam pushes the melt away from the heating center. Flow structures of the melt and “squeezed” grains (Figure 5) are the result

130

Barbara Sponholz

Figure 2. North-East Niger with localization of rock fulgurites of Emi Bao and Achelouma and the common distribution of dune sand fulgurites in the Grand Erg de Bilma and the Erg du Ténéré.

of this effect. Nevertheless, the pressure of the surrounding rock and probably also the high density of the rock without important porosity reduce the effect of explosion. The result of the lightning strike remained limited to the mineral transformation in a restricted zone and cracking of the sandstone layer. Lechatelierite is present in this sample from Emi Bao, but in the thin section it is not the dominant transformation of quartz. However, microcrystalline quartz dominates in the lightning path area. This formation of microcrystalline quartz is an effect of the cooling of the melt and depends on the composition of the melt and the duration of cooling (Vincent, 2007). If incipient recrystallization of lechatelierite is regarded as an indicator of long-term transformation of natural glass, the Emi Bao fulgurite represents a lightning strike that does not correlate with the Late Pleistocene or Holocene climatic conditions, but belongs to an earlier stage. 7.2.2 The rock fulgurite near the oasis of Achelouma Topographic position: 22°18'42'' N, 12°40'38,9'' E, 672 m asl. Stratigraphic situation: Post Tassilien to Grès de Nubie (Nubian sandstone) (Late Palaeozoic to Early Mesozoic, after BRGM, 1966).

New discovery of rock fulgurites in the Central Sahara

131

Figure 3. Thin section of a dune sand fulgurite, Grand Erg de Bilma (Niger). Quartz grains melted, transformed into lechatelierite and solidified with preservation of their contours. Matrix between the grains consists of lechatelierite, too.

Figure 4. Thin section of the rock fulgurite sampled at Emi Bao. Lechatelierite (L) and microcrystalline quartz (MQ) matrix are visible. In the lower part of the photograph remaining quartz grains are visible. Width of the photograph: 1 mm (xpl).

132

Barbara Sponholz

Figure 5. Single quartz grains (Q) that partially melted and subsequently “squeezed” by the explosion pressure caused by sudden vaporization of pore water during the lightning strike (Emi Bao). Width of the microphotograph: 500 μm.

The rock fulgurite is situated on a wadi terrace, 2 m above the recent valley floor. The wadi is tributary to the Enneri Achelouma which it joins from the North, and its length is about 10 km, its width at the mouth about 2,5 km. The recent wadi bed however is restricted to a narrow valley of just a few meters, reaching a maximum about 30 m in width. On the other hand, upstream morphology and coarse pebble deposits indicate at least single high-energy floods also in the recent past. Precipitation is small and episodic (no climatic data is available for this area, but mean annual rainfall in the Central Sahara is below 50 mm (DMN) (e.g. Bilma: mean precipitation 19 mm/a; Annuaire Météorologique du Niger, 1991). However, during the winter months heavy rains may occur and may cause fluvial dynamics in wadis and flooding of roads. According to locals fulgurite formation seems probably due to a lightning strike during the years of rebellion in the North of Niger in the first half of the 1990s. The fulgurite rock (Figure 6) is situated on the wadi river terrace in an upright position. Before fragmentation by lightning, it was a compact, almost round boulder of 70 cm in diameter. This size is about twice the maximum size of other fluvial terrace boulders nearby at the surface. Therefore the struck boulder was one of the prominent points on the wadi terrace during the thunderstorm. As the localization of lightning strikes to the ground is determined by strong electrical fields or by prominent reliefs, this prominence of the boulder was probably the reason for the lightning strike’s localization. Today, seven large fragments of the boulder are oriented as shown in figure 6. The western part of the boulder remained in the original upright position. Its east-oriented face corresponds to the main fragmentation face. The six other large fragments were pushed

New discovery of rock fulgurites in the Central Sahara

133

Figure 6. Rock fulgurite near Achelouma. The upright fragment is still in its original position as it was struck by lightning. The other large fragments were dislocated by the lightning strike and the explosion of pore water.

by the lightning strike like petals in centrifugal direction to the North, West, Southwest and South. Their base is at a distance of 30 to 50 cm from their original position. By their arrangement, orientation and also by their size it is evident that today they are still in the same position as immediately after the lightning strike event. Later flooding or other morphological processes did not affect them. At the top of the upright fragment, the point of the lightning impact is still visible (Figure 6). From this entrance point of the lightning into the rock, the main fragmentation face goes vertically downwards and designs the way of further fragmentation. The struck quartz sandstone is compact and does not show important fissures, fractures or loose bedding planes. On all the fragments, orientation of the fracture surfaces does not correlate to any major rock structure. That means that the fragmentation was the result of the explosion pressure, as well as the dislocation of the large fragments over up to 50 cm. As the boulder was isolated on the wadi terrace, no pressure of surrounding solid rock could prevent fragmentation and dislocation of the fragments as it was the case at the Emi Bao site. Even more evident than on the Emi Bao rock fulgurite (see above) is the change in colour on the Achelouma rock fulgurite. By its iron content the original quartz sandstone has a yellowish-brown colour and a thin weathering crust of dark iron/manganese oxides on the surface. The lightning strike caused a change in colour from bright yellow—even whitish—to dark violet, with all transitions from yellowish-brown to reddish. This change of colour concerns irregular paths and rims of up to 5 cm width throughout the fragments. Thin section analyses of various “typical” parts of the rock fulgurite from Achelouma reveal in general the same characteristics as those already mentioned for the Emi Bao rock fulgurite: presence of lechatelierite, flow structures and deformation of partially melted quartz grains and mobilisation of iron. However, by the possibility of more complete

134

Barbara Sponholz

sampling at the exposed fragments and the better representation of material variety in thin sections, further details become evident. The bright (white to yellowish-brown) and the dark (reddish to dark violet, even black) parts of the rock differ mainly in their content in iron. The dark parts contain lots of minute haematite crystals, magnetite cubes are distributed within the lechatelierite matrix (Figure 7). As Fe remains a non-volatile component, it becomes incorporated into the silicate melt. At temperatures in excess of 2.000 K—that frequently occur at lightning strikes to the ground inside the substratum, see above (Feldmann 1988)—the Fe2O3 and silicate melts are immiscible (Anonymus, 2006). During the subsequent cooling of the melt, the minute haematite or magnetite crystals form from the Fe2O3 melt. At the present stage of analyses, the presence of silicides, as they are described from fulgurites by Anonymus (2006), is not yet proofed for the Achelouma rock fulgurite, but further analyses will focus on this point. The more frequent the haematite crystals are in the fulgurite, the higher the content of iron oxides was during the lightning strike. The lighter coloured parts of the rock, however, are characterised by a lower content in iron and by the presence of iron hydroxides. So they correspond to the more water-bearing parts of the rock, which were situated near the bottom of the boulder and always in contact with the terrace floor. Mobilization of iron by enforced chemical weathering processes at the bottom of the boulder may be explained by hydrological and microclimatical differences. During and after episodic floods, the lower part of the boulder got more soil humidity that enabled iron hydroxide formation and iron mobilization before the lightning struck it. Whether the mobilized iron was moved by evaporation from the lower part to the surface respectively to the upper part of the boulder remains unclear.

Figure 7. Haematite (H; left; xpl) and magnetite (M; right; ppl) crystals embedded in lechatelierite matrix. (Achelouma). Width of the microphotographs: 500 μm.

New discovery of rock fulgurites in the Central Sahara

135

7.3 PALAEOCLIMATIC IMPLICATION OF DUNE SAND AND ROCK FULGURITES In contrast to the palaeoclimatic implication of the dune sand fulgurites in the Erg of Bilma and the Erg of Ténéré south of 18°N, the rock fulgurites at the Emi Bao and the Achelouma implicate another interpretation. The dune sand fulgurites indicate monsoonal thunderstorms up to 18°N during the Mid-Holocene (Sponholz et al., 1993). Their localization in midslope position of interdune depressions needed highly contrasting hygric seasonality with coexistence of permanent lakes/groundwater levels and overlying dry dune sands. The rock fulgurites further north do not require any special hydrological conditions. They correspond to the lightning strike risk of exposed topographic points. Lightning strike may have occurred under humid as well as under arid climatic conditions with just ephemeric thunderstorm events. In consequence, the two analysed rock fulgurites of Emi Bao and Achelouma fit with the palaeoclimatic interpretation of dune sand fulgurites in the Central Sahara published by Sponholz et al. (1993) that proposes seasonal northward movement of tropical (monsoonal) air masses up to 18°N during the Mid-Holocene. But the described rock fulgurites are not necessarily correlated in time and/or (palaeo-) climatic conditions to those dune sand fulgurites. 7.4 PERSPECTIVE Rock fulgurites are more difficult to detect in the field than dune sand fulgurites. Nevertheless, by their change in colour, at least iron-bearing sandstones are transformed in a typical way by lightning strikes as it was described above. Rock fulgurites of this type may be found by the change in colour in sandstone outcrops. As iron-bearing sandstones are common in the rock record of Northern and Western Africa, the detection of rock fulgurites and their topographical position together with absolute dating will be an important step in future fulgurite research. Although there are a lot of open questions about fulgurites, they are reliable indicators for lightning strike to the ground—and therefore for thunderstorm activity, in most of all cases related to rainfall (Proctor, 1991). This single way of formation makes all types of fulgurites relevant palaeoclimatic indicators, despite of their rareness. REFERENCES Annuaire Météorologique du Niger, 1991 (Niamey). Anonymus, 2006, without title. http://www.turnstone.ca/fulgur.htm, 07.12.2007. BRGM: 1966, République du Niger, Carte Géologique, 1:2.000.000 (Paris). DMN: Diréction Météorologique du Niger. After Sani, I.M., 2007, unpublished, Niamey. Ege, C., 2005, What are fulgurites and where can they be found? Utah Geolog. Survey, Survey Notes, 37, 1. http://geology.utah.gov/surveynotes/gladasked/gladfulgurites.htm, 03.02.2007. Feldmann, V., 1988, Comparative characteristics of impactite, tektite and fulgurite glasses. In: Konta (ed.): International Conf. on Natural Glasses, (Prague, 1987), pp. 215–220. Julien, A.A., 1901, A study of the structure of fulgurites. Journal of Geology, 9, pp. 673–693. Lacroix, A., 1915, Le silice foudré considerée comme mineral (Lechatelierite). Bulletin de la Societé francaise de Minéralogie, 38, pp. 1–182. Lacroix, A., 1931/1932, Nouvelles observations sur les fulgurites exclusivement siliceuses du Sahara. Bulletin de la Societé francaise de Minéralogie, 54, pp. 75–79.

136

Barbara Sponholz

Navarro-Gonzales, R., Mahan, S.A., Singhvi, A.K., Navarro-Aceves, R., Rajot, J.L., McKay, Chr., P., Coll, P. and Raulin, F., 2007, Paleoecology reconstruction from trapped gases in a fulgurite from the Late Pleistocene of the Libyan desert. Geology, 35, 2, pp. 171–174. Norin, J., 1986, Geomorphpological effects of lightning. Zeitschrift für Geomorphologie N.F., 30, pp. 141–150. Proctor, D.E., 1991, Regions where lightning flashes began. Journal of Geophysical Research, 96, D3, pp. 5099–5112. Rakov, V.A., 1999, Lightning makes glass. 29th Annual Conf. of the Glass Art Society, Tampa/Florida. http://plaza.ufl.edu/rakov/Gas.html, 07.12.2007. Sponholz, B., 2004, Fulgurites as palaeoclimatic indicators—the proof of fulgurite fragments in sand samples. Lecture Notes in Earth Sciences, 102, pp. 73–78. Sponholz, B., Baumhauer, R. and Felix-Henningsen, P., 1993, Fulgurites in the Southern Central Sahara, Republic of Niger and their palaeoenvironmental significance. The Holocene, 3, 2, pp. 97–104. Vincent, E., 2007, Ageing, rejuvenation and memory: the example of spin glasses. Lecture Notes in Physics, 716, pp. 7–60.

CHAPTER 8

Fluvial geomorphology and palaeohydrology of a small tributary of the Plateau de Mangueni, NE Niger Jan Krause and Brigitta Schütt Department of Earth Sciences, Physical Geography, Freie Universität, Berlin, Germany ABSTRACT: The Seeterrassental is a small catchment in the Plateau de Mangueni in the presently hyperarid Central Sahara. Investigations of channel and catchment geomorphology are the base for the determination of Holocene runoff dynamics continuing the studies of Grunert (1983) on regional landscape history. Past and present discharge calculations integrate flow velocity estimation and channel geometry measurements. Flow velocity is determined as flow velocity of bankfull discharges applying the Manning-Strickler equation and is determined as flow velocity of peak discharges applying the Costa equation. For the Holocene three different channel levels can be recorded, including the present channel bed. Holocene fluvial incision started 7.840 a cal BP, dated by the dissected Early Holocene slack water sediments. Terrace levels T2 (older) and T1 (younger) occur as erosion terraces, as well as the present channel bed indicates active incision. Differences in flow dynamics for the Early Holocene terrace level T2, the subsequent terrace level T1 and the present channel point out to the influence of Holocene aridisation on land forming processes. Results for all three terrace levels show a downstream decrease in flow velocity and discharge, caused by downstream water loss in the alluvial channel bed. Linking the derived discharges with palaeoclimate data is the prerequisite to identify past runoff dynamics.

8.1 INTRODUCTION 8.1.1 Objective It is well known that fluvial landforms are important—locally even the most important— geomorphologic elements in arid and hyperarid areas. However, analyses of windblown processes dominate in geomorphologic publications about deserts. Studies on fluvial geomorphology are rare and mainly focus on the easily accessible deserts in the American basin-and-range province (e.g., Abrahams et al., 1998; Montgomery and Gran, 2001). Only a few studies deal with fluvial dynamics, such as published for the Negev desert and Dead Sea region of Israel (e.g., Greenbaum et al., 2000; Enzel et al., 2003; Kuhn et al., 2004). As for geomorphologic research in the Sahara, there are hardly any studies on fluvial landforms except for a few that focus on small-scale landform development, mainly in escarpment areas (Baumhauer and Hagedorn, 1990; Baumhauer, 1991). Yet, water is one of the most important factors of environmental change in arid regions, and fluvial processes are the dominant landscape-forming agents in deserts. These fluvial forms give us an indication of discharge rates—locally the only indication on the intensity of past and present runoff processes in these areas. This paper focuses on the fluvial

138

Krause and Schütt

geomorphology of the Plateau de Mangueni in the presently hyperarid Central Sahara. By analysing fluvial landforms, channel geometry and bedload characteristics, a first attempt will be made to achieve a systematic investigation of fluvial landforms. The geometry of drainage basins and channels will be integrated as a step towards estimating channel flow velocity and peak discharge. However, it is difficult to validate these estimates because all channels are ephemeral owing to the erratic character of rainfall events in this area. The research presented here is based on a DFG-funded expedition in 2006 which took place within the bundle project of the universities of Freie Universität Berlin, Würzburg and Giessen. 8.1.2 The regional geologic-geomorphologic setting The study site is a small catchment area of 31,5 km² located in the Central Sahara in the northeastern part of the Republic of Niger, close to the military post of Madama (approximately 22°20'N, 12°40'E; Figure 1). It drains an apron of the escarpment of the Plateau de Mangueni and is tributary to the Achelouma valley (Figure 2). In the following an overview of the regional natural setting is given. A large-scale geomorphologic investigation, focusing on fluvial landforms, is reported subsequently in chapter 3. The region’s climate is hyperarid: precipitation events reaching this region are erratic and even fail to occur for decades (Dubief, 1963). Owing to the southward displacement of the ITCZ (Intertropical Convergence Zone), the low rainfall in the Northern Sahara comes from the westerlies during the northern hemispherical winter and from premonsoonal humid air masses that reach this area from Southwest in springtime (Wininger, 1975). However, precipitation events from the westerlies have little impact here (Grunert, 1983). Busche (1998) reports an annual precipitation of 0–10 mm for the study site. The nearby meteorological station in Madama (21°56'N, 13°38'E) recorded a mean annual precipitation of 15,1 mm for the 1939–1943 observation period. The meteorological station in Faya, Chad (19°06'N, 19°02'E), located 640 km southeast of Madama, shows annual precipitation similar to Madama (Pachur and Altmann, 2006). Nevertheless, during the observation period mentioned, Madama station recorded a heavy rainfall event with roughly 41,1 mm during one day in May 1943 (Grunert, 1983). Due to the extreme aridity of the region the vegetation found in the region consists of contracted vegetation, dominated by Acacia-Panicum and patchy Achab flora (Schulz, 1988). The structural setting of the region is consistent with the overall structural setting of the Central Sahara, which is dominated by alternating basins and ridges, mostly corresponding to anticlines. Faults strike predominantly SSE–NNW and NNE–SSW. The study site is located at the southern boundary of the Edeyen (Tubu erg) of Murzuk, a structural basin mainly built of Cambrian to Cretaceous sediments, which adjoins the Tibesti Mountains to the SE (Schulz, 1980). The centre of the Murzuk basin is filled by Quaternary dune fields. Cuestas and ridges encircle the basin (Grunert, 1983; Carte Géologique République du Niger, 1965; Busche, 1998). To the South of the Murzuk basin, a collapsed NW–SE striking anticline, corresponding to a graben structure, margins the Murzuk basin. Along the graben floor with strike direction to the SE the Achelouma valley developed, separating the Plateau de Mangueni in the NE from the Plateau du Djado in the SW (Figure 2). The Plateau du Djado is composed of Palaeozoic (Cambrian–Ordovician) sediments, whereas sediments building the Plateau de Mangueni belong to the Cretaceous “Continenal Intercalaire” and the Tertiary “Continental Terminal”, which here both consists of mud- and sandstones. Bedrock outcroppings along the Achelouma valley are Upper Carboniferious limestones and Lower Carboniferous sandstones (Grunert, 1983; Carte Géologique République du Niger, 1965; Busche, 1998;

Fluvial geomorphology and palaeohydrology of a small tributary, NE Niger 139

Figure 1. Aerial photo of the Seeterrassental and its location in West Africa. The dashed line recorded by differential GPS represents the thalweg divided into different segments of the valley (IGN NF 33 XIII–Mosaik, 1955).

Pachur and Altmann, 2006). The surface of the Plateau de Mangueni dips to the N–NE, whereas the surface of Plateau du Djado dips to the South (Pachur and Altmann, 2006). The faults in the Plateau du Djado mainly strike NW/SE. At the Plateau de Mangueni the major joint direction is N–S with a second accumulation in NE–SW direction (Figure 2). The escarpment areas of the Plateaus de Mangueni and Djado are both highly dissected by rivers tributary to the Achelouma valley, and both escarpment areas are also strongly slump-affected (Grunert, 1983). At present the channel of the Achelouma valley has only ephemeral flow and drains south–eastward, drying up before it ends in the basin of Madama. The size of the fans of both the receiving and the tributary streams documents a much more humid climate in the past (Grunert, 1983). The study site is the so-called Seeterrassental (Grunert, 1983), a small drainage basin with an area of 31,5 km² and 13 km in length, draining the escarpment of the Plateau de Mangueni. The receiving stream is the Achelouma valley (Grunert, 1983; Carte Géologique République du Niger, 1965; Busche, 1998).

140

Krause and Schütt

Figure 2. Three dimensional illustration of the Achelouma valley and the surrounding plateaus based on Landsat ETM+ data blended with STRM data (10-times vertically exaggerated).

8.2 METHODS 8.2.1 Field measurements The groundwork of the fluvial analysis of the Seeterrassental is the detailed field survey recorded by differential GPS. The measurement system, a combination of Magellan and Ashtech equipment, is used with one base station with external antenna and two handheld GPS. The maximum distance between the recording GPS and the base station was always smaller than 14 km. Regular control of the ‘positional dilution of precision’ (PDOP), a measure for the accuracy of the 3-dimensional position, is presumed (Gurgel, 1991). To increase accuracy, the recorded data are post-processed according to the orbital variations of the satellite, the inaccuracy of the satellite time and the atmospheric influence, all recorded by the base station (Willgalis, 2005). During field work the longitudinal profile was recorded. The absolute error in all three dimensions cannot be quantified as the absolute position of the base station was not known. In addition 35 channel cross profiles including several valley cross profiles were taken. For the analysis of mean flow velocity at several channel cross profiles, 46 pebble clusters were located. For each pebble in the pebble clusters length, width and height were measured (a-b-c axis). The pebble cluster position is linked to the correspondent terrace level in the correspondent channel cross profile. For characterization of flow dynamics riffles are located along the length profile (Gurnell and Gregory, 1995). The corresponding pools could not be recorded owing to local sand cover.

Fluvial geomorphology and palaeohydrology of a small tributary, NE Niger 141

The stream sinuosity (SI) describes the ratio between stream length and the straight-line distance of a stream section as the intensity of the meandering of a stream. The stream length is the real length of the stream including all bends, and the straight line is the linear connection between the section start and its end (Wagenschein, 2006). 8.2.2 Catchment morphometry For characterisation of the Seeterrassental catchment, its discharge-relevant morphometrical parameters are determined based on the differential GPS measurements (see above). The first parameter is the dimension of the catchment area’s form. The stretched form of a catchment can be quantified by the elongation ratio (E) after Schumm (1960) (cf. also Gregory and Walling, 1973):

( )

E = 2 Aπ

0 ,5

LE

(1)

The catchment area (A) is given in km² and the catchment length (LE) in km. An alternative method of characterizing the catchment area’s form is the elongation ratio after Horton (1932) also known as form factor (Rf ) (Cyffka, 1991): R f = A L2E

(2)

The second parameter is drainage density, which is the ratio between the accumulated stream length (LS) and the area size (A). It displays the influence between catchment and discharge. Gregory and Walling (1973) define the drainage density (D) as: D = LS A

(3)

The downstream variations of the channel cross profiles (channel depth, channel width and channel forms) are defined by regression analysis. These parameters can be split up for the different channel generations (Gregory and Walling, 1973; Kim, 1989; Richards, 1982). 8.2.3 Discharge calculation The hydraulic settings of the Seeterrassental give the input for the calculation of discharge values. For this purpose, the channel cross and longitudinal profiles separated into the present channel and older channel levels (terraces) are used to approximate discharge volumes. Q = v × A [ m 3 / s]

(4)

This formula is used to calculate the peak discharge (Q) of a cross sectional area in which water flows (A [m²]) (Gregory and Walling, 1973; Jones, 1997). To approximate flow velocity (v [m/s]) two different approaches are chosen comparative. The ManningStrickler approach uses the hydraulic factors derived from channel cross and length profile to calculate flow velocity (v [m/s]) of bankfull discharge (Naudascher, 1992): v = kST × Rh 2 3 × I 01 2

[ m / s]

(5)

142

Krause and Schütt

The roughness parameter (kST), derived from Manning’s n [1/n], is set to 30 [m1/3/s] according to the situation in the river bed (Chow, 1995; Naudascher, 1992). Cross and longitudinal profiles give the information for the parameters Rh (wetted perimeter [m] = A/(2 × channel depth + channel width)) and I0 (base slope [m/m]). Supplementary, the Costa approach is applied, where maximum flow velocity (v [m/s]) is calculated on the base of pebble size (Costa, 1983). v = 0,18 × DI 0 , 487

[ m / s]

(6)

This relation is only valid for materials or pebbles where the mean axis (DI) ranges from 50 to 3.200 mm (Costa, 1983). 8.3 RESULTS AND DISCUSSION 8.3.1 Geomorphology The geomorphologic map in figure 3 is produced by combining remote sensing data with field investigations. The predominant landscape element is the roof area (dip slope) of the Plateau de Mangueni with its SW facing escarpment and the affiliating graben. The valley mapped is the so-called Seeterrassental, a tributary to the Achelouma valley. The valley bottom of the Seeterrassental with its river bed is framed by hillslopes which are strongly slump affected. The topography of the valley cross section shows distinctly the influence of mass movements on the local morphology (cf. figure 4). In its upper course, a Late Quaternary landslide caused a damming situation, so that a natural storage lake existed here over centuries in the Early Holocene—eponymous for the Seeterrassental (Grunert, 1983). A lake outburst caused the dissection of the slack water sediments of this lake and the erosion of extended parts of the damming landslide. At present, the outburst is documented by a gorge-like channel dissected into the slack water deposits. A geomorphologic situation comparable to this occurs in the middle course of the Seeterrassental, at the transition between middle to lower course (Figure 1). However, strong erosion caused that remnants of the damming situation and corresponding slack water deposits are largely removed area-wide. Hence, it cannot be excluded that such situation existed also in other sections of the Seeterrassental, but with their remnant landforms and sediments eroded subsequently. The bars marked in the thalweg in figure 5 indicate the location and thickness of slack water sediments. The heights of the bars are grouped by the difference in altitude between the top edge of the sediments and the channel bed. A concentration of slack water sediments occurs in the upper course (UC) and in the middle course 1 (MC1)—possibly indicating former damming situations. In the transition zone between hillslopes and dip slope regressive erosion leads to a disintegration of the dip slope. At its junction into the graben of Achelouma valley the channel of the Seeterrassental deposited an expanded fan (Figure 3). The satellite image reveals a minimum of two fan generations where the older fan is also fed by small tributaries originating from the East and West affiliating escarpment and which are not part of the Seeterrassental drainage system. Additionally, at its southeastern margins the fan is slump affected from the eastward affiliating escarpment. Aeolian deposits occur as small lee dunes underneath the escarpment edge in the headwater areas. 8.3.2 Longitudinal profile The longitudinal profile of the thalweg is of a stretched—concave shape and changes into a convex shape where the alluvial fan (F) starts (Figure 5). The most concave shape occurs in

Fluvial geomorphology and palaeohydrology of a small tributary, NE Niger 143

Figure 3. Geomorphologic map of the Seeterrassental with a joint diagram of the dislocations of the Plateau de Mangueni.

the upper course (UC). In the upper and middle courses also two shortenings of the length profile are evident: at the gorge-like outburst channel and in the channel’s middle course, in the transition between MC1 and MC2. The position of erosional riffle forms is marked in figure 5 with bars beneath the thalweg. Table 1 shows the riffle frequency for each course section as shown in figure 6. As expected, riffle frequency decreases downstream (Kieffer, 1990). The local increase of riffle frequency in the fan corresponds to the braided character of the channel

144

Krause and Schütt

Figure 4. Valley cross profile of the lower Seeterrassental derived from SRTM data combined with differential GPS measurements. The exact location is marked in figure 6.

Figure 5. Length profile of the thalweg and its segmentation (cf. figure 1). Locations of channel cross profiles (cf. Figure 1), riffle positions and slack water sediments are marked along the thalweg. Slack water sediments are grouped by thickness as evident from field mapping into 5 classes, comprising 0–10 m (0–0,5 m, 0,5–1 m, 1–2 m, 2–3 m, 3–5 m, 5–10 m).

in the section (Schumm, 2005). Next to this, already in the transition from the lower course to the fan area (TF-LC) riffle frequency is increased. In this section two larger tributaries disemboguing from the right and from the left increase runoff, and thus cause increased flow velocity with coinciding increased turbulences, documented in an increased riffle frequency (Kieffer, 1990). Riffle frequency is defined by, for example, Ahnert (1996) as

Fluvial geomorphology and palaeohydrology of a small tributary, NE Niger 145 Table 1. Riffle frequency and sinuosity for the different course sections.

Course section F TF-LC LC MC1 MC2 UC

Riffle frequency (×100 m–1)

Sinuosity [m/m]

0,932 1,195 0,399 0,571 0,931 4,882

1,047 1,034 1,143 1,071 1,104 1,218

the frequency of riffles within a stream section calculated per 100 m. Barbour et al. (1999) see therein the possibility to display the heterogeneity occurring in a stream. 8.3.3 Catchment and channel morphometry For the Seeterrassental the elongation ratio (E) has the value of 0,62. After Kim (1989) a more stretched catchment form shows a smaller elongation ratio than a round one. Correspondingly, after Schumm (1960) E depends on the average slope of the drainage and increases from 1,0 in flat terrain to 0,6 in areas with strong relief. The elongation ratio (E) also indicates that catchments with high E values have high infiltration capacity and low runoff. In contrast, catchments with low E values are vulnerable to high erosion and sedimentation load (Obi Reddy et al., 2004). The form factor after Horton (Rf) reaches for the Seeterrassental values of about 0,3. Catchments with this low Rf have less side flow for shorter duration and high main flow for longer duration and vice versa (Obi Reddy et al., 2004). Both, elongation ratio and form factor are valid for the present day situation only. The analysis of the drainage density for the Seeterrassental results in the value of 5,96 km/km². Reference values from other hyperarid areas are not available. Values of drainage density data from subarid to dry-subhumid areas range between 5,1–9,4 km/km² in igneous rocks (McCoy, 1971) and 2,9–33,4 km/km² in granite or gneiss, both in Colorado, USA (Melton, 1957). However, it is well known that drainage density is a morphometric bulk parameter, reflecting a multiply of influencing factors: After Langbein (1947) the drainage density in humid regions is greater than in arid regions, or rather reflects annual precipitation or precipitation-evaporation ratio respectively (Gregory, 1976). Beyond, drainage density is directly affected by bedrock character (Gregory, 1976) as well as the relief and the duration of drainage network development (Oguchi, 1998). Examining the outline of the Seeterrassental it has to be noticed that the channel runs poor winding in its upper and middle course (UC, MC1, MC2; cf. table 1). Locally it becomes the character of a meadow meander, but the distinct character of valley meanders lacks. In its lower course the valley floor shows a marked downstream widening where the fan deposits start. The predominant stretched outline of the Seeterrassental is documented by the sinuosity values close to 1 (Table 1; after Schwaller and Tölle, 2005) and reflects the altogether high erosion potential and high runoff dynamics of the Seeterrassental (Kim, 1989). For the Seeterrassental channel cross profiles are recorded along the main channel (cf. figure 6). In the upper course (UC) the profiles show a steep channel incision and a small channel width (Q11-Q13, Figure 6). This character is emphasized by the gorge-like

146

Krause and Schütt

Figure 6. Channel cross profiles of the main channel within the different valley segments (cf. figures 1 and 5). The locations of the channel cross profiles as well as the valley cross profile (cf. figure 4) are marked in the stream network.

outburst channel in Q11. In contrast, in the upper part of the middle course (MC2; e.g. Q8) channels are wide and shallow, whereas the depth of the channel increases in the lower part of middle course (MC1; e.g. Q6). The more the channel bed widens, the channel tends to furcate. Also in the fan area wide channel beds with local furcations are characteristic (Q1 and Q2; cf. figures 6). For the main parameters channel depth (D), channel width (W) and channel form (f), split up the three channel generations recorded, functional relations are shown in table 2 (for terrace levels see figure 7). For most of the present day channel bed (FB) and for the channel bed corresponding to the first terrace level (T1) a downstream decreasing channel

Fluvial geomorphology and palaeohydrology of a small tributary, NE Niger 147 Table 2. Regression analysis results for downstream variation of channel width, channel depth and form ratios. Variables mentioned in the table: W = channel width [m], D = channel depth [m], L = channel length [km], f = form ratio = W/D (n = 13); for exponential equation a = regression coefficient, b = exponent.

FB

T1

T2

Regression of channel width W = a × Lb

Regression of channel depth D = a × Lb

Regression of form ratio f = a × Lb

a = 4,1705 b = 0,6005 R² = 0,2733 a = 35,216 b = –0,118 R² = 0,0097 a = 33,594 b = 0,3759 R² = 0,2101

a = 0,4176 b = –0,2841 R² = 0,2779 a = 0,5506 b = –0,2974 R² = 0,4539 a = 0,3985 b = 0,2499 R² = 0,4134

a = 9,987 b = 0,8846 R² = 0,3337 a = 63,964 b = 0,1794 R² = 0,029 a = 84,306 b = 0,126 R² = 0,0156

Figure 7. Idealized channel cross profile including the present channel generation (FB) and the two overlying older channel generations (T1 and T2). The pebble clusters (Cl) are indicated in the respective level.

depth can be noticed, while for the channel bed corresponding to the second terrace level (T2) the reverse relationship occurs. The downstream increase of the channel bed width is recorded for FB and T2, while for T1 this relationship behaves almost constantly along the river course. The channel width-depth ratio (form ratio) increases for the most recent channel downstream, while for T1 and T2 again a more or less constant behaviour can be recorded. In channel beds made up of uncongested materials the channel width-depthratio is inversely proportional to the contents of silt and clay in the bank material. The higher the proportion of the silt and clay fraction, the less stable are the channel banks and the easier is a broadening of the channel due to lateral erosion (Schumm, 1960). At the same time incision into the material decreases (Ahnert, 1996). This assumption implies that, due to downstream increase of the form ratio, for FB a decrease of the silt and clay proportion leads to higher lateral erosion with a simultaneous decrease in channel incision (Schumm, 1960). However, for all values shown in table 2 the determination coefficients lack significance (α > 0,05). Nevertheless, functional relations are not necessarily synonymous with causally determined relations (Ahnert, 1996). As shown in the geomorphologic map (Figure 3) and the aerial photo (Figure 1) the Seeterrassental is strongly affected by landslides originating from the hillslopes. These landslides also

148

Krause and Schütt

affect the downstream channel character, varying in magnitude for the different channel generations. Next to this, comparative studies on channel geometry mostly deal with subhumid and humid test sites where channels are perennial and the system properties, such as transport capacity, are not fully loaded (among others Kim, 1989; Golden and Springer, 2006; Stewardson, 2005). In contrast, in ephemeral channels experiencing runoff the transport load is distinctly higher than in mid-latitude perennial channels (Reid, 2002). Next to this, not all runoff events affect the entire catchment, so a downward propagation of channel forming peak discharges is not obligatory. The resolution of the data is most precise available for the moment. SRTM data are especially lacking a good resolution in z direction for this catchments scale. 8.3.4 Discharge calculation Discharge estimations (eq. 4) are carried out for those channel cross profiles where next to the geometric information pebble clusters in the corresponding terrace level were available to estimate maximum transport capacity. The discharge estimations reproduce (1) bankfull discharge as calculated applying the Manning-Strickler-equation (eq. 5) for the flow velocity estimation and (2) peak discharge as calculated applying the Costa-approach (eq. 6) for the flow velocity estimation. These discharge calculations were accomplished for the most recent channel and for the subrecent channels as discernible by the two lowermost terrace levels T1 and T2. Present day channel For all cross profiles incorporated into the runoff estimations persistently values of bankfull discharge (application of Manning-Strickler equation; eq. 5) remain below values of peak discharge (Costa-approach; eq. 6) (Figure 8). Comparing both values percentual difference between estimated bankfull discharge and estimated peak discharge increases downstream. The percentual difference increases from 38,5% in the upper course (Q12) to 78% in the lower course (Q2). For three cross profiles in the lower course (Q1, Q1.1, Q2), with almost the same catchment size, the percentual difference is almost identical (from 75,5% to 78%)—corresponding to the quasi-constant catchment area. Errors in discharge calculation as well applying the Manning-Strickler as the Costa-approach are caused by uncertainties estimating the margins of bankfull discharge in the cross profiles investigated. The increase of the percental difference from upper to lower course indicates a change of hydraulic conditions. In this context the braided channel system occurring in the alluvial fan area needs to be highlightened as here the numerous anabranches make it most likely that the cross profiles recorded are not complete. Additionally, a potential error applying the Costa-approach is the incorrect allocation of pebble clusters to a certain channel or terrace level. Validity check To check the quality of the estimated runoff values the catchment-influenced specific peak discharge [l/(s * km²)] and the height of peak discharge [mm/h] are calculated corresponding to standardized values. The results are calculated in an hourly resolution and show the effective amount of water that leads to surface runoff measured in the cross profiles corresponding to a pour-point as an imaginary gauging station. The specific peak discharge decreases from the upper course to the lower course (cf. table 3), a behaviour well known for catchments in dry regions (Graf, 1983). The values derived for the specific bankfull discharge show the same behaviour.

Fluvial geomorphology and palaeohydrology of a small tributary, NE Niger 149

Figure 8. Results of the flow velocity and discharge calculation with the Manning-Strickler approach and the Costa approach for five selected channel cross profiles in the present channel bed (FB).

Table 3. Calculation of the specific peak discharge [l/(s * km²)] and the height of peak discharge [mm/h] for five selected channel cross profiles in the present channel bed (FB) using the Costa approach. (for locations see figures 5 and 6).

Q 12 Q4 Q2 Q 1.1 Q1

q [l/(s * km²)]

hA [mm/h]

2.669,35 2.601,10 2.401,59 2.118,20 2.165,38

13,21 2,16 1,45 0,43 0,60

To validate the output the input needs to be considered. Table 4 shows the precipitation data of the three neighbouring meteorological stations Bilma (Niger), Djanet and Tamanrasset (both Algeria). The three maximum precipitation events since 1988 of each station are extracted from the Globalsod database (http://gcmd.nasa.gov/records/ GCMD_GLOBALSOD_NCDC.html). The authors are aware that a comparability of

150

Krause and Schütt Table 4. The three maximum daily precipitation values of the meteorological stations Bilma (Niger), Djanet and Tamanrasset (both Algeria) (data source: Globalsod 1994–2004).

Bilma 18°41'N 21°55'Ε 102,1 mm/d (01/1994) 62,0 mm/d (06/2003) 16,0 mm/d (08/2004)

Djanet 24°33'Ν 9°29'E

Tamanrasset 24°02'N 5°57'E

99,1 mm/d (05/2002) 79,0 mm/d (08/1994) 70,1 mm/d (02/1995)

69,1 mm/d (05/2002) 48,0 mm/d (02/1998) 41,9 mm/d (04/1996)

these meteorological stations and topographic positions is restricted. However, a statistical analysis of the daily precipitation data from the stations Tamanrasset and Bilma shows comparable character of rainfall data. Nevertheless, the maximum daily precipitation at Bilma weather station with 102,1 mm in January 1994 indicates, that this rainfall event was brought by the westerlies such as by a cold pool effect (Línes Escardo, 1970). However, statistic shows that winter rainfall in Bilma is most unlikely also in comparison to the stations Djanet and Tamanrasset, which mostly have extreme precipitation during Northern Hemisphere winter or spring (Wininger, 1975). While at the weather stations Djanet and Tamanrasset precipitation is brought by the westerlies, in Bilma it is brought by the SW monsoon (Dubief, 1963). All precipitation data shown in table 4 are daily scaled, but most likely rainfall amounts resulted from singular events, which last from 45 minutes in the northern summer term up to 6 hours in the northern winter term (Dubief, 1971). In the following based on the maximum precipitation data as shown in table 4 a four-hours lasting precipitation event of 65 mm in total is assumed to come about the Seeterrassental. Due to the rocky character of the drainage basin’s headwater area and the altogether strong relief, short reaction time is assumed for runoff generation. Corresponding to this hypothesised short time-lag direct response of runoff generation on precipitation is assumed. Thus, it is assumed that a four hour rainfall event with 65 mm in total effects direct runoff with a duration of approximately four hours. Corresponding to the calculations, drainage basin related calculations of peak discharge height [mm/h] a four hour flood would lead to a four hours lasting peak discharge height of 13,21 mm/h in the headwater area (Q12)—totalling 52,84 mm runoff in four hours. In this case 81% of the fallen precipitation would lead to direct runoff. This value corresponds to the runoff coefficient and decreases downstream up to 3%. This decrease displays the change in underlying material from the rocky headwater catchment to the accumulation areas of the braided streamlet system in the lower course. In the lower course catchments the runoff response to precipitation decreases due to higher infiltration rates of accumulation area as well as the buffering effect of anabranch situation in the braided streamlet system. Belz (2000) shows monthly runoff coefficients for a Maghreb catchment from 7% to 34%. Puigdefabregas (1999) calculates for an akin event of 66 mm, but with a reduced intensity, for the ephemeral Rambla Honda test site (SE-Spain) runoff coefficients up to 19%.

Fluvial geomorphology and palaeohydrology of a small tributary, NE Niger 151

Discharge calculation of the palaeochannels To estimate the mean peak discharge and bankfull discharge of the palaeochannels, hydraulic conditions of these palaeochannels as deducible from fluvial terraces were used. In the Seeterrassental the cross profile shows allover a uniform shape (Figure 7): Along both river banks above the present day channel bed corresponding terrace levels can be detected, which indicate older channel beds. As evidence of the uniform shape the longitudinal profiles of the left and right bank of each terrace level run parallel. For discharge gross ups the cross sectional area of the successional, younger channel is not included as it is assumed that the successional channel incised into the bed of the preceding channel. Further on, it is implied that the catchment area of the Seeterrassental did not vary significantly compared to the present day situation. The longitudinal inclination of the palaeochannel is derived from differential GPS length profiles measurements of the terrace levels T1 and T2 (Figure 7). Additionally, channel cross profile measurements as shown in figure 6 also complete the measurement of the terrace levels T1 and T2. Like for the estimations of present day discharge (see paragraph above; cf. figure 8) a downstream decrease in discharge can be noted for the older channel generations as well. This is distinctly shown by the discharge estimations for the channel corresponding to the terrace level T1 (cf. figure 9, left). The oldest channel generation considered here (T2, cf. figure 9, right) also shows a downstream runoff decrease, but more moderate than to be observed presently and for terrace level T1. For both palaeochannel generations T1 and T2, again the differences between the bankfull discharge estimations (ManningStrickler) and the peak discharge estimations (Costa) are obvious (Figure 9). The ratio between both values are consistently smaller than for the values of the most recent channel, corresponding to more adjusted values between peak discharge and bankfull discharge. Values for T1 range from 31% in the upper course to 80% in the lower course, has mostly higher percentual differences than T2, where they increase from 10% to 55,5%. However, in contrast to the present day channel runoff coefficients increase downstream. In general, both palaeochannel generations show higher discharge values than the most recent channel. The topmost channel shows the highest discharge values.

Figure 9. Results of the flow velocity and discharge calculation with the Manning-Strickler approach and the Costa approach for five selected channel cross profiles for the channel bed corresponding to the first (T1, left) and the second terrace level (T2, right).

152

Krause and Schütt

Slack water deposits at cross section Q11 (Figure 6) are dated for a maximum age of 9.480 a cal BP for the basal slack water sediments and of 7.840 a cal BP for the youngest slack water sediments (Grunert, 1983). Correspondingly, incision took place after 7.840 a cal BP. As Terrace levels T1 and T2 also can be found in cross profile Q11 these terrace levels are younger than slack water deposition. Discharge values highest for the assumed oldest palaeochannel, and showing lower values for the successional palaeochannel in T1 correspond to the idea of an aridisation of the Sahara since the Early Holocene, corresponding to decreasing precipitation and according to this decreasing runoff rates (among many others Baumhauer et al., 2004; Gasse et al., 1990; Renssen et al., 2006; Flohn and Nicholson, 1980; Geyh and Jäkel, 1973; Servant and Servant-Vildary, 1980; Gasse and Van Campo, 1994; Foley et al., 2003; Lezine et al., 1990; Pachur and Hoelzmann, 1991; de Menocal et al., 2000). Relatively small difference between estimated peak discharge and bankfull discharge for terrace levels T1 and T2 in comparison to the most recent channel FB indicates that magnitude of flood events was less pronounced than today (Zielhofer et al., 2004). Data indicate a trend of an increasing discrepancy between magnitude of flood events and bankfull discharge from past to present. Again, this points to the Holocene aridisation of the area coinciding with increasingly erratic rainfall and an increasing variability of daily rainfall magnitude (Weischet and Endlicher, 2000; Nicholson, 2001, 2005). 8.4 CONCLUSIONS It is well known that highest flood events occur in the drylands. Crippen and Bue (1977) and Costa (1987) document that the highest maximal floods in the USA occur in the arid and semiarid zones. Such investigations are unkown for the Sahara, the largest successional desert on Earth. Discharge estimations of the most recent channel FB and the two preceding palaeochannel generations T1 and T2 give a first insight into runoff behaviour of a small catchment in the present day hyperarid Central Sahara. Comparable studies by Schick and Lekach (1987) and Greenbaum et al. (2000) show similar results, which underline that the range of the estimation carried out in this paper, are plausible. However, temporal matching of the palaeodischarge is required. For dating landscape development, slack water sediments deposited in the backwater situation of landslides were sampled for radiocarbon dating within the Seeterrassental (corresponding to location Q11, figure 6) by Grunert (1983) and Baumhauer in 2006 (unpublished). Dating results point to ages of maximum 9.480 a cal BP for the basal slack water sediments and of 7.840 a cal BP for the youngest slack water sediments (Grunert, 1983). Thus, increased land slide dynamics and the deposition age of slack water deposits correlate with the Early Holocene wet phase or African Humid Period (AHP). From theses age determinations it can be concluded that the channel incision, as already visible for terrace level T2, is post-Early Holocene. Renssen et al. (2006) hindcasts precipitation levels of about 290 mm/a for the periods 9 to 7,5 ka cal BP. In this phase the Western Saharan and Sahelian climate was affected by (relatively) high summer temperatures and enhanced precipitation. Grunert et al. (1991) as well as Flohn and Nicholson (1980) confirm this by precipitation estimations of 300–400 mm/a and 250–400 mm/a, respectively. The increased humidity was caused by the intensification of the African monsoon due to changes of the orbital parameters, with a maximum around 10,6 ka cal BP (de Menocal et al., 2000). Several authors estimate that precipitation over North Africa was 40% higher than today (de Menocal et al., 2000; Kutzbach and Guetter, 1986; Prell and Kutzbach, 1987). As the datings prove the deposition events, the erosion events can only be estimated. Nevertheless it is possible that after deposition precipitation was still sufficient for the first insection phase, for example, the channel generation corresponding to terrace T2.

Fluvial geomorphology and palaeohydrology of a small tributary, NE Niger 153

For the Mid Holocene wet period with three arid-to-humid transitions Damnati (2000) hindcasts 210 mm precipitation per year. This value is confirmed by the model outputs by Kutzbach and Guetter (1986) with estimated precipitation values of 110–200 mm/a. In comparison with the wet phases a link between the first terrace level to the Mid Holocene wet phase is possible but cannot be proven. Beside these major wet phases the annual precipitation characteristics show a decrease after the AHP. At the same time an increase in precipitation variability can be identified (Renssen et al., 2006). The present channel is a result of more or less present rainfall events, documented by the sand structure and the lack of erosion. The meteorological databases from WMO and Globalsod show precipitation values of about 150 mm/a for the 1950–2000 period (Vose et al., 1992). In summary, annual precipitation decreased continuously during the Holocene. Nevertheless, an increase of precipitation variability can be assumed. The two approaches applied to estimate flow velocity provide different information: the Manning-Strickler approach provides estimations of flow velocity during bankfull discharge, while the Costa approachs provides information on the flow velocity during peak discharge. Proportion between both values points out that estimated values are plausible. Errors in discharge calculation as well applying both approaches for flow velocity estimation are given by uncertainties estimating the margins of bankfull discharge in the cross profiles investigated. Next to this, the multitude of anabranches in the middle and especially lower course of the Seeterrassental generate high inaccuracies by cross-profile measurements. A main character of the channel bed is the frequent change of erosion and accumulation zones from the headwater area to the apex of the fan. This causes a sequence of sediment traps corresponding to a sediment cascade along the channel, alternating in magnitude and frequency. The channel length-discharge relationship shows upstream an increase of discharge, changing in its middle course into downstream continuous declining discharge while downstream the drainage area continuously increases. These results show an eventinduced change of the channel characters. In the upper course the discharge behavior is characterized by short reaction times due to the bedrock character of the subcatchment and the strong relief differences, leading to runoff coefficients from about 80%. The strong change in underlying material combined with repeatedly emerging damming situations along the valley as documented by the landslide remnants cause a decrease of the runoff coefficients. A specific feature is the downstream decrease of discharge as documented by both, the Manning-Strickler and the Costa approach. The drainage basin geomorphology gives reason: In the upper course the channel runs in bedrock, Cretaceous sandstones with poor infiltration capacity. In contrast in the middle and lower course the channel is running in alluvial accumulation bodies, which have high infiltration rates and high pore volume. Consequently it is assumed that downstream decreasing runoff is caused by water loss due to infiltration. An influence of evaporation on the downstream decreasing runoff cannot be excluded (Schick, 1988); however, references on such processes are poorly available. Linking of channel morphometry with palaeoclimate data is the prerequisite to identify past runoff dynamics. Coupling these data with other palaeoenvironmental proxies in future will increasingly and enhanced provide information on the effect of changing environmental parameters on process dynamics. ACKNOWLEDGEMENTS We would like to thank the German Research Foundation (DFG) for their financial support of the ‘Limnosahara’ research project (Schu 949/8), funded since 2005. Thanks are also due to Nicole Marquardt and Katharina Ducke for their great support in preliminary work within this project.

154

Krause and Schütt

REFERENCES Abrahams, A.D., Li, G. and Parsons, A.J., 1998, Rill hydraulics on a semiarid hillslope, Southern Arizona. Earth Surface Processes and Landforms, 21, 1, pp. 35–47. Ahnert, F., 1996, Einführung in die Geomorphologie (Stuttgart: Verlag Eugen Ulmer), pp. 1–440. Baker, V.R., Bowler, J.M., Enzel, Y. and Lancaster, N., 1995, Late Quaternary palaeohydrology of arid and semi-arid regions. In Global Continental Palaeohydrology, edited by Gregory, K.J., Starkel, L. and Baker, V.R. (Chichester), pp. 203–231. Barbour, M.T., Gerritsen, J., Snyder, B.D. and Stribling, J.B., 1999, Rapid bioassessment protocols for use in streams and wadeable rivers: periphyton, benthic macroinvertebrates and fish. Second edition, EPA 841-B-99-002, U.S. Environmental Protection Agency, Office of Water, Washington, D.C. Baumhauer, R., 1991, Palaeolakes of the South Central Sahara: problems of palaeoclimatological interpretation. Hydrobiologia, 214, pp. 347–357. Baumhauer, R., Schulz, E. and Pomel, S., 2004, Environmental changes in the Central Sahara during the Holocene—the Mid-Holocene transition from freshwater lake into sebkha in the Seggedim depression, NE-Niger. Lecture Notes in Earth Sciences, 102, pp. 33–47. Baumhauer, R. and Hagedorn, H., 1990, Problems of ground water capture in the Kawar (Niger). Applied Geography and Developement, 36, pp. 99–109. Belz, S., 2002. Nutzung von Landsat Thematic Mapper Daten zur Ermittlung hydrologischer Parameter. Mitteilungen des Institutes für Wasserwirtschaft und Kulturtechnik der Universität Karlsruhe (TH), 206, pp. 1–150. Busche, D., 1998, Die zentrale Sahara—Oberflächenformen im Wandel, (Gotha: Justus Perthes Verlag), pp. 1–284. Chow, V.T., Maidmant, D.R. and Mays, L.W., 1988, Applied Hydrology, (New York: McGraw-Hill Book Company), pp. 1–572. Chow, V.T., 1959, Open-channel hydraulics, McGraw-Hill Book Company, New York. pp. 1–680. Costa, J.E., 1987, Hydraulics and basin morphometry of the largest flash floods in the conterminous United States. Journal of Hydrology, 93, pp. 313–33. Costa, J.E., 1983, Paleohydraulic reconstruction of flash-flood peaks from boulder deposits in the Colorado Front Range. Geological Society of American Bulletin, 94, pp. 986–1004. Crippen, J.R. and Bue, C.D., 1977, Maximum floodflows in the conterminous United States. US Geological Survey Water-supply Paper, 1887, pp. 1–52. Damnati, B., 2000, Holocene lake record in the Northern Hemisphere of Africa. Journal of African Earth Sciences, 31, 2, pp. 253–262. De Menocal, P., Ortiz, J., Guilderson, T., Adkins, J., Sarntheim, M., Baker, L. and Yaruskinsky, M., 2000, Abrupt onset and termination of the African Humid Period: Rapid climate response to gradual insolation forcing. Quaternary Science Reviews, 19, pp. 347–361. Dubief, J., 1971, Die Sahara, eine Klima-Wüste. In Die Sahara und ihre Randgebiete— Darstellung eines Naturraumes, I. Band Physiogeographie, edited by Schiffers, H. (München: Weltforum Verlag), pp. 227–348. Dubief, J., 1959/1963, Le climat du Sahara. Université d´Alger. Institution de Recheres Sahariennes, 1 (1959), pp. 1–312, 2 (1963), pp. 1–274. Enzel, Y., Bookman, R., Sharon, D., Gvirtzman, H., Dayan, U., Ziv, B. and Stein, M., 2003, Late Holocene climates of the Near East deduced from Dead Sea level, variations and modern regional winter rainfall. Quaternary Research, 60, pp. 263–273.

Fluvial geomorphology and palaeohydrology of a small tributary, NE Niger 155

Felix-Hennigsen, P., 1992, Frühholozäne Feuchtzeitböden auf Altdünen der Ténéré und des Tchigai-Berglandes, Ost-Niger. Würzburger Geographische Arbeiten, 84, pp. 97–129. Flohn, H. and Nicholson, S., 1980, Climatic fluctuations in the arid belt of the “Old World” since the Last Glacial Maximum; possible causes and future implications. Palaeoecology of Africa, 12, pp. 3–22. Foley, J.A., Coe, M.T., Scheffer, M. and Wang, G., 2003, Regime shifts in the Sahara and Sahel: Interactions between ecological and climatic systems in Northern Africa. Ecosystems, 6, pp. 524–539. Gasse, F. and van Campo, E., 1994, Abrupt post-glacial climate events in West Asia and North Africa Monsoon domains. Earth Planetary Science Letter, 126, pp. 435–456. Gasse, F., Téhet, R., Durand, A., Gilbert, E. and Fontes, J.C., 1990, The arid-humid transition in the Sahara and the Sahel during the last deglaciation. Nature, 346, pp. 141–146. Geyh, M.A. and Jäkel, D., 1973, Late Glacial and Holocene climatic history of the Sahara desert derived from a statistical assey of 14C dates. Palaeogeography Palaeoclimatology Palaeoecology, 15, pp. 205–208. Golden, L.A. and Springer, G.S., 2006, Channel geometry, median grain size, and stream power in small mountain streams. Geomorphology, 78, pp. 64–76. Graf, W.L., 1983, Flood related changes in an arid region river. Earth Surface Processes and Landforms, 8, pp. 125–141. Greenbaum, N., Schick, A.P. and Baker, V.R., 2000, The palaeoflood record of a hybrid catchment, Nahal Zin, Negev Desert, Israel. Earth Surface Processes and Landforms, 25, pp. 951–971. Gregory, K.J., 1976, Drainage networks and climate. In Geomorphology and climate, edited by Derbyshire, E. (London: John Wiley & Sons), pp. 289–315. Greigert, J. and Pougnet, R., 1967, République du Niger, carte géologique 1:2 Mio., Bureau des Recherches Géologique et Minières, Niamey, Niger. Grunert, J., 1983, Geomorphologie der Schichtstufen am Westrand des Murzuk-Beckens (zentrale Sahara). Relief, Boden, Paläoklima, 2, (Berlin, Stuttgart), pp. 1–271. Grunert, J., Baumhauer, R. and Völkel, J., 1991, Lacustrine sediments and holocene climates in the Southern Sahara: The example of palaeolakes in the Grand Erg of Bilma (Zoo Baba and Dibella, Eastern Niger). Journal of African Earth Science, 12, 1–2, pp. 133–146. Gurgel, K.W., 1991, Erfahrungen mit dem Satelliten-Navigationssystem, GPSGenauigkeiten an Land und auf See. Ocean Dynamics, 44, 1, pp. 35–49. Gurnell, A.M. and Gregory, K.J., 1995, Interactions between semi-natural vegetation and hydrogeomorphologic processes. Geomorphology, 13, pp. 49–69. Institut Géographique National (IGN), Photothèque Nationale, 1955, République du Niger, mission aérienne NF 33 XIII (Achelouma), Numbers: 167–169, 194–197, 214 and 215. Jones, J.J.A., 1997, Global Hydrology—Processes, resources and environmental management, (Essex: Pearson Education Limited), pp. 1–399. Kieffer, S.W., 1990, Hydraulics and geomorphology of the Colorado River in the Grand Canyon. In Grand Canyon Geology, edited by Beus, S. and Morales, M. (New York: Oxford University Press), pp. 333–383. Kim, J.-W., 1989, Funktionale Fluvialmorphologie der Kall. Aachener Geographische. Arbeiten, 21. Kuhn, N.J., Yair, A. and Kasanin-Grubin, M., 2004, Spatial distribution of surface properties, runoff generation and landscape development in the Zin Valley Badlands, Northern Negev, Israel. Earth Surface Processes and Landforms, 29, pp. 1417–1430. Kutzbach, J.E., and P.J. Guetter, 1986, The influence of changing orbital parameters and surface boundary conditions on climate simulations for the past 18.000 years. Journal of the Atmospheric Sciences, 43, 16, pp. 1726–1759.

156

Krause and Schütt

Langbein, W.B., 1947, Topographic characteristics of drainage basins. United States Geological Survey, Water-Supply Paper, 968, C, pp. 125–157. Lezine, A.-M., Casanova, J. and Hillaire-Marcel, C., 1990, Across an early Holocene humid phase in Western Sahara, pollen and isotope stratigraphy. Geology, 18, pp. 264–267. Lines Escardo, A., 1970, The climate of the Iberian Peninsula. In Climates of Northern and Western Europe, edited by Wallén, C.C. (Amsterdam: Elsevier Publishing Company), pp. 195–239. McCoy, R.M., 1971, Rapid measurement of drainage density. Bulletin of the Geological Society of America, 82, pp. 757–761. Melton, M.A., 1957, An analysis of the relations among elements of climate, surface properties and geomorphology. Office of Naval Research, Geography Branch, Project NR 389–042: Technical Report 11, Columbia University. Montgomery, D.R. and Gran, K.B., 2001, Downstream variations in the width of bedrock channels. Water Resources Research, 37, 6, pp. 1841–1846. Naudascher, E., 1992, Hydraulik der Gerinne und Gerinnebauwerke, Springer Verlag, Wien, New York, 2nd Edition. Nicholson, S., 2001, Climatic and environmental change in Africa during the last two centuries. Climate Research, 17, pp. 123–144. Nicholson, S., 2005, On the question of the ‘‘recovery’’ of the rains in the West African Sahel. Journal of Arid Environments, 63, pp. 615–641. Obi Reddy, G.P., Maji, A.K. and Gajbhiye, K.S., 2004, Drainage morphometry and its influence on landform characteristics in a basaltic terrain, Central India—a remote sensing and GIS approach. International Journal of Applied Earth Observation and Geoinformation, 6, pp. 1–16. Oguchi, T., 1998, Drainage density and relative relief in humid steep mountains with frequent slope failure. Earth Surface Processes and Landforms, 22, 2, pp. 107–120. Pachur, H.-J. and Altmann, N., 2006, Die Ostsahara im Spätquartär—Ökosystemwandel im größten hyperariden Raum der Erde, (Berlin, Heidelberg: Springer Verlag), pp. 1–662. Pachur, H.J. and Hoelzmann, P., 1991, Paleoclimatic implications of Late Quaternary lacustrine sediments in Western Nubia, Sudan. Quaternary Research, 36, pp. 257–276. Prell, W.L. and Kutzbach, J.E., 1987, Monsoon variability over the past 150.000 years. Journal of Geophysical Research, 92, pp. 8411–8425. Renssen, H., Brovkin, V., Fichefet, T. and Goosse, H., 2006, Simulation of the Holocene climate evolution in Northern Africa: The termination of the African Humid Period. Quaternary International, 150, pp. 95–102. Reid, I., 2002, Sediment dynamics of ephemeral channels. In Dryland rivers: hydrology and geomorphology of semi-arid channels, edited by Bull, L.J. and Kirkby, M.J. (John Wiley & Sons, Ltd.), pp. 107–128. Richards, K.S., 1982, Rivers: form and process in alluvial channels, (London: Methuen) pp. 1–361. Servant, M. and Servant-Vildary, S., 1980, L’ environment Quaternaire du Bassin du Tchad. In Sahara and the Nile, edited by Williams, M.A. and Faure, H., pp. 133–163. Schick, A.P., 1988, Hydrologic aspects of floods in extreme arid environments. In Flood geomorphology, edited by Backer, V.R., Kochel, R.C. and Patton, P.C., (Jon Wiley & Sons), pp. 189–230. Schick, A.P. and Lekach, J., 1987, A high magnitude flood in the Sinai desert. In Catastrophic Flooding, edited by Mayer, L. and Nash, D., (Winchester, MA: Allen and Unwin), pp. 381–410. Schulz, E., 1988, Der Südrand der Sahara. Würzburger Geographische Arbeiten, 69, pp. 167–210.

Fluvial geomorphology and palaeohydrology of a small tributary, NE Niger 157

Schulz, E., 1980, Zur Vegetation der östlichen Sahara und zu ihrer Entwicklung im Holozän. Würzburger Geographische Arbeiten, 51, pp. 125–162. Schumm, S.A., 2005, River variability and complexity. Cambridge University Press, pp. 1–220. Schumm, S.A., 1960, The shape of alluvial channels in relation to sediment type. United States Geological Survay, Professional Paper, 352, B, pp. 17–30. Schwaller, G. and Tölle, U., 2005, Einfluss von Maßnahmen der Gewässerentwicklung auf den Hochwasserabfluss. Bayrisches Landesamt für Wasserwirtschaft (Ed.), Materialien, 122. Stewardson, M., 2005, Hydraulic geometry of stream reaches. Journal of Hydrology, 306, pp. 97–111. Vose, R.S., Schmoyer, R.L., Steurer, P.M., Peterson, T.C., Heim, R., Karl, T.R. and Eischeid, J., 1992, The Global Historical Climatology Network: long-term monthly temperature, precipitation, sea level pressure, and station pressure data. ORNL/ CDIAC-53, NDP-041. Carbon Dioxide Information Analysis Center, Oak Ridge National Laboratory, Oak Ridge, Tennessee. Wagenschein, D., 2006, Einfluss der Gewässermorphologie auf die Nährstoffretention— Modellstudie am Beispiel der mittleren Weißen Elster. Dissertation, Cottbus, Brandenburgische Technische Universität Cottbus. pp. 1–110. Weischet, W. and Endlicher, W., 2000, Regionale Klimatologie (Stuttgart, Leipzig: Teubner). Wininger, M., 1975, Bewölkungsuntersuchungen über der Sahara mit Wettersatellitenbildern. Geographica Bernensia, Reihe G, 1, pp. 1–149. Willgalis, S., 2005, Beiträge zur präzisen Echtzeitpositionierung in GPS Referenzstationsnetzen, Dissertation, Hannover, pp. 1–169. Zielhofer, C., Faust, D., Escudero, R.B., Diaz del Olmo, F., Kadereit, A., Moldenhauer, K.-M. and Porras, A., 2004, Centennial-scale Late-Pleistocene to Mid-Holocene synthetic profile of the Medjerda Valley, Northern Tunisia. The Holocene, 14, 6, pp. 851–861.

CHAPTER 9

Palaeoecology of the Giant Catfish (Arius gigas, Ariidae) in Holocene Saharan and tropical West African waters Hélène Jousse Naturhistorisches Museum Wien, Säugetiersammlung, Wien, Austria Wim Van Neer Royal Belgian Institute of Natural Sciences, Brussels, Belgium and Katholieke Universiteit Leuven, Laboratory of Animal Biodiversity and Systematics, Leuven, Belgium ABSTRACT: The Giant Catfish Arius gigas is an endemic species of West African freshwaters that is almost extinct today, and its way of life is poorly known to ichthyologists. However, this species is known from the Holocene archaeofaunal record, in particular from the Niger basin. The skeletal anatomy of the Giant Catfish described in this paper should facilitate its future identification within palaeo-ichthyological assemblages. In addition, the species’ occurrence is studied from a palaeogeographical and palaeoecological point of view. A. gigas certainly has ecological requirements similar to the related large carnivorous fish inhabiting well oxygenated waters, and would not tolerate shallow, muddy and stagnant ecotopes of marginal waterways. By over fishing such a large species, humans contribute to the lowering of its reproduction potential, and to its recent drastic decline.

9.1 INTRODUCTION The Giant Catfish, Arius gigas Boulenger 1911 is the only freshwater fish of the family Ariidae (Siluriformes), all the other species being marine. It is endemic to West Africa where it has been captured in the Niger, Benue, Volta and Ouémé Rivers (Figure 1). According to Daget (1988) and Lévêque (1999a), the barrier formed by the Gauthiot Falls on the Benue system explains why some species from the Niger that exist in the Benue, including A. gigas, never reached the Chad basin, or other Eastern rivers, during more humid spells. Daget and Stauch (1963) mentioned that fishermen were still fishing A. gigas in the mouth of the Faro River in Cameroon. Arius gigas has an elongate and rounded body, a slightly flattened and very wide head, with a moderate median keel. The cranial roof bones are ornamented by rugosities on their external surfaces. The characters in the gross anatomy that differentiate the species from other Ariidae are the triangular shape of the palatine tooth plates, and the large and distally truncated occipital process (Daget, 2003). Body lengths of about two metres have been reported in the Niger River during the 20th century (Daget et al., 1988), and the present study will show that it also reached that size in the past.

160

Jousse and Van Neer

Figure 1. Distribution of the Giant Catfish in modern West African rivers after Daget (2003) and during Holocene times. NID: Niger Inland Delta.

Specimens are very rare today, and this was already the case in the 1950s in the Upper Niger (Daget, 1954). At that time it was still common downstream of Lake Debo, in the Niger Inland Delta in Mali. Very few specimens of A. gigas are preserved in natural history collections, and they are mainly preserved in alcohol, preventing observation of their skeleton. A dry skeleton of a modern specimen in the British Museum collections and the numerous remains of A. gigas from the Neolithic site of Kobadi in Mali gave a unique opportunity to analyse the skeletal anatomy of the species. Some characteristics of its palaeoecology, including size and geographical distribution, are inferred from its occurrences in past faunal assemblages in West Africa. The archaeofaunal data demonstrate that the species prospered during the Holocene, and that it was a valuable food resource for humans in the past. 9.2 MATERIAL AND METHODS The skeletal anatomy of Arius gigas has never been studied in detail. This may be due to its rarity and endemicity, and the resulting low number of specimens preserved in natural history collections. The present description of its osteomorphological features and the schematic drawings of its main bones are partially based on the specimen BMNH 1863.12.9.2 of the British Museum in London. Since this specimen was formerly mounted for an exhibition, its skeleton is still partially articulated. As a result not all the articular surfaces could be seen and some bones could not be observed from all sides. However, this is compensated by the Holocene site of Kobadi mentioned below, that yielded several bone elements with their articular surfaces intact that allowed observation of their morphology (Jousse et al., 2008). In order to address issues such as the former ecology and human exploitation of the Giant Catfish, it is crucial to document the species’ size through time. This can be done

Palaeoecology of the Giant Catfish (Arius gigas, Ariidae)

161

by direct comparison of fossil bones with modern specimens of known body length, or by using a regression formula that correlates a given bone measurement and fish length. In the case of other African freshwater taxa such as Lates niloticus, Tilapiini and Clarias sufficient modern specimens were available to calculate such regressions (Van Neer, 1989; Van Neer and Lesur, 2004). Of Arius gigas only one skeleton with known body length is available. Therefore, equations were established using several Ariidae species, a procedure that has been applied previously in other fish families (Desse and Desse-Berset, 1996). All available Ariidae skeletons of individuals with known body length stored in the collections of the Institut für Paläoanatomie und Geschichte der Tiermedizin of Munich, of the Royal Belgian Institute of Natural Sciences in Brussels, and of the Centre d’Etudes Préhistoire, Antiquité, Moyen Age (CÉPAM) at Valbonne, Sophia Antipolis, were measured. This material included a total of 21 skeletons: Arius thalassinus (8 specimens), Arius bilineatus (2), Arius parkii (2), Arius heudelotii (1), Arius latiscutatus (1), Arius maculatus (1), Arius gigas (1) and Arius sp. (5). The Arius gigas BMNH specimen was very helpful since it represents the only specimen with a size exceeding one metre (129 cm). Fish grow throughout life, following an exponential growth model. When they reach very large sizes, they tend to become greater in volume, but body length increase becomes minimal. Because no very large Ariidae of about two metres are available in the visited collections, we lack the data that would extrapolate the maximum end of the size estimation regression. We are aware, therefore, of the fact that the equations calculated from modern specimens will tend to somewhat overestimate the size of the largest fish (Jonsson, 1994). The equations established for each bone measurement are beyond the scope of this paper, and will be published elsewhere. The past occurrence of A. gigas in Africa during the Holocene is documented by the following sites, from which the fish remains, with the exception of Ntereso, were analysed by the authors (Figure 1, table 1). Kobadi (Mali): The Sahelian site of Kobadi, located on the western edge of the former Niger Inland Delta, is particularly rich in ichthyofauna. At about 2.000–1.400 cal years BC, A. gigas (n = 1.080) constitutes the third most abundant species from the site, after the Nile perch Lates niloticus and the catfish Auchenoglanis sp. (Jousse et al., 2008). Some skeletal elements suggest the presence of very large specimens, up to two metres in length. Hassi el Abiod (Mali): Within the faunal remains of the Neolithic sites near Hassi el Abiod in the Malian Sahara, A. gigas is abundant (n = 401). Numerically speaking, it constitutes the second or third most important fish taxon exploited by humans, after Lates niloticus and the catfish Clarias sp. (Van Neer and Gayet, 1988). The region was formerly linked to the Niger Basin between 6.000 to 2.000 cal years BC, when the climate was more humid than today, and the Niger floods inundated northerly located lacustrine areas (PetitMaire and Gayet, 1984). Tiabel Goudodie (Mali): This site belongs to the same cultural tradition as Kobadi, and dates to 2.570–2.350 cal years BC. It yielded some remains of A. gigas, but the detailed study has not been completed and no quantitative data are available. Its occurrence, however, confirms the species’ presence in the Niger Inland Delta during the Late Holocene (MacDonald, 1994; MacDonald and Van Neer, 1994). Kolima (Mali): In the site of Kolima, also linked to the former expansion of the Niger Inland Delta, A. gigas is very scarce. Its presence is attested by just one bone, in level 2 of Kolima Unit 2 (Van Neer, personal communication). This site is slightly younger than Kobadi, about 1.300 to 500 cal years BC (Phase II) (MacDonald, 1994; MacDonald and Van Neer, 1994). Ntereso (Ghana): The site of Ntereso belongs to the Kintampo complex and lies near the White Volta River. The presence of Giant Catfish has been reported from the fauna at

Genus/ species

Polypteridae Polypterus sp. Osteoglossidae Heterotis niloticus Mormyridae Mormyrus sp. Hyperopisus bebe Mormyrops sp. Gymnarchidae Gymnarchus niloticus incl. Characidae Alestes/Brycinus Hydrocynus sp. Citharinidae Distichodus sp. Citharinus sp. Cyprinidae Labeo sp. Barbus sp. Siluriformes Bagridae Bagrus sp.

Family/ Order

2 – – 94 – 1

18 – 3 8 7 59 1 1.561 33 –

3 128

– 28

2

10 26 2

– – –

– – –

15 263

3 14

Hassi el Kobadi Abiod

1

– – – – – – – – – 1

– 1

– – –

–

Kolima KL2lev2

– – – – – – – – – –

– –

– – –

– –

Tiabel Goudodie

– + – – – – + – – –

– –

– – –

– –

Ntereso

0,1 – – – 0,1 – – 3,5 – 0,04

– 1,1

– – –

0,1 0,5

0,2 – 0,03 0,1 0,1 0,6 0,01 14,9 0,3 –

0,03 1,2

0,1 0,2 0,02

0,1 2,5

Hassi el Kobadi Abiod % % lev2 %

– – – – – – – – – 0,6

– 0,6

– – –

– 0,6

– – – – – – – – – –

– +

– – –

+ +

– – – – – – – – – +

– –

– – –

– –

Kolima Shallow Open KL2water water fish fish fish

Table 1. Arius gigas occurrences among Holocene fish fauna in number of specimens. For Hassi el Abiod, Kobadi and Kolima, percentages have been calculated of the taxa that are typical for open and shallow water, respectively.

162 Jousse and Van Neer

Van Neer & Gayet, 1988

References

–

Jousse et al., 2008

10.468 1.507 6.099 7.606

11.206

3.749 756 1

109 140 1.245 344 856 19 1.080 1

42 3 41 769 19 – 401 11 1025 142 62

–

–

2659 967 1131 2098

Synodontis sp. Schilbe sp. Arius gigas Parachanna obscura Lates niloticus Tilapinii Tetraodon lineatus

Bagrus cf. docmak Chrysichthys sp. Clarotes laticeps Auchenoglanis sp.

Total identified Total shallow water fish Total open water fish Total

Pisces

Tetraodont.

Perciformes

Clariidae Mochocidae Schilbeidae Ariidae Channidae

Bagridae

– – – –

– – – –

–

+ – –

+ – – –

– + + + + – + –

+

– – – – – – + –

–

unpubl., MacDonald Carter & Van Neer & Van Flight, 1994 1972

157 114 42 156

–

11 34 –

– – 12 78 18 – 1 –

–

– 46,1 53,9 –

–

38,5 5,3 2,3

1,6 0,1 1,5 28,9 0,7 – 15,1 0,4

–

– 19,8 80,2 –

–

35,8 7,2 0,01

1,0 1,3 11,9 3,3 8,2 0,2 10,3 0,01

–

– 73,1 26,9 –

–

7,0 21,7 –

– – 7,6 49,7 11,5 – 0,6 –

–

– – – –

–

– + –

– – – + – – – +

–

– – – –

–

+ – –

+ + + – + – – –

–

Palaeoecology of the Giant Catfish (Arius gigas, Ariidae) 163

164

Jousse and Van Neer

about 2.000–1.300 cal years BC, but no quantitative data exist (Carter and Flight, 1972; Gautier and Van Neer, 2005). 9.3 ANATOMY The skeletal anatomy of Arius gigas is very similar to that of A. heudelotii and A. latiscutatus, marine members of the family inhabiting the African Atlantic coast, although they do not reach such large sizes. For faunal analysts acquainted with African freshwater fish, the bones of A. gigas also recall those of Bagrus, the two taxa presenting very similar anatomical features. The following descriptions focus on the skeletal elements that are most frequently preserved in the fossil record and that are the most diagnostic for identification. Atlantico-African Ariidae have a skull that is shorter than the very elongated one of the Pacific Ocean form A. thalassinus. The supraoccipital (soc) of Arius gigas (Figure 2) is relatively short and moderately vaulted. Its posterior part has a truncated triangular shape. The temporal fossa (ft) is very large between the pterotic (pto) and the posttemporal (ptm). The latter bone fully participates in the strong cephalic shield and contributes largely to the skull’s width. The morphology of the frontal (fr) is very variable among Ariidae species. That of A. gigas is elongated and bifurcates in its anterior part. There it forms a lateral, thin, bridge, leading to the prefrontal, and delineating a long fenestra. The left and right frontals are posteriorly connected by a short suture of about 1/5 of the mesial length of the bone. The rest of the frontal bones border the anterior fontanelle (fo), which is very long and wide. The prefrontal (pfr) is well developed laterally, with a strong protuberance that is slightly curved backwards. Compared to other Ariids, the dermethmoid (deth) is very flat and large. The two anterior transverse processes of the dermethmoid are flattened and there is a slight notch between them (Figure 2a-1). On the dorsal side of the dermethmoid is a ridge that runs parallel to the anterior margin of the bone (Figure 2a-2). The granulations that ornament the dorsal side of the cranial roof bones cover the back of the cephalic shield, and extend as far as the frontal. The granulations show radiating patterns, especially on the supraoccipital, sphenotic (sphot) and frontal. They also decorate the dorsal face of the nuchal shield (ns). Longitudinal grooves are particularly pronounced between the anterior part of the supraoccipital and the posterior part of the frontals. The prefrontal displays numerous ridges and a complex relief, but tubercles are lacking. The internal side of the skull displays the so-called “crucifix” shape, typical for Ariidae fish. This feature is particularly orthogonal in A. gigas. The posterior part is mainly characterised by a massive basioccipital (boc) and a wide parasphenoid (pas) (Figure 2b). The Weberian apparatus (Wb ap) is a complex formed by the fusion of the first four vertebrae, thought to help in supporting the weight of the skull (Gregory, 1959). Its development results in a posterior extension of the skull. The transverse processes of the vertebrae extend laterally beyond the skull border. The palatine tooth plates (pp) are triangular in shape (Figure 2b), a specific character mentioned by Daget (2003). The premaxillae (pmx) are reduced to two quadrangular plates (Figure 2c). The whole surface of the palatine and premaxillary plates is dentigerous. The morphology of the dentary (Figure 2e–g) has mainly been observed on the modern specimen of the BMNH, as this bone element is often broken in the fossil material. The dentary has a robust and thick keel (Figure 2f-3) on its inferior margin, that extends over the whole length of the bone as far as the lower aboral process (Figure 2f-4) that is in anatomical connection with the articular. As a result of the heavy development of this keel, the anterior

Palaeoecology of the Giant Catfish (Arius gigas, Ariidae)

165

Figure 2. Schematic drawing of Arius gigas cranial skeleton. a. Skull, dorsal view; b. Skull, ventral view. The left palatine plate was removed to observe the vomer below it; c. Left premaxilla, ventral view; d. Right articular, lateral view; e. Right lower jaw (dentary and articular), dorsal view; f. Right dentary, ventral view; g. Right lower jaw, medial view; h. Right quadrate, anterior view; i. Right quadrate, lateral view; a–i. from the BMNH specimen. See main text for the abbreviations.

166

Jousse and Van Neer

part of the dentary has a triangular section. The dentigerous surface of the dentary is wide, and narrows abruptly at its posterior end. The articular is short and in dorsal view numerous foramina (Figure 2e-5) are observed behind the articular surface. The quadrate has a wide articular condyle with a regular outline (Figure 2h–i). The opercular shows very dense ornamentation covering its whole external surface, except for a narrow, posterior, band. The inner surface is also heavily ornamented just below the articulation and the crista opercularis is particularly strong (Figure 3c-1). The anterior margin is very long and rectilinear. The processus supra-articularis is thick, and the articular fossa is elliptic compared to other Ariidae where it is much more elongated. As in the skull roof bones and the opercular, the external side of the humeral process of the cleithrum is also ornamented (Figure 4a, d). There is individual variation in the ornamentation, from low asperities to organised granulations that can interconnect and fuse into crests. Van Neer and Gayet (1988) illustrated such a cleithrum of a fossil A. gigas with crested ornamentation. The posterior margin of the humeral process (Figure 4c-1) is straight, whereas it is more bent in other Ariids. The dorsal fin spine is very massive (Figure 4e–f). Proximally, two major rows of granulations are visible but distally they gradually merge to finally form a single row. On the lateral side of the major rows of granulations, smaller rough patches and numerous striations are observed. Its ventral side is denticulate. The median articular process is strong, well rounded, and delimits a wide median foramen. The pectoral fin spine is also robust (Figure 4g). The articular condyle (Figure 4g-2) is thick and wide, and the dorso-lateral processus (Figure 4g-3) is well developed and high. The body of the spine is heavily compressed in a dorso-ventral direction. Granulations are only seen along the anterior side. As in the dorsal spine, there are two rows of granulations close to the articulation that gradually merge into a single row. On the dorsal and ventral sides, the ornamentation is limited to fine striations. The posterior side shows small indentations along the midline. Similar characteristics were observed by Van Neer and Gayet (1988) and Gayet and Van Neer (1990) on dorsal and pectoral fins of Arius gigas specimens from the Holocene sites in Northern Mali. With better preserved remains, they were able to describe age-related changes in the ornamental features of the spines. The first four vertebrae comprise the Weberian apparatus (Figure 5a). In the BMNH specimen, the vertebral column was composed of 48 vertebrae, with in addition the terminal urostyle. The centra of the first precaudal vertebrae have a very simple morphology. The outer surface is smooth, and there is much individual variation in the outline of the vertebral centra. They can be rounded, oval, generally compressed dorsoventrally, and they can also show polygonal outlines (Figure 5d). The ventral side is slightly curved inwards. The midpoint of the concentric growth lines of the centrum is situated in the dorsal part of the vertebra. More posteriorly located precaudal vertebrae become progressively more square in outline with the midpoint of the growth lines migrating towards the middle of the centrum (Figure 5e). With the parapophyses gradually taking a more ventral position (Figure 5e), the ventral side of the centres present more pronounced foramina (Figure 5c). The caudal vertebrae show less variation in shape (Figure 6a). They become progressively shorter towards the urostyle and the last three or four vertebrae are almost fused in the BMNH specimen. This individual also presents a fusion of two caudal vertebrae, which should be considered as a pathology (Figure 6a-1). On the ventral side of the centrum three elongated and parallel foramina occur, with the two lateral ones being occasionally divided in two (Figure 6d-2). The outline of the caudal vertebrae is well rounded, with very slightly inward curving dorsal and ventral sides (Figure 6c–f).

Palaeoecology of the Giant Catfish (Arius gigas, Ariidae)

167

Figure 3. Schematic drawing of Arius gigas bones. a. Left hyomandibular, lateral and medial views; b. Urohyale, dorsal view; c. Left opercular, lateral, anterior and medial views; d. Left interopercular, lateral and medial view; e. Right hyoid, medial, anterior and lateral views; f. Left basipterygium, ventral and dorsal view; a, c–f from the BMNH specimen; b from Kobadi fossil.

168

Jousse and Van Neer

Figure 4. Schematic drawing of Arius gigas bones. a–d. Left cleithrum, lateral view (a), medial view (b), detail of the articulation with the spine (c), detail of the ornamentation on the humeral process (d); e. Dorsal fin spine and nuchal shield, dorsal and left lateral views ; f. Dorsal fin spine, dorsal, ventral and right lateral views; g. Left pectoral fin spine, posterior, anterior, articular and proximal views; a, b, e from the BMNH specimen; c, d, f, g from Kobadi fossil remains.

Palaeoecology of the Giant Catfish (Arius gigas, Ariidae)

169

Figure 5. Schematic drawing of Arius gigas precaudal vertebrae. a. Weberian apparatus, caudal view; b. Anterior precaudal vertebrae, left lateral view; c. Anterior precaudal vertebrae, ventral view; d. Anterior precaudal vertebrae, cranial views; e. Precaudal vertebrae, cranial views; b, c from the BMNH specimen; a, d, e from Kobadi fossil remains.

170

Jousse and Van Neer

Figure 6. Schematic drawing of Arius gigas caudal vertebrae. a. Partial caudal series, left lateral view; b. Caudal vertebrae, ventral view; c–f. Caudal vertebrae, cranial/left lateral/ventral views. a, b from the BMNH specimen; c–f from Kobadi fossil remains.

From table 2, some observations can be made concerning the survival chances of the various Arius gigas bones in an archaeological context. Because fish were processed before consumption, and due to decomposition of the food remains after their disposal, no completely articulated skeletons were found. The only articulated parts that are sometimes preserved are series of the short precaudal vertebrae, which are sometimes found in a connected sequence of 4–6 vertebrae. The most robust elements are from the pectoral girdle (cleithrum, pectoral spine), and from the area immediately posterior to the skull (dorsal spine, precaudal vertebrae). Roof bones are well preserved, but also easily identified to that species because of their typical external ornamentation. The large size of most of the individuals also explains the excellent preservation of bones, since larger, well ossified, bones will better resist the various taphonomic processes. It is particularly the articulations, such as in the articular, quadrate or dentary that have the best preservation chances thanks to their robustness. 9.4 PRESENT AND PAST ECOLOGY According to Daget (2003), the Giant Catfish Arius gigas was already a threatened species in the 1950s, and it is now considered in danger of total extinction. Ichthyologists specialising in West African fish admit that the species is poorly known, and most of them have never even observed a specimen. Daget (1954) described the species from the Upper Niger, and reported that the stomach contents of a juvenile consisted of insects and fish scales. According to the accounts mentioned by Ligers (1966, p. 25), of fishermen in the

Palaeoecology of the Giant Catfish (Arius gigas, Ariidae)

171

Table 2. Preserved skeletal elements of Arius gigas at Kobadi (Mali) and the observed fire marks on the bone.

Skeletal element

n

cleithrum 100 precaudal vertebra 97 pectoral spine 87 roof bones 57 dorsal spine 49 articular 49 coracoid 46 caudal vertebra 44 dentary 43 quadrate 42 frontal 36 spine 30 vertebra 23 supraoccipital 20 posttemporal 20 sphenoticum 20 hyomandibular 18 pterygiophore 17 hyoid 14 basioccipital 13 basisphenoid 13 Weberian apparatus 13 maxilla 12 opercular series 10 symplecticum 5 vomer-mesethmoid 4 pteroticum 3 premaxilla 2 parasphenoid 2 Total

889

Not burnt

Partly black

Completely blackened Charred

97 91 80 49 47 48 45 44 41 42 34 26 20 15 19 18 14 15 14 13 12 12 11 9 4 3 3 2 2

– 2 – 1 – – – – – – – – – – 1 – – – – – 1 – 1 – – – – – –

3 4 5 5 1 1 1 – 2 – 2 3 1 4 – 2 1 1 – – – – – 1 1 1 – – –

– – 2 2 1 – – – – – – 1 2 1 – – 3 1 – – – 1 – – – – – – –

830

6

39

14

Niger Inland Delta in Central Mali, it is a large carnivorous fish that favours small fish such as Brycinus leuciscus. In 1992 an exceptionally large Giant Catfish was captured by fishermen using a dugout in the River Niger near Niamey, suggesting that the species still occurs in some parts of the Niger, away from its Inland Delta. The head of this specimen was donated to the Musée National du Niger at Niamey for its conservation. It is now on display in the office of its Director. The skull is preserved with the skin, and the dorsal and pectoral fins still connected (Figure 7). Measurements of the skull and some bones give an estimation of the specimen total length around 150 cm (Table 3). Although the species is believed to be present today in the Niger and the Volta Rivers, its rarity has resulted in poor knowledge of its ecological requirements. The fossil data can

172

Jousse and Van Neer

Figure 7. Skull of an Arius gigas specimen (dorsal and left lateral views) fished in the Niger River in 1992 and displayed in the Musée National du Niger in Niamey.

contribute to this issue, however, since there are some ichthyological assemblages where A. gigas occurs and sometimes dominates. With only five Holocene archaeological contexts of West Africa yielding skeletal remains, the record is not very rich (Table 1; Figure 1), but it illustrates the species’ wider distribution in West Africa during the Holocene. Among the catfish, it is one of the most frequently caught species in the Saharan sites near Hassi el Abiod and in every excavated level and square at Kobadi. At Kolima, which lies in the same palaeolacustrine environment as Kobadi, but which is younger, Arius gigas is only recorded once on the whole site. This low frequency cannot be explained exclusively by the lower total number of fish bones at Kolima compared to Kobadi: it appears that the Giant Catfish remains represent only 0,6% of all identified fish bones at Kolima, versus 10,3%

Palaeoecology of the Giant Catfish (Arius gigas, Ariidae)

173

Table 3. Size estimation of the Giant Catfish from the Musée National du Niger, Niamey.

Bone dimension

mm

Individual length estimation (cm)

Skull length Skull width (frontal) Skull width (sphenoticum) Supraoccipital width Dorsal spine length

475 145 169 93 215

145,07 171,05 151,24 119,12 150,9

Estimation mean

147,48

in Kobadi. It is more likely that this diachronic trend illustrates the change of hydrological conditions due to climatic degradation during the Late Holocene, which was inhospitable for the species. Another archaeofaunal fact that helps understanding the species’ ecology is the size that it reached formerly. For the specimens of Hassi el Abiod in Northern Mali, no detailed data are available concerning the size distribution of Arius gigas. It is, however, reported to range between 30–40 and 100 cm in length (Van Neer and Gayet, 1988). The unpublished bone of Kolima 2-level 2 was estimated to represent a very large specimen, of about 150 to 200 cm length (Van Neer, personal data). The only population that can provide an idea of its size distribution is the fauna from Kobadi. Total length was obtained from the regression formulae that were established for the Ariidae family, or—for fragmented bones—by direct comparison to Ariidae skeletons of known body length. The size distribution, in classes of 10 cm total length, follows a normal pattern, with a mean at 91 cm length (Figure 8). Some specimens occur of large size, exceeding 150 cm and, more exceptionally, reaching the size of two metres. This gives the species the status of one of the largest fish from African freshwaters, with only the Nile perch Lates niloticus being able to reach such a

Figure 8. Size (total length) profile of the Holocene population of Arius gigas at Kobadi (Mali, Niger Inland Delta).

174

Jousse and Van Neer

size. Compared to its rarity in modern ecosystems, this illustrates the importance of the species’ decline in the recent past. In the archaeofaunas, but also in fishing reports such as Ligers (1966), A. gigas is seen to be often associated with other large carnivorous fish such as Hydrocynus brevis, Gymnarchus niloticus, Auchenoglanis, Clarotes, Bagrus, and especially Lates niloticus. In the past record presented here (Table 1), Bagrus is lacking almost completely, while it is very common in the modern West African freshwaters. It is tempting to consider its near absence as a result of competition between these two catfish taxa. Several of the large predatory fish mentioned above inhabit well oxygenated waters. The site of Kolima, that yielded the smallest amount of Arius bones, also has the smallest proportion of open water fish (about 27%). At Hassi el Abiod and Kobadi, the proportion of open water species is about 54% and 80% respectively. Possibly, Arius gigas had ecological requirements similar to those of other large predatory fish that prefer the open waters and that do not, or only occasionally during the flood season, enter marginal waters. We suggest that A. gigas would not tolerate very shallow, muddy and stagnant waters.

9.5 EXPLOITATION BY HUMANS In the Holocene archaeological sites in Mali, fishing was one of the main subsistence activities of the populations living near rivers and lakes. These activities date back to the Early Holocene, in the Saharan sites, where fish remains are abundant and occur together with other vertebrates such as crocodiles, turtles, hippos, and antelopes (Van Neer and Gayet, 1988; MacDonald and Van Neer, 1994; Jousse, 2004). Arius gigas was one of the most frequently captured species and, because of its large size, may have contributed significantly to the human diet. Some fishing gear was found on those sites, mainly harpoons and hooks made of bone (Raimbault, 1994), that could serve for catching such large catfish, as well as the abundant Nile perch. Harpoons and hooks were also the fishing gear used for Arius fishing by the Sorko or Bozo, traditional and specialised fishermen still exploiting this fish in the 1950s in the Niger Inland Delta of Central Mali (Ligers, 1966). They traditionally used large and bifid harpoons, now made of metal, attached to a wooden stick that was connected to a floater, and attached to the dugout by two or three ropes. They attracted fish using a strange instrument called a “hojo”, made of wood, plants, bird feathers, and the following fish bones: a skull roof of Heterobranchus and a dorsal spine of Arius gigas. By shaking it in the water, the spine rubs against the cranial bone, and the resulting lapping imitates the sound of a fish that attracts the other ones. When fish approach the boat, fishermen harpoon them. Fish reportedly captured this way were Hydrocynus brevis, Hepsetus odoe, Gymnarchus niloticus, Lates niloticus, Heterobranchus, Arius gigas, Distichodus brevipinnis, Clarotes and Bagrus (Ligers, 1966, pp. 90–93). These fishermen had two kinds of hooks: fish gorges made of wood, measuring about 80 mm in length and 4 mm in diameter, and some curved hooks with various shapes, nowadays made of metal. Hooks are baited with fish flesh or intestines of fish or chicken, and are attached to a line that is laid out across the river and left passively for the night. In the morning, fish such as A. gigas, H. odoe, G. niloticus, L. niloticus and Heterobranchus are found caught. Line fishing from the river bank yielded all kinds of fish, but especially H. brevis, G. niloticus, Clarotes, Clarias anguillaris, Heterobranchus longifilis, A. gigas, L. niloticus, Malapterurus electricus (Ligers, 1966, pp. 122–123). Together with Gymnarchus niloticus and Heterotis niloticus, Arius gigas meat is reported to be preserved in dried form. Large quantities of meat are cut in strips and

Palaeoecology of the Giant Catfish (Arius gigas, Ariidae)

175

sun-dried over a period of three days. Thereafter, it is hung in the house and is then progressively consumed (Ligers, 1966, p. 197). At Kobadi, the Giant Catfish remains are sufficiently abundant to allow a study of the consumption patterns by humans. A detailed analysis of the preserved bones on the sites showed that there were no traces of disarticulation or cutting on the bone surface (Jousse et al., 2008). The fire traces (Table 2) left on the bones showed no clear pattern: all skeletal elements can be affected, or not, by combustion. It is concluded that most of the fire marks are due to post-depositional modification, when bones were thrown in—or unintentionally affected by—fire places. It is believed that the people at Kobadi processed the fish using very simple techniques, mainly cooking or grilling, and that they threw the refuse in the midden deposit or hearths (Jousse et al., 2008). The Arius bones at Kobadi are not exclusively food refuse. The site also yielded two fin spines that were polished and used as tools, whose function is not clear since the objects are broken (Jousse et al., 2008). The spines are solid and rectilinear and may have been used as a smoother, for example.

9.6 ARIUS GIGAS DECLINE Since the Holocene, both climatic changes and human activities have disturbed terrestrial and hydrological ecosystems (Lévêque, 1999b). Arius gigas is a good, but unfortunate, example of a species that suffered from both types of impact. According to Ligers (1966, p. 25), the Bozo fishermen used to concentrate on taking large carnivorous fish when they were considered to become too abundant and regarded as a threat to the survival of Brycinus leuciscus, which they much appreciate and intensively fish. Among the targeted piscivorous fish figure A. gigas and Malapterurus electricus. Ligers also wrote in 1966 concerning the frequency of the fish in the Niger Inland Delta: « . . . Arius gigas, qui était beaucoup plus fréquent il y a quelques générations et dont on capturait au moins un spécimen chaque matin.» Compared to small fish, larger species are more vulnerable to fishing pressure because they need more time to reach maturity. For instance, at 20 cm, a fish that will reach a maximum size of 30 cm will already be sexually mature and have reproduced. A fish that can reach a meter and more will not. So when fishing intensively, people would also take juvenile individuals of larger species, and thus contribute to the lowering of their reproduction potential and consequently negatively affect the populations (Lévêque, 1999b). The size at maturation of A. gigas is unknown, but may have been large and can have varied geographically. It has been established, for instance, that Clarias gariepinus can reach maturity when 1 to 3 years old (25–45 cm) in lakes and rivers of the Transvaal and Zimbabwe, but when 70–80 cm and more in Le Roux Dam and Verwoed Dam in South Africa (Quick and Bruton, 1984). Its slow rate of reproduction may also have hampered a rapid adaptation of the population to ecological stress. Arius gigas may have been over-fished because of its large size, its appreciated tasty meat and because it was considered a harmful predator of other fish. Since it was reported to measure about two metres at the beginning of the 20th century (Daget et al., 1988), the major decline of the species may have occurred during that century mainly as a result of increased fishing pressure. Fishing activities were encouraged for commercial purposes and developed in the Niger Inland Delta with the introduction of modern fishing gear such as synthetic nets with smaller mesh and more robust fishing vessels (Laë et al., 1994).

176

Jousse and Van Neer

ACKNOWLEDGEMENTS We address our thanks to P. Campbell from the fish collection of the British Museum of London, to J. Peters from the Institut für Paläoanatomie, Domestikationsforschung und Geschichte der Tiermedizin in Munich, and to J. Desse and N. Desse-Berset from CEPAM, Valbonne, for access to Ariidae skeletons in their collections. The specimen from Niamey was analysed in the office of M. Kelessi, director of the Musée National du Niger in Niamey, who kindly proposed to examine the Arius gigas. S. Hamilton-Dyer (Southampton) is acknowledged for the correction of the English text. This research is part of a post-doctoral project financed by a fellowship of the Fondation Fyssen, Paris, attributed to H.J. and carried out at the Institut für Paläoanatomie in Munich. The study of the Belgian collections received support from the SYNTHESYS Project (BE-TAF-1400) financed by European Community Research Infrastructure Action under the FP6 “Structuring the European Research Area” Programme. REFERENCES Carter, P.L. and Flight, C., 1972, A report on the fauna from the sites of Ntereso and Kintampo rock shelter six in Ghana: with evidence for the practice of animal husbandry during the second millenium B.C. Man, 7, pp. 277–282. Daget, J., 1954, Les poissons du Niger Supérieur. Mémoires de l’Institut Français d’Afrique Noire, 36, pp. 1–391. Daget, J., 1988, Systématique. In: Biologie et écologie des poissons d’eau douce africains, edited by Lévêque, C., Bruton, M.N. and Ssentongo, G.W., (Paris: ORSTOM, Travaux et Documents, 4), pp. 15–34. Daget, J., 2003, Ariidae. In: Faune des poissons d’eaux douces et saumâtres de l’Afrique de l’Ouest, edited by Lévêque, C., Paugy, D. and Teugels, G.G., (Paris: ORSTOM, ed. 2, vol. 2), pp. 269–276. Daget, J., Gaigher, I.C. and Ssentongo, G.W., 1988, Conservation. In: Biologie et écologie des poissons d’eau douce africains, edited by Lévêque, C., Bruton, M.N. and Ssentongo, G.W., (Paris: ORSTOM, Travaux et Documents, 4), pp. 481–491. Daget, J. and Stauch, A., 1963, Poissons de la partie camerounaise du Bassin de la Bénoué. Mémoires de l’Institut fondamental d´Afrique Noire, 68, pp. 85–107. Desse, J. and Desse-Berset, N., 1996, Archaeozoology of groupers (Epinephelinae). Identification, osteometry and keys to interpretation. Archaeofauna, 5, pp. 121–127. Gautier, A. and Van Neer, W., 2005, The continuous exploitation of wild animal resources in the archaeozoological record of Ghana. Journal of African Archaeology, 3, pp. 195–212. Gayet, M. and Van Neer, W., 1990, Caractères diagnostiques des épines de quelques silures africains. Revue de zoologie africaine, 104, pp. 241–252. Gregory, W.K., 1959, Fish skulls. A study of the evolution of natural mechanisms, (Florida: E. Lundberg), pp. 76–481. Jonsson, L., 1994, The estimation of fish size from bones and otoliths. Some methodological considerations and a simplified method for use with limited comparative material. Offa, 51, pp. 379–383. Jousse, H., 2004, Impact des variations environnementales sur la structure des communautés mammaliennes et l’anthropisation des milieux: exemple des faunes holocènes du Sahara occidental. Documents des laboratoires de Géologie de Lyon, 160, pp. 1–273. Jousse, H., Obermaier, H., Raimbault, M. and Peters, J., 2008, Late Holocene economic specialisation through aquatic resource exploitation at Kobadi in the Méma, Mali. International Journal of Osteoarchaeology. doi:10.1002/oa.956.

Palaeoecology of the Giant Catfish (Arius gigas, Ariidae)

177

Laë, R., Maïga, M., Raffray, J. and Troubat, J.-J., 1994, Evolution de la pêche. In La pêche dans le Delta Central du Niger, edited by Quensière, J., (Paris: ORSTOM-KharthalaIER), pp. 143–163. Lévêque, C., 1999a, Biogéographie et mise en place des faunes ichthyologiques actuelles. In Les poissons des eaux continentales africaines, edited by Lévêque, C. and Paugy, D., (Paris: IRD), pp. 69–81. Lévêque, C., 1999b, Impact des activités humaines. In Les poissons des eaux continentales africaines, edited by Lévêque, C. and Paugy, D., (Paris: IRD), pp. 365–383. Ligers, Z., 1966, Les Sorko (Bozo), Maîtres du Niger: Etude ethnographique. II- La pêche, (Paris: Librairie des Cinq Continents), pp. 1–203. MacDonald, K.C., 1994, Socio-economic diversity and the origins of cultural complexity along the Middle Niger (2000 BC to AD 300). Unpublished PhD Thesis (Cambridge: University of Cambridge). MacDonald, K.C. and Van Neer, W., 1994, Specialised fishing peoples in the Later Holocene of the Méma (Mali). In Proceedings of the 7th meeting of the ICAZ, Fish remains working group, edited by Van Neer, W. (Tervuren: Annales du Musée royal de l’Afrique Centrale, Sciences zoologiques), pp. 243–351. Petit-Maire, N. and Gayet, M., 1984, Hydrographie du Niger (Mali) à l’Holocène ancien. Comptes Rendus de l’Académie des Sciences, 298, II, pp. 21–23. Raimbault, M., 1994, Sahara malien: environnement, populations et industries préhistoriques. Unpublished thesis (Aix-Marseille: Université de Provence). Quick, A.J.R. and Bruton, M.N., 1984, Age and growth of Clarias gariepinus (Pisces: Clariidae) in the P.K. le Roux Dam, South Africa. South African Journal of Zoology, 19, pp. 37–45. Van Neer, W., 1989, Contribution à l’ostéométrie de la perche du Nil Lates niloticus (Linnaeus, 1758). Fiches d’ostéologie animale pour l’archéologie, ADPCA, 5, pp. 1–17. Van Neer, W. and Gayet, M., 1988, Etude des poissons en provenance des sites holocènes du Bassin de Taoudenni-Araouane (Mali). Bulletin du Muséum national d’Histoire naturelle, 10, section C, 4, pp. 343–383. Van Neer, W. and Lesur, J., 2004, The ancient fish fauna from Asa Koma (Djibouti) and modern osteometric data on three Tilapiini and two Clarias catfish species. Documenta Archaeobiologiae, 2, pp. 141–160.

CHAPTER 10

Neolithic Domestication and Pastoralism in Central Sahara: The cattle Necropolis of Mankhor (Tadrart Algérienne) Michel Tauveron Independent Archaeologist, Lissac sur Couze, France Karl Heinz Striedter Frobenius-Institut, Frankfurt am Main, Germany Nadjib Ferhat C.N.R.P.A.H., Algiers, Algeria ABSTRACT: Neolithic Pastoralism in Central Sahara is much in evidence by thousands of rock paintings and rock engravings. However, remains of cattle are extremely rare. So the discovery of an extensive cemetery of cattle is of highest interest. Excavations brought to light well-conserved bones which enabled palaeozoological examination and radiocarbon dating. The burial of the cattle remains was obviously governed by a set of rules suggesting its ritual character. Evidently, cattle was not only an economic factor at this time, but had social and/or socio-political implications comprising social rites. Some forms of such rites have been observed still in recent pastoral societies.

10.1 INTRODUCTION During a long period of the Late Pleistocene, Central Sahara was not populated, or at least very sparsely. From the end of the Pleistocene period on, when a change in climate provided for more favourable conditions of human subsistence, a resettlement took place. The Holocene saw an important evolution of human culture with the development of new technologies such as new lithic industries and the invention of pottery. During this period, known as the Neolithic period, a new form of economy appeared: Pastoralism, which had its climax in the Middle Holocene. It seems almost certain that domestication of the Bos species took place in Central Sahara. However, as the conditions of bone conservation in the Sahara are generally unfavourable, direct dating of animal remains was not possible. The discovery of a real cattle cemetery with well-conserved bones opened new perspectives. 10.2 THE SITE OF MANKHOR In the south-eastern section of the Algerian Sahara, between the Tassili n’Ajjer in the North and the Djado plateau (Republic of Niger) in the South, the Tadrart is an important sandstone massif extending over approximately 300 km. Its occidental slope is abrupt and deeply incised by gorges. Because of the presence of a granite uplift in the West, it is

180

Michel Tauveron et al.

bordered by a series of basins. This is where discharge some wadis whose continuation towards the Admer plain is blocked. The basin of Mankhor is one of the most important in the area. Being situated in a region crossed by series of crevices, it has profited through millenaries from exogenous water alimentation which is imbibed by the sands in this zone. In prehistoric times, during moments of less aridity, these rising waters permitted the development of important vegetation. In the northern part of the basin, some swampy formations left traces in the form of darkly coloured banks. In the spring of 1995 (Mission O.P.N.T., Frobenius-Institut, directed by K.H. Striedter and M. Tauveron), a particular prehistoric site was discovered. Over about 80 ha are scattered, without apparent organisation and with an approximate medium density of 1/100 m2, with numerous stone accumulations that look like ancient fireplaces (Figure 1). Artefacts appear rarely, without significant density, except for some potsherd accumulations in a limited zone. The artefacts seem to be the same as those on the pastoral Neolithic sites in the area. Numerous animal bones appearing at the surface, sometimes in heaps of about 60 cm in diameter (most quickly identified as cattle bones for most of them) mark the uniqueness of the site. The first investigations revealed that only some of the stone accumulations show traces of burning and could be considered as fireplaces; the potsherd accumulations have shown an intimate association with burned remains, especially bones, and two of them were removed for laboratory analysis. A frontal part of cattle was also removed for identification of the species. This finding showed a partial burying with the clear delimitation of a pit. This fact incited us to look for a more complete view on the presence of the faunal remains: we decided to carry out an excavation centred on a cattle mandible, which seemed to be partially included in the sediment. This excavation yielded a rather complete skeleton (a part of the skull is lacking) of some form of cattle, which was in a prime conserved state. It was associated with broken pottery and disposed in superimposed pieces, each in anatomic

Figure 1. Part of the site of Mankhor, with concentration of stone accumulations (scale bar of 2 m in the foreground).

Neolithic Domestication and Pastoralism in Central Sahara

181

Figure 2. Last layer of the inhumation excavated in 1995.

connection, in a 1,5 m deep pit measuring 0,8 × 0,8 m. From top to bottom: the skull with two cervical vertebras and the pottery, anterior members, thorax and vertebral column down to the lumbar, the rest of vertebral column with the basin, and lastly, the inferior members (Figure 2). The members were sawed above the metapodialia and none of the bones showed any trace of burning. Such a disposition could only be the result of ritual inhumation, and it is probable that the meat was not completely consumed, as indicated by the ribs in connection. As few cattle inhumations were known in Central Sahara (Clark et al., 1973; Paris, 1997; Roset, 1987), this preliminary result stimulated more extensive excavations, which began in the spring of 1996 (Mission O.P.N.T.–C.N.R.P.A.H., Frobenius-Institut, directed by N. Ferhat). 10.3 THE EXCAVATIONS AT MANKHOR This excavation campaign has provided a better comprehension of the site structure, which appears to be subdivided into four distinct zones. One, near the cliff, with an important presence of actual sand and some artefacts in the surface without any concentration, seems to have been only an area of circulation. The second, in the eastern part of the site, shows concentrations of three to five stone accumulations and a few stone artefacts as well as rare tools and pottery; it is separated from the third part of the site by a rather empty zone. The central part of the site forms the third zone, characterised by the most important concentration of stone accumulations and bones. The density of artefacts is more important here, but remains low; it is never more than 5/m2 including debitage, with the exception of potsherd concentrations. The fourth zone includes the western part of the site. It shows

182

Michel Tauveron et al.

smaller stone accumulations, less concentrated as in the third zone, with less bone remains and ceramics at the surface, but with somewhat more stone tools (especially scrapers and arrow points). Only the third and fourth zones have been partially excavated. The first excavated area is located in the middle of the third zone, which presents the most important concentration of stone accumulations. There are three different types: (1) Important circular fireplaces, roughly one meter in diameter, delimited by decimetric burned stones, with a stratum of at least 25 cm of ashes and charcoals mixed with calcined splinters of bones; surrounding them are various potsherds, stone artefacts (mostly flakes), and other burned splinters of bones; (2) small circular fireplaces, roughly 60 cm in diameter, identified by their delimitation, a partial cover by burned stones and a darker colour of sediment due to the presence of ashes and micro-charcoals; (3) subcircular stone accumulations without any trace of fire, covering pits with bone deposits. In two cases, we found three different bone deposits in the pits, with a light horizontal dislocation and a centimetric sterile layer between them, which suggests successive deposits and a system to recover the precise place of the pit. In the other pits, there was only one bone deposit. All have shown a complex structure, each time on the same basis. Compact packets of bones (Figure 3), usually 20 to 30 cm high and 50 to 60 cm in diameter, are delimited for the perimeter and the bottom by respectively oblique and flat-crossed rib fragments, sometimes with other flat bones (often omoplates), with small stones often securing the whole arrangement. The remains of long bones are in the centre, almost exclusively heads, the vertebras, often several of them in connection, the extremities of members, with the last two phalanxes always quite flat. The mandibles, always in two parts, may be placed in an oblique manner, the rear part up, or in flat position and crossed (Figure 4). The skull, when complete, is carefully blocked on the border of the heap, sometimes vertically with the rear part facing up; if it is fragmented, its remains occupy the centre in the bottom.

Figure 3. Excavated bone heap «C», it’s probable that the compactness is related to the use of some type of bag, of organic material.

Neolithic Domestication and Pastoralism in Central Sahara

Figure 4. Excavated bone heap «D», well-conserved mandibles in crossed position and maxilla.

Figure 5. Combination of two stone and one bone accumulation which forms an equilateral triangle.

183

184

Michel Tauveron et al.

The second excavated area concerns an alignment of bone accumulations, marked by less than 10 small stones around each, in a WNW–SSE direction, still in the third zone. The bones showed the same disposition as in the pits, but were not so well conserved. The third area excavated is in the fourth zone, where, looking for a reference system, a similar relative disposition between two stone accumulations and one bone accumulation has been ascertained in several cases. They were each situated at the top of an equilateral triangle, also with a similar orientation grouped around 30–35° N for the side, limited by the two stone accumulations which were small fireplaces (Figure 5). On the surface, bones appear here only in the form of some burned splinters, but the excavation revealed the same disposition as in the previous cases. The most important difference is that of burning, probably after the deposit, as the surrounding sediment is quite indurated (by heating), with the most severely burned part in the upper centre. 10.4 OBSERVATIONS AND RESULTS The complete animal excavated in 1995 was identified by Pr. C. Guérin (personal communication, 1996) as Bos taurus ibericus and was dated as originating from 4.870 ± 120 BP (Pa, 1457, thanks to J.F. Saliège), but we did not unearth additional inhumation of this type. The bone’s conservation was better than in all other deposits and only two graves of the fourth zone give a little bit younger date (5 years later!), marking the end of the site frequentation. In the other cases, two major rituals appear, one with burning, the other without. Two secondary series exist in this last case: One with the association of potsherds, the second without. In contrast with surface ceramic, most of these potsherds, like the pottery discovered in 1995 (Figure 6) are of very poor quality

Figure 6. After reconstruction, nearly complete pottery found in the inhumation excavated in 1995.

Neolithic Domestication and Pastoralism in Central Sahara

185

(paste, mounting, decoration and firing), but with the same shapes and decorations, which suggest a special fabrication for the ritual, or they were broken before their deposit, as attested by polished fractures. A second set of palaeontological identification is still in progress: The preliminary results indicate the presence of large-sized cattle, probably Bos primigenius, as well as a second smaller one, which has not yet been clearly identified, but could be a female type of the same species or Bos taurus (ibericus). Most of the removed bones are those of young cattle, as indicated by tooth attrition and partial ossification of conjugation cartilages. For eight of the studied animals, all the bones are found in the pit (even if they are not complete), with the systematic exception of the frontal part of the head with horns. Scratches appearing on some ribs are probably scraping traces. The whole set of radiocarbon dates is included in a fork from 5.525 to 4.865 B.P., 4.490 to 3.510 cal. B.C. (1 σ⋅ cal. 10 dates on H.C.A.: Pa 1455, Pa 1705, Pa 1707, Pa 1709 to 1715). 10.5 INTERPRETATION The careful and recurring disposition of the bones and a nearly similar age of the buried cattle provide an interpretation of the inhumations as that of cattle. The poor density of the artefacts and the presence of numerous fireplaces always in proximity with these inhumations indicate a place exclusively dedicated to a ritual activity. This concurs with some of the interpretations proposed by S. Di Lernia (2006) for Libyan cattle inhumations from the Messak plateau (but not all of them): There are no true megalithic structures here, even if there are some small stones arrangements upon or beside the graves, and no human inhumation was found on the site. Some differences can also be noted with the animal burials in the Republic of Niger: If a common background of a cattle cult appears probable, it is with a great variability from one place to another, which is confirmed by rock art manifestations. The presence of at least three different series of inhumation in the same global tradition could be interpreted as a juxtaposition of different facies or as chronological evolution. Nevertheless, the hypothesis of a “collective ritual that emerged within Saharan pastoral societies to face an uncertain climate” (Di Lernia S., 2006) appears plausible (Ferhat et al., 1998). Rituals where cattle play the prominent part are known until today in the African Sahel. P. Paris (1997) reports a rite among the Wodaabe (Republic of Niger), the last really nomadic group of Western Africa. A “sacrificed” bull marks the essential part of the rite which confirms or establishes alliances between different groups. It seems that this rite has exclusively social or socio-political implications. 10.6 CONCLUSION In conclusion, for the first time in this area, not far from Tin Hanakaten rock shelter, there are complete skeletons of cattle; the encountered problems to identify great Bos taurus or small Bos primigenius demonstrate the difficulty in underlining domestication criteria for Bos species in the lack of any horn element. Rock paintings of the area show a great domestic Bos in a pastoral context; this type of Bos was previously called Bos africanus and has specific horns in lyra shape. As there is no prior bone identification of this animal, could it be the living model of some of the ambiguous Bos from Mankhor, and/or a possible answer to the palethnologic questions due to the presence of Bos primigenius? Small size Bos taurus (ibericus), clearly domestic, is well known in rock art throughout the bovidian period, but increasingly more prevelant in the middle and final sequences. All of these

186

Michel Tauveron et al.

elements give Mankhor an unusual dimension, which is of utmost importance for the understanding of Central Saharan pastoral Neolithic and Cattle Keepers’ environment adaptation process, including their spiritual and cultural development. REFERENCES Clark, J.D., Williams, M.A.J. and Smith, A.B., 1973, The geomorphology and archaeology of Adrar Bous, Central Sahara: a preliminary report. Quaternaria, 17, pp. 245–297. Di Lernia, S., 2006, Building monuments, creating identity: Cattle cult as a social response to rapid environmental changes in the Holocene Sahara. Quaternary International, 151, pp. 50–62. Ferhat, N., Chaid-Saoudi, Y. and Aumassip, G., 1998 (2000), Le site de Mankhor et ses relations au culte du taureau dans le Sahara central. Animaux et Rites: Perception et Reproduction de la vie. C.N.R.P.A.H., Algiers, pp. 11–18. Ferhat, N., Striedter, K.H. and Tauveron, M., 1996, Un cimetière de bæufs dans le Sahara central: la nécropole de Mankhor. La préhistoire de l’Afrique de l’Ouest. Sépia, SaintMaur, pp. 103–107. Guérin, C. (Université C. Bernard-Lyon 1, France), private letter from 1996/01/20. Paris, F., 1997, Les inhumations de Bos au Sahara méridional au Néolithique. Archéozoologia, IX, pp. 113–122. Paris, F., 2000, African livestock remains from Saharan mortuary contexts. The Origins of African Livestock. Blench R., MacDonald K. (Eds.), UCL, London, pp. 111–126. Paris, P., 1997, Ga’i ngaanyka ou les taureaux de l’Alliance. Description ethnographique d’un rituel interlignager chez les Peuls Vod’aab’e du Niger. Journal des Africanistes, 67, 2, pp. 71–100. Roset, J.P., 1987, Néolithisation, néolithique et post-néolithique au Niger nord-oriental. Bulletin de l’Association Française pour l’Etude du Quaternaire, 4, pp. 203–214.

CHAPTER 11

The microstratigraphy and micromorphology of a Holocene palaeolake in Southern Tunisia Abdelhakim Abichou CGMED, Faculty of Human and Social Sciences, Tunis, Algeria ABSTRACT: The cartographic surveys and the micromorphological investigations of the sediments of the sebkha Erg el Makhzen demonstrate that its basin, which had functioned during the Early and Middle Holocene under the conditions of inundation and drought, ended in the creation of three lacustrine facies. The first lacustrine optimum, passing from the Pleistocene to the Early Holocene period, bears witness to the accumulation of goethite and haematite as well as to the creation of dense and pisolithic ferricrusts. Around 8.060 ± 150 BP, the second lacustrine optimum is dominated by the increase in organic and carbonaceous-sulphate fluvial-lacustrine deposits. Around 6.500 ± 70 BP, the carbonaceous third lacustrine optimum shows the compaction of oncolothic crusts and of the flat stromatolithic crusts. The increasing climatic aridity around 5.880 ± 127 BP heralds the drying out of the palaeolake of Erg el Makhzen and the accumulation of dune deposits. The return of fluvial sedimentation during the Later Holocene period engenders the formation of an enormous spreading cone at the mouth of the wadi.

11.1 INTRODUCTION Erg el Makhzen Sebkha spreads between 10°55' longitude and between 32°54' and 32°56' latitude N (Figure 1). The Sebkha’s basin occupies the navel of post-Triassic tectonic subsidence and its bed consists exclusively of the Triassic red sandstone of “Kirachaou”.

Figure 1. Location of the Sebkha Erg el Makhzen.

188 Abdelhakim Abichou

The progressive deterioration of the bed illustrates the passage from a bare sandstone system of monoliths to a system of ferri-crusts developed discordantly with the geological substratum (Figure 2). A priori, the forced deterioration of sandstone, due to the immersion of the basin, leads to the marking of the epigyny of the sandstone by the petrographic trace of iron.

Sebkha Erg el Makhzen Contact zone of the iron crust/sandy substratum

Polished section

Thin section: Iron crust sandstone

Composite pisolith. Each composite pisolith encloses serval amorphous iron nucleoides, each with a fibro-radar goethite cortex. The whole structure is enclosed in a single manganese ring.

Dense ferricrust

The discordant contact between the altered sandstone and the crust encloses an amorphous quartz sandstone and a layer of ferriginous clays

Microstratigraphy of the crust. The upper part of the crust shows laminary haematite structures.

Iron pisolith

Figure 2. Micromorphology of dense and vascular ferricrusts.

The microstratigraphy and micromorphology of a Holocene palaeolake in Southern Tunisia

189

11.2 THE DETERIORATION OF THE SANDSTONE BED AND THE FORMATION OF FERRI-CRUSTS The micromorphological traits of the two crusts represent a typical example of the relative accumulation of goethite and haematite following the subtraction of silica, alumina and the bases. It concerns a global transformation process of the sandy substratum’s architecture, where iron appears progressively in the spaces left by kaolinite and the skeleton of quartz. In thin section, the dense ferri-crust is black, dense and unbroken in appearance (Figure 2). The ETM microprobe analyses reveal goethite to be the primary mineral constituent. The structure of the dense ferri-crust ensues from the combination of homogenous micropisoliths of 15 μm in diameter. Each pisolith is composed of a minuscule nucleus of amorphous iron, a single fibro-radial crystal structure of goethite and a single manganese-based peripheral ring. However, in thin section, the vacuolaminary crust—unconformous to the Kirchaou sandstone—, is defined by the combination of a string of pisoliths greater than 200 μm in diameter, where the size of the interpisolithic spaces can reach up to 2 mm. Each pisolith is endowed with a nucleus made up of more than 90% amorphous iron, allied with 1 to 2% alumina and less than 5% manganese. The cortex of the fibro-radiar goethite crystal pisoliths is organised into several concentric rings. The smallest pisoliths have at least two concentric rings between 30 and 50 μm in diameter. The largest pisoliths may contain more than six concentric rings in excess of 100 μm in diameter. The peripheral rings made by the neoformation of manganese hydroxide indicate that the formation of the pisoliths takes place in a confined environment of a hydromorphic nature. All things considered, the results of the microanalyses and of the cartography allow us to conclude that the formation of ferricrusts takes place in a lacustrine milieu distinguished by a fairly strong water current capable of transporting pisoliths over long distances.

11.3 THE EXPANSION OF FLUVIO-LACUSTRINE DEPOSITS The second lacustrine optimum illustrates the accumulation in the basin of Erg el Makhzen of fluvio-lacustrine deposits carbon-dated to 8.070 ± 150 BP (Figure 3). The lithostratigraphy of the sites evokes a process of fluvio-lacustrine sedimentation in clear and calm waters. The concentration of empty, cylindrical root moulds of Phragmites would indicate water of less than 2% salinity. These roots served the role of filters restricting the dispersion of fluvial detritus. The diffractometry of the carbonates and sulphates also enables the appearance of gypsum and of calcite, a cortege of three hydrated slats composed of bassanite (CaSO4 × ½ H2O), of bloenite (NaSO4 × MgSO4 × H2O) and of epsomite (MgSO4 × 7H2O). The mineralogy of the clays shows a cortege situated around the kaolinite, illite, palygorskite and serpentine. The micromorphological analysis of a thin section reveals that the superposition of three layers is also responsible for the increase of deposits. At the base a layer of detritus of quartz silt and of clays was around the kaolinite, illite, palygorskite and serpentine. The micromorphological analysis of a thin section reveals that the superposition of three layers is also responsible for the increase of deposits. At the base a layer of detritus of quartz silt and of clays was around the kaolinite, illite, palygorskite and serpentine. The micromorphological analysis of a thin section reveals that the superposition of three layers is also responsible for the increase of deposits. At the base a layer of detritus of quartz silt and of clays was cellulose. The third gypsum layer closes the circle of fluvial

190 Abdelhakim Abichou Wadi Cherchera

Fluvio-lacustrine deposits

Organic and charcoal layers Detritic and evaporitic layer

Evaporitic layer

Layer’s accretion

Detritic layer

Thin section: Fluvio-lacustrine organisation

Figure 3. Wadi Cherchera. Fluvio-lacustrine deposits.

evaporation. The crystallisation of the gypsum took place due to its concentration in Ca2+ and in SO42– and alongside the shrinking of the water table. In thin section, the skeleton of the crystallomorphic gypsum layer encloses an s-matrix of transparent gypsum crystals, monocrystaline along the axis of completed growth. The fluvio-lacustrine sedimentation had a continuous and regular flux and within that continuity each triplet of layers constituted a separate sedimentary cycle.

The microstratigraphy and micromorphology of a Holocene palaeolake in Southern Tunisia

191

11.4 THE CARBONACEOUS DEPOSITS OF THE MIDDLE HOLOCENE The third lacustrine optimum recorded in the basin of Erg el Makhzen reveals around 6.970 ± 70 (Ballais 1991) the mobilisation of significant amounts of carbonates supplied by karstic waters concentrated in ions of Ca2+. The results of the microprobe microanalyses

Figure 4. Facies and microfacies of stromatolithic carbonates.

192 Abdelhakim Abichou

interfaced at ETM indicate that crusting close to water sources is oncolithic, whilst those of the flood basin correspond to the flat stromatolithic crusts. The lithification of the oncolithic crusts results from the disposition of a niche of homometric oncoliths between 200 and 250 μm in diameter (Figure 4). The micromorphology shows that each oncolithic corresponds to a calcitic concretion of symmetrical growth, of which the radiar-structured cortex developed around a microcristaline nucleus of calcite. The condensed contents of the grain-supported nucleus display a drowned micritic central pole within a thick crown of microsparitic calcite, of very slight magnesium content. The cortex of the oncolith shows a pattern of synchronic growth and overlapping spikes of calcic crystals between 100 and 120 μm. The duration of the growth of each oncolith is revealed by the nature of one or more cortex. At the end of the process, the oncolith is enclosed within a single ring whose irregular contour ranges between 20 to 30 μm in thickness. Polarised light enables us to see the disintegration of the rings, the result of pressure centred on the seams of the oncoliths. However, the two or three oncoliths will in turn gather within the same single ring. The microprobe and ETM analysis brings out the rings of an amorphous, flaky gel entirely consistent of manganese hydroxide, which argues in favour of the perseverance of a hydromorphous environment. Throughout the Middle Holocene the return of water to Erg el Makhzen basin is shown by the formation of a flat stromatolithic crusting in discord with the sandstone substratum. The systematic analysis of the thin section microprofiles by microprobe and the ETM indicates a crust formation composed of microsequences, where each of them comprises

Figure 5. The Mid Holocene palaeodune.

The microstratigraphy and micromorphology of a Holocene palaeolake in Southern Tunisia

193

four layers of different mineralogy. The white base layer is composed of a cryptocrystalline calcite mud. The second sparitic layer is composed of acicular calcite in excess of 200 μm implanted perpendicular to the previous layer. Under polarised light, the third layer shows up bright red, comprised as it is of a strong concentration of haematite. According to the appearances of the microstructures, it becomes apparent that the precipitation of iron took place under the strict conditions of a strong concentration of algae (Rivularia haematite). The dark black colour of the fourth layer corresponds to the precipitation of manganese hydroxide. 11.5 THE CLIMATIC ARDIFICATION AND THE GENERALISATION OF DUNE DEPOSITS THROUGHOUT THE MIDDLE HOLOCENE The aridification of the climate around 5.880 BP reveals the incision of the wadi terraces, the destruction of the fluvio-lacustrine barrages and the predominance of aeolian morphogenesis. The paleo-dune of the wadi Cherchera encrusted itself in the incisions made in the terraces of the wadi Cherchera and in the fluvio-lacustrine deposits (Figure 5). The dune deposits are organised on several stratified levels where beds of fine sand alternate with beds of gypsum rosettes recording the fluctuations in the level of the water table of the Sebkha. 11.6 THE HYDRO-AEOLIAN SEDIMENTATION OF THE LATE HOLOCENE The morphodynamic contrast between the two sides of the sebkha’s basin is reflected by the type of the infill of the eastern part of sebkha. A heterometric saprolite followed by a fine ferruginised arena characterizes the base of the infill. This is followed by the appearance of the pseudo-regs resulting from the deterioration of the Triassic clay sandstone. On top, the sequence encloses strictly fluviatile sediments of a laminated structure, a reorganisation of the fluvio-lacustrine deposits, the extension of the gravel beaches and the formation of pulverised surfaces of fech-fech. This part of the sebkha, having collapsed to a greater extent, sends the surplus sediments towards the wadis and coastal sebkhas. 11.7 CONCLUSION Due to its topographical position and its behaviour, the sebkha Erg el Makhzen operated as an axis of the anterior river basin. Its special position accounts for the importance of the multiplication of lacustrine facies recorded in this basin. On the chrono-lithological axis of figure 6 the ferri-crust, the fluvio-lacustrian deposits and the stromatoliths are attributed to the end of the Pleistocene to the Early Holocene, and to the Middle Holocene. This results from mapping of the formation of flood basins, and from the microanalyses. All formations of the sebkha Erg el Makhzen are taken as indicators of one biostatic phase. Tectonic activities caused a period of the emergence of springs favouring the lacustrian environment of the basin. From this stem the presence of the condition caused favourable conditions of bio-crystallisation of carbonates during the Early and Middle Holocene. The absence of detritic sedimentation in this section indicates a stability of the superficial formations of the upper catchment. The aridification of the climate around 5.500 BP is a period of rhexistasis. It is responsible for the disorganisation of the drainage system and for the current fluvial in-fill of the sebkha Erg el Makhzen.

Figure 6. Chrono-stratigraphy of Sebkha Erg el Makhzen.

194 Abdelhakim Abichou

The microstratigraphy and micromorphology of a Holocene palaeolake in Southern Tunisia

195

REFERENCES Abichou, A., 2002, Les changements de paysages du basin-versant de l’oued Tataouine Fessi (sud-est tunisien): Etude multiscalaire et micromorphologie des remplissages des sebkhas et etudes des états de surface. PHD, Bordeaux3, Université Michel de Montaigne, pp. 1–417. Ballais, J.L., 1991, Evolution holocène de la Tunisie saharienne et présaharienne. Méditerranée, 4, pp. 31–38. Bouaziz, S., 1986, La déformation dans la plate-forme du Sud-tunisien (Dahar et Jeffara). Approche multiscalaire et pluridisciplinaire. Thèse de 3ème cycle. Univ. Tunis, (Tunis), pp. 1–180. Casanova, J.N., 1986, Les stromatolites continentaux: paléoécologie, paléohydrologie, paléoclimatologie. Application au Rift Grégory. Th. Doc. Etat. Uni. Aix-Marseille II. I, pp. 1–256, II, pp. 1–59. Fromm, R.M., 1993, Geologische Kartierung und mineralogisch-geochemische Untersuchungen in Südtunesien. Diplomarbeit, Geologische Karte Ksar el Aïne, 1:10.000, Mineralogisches Institut, Universität Karlsruhe, pp. 1–112. Purser, B.H., 1980–1983, Sédimentation et diagenèse des carbonates nétritiques récents. Edt. Techn., Ecole. Nat. Sup. du pétrole et des moteurs. Schulz, E., Abichou, A., Hachicha, T., Pomel, S., Salzmann, U. and Zouari, K., 1995, Climate and man. Questions and answers from both sides of the Sahara. Second. Symposion on African Palynology, Tervuren, Belg. Publ. Occas. CIFEG, 31, pp. 35–47. Stengele, F., 1993, Mineralogische Untersuchungen an quartären Sebkha-Sedimenten Südtunisiens, Geologische Karte des Jbels Rehach Schichtuferandes im Bereich der Playa Areg el Makhzen, Diplomarbeit, Mineralogisches Institut, Teil II, Karlsruhe, pp. 1–69. Tardy, Y., 1993, Pétrologie des latérites et des sols tropicaux. Paris, Masson, pp. 1–459. Tucker, M.E. and Wright, P.V., 1990, Carbonate Sedimentology. London, Blackwell, pp. 1–420.

CHAPTER 12

Different dimensions of recent vegetation dynamics of North and West Africa Brian Beckers and Brigitta Schütt Department of Earth Sciences, Physical Geography, Freie Universität, Berlin, Germany ABSTRACT: The Sahara and its adjacent regions have been subjected to several ecological and climate variations since the Mid-Holocene. Variations in the 20th century gave rise to controversial debates about the causes and the character of these environmental changes. This study investigates the relationship and character of vegetation and rainfall dynamics using AVHRR NDVI data, GPCP rainfall estimates and weather station data for the period 1983 to 1999. The research area is a N–S transect reaching from the Mediterranean Sea to the Gulf of Guinea. The findings back the widespread assumption that rainfall is the main causative factor of vegetation dynamics in North and West Africa. They also confirm regional greening trends in the Sahel, but found negative vegetation trends in the Sudan zone. Secular vegetation trends in the Mediterranean region are on a local scale and are most likely human induced.

12.1 INTRODUCTION For the Sahara and its adjacent regions several ecological and climatic “regime shifts” have been shown to have occurred since the Early Holocene (Foley et al., 2003; see also Schütt and Krause, this volume) and have been assumed—but without correct dating—for the Quaternary (cf. Busche, 1973, 1998). Between 14.500 to 5.500 years ago the region was more humid than today. The Saharan-Sahelian border lay approximately 600 km further north than today, at approx. 23°N. Since the Mid-Holocene, lakes and mesic vegetation communities have disappeared successively in the Sahara, hence indicating a transition to much more arid conditions (Lezine et al., 1990; Pachur and Hoelzmann, 1991; de Menocal et al., 2000). Even with this general aridisation trend regional climate was repeatedly subject to alternations between arid and more humid periods, which are assumed to have been triggered by the Milanchovitch variations and amplified by positive feedback mechanisms between vegetation and atmosphere (Street-Perrott et al., 1990; Claussen, 2003). Since 2000 years ago the climate of Africa has been generally similar to that of today, yet—albeit to a lesser extent and probably triggered by Sea Surface Temperature (SST) variations—wetter periods have been alternating with drier ones (Hastenrath, 1990; Ward et al., 1991; Nicholson, 2000). Although these alternations between drier and more humid phases are considered to be normal in the arid and semi-arid circum-Saharan belt, the prolonged recurring droughts of the Sahel since 1969 have been outstanding in their length and impact in the 20th century. The droughts caused widespread human suffering that was exacerbated by political unrest and mismanagement. These droughts led to a still ongoing debate about desertification (Mainguet, 1994; Thomas, 1997). On the basis of rather local and short term observations, Lamprey (1988) proposed an encroachment of the Sahara into the Sahel and assumed this to be an irreversible process. However, this proposal

198

Beckers and Schütt

of a large-scale environmental degradation has been proved wrong by many recent studies on long-term vegetation and rainfall dynamics in the Sahel. For example, Nicholson (2005) showed that rainfall in the Sahel recovered from the post-1969 droughts and has reached the long-term mean. Studies using remote sensing techniques have found greening trends in the Sahel on a regional scale since the 1980s. Moreover, these studies point out that large-scale environmental change in this region is largely driven by rainfall variability (e.g. Olson et al., 2005; Herrmann et al., 2005; Anyamba and Tucker, 2005). This controversy, together with the fact that environmental change in this region can have such severe impacts, highlights that knowledge about the nature and the causative factors of ecosystem dynamics, and the associated vegetation dynamics is necessary to assess environmental changes accurately. In this context the main objectives of this study are to (1) gain knowledge about the distribution and the character of regional vegetation dynamic patterns of North and West Africa, (2) further analyse vegetation trends in the study area, (3) detect possible shifts of vegetation dynamic patterns, and (4) characterise the relationship between precipitation and vegetation changes. A link is hypothesised between secular trends in vegetation greenness and shifts in regional vegetation dynamic patterns. The study uses remote sensing techniques and ground measurements of rainfall. Research was carried out along a N–S transect across the hyperarid Central Sahara, reaching from the Mediterranean to the Guinean Gulf. 12.2 STUDY SITE AND NATURAL SETTING The transect examined intersects the central part of Africa’s climate- and ecoregions north of the equator. It extends southwards from the Gulf of Sidra in the Mediterranean Sea at about 34°N to the Adamaoua Mountains in Cameroon at 6°N and is approximately 30° latitude and 12° longitude in size (see figure 1). The description of the present vegetation in this study is largely based on its physiognomy. The regional terms of the main vegetation units are adopted from White’s Vegetation Map of Africa (1983). Furthermore, the vegetation described here is the potential natural vegetation rather than the actual vegetation. This is due to the ever increasing influence and activities of humans that have modified the “natural” vegetation (Anhuf and Frankenberg, 1991; Foley et al., 2003). In Northern Africa the spatial distribution of rainfall and its seasonality are determined by the general circulation pattern. In the summer of the northern hemisphere the ITCZ (Intertropical Convergence Zone) and an associated low pressure cell over the Sahara lie between the moist SW monsoon and the dry NE trade wind or harmattan. In the winter months of the northern hemisphere the ITCZ is located further south, while a high pressure cell forms over the Sahara. During this time the northern part of the working area is affected by the rainbearing mid-latitude westerlies, whereas the dry NE trade wind influences the southern part. Hence, the working area separates into a winter and a summer rainfall regime, with the Sahara in between (Weischet and Endlicher, 2000; Nicholson, 2001). Mediterranean climate. A small coastal strip of wetter climate with a distinct seasonality separates the Mediterranean Sea from the arid Sahara. In the working area the Mean Annual Rainfall (MAR) reaches from 100 mm in the coastal lowlands to 250 mm in the coastal mountains (i.e. Jabal Natusah and Jabal Akhdar, Atlas in Tunisia), and rainfall occurs in the northern hemisphere winter months (Le Houerou, 2001). According to the phytogeographical classification of White (1983) this region belongs to the Mediterranean/ Sahara regional transition zone. The broader vegetation consists of grass and shrub steppes. Forest virtually only occurs in the coastal mountains (White, 1983; Le Houreou, 2001). Desert climate. South of the 100–150 mm isohyete the distribution of the vegetation becomes diffuse, passing into contracted vegetation further south and hence marking the

Different dimensions of recent vegetation dynamics of North and West Africa

199

Figure 1. Map of Africa north of the Equator with relief, international boundaries and major landmarks in the study area. The black polygon indicates the working area. Note that the extent of Lake Chad is exaggerated (outlines refer to ∼1960) (map source: GTOPO30).

transition to the Sahara (Breckle, 2002; Schultz, 2005). The Sahara itself is under the influence of the subtropical high-pressure belt with little influence of either the westerlies or the ITCZ (Weischet und Endlicher, 2000). Here, where rainfall is erratic, MAR ranges between 50–100 mm, with parts of the Sahara receiving no rainfall for years. Most of the rainfall is associated with mid-latitude westerlies in combination with tropical easterlies and is hence of extra-tropical origin (Nicholson, 1980 and 2000). The vegetation consists of perennial and ephemeral xerophytes. The former are restricted to areas where groundwater reaches the surface (i.e. oases) or to areas with runoff (i.e. wadis (Arabic), koris (Tamashek), koris (Tubu)). The short growing cycle of the latter is linked to rainfall (i.e. Acheb, Arabic for flowering desert). In large parts of the Sahara ephemerals are almost the only vegetation. Most of the vegetation is to be found in the Sahara mountains (i.e. Tibesti, Hoggar), which receive higher amounts of rainfall than their surroundings (Meadows, 1999; Le Houerou, 1997; Schultz, 2005). Besides these vascular plants, photosynthetic active microphytic communities can be found throughout arid and semiarid areas where vascular plants are absent (Karnieli et al., 2002b). Dry and Thorn Savanna Climate. South of the Sahara, the contracted vegetation distribution gives way to a dispersed cover. Here, at about 17°N, is the transition to the Sahel. According to Le Houerou (1980) the Sahel (Arabic for shoreline, coast) can, from an ecological and geobotanical point of view, be subdivided into three subzones: (1) Saharo-Sahelien with a MAR of 100–200 mm, (2) Typical Sahelien, MAR 200–400 mm, (3) Sahelien-Sudanian, MAR 400–600 mm. The Sahel’s rainfall regime is characterised by

200

Beckers and Schütt

a northward decreasing unpredictability and interannual to decadal variability of rainfall, thus figures of MAR for the Sahel are almost meaningless (Hulme, 2001). Therefore, the above-mentioned MAR values serve only for orientation purposes. As the rainfall increases southwards, so does the length of the rainy season. It ranges from 1,5 months in the northern to 3,5 months in the southern part of the area under consideration. The rainy season is from July to September. Most of the rainfall in the Sahel is associated with easterly waves and cloud clusters (Weischet and Endlicher, 2000; Nicholson, 2001). The vegetation in the Saharo-Sahelien part consists of perennial grasses with a patchy distribution. Further south, in the Typical Sahelian subzone thorny trees occur. The grass cover is perennial. The vegetation of the southward succeeding Sahelian-Sudanian subzone is characterised by a tree/shrub savanna (LeHoureou, 1980; White, 1983). With increasing rainfall these grasslands with scattered trees or bushes pass into the Sudanian regional centre of endemism (White 1983). Wet Savanna climate. The Sudanian region can be subdivided into a northern and a southern part. MAR varies between 600 mm in the northern part to 1.500 mm in the southern part. The climate is characterised by a distinct seasonality. The rainy season is between May and September (Weischet and Endlicher, 2000). Different types of savannas are the prevailing vegetation. These are from north to south: savanna grasslands, savanna parklands and savanna woodlands. Inconsistently with the above remarks on natural vegetation, however, savanna ecosystems are assumed to be a result of cultivation (e.g. grazing, fire, forestry) rather than a climatic climax (White, 1983; Adams, 1999; Meadows, 1999).

12.3 DATASETS AND METHODS 12.3.1 Datasets NDVI Since vegetation is not directly measurable with multi spectral sensors Vegetation Indices (VI) have been developed as a proxy for the photosynthetic capacity of vegetation (Tucker, 1979). The most widely used vegetation index to detect and monitor vegetation is the Normalised Difference Vegetation Index (NDVI) (Cracknell, 1997). The NDVI exploits the fact that, unlike most other natural surfaces, photosynthetic active vegetation absorbs a good portion of the Visible Light (VIS = 0,22–0,68 μm) and reaches maximum reflectance in the Near Infrared (NIR = 0,73–1,1 μm). Using this attribute, the NDVI is generated by the ratio of the reflectance value R in the indicated band (commonly red and NIR) of applicable satellite sensors and is defined as follows: NDVI = (R_NIR – R_red )/(R_NIR + R_red ). By this, photosynthetic active plant surfaces can be distinguished from other surfaces. High photosynthetic capacity and dense vegetation cover correspond to high NDVI values. The NDVI ranges from –1 to 1, values <0 represent snow and water) (Tucker, 1979; Holben, 1986; Myneni et al., 1995; Karnieli et al., 2002b). Whereas the NDVI is a good proxy for vegetation greenness (Herrmann et al., 2005) and largely correlates with biomass and Leaf Area Index (Tucker, 1979; Sellers, 1985), it is also sensitive to soil humidity and soil colour (with increasing influence at lighter vegetation cover), atmospheric effects and observation geometry (Huete, 1988; Los et al., 1994; Herrmann et al., 2005). The NDVI may vary in the order of 0.07–0.08 owing to differences in soil humidity and soil type (Los et al., 1994). Studies by Huete (1988) and Smith et al. (1990) found that the effect of soil type and humidity on the NDVI becomes greatest if the vegetation cover is <30%–40%. Several attempts have been made to minimise some of these problems. A common approach to decrease short-term atmospheric effects (e.g. cloud cover) is to use Maximum Value

Different dimensions of recent vegetation dynamics of North and West Africa

201

Composites (MVC) (Gutman, 1989). To minimise the effect of soil attributes on the NDVI, alternative vegetation indices were developed, e.g. the Modified Soil Adjusted Vegetation Index (Qi et al., 1994) and the Soil Adjusted Vegetation Index (Huete, 1988). These vegetation indices use a soil factor to assess the effect of the soil properties. However, knowledge of the attributes of the respective soils is needed, and these vegetation indices did not result in the creation of a long-term time series for universal application (Herrmann et al., 2005). In this study, we use the NDVI derived from the Advance Very High Resolution Radiometer Sensor (AVHRR) onboard of the National Oceanic and Atmospheric Administration (NOAA) satellite series as a proxy for vegetation greenness. The dataset used in this study is provided by the Global Inventory Modelling and Mapping Studies (GIMMS) Group (Tucker et al., 2005), available via the IRI Data Library, http://ingrid.ldeo.columbia.edu/. The AVHRR data has several biases (Slayback et al., 2003) that have been minimised or eliminated by the pre-processing performed by the GIMMS group. This pre-processing includes: correcting residual sensor degradation and sensor intercalibration differences, effects of changing solar zenith and viewing angles, volcanic aerosols, atmospheric water vapour and cloud cover. It is reported to be of superordinate quality regarding the biases that affect AVHRR data (Tucker et al., 2005). It has a spatial resolution of 8 km and is a 10-day MVC (Maximum Value Composite) to remove atmospheric effects (Gutman, 1989). The NDVI values of the dataset range from 0 to 1, values <0 are set to missing data (Tucker et al., 2005). Rainfall data One of the major problems of research on environmental issues in arid environments is the frequent absence of reliable long-term information on precipitation. This is especially the case for the region examined in this paper. Rain gauges are sparse here, and the available data of the existing ones are often incomplete (Nicholson, 2005; Olson et al., 2005). It was therefore decided to use two datasets for precipitation analysis: (1) Monthly station precipitation data, provided by the Climate Prediction Center (CPC), (2) monthly gridded satellite precipitation estimates from the Global Precipitation Climatology Project (GPCP) with a spatial resolution of 2,5° * 2,5° (available via the IRI Data Library, http://ingrid.ldeo.columbia.edu/). The latter is generated by the combination of satellite data, global climate models and rain gauges (Xie and Arkin, 1997; Nicholson, 2005). Although error estimates are high in the study area (the GPCP dataset is provided with additional error estimates) and its coarse spatial resolution, GPCP rainfall estimates are shown to be useful for analysing precipitation dynamics on a regional scale (Nicholson et al., 2001, Herrmann et al., 2005). 12.3.2 Methods Pre-processing The NDVI dataset was tested for consistency by visually interpreting the spatially averaged NDVI dataset (Schönwiese, 2000). We found a significant shift in the time series from the middle of 2001 onwards (Figure 2). Olson et al. (2005) and Eklundh and Olson (2003) found a shift in NDVI time series derived from the same sensors after 1999, assuming this to be due to the late afternoon overpass of NOAA-14 (Gutman, 1999). To make sure the problems with NOAA-14 do not influence the findings, the examined time series was restricted to the period 1983 to 1999. Another possible disturbing signal was found in pixels located in the Bodélé Depression (Figure 3). Neither the cause of this rather “artificial” signal could be found in literature nor whether it affects the entire dataset.

202

Beckers and Schütt

Figure 2. Spatially averaged NDVI of the working area. Note the shift from the middle of 2001 onwards.

Figure 3. Temporal profile of the averaged NDVI pixels located in the Bodélé Depression (∼17°N/18°E).

Besides the pre-processing by the GIMMS Group (Tucker et al., 2005), the 10-day MVC, 8 km, NDVI data were converted from Albers Conical Equal Area Projection to Geographic Lat. Long. with a spatial resolution of 0,1° and monthly averaged to match with the temporal resolution and projection of the precipitation data. All pixels of the NDVI dataset with missing data in the corresponding time trajectory were removed from the dataset. This resulted in a 341 * 151 * 204 (x * y * months) pixel NDVI dataset covering 17 years. To examine the association of rainfall and NDVI, the 2,5° * 2,5° GPCP data were resampled to 0,1° * 0,1° to fit in with the NDVI data, as done by Herrmann et al. (2005). This approach leads to a replication of GPCP pixel with the same value but keeps the finer spatial resolution of the NDVI dataset (Herrmann et al., 2005). The CPC station precipitation data in the study area were tested for missing data, and a plausibility check was performed. Only 29 out of 90 available stations showed

Different dimensions of recent vegetation dynamics of North and West Africa

203

usable data >90% in the respective dataset for the period 1983 to 1999, with many of them located in Tunisia. Only stations with less than 2 missing data in the examined period were used for trend analysis. Of these, 7 stations were selected based upon the results of the “per pixel” trend analysis. Because in the first place no significant trends could be found in the respective CPC data, the long-term mean (average of 1983 to 1999) of the corresponding season was computed and trend analysis performed on the departure of the seasonal mean from the long-term average. First attempts to perform cluster analysis did not lead to plausible results. This was assumed to be due to the high number of pixels with low “Signal to Noise Ratio” (SNR) (Coppin et al., 2004) located in the region of the Sahara. Therefore, the pixels assumed to be greatly affected by noise due to low vegetation cover were removed. To remain sensible to processes in desert margins, the following rather low thresholds were chosen to eliminate pixels with low SNR: ΔNDVI < 0,04 and NDVImax < 0,1 in the indicated time trajectory (for the discussion on ΔNDVI thresholds of detection see Tucker et al., 2005). Classification Cluster analysis was performed to identify representative patterns in NDVI dynamics on a regional scale. Working with annual means is considered inappropriate in the working area, for fluctuations of vegetation and their driving climatic forces characterise great parts of the study area (Hulme, 2001). To allow for this circumstance, clustering was applied by using the adjusted 17-year monthly NDVI time series. The cluster analysis was performed using the k-means algorithm (MacQueen, 1967). This involves two major problems: (1) The algorithm requires a prior knowledge of k, (2) k-means is sensitive to the initial starting conditions, because the algorithm may converge to one of numerous local minima (Hand and Kraznowski, 2005; Bortz, 2005). Methods to determine k are often developed if background knowledge about k is not available. Since these are all heuristic methods (Sugar and James, 2003), k was determined on the basis of background knowledge of the number of ecoregions (see section 12.2), the purpose of the cluster analysis and sampling. As similarities are assumed between regional patterns of vegetation dynamics and those of ecoregions in the study area, we set the number of clusters at between 6 and 10. It is assumed that the major factor discriminating cluster in NDVI dynamics in the study area is the shape of the profile of the time trajectory of each pixel. Therefore, correlation was chosen as the distance measurement. Following the proposal of Sugar and James (2003), k-means was run 20 times with 6 to 10 ks. Change detection For the purpose of detecting changes in the cluster distribution of the NDVI dataset over the examined period, the pixels of the first eight years (1983–1990) and the following nine years (1991–1999) were assigned separately to the corresponding section of the centroids obtained by clustering the 17-year period. This approach is assumed to detect changes in the time trajectories of the pixels. In other words, if the first half of a time series of one pixel resembles the shape of one centroid and the second half resembles another centroid, the pixels will switch their assignment, hence marking a change in vegetation dynamics. In order to visualise the changes of the cluster distribution, a map was made by simply subtracting the matrix of the first period from the second one. Pixels changing their assignment from a class with lower mean NDVI to one with higher mean NDVI were termed positive shifts and vice versa as negative shifts. Pixels that do not change their assignment are defined as 0. This does not necessarily mean that these pixels have positive, negative or no trends in NDVI.

204

Beckers and Schütt

Extraction of temporal profiles In order to analyse vegetation and precipitation dynamics on a regional scale, temporal profiles of the NDVI dataset and the resampled GPCP dataset were extracted by spatially averaging the pixels of both datasets based on their assignment to a cluster achieved by clustering the 17-year NDVI dataset. Note that the time series generated largely correlate with the corresponding cluster centroids. Trend analysis, linear correlation and analysis of the growing season Regression analysis was performed to detect long-term trends in rainfall and NDVI. Large parts of the region examined in this study are characterised by a strong seasonality in precipitation and vegetation productivity. Another component of the time series is shortterm fluctuations. Moreover, the influence of noise increases with decreasing MAR and vegetation cover (see section 12.4). These components of the time series might obscure longterm trends in the time series (Schönwiese, 2000; Borradaile, 2003). To distinguish possible trends, rainy/growing season means of each of the broader ecoregions were generated. Rather large time frames were chosen for the length of the seasons to reduce influences due to possible interannual shifts in the seasonal cycle (Olson et al., 2005). Based on section 2. three general rainy/growing seasons can be distinguished in the working area: October to April for the Mediterranean region, July to September for the Sahel, and May to September for the Sudan zone. Trend analysis was then performed by fitting linear functions through the seasonal means, and the trend slopes were calculated (1) of the spatially averaged NDVI and GPCP pixels of the corresponding clusters, (2) for each of the pixels of the NDVI dataset, (3) on the adjusted CPC station rainfall data, and (4) for averaged 3 * 3 NDVI grid cell windows around selected stations of the CPC dataset to analyse the relationship of the CPC rainfall data and NDVI in addition to the regional analysis, as done by Olson et al. (2005). The extracted grid cells correspond to an area of ∼25 * 25 km. To test the association of rainfall and NDVI, Pearson’s correlation coefficient was computed for the two variables with different time lags (1–3 month rainfall prior to NDVI) as done by Nicholson et al. (1990) and Herrmann et al. (2005). This was operated for the spatially averaged NDVI and the converted GPCP dataset. The time lag is related to the fact that the vegetation growth does not react directly to rainfall, but rather to soil moisture which is a several weeks to months integral of rainfall (Malo and Nicholson, 1990). The trends and the correlations were tested for significance on a 0,5% level. For the purpose of visualisation the per pixel trends of step (2) were converted in NDVI changes throughout the study period as done by Herrmann et al. (2005). These are expressed in percentage relative to the value at the starting point of the linear trend line. To detect a possible shift in the length of the regional growing seasons, the start of the growing season was established by ascertaining the month of steepest positive slope in the time series of each year, and its end by determining the month of steepest negative slope (e.g. White et al., 1997). 12.4 RESULTS 12.4.1 Classification The elimination of the pixels removed almost all pixels of the NDVI dataset located in the Sahara, except for isolated pixels in the irrigated region of Murzuq, the Tibesti and Al Haruj Al Aswad (see figure 4, map A). The best result with regard to the presumptions made in section 3.2 was achieved by using 8 cluster centres (k). More than 8 k led to a

Different dimensions of recent vegetation dynamics of North and West Africa

205

Figure 4. Maps of regional vegetation (i.e. NDVI) dynamic patterns in different periods (A, B, C) in the working area. Large parts of the Sahara are omitted due to classification problems. Map D indicates the differences between map B and map C. + is defined as a shift from one class with lower mean NDVI to one class with higher NDVI. – as the opposite, and 0 as no change.

further partition of the Sahel, whereas the similarity of those cluster centres increased. Less than 8 k did not lead to a partitioning of the Mediterranean region. Performing the 20 randomly chosen starting points did not lead to significant changes of the cluster centres. The result of clustering the vegetation dynamics of the 17-year NDVI dataset is shown in figure 4, map A: Class 1, located in the Mediterranean region, encompasses the northern declivity of the Tunisian Atlas and the adjacent coastal lowlands, the Jabal Natusah in the hinterland of Tripoli and, connected by a narrow coastal strip, the Jabal Akhdar. Except for the coastal strip, this class represents areas of highest rainfall and biomass production in the Mediterranean region of the study area (see section 2). Class 2 covers the coastal lowlands around Kairouan, the highland steppe of Jabal Nafusah south of Tripoli and the coastal lowlands of Djefara. According to White’s Vegetation Map of Africa (1983), these are mainly anthropogenic landscapes. The southern- and easternmost parts of class 2 are located on the regs and hamadas of the northern Sahara. The classes 3 to 5 are located in the Sahel. The classes narrow northwards, representing the increasing North–South gradient in vegetation greenness and the increasing ecological gradient typical for desert margins (Schultz, 2005). The trisection of vegetation dynamic classes corresponds to the three ecological subdivisions of the Sahel (see section 12. 2, Dry and Thorn Savanna Climate). Class 3 extends as a narrow latitudinal belt along 16°N and sprawls northwards on the western side of the study area accordant to the Aïr Mountains. Classes 6 and 7 divide the Sudan zone into a northern and a southern part at 12°N (see section 12.2, Wet Savanna Climate). The northern border of class 6 is located at 14°N. Class 8 encompasses, in the terms of White (1983), roughly the Sudanian woodland and the Guineo-Congolian mosaic of rain forest and secondary grassland.

206

Beckers and Schütt

12.4.2 Analysis of the temporal profiles The NDVI of class 1 (Figure 5A) shows a distinct seasonality with minor interannual variability and a low mean NDVI of 0,17. Mean Annual Rainfall (MAR) is at about 220 mm. The growing season starts in October/November and lasts till April/May, whereas the rainy season starts and ends one month earlier, both without changes in the seasonal cycle. The NDVI correlates best with rainfall for a two month time lag (correlation coefficient 0,8, n = 202). Both the NDVI and the GPCP time series exhibit no significant trend. The NDVI of class 2 (Figure 5B) shows a more complex seasonal cycle with variable numbers of annual peaks in the winter months and a distinct interannual variability. The NDVI averages 0,12, and MAR reaches 180 mm. While a correlation with rainfall is missing, the aboveaverage amplitude of the NDVI growing season (i.e. seasons 1985/86 and 1995/96) seems to be linked with above-average rainfall. Again, neither the NDVI nor the rainfall shows a significant trend. Class 3 (see figure 5C) has the lowest mean NDVI of all classes, totalling 0,08. This coincides with the lowest MAR of 100 mm. On this low level interpretation of the profile has to be taken with caution because of the relative strong influence of noise. Precipitation peaks regularly in July, marking the departure of the otherwise erratic month-to-month NDVI profile. Correlation between NDVI and rainfall could not be found, and significant trends are absent in both time series. But again above-average amplitude in NDVI seems to be linked with above-average precipitation. In 1994 the annual precipitation reaches >200 mm, resulting in an increase in the NDVI amplitude of the following months. Moreover, the annual minima of the profile appear to have a positive trend in the following years. This seems to occur again in 1999. Similar to class 3, the NDVI of class 4 (mean NDVI 0,1, and MAR 110 mm) is lacking a distinct seasonality (Figure 5D). The annual peaks are diffuse, but from 1988 onwards recurring positive NDVI fluctuations are visible (1988, 1994, 1999). They are clearly related to fluctuations in precipitation. These fluctuations are also the cause of a significant positive trend in the month of July to August, both in precipitation and NDVI. Moreover, these fluctuations appear to trigger a slight increase of the NDVI of the following years. Below average annual NDVI occurred in 1984 and 1991. The NDVI and rainfall of class 5 (mean NDVI 0,15, MAR 260 mm) (Figure 5E) correlate best with a two month time lag (correlation coefficient 0,8, n = 202). Unlike class 4, the NDVI shows a distinct seasonality with NDVI peaks in July/August. Fluctuations related to rainfall occurred in the same years as in class 4. Similar to class 4 these fluctuations cause an increase in annual NDVI of the following years. Both the rainfall and the NDVI exhibit positive trends in the months July to September. The growing season of class 6 (mean NDVI 0,25, MAR 480 mm) (Figure 5F) lasts for about three months. It starts in July, the month when rainfall reaches maximum. The same fluctuations visible in class 4 and 5 also occur in class 6, although one month earlier. The correlation (0,7, n = 202) between NDVI and rainfall is again best with a two month time lag. Both the NDVI and the rainfall time series show significant positive trends. The NDVI and the rainfall of class 7 (mean NDVI 3,8, MAR 876 mm), (Figure 5G) show significant positive trends. The growing season lasts from July to October/November, whereas the end of the growing season appears to shift from October to November from 1994 onwards. The NDVI of class 8 (mean NDVI 0,53, MAR 1.350 mm) (Figure 5H) exhibits a significant negative trend, whereas precipitation has a significant positive trend. The NDVI has a distinct peak until 1992. Then the profile develops a twofold peak that does not seem to be related to rainfall. The length of the growing season is stable, lasting from April to November. Rainfall correlates best with NDVI on a one month time lag with a correlation coefficient of 0,9 (n = 203).

Different dimensions of recent vegetation dynamics of North and West Africa

207

Figure 5a. Temporal profiles of spatially averaged vegetation dynamic classes and the corresponding GPCP rainfall estimates—classes 1 to 4. The classes refer to figure 4, map A.

208

Beckers and Schütt

Figure 5b. Temporal profiles of spatially averaged vegetation dynamic classes and the corresponding GPCP rainfall estimates—classes 5 to 8. The classes refer to figure 4, map A.

Different dimensions of recent vegetation dynamics of North and West Africa

209

Figure 6. Seasonal rainfall departure of the long-term seasonal mean of selected stations. The NDVI trend refers to a 24 * 24 km grid cell window around the corresponding weather station. The location of the stations is shown in figure 8.

Positive NDVI trends are apparent across the Sahelian part of the study area between 12°N and 16°N in the months May to September, and extend further southwards in the months July to September. Highest positive vegetation trends are to be found on the northern former outline of Lake Chad (∼until 1960) and on patches at about 17°E and 14°N. The most noticeable positive shifts in the Sub-Saharan part of the study area occur as latitudinal belts between classes 3 and 4 and classes 4 and 5. The longitudinal extent of shifts narrows further south. On the easternmost side of the study area rather broad positive shifts are to be found between all classes of the Sub-Saharan part. Notwithstanding, narrow negative shifts occur between classes 7 and 6 and classes 6 and 5 west of 20°E. Between classes 7 and 8 positive shifts dominate, but are intersected by negative changes. Isolated pixels located in the Sahara north of 28°N exhibit positive shifts. Pixels in the Sahara south of this latitude do not switch their assignment. In the Mediterranean part, a major negative shift between classes 1 and 2 occurs in the region of Tripoli. A narrow positive shift is to be found in the Atlas Mountains in Tunisia. 12.4.4 Trend analysis The results of the trend analysis are shown in figure 7. NDVI trends of the Sub-Saharan part in the winter seasons (October to April) are negligible and hence not shown. For the

210

Beckers and Schütt

Figure 7. NDVI trends 1983 1999. The areas with negligible trends are omitted in the respective period.

same reason the Mediterranean region in the months July to September is omitted. The results of the trend analysis of the CPC station data and the corresponding NDVI grid cell windows are shown in figure 6. Stations without significant rainfall trend are omitted in this figure. The positive NDVI trends coincide with positive rainfall trends at the station Maine Soroa in Niger (13°12' N, 12°02'E), which exhibit a positive deviation from the long-term mean from 1988. The time series of the surrounding NDVI grid cell window shows a similar positive trend throughout the period, but exhibits negative deviations from the long-term mean that at least in the amplitude are not related to rainfall (e.g. 1993, 1997).

Different dimensions of recent vegetation dynamics of North and West Africa

211

Table 1. The vegetation dynamic classes (see section 12. 4.) in comparison with the new Land Cover Map of Africa (after Mayaux et al. (2004) and Vegetation Map of Africa (White, 1983)).

Classes (Fig. 4A)

Corresponding land cover types according to the LCA* (Mayaux et al., 2004)

Class 1

Cropland cover >50%

Class 2

Open deciduous wood- and shrub-land Sparse grassland

Class 6

Sandy and stony desert Stony desert Stony desert Sparse grassland Open grassland Open grassland with sparse shrubs Closed grassland

Class 7

Cropland >50% Open woody vegetation

Class 3 Class 4 Class 5

Class 8

Open deciduous shrub land Deciduous woodland with sparse trees Deciduous woodland mosaic forest/savanna

Corresponding major phytochoria according to White’s Vegetation Map of Africa (White, 1983) Mediterranean regional centre of endemism Mediterranean/Sahara regional transition zone Mediterranean/Sahara regional transition zone Sahara regional transition zone Sahel regional transition zone Sahel regional transition zone Sahel regional transition zone

Sudanian regional centre of endemism Sudanian regional centre of endemism Guinea/Sudanian regional transition zone

South of 10°N, negative NDVI trends appear in the months July to September, extending over almost the entire region. It is remarkable that the highlands of the region, e.g. the northern parts of the Adamaoua Mountains (Figure 1), the Mandara Mountains (∼10°N, 13°E), the Jos Plateau (∼9°N, 9°E) and the Salamat floodplains (∼19°N, 20°E), are areas without negative NDVI trends throughout the months July to September. In the period from May to September positive NDVI trends predominate and negative NDVI trends are restricted to isolated patches in the southernmost part of the study area. This region refers to the Sudan zone and the southern parts to the Sudanian/Guinea transition zone. Contrary to the negative NDVI trends, the rainfall of the weather station Bossangoa in the Central African Republic (6°28'N, 17°27'E) exhibits a positive rainfall trend in both rainy seasons. The station Garoua (9°21'N, 19°22'E), located 500 km northwest of Bossangoa, shows a positive rainfall trend in the periods July to September, whereas a significant rainfall trend is lacking for the months May to September. In the Mediterranean region positive NDVI trends are apparent in the summer and the winter seasons, on the lowlands around Kairouan and the adjacent southern declivity of the Atlas in Tunisia. Near Tripoli positive NDVI trends occur in the winters. Whereas

212

Beckers and Schütt

the positive NDVI trends of the region around Kairouan persist in the summer seasons, the NDVI trends of the summer season in the vicinity of Tripoli are negative. The rainfall data of the weather station at Kairouan (35°40'N, 10°06'E) lacks a significant rainfall trend in the winter months, but exhibits a positive rainfall trend in the summer months. Contrary to the rainfall, the NDVI covering this station largely increases from 1988 to 1990 and remains on a rather high level from then on. In the Al Akhdar Mountains near Benina isolated pixels exhibit positive NDVI trends in the winter seasons and negative NDVI trends in the summer seasons. The stations of Tripoli and Benina exhibit no significant rainfall trends in either the summer seasons or the winter seasons. 12.5 DISCUSSION 12.5.1 Classification in comparison with the “New land cover map of Africa” Cluster analysis of the NDVI dynamics was performed using correlation as the distance measurement. According to the distribution of the classes, this resulted in a map that is broadly similar to land cover maps of Africa, e.g. the “New land cover map of Africa” (further denoted as LCA) (Mayaux et al., 2004). This agrees with the successful attempts made to classify vegetation and land cover types, using remotely sensed vegetation dynamics as one of the discriminating factors (e.g. de Fries and Townshend, 1994; Benedetti et al., 1994). It is understood that vegetation and land cover units are permanently subject to change and each recording of vegetation is only momentary (Schulze et al., 2002). Nonetheless, assumptions on the land cover type of the classes and quality assessment of the classification was made in comparison with the “New land cover map of Africa” (http://www.gvm.jrc. it/glc2000.htm; Mayaux et al., 2004). The LCA map was built using the spectral response and the temporal profiles of the vegetation cover, digital image processing, radar data and thermal sensors and local knowledge (Mayaux et al., 2004). The main difference of the two maps is the extent of what is supposed to be desert: the desert/grassland border of the LCA is located at approximately 14°N, whereas the border of class 3 is located 1° further north, around 15°N (Figure 4). In contrast, classes 3 and 4 are located on stony desert according to the LCA. Temporal profiles confirm that these arid areas show long-term detectable vegetation dynamics (Figure 5a). Moreover, the recurring fluctuations of the temporal profile of class 4 can be tracked further south to class 6. This makes it likely that the region represented by class 4 is part of the rainfall and ecosystem regime of the Sahel. In the Mediterranean region class 2 is partly located on sandy and stony desert and locally corresponds to sparse grassland—both according to the LCA. The erratic and ambiguous character of the temporal profile suggests that class 2 is a mixture of pixels with great differences in temporal behaviour and is hence not suitable to characterise the vegetation dynamics of the region in which it is located. With regard to LCD, the area covered by class 5, located in the Sub-Saharan part of the study area, is characterised by sparse grassland in the north, changing to open grassland, and open grassland with sparse shrubs in the south. The sharp increase and distinctive peaks in NDVI each July imply that the entire region receives rainfall at the same time every year. Class 6 covers closed grassland areas and areas with cropland >50% coverage. The temporal profile shows a similar, but more distinctive, behaviour to that of class 5. Class 7 is located on croplands with open woody vegetation and open deciduous shrub land. Class 8 encompasses deciduous woodland with sparse trees, deciduous woodland and mosaic forest/savanna. The smooth and regular temporal profiles of both classes indicate that these are suitable to represent the vegetation dynamics of the regions.

Different dimensions of recent vegetation dynamics of North and West Africa

213

12.5.2 Change detection To detect changes in the cluster distribution the NDVI dataset was split up into two periods: 1983 to 1990 and 1991 to 1999. The pixels were assigned to the respective period of the cluster centroids obtained by clustering the 17-year period. Correspondingly, shifts in the cluster distribution indicate a change in the temporal behaviour of the pixels through time. The resulting change map (Figure 7) confirms the widespread assumption that temporal and large-scale shifts in vegetation distribution and dynamics are to be found at the boundaries of ecoregions, particularly in arid and semiarid areas (e.g. Ellen et al., 1998; Schulze et al., 2002). Moreover, the distribution of the shifts largely coincides with positive/negative patterns of NDVI trends shown in figure 5b. To discuss and assess shifts, time trajectories of arbitrarily chosen pixels that change their assignment were visually interpreted. The positive shifts between classes 3 and 4 appear to be related to the recurring fluctuations of rainfall in 1995. The general character of the time trajectory of the corresponding pixels (e.g. no distinct seasonality) does not change. Therefore, this shift only gives an indication about the northern extent of the rainfall fluctuations. The shift between classes 4 and 5 can be explained by the development of a distinct annual July-peak in 1990. This indicates a northern displacement of the tropical rain belt over West Africa. This is in agreement with findings of Nicholson (2005) who stated that the influence of the rain belt extends further north from 1990. The pixels switching the assignment from class 7 to class 8 appear to develop a similar twofold peak in the growing season. Changes between classes 1 and 2 could not be assessed, because class 2 is assumed to be non-representative. Notwithstanding, the negative change between classes 1 and 2 agrees with a negative NDVI trend. Visual interpretation of pixels exhibiting negative shifts between classes 6 and 5 and classes 5 and 4 failed, because obvious changes in the shape of the temporal profiles could not be found. 12.5.3 Temporal profiles The results of the correlation analysis of GPCP data and NDVI agree with previous studies on the relationship of vegetation and rainfall in the Sahelian part of the study area (e.g. Nicholson 1990; Olson et al., 2005; Herrmann et al., 2005). Furthermore, they agree with the general assumption that vegetation dynamics and biomass production in the study area is largely driven by rainfall. Nicholson (1990) showed that the relationship between rainfall and NDVI in the Sahel and East Africa is good on seasonal and annual time scales for areas where MAR ranges from ∼200 mm to 1.200 mm. This confirms the finding in this study that a correlation between rainfall and NDVI could not be found in the classes with MAR <200 mm. Herrmann et al. (2005), also using NDVI and GPCP datasets, established that rainfall is the dominant causative factor of vegetation greenness in the Sahel. Nevertheless, the same authors suggest that human activities contribute to the distribution and state of vegetation in the Sahel (Herrmann et al., 2005). The fluctuations of NDVI in the Sahel in 1988, 1994 and 1999 are strongly related to rainfall fluctuations as well as the below-average NDVI in 1984 and 1991. This is in agreement with findings of Anyamba and Tucker (2005) and Nicholson (2005). Anyamba and Tucker (2005) compared the fluctuations of NDVI with fluctuations of ENSO and found that these are not directly linked. A positive vegetation rainfall feedback in the Sahel as simulated by Zeng et al. (1999) was not confirmed by visual interpretation of the

214

Beckers and Schütt

temporal profiles. The relationship of NDVI and rainfall of class 1 is in agreement with the results of Wellens (1997) who analysed this relationship in Tunisia with NDVI data. 12.5.4 Secular per pixel trends The positive NDVI trends in the Sahel are consistent with previous findings of Anyamba and Tucker (2005), Olson et al. (2005) and Herrmann et al. (2005). Furthermore, these positive NDVI trends agree with the positive rainfall trends of the temporal profiles in the region and the positive rainfall trends of the Maine Soroa station. Despite widespread agreement that the positive NDVI trends in the Sahel are mainly driven by positive rainfall trends (see section 2), other weaker causative factors of the secular vegetation trends are assumed by several authors. For example Olson et al. (2005) compared the NDVI trends to statistics of urban growth rates and figures on internally displaced people. Negative NDVI trends of the Sudanian/Guinean zone occur in the months July to September (Figure 7). They are contrary to the positive or absent (not significant) rainfall trends, both in the CPC data of the stations at Garoua and Bossangoa and the temporal rainfall profile of classes 7 and 8. A negative NDVI trend due to a shift in the seasonal peak is improbable, for the season peaks regularly in August throughout the study period. This and the pattern of negative NDVI trends indicate that there might be another weaker causative factor of vegetation greenness. A possible human factor that contributes to negative NDVI trends was not possible to assess because suitable population or economic figures were not available. In the region of Murzuq east of Al Haruij Al Aswad a group of isolated pixels show positive NDVI trends. These pixels correspond to irrigation areas in the Murzuq basin. This becomes obvious if Landsat MSS images (path 186 188; row 41 42) of 1970 are compared to Landsat TM images of 1980 and 1990. The intensification of irrigation appears in particular as an increase of circular fields, a characteristic pattern of centre pivot irrigation (Tarjuelo et al., 1999). The absence of a significant rainfall trend in the data of the weather station of Sebha confirms this assumption. In the Sahara pixels located along 17°N show negative NDVI trends. Whereas these NDVI trends are relative strong (α < 0,01), the absolute NDVI change throughout the study period is <0,02. This also applies to the positive NDVI trends that distinguish the Tibesti and the Al Hauruj Al Aswad from their surroundings. Moreover, the vegetation distribution of the Tibesti is reported to be contracted and appears diffuse only in altitudes above 2.000 m asl (Walter, 1990). Accordingly it is most likely that these positive NDVI trends are not related to significant vegetation change. Nevertheless, at least the positive NDVI trends in Tibesti and Al Haruj Al Aswad are not likely to be accidental; hence NDVI trends related to problems with AVHRR data are implausible (see section 12.3). A possible factor could be increased soil humidity or the specific reflectance character of the soil/rock surface of the Tibesti and Al Haruj Al Aswad, which influences the NDVI. Furthermore, photosynthetic active mycrophytic communities may affect the NDVI. Karnieli et al. (1999) showed that the NDVI of active mycrophytic communities can be as high as 0,3 and pointed out that this may lead to an overestimation of ecosystem productivity. However, the photosynthetic activity of these mycrophytic communities is dependent on moisture availability (Karnieli et al., 1999). A possibly corresponding significant rainfall trend in the Tibesti and Al Haruj Al Aswad cannot be verified owing to the lack of rainfall data. The positive NDVI trends around Kairouan are contrary to the rainfall trend in the temporal profile of class 1 which lacks a significant rainfall trend. However, they agree with the positive rainfall trend at the Kairouan weather station. The temporal NDVI profile neighbouring the weather station shows the typical pattern for areas with increasing

Different dimensions of recent vegetation dynamics of North and West Africa

215

irrigation (Ozdogan et al., 2003). Furthermore, irrigation statistics of Tunisia show an increase in irrigation activity in the governorate of Kairouan since 1990 (FAO, 2000). The negative NDVI trends around Tripoli and Benina are difficult to assess. Literature on contemporary vegetation dynamics in the Mediterranean, part of the study area as well as reliable statistics on population growth or agriculture are rare and mostly available on the national Libyan scale. However, these negative vegetation trends could be the continuance of land degradation due to overgrazing and deforestation since Roman times in the Mediterranean region proposed by White (1983) and Le Houerou (2001), for example. 12.5.5 Discussion of methods The unsupervised classification approach (here: k-means) leads to generally plausible results compared to the “New land cover map of Africa” (Mayaux et al., 2004) and offers advantages, especially where field information is not available. However, pixels which were assumed to have low SNR (Signal to Noise Ratio) had to be removed for classification, hence revealing one limit of this approach in the study area: coping with very low SNR pixels. Because of this, a continuous classification of the dataset could not be made with this approach. The low thresholds, used to remove pixels with low SNR, were based on rather global assumptions on ΔNDVI thresholds of detection, but these thresholds differ from dataset to dataset and are dependent on the respective landscape attributes (Cihlar, 2000; Tucker et al., 2005). Moreover, (1) the clustering is based on different characteristics of the pixel, including noise and background signals, making a consistent class description impossible, (2) the need for an a priori knowledge of the number of clusters may lead to an over or under-segmentation of the dataset, (3) the heuristic character of the k-means algorithm hinders the repeatability and the comparability of results (Geerken et al., 2005). These problems are likely to affect the classification made. Especially the classes located in the more arid areas (i.e. desert margins) are assumed to contain largely different vegetation dynamic patterns and noise because of their erratic character regarding the distribution and dynamics of vegetation. More cluster centres are assumed to be more suitable for partitioning these areas of steepest ecological gradients. Nevertheless, the initial approach of classifying long-term time series of NDVI with correlation as a discriminating factor seems to be a suitable and feasible way to gain knowledge about regional patterns of vegetation dynamics in the study area. The extraction of temporal NDVI (Figure 4) and precipitation profiles (Figure 6) is dependent on the quality of the initial classification. Furthermore, the resampling of the GPCP dataset resulted in an incidental overlapping use of the same precipitation grid cell. This problem is particularly strong in the rather narrow classes of 3–5. Finally the change detection approach used in this study led to several inaccuracies: the dispartment of the dataset is arbitrary. Therefore assumptions on the exact time of the changes cannot be made and the accuracy of the change detection is largely dependent on the initial classification and the obtained cluster centroids. Hence, it is not possible to assess changes between classes assumed to be non-representative. 12.6 CONCLUSIONS This study aims to analyse vegetation dynamics and their relationship to rainfall variability in North and West Africa from 1983 to 1999. The data base was a ten-day NDVI dataset with a spatial resolution of 8 km, a 2,5° * 25° monthly GPCP dataset and ground measurements of rainfall. It is shown that the distribution of regional vegetation dynamic patterns generally

216

Beckers and Schütt

coincides with large-scale land cover types, especially in the sub-Saharan part of the study area. Change in the distribution of the vegetation dynamic patterns correlates with secular vegetation trends in this period. The timing of growing cycles on a regional scale is constant in almost all parts of the study area. The above-mentioned findings suggest that secular vegetation trends and shifts in vegetation dynamic patterns in the Sahelian part of the study area are related to shifts in the position of regional rainfall and ecosystem regimes rather than to the development of ”new” ones. However, a 17-year NDVI dataset is too short to assess accurately long-term environmental variability or even “regime shifts” in the sense of Foley et al. (2003). The study also adds to previous results on the strong positive relationship of land use, rainfall and positive NDVI trends in the Sahel (e.g. Nicholson, 1990; Anyamba and Tucker, 2005). Regional negative vegetation trends were found in the Sudan zone. Although vegetation dynamics largely correlates with rainfall dynamics in this region the negative NDVI trends are contrary to positive rainfall trends in the Sudan zone. In the Mediterranean region vegetation changes are most likely to be related to human activities such as irrigation and deforestation. However, like previous studies on large-scale environmental changes in the region this study also lacks reliable population and economic figures that fit spatially and temporally to the corresponding vegetation changes. Hence assumptions about the human factor and its influence on large scale vegetation dynamics remain speculative. Moreover, this study shows that a continuous unsupervised classification of vegetation dynamics using k-means and the temporal profiles of a monthly NDVI dataset as the discriminating factor is not suitable for the study area. Finally we found an artificial signal in the Bodélé Depression whose effects on NDVI data require further investigation. ACKNOWLEDGEMENTS We would like to thank the German Research Foundation (DFG) for financial support of the ‘Limnosahara’ research project (Schu 949/8), funded since 2005. REFERENCES Adams, M.E., 1999, Savanna environments, In: Goudie, A.S., Adams, W. and Orme, A.R. The Physical Geography of Africa. Oxford University Press. Allen, C.D. and Breshears, D.D., 1998, Drought-induced shift of a forest–woodland ecotone: rapid landscape response to climate variation. Proceedings National Academic Science USA, 95, pp. 14839–14842. Anhuf, D. and Frankenberg, P., 1991, Die naturnahen Vegetationszonen Westafrikas. Die Erde, 122, pp. 243–265. Anyamba, A. and Tucker, C.J., 2005, Analysis of Sahelian vegetation dynamics using NOAAAVHRR NDVI data from 1981–2003. Journal of Arid Environments, 63, pp. 596–614. Benedetti, R., Rossini, P. and Taddei, R., 1994, Vegetation classification in the Middle Mediterranean area by satellite data. International Journal of Remote Sensing, 15, pp. 583–596. Borradaile, G., 2003, Statistics of earth science data. Their distribution in time, space, and orientation. Springer, Berlin, Heidelberg, New York. Bortz, J., 2005, Statistik. Heidelberg, Springer Medizin Verlag, 6. Auflage. Breckle, S.-W., 2002, Walters’s vegetation of the earth. The ecological systems of the geo-biosphere, Springer, Berlin, Heidelberg, New York. Busche, D., 1973, Die Entstehung von Pedimenten und ihre Überformung, untersucht an Beispielen aus dem Tibesti-Gebirge, République du Tchad. Arbeit aus der Forschungsstation Bardai/Tibesti.

Different dimensions of recent vegetation dynamics of North and West Africa

217

Busche, D., 1998, Die zentrale Sahara. Klett, Stuttgart. Charney, J., Stone, P.H. and Quirk, W.J., 1975, Drought in the Sahara: a biogeophysical feedback mechanism. Science, 187, pp. 434–435. Cihlar, J., 2000, Land cover mapping of large areas from satellites: status and research priorities. International Journal of Remote Sensing, 21, pp. 1093–1114. Claussen, M., Brovkin, V., Ganpolski, A., Kubatzki, C. and Petroukhov, V., 2003, Climate Change in Northern Africa: the past is not the future. Climatic Change, 57, pp. 99–118. Coppin, P., Jonckheere, I., Nackaerts, K., Muys, B. and Lambin, E., 2004, Digital change detection methods in ecosystem monitoring: A review. International Journal of Remote Sensing, 25, pp. 1565–1596. Cracknell, A.P., 1997, The Advanced Very High Resolution Radiometer. London, Taylor and Francis. de Fries, R.S. and Townshend, J.R.G., 1994, NDVI-derived land cover classification at global scales. International Journal of Remote Sensing, 15, pp. 3567–3586. de Menocal, P.B., Ortiz, J., Guilderson, T., Adkins, J., Sarnthein, M. and Baker, L., 2000, Abrupt onset and termination of the African humid period. Rapid climate response to gradual insolation forcing. Quaternary Science Review, 19, pp. 347–361. Dierschke, H., 1994, Pflanzensoziologie: Grundlagen und Methoden. Stuttgart, Ulmer. Eklundh, L. and Olsson, L., 2003, Vegetation index trends for the African Sahel 1982–1999. Geophysical Research Letters, 30. FAO, 2000. http://www.fao.org/nr/water/aquastat/irrigationmap/tn/index.stm (12.10.2007) Foley, J.A., Coe, M.T., Scheffer, M. and Wang, G., 2003, Regime shifts in the Sahara and Sahel: interactions between ecological and climatic systems in Northern Africa. Ecosystems 6, Gebrüder Borntraeger, Berlin, Stuttgart, pp. 524–539. Geerken, R., Zaitchik, A. and Evans, J.P., 2005, Classifying rangeland vegetation type and coverage from NDVI time series using Fourier Filtered Cycle Similarity. International Journal of Remote Sensing, 26, pp. 5535–5554. Gutman, G., 1989, On the relationship between monthly mean and maximum-value composite normalized vegetation indices. International Journal of Remote Sensing, 10, pp. 1317–1325. Gutman, G., 1999, On the use of long-term global data of land reflectance’s and vegetation indices derived from the advanced very high resolution radiometer. Journal of Geophysical Research, 104, pp. 6241–6255. Hand, D.J. and Krzanowski, W.J., 2005, Optimising k-means clustering results with standard software packages. Computational Statistics & Data Analysis, 49, pp. 969–973. Hastenrath, S., 1990, Prediction of the Northeast Brazil rainfall anomalies. Journal of Climate, 3, pp. 893–904. Herrmann, S.M., Anyamba, A. and Tucker, C.J., 2005, Recent trends in vegetation dynamics in the African Sahel and their relationship to climate. Global Environmental Change, 15, pp. 394–404. Holben, B.N., 1986, Characteristics of maximum-value composite images from temporal AVHRR data. International Journal of Remote Sensing, 7, pp. 1417–1434. Huete, A.R., 1988, A soil-adjusted vegetation index (SAVI). Remote Sensing of Environment, 25, pp. 295–309. Hulme, M., 2001, Climatic perspectives on Sahelian desiccation: 1973–1998. Global Environmental Change, 11, pp. 19–29. Karnieli, A. and Dall’olmo, G., 2002a, Monitoring phenological cycles of desert ecosystems using NDVI and LST data derived from NOAA-AVHRR imagery. International Journal of Remote Sensing, 23, pp. 4055–4071. Karnieli, A., Gabai, A., Ichoku, C., Zaady, E. and Shachak, M., 2002b, Temporal dynamics of soil and vegetation spectral responses in a semi-arid environment. International Journal of Remote Sensing, 23, pp. 4073–4087.

218

Beckers and Schütt

Karnieli, A., Kidron, G.J., Glaesser, C. and Ben-Dor, E., 1999, Spectral characteristics of cyanobacteria soil crust in semiarid environments. Remote Sensing of Environment, 69, pp. 67–75. Lamprey, H.F., 1988, Report on desert encroachment reconnaissance in Northern Sudan: 21 October to 10 November 1975. Desertification Control Bulletin, 17, pp. 1–7. Le Houerou, H.N., 2001, Biogeography of the arid steppeland north of the Sahara. Journal of Arid Environments, 48, pp. 103–128. Le Houerou, H.N., 1980, The rangelands of the Sahel. Journal of Range Management, 33, pp. 41–46. Le Houerou, H.N., 1996, Climate, flora and fauna changes in the Sahara over the past 500 million years. Journal of Arid Environments, 37, pp. 619–649. Lezine, A.-M., Casanova, J. and Hillaire-Marcel, C., 1990, Across an Early Holocene humid phase in Western Sahara, pollen and isotope stratigraphy. Geology, 18, pp. 264–267. Los, S.O., Justice, C.O. and Tucker, C.J., 1994, A global NDVI data set for climate studies derived from the GIMMS continental NDVI data. International Journal of Remote Sensing, 15, pp. 3493–3618. MacQueen, J.B., 1967, Some methods for classification and analysis of multivariate observations. Proceedings of the Fifth Symposium on Math, Statistics, and Probability Berkeley, CA: University of California Press. Mainguet, M., 1991, Desertification: Natural Background and Human Mismanagement. Springer, Berlin. Malo, A. and Nicholson, S.E., 1990, Study of rainfall and vegetation dynamics in the African Sahel using Normalized Difference Vegetation Index. Journal of Arid Environment, 19, pp. 1–24. Mayaux, P., Bartholomé, E., Fritz, S. and Belward, A., 2004, A new landcover map of Africa for the year 2000. Journal of Biogeography, 316, pp. 861–77. Meadows, M.E., 1999, Biogeography. In Goudie, A.S., Adams, W. and Orme, A.R.: The Physical Geography of Africa. Oxford University Press. Myneni, R.B., Hall, F.B., Sellers, P.J. and Marshak, A.L., 1995, The interpretation of spectral vegetation indices. IEEE Transactions on GeoScience and Remote Sensing, 33, pp. 481–486. Nicholson, S., 1980, African environmental and climatic changes and the general atmospheric circulation in Late Pleistocene and Holocene. Climatic Change, 2, pp. 313–348. Nicholson, S., 1990, A comparison of the vegetation response to rainfall in the Sahel and East Africa, using normalized difference vegetation index from NOAA AVHRR. Climatic Change, 17, pp. 209–241. Nicholson, S., 2000, The nature of rainfall variability over Africa on time scales of decades to millennia. Global and Planetary Change, 26, pp. 137–158. Nicholson, S., 2001, Climatic and environmental change in Africa during the last two centuries. Climate Research, 17, pp. 123–144. Nicholson, S., 2005, On the question of the “recovery” of the rains in the West African Sahel. Journal of Arid Environments, 63, pp. 615–641. Nicholson, S., Tucker, C.J. and Ba, M.B., 2001, Desertification, drought, and surface vegetation: an example from the West African Sahel. Bulletin of the America Meteorological Society, 79, pp. 815–829. Nicholson, S.E., Davenport, M.L. and Malo, A.R., 1990, A comparison of vegetation response to rainfall in the Sahel and East Africa using normalized difference vegetation index from NOAA-AVHRR. Climate Change, 17, pp. 209–241. Olson, L., Eklundh, L. and Ardö, J., 2005, A recent greening of the Sahel—trends, patterns and potential causes. Journal of Arid Environments, 63, pp. 556–566.

Different dimensions of recent vegetation dynamics of North and West Africa

219

Ozdogan, M., Woodcock, C.E. and Salvucci, G.D., 2003, Monitoring changes in irrigated lands in Southeastern Turkey with remote sensing. Geoscience and Remote Sensing Symposium 2003, IEEE International, 3, pp. 1570–1572. Pachur, H.J. and Hoelzmann, P., 1991, Paleoclimatic implications of Late Quaternary lacustrine sediments in Western Nubia, Sudan. Quaternary Research, 36, pp. 257–276. Schönwiese, C.-D., 2000, Praktische Statistik für Meteorologen und Geowissenschaftler. (Berlin, Stuttgart: Gebrueder Borntraeger), pp. 1–298. Schultz, J., 2005, The ecozones of the world. The ecological divisions of the geosphere. Springer, Berlin, Heidelberg, New York. Schulze, E.D., Beck, E. and Müller-Hohenstein, K., 2002, Plant ecology. Spektrum Akademischer Verlag. Sellers, P.J., 1985, Canopy reflectance, photosynthesis and transpiration. International Journal of Remote Sensing, 6. Slayback, D.A., Pinzon, J.E., Los, S.O. and Tucker, C.J., 2003, Northern hemisphere photosynthetic trends 1982–99. Global Change Biology, 9, pp. 1–15. Smith, M.O., Ustin, S.L., Adams, J.B. and Gillespie, A.R., 1990, Vegetation in desert: I. A regional measure of abundance from multispectral images. Remote Sensing of Environment, 31, pp. 1–26. Street-Perrott, F., Mitchell, J., Marchand, D. and Brunner, J., 1990, Milankovitch and albedo forcing of the tropical monsoon: a comparison of geological evidence and numerical simulations for 9.000 yr BP. Trans. R. Soc. Edin. (Earth Science), 81, pp. 407–27. Sugar, C. and James, G., 2003, Finding the number of clusters in a data set: an information theoretic approach. Journal American Statistic Association, 98, pp. 750–763. Tarjuelo, J.M., Montero, J., Honrubia, F.T., Ortiz, J.J. and Ortega, J.F., 1999, Analysis of uniformity of sprinkle irrigation in a semi-arid area. Agricultural Water Management, 40, pp. 315–331. Thomas, D., 1997, Science and the desertification debate. Journal of Arid Environments, 37, pp. 599–608. Tucker, C.J., 1979, Red and photographic infrared linear combination for monitoring vegetation. Remote Sensing of the Environment, 8, pp. 127–150. Tucker, C.J., Pinzon, J.E., Brown, M.E. and Slayback, D.A., 2005, An extended AVHRR 8-km NDVI dataset compatible with MODIS and SPOT vegetation NDVI data. International Journal of Remote Sensing, 26, pp. 4485–4498. Walter, H., 1990, Vegetation und Klimazonen: Grundriß der globalen Ökologie. Stuttgart, Ulmer, 1990. Ward, M.N. and Follard, C.K., 1991, Prediction of seasonal rainfall in the North Nordeste of Brazil using eigenvectors of sea-surface temperature. International Journal of Climatology, 11, pp. 711–743. Weischet, W. and Endlicher, W., 2000, Regionale Klimatologie. Teubner, Stuttgart, Leipzig. Wellens, J., 1997, Rangeland vegetation dynamics and moisture availability in Tunisia: an investigation using satellite and meteorological data. Journal of Biogeography, 24, pp. 845–855. White, F., 1983, The vegetation of Africa. A descriptive memoir to accompany the UNESCO/ AETFAT/UNSO vegetation map of Africa. UNESCO, Paris, France. White, M.A., Thomton, P.E. and Running, S.W., 1997, A continental phenology model for monitoring vegetation responses to interannual climatic variability. Global Biogeochemical Cycles, 11, pp. 217–234. Xie, P. and Arkin, P.A., 1997, Global precipitation: A 17-year monthly analysis based on gauge observations, satellite estimates, and numerical model outputs. Bulletin of the American Meteorological Society, pp. 2539–2558. Zeng, N., Neelin, J.D., Lau, K.-M. and Tucker, C.J., 1999, Enhancement of interdecadal climate variability in the Sahel by vegetation interaction. Science, 286, pp. 1537–1540.

CHAPTER 13

Landmines, drugs and justice. The recent history of two Saharan mountains (Adrar des Iforas/Mali and Air Mts./Niger) Issa Ousseïni and Aboubacar Adamou Département de Géographie, Faculté des Lettres, Université Abdou Moumouni, Niamey, République du Niger Erhard Schulz Geographical Institute, University of Würzburg, Germany ABSTRACT: The present rebellions of the Tuareg against the central governments in Niger and Mali may for a greater part be explained on the base of the different ecological and economical situation of the Adrar des Iforas (Mali) and Air Mountains (Niger) (Figure 1). The Tuareg developed different strategies to cope with the limited resources of the desert environment. The Kel-n-Adrar in Northern Mali evolved a pastoral economy combined with a diversified gathering of permanent and spontaneous vegetal resources without a noteworthy horticulture. It resulted in a certain independence of alimentation as well as their social and political behaviour. This economy is also characteristic for the Tuareg of the Tamesna plain between the mountains and the western rim of the Air Mountains The Kel Owey of the central Air developed a tripartite economy based on horticulture, camel and small animal breeding as well as on a triangular trade between North-eastern Niger, Air Mountains, Southern Niger and Northern Nigeria. This trade assured the food supply, which could not be provided sufficiently from the local sources. Thus, they are not independent but incorporated into the economic system of Niger. The drought periods of the last decades as well as the economic changes in Mali and in Niger resulted into a principal thread of both societies. However, external economic forces also influence the present rebellions. In Mali both parties negotiate to find a solution for this conflict. In Niger both conflict sides have not the force or will to negotiate in order to leave the political impasse. Moreover the use of landmines represents a long-time threat for the civil population and an economic recovery. In another consequence, scientific fieldwork will be difficult if not impossible in the future.

13.1 INTRODUCTION The amelioration of the ecological status of the Sahel and Southern Sahara, which has been discussed as a “greening of the Sahel” in the last years (cf. Herrmann and Hutchinson, 2005) improved the pasture conditions in the some parts of the region and thus the economic situation of the population. However, regional conflicts will largely counteract these chances. During the last years a civil war started again in Northern Mali and Niger. The peace agreements of 1992/1995, which terminated the rebellion of the early 1990ies were no longer respected neither in the Air Mountains nor in the Adrar des Iforas. However, the two main Tuareg groups acted differently. After several clashes with the Malian army

222

Issa Ousseïni et al.

another treaty was signed in Algiers (Minist. Malien d’Ext., 2006), in which the central government of Mali finally accepted to fulfil the conditions of the peace treaty of 1992 conceding a certain internal autonomy, integration of rebellion members into the army as well as infrastructural measurements in the Kidal region. In Niger the situation is different. A new “Tuareg” rebellion army was formed in 2006 and succeeded in several attacks against the Niger armed forces, civil persons, and tourists as well in capturing logistic material, arms and vehicles. Fights and mutual attacks were concentrated on the western foreland of the Air Mountains between Arlit and Agadez, the northern Air Mountains as well as in the region of Agadez itself. Apart from material losses a great number of people were killed and and civil security collapsed in the region. In contrast to the civil war in the early 1990ies antipersonnel and antitank mines are used in great numbers in this conflict. Both sides accuse each other for use of landmines denying it for themselves. At the moment no end of the conflict is visible. The rebellion army declares to fight for justice (Nigrian Movement for Justice: MNJ) and for a reasonable participation of the population on the profits taken out of the mineral resources. However, on their internet side (http://M-N-J.blogspot.com) they adopted a romantic vocabulary of violence announcing their military success against the Niger army. They mention ancient plans of an autonomous region/state or Tuareg republic retracing to former ideas of a Central Saharan political entity as the OCRS (Organsiation Commune des Régions Sahariennes) (cf. Gregoire, 1999). The rebels claim official negotiations with the central government, which the Niger government strictly refuses in declaring that there is no idea to negotiate with bandits employing a similar diction and vocabulary as the rebellion movement (JA, 2008b). The situation got worse during the last year and the government declared the martial status over the Agadez region, finally the North of Niger. Since August 2007 only very few and contradictory information is available, the government interdicts reports on the region, also arresting several journalists being false and state endangering publicity (Le Monde, 15.1.08). Thus, at present there is no end of this conflict visible. This article tries to explain the evolution of the conflict in regard of the exploitation of different physical resources as well as with the external conflict sources (see also Schulz et al., 2000). In addition it will explain why fieldwork for geographical/ecological investigations will be profoundly threatened. 13.2 THE PHYSICAL SETTING. THE BASE OF PEOPLE’S ACTION The Tuareg of the Adrar des Iforas in Northern Mali and of the Air Mountains in Northern Niger (Figure 1) are forced to arrange with a desert environment with only limited resources, if they want to earn their living from the region itself. Compared to the Ahaggar with a change to semidesert from 2.000 m asl on, the two mountains have their Saharan character in common, however the Air Mountains show an altitudinal change from desert vegetation to savanna (Figure 2). Both mountains receive a similar amount of rainfall rarely exceeding 120 mm per year. The Adrar des Iforas only reaches to about 900 m asl in altitude thus providing only poor permanent vegetation and tree pasture, whereas the resources of the Air Mountains are much more diversified also with a series of mountain pastures. However the forelands of the Adrar des Iforas provide excellent and regular pastures in large extensions. Consequently the Kel-n-Adrar developed an economy mainly based on a nomadic exploitation of the pastures in the western and southwestern foreland of the Adrar des Iforas principally relying on milk and milk products. There are very few settlements and only recently they started with some garden exploitation. Klute (1992) described this

The recent history of two Saharan mountains

223

Figure 1. The situation of the Adrar des Iforas in N-Mali (left) and of the Air Mountains in N-Niger (right) also showing the climate diagrams of Kidal (Mali) and Agadez (Niger) (after Department of Geography Niamey 2000, modified.).

Figure 2. The altitudinal change of vegetation in the Sahara. Comparison of the Adrar des Iforas (left), Ahaggar (above) and Bagzan/Air Mountains (below) (Schulz et al., 2001, modified).

224

Issa Ousseïni et al.

as a “milk economy”. In normal years these pastures provide certain autonomy for the Kel-n-Adrar. However, in periods of severe drought they cannot rely on an additional tree pasture in the face of the rareness of this vegetation type (see also Sidiyene, 1996; Voss and Kral, 1994a, b). The tree pasture is restricted to some wadis and mostly present in the mountains, as shown in figure 3. In the forelands of the mountains and plains of Tilemsi, Tamesna and Ténéré the achabs provide short time but often extended pasture, which may also incorporate wild millet fields allowing collection of grains up to 250 kg/ha (Schulz and Adamou, 1997).

Figure 3. The pasture resources in the Sahara. Contracted wadi vegetation in the mountains with Acacia raddiana, A. nilotica and Panicum turgidum (a), achabs of Cyperus conglomeratus (b) and wild millet fields with Sorghum aethiopicum (c).

The recent history of two Saharan mountains

225

In contrast to the Kel-n-Adrar the Tuareg of the Air Mountains especially the Kel Evey, could not rely only on these resources and developed early a diversified economy based on an oasis exploitation with villages, animal breeding mainly of goats and camels and caravan trade, which assures the necessary supply of millet. Thus, they are not autonomous, but they got part of the economy of the agriculture of the savannas of Niger and Northern Nigeria and thus, based on their mobility they know the state of Niger best (Adamou, 1979; Bernus, 1981, 1991; Hamani, 1989; Spittler, 1989; Schulz et al., 2001). 13.3 THE EVOLUTION OF THE CONFLICT. BETWEEN ADAPTATION AND EXTERNAL PRESSURE An understanding of the present conflict demands knowledge on the historical evolution of the economic and social situation from the times of colonial regime on. It is necessary to evoke the different resource managements, their natural limits and the various actions of the Tuareg out of these basic facts. In addition the interferences from outside, under colonial rule and later on southern central regimes and recently the internationalisation of conflicts shall be considered too. Thus, finally one may distinguish five main periods in the history of the Tuareg societies. The figures 4 to 9 will explain the different economic and social relations and political events in the Adrar des Iforas and the Air Mountains during the last hundred years in order to give an integral picture of this region. 13.3.1 A collapse of the traditional social and economic systems of the Tuareg of Kel-n-Adrar and Kel-Air The beginning of the 20th century saw a general transformation of the social and economic systems in the Southern Sahara (Figure 4). The main activity of the Tuareg groups was animal breeding (camels, goats, sheep and cattle). It was organised in a semi nomadic and nomadic exploitation of large areas of the Sahara. Gathering, some oasis culture as well as hunting completed the economic base. The Tuareg clans still had their traditional hierarchy (cf. Bernus, 1967; Spittler, 1989). Animal breeding was arranged on a family or clan base and the work distribution mainly relied on slaves or dependent people. Traditionally they were captured during razzias or by ordinary robbery. Only some Tuareg groups practised the caravan trade. However, the control of caravan routes or the toll collection was incorporated in a general razzia system, which could provide a surplus of animals and work force by slaves. In this system the Tuareg obeyed some rules as they refused to kill the people, only taking animals, goods and slaves with a later chance of negotiation or of a counter-razzia. The Tuareg of the Air Mountains were also incorporated into an international trade in providing transport and orientation facilities for the caravans departing from Agadez to the North and Northeast. Simple robbery always was a second way to exploit the international caravans. In parallel, the Kel Ewey-system was based on the triangular trade combining a Ténéré-branch with the exchange of millet against dates and salt with a Damergou-branch in Southern Niger to exchange salt against millet and to obtain the provision of commercial goods necessary in the Air Mountains. In some years the salt caravan could reach up to 18.000 camels (Spittler, 1989). The extension of the French colonial government—in Niger as a military administration—provoked a violent opposition in the first decades of the 20ies century both in the Adrar des Iforas as in the Air Mountains. The colonial administration brought an end to the former sovereignties, it established a tax system, abolished slavery and the

226

Issa Ousseïni et al.

Figure 4. The beginning of the colonial period. The collapse of the traditional social and economic systems of the Tuareg Kel-n-Adrar and Kel-Air. This and the following schemes explain the economic and social situation of the Adrar des Iforas (left) and the Air Mts and Kawar region (right) and their mutual relations at different times.

razzia economy and tried to install a free migration and travel in these regions. Moreover, the international trade exchange was stepwise moved from the Kano-Tripolis route to the Gao-Algiers connection. In addition the French administration required great numbers of camels such as for the Tibesti-mission in 1913. Compensation was given only to the clan leaders. A series of battles started in the region between the Adrar des Iforas and the Menaka

The recent history of two Saharan mountains

227

region in Mali as an answer to these measures. A general upheaval under the Tuareg leader Kavcen followed 1917 in the Air Mountains, which lasted up to 1919. It finally led to the collapse of the Tuareg society. This both extended to the physical and to the psychological dimension. The Air Mountains were evacuated and people were concentrated under miserable conditions around Agadez. Moreover, the intellectual leadership was annihilated during this period too. Thus, this period still remained as a deep horror in the memory of the Kel Air and directed their behaviour for long periods (see also Adamou, 1979; Hamani, 1989; Spittler, 1989). 13.3.2 Restoration and development during the colonial regime up to the time of independence The period from the 1920ies to the beginning of the 1960ies was characterised by the reinstallation of the Tuareg population in the Air Mountains, whereas many of the Kel-nAdrar people left for Algeria or to the western forelands of the Air Mountains and further on to Libya (Figure 5). A certain part of the Air people accompanied them. They also moved to several regions of Niger such as Zinder, Maradi or Tahoua. During this decade the razzia diminished step by step and the circulation and regional economy became stable. The main risks for the people changed from the fear of razzia to the fear of drought. The climatic conditions ameliorated and the Tuareg could restore their traditional economy. It also included the caravan trade between the Kawar, Air Mountains and the South. The international trade and transport connection changed to the West with the establishment of the Gao-Adrar-Bechar-Oran or Algiers route. Even though the region was considered as remote and marginal by the colonial administration, one tried to explore its mineral or vegetal resources in order to assure a provision of the metropolitan region especially in times of war. Some drought periods in the 1930ies provoked the contradictory discussion on expanding deserts (cf. Herrmann and Hutchinson, 2005; see also below). In this period the colonial administration established an education system, a system of public health combined with a successful vaccination program from the 1950ies on, the construction of tracks and roads and finally of an infrastructure in the rural zones. After the independency both states had a different political vision (socialist in Mali and capitalist in Niger). However, their financial resources were insufficient to fulfil their own propaganda as well as peoples hope. The Tuareg preserved a deep mistrust against the colonial administration and later on to the central government and refused to send their children into school. It finally led to their exclusion of the administration, which was taken by other ethnic groups. In the early 1960ies the Kel-n-Adrar started an upheaval, which was mercilessly suppressed by the socialist Malian regime leading to a nearly complete escape of the people from the Adrar. It finally was the second wave of exodus to Libya, especially of young people. The end of the 1960ies was marked by a steady development of the traditional economies but also by first steps to a modern economy in providing energy, health system and transport facilities. 13.3.3 The period of alternating droughts, a part wise collapse of the traditional economy as well as the development of a centralised state and an exploitation of mineral resources During this period access to remote areas was possible parallel to the traditional caravan trade. The commercial exchange with Algeria developed, which was mainly based on the exchange of subsidized food from Algeria (Figure 6).

228

Issa Ousseïni et al.

Figure 5. Restoration and development during the colonial regime up to the time on independence.

The exploration and opening of the uranium mines in the North of Niger at Arlit also allowed the government to invest in infrastructure from roads to education. Taxes were no longer claimed in Niger. The industrial activities of the uranium mines attracted a great number of peoples from the South, but the Tuareg population only took part indirectly in having a better market for their animals, animal and vegetal products. The traditional system of salt caravans was concurrenced by lorries, but for a certain period it was protected by an interdiction to transport the Bilma salt by trucks. The first severe drought periods in the 1970ies and especially in 1984 brought a definite strike to the traditional societies and economies of the Tuareg population.

The recent history of two Saharan mountains

229

Figure 6. The period of alternating droughts, a part wise collapse of the traditional economy as well as the development of a centralised state and an exploitation of mineral resources.

Most of the animals died, forcing people to leave the mountains to seek help around cities or to leave the country for Algeria, Libya or Mauritania, where large refugee camps were built near the borders. Here they remained for a long period. Having lost their economic basis the societies collapsed as people realised that they could not help themselves any longer. A greater part of the young generation lost their confidence and consequently their respect to the traditional society system and they left for Libya or Algeria to look for a new physical and social base. Also many of the former servants definitely left their clans and the region too. The rising industry around the uranium exploitation at Arlit could not compensate these economic losses and social calamities. In all Saharan countries the climatic and

230

Issa Ousseïni et al.

economic situation was comparable. It was during the drought in 1984 that for the first time since the Kawcen war in 1917 the Bilma salt caravan did not take place. A great wave of international help and intervention was mobilised. It also created an image of these countries as being not able to help and solve these problems for themselves even though the mineral exploitation gave a certain economic boom. Moreover in this situation the exodus of the disillusioned young generation created a brisance for the next decades. The climatic conditions ameliorated at the end of the 1980ies, and at that time one started to talk about a greening of the Sahel (Herrmann and Hutchinson, 2005). The economic system of the Tuareg could recover for a certain part, whereas the general economic situation of Niger declined dramatically. After the Chernobyl catastrophe in 1986 the uranium production was near to a collapse. However, the international trade developed on the three routes to Algeria and Libya as well as the salt caravan to Bilma did. On this background the neighbour states such as Mauritania, Algeria and Libya claimed for a repatriation of the refugees still living in the camps near the respective borders. 13.3.4 Amelioration of the physical conditions but clash of civil war 1990 the situation changed profoundly. A well-organised system of “smuggle” trade developed between Nigeria and later on Benin and Algeria and/or Libya heading for Tamanrasset or Sebha (Gregoire, 1999). It was based on the large-scale “trade” or reexportation system of cigarettes accompanied stepwise by the trade of drugs, arms and men (Figure 7). Apart from a regional working exodus the stream of clandestine emigrants evolved with the objective to find work in Libya and or finally in Europe. In parallel, the trade of arms developed too, because the Sahara was still rich in weapons from the various civil wars in Chad and Marocco/Westsahara as well as from the Libyan tentative to build up an “Islamic Army” (cf. Bensaad, 2002; Gresh et al., 2006; Pliez, 2002). Various authorities of the respective states were more ore less involved in this multiple trade in order to assure an uninterrupted flow of goods and passengers. There are estimations, that the financial dimension of this trade passed that of the uranium trade (Gregoire, 1999). Negotiations of the governments of Mali and Niger with Algeria and Libya 1990 resulted into a massive return of refugees into their home countries. Contrary to the official promises only very little help and infrastructure was prepared for the returned people. Out of disputes and quarrels with local policemen an enormous tension developed and military interventions resulted in massacres of civilians. An upheaval and rebellion of various fractions followed both in Niger and Mali. With massive reactions and counterattacks of the Malian and Nigrian armies the situation evolved to a civil war, where both sides regularly attacked the civil population in accusing them of cooperation of the respective adversary. This war was mainly directed from outside by an effective propaganda in Europe presenting a romantic scheme of suppression and salvation (Dayak, 1992; Salifou, 1993). In 1990 in Niger and in Mali National Conferences were installed. It changed the habit of the population. People started with a democratic behaviour and did no longer accept a rough central government especially with the experience of one year’s retreat of the government from the North. This developed on the background of overall presence of weapons, and their necessary financing by the drug trade, which supported the civil war for years. In a first phase of the upheaval political illusions were active such as a far going autonomy of the Tuareg region. Mutual contacts and cooperation of different rebellion groups or liberation armies developed during the war. An important part of the rebellion fractions was based on returned young Tuaregs from Libya and Algeria, which did no longer accept a traditional order of their society and which for years had their important

The recent history of two Saharan mountains

231

Figure 7. Amelioration of the physical conditions but clash of civil war.

experience only in war and attacks. They called themselves “ishumac” derived from “chomage”—non employment (Klute et al., 1996). Attacks on roads, telecommunication installations, military camps or power plants destructed the base of the civil and military infrastructure and cut the civil population from the supply of basic necessities especially in the health sector. Finally both parts were not able to achieve a military solution. As a result some hundred of thousands of people were concentrated in and around Agadez looking for shelter and help from international organisations. Also many Tuareg left the mountains with their animals to exploit the pastures in the Southwest of the mountains.

232

Issa Ousseïni et al.

But here they came in concurrence and conflicts with other groups even though the climatic and consequently pasture conditions developed positively. 13.3.5 Loss of illusions, attrition, final treaties, recovery and hunger catastrophe All sides of the civil war in Mali and Niger were not able to obtain a military solution, and the situation degraded to a long attrition on the back of the civil population (Figure 8). After long and complicated negotiations under the mediations of the president of Burkina Faso and Algeria final treaties for Mali were signed in 1992 in Algiers. For Niger it lasted from 1995 with a pre-agreement to 1997 to have all different fractions to sign a treaty. The situation was also complicated by the military putsch of Mainassara Bare in Niger 1995 leading to a diplomatic isolation of the country. Consequently the government developed the contacts and relations to Libya also supporting the suppression of the embargo in 1996. The results of the peace treaties were different in Mali and Niger. In Mali both parts agreed in certain autonomy of the region, in an improvement of infrastructure, in an incorporation of the rebels in the national army and in an enforcement of the local administration. In the following years the Malian army quasi completely retreated from the region declaring the region as out of control of them. Thus also the international route from Algeria via Tessalit and Gao fell more or less apart. A certain part of the Kel-n-Adrar decided not to return from the refugee camps in the South and to install definitely outside of the Adrar des Iforas starting horticulture or other activities. Moreover from 1996 on in the vicinities of River Niger an irrigation project for rice cultivation was launched, which succeeded in the next decade to reach an autonomy in rice production (Mali-Nord, 2008). In Niger the main conditions of the treaties were as follows: Decentralisation should lead to a reorganisation of the state. Local and regional administration should be installed, being directly elected. Moreover, the people of the Northern regions should participate on the benefits of its resources. An integration of members of the different rebellion fractions into the army or into various administration authorities was another point. During the following years the reorganisation of the state was persecuted (cf. Ousseïni et al., 2000), although it took some years before elections on the different levels of administration could be organised. It also was necessary to reintroduce a tax system in order to have the necessary base of the administration. In the Air Mountains the Tuareg found themselves in a certain circle of confinement, because only a few working places or chances of income were available outside of the mountains. Thus, they were forced to maximise all kind of exploitations ranging from horticulture, pasture to fibre cutting in order to get raw material for basket fabrication. Combined with the transformation of the hydrological regime of several wadis, these activities transformed several parts of the mountains into a cultural landscape (see figure 9). The mineral exploitation in the mountains could hardly create any employment. The ameliorating climatic conditions however helped in this situation. Big hope was set on the re-development of tourism (see Adamou and Morel, 2005), but a remaining insecurity impeded it. The international trade definitely changed to the Agadez-Sebha route. This trade and/or smuggling was based on an old organisation system retracing the activities of the Senoussia brotherhood and some families of the Ouled Sliman installed in Niger and Chad since the 1930ies. Pliez (2000) insisted well that the exodus of Sub-Saharan people was new and could not be explained as a reactivation of a former transsaharan nomadism. The benefits however for the population were locally restricted. The Kawar oases emerged as active markets, and also the public health system was reconstructed in these places.

The recent history of two Saharan mountains

233

In 2004/2005 an unexpected hunger catastrophe hit the Sahel countries and especially Niger was heavily shocked (UN, 2005). Rainfall was not extremely low but its repartition was precarious. In some regions it started normal, but stopped after the first weeks to return only in September, too late for the growing of millet. In certain regions the harvest failed completely. Moreover, the favourable climatic conditions of the previous years provoked an increase of the locust population and huge swarms attacked the whole region. All the traditional measurements of exploiting the vegetation for additional food failed too. It resulted in an acute famine, affecting millions of people and especially children. People concentrated near the main roads seeking for rescue and shelter (Figure 8). Help from outside started very late and also the Niger administration neglected this situation too long. This also was the blame coming from international organisations.

Figure 8. The loss of illusions, attrition, final treaties and hunger catastrophe.

234

Issa Ousseïni et al.

Figure 9. The present situation in the Kori Telwa/Air Mountains The intensive garden cultivation and especially the deeply inserted walls against side erosion created a cultural landscape.

However one point was always ignored: The political philosophy of the World Bank was privatisation. The governments had to cease their economic activities as a base of credit negotiations. Thus, also the former existing system of cereal storage by the state owned office OPVN (Office des Produits du Niger) was abandoned and the warehouses now served to store cigarettes for the privately organised international trade (s.a.). Finally, people realised a slow implementation of the peace treaty, a complete failure in the Nigrian catastrophe management and their own marginal participation on the economic activities in the region even in the last years there was a positive discrimination for Tuareg people for positions in the administration. 13.3.6 Present impasse. A civil war with no visible solution As already mentioned, the North of Niger is under military law since August 2007 and both conflict parties act in a comparable martiality also employing the same vocabulary of romantic violence. The civil population is trapped and quasi taken as hostages (Figure 10). The economic and logistic isolation of the population in the Air Mountains enforces the already existing overexploitation of the natural resources as well as horticultural activities in order to assure the daily life in a situation where access to markets to sell the garden products or to buy necessary goods are nearly cut (cf. Anthelme et al., 2005). However, the situation is different from the rebellion of the 1990ies. The rising prices of primary resources of the last years induced an enlargement of exploration activities. Especially Chinese companies are active in the petroleum sector. There are plans to enlarge the uranium exploitation; a new mine is also up for discussion. However, there is a rising concern of the civil population about the present and in future enlarged health

The recent history of two Saharan mountains

235

Figure 10. The present situation in Niger. A civil war with no visible solution.

risk out of the contamination by the uranium ore dust (Ousseïni, 2008). This combined the dissatisfaction of the people over the slow process of the governments reorganisation with the feeling not to participate in the new activities in the primary sector. Also the international scene changed profoundly: The Central and Southern Sahara became the theatre of various attacks mostly interpreted as terror acts of the various groups active in the greater region. Moreover since some years the “Pansahel” and “Africom” activities of the US army are to equip and to train the armies of Algeria, Morocco, Senegal, Nigeria and Mali, Niger, Mauritania, Tunisia and Chad against any kind of terror activities (JA, 2008a). On the other side there are still plenty of weapons in the Sahara hidden since the last rebellions. The traffic of arms is active financially based in the trade of drug and persons. The conflict got enlarged and it is no longer to understand in the regional scale. It was combined to the question of primary resources and the anti-terror politics. Thus, the situation in Niger is different from Mali, where both parts could find a mediated agreement (Minist. des Maliens de l’Ext., 2006)—even still or most recently again (JA, 2008c) there is some fighting of the rebel groups with the Malian army. Contrary to that in Niger, all kind of negotiations are refused by the central government (cf. JA, 2008b). In fact the civil population got kept in an impasse. The most severe fact however is the large-scale use of antipersonnel and antitank mines, which will threaten the people for long time, even after an uncertain treaty. 13.4 CONCLUSIONS The recent amelioration of the pasture resources of Southern Sahara and Sahel could also ameliorate the economic condition of the people. However the present political conflicts limit and impede the chances to exploit them. A short screening of the historical development of

236

Issa Ousseïni et al.

the Tuareg economies evidenced a double dependency as well on the environmental factors as on the political or sociological evolution. The Kel-n-Adrar and Kel Ewey adapted in different ways to the desert environment depending on the amount of grass or tree pasture. Whereas the Kel-n-Adrar evolved certain autonomy the Kel Ewey opted for a diversified economy, which incorporated them also in the economies of the peasants of Southern Niger and Northern Nigeria. However, the Tuareg largely refused to incorporate into the colonial and later on education and administration system. The civil war in the 1990ies seems to be a result of the long time accumulated feelings of marginalisation, especially in crises and of being separated from the society of the country. In addition some hardliners in the army provoked a sudden acceleration of conflict. During the war it became evident that the Kel-n-Adrar were more able to negotiate, especially, when the government of Mali had the personalities to follow the mediated negotiation up to a peace treaty. Even in the present rebellion these circumstances became evident again (JA, 2008c). In Niger the situation is different: There is a growing importance of the international political and economical scene. Moreover, in Niger the central government does not follow the Mali example and refuses to negotiate. In all there is a triple risk of marginalisation for the Tuareg societies of Northern Mali and Northern Niger. The one lies in the historical development of their societies. The second is the growing importance of the international conflicts and economic developments, where the Tuareg of Niger may find them manipulated from these internationally acting groups. A third lies in the “ishumaq”-movement. It represents about two young generations, which are no longer integrated in the traditional Tuareg society. Their training is warfare and thus only with difficulties they are to integrate in a normal civil life. Again the two regions develop differently. The Tuareg of the Adrar des Iforas still may attain a certain autonomy in food. Their relations to the South of Mali are not as strong as those of the Kel-Air in Niger, and there are no major mineral resources in their region, which interest the central government. In addition, the “Mali-Nord” project (s.a.) provided a stable base for the refugees of the previous civil war to earn their living. In contrast to the Adrar des Iforas the uranium and petrol resources of Northern and Eastern Niger alerted both the central government of Niger and the rebel groups for its value and vulnerability. The central government opted for the military way to protect the economic chances of new mineral resources (JA, 2008 b). However, no military solution for both conflict parties is to see. The civil population suffers from the economic decline caused by the closure of the region and the general insecurity. The conflict is also embedded in the combination of an international arm and drug trade, in the activities of various fundamental and/or terrorist groups of the whole region as well as in the counteractions of the American Army as PanSahel or Africom (JA, 2008 a). Moreover, one should not forget that the present civil war in Niger got its own brisance by the large-scale use of landmines, which represent a long time danger for the civil population. There is a risk, that for Northern Niger it will go the Angola way. Important exploitation centres will remain highly protected as well as the major circulation axes. But the civil population would suffer any more from a general insecurity of the territory and of the economic decline especially concerning the once promising development of tourism (see Adamou and Morel, 2005). It will also impede the reinstallation of ambitious conservation and development projects like the Air-Ténéré National Reserve (Anthelme et al., 2005; Giazzi, 1996), which already felt apart in the first Tuareg rebellion in Niger. There is another consequence of the conflicts and especially of the employment of landmines: Scientific investigations will largely be impeded if not be impossible, especially

The recent history of two Saharan mountains

237

concerning geographic or ecologic fieldwork. Thus, probably another part of the Sahara will be closed for such investigations, as it is already the case for Northern Chad, Southern Sudan or Northern Mali. ACKNOWLEDGEMENTS We are grateful to K. Wepler for his help with the cartography. REFERENCES Adamou, A., 1979, Agadez et sa région. Etudes Nigriennes, 44, pp. 1–358. Adamou, A. and Morel, A., 2005, Niger. Agadez et les montagnes de la Air, (Grenoble. Eds. de la Boussole), pp. 1–191. Anthelme, F., De Boissieu, D. and Mato, M.-W., 2005, Conditions écologiques et socioéconomiques de la Réserve naturelle nationale de l´Air-Ténéré et ses connexes. Republique du Niger, PNUD, GEF (Niamey), pp. 1–104. Bensaad, A., 2002, La grande migration africaine à travers le Sahara. Méditerranée, 3, 4, pp. 41–52. Bernus, E., 1967, Cueillette et exploitation des ressources spontanées du Sahel nigérien par les Kel Tamasheq. Cah. ORSTOM. ser. hum., 4, pp. 31–52. Bernus, E., 1981, Touaregs nigériens. Unité culturelle et diversité régionale d´un peuple pasteur. Mém. ORSTOM, 94, pp. 1–507. Bernus, E., 1991, Les montagnes touaregues. Revue de Géographie alpine, 1, pp. 117–130. Felix-Katz, J., 1980, Analyse éco-énergetique d´un elevage nomade (Touareg) au Niger, dans la région de l`Azawak. Annales de Géographie, 491, pp. 57–72. Giazzi, F., 1996, La réserve naturelle nationale de l´Air et du Ténéré (Niger). MH/E,WWF, UICN, (Gland), pp. 1–678. Gregoire, E., 1999, Tuareg du Niger, (Paris: Karthala), pp. 1–339. Gresh, A., Radvany, J., Recacewicz, P.H., Samary, C. and Vidali, D., 2006, L´ atlas du Monde Diplomatique, (Paris), pp. 1–194. Hamani, D., 1989, Au carrefour du Soudan et de la berberie: Le sultanat touareg de l´Ayar. Etudes Nigériennes, 55, pp. 1–521. Herrmann, L., Vennemann, K., Stahr, K., von Oppen, M., 2000, Atlas of natural and agronomic resources of Niger and Benin. Dept. Geography, Univ. A. Moumouny, Niamey and Univ. Stuttgart-Hohenheim. Herrmann, S.M. and Hutchinson, C.F., 2005, The changing context of the desertification debate. Journal of Arid Environments, 63, pp. 538–555. JeuneAfrique, 2005/2008, Operation Pansahel. http://www.Jeuneafrique.com/jeuneafrique_/ article_jeune_afrique.fr Jeune Afrique, 2008a, Au secours! Les Américains debarquent, 2438, pp. 22–29. JEUNE Afrique, 2008b, Ce que fait courit Tandja, 2454, pp. 32–34. Jeune Afrique, 2008c, Prorité à la médiation, 2473, pp. 71–72. Klute, G., 1992, Die schwerste Arbeit der Welt. (München; Trickster), pp. 1–277. Klute, G., Weingärtner, L. and Weingärtner, P., 1994, Der “Tuaregkonflikt” in der Republik Niger. Rapport. (Rottenburg), pp. 1–55. Mali-Nord, 2008, http://www.mali-nord.de/home.html (05.05.2008). Ministère des Maliens de l’Exterieur et de l’Integration Africaine, 2006, Accord’Alger: Kafougouna devant les députés. http://www.maliensdelexterieur.gov.ml/cgibin/view_ article.pl?id=415

238

Issa Ousseïni et al.

Nicolaisen, J., 1963, Ecology and culture of the pastoral Tuareg. Nationalmuseets. srkifter. Ethnik. raekke. IX, (Copenhagen), pp. 1–548. Ousseïni, I., 2008, Impact environnemental du projet Imourarten. Les inquietudes des populations dans Agadez. Republicain-Niger, 15. 5. 2008, p. 2. Ousseïni, I., Maman, A.H., Lawali, D., Alkassoumi, Djibo, M. and Waziri, M., 2002, Atlas National du Niger. (Niamey; DADT), pp. 1–59. Pliez, O., 2000, Nomades d´hier, nomades d´aujourd´hui. Les migrants africains reactiventsils les territoires nomades au Sahara? Annales de Géographie, 625, pp. 688–707. Pliez, O., 2002, Vieux reseaux et nouvelles circulations entre les deux rives du Sahara. Méditerranée, 3, 4, pp. 31–40. Salifou, A., 1993, La question touarègue au Niger. (Paris; Karthala), pp. 1–207. Schulz, E. and Adamou, A., 1988, Die Vegetation des Air-Gebirges in Nord-Niger und ihre traditionelle Nutzung. Giessener Beiträge zur Entwicklungsforschung, I, 17, pp. 75–86. Schulz, E. and Adamou, A., 1997, Die Grenzen der “neolithischen Revolution”. Gab es einen frühen Ackerbau in der Sahara? Würzburger Geographische Arbeiten, 92, pp. 71–96. Schulz, E., Adamou, A. and Mahamane, K., 1994, Le Mt. Bagzane dans l’Air. Montagne refuge et seuil écologique. Revue Géographie Alpine, II, pp. 43–63. Schulz, E., Adamou, A. and Ousseïni, I., 2001, Air et Adrar des Iforas: une comparaison de deux montagnes du Sud du Sahara et leur evolution actuelle. Espaces Tropicaux, 16, pp. 319–323. Sidiyene, E.A., 1996, Des arbres et arbustres spontanés de l’Adrar des Iforas (Mali), (Paris: ORSTOM, Cirqad), pp. 1–137. Sommer, E., 2006, Kel Tamahek, (Schwülper; Cargo-Verlag), pp. 1–310. Spittler, G., 1989, Dürren, Krieg und Hungerkrisen bei den Kel Ewey (1900–1985). Studien zur Kulturkunde, 89, pp. 1–199. United Nations, 2005, Humanitarian crisis in Niger. (Paris), 1 map. Voss, F. and Krall, S., 1994a, Principaux biotopes du criquet pelerins dans le Nord du Tilemsi/Mali., (Berlin, Eschborn), 1 map. Voss, F. and Krall, S., 1994b, Principaux biotopes du criquet pelerins dans l’Adrar des Iforas (Mali) (Berlin, Eschborn), 1 map.

CHAPTER 14

The Sahelian and Saharan dune systems of Niger. A comparison of their granulometric characteristics Ibrahim Mamane Sani and Issa Ousseïni Department of Geography, University Abdou Moumouni, Niamey, Niger ABSTRACT: A comparative analysis is presented of sand dunes situated between 13° and 21°N in the eastern part of Niger. It is limited to the grain size characterization of their quartzes. The total distribution in their particle sizes confirms, that in spite of being highly prone to wind erosion, the Saharan ergs could not be the primary source of Saharan dusts because of their extremely low content of very fine fractions. Although limited to the central portions of the grain size distributions, all the parameters indicate a progressive and fairly continuous refinement of the material and its sorting from the hyper-arid zone towards the Sahel. It means, that in terms of sensitivity to wind erosion, Sahelian ergs are more prone to wind erosion than those of the Sahara without vegetation cover. This tendency of parameters, interpreted only as an indicator of sorting by wind action, is consistent with models—quite frequent in the literature—of interconnected synoptic-scale transport in the Sahara-Sahel. It stands in contradiction to some sketches of mineralogical correlations between the ergs, which give evidence unlike wind-blown dusts, that sands are largely autochthonous. However, some current reconsiderations of quartz grains sensitivity to weathering suggest a hypothesis that reconciles the discontinuities in heavy minerals (autochthony of sands) with the continuity in the refinement of the grain sizes. Research perspectives should focus on enhancing the understanding of erosion mechanisms in situ and on defining more precisely the various mineralogical provinces of sediments as well as their relationship with the bedrock.

14.1 INTRODUCTION Knowledge of the mechanisms of mobilisation of dust and sand by wind is very important for understanding the Saharan and Sahelian environments, such as soil, climate and weather as well as the frequencies of epidemics. Continuous wind transport and its suspension of aerosols are evident from the hyper-arid zone onwards. But what is about the sandy material? The answer concerns all aspects of infrastructure (roads, barrages and settlements) and also resource protection in the South. It is of special importance against the background of observed tendencies of accelerating frequency of drought and increasing demographic pressure. Intensified soil erosion characterises the present-day landscape evolution of the Sahel. Wind erosion is important for this dynamic process, especially in the systems of stabilised dunes. Unlike fluvial erosion, which is active only during the rainy season, wind erosion is important throughout the year. It is caused by the semiarid climate and the reduced plant cover due to human impact such as wood cutting or overgrazing. All this makes an evaluation of wind erosion difficult. There is always a mosaic of non-affected soil surfaces and eroded and sand-covered surfaces at different scales, depending on aerodynamic factors and the

240

Sani and Ousseïni

state of soil surfaces. During heavy runoff there is an interaction of water and wind erosion. The plurality of these factors makes the evaluation of wind erosion difficult. However, it is possible to determine the intensity of wind erosion in an indirect way by some factorial indicators such as aerodynamic parameters as well as characteristics of soil surface and sediments. Sediment characteristics are a good example for normalised scientific treatment. 14.2 SAMPLING AND LOCALISATION Sampling on dune areas was done in all five bioclimatic zones (Figure 1): the NorthSahelian Zone around Mounio (Goure) and Manga (Maine Soroa), east of Agadez near the Tigidit escarpment and in the Ténéré for the arid Sahara, as well as around Dao Timni (Djado) and Emi Fezzan (Mangueni) in the hyper-arid Sahara. With the exception of some small deposits in Mounio and Manga, sampling was arranged to cover the sites exposed to the Harmattan (Northeast to East) and to West to Northwest. Most of the samples were taken from wind-activated surfaces. In the Sahelian zone, where soil development is able to differentiate the dune bodies, some additional samples were collected in different homogeneous levels. These ergs are all situated between 10° and 14° E and 13° to 21°40' N. The form and shape of the dune bodies vary from one erg to another and also in the same erg. Table 1 gives the morphological types following the geomorphologic terms adopted by Mainguet and Callot (1978). The barchans seem to be limited to arid and hyper-arid zones. By contrast, aklés are typical for the semiarid areas. Often they are pear-shaped and rugged due to the discontinuous plant cover. Dune belts however are visible in arid zones as well as in hyperarid zones.

Figure 1. The bioclimatic zones of Niger.

The Sahelian and Saharan dune systems of Niger

241

Table 1. Localisation and morphologic characteristics of the studied ergs.

Region

Latitude

Bioclimatic zone

Manga Mounio

13°19' N

Semi-arid

14°02' N

Tiguidit Ténéré Djado Manguéni

16°56' N 18°14' N 20°56' N 21°39' N

General status of the erg

Morphologic type

Fixed erg, remobilised on surface

Aklé—mix of transversal and longitudinal dunes Barchans Longitudinal dune Barchans

Arid Hyper-arid

Active erg

Table 2. Rainfall (1951–2001, in mm) of typical stations from the different bioclimatic zones (Data: DMN, Niamey).

Stations Jan. Feb. Mar. Apr. May June

July

Mainé SoroaGoure Agadez Bilma

Aug.

Sept. Oct. Nov. Dec. Total

0,0

0,0

0,1

2,7

10,3 32,3

106,9 151,0

55,6

7,4

0,0

0,0

366,4

0,0 0,0 0,1

0,0 0,1 0,1

0,0 0,2 0,0

1,7 1,2 0,5

12,4 25,1 6,4 10,7 0,3 1,1

102,6 132,7 40,7 62,5 1,8 7,1

53,3 14,0 1,1

2,2 0,3 0,2

0,0 0,0 0,0

0,0 0,0 0,0

330,8 136,1 12,1

Rainfall in the semiarid zone (Table 2, Maine, Soroa-Goure) is about 300 mm per year. A sparse shrub savanna of the North Sahel allows a fixation of the dunes. However, clearance for agriculture and overgrazing triggered a remobilisation by the wind over large areas. Dune ridges got inverse during the rainy season showing that rainfall and plant cover were insufficient to reduce the wind’s effectiveness during monsoon times. Precipitation less than 150 mm (Agadez) is characteristic for the arid zone. It allows a contracted shrub savanna in the southern depressions. In the Ténéré aleatoric rainfalls are characteristic allowing the achab pastures. Thus more or less the whole area is exposed to wind erosion and mostly to the dominant Harmattan. On rare occasions the dune ridges may be inverted during the rainy season and the monsoon. In the hyper-arid zone (Bilma) the rainfalls are very aleaotoric, which makes the mean of 12,1 mm very indicative. For the most part, precipitation is monsoonal; however, in January–February intrusions of Mediterranean air masses may be observed too. Thus, vegetation is limited to oases, and achabs are rare also. The Harmattan is exclusively responsible for the wind activity. 14.3 RESULTS The most classic and common characteristics are granulometric characteristics, taken from relative frequencies and cumulation of the different grain size classes (Rivière, 1977). Fines are made of silt and clay smaller than 0,040 mm (AFNOR norm).

242

Sani and Ousseïni

Figure 2. Proportions of silt and clay in the dune fields along their latitude.

Dune systems contain less silt and clay along the aridity gradient, which increases with latitude. Proportionally they vary from more than 1% in the southern fixed dune systems of Mounio and Manga, to 0,3–0,5% in the Ténéré to less than 0,2% in the Mangueni region (Figure 2). However, there are some exceptions: the larger content of fines in the North corresponds to the proximity to lithological sources rich in fine grained material: lake sediments, or outcrops of highly weathered rocks. By contrast, the lower content in the South corresponds to wind-exposed sites or to reworked alluvial sands. There is a good hierarchy of modal values increasing with latitude (Table 3). In the South the modal values are at 0,125 to 0,315 mm (fine sand), at 0,630 mm (medium sand) in the Ténéré and at 1 mm (coarse sand) on the Djado and Mangueni plateaus. In arid and hyper-arid zones there is a pronounced differentiation between the windward sides with coarser modal values to lee sides with poorer values. This is less pronounced in the semiarid zone where the differentiation is only visible with middle values. However the middle values have a gradient conform to their variation. As regards the fixed ergs of the South, the means show a surface material, which is finer than in the deeper parts in Mounio but inverse in the Manga region. The present reworked eolian deposits of the Mounio (A) as well as the lower dune levels (B, BC) are also much coarser than in the Manga region. The barchans of the hyper-arid regions (Mangueni, Djado) show the most developed sorting (Figure 3, Table 3) with a clear tendency towards the coarse elements, especially on windward sides which are better sorted than the leeward sides. The barchans of the Ténéré bordering the Air Mountains show a less well-developed sorting, even the sand material is more fine. The longitudinal dunes of the Ténéré show similar characteristics to the barchans (Figure 4, Table 4) with a more pronounced differentiation towards the fine sands. The Manga samples are comparable to the windward parts of the Ténéré dune belts in their sorting near to fine sands. In the two dune fields of the Ténéré and Manga the longitudinal dunes are distinguished by their windward-leeward difference and their better sorting of the leeward sands. Concerning the reworked parts of the aklé zones in the South (Figure 4, Table 4) the Mounio dune field is characterised by its preponderance of coarse sand. In the Manga region, however, the sorting is far more developed (as for example in the Ténéré) and

The Sahelian and Saharan dune systems of Niger

243

Figure 3. Examples of cumulated frequency curves of surface sand on barchans (av = windward, sv = leeward).

tends to fine sands and coarse silt. This sorting is better developed on the actual surface sediments and the deep BC-horizon than in the pedoplasmatic (AB, B) horizons. 14.4 DISCUSSIONS AND PERSPECTIVES These preliminary results disproof the hypotheses which reject the idea of the dune fields from the Sahara on as a primary dust source (Herrmann et al., 1997) because they are

Table 3. Granulometric modes and means.

Ergs and dominant dune types

Modes (mm)

Means (mm) (P84 + P16)/2

windward

leeward

windward

leeward

Barchans Manguéni Djado Ténéré

1,000 1,000 0,250 to 0,630

0,500 0,315

0,904 0,653 0,325

0,325 0,240 0,168

Longitudinal dunes and aklé Ténéré Mounio/reworked N1 N2 Manga/dune belt Manga/reworked N1 Manga N2 Manga N3

0,630 0,315 0,315 0,125 to 0,250 0,125 to 0,250 0,125 to 0,250 0,125 to 0,250

0,250

0,584 0,228 0,332 0,163 0,163 0,153 0,151

0,204

0,139

244

Sani and Ousseïni Table 4. Sorting of different materials.

Ergs and dominant dune types

Sorting index (√P75/P25)

Standard deviation (P84 − P16)/2

windward

leeward

windward

leeward

Barchans Manguéni Djado Ténéré

1,250 1,234 1,565

1,261 1,465 1,506

0,249 0,313 0,225

0,155 0,151 0,083

Longitudinal dunes and aklé Ténéré Manga/dune belt Mounio/reworked N1 N2 Manga/reworked N1 Manga N2 Manga N3

1,439 1,448 1,445 1,4 1,319 1,358 1,343

1,366 1,303

0,193 0,072 0,107 0,161 0,028 0,059 0,056

0,091 0,049

deprived by their fines. However, the dust sources should be searched in the ancient lake depression and areas of heavily weathered and denudated rocks in the Sahara as well as in the Sahel. The rising of granulometric parameters from Djado and Mangueni down to Manga normally is interpreted as being used up during transport suggesting a continuous sand transport between the Sahara and its borders. The schemes of sediment budgets partly deduces from meteorological observations (Mainguet, 1977, 1996) but also from field observation in the connection areas of some dune fields (Mainguet and Callot, 1978; Stengel, 1992) will reject this hypothesis.

Figure 4. Examples of cumulated frequency curves of surface sand on longitudinal dunes (av = windward, sv = leeward).

The Sahelian and Saharan dune systems of Niger

245

Figure 5. Examples of cumulated frequency curves of surface sand on aklés of Manga and Mounio (A = reworked surface material, B = in situ soil material (horizon B), BC groundwater influence).

Some correlations between the ergs (Pfeiffer, 1992) and local geological substratum (Ousseïni, 1986) are pointing to an autochthony of the sand masses. This argumentation is coherent to the results presented above if one rejects the interpretation on a continuous granulometric evolution but stresses a heritage in the context of a progressive and differentiated alteration, more dependent on hydroysis in the South and limited by the desaggregation of grains in the North. The recent reflections on the proved sensibility of the siliceous elements to alteration (Pope, 1995; McFarlane et al., 2007) will support this hypothesis. To get a better knowledge of the dynamic of eolian mobilisation of sand it will be necessary to have middle and long term observations at the place. In addition an identification of minerals of the sedimentary provinces in a finer scale particularly of the fine and better mobilisable sand masses and their lithological affiliations is needed. ACKNOWLEDGEMENTS We are indebted to the Programme de Lutte contre l’Ensablement (Mainé Soroa), the Programme de Recherche sur l’Ensablement des Cuvettes (Gouré) and the Institute of Geography, University of Würzburg, for help with sampling and data processing. Thanks go to A. Beck, Berlin, for the translation into English language. REFERENCES Herrmann, L., Stahr, K. and Sponholz, B., 1997, Identifizierung trockenzeitlicher und regenzeitlicher Staubquellen im östlichen Westafrika. Würzburger Geographische Arbeiten, 92, pp. 189–211. Mainguet, M., 1977, Analyse quantitative de l’extrémité sahélienne du courant éolien transporteur de sable au Sahara nigérien. C. R. Aca. Sc., 285, pp. 1029–1032.

246

Sani and Ousseïni

Mainguet, M., 1996, The Saharo-Sahelian global wind action system: one facet of wind erosion analysed at a synoptic scale. In: Wind erosion in West Africa: the problem and its control. Proceedings of International Seminar, (Stuttgart-Hohenheim: Margraf Verlag), pp. 7–22. Mainguet, M. and Callot, Y., 1974, Air photo study of typology and interrelations between the texture and structure of dune patterns in the Fachi-Bilma Erg, Sahara. Zeitschrift für Geomorphologie N.F., Suppl.-Bd., 20, pp. 62–68. Mainguet, M. and Callot, Y., 1978, L’Erg de Fachi-Bilma (Tchad-Niger) contribution à la connaissance de la dynamique des Ergs et des dunes des zones arides chaudes. CNRS Mémoire et documents. Nouvelle série, 18, (Paris), pp. 1–184. McFarlane, M.J., Coetzee, S.H., Kuhn, J.R., Vanderpost, C.H.M. and Eckardt, F.D., 2007, In situ rounding of quartz grains within an African surface weathering profile in NorthWest Ngamiland, Botswana. Zeitschrift für Geomorphologie N.F., 51, 3, pp. 269–286. Ousseïni, I., 1986, Etude de la répartition des formations quaternaires et interprétation des dépôts éoliens dans le Liptako Oriental (Rép. du Niger). Thèse 3e cycle, Paris 6, (Paris), pp. 1–233. Ozer, P., 2002, Les lithométéores au Niger: mise au point. Würzburger Geographische Arbeiten, 97, pp. 7–32. Pfeiffer, L., 1992, Heavy mineral analysis of dune sands of Africa and Central Asia (South Sahara, Sahel, Namib, Taklamakan). Würzburger Geographische Arbeiten, 84, pp. 347–359. Pope, G.A., 1995, Newly submicron-scale weathering in quartz: geographical implications. Professional Geographer, 47, 4, pp. 375–389. Rivière, A., 1977, Méthodes granulométriques: techniques et interpretations (Paris, New York, Barcelone, Milan, Masson), pp. 1–167. Stengel, I., 1992, Zur äolischen Morphodynamik von Dünen und Sandoberflächen. Würzburger Geographische Arbeiten, 83, pp. 1–363. Stengel, I., 1992, Morphodynamic types of longitudinal dunes in the Ténéré desert and the Erg of Bilma (Republic of Niger). Würzburger Geographische Arbeiten, 84, pp. 147–168.

Regional/Location Index

Abalak 95 Achelouma valley 138–140, 142 Adamaoua Mountains 198, 211 Adrar des Iforas 73, 221–223, 225, 226, 232, 236 Agadem 3, 74, 75 Agadez 75, 95, 222, 223, 225, 227, 231, 232, 240, 241 Ahaggar 7, 71, 73, 222, 223 Ain Galaka 107–109, 113, 115, 123 Air Mountains 221–223, 225, 227, 232, 234, 242 Aleg 76, 77 Algeria 4, 27, 28, 70, 149, 150, 179, 187, 227, 229, 230, 232, 235 Angamma escarpment 107–110, 113, 115, 118–120, 122–124 Bahr el Ghazal 107, 108, 114, 115 Benue 159 Bilma 1–10, 32, 40, 74, 75, 128, 130–132, 135, 149, 150, 228, 230, 241 Bodélé Depression 13, 201, 202, 216 Cameroon 68, 159, 198 Central Sahara 1–7, 9–13, 23, 24, 27, 28, 30–32, 39, 40, 64, 65, 70, 92, 102, 127, 129, 132, 135, 137, 138, 152, 179, 181, 186, 198, 222 Chad 1–5, 7, 12, 13, 30–33, 40, 72, 73, 75, 81, 82, 91, 92, 95, 96, 98–100, 102, 107–109, 112–114, 116, 117, 120, 122, 124, 138, 159, 199, 209, 230, 232, 235, 237 Chad Basin 1–5, 7, 12, 13, 40 Cherchera 190, 193 Djado 1, 3–5, 7, 10, 40, 60, 138, 139, 179, 240–244 Djanet 149, 150

Emi Fezzan 240 Erg de Fachi 1, 7 Erg de Ténéré 1 Erg el Makhzen 187–189, 191–194 Faro River 159 Faya-Largeau 107–109 Ghana 161 Grand Erg de Bilma 1, 7, 128, 130, 131 Gulf of Guinea 197 Gulf of Sidra 198 Hassi el Abiod 161, 162, 172–174 Kel Ewey 225, 236 Kel-n-Adrar 221, 222, 224–227, 232, 236 Kidal 222, 223 Kiffa 95, 98 Kirachaou 187 Kobadi 96, 99, 160–162, 167–175 Kolima 161, 162, 172–174 Lake Chad 2, 3, 5, 13, 32, 33, 40, 81, 92, 107, 108, 117, 122, 199, 209 Maghreb 26, 27, 30, 31, 150 Mali 4, 26, 27, 73, 75, 160, 161, 166, 171, 173, 174, 221–223, 227, 230, 232, 235–237 Manga 12, 92, 240–245 Mangueni 1, 4, 6, 7, 10, 80, 137–139, 142, 143, 240, 242, 244 Mankhor 179, 180, 185, 186 Massif d’Agadem 3 Massif du Termit 28 Mauritania 26, 63, 75, 76, 79, 83, 84, 91, 92, 95, 98, 101, 102, 229, 230, 235 Mediterranean Sea 197, 198 Mounio 240–245

248

Regional/Location Index

Moussoro 96, 99, 109 Murzuq Basin 1, 3–7, 14 Niamey 1, 39, 63, 95, 96, 98, 171–173, 176, 221, 223, 239, 241 Niger 3–5, 7–9, 12, 13, 27, 28, 30, 31, 39, 40, 63, 68, 71–75, 80, 81, 84, 91–93, 95–99, 101, 102, 112, 114, 117, 124, 127–132, 137–139, 149, 150, 159–161, 170–176, 179, 185, 210, 221–223, 225, 227, 228, 230, 232–236, 239, 240 Niger basin 159 Niger Inland Delta 160, 161, 171, 173–175 Niger River 31, 159, 172 Nouakchott 10, 29, 76–78, 92, 102 Ntereso 161, 162 Oceanic Sahara 63 Ouémé 159 Plateau de Mangueni 137–139, 142, 143 Plateau du Djado 3, 138, 139 Plateau du Tchigai 7 Sahara-Sahel boundary 1 Sahel 1, 9–13, 27–31, 40, 60, 70, 71, 73, 76, 77, 80–83, 91–93, 95–103, 107, 108, 111, 112, 114, 115, 117, 123, 152, 161, 185, 197–200, 204, 205, 209, 211–214, 216, 221, 230, 233, 235, 236, 239–241, 244

Sahelian Zone 27, 240 Seeterrassental 137, 139–145, 147, 150–153 Seggedim 7, 10, 12, 39, 40, 42, 45, 47, 48, 52, 54, 56, 59, 60, 127, 129 Sudan 10, 12, 23, 27, 28, 30, 31, 40, 68, 71–73, 75, 76, 80, 82, 83, 92, 102, 117, 197, 199, 200, 204, 205, 211, 214, 216, 237 Sudanian-Guinean-Zone 27, 31 Tadrart 179 Tahoua 95, 227 Tamanrasset 149, 150, 230 Tassili n’Ajjer 7, 179 Tiabel Goudodie 161 Tibesti Mountains 4, 6, 13, 14, 28, 32, 73, 138 Tigidit 74, 75, 240 Tunisia 63, 68, 70, 83, 187, 198, 203, 205, 209, 211, 214, 215, 235 Volta River 161, 171 Wadi Cherchera 190 Western Sahara 10, 11, 26, 28, 30, 31, 39, 64, 70, 79, 152

Subject Index

Acacia 40, 50, 52, 68, 71, 73–78, 80, 82, 83, 96, 138, 224 achab 63, 70–73, 75, 78, 82, 84, 108, 224, 241 Aerva 71 African Humid Period 28, 30, 152 A-horizon 13, 121 aklé 240, 242–245 aleatoric rainfall 71, 72, 241 ancient dune 8–13, 91–93, 95–97, 102, 103 anterior river basin 193 aquifer 32, 56, 60 archaeofauna 159, 160, 173, 174 Arenosol 95–101, 111, 112, 114, 123 arid period 28, 92, 97 aridisation 24, 29, 30, 115, 137, 152, 197 Ariidae 159, 161, 163, 164, 166, 173, 176 Aristida 71, 75 Aristida mutabilis 75 Arius gigas 159–176 Artemisia 68–70, 73 Arthrophytum 70 Auchenoglanis 161, 174 Auchenoglanis sp. 161, 163 AVHRR 197, 201, 214

Cassia 71 cation exchange capacity 121 cattle 179–181, 185, 225 cemetery 179 charcoal 42, 48–50, 52, 59, 60, 182, 190 Chenchrus biflorus 75 civil war 3, 80, 221, 222, 230–232, 234–236 Clarias sp. 161 Clarotes 163, 174 clay 7, 10–12, 33, 42–44, 47–49, 52, 56, 70, 94, 95, 97, 108, 111–113, 115, 117–119, 121–123, 147, 188, 189, 193, 241, 242 Climate Prediction Center 201 coarse sand 40, 71, 72, 108, 242 Community Climate Model 30 Continental Terminal 4, 5, 95, 96, 138 continentality 3, 13, 27 corings 39 Costa approach 142, 148, 149, 151, 153 Cretaceous 3, 4, 5, 75, 138, 153 cuesta 32, 40, 52, 74, 75, 84, 95, 127, 129 cultural landscape 76, 232, 234 Cupressus 73

Bagrus 162–164, 174 Balanites 71, 75–77 barchan 108, 110, 117, 240, 242, 243 basin 3, 5, 7, 31–33, 39, 52, 59, 60, 137–139, 150, 153, 159, 180, 181, 187–189, 191–193, 214 bedrock 7, 39, 44, 45, 57, 58, 145, 153, 239 B-horizon 13, 110, 113 Bos africanus 185 Bos primigenius 185 Bos taurus 184, 185

deflation 8, 9, 31, 84, 93, 103, 107, 110, 111, 117, 123 depression 1, 3, 5, 7–9, 10, 12, 32, 39, 40, 47, 48, 52, 54, 56, 60, 64, 70, 71, 73–75, 77, 108, 115, 117, 124, 129, 135, 241, 244 Desert Climate 198 desert vegetation 73, 222 desertification 73, 93, 95, 96, 103, 197 Desmostachya bipennata 40 detritic sedimentation 193 diatom 6, 7, 13, 33, 48–50, 52, 54, 56, 100, 108, 114, 115 diatomite 7, 100, 108 dip slope 142 discharge volumes 141 drainage density 141, 145

Cambrian 138 Capparis 71, 74, 75, 83 caravan trade 225, 227 Carboniferious 138

250

Subject Index

drainage system 142, 193 drug trade 230, 236 Dry and Thorn Savanna Climate 199, 205 dune fields 12, 64, 91, 108, 110, 124, 127, 138, 242–244 dune sand 8, 10, 12, 13, 40, 41, 68, 70, 95, 96, 97, 108, 115, 117, 127–131, 135 dune systems 239, 242 dunes 1, 8–13, 28, 77, 78, 91–93, 95–97, 100, 102, 103, 108, 111, 112, 114, 117, 124, 142, 239, 241, 242, 243, 244 ecological stress 175 economy 12, 179, 221, 222, 224–227, 229, 236 elongation ratio 141, 145 endemicity 160 Ephedra 70, 73, 79 ephemeral 123, 138, 139, 148, 150, 199 erg 1, 4, 5, 7, 10, 14, 24, 30, 32, 33, 40, 41, 47, 65, 71, 74, 75, 77, 79, 80, 83, 91, 93, 95, 103, 107, 108, 132, 138, 153, 166, 185, 193, 198, 203, 215, 225, 227, 232, 239–242, 245 Erg el Makhzen 187–189, 191–194 Fagonia 75 fan 139, 142–146, 148, 153 fault 139 Fe-accumulation 110 Fe-compounds 115 Fe-isotopes 124 ferri-crust 188, 189, 193, 209 Fe-sandstones 113 fine sand 71, 72, 111, 193, 242, 243 fines 6, 28, 73, 91, 93, 96–98, 100, 102, 111, 117, 242, 244 fishing gear 174, 175 fixed dune system 242 fluvial erosion 96, 239 fluvial geomorphology 137 fluvio-lacustrian deposits 193 fragmentation 23, 27, 48, 128, 129, 132, 133 fulgurites 107, 127–130, 134, 135 Giant Catfish 159–161, 170–173, 175 gley 8, 13, 93, 110, 115, 117, 118, 124 Global Precipitation Climatology Project 201 GPCP 197, 201, 202, 204, 206–208, 213, 215

graben 3, 138, 142 granulometric characteristics 239, 241 granulometry 119 grassland 12, 64, 68, 81, 200, 205, 211, 212 groundwater 4, 5, 13, 25, 32, 33, 40, 52, 54, 56, 57, 60, 64, 65, 70, 71, 73, 80, 82, 83, 84, 122, 123, 135, 199, 245 Gymnarchus niloticus 174 Harmattan 111, 240, 241 headwater 31, 32, 142, 150, 153 Holocene 1, 6–13, 23, 24, 27–32, 39, 40, 59, 65, 68, 80, 91, 92, 95–97, 100, 102, 103, 107, 108, 112, 130, 135, 137, 142, 152, 153, 159, 160, 161, 162, 166, 172–175, 179, 187, 192, 193, 197 horticulture 80, 221, 232 human impact 12, 80, 81, 239 humid period 27, 30, 31, 81, 82, 91–93, 97, 102, 103, 197 Hydrocynus brevis 174 Hyphaene 41, 74 ichthyofauna 161 Imperata cylindrica 40 Inductively Coupled Plasma Optical Emission Spectrometry 41 infrastructure 227, 228, 230, 231, 232, 239 inselberg 5, 79, 91, 95 Intertropical Convergence Zone 24, 138, 198 iron oxide sedimentation 50 Juncus 41 karstification 5 Kirachaou 187 lacustrine deposit 7, 187, 189, 190, 193 lacustrine optimum 187, 189, 191 lacustrine sediment 23, 60, 108–110, 111, 114, 115, 117, 119, 124, 189, 190 lake sedimentation 52 lake stages 39, 45, 59 lake-sediments 6, 7, 10, 13, 33, 39, 47, 100, 107, 242 landmines 83, 221, 222, 236 Landsat TM 110, 113, 214 landslide 5, 6, 142, 147, 152, 153 Last Glacial Maximum 28, 30 Lates niloticus 161, 163, 173, 174

Subject Index

lechatelierite 127–131, 133, 134 Leptadenia 71, 75–77, 80 lightning 60, 127–130, 132–135 longitudinal dune 1, 92, 96, 108, 241, 242, 244 Macchia 68 Maerua 50, 71, 73–76, 78, 79, 83 Manning-Strickler approach 141, 149, 151, 153 MAR 198–200, 204, 206, 213 mass movements 142 Mean Annual Rainfall 198, 206 Mediterranean Climate Mesozoic 3, 127, 129, 130 Messak Sandstone 4, 5 meteorological station 65, 138, 149, 150 micromorphology 187, 192 microprobe 189, 191, 192 microstratigraphy 187 Middle Terrace 6, 7 Mid-Holocene 10, 11–13, 30–32, 39, 40, 91, 135, 197 Milanchovitch variations 197 millet 12, 224, 225, 233 mineral exploitation 230, 232 monsoon 3, 10, 11, 24, 27–30, 40, 60, 63, 65, 83, 135, 138, 150, 152, 198, 241 morphodynamics 12–14, 33 NDVI data 197, 201–205, 213–215, 216 Neolithic 7, 9, 14, 29, 40, 80, 92, 160, 161, 179, 180, 186 Nile perch 161, 173, 174 nomadic 12, 185, 222, 225 nomadism 80, 232 oak forest 68 Oasis 7, 13, 29, 54, 74, 80, 92, 107, 108, 113, 115, 129, 130, 225 Olea 68, 70, 73 Organic matter content 121 OSL dating 13, 91, 92, 94, 96, 103, 117 OSL-ages 102 osteomorphological features 160 palaeochannel 151, 152 palaeodischarge 152 palaeodune 8–10, 13, 192 Palaeohydrology 137 palaeolake 7–9, 27, 28, 32, 48, 187 palaeolake sediments 7

251

palaeopedological 6, 10, 11, 107, 114 palaeoprecipitation 30, 31 palaeosols 92, 102, 103 Palaeozoic 3, 127, 129, 130, 138 Panicum 40, 52, 71, 73–75, 77, 78, 83, 138, 224 pastoralism 80, 179 pasture 64, 72, 75, 80, 221, 222, 224, 231, 232, 235, 236, 241 peak discharge 137, 138, 141, 148–153 pebble 109, 124, 128, 132, 140, 142, 147, 148 pedogenesis, 108 pedogenic oxides 97, 100 Pennisetum millet 12 Phoenix 41, 74 pH-values 42, 45, 55, 56, 91, 94, 100, 103, 113–115 Piliostigma 76 pisoliths 189 Pleistocene 4–11, 13, 14, 39, 59, 91, 92, 95–97, 100, 102, 103, 115, 124, 130, 179, 187, 193 Pliocene 5, 115 potsherd 180–182, 184 potsherd accumulations 180 precipitation 3, 6, 8, 9, 23, 24, 28, 29, 30–33, 40, 52, 59, 60, 64, 65, 70, 71, 83, 84, 91, 95, 100, 103, 132, 138, 145, 149, 150, 152, 153, 193, 198, 201, 202, 204, 206, 215, 241 pseudotachylyte 128 quartz 5, 44, 47, 50, 52, 56, 94, 109, 127–133, 188, 189, 239 Quaternary 4, 6, 9, 12, 13, 23, 39, 92, 138, 142, 197 Quercus 68, 69 radiocarbon dating 42, 57, 59, 60, 92, 152, 179 rainfall variability 198, 215 razzia 225–227 rebellion 80, 132, 221, 222, 230, 232, 234–236 Redness Ratio 93, 97 regression analysis 141 regressive erosion 142 remote sensing 107, 108, 110, 113, 124, 142, 198 rhizoconcretions 8, 9, 13, 110 ridge 8, 75, 95, 96, 138, 164, 241

252

Subject Index

riffle frequency 143, 144 ritual inhumation 181 rock fulgurites 127, 129, 130, 135 rock paintings 179 roughness parameter 142 runoff 9, 32, 40, 52, 64, 70, 83, 137, 144, 145, 148, 150–153, 199, 240 salt caravan 225, 228, 230 Salvadora 71, 74, 75, 76 sand fulgurites 127–130, 135 sandstone 5, 39, 40, 42, 44, 48, 56, 74, 75, 83, 95, 108–113, 115, 127, 129, 130, 133, 135, 138, 153, 179, 187–189, 192, 193 sapropel savanna vegetation 12, 40, 59, 79 scarpfoot depression 5 sebkha 10, 39–42, 47, 48, 50, 52, 54, 59, 60, 129, 187, 193 sebkha environment 39, 40, 47, 50 sebkha sedimentation 42, 47, 54 semidesert 63, 64, 68–71, 73, 77–79, 83, 222 serir 40, 108, 109, 111, 113, 118, 124 serir-surfaces 108, 109, 113, 123, 124 silt 40, 48–50, 52, 97, 108, 110, 111, 115, 147, 189, 241–243 skeletal anatomy 159, 160, 164 slack water sediments 6, 137, 142, 144, 152 soil development 13, 91–93, 102, 103, 240 soil fertility 103 soil stability 93, 103 Solenostemma 71 soluble salts 91, 92, 97, 100–103 Sporobolus spicatus 40 steppe 64, 68–70, 81–83, 198, 205 Stipa 68–71, 74, 75, 77, 78

Stipagrostis 68, 70, 71, 74, 75, 77, 78 stone accumulations 180–182, 184 stone artefacts 96, 181, 182 stream sinuosity 141 stromatoliths 193 Swamp-iron crusts 8 Tamarix 41, 71, 76, 77 Tamarix canariensis 41 tectonic movements 5 Terrace 6, 132–134, 137, 140, 146–148, 151, 152, 153 Tertiary 3, 4, 5, 6, 95, 96, 108, 138 thin section 39, 41, 59, 128–130, 134, 189, 190, 192 Tuareg 221, 222, 225–228, 230–232, 234, 236 Tunisia 63, 68, 70, 83, 187, 198, 203, 205, 209, 211, 214, 215, 235 uranium exploitation 229, 234 vegetation dynamics 197, 198, 203, 205, 212, 213, 215, 216 vegetation map 63, 65, 68, 71, 84 vegetation trends 197, 198, 209, 214–216 wadi 52, 54, 64, 70, 71, 73–75, 78, 92, 95, 96, 132, 133, 180, 187, 193, 199, 224, 232 weather station data 197 Wet Savanna Climate 205 wind erosion 107, 123, 124, 239–241 Woodland 64, 81, 205, 211, 212 XRD diffractogramme 122 XRF analyses 44 yardang 7, 107–110, 124