the archaeology of drylands: living at the margin (one world archaeology)
TRANSCRIPT
THE ARCHAEOLOGY OF DRYLANDS
The One World Archaeology (OWA) series stems from conferences organized by theWorld Archaeological Congress (WAC) an international non-profit making organization which provides a forum of debate for anyone who is genuinely interested in or has aconcern for the past All editors and contributors to the OWA series waive any fees theymight normally receive from a publisher Instead all royalties from the series are receivedby the World Archaeological Congress Charitable Company to help the wider work ofthe World Archaeological Congress The sale of OWA volumes provides the means forless advantaged colleagues to attend World Archaeological Congress conferencesthereby enabling them to contribute to the development of the academic debatesurrounding the study of the past
The World Archaeological Congress would like to take this opportunity to thank all editors and contributors for helping the development of world archaeology in this way
ONE WORLD ARCHAEOLOGY Series Editor (Volumes 1ndash37) Peter JUcko
Academic Series Editors (Volume 38 onwards) Martin Hall and Julian Thomas Executive Series Editor (Volume 38 onwards) Peter Stone
1 What is an Animal TIngold (ed)
2 The Walking Larder Patterns of domestication pastoralism and predation JClutton-Brock
3 Domination and Resistance DMiller MJRowlands and CTilley (eds)
4 State and Society The emergence and development of social hierarchy and political centralization JGledhill BBender and MTLarsen (eds)
5 Who Needs the Past Indigenous values and archaeology RLayton (ed)
6 The Meaning of Things Material culture and symbolic expression IHodder (ed)
7 Animals into Art HMorphy (ed)
8 Conflict in the Archaeology of Living Traditions RLayton (ed)
9 Archaeological Heritage Management in the Modern World HFCleere (ed)
10 Archaeological Approaches to Cultural Identity SJShennan (ed)
11 Centre and Periphery Comparative studies in archaeology TCChampion (ed)
12 The Politics of the Past PGathercole and DLowenthal (eds)
13 Foraging and Farming The evolution of plant exploitation DRHarris and GCHillman (eds)
14 Whatrsquos New A closer look at the process of innovation SE van der Leeuw and RTorrence (eds)
15 Hunters of the Recent Past LBDavis and BOKReeves (eds)
16 Signifying Animals Human meaning in the natural world RGWillis (ed)
17 The Excluded Past Archaeology in education PGStone and RMacKenzie (eds)
18 From the Baltic to the Black Sea Studies in medieval archaeology DAustin and LAlcock (eds)
19 The Origins of Human Behaviour RAFoley (ed)
20 The Archaeology of Africa Food metals and towns TShaw PSinclair BAndah and AOkpoko (eds)
21 Archaeology and the Information Age A global perspective PReilly and SRahtz (eds)
22 Tropical Archaeobotany Applications and developments JGHather (ed)
23 Sacred Sites Sacred Places DL Carmichael JHubert BReeves and ASchanche (eds)
24 Social Construction of the Past Representation as power GCBond and AGilliam (eds)
25 The Presented Past Heritage museums and education PGStone and BLMolyneaux (eds)
26 Time Process and Structural Transformation in Archaeology SEvan der Leeuw and JMcGlade (eds)
27 Archaeology and Language I Theoretical and methodological orientations RBlench and MSpriggs (eds)
28 Early Human Behaviour in the Global Context MPetraglia and RKorisettar (eds)
29 Archaeology and Language II Archaeological data and linguistic hypotheses RBlench and MSpriggs (eds)
30 Archaeology and Anthropology of Landscape Shaping your landscape PJUcko and RLayton (eds)
31 The Prehistory of Food Appetites for Change CGosden and JGHather (eds)
32 Historical Archaeology Back from the edge PPAFunari MHall and SJones (eds)
33 Cultural Resource Management in Contemporary Society Perspectives on managing and presenting the past FP MacManamon and AHatton (eds)
34 Archaeology and Language III Artefacts languages and texts RBlench and M Spriggs (eds)
35 Archaeology and Language IV Language change and cultural transformation RBlench and MSpriggs (eds)
36 The Constructed Past Experimental archaeology education and the public PGStone and PPlanel (eds)
37 Time and Archaeology TMurray (ed)
38 The Archaeology of Difference Negotiating cross-cultural engagements in Oceania RTorrence and AClarke (eds)
39 The Archaeology of Drylands Living at the margin GBarker and DGilbertson (eds)
THE ARCHAEOLOGY OF DRYLANDS
Living at the margin
Edited by
Graeme Barker and David Gilbertson
London and New York
First published 2000 by Routledge 11 New Fetter Lane London EC4P 4EE
Simultaneously published in the USA and Canada by Routledge 29 West 35th Street New York NY 10001
Routledge is an imprint of the Taylor amp Francis Group
This edition published in the Taylor amp Francis e-Library 2005
ldquoTo purchase your own copy of this or any of Taylor amp Francis or Routledges collection of thousands of eBooks please go to
wwweBookstoretandfcoukrdquo
copy 2000 Selection and editorial matter Graeme Barker and David Gilbertson individual chapters the contributors
All rights reserved No part of this book may be reprinted or reproduced or utilized in any form or by any electronic
mechanical or other means now known or hereafter invented including photocopying and recording or in any information
storage or retrieval system without permission in writing from the publishers
British Library Cataloguing in Publication Data A catalogue record for this book is available from the
British Library
Library of Congress Cataloging in Publication Data The archaeology of drylands living at the margin[edited by]
Graeme Barker and David Gilbertson p cm (One world archaeology 39)
Includes bibliographical references and index 1 Social archaeology 2 Landscape archaeology
3 Human ecology 4 DesertsmdashHistory 5 Land settlement mdashHistory 6 Land settlement patterns PrehistoricmdashHistory
7 Arid regions agriculturemdashSocial aspectsmdashHistory 8 Climatic changesmdashHistory I Barker Graeme
II Gilbertson DD III Series CC724A735 2000
9301ndashdc21 00ndash038257
ISBN 0-203-16573-X Master e-book ISBN
ISBN 0-203-26029-5 (Adobe e-Reader Format) ISBN 0-415-23001-2 (Print Edition)
Contents
List of figures ix List of tables xiii List of contributors xv Series editorsrsquo foreword xxv Preface xxvii
Part I Introduction
1 Living at the margin themes in the archaeology of drylands Graeme Barker and David Gilbertson 3
2 The dynamic climatology of drylands Greg Spellman 18
Part II Southwest and Central Asia
3
The decline of desert agriculture a view from the classical period Negev Steven ARosen
44
4
Farmers herders and miners in the Wadi Faynan southern Jordan a 10000-year landscape archaeology Graeme Barker
62
5
Differing strategies for water supply and farming in the Syrian Black Desert Paul Newson
85
6
Irrigation agriculture in Central Asia a long-term perspective from Turkmenistan Mark Nesbitt and Sarah OrsquoHara
101
Part III Sahara and Sahel
7
Conquests and land degradation in the eastern Maghreb during classical antiquity and the Middle Ages Jean-Louis Ballais
121
8
Success longevity and failure of arid-land agriculture Romano-Libyan floodwater farming in the Tripolitanian pre-desert David Gilbertson Chris Hunt and Gavin Gillmore
133
9 Twelve thousand years of human adaptation in Fezzan (Libyan Sahara) David Mattingly 156
10 Farming and famine subsistence strategies in Highland Ethiopia Ann Butler and ACatherine DrsquoAndrea 174
Part IV Eastern and southern Africa
11 Engaruka farming by irrigation in Maasailand cAD 1400ndash1700 John EGSutton 195
12 The agricultural landscape of the Nyanga area of Zimbabwe Robert Soper 214
13
Fifteenth-century agropastoral responses to a disequilibrial ecosystem in southeastern Botswana John Kinahan
227
14
Islands of intensive agriculture in African drylands towards an explanatory framework Mats Widgren
246
Part V North and Central America
15
Prehistoric agriculture and anthropogenic ecology of the North American Southwest Paul EMinnis
264
16
The role of maguey in the Mesoamerican tierra friacutea ethnographic historic and archaeological perspectives Jeffrey RParsons and JAndrew Darling
280
Part VI Europe
17 Traditional irrigation systems in dryland Switzerland Anne Jones and Darren Crook 307
18
Desertification land degradation and land abandonment in the Rhocircne valley France Sander van der Leeuw
327
Index 346
Figures
11 The world map of drylands 312 A Roman-period fortified farm northwest Libya 413 The location of the case studies in this volume 514 Drowning in drylandsmdashtwo vehicles sunk in a flash-flood 721 Thermal regimes in two dryland locations Aswan Egypt and Jacobabad
Pakistan 2222 Mean monthly relative humidity at four locations 2423 The rainshadow effect leading to aridity 2524 The Hadley Cell circulation of the tropical northern hemisphere 2625 The structure of the trade wind atmosphere 2826 The interaction between the subtropical westerly flow and the tropical
easterlies leading to the creation of Saharan depressions 3027 The monthly progression of the East African Low-Level Jet Core 3228 The tracks of Sudano-Saharan depressions over the Sahara 3431 Terraced dam system in the central Negev 4532 The wine press at Shivta (Subeita) 4633 Sketch of the Byzantine town of Shivta (Esbeita or Subeita) 4734 Map of the general settlement system of the central Negev during the
Late Byzantine and Early Islamic periods 4835 View of the Byzantine town of Avdat (looking north) 4936 Elaborate raised field and dam system on Nahal Lavan 5037 The early Islamic village of Sede Boqer in the central Negev 5241 The location of Wadi Faynan within its region 6342 Looking northeast across part of the ancient field system to Khirbet
Faynan 6443 The survey area of the Wadi Faynan Landscape Survey 6544 Ethnoarchaeological survey the typical site of a winter bedouin tent in
Wadi Faynan 6745 The settlement locations of the first farmers in the Wadi Faynan 7046 Part of the Wadi Faynan field system WF4 showing the early bronze age
and the classical landscapes 7247 A field system on the northern side of the Wadi Faynan 7348 A field map of part of the field system WF4 7649 The distribution of copper through sediments accumulated behind the
Khirbet Faynan barrage 77410 Section through a Roman-period water conduit channel 79
51 The Hauran and the Harra regions of Syria
8652 Plan of the Roman-period settlement of Ad-Diyatheh and its water
channels and field systems 8853 The Roman-period settlement of Ad-Diyatheh and its channel walls 8954 Plan of the main irrigated zone along the Wadi Sham by the Roman-
period settlement of al-Namara 9255 View of the main irrigated zone along the Wadi Sham at al-Namara 9356 Canal 3 at al-Namara viewed from the east 9357 The ancient reservoir at Qasr Burqursquo 9558 Air photograph of Qasr Burqursquo and its reservoir 9561 Turkmenistan showing locations mentioned in Chapter 6 10362 Bronze and iron age settlement in the Merv oasis Turkmenistan 10771 The eastern Maghreb showing locations mentioned in Chapter 7 12272 Flood deposits at Ksar Rhilane (Tunisia) 12473 The modern Roman aqueduct crossing Wadi Bou Jbib Carthage 12674 Holocene terrace of Wadi Cheacuteria-Mezeraa (Algeria) 12775 Holocene terraces in the Wadi el Akarit (Tunisia) 12881 Tripolitania northwest Libya showing the principal landforms and
settlements and the location of the UNESCO Libyan Valleys Survey 13482 RomanomdashLibyan floodwater farming in the Wadi Gobbeen 13483 Simplified distribution of early RomanomdashLibyan farms 13684 A RomanomdashLibyan fortified farm (gasr) and its satellite buildings at
Ghirza 14485 Model of RomanomdashLibyan agriculture 14686 Walls in the desert wall systems in the Wadi Mimoun 14791 Map showing the location of the Fezzan and the area of most detailed
survey around Germa 15792 The major climatic fluctuations of the Holocene in the Libyan Sahara 15993 The settlement of Germa (ancient Garama) the capital of the Garamantes 16194 Schematic cross-section of a foggara 16395 Model of the neolithic landscape around Germa 16496 Model of the evolved Garamantian landscape around Germa 16797 Model of the medieval landscape around Germa 16898 Model of hypothetical future direction of settlement and farming in
Fezzan 169101 Map of Ethiopia showing Adi Ainawalid in Tigrai province 175102 Residential compound near fields Adi Ainawalid 176103 Intercropped bread and durum wheats near Mai Kayeh Tigrai 179104 Harvesting grasspea by hand uprooting Adi Ainawalid 181105 First threshing of teff Adi Ainawalid 182106 Winnowing teff Adi Ainawalid 183
107 Grain storage jars Adi Ainawalid 184111 The Rift Valley and Crater Highlands of northern Tanzania 196112 Engaruka and the Rift Valley escarpment 197113 The Engaruka escarpment from the east 198114 Engaruka south fields 200115 Grid of feeder furrows and levelled field plots 200116 The support for the lsquogreat northernrsquo canal 201117 The embanked causeway of the lsquogreat northernrsquo canal 202118 An angular cairn 203119 Sonjo wooden house with thatch dome 212121 Location of the Nyanga area Zimbabwe 215122 Terraced hillsides in the Nyanga lowlands 216123 Vertical aerial photograph of cultivation ridges 219131 The regional setting of Letsibogo southeastern Botswana 229132 The distribution and linkage of Khami period sites at Letsibogo 233133 Plan and section of Letsibogo Site 125 234134 Distribution of soil nutrient values at Letsibogo Site 125 236141 Eastern and southern Africa showing sites mentioned in Chapter 14 248142 An intensively cultivated landscape at Kwermusl 251143 Preparing the field at Kwermusl 252144 Piles of manure from stalled cattle Mama Issara 252145 An irrigation channel above Tot in Marakwet Kenya 253146 An irrigation canal under repair above Chesoi Marakwet 254151 The North American Southwest 265152 Prehistoric Hohokam communities and irrigation systems in the Phoenix
basin of the Salt river 267153 Aerial photograph of prehistoric trincheras (checkdam) fields near Casas
Grandes Chihuahua Mexico 268154 Gridded gardens of fields near Safford Arizona 269155 A rock mulch field near Casas Grandes Chihuahua Mexico 270156 An excavated rock pile from the field shown in Figure 155 270161 Middle America showing the approximate extent of the tierra friacutea 281162 Field of cultivated magueys 282163 Castrating a mature maguey plant 286164 Spinning maguey fibre 290165 Pre-columbian spindle whorls used for spinning maguey fibre 291166 Use of modern iron scraper for extracting maguey fibre 293167 Examples of pre-columbian trapezoidal tabular basalt scrapers 294168 Experimental use of pre-columbian trapezoidal tabular basalt scraper 295169 A pre-columbian scraper plane 295
1610 Modern iron scraper and pre-columbian obsidian scrapers 296171 The Valais canton Switzerland showing places mentioned in Chapter 17 308172 Distribution of agricultural land in Vernamiegravege during the 1960s 309173 The Grand Bisse de Lens 313174 The irrigation sectors in Vernamiegravege 317175 Distribution of water during the first tour from the bisses of Vernamiegravege
5th May-8th June 1964 318176 A tessel used by members of the consortage of the Grand Bisse de Lens 320181 The middle and lower Rhocircne valley showing the progress of Roman
colonization 329182 Settlement trends in the middle and lower Rhocircne valley 50 BC-AD 600 331183 The persistence of settlements in the middle and lower Rhocircne valley
through different occupation periods 333184 GIS maps of the Haut Comtat 335185 The three levels of the investigation into modern-day urbanmdashrural
dynamics in southern France 338186 Relations between cities individual communes and their contexts 339187 Differences in context occurring among towns of similar andor different
sizes in the Haut Comtat 340
Tables
21 The regional distribution of world drylands 1922 Rainfall regimes at selected dryland stations 1923 Estimates of the land area of arid lands 2024 Mean relative humidity at various isobaric levels in the Sahara and the
Arabian peninsula 2325 Seven ITCZ zones 2926 The extent and severity of desertification 3561 Simplified chronological chart of prehistoric settlement in Turkmenistan 11262 Simplified chronological chart of settlement in the Merv oasis in the
historic period 11371 Morphoclimatic evolution in the eastern Maghreb during the later
Holocene 12281 Farm products of the Tripolitanian pre-desert first to fifteenth centuries
AD 141101 Crops cultivated at Adi Ainawalid Tigrai Ethiopia 178102 Crops no longer cultivated at Adi Ainawalid Tigrai Ethiopia 186131 Selected radiocarbon measurements from Letsibogo 231 132 Faunal taxa from Letsibogo Site 125 237161 The prehispanic chronology of central Mexico 283171 Approximate numbers of named irrigators using the Grand Bisse de Lens
lsquoaqueductis communirsquo in 1457 314172 Examples of tours with the number of droits and sequence of irrigation
hours 316173 Examples of bisse disputes 321181 Evolution of the lsquonature-culturersquo debate over the last thirty years 341182 The different approaches of the historical and natural sciences to the
reconstruction of the past 342183 The opposition between analytical and integrative approaches in research 342
Contributors
JEAN-LOUIS BALLAIS is Professor of Physical Geography at the Universiteacute de Provence Aix-en-Provence (France) His principal research interests focus onHolocene Mediterranean erosion (he co-directed the study of erosion and geosystemshistory in classical antiquity and the Middle Ages for the EU-funded Archaeomedesproject described in Chapter 18) and present-day erosion desertification and landdegradation in the south of France and in the Maghreb Relevant recent publicationsinclude lsquoAeolian activity desertification and the ldquoGreen Damrdquo in the Ziban range Algeriarsquo in ACMillington and KPye (eds) Environmental Change in DrylandsBiogeographical and Geomorphological Perspectives 177ndash98 Chichester John Wiley and Sons 1994 lsquoThe south of France and Corsicarsquo in AJConacher and MSala (eds) Land Degradation in Mediterranean Environments of the World 29ndash39 Chichester John Wiley and Sons 1998 and (with J-CMeffre) Le Plan de Dieu (Nord-Vaucluse) Geacuteoarcheacuteologie et Histoire drsquoun Paysage Anthropiseacute Etudes Vauclusiennes 15 1996 (Institutional address Institut de Geographic lsquoUniversiteacute de Provence (Aix-Marseiile I) 29 Avenue Robert Schuman 13621 Aix-en-Provence France)
GRAEME BARKER (BA PhD University of Cambridge) taught prehistoric archaeologyat the University of Sheffield (1972ndash84) and was then Director of the British School at Rome (1984ndash88) before taking up his appointment as Professor of Archaeology at theUniversity of Leicester (UK) where he is currently Dean of the University GraduateSchool His principal research interests have been in the archaeology of subsistenceand agriculture with a special focus first on archaeozoology but later in landscapearchaeology He has conducted fieldwork in Italy Mozambique and the formerYugoslavia and has directed inter-disciplinary field projects in Italy Libya andcurrently in Jordan His publications include Landscape and Society PrehistoricCentral Italy London Academic Press 1981 (with RHodges) Archaeology and Italian Society Oxford British Archaeological Reports 1981 Prehistoric Communities in Northern England Sheffield University of Sheffield 1981 Prehistoric Farming in Europe Cambridge Cambridge University Press 1985 (with CSGamble) Beyond Domestication in Prehistoric Europe London Academic Press 1985 (with JLloyd) Roman Landscapes Archaeological Survey in the Mediterranean Region London British School at Rome 1991 A Mediterranean Valley Landscape Archaeology andAnnales History in the Biferno Valley London Leicester University Press 1985 (twovolumes) (with DGilbertson BJones and D Mattingly) Farming the Desert The UNESCO Libyan Valleys Archaeological Survey Volume One Synthesis VolumeTwo Gazetteer and Pottery Paris UNESCO 1996 (with TRasmussen) The Etruscans Oxford Blackwells 1998 The Companion Encyclopedia of ArchaeologyLondon Routledge 1999 and (General editor with DMattingly) Mediterranean Landscape Archaeology Oxford Oxbow 2000 (five volumes) He was elected a
Fellow of the British Academy in 1999 (Institutional address the Graduate SchoolUniversity of Leicester Leicester LE1 7RH UK)
ANN BUTLER is an Honorary Lecturer at the Institute of Archaeology UniversityCollege London with a BSc degree in Botany (London University) an MA inArchaeology (Manchester University) and a PhD in Archaeology (London University)Her research interests centre on legumes as a human resource in the temperate OldWorld and include ancient diet and nutrition plant domestication and crop dispersalstraditional agriculture and sustainable farming Her fieldwork has been conducted inEurope Southwest Asia and Highland Ethiopia Her current research focuses on theevidence for the domestication and exploitation of legume crops Her recentpublications include lsquoPulse agronomy traditional systems and implications for early cultivationrsquo in PCAnderson (ed) Preacutehistoire de lrsquoAgriculture 67ndash78 Paris CNRS 1998 and lsquoTraditional seed cropping systems in the temperate Old World models forantiquityrsquo in CGosden and JHather (eds) The Prehistory of Food 473ndash77 London Routledge 1999 (Institutional address Institute of Archaeology University CollegeLondon 31ndash34 Gordon Square London WC1H OPY UK)
A CATHERINE DrsquoANDREA is an Associate Professor of Archaeology at Simon Fraser University British Columbia Canada She completed a BSc in Anthropology at theUniversity of Toronto an MSc in Bioarchaeology at University College London and aPhD in Anthropology at the University of Toronto Her research interests includepalaeoethnobotany ethnoarchaeology and early agrarian societies in Africa and the FarEast She is currently conducting ethnoarchaeological and palaeoethnobotanicalresearch in northern Ethiopia as well as collaborating on an excavation in northernGhana Her recent publications include lsquoThe dispersal of domesticated plants into northeastern Japanrsquo in CGosden and JHather (eds) The Prehistory of Food 163ndash83 London Routledge 1999 and (with DELyons Mitiku Haile and EA Butler)lsquoEthnoarchaeological approaches to the study of prehistoric agriculture in the Ethiopian Highlandsrsquo in Mvan der Veen (ed) The Exploitation of Plant Resources inAncient Africa 101ndash22 New York Plenum Publishing Corporation 1999 (Institutional address Department of Archaeology Simon Fraser University BurnabyBritish Columbia Canada V5A 1S6)
DARREN CROOK is currently a Postdoctoral Fellow in the Department of GeographyUniversity of Liverpool working on historical impacts of land use and climate onhydrology in a pre-alpine landscape funded by the Leverhulme Foundation Hegraduated with a BSc in Human Ecology from the University of Huddersfield HisPhD also from the University of Huddersfield dealt with the sustainability of the bissemountain irrigation system in the Valais Switzerland Publications from this include(with AM Jones) lsquoTraditional irrigation and its importance to the tourist landscape of Valais Switzerlandrsquo Landscape Research 24 (19) 199949ndash65 and (with AM Jones) lsquoTraditional water management in a developed world context an example from the Valais Switzerlandrsquo Mountain Research and Development 19 (2) 199979ndash99 (Institutional address Department of Geography University of Liverpool RoxbyBuilding Liverpool L69 3BX UK)
JANDREW DARLING completed his PhD in Anthropology at the University ofMichigan in 1998 and is currently an archaeologist for the Gila River Indian
Community and Board Member for the Mexico-North Research Network Cd Chihuahua Mexico He has conducted fieldwork in Zacatecas Mexico (1988-present) on the north coast of Peru (1989ndash90) in southeastern Hungary (1987) and the NorthAmerican Southwest (1984ndash1987) His research interests include compositionalstudies exchange regional interaction and ritual in prehistoric and ethnographiccomplex societies including parallel archival investigations on the development ofAmerican archaeology in Mexico during the early twentieth century Significantpublications include lsquoAnasazi mass inhumation and the execution of witches in theAmerican Southwestrsquo American Anthropologist 100 19931ndash21 (with MGlascock) lsquoAcquisition and distribution of obsidian in the north-central frontier of Mesoamericarsquo in ECRattray (ed) Rutas de Intercambio en Mesoamerica III Coloquio Pedro Posch Gimpera 345ndash64 Mexico DF Universidad Nacional Autonoma de Mexico 1998lsquoTrace element analysis of the Huitzila and La Lobera obsidian sources in the southern Sierra Madre Occidential Mexicorsquo Journal of Radioanalytical and Nuclear ChemistryArticles 196 (2) 1995243ndash52 and lsquoNotes on obsidian sources of the southern Sierra Madre Occidentalrsquo Ancient Mesoamerica 4 1993245ndash53 (Institutional address Mexico-North Research Network 16 de Septiembre 402 Cd de Chihuahua Chihuahua Mexico CP 31020)
DAVID GILBERTSON is Head of the School of Conservative Sciences at the Universityof Bournemouth (UK) and Distinguished Visiting Scholar in the Department ofGeography and Environmental Studies at the University of Adelaide (Australia) Hegraduated in Environmental Sciences at the University of Lancaster and gained hisPhD and DSc in Quaternary and Archaeological Geology from the University ofBristol Previous appointments have included a Senior Fulbright Scholar at theUniversity of Arizona the Directorship of the MSc Programme in EnvironmentalArchaeology and then Head of the Research School in Archaeology andArchaeological Science at the University of Sheffield where he became Reader thenProfessor Director of the Institute of Geography and Earth Science at the University ofWales Aberystwyth and Professorial Research Fellow University CollegeNorthampton In addition to dryland environments past and present his researchinterests include coastal geomorphology environmental change and caves andenvironmental archaeology in general His principal publications include (withRDSJenkinson) In the Shadow of Extinction The Quaternary Geology andPalaeoecology of the Lake Fissures and Smaller Caves at Creswell Crags Sheffield University of Sheffield Monographs in Prehistory 1984 (with DJBriggs andGRCoope) The Chronology and Environmental Framework of Early Man in the Upper Thames A New Model Oxford British Archaeological Reports 1985 Run-Off Farming in Rural Arid Lands Applied Geography Theme Volume 6 (1) 1986 (withGBarker BJones and DMattingly) Farming the Desert The UNESCO LibyanValleys Archaeological Survey Volume One Synthesis Volume Two Gazetteer andPottery Paris UNESCO 1996 and (with MKent and JPGrattan) The Outer Hebrides The Last 14000 Years Sheffield Sheffield Academic Press 1996(Institutional addresses School of Conservation Sciences University of BournemouthBournemouth BH12 5BB UK and Department of Geography and EnvironmentalStudies University of Adelaide South Australia 5005)
GAVIN GILLMORE is Senior Lecturer in Earth Science at University CollegeNorthampton His research interests include the application of fossil studies topalaeoenvironmental analysis ecotoxicology of cave environments and microfossiland stratigraphic studies of Quaternary sedimentary basins He has worked extensivelyfor oil exploration companies producing many consultancy reports on JurassicCretaceous and Tertiary-Quaternary microfossil assemblages Some current papers include (with MSperrin PPhillips and ADenman) lsquoRadon hazards geology and exposure of cave users a case study and some theoretical perspectivesrsquo Ecotoxicology and Environmental Safety 2000 (with KASmith and SSinclair) lsquoPalaeoenvironmental and biostratigraphical significance of Ostracoda from the Milton Formation (Quaternary) Northamptonshire UKrsquo Proceedings of the Geologists Association 2000 and (with TKjennerud and RKyrkjeboslash) lsquoThe reconstruction and analysis of palaeowater depths a new approach and test of micropalaeontologicalapproaches in the post-rift (Cretaceous to Quaternary) interval of the Northern NorthSearsquo in OJMartinsen and TDreyer (eds) Sedimentary Environments OffshoreNorwaymdashPalaeozoic to Recent Oslo Norwegian Petroleum Society Special Publication 2000 (Institutional address School of Environmental Science UniversityCollege Northampton Northampton NN2 7AL UK)
CHRIS HUNT is currently a Senior Lecturer in the Division of Geographical Sciences atthe University of Huddersfield His academic education led him from a degree inGeogaphyGeology and an MSc in Micropalaeontology (Palynology) at the Universityof Sheffield to a PhD at Aberystwyth (Wales) on the Pleistocene history of an area inSomerset He has published extensively in the area of environmental archaeology inBritain Europe North Africa and the Middle East Recent publications include (withGBarker DDGilbertson and DMattingly) lsquoRomano-Libyan agriculture integrated modelsrsquo in GBarker DGilbertson BJones and DMattingly Farming the Desert The UNESCO Libyan Valleys Survey Volume One Synthesis 265ndash90 Paris UNESCO 1996 (with SCampbell JScourse and DHKeen) The Quaternary of South West England Chichester Chapman amp Hall Geological Conservation Review Series14 1998 and (with DDGilbertson) lsquoContext and impacts of ancient catchment management in Mediterranean countries implications for sustainable resource usersquo in DWheater and CKirby (eds) Hydrology in a Changing Environment Volume II 473ndash83 Chichester John Wiley and Sons 1998 (Institutional address Division ofGeographical Sciences University of Huddersfield Queensgate Huddersfield HD13DH UK)
ANNE JONES is currently a Senior Lecturer in the Division of Geographical Sciences atthe University of Huddersfield and Head of Division She graduated with a BSc inGeography from Queen Mary College (University of London) and then obtained herDPhil thesis at the University of Oxford studying immigrant communities in MarseilleFrance Subsequently she has held posts at the Open University Liverpool Universityand Cambridgeshire College of Arts and Technology (now Anglia PolytechnicUniversity) Her research interests focus around the inter-relationship between demography and the allocation of scarce resources in marginal environments in alpineEurope the Mediterranean and Africa Her principal publications include lsquoExploiting a marginal European environment population control and resource management under
the Ancien Regimersquo Journal of Family History 16 (4) 1991363ndash79 (with COHunt) lsquoWalls wells and water supply aspects of the cultural landscape of Gozo Maltese Islandsrsquo Landscape Issues 11 (1) 199424ndash9 and (with COHunt and DSCrook) lsquoTraditional irrigation strategies and their implications for sustainable livelihoods in semi-arid areas examples from Switzerland and the Maltese islandsrsquo in HWheater and C Kirby (eds) Hydrology in a Changing Environment Volume II 485ndash94 Chichester John Wiley and Sons 1998 (Institutional address Division ofGeographical Sciences University of Huddersfield Queensgate Huddersfield HD13DH UK)
JOHN KINAHAN (PhD 1989 Witwatersrand) is an independent consultant inarchaeological and palaeoenvironmental studies and has authored more than fortyscientific papers in diverse fields Recent publications include Pastoral Nomads of the Central Namib Desert The People History Forgot Windhoek New Namibia Books 1991 lsquoThe rise and fall of nomadic pastoralism in the central Namib desertrsquo in TShaw PSinclair BAndah and AOkpoko (eds) The Archaeology of Africa Food Metals and Towns 372ndash85 Routledge One World Archaeology 20 1993 and lsquoA new archaeological perspective on nomadic pastoralist expansion in south-western Africarsquo in JEGSutton (ed) The Growth of Farming Communities in Africa from the Equator Southwards 211ndash26 Nairobi British Institute in Eastern Africa 1996 He is currently attached to the University of Uppsala in Sweden while carrying out research on thelong-term environmental impacts of nomadic pastoralism in Namibia and Tanzania (Institutional address Quaternary Research Services PO Box 22407 WindhoekNamibia and Department of Archaeology and Ancient History St Erikstorg 5University of Uppsala Uppsala S75310 Sweden)
SANDER VAN DER LEEUW presently Professor in the History and Archaeology ofTechniques at the Sorbonne in Paris followed a university education in History andArchaeology at the University of Amsterdam (PhD 1976) and has taught at theUniversities of Leiden and Amsterdam in the Netherlands Reading and Cambridge inthe UK Michigan (Ann Arbor) Massachussetts (Amherst) and the Santa Fe Institute inthe USA and the Australian National University He has undertaken fieldwork inSyria Holland the Philippines and France His main research interests embrace thetechnology of ancient pottery making regional and spatial archaeology the relationsbetween people and their environment through time as well as in the present and theuse of dynamical systems modelling for understanding social systems Among hispublications are Studies in the Technology of Ancient Pottery Amsterdam University of Amsterdam Institute for Pre- and Protohistory 1976 (with ACPritchard) The Many Dimensions of Pottery Amsterdam University of Amsterdam Institute for Pre-and Protohistory 1982 (with RTorrence) Whatrsquos New A Closer Look at the Process of Innovation London Unwin Hyman One World Archaeology 14 1987 (withJMcGlade) Time Process and Structured Transformation in Archaeology London Routledge 1997 and The ARCHAEOMEDES Project Understanding the Natural andAnthropogenic Causes of Land Degradation and Desertification in the MediterraneanLuxemburg Publications Office of the European Union 1998 (Institutional addressBoit 05 Maison de lrsquoArcheacuteologie et de lrsquoEtnologie 21 Alice de lrsquoUniversiteacute 92023 Nanterre France)
DAVID MATTINGLY is Professor of Roman Archaeology in the School ofArchaeological Studies at the University of Leicester and has conducted fieldwork inNorth Africa the Near East the Mediterranean and Britain He took his BA and PhD atthe University of Manchester the latter a study of Roman Libya Following the tenureof a British Academy Post-Doctoral Fellowship at the University of Oxfordresearching Roman-period olive oil production he taught at the University of Michigan (Ann Arbor) before joining Leicester in 1992 He has published extensively on landscape archaeology especially of arid zones the Roman empire and its impacton people and environment and olive oil production and trade in the ancient world Inaddition to his current field project in Fezzan (Libya) which he discusses in his chapter(Chapter 9) he is also collaborating with Graeme Barker and David Gilbertson in the Wadi Faynan Landscape Survey Jordan (Chapter 4) His principal publicationsinclude (with JALloyd) Libya Research in Archaeology Environment History andSociety 1969ndash1989 London Society for Libyan Studies 1989 (with GDBJones) An Atlas of Roman Britain Oxford Blackwell 1993 Tripolitania London Batsford 1995 (with GBarker DGilbertson and BJones) Farming the Desert The UNESCO Libyan Valleys Archaeological Survey Volume One Synthesis Volume Two Gazetteer andPottery Paris UNESCO 1996 Dialogues in Roman Imperialism Power Discourseand Discrepant Experience in the Roman Empire Portsmouth RI 1997 and (with MGillings and Jvan Dalen) Mediterranean Landscape Archaeology 3 Geographical Information Systems and Landscape Archaeology Oxford Oxbow 2000 (Institutional address School of Archaeological Studies University of Leicester Leicester LE17RH UK)
PAUL EMINNIS is an Associate Professor at the University of Oklahoma His researchinterests include paleoethnobotany human ecology social evolution human responsesto food shortages the relationships between archaeology and biodiversity and theprehistory of the North American Southwest For the past decade he has co-directed an archaeological project in northwestern Chihuahua Mexico to understand the regionalsetting of Casas Grandes one of the most complex prehistoric polities in NorthAmerica His books include Social Adaptation to Food Stress Chicago University of Chicago Press 1985 (with CRedman) Perspectives on Southwestern PrehistoryBoulder CO Westview Press 1990 (with WElisens) Biodiversity and Native AmericaNorman University of Oklahoma Press 2000 and Ethnobotany A Reader Norman University of Oklahoma Press 2000 (Institutional address Department ofAnthropology University of Oklahoma Norman OK 73019 USA)
MARK NESBITT is an ethnobotanist at the Centre for Economic Botany Royal BotanicGardens Kew An undergraduate degree in agricultural botany at Reading Universitywas followed by postgraduate training in archaeobotany at the Institute ofArchaeology University College London Since 1985 he has been involved in a widerange of archaeological fieldwork in the Near and Middle East including Turkey IraqBahrain and Turkmenistan His research interests include the origins development andsustainability of crop husbandry in arid lands the evolution of Old World cereals andarchaeological and ethnographic evidence for wild plant foods in the temperate zonesPublications include lsquoArchaeobotanical evidence for early Dilmun diet at Saar Bahrainrsquo Arabian Archaeology and Epigraphy 4 199320ndash47 lsquoPlants and people in
ancient Anatoliarsquo Biblical Archaeologist 58 199568ndash81 (with DSamuel) lsquoFrom staple crop to extinction The archaeology and history of the hulled wheatsrsquo in SPadulosi KHammer and JHeller (eds) Hulled Wheats 41ndash100 Rome IPGRI 1997 (Institutional address Centre for Economic Botany Royal Botanic Gardens KewRichmond Surrey TW9 3AE UK)
PAUL NEWSON is a PhD student at the University of Leicester After a first career as agraphic designer he took his BA and MA in archaeology at University CollegeLondon Funded by the Natural Environment Research Council he is preparing hisPhD thesis at Leicester on water management strategies in Roman Arabia and theirimplications for understanding processes of Romanization combining field studies ofdryland water-management systems of the kind discussed in his chapter (Chapter 5) with a detailed analysis of Roman-period field systems in the Wadi Faynan in southernJordan using Geographical Information Systems In 1999 he was Acting AssistantDirector of the Council for British Research in the Levantrsquos British Institute at Amman for Archaeology and History (Institutional address School of Archaeological StudiesUniversity of Leicester Leicester LE1 7RH UK)
SARAH OrsquoHARA is a Reader in Environment and Society in the School of GeographyUniversity of Nottingham She completed a BSc in Physical Geography and Geologyat the University of Liverpool an MSc in Geography at the University of AlbertaCanada and a DPhil in Geography at the University of Oxford Her research interestsinclude environmental reconstruction humanenvironmental interactions and waterresource management in the worldrsquos drylands Her research has been carried out in Albania Canada Iran Mexico Turkmenistan and Uzbekistan More recently she hasbegun collaborating with Professor Julian Henderson on the Ancient Raqqa IndustrialProject in Syria Recent publications include (with DThomas MD Bateman andDMershahi) lsquoDevelopment age and environmental significance of a Late Quaternarysand ramp central Iranrsquo Quaternary Research 48 1997155ndash61 (with THannan) lsquoIrrigation and water management in Turkmenistan past systems present problems and future scenariosrsquo Europe-Asia Studies 51 199921ndash41 and lsquoLearning from the past water management in Central Asiarsquo Water Policy (forthcoming 2000) (Institutional address School of Geography University of Nottingham NottinghamNG7 2RD UK)
JEFFREY RPARSONS (PhD in Anthropology University of Michigan 1966) iscurrently Professor of Anthropology and Curator of Latin American Archaeology atthe University of Michigan Ann Arbor Michigan USA In addition to ongoingfieldwork in the Valley of Mexico since 1961 he has also worked in Guatemala (TikalPeten 1966) Peru (Chilca central coast 1969ndash1970 and Junin central highlands 1975ndash76) Iceland (Eyaforur 1985) and northwest Argentina (Jujuy 1995) His research interests include the development of pre-industrial complex society settlement pattern studies archaeological ethnography and (in the Andes) long-term relationships between herders and agriculturalists Current plans include fieldwork onthe pre-hispanic utilization of lacustrine resources in the Valley of Mexico and pre-hispanic regional organization in the Peruvian central highlands Significantpublications include (with CMHastings and Ramiro Matos M) lsquoRebuilding the state in highland Peru herder-cultivator interaction during the Late Intermediate Period in
the Tarama-Chinchaycocha regionrsquo Latin American Antiquity 8 1997317ndash41 (with GMastache RSantley and MC Serra) Arqueologiacutea Mesoamericana Homenaje a William TSanders Mexico DF Institute Nacional de Antropologiacutea e Histoacuteria 1996 lsquoPolitical implications of pre-hispanic chinampa agriculture in the Valley of Mexicorsquo in HHarvey (ed) Land and Politics in the Valley of Mexico A Two Thousand-Year Perspective 17ndash42 Albuquerque University of New Mexico Press 1991 and (withMH Parsons) Maguey Utilization in Highland Central Mexico An ArchaeologicalEthnography Ann Arbor University of Michigan Museum of AnthropologyAnthropological Paper No 82 1990 (Institutional address Department ofAnthropology University of Michigan Ann Arbor Michigan MI 48104 USA)
STEVE AROSEN teaches archaeology in the Department of Bible and Ancient NearEast at Ben-Gurion University in Beer-Sheva Israel He received his Bachelorrsquos degree from the University of California at Berkeley in mathematics and anthropologyand his graduate degrees in anthropology from the University of Chicago Prior to hiscurrent position he worked for eight years as a survey archaeologist for theArchaeology Survey of Israel in the Negev His research interests include desertadaptations the archaeology of pastoral nomadism Levantine prehistory and lithicanalysis His major publications include Lithics After the Stone Age Walnut Creek Altamira Press 1997 (with GAvni) The lsquoOded Sites Investigations of Two EarlyIslamic Pastoral Camps South of the Ramon Crater Beer-Sheva Ben-Gurion University Press 1997 and Archaeological Survey of Israel Map of Bersquoerot OdedBeer-Sheva Ben-Gurion University Press 1994 (Institutional address ArchaeologyDivision Ben-Gurion University PO Box 3653 Beer-Sheva 84105 Israel)
ROBERT SOPER (MA Cambridge UK) is a Senior Research Fellow in Archaeology atthe University of Zimbabwe Between 1962 and 1985 he worked for the NigerianAntiquities Department the British Institute in Eastern Africa Ibadan University andthe University of Nairobi His principal research interests have included the laterprehistory of East Africa (especially the Early Iron Age and later ceramics) the site ofOyo Ile in Nigeria and Great Zimbabwe tradition sites in northern ZimbabweSignificant publications include lsquoA general review of the Early Iron Age in thesouthern half of Africarsquo Azania 6 19725ndash37 lsquoRoulette decoration on African potteryrsquo African Archaeological Review 3 198529ndash51 lsquoThe palace at Oyo Ile western Nigeriarsquo West African Journal of Archaeology 22 1993 and (with BEKipkorir and JWSsennyonga) The Kerio Valley Past Present and FutureNairobi University of Nairobi Institute of African Studies 1983 He has conductedresearch on the Nyanga terrace complex since 1993 and a monograph on this isforthcoming (Institutional address History Department University of Zimbabwe POBox MP 167 Mount Pleasant Harare Zimbabwe)
GREG SPELLMAN is a lecturer in the School of Environmental Science at UniversityCollege Northampton (UK) After a BA in Geography at the University of Sheffieldhe took a PGDip in Applied Meteorology and Climatology at the University ofBirmingham and an MA in Professional Education at the University of Leicester Hisresearch interests include synoptic climatology extreme hydrological events in theIberian peninsula and the meteorology of air pollution He is currently researching onthe synoptic climatology of Spain particularly the evaluation of various downscaling
methods in order to improve regional climate change scenarios Recent publicationsinclude lsquoAn application of artificial neural networks to the prediction of surface ozone concentrations in the United Kingdomrsquo Applied Geography 19 1999123ndash36 and lsquoInvestigating the synoptic climatology of precipitation in Mallorca Spainrsquo Journal of Meteorology 23 1998117ndash30 (Institutional address School of Environmental Science University College Northampton Northampton NN2 7AL UK)
JOHN EGSUTTON (MA Oxford PhD East Africa FSA) was Director of the BritishInstitute in Eastern Africa from 1983 to 1998 and was previously Professor ofArchaeology at the University of Ghana He has been mainly concerned with laterarchaeology and its contribution to African history with a special interest in fieldsystems and agricultural technology His first visit to Engaruka in the northernTanzanian Rift Valleymdashthe subject of his contribution to this volumemdashwas in 1963 while a research scholar of the British Institute Later as a lecturer at the University ofDar es Salaam he continued the investigation of the Engaruka irrigation-agricultural settlement with student assistants That study has been extended in more recent yearswithin a broader comparative project on African field systems and cultivationstrategies Engaruka and other prominent sites are described in detail in Archaeological Sites of East Africa Four Studies (special volume 33 of Azania for 1998) and there is a shorter illustrated account in A Thousand Years of East Africa (Nairobi British Institute of East Africa 1990) Another special volume of Azania (2930 1995) The Growth of Farming Communities in Africa from the Equator Southwards surveyed the present state of research on the Early Iron Age and the Bantu agricultural expansion(Address 118 Southmoor Road Oxford OX2 6RB UK)
MATS WIDGREN teaches at Stockholm University where he is a Professor in HumanGeography He received his PhD in Stockholm in 1983 and has researched on thehistorical geography of agricultural landscapes from the Iron Age to the present in Sweden and in southern and eastern Africa Among his publications are lsquoIs landscape history possiblersquo in PUcko and RLayton (eds) The Archaeology and Anthropology of Landscape 94ndash103 London Routledge 1999 lsquoFields and field systems in Scandinavia during the Middle Agesrsquo in GAstill and JLangdon (eds) Medieval Farming and TechnologymdashThe Impact of Agricultural Change in Northwest Europe173ndash92 Leiden Brill 1997 lsquoStrip fields in an iron age context a case study from Vaumlstergoumltland Swedenrsquo Landscape History 12 19905ndash24 and Settlement and Farming Systems in the Early Iron Age A Study of Fossil Agrarian Landscapes inOumlstergoumltland Sweden Stockholm Almquist amp Wiksell 1983 His most importantworks in Swedish are his contribution to the first volume of the agrarian history ofSweden Jordbrukets foumlrsta femtusen aringr (1998) and a book on medieval field systems inBohuslaumln Sweden (1995) (Institutional address Department of Human GeographyStockholm University S-106 91 Stockholm Sweden)
Foreword
One World Archaeology is dedicated to exploring new themes theories and applicationsin archaeology from around the world The series of edited volumes began withcontributions that were either part of the inaugural meeting of the World ArchaeologicalCongress in Southampton UK in 1986 or were commissioned specifically immediatelyfollowing the meetingmdashfrequently from participants who were inspired to make their own contributions Since then the World Archaeological Congress has held three furthermajor international congresses Barquisimeto Venezuela (1990) New Delhi India(1994) and Cape Town South Africa (1999) It has also held a series of more specialisedlsquointercongressesrsquo focusing on Archaeological ethics and the treatment of the dead(Vermillion USA 1989) Urban origins in Africa (Mombasa Kenya 1993) and The destruction and restoration of cultural heritage (Brac Croatia 1998) In each case these meetings have attracted a wealth of original and often inspiring work from manycountries
The result has been a set of richly varied volumes that are at the cutting edge of (frequently multi-disciplinary) new work and which provide a breadth of perspective thatcharts the many and varied directions that contemporary archaeology is taking
As series editors we should like to thank all editors and contributors for their hard workin producing these books We should also like to express our thanks to Peter Ucko theinspiration behind both the World Archaeological Congress and the One WorldArchaeology series Without him none of this would have happened
Martin Hall Cape Town South AfricaPeter Stone Newcastle UK
Julian Thomas Manchester UK June 2000
Preface
This book stems from a symposium organized by the editors on the archaeology ofdrylands held at the World Archaeological Congress at Cape Town in January 1999 TheCongress provided the opportunity to bring together scholars working on the archaeologyof different regions of the worldrsquos drylands to pool experiences and in particular toinvestigate the extent to which we could discern common themes
Although over a third of the worldrsquos population today lives in arid and semi-arid lands there are many gaps in our understanding about how fragile or resilient these regions arefor human settlement To fill these gaps we need to answer questions that are likely to beof very great significance for the global community in the twenty-first century Many dryland regions have abundant remains of ancient settlement and people have oftenspeculated that the actions of farmers and herders in the past must have been important increating the degraded landscapes of the present For decades the debate has beencharacterized more by speculation than informed debate and by a propensity to argue forsimple processes of cause and effect in terms of climatic change or humanly inducedenvironmental degradation In the past fifteen years or so however inter-disciplinary archaeological and palaeoecological studies (especially when employed within integratedresearch frameworks) have demonstrated their potential to move the debate forwards byproviding detailed case studies of how ancient societies actually exploited drylandlandscapes how they interacted with them and the complex environmental and socialcontexts in which they variously succeeded or failed Moreover as we conclude inChapter 1 the advancement of understanding about past dryland societies andenvironments and of the complexity of their interactions with each other has a particularurgency given the way in which changing political agendas have been prone to eitherdemonize or sentimentalize them
Nine papers were given at the Cape Town symposium and a series of common themesabout dryland settlement emerged from the discussion Other papers were thencommissioned by the editors so that the collection as a whole would draw on the archaeology of different kinds of drylands throughout the world (the locations of whichare shown in Figure 13) different periods of the past and different kinds of societies butall addressing the issues we had identified at Cape Town to give a comparativeperspective Common themes though as we discuss in Chapter 1 do not equate with similar solutions to dryland living or similar responses to threats and opportunities thearchaeology of drylands is eloquent testimony perhaps most of all to peoplersquos ingenuity as well as to their resilience
The editors would like to express their thanks to the organizers of the 1999 World Archaeological Congress for their invitation to organize the Drylands Archaeologysymposium and for every assistance from them during the conference We would alsolike to thank the Society for Libyan Studies for a generous grant towards our two fares toCape Town augmented in the case of GB by a grant from the Staff Development Fund of
the Faculty of Arts of the University of Leicester We are very grateful to all thecontributors to the volume for their patience especially those who contributed to thesymposium whose debates helped frame the discussion document we then circulated tothem and to the authors of the papers commissioned afterwards and who have remainedcommitted to our idea of the integrated comparative volume despite the timelag since theconference Finally we would like to dedicate this book to the memory of Barri JonesBarri was Professor of Archaeology at the University of Manchester and introduced usboth to dryland archaeology in the UNESCO Libyan Valleys Survey He died on the eveof his retirement in the summer of 1999 leaving a legacy of a major scholarly output ofbooks and papers an army of professional and amateur archaeologists enthused with hispassion for the subject and for his desert companions in particular memories of hair-raising adventures in his company He was a frenetic personality who was bothenchanting and exasperating to work withmdashhe was notorious for doing too many things at once mostly while nominally in control of a Landrover Amongst his many talentsthough he had an extraordinary topographical eye he was liable to get us lost in somedesert waste to visit an archaeological site once noted by a traveller 100 years ago andthrash the vehicle in the process but he was also by far the best person to be with to getsafely back to camp again
Graeme Barker and David Gilbertson January 2000
Part I INTRODUCTION
1 Living at the margin themes in the archaeology
of drylands GRAEME BARKER AND DAVID GILBERTSON
INTRODUCTION
Drylands cover 40 per cent of the land area of the Earth their total area is about 60million km2 of which about ten million km2 are hyper-arid deserts (Fig 11) Drylands support over one fifth of the worldrsquos population and arid and semi-arid lands together over a third Living conditions vary from the most affluent and profligate to thedesperately poormdashin some cases in close proximity The political stability and ecological economic and social sustainability of dryland settlement are among the mostdaunting challenges confronting the global community in the twenty-first century water seems likely to be a primary flashpoint for disputes between neighbouring states withdryland irrigation systems under strain from fast-growing populations
Figure 11 The world map of drylands Source After UNEP 1992
and with environmental refugees from global warming predicted to be in the order of 150million by the year 2050 under the business-as-usual scenario (Houghton 1997) thecatastrophic consequences will be particularly acute for dryland populations
Many dryland regions have archaeological remains suggesting that once upon a time
there must have been intensive phases of settlement in what are now dry and degradedenvironments (Fig 12) People have often speculated about what must have happened to turn past glories into present-day barrenness generally dividing in favour of climaticchange or human agency as the primary culprit Perhaps the climate shifted to greateraridity Or was it that people sowed the seeds of their own destruction through their follyfor example by developing irrigation systems that caused salinization or by stripping thelandscape for fuelwood or by allowing their livestock to over-graze the vegetation In general the debate has been characterized more by confident assertion than well-founded argument Furthermore as we discuss later in this chapter contemporary ecologicaltheory suggests that relations between dryland environments climate and people are byno means simple (Beaumont 1993) drylands can sometimes be remarkably resilient forexample recovering relatively quickly from over-exploitation and simple procedures byfarmers can often protect against the latter (Mortimore 1998 Tiffen et al 1994) These findings are at odds with the simplistic models that have tended to dominate thearchaeological literature about how climate and people may or may not have affecteddrylands in the past
Figure 12 A Roman-period fortified farm on the desert margins of Tripolitania northwest Libya
Photograph GBarker
Modern inter-disciplinary archaeology especially when working in conjunction withother social and environment sciences has the potential to move the debate forwardArchaeology deals with the entire human past its geographical scope is regionallyspecific but worldwide its scale of enquiry ranges from distributions and processes ofchange at the global scale and over millennia down to the actions of individuals We canuse the techniques of landscape archaeology to understand how different kinds ofsocieties whether recent or remote in time exploited the different dryland environments
The archaeology of drylands 4
of the world We can characterize the risks and opportunities confronting those societiesidentify the solutions they reached and often the reasons for them as well as monitoringthe short- and long-term effects of those solutions By developing a more sophisticated understanding than has hitherto characterized the debate about variability in past land usestrategies and the reasons for their successes and failures archaeology can contributeeffectively to modern debates about desertification and the sustainability of drylandsettlement (Beaumont 1993)
The World Archaeological Congress held at Cape Town in January 1999 provided an ideal opportunity to explore these issues from the perspectives of various scholarsworking on the archaeology of different regions of the worldrsquos drylands The symposium focused on nine contributions discussing work in the Near East North and sub-Saharan Africa and North America all of which are represented in this volume A series ofcommon themes about
Figure 13 The location of the case studies in this volume numbers refer to chapter numbers
dryland settlement rapidly emerged and the papers presented at the symposium wererewritten and further papers commissioned to address these common issues within acomparative perspective in this book with case studies drawn from different kinds ofdryland regions throughout the world (Fig 13) Common themes though as we discussbelow do not equate with similar solutions to dryland living or similar responses to risksand opportunities
Living at the margin themes in the archaeology of drylands 5
THEMES IN DRYLAND ARCHAEOLOGY
The term lsquodrylandsrsquo obviously fixes attention upon low precipitation Commonknowledge emphasizes that the climatic significance of this shortage depends upon theother aspects of the atmospheric environmentmdashthe radiation budget thermal regimewind regime the sources and pathways of moisture including fog as well as the manyother components of the biosphere and lithosphere that play significant parts in thehydrological cycle The meteorology and climatology that produce drylands are notsimple (Spellman Chapter 2) understanding them requires an appreciation of variability of precipitation and drought in both space and time Rainfall in many drylands istypically characterized as consisting of erratic short localized downpours of highintensity Low average precipitation totals are associated with notable variability Fierceand localized downpours creating sudden and dangerous floods are the primary resourcebase that many indigenous peoples have had to utilize to maintain themselves their cropsand their animals for millennia (as well as environmental hazards for archaeologistsworking in arid lands see Fig 14) However it is the prospect of prolonged and severe drought that dominates thinking about drylands Instrumental historical andpalaeoenvironmental records show that episodes of severe drought lasting decades ormore in length have not been uncommon in many drylands over the last few thousandyears (eg Bureau of Meteorology nd Fritts 1991 Lamb et al 1995 Nicholson 1994) Ingenuity flexibility and enterprise have been required from individuals communitiesand organizations to negotiate their survival in the face of such uncertainty and risk
It is important to remember that apart from such fluctuations at the scale of seasons and decades modern drylands have also been part of tremendous fluctuations in climateoperating at the global scale in the remote past from major shifts in the worldrsquos oceanographic and atmospheric systems The period from approximately 18000 to10000 years ago the last phase of the Pleistocene (the lsquoIce Agersquo) saw the last major ice advances of glaciers and icesheets in the high latitudes Regions of the world that nowenjoy a temperate climate such as Europe and North America were cold and aridRegions such as the Sahara were hyper-arid long-term reductions of rainfall reached well beyond the present desert margin as far south as present-day Nigeria with much of interior North Africa having to be abandoned by human populations However between10000 and 8000 years ago during
The archaeology of drylands 6
Figure 14 Drowning in drylandsmdashtwo vehicles of the UNESCO Libyan Valleys Survey sunk in a flash-flood
Photograph GBarker
the opening millennia of the Holocene or Postglacial the environment across the Saharawas notably wetter than occurs today with the development of lakes and woodland infavoured locations such as those that are now desert oases and savannah-like habitats on the surrounding plateaux (Barker et al 1996 and see Mattingly Chapter 9) The people who colonized these places were able to live by plant and shellfish gathering fishing andhunting not just steppe animals like gazelle but wetland species such as turtle hippo andcrocodile (Cremaschi and di Lernia 1998 Wendorf and Schild 1980) In the Near Eastwetter environments at this time were the context for the development of mixed farmingsystems of the kind found in the Wadi Faynan in southern Jordan (Barker Chapter 4) and the first farming in Turkmenistan may also have developed in wetter conditions thantoday (Nesbitt and OrsquoHara Chapter 6)
Desiccation started to develop about 6000 years ago in North Africa and the NearEast with people responding differently In the Sahara people shifted to cattle andsheepgoat herding (Barich 1987 1998 Wendorf et al 1984 1989) whereas the farmers of Wadi Faynan started to experiment with methods of trapping and storingwater which were the beginnings of recognizable systems of dryland farming there(Chapter 4) It was also about this time in the fourth millennium BC that prehistoric farmers in Turkmenistan started to build canals to divert floodwaters and via smallfeeder channels (aryks) bring it to their fields (Chapter 6) So far as we can tell it was not until notable aridification developed in the Sahara around 4500 years ago asseasonal streams replaced perennial streams salt pans replaced fertile lake floors and themodern regime of flash-floods and droughts developed that similar systems of dry farming were developed by farmers living in the oases (Chapter 9) and on the desert margins such as the Tripolitanian Pre-desert (Barker et al 1996)
Archaeological evidence for the development of irrigation-based farming is a recurrent theme throughout this book because whilst far from all drylands are warm or hot for all
Living at the margin themes in the archaeology of drylands 7
or part of the year the provision of an adequate and reliable supply of water in warm andsunny regions has been a goal for innumerable communities over time Rainwaterharvesting and floodwater farming are described for North America Africa and Asia in avariety of chapters Four major themes emerge from these case studies the first three ofwhich are closely inter-related
The first is the resilience of many farmers in antiquity to cope with harsh and risk-prone arid environments and their climatic vicissitudes over long periods of time Thislongevity of occupation points to an inherent robustness of many of these pastcommunities their attitudes and ways of life The second is the repeated evidence for thesimilarity of the lsquobuilding blocksrsquo or lsquotacticsrsquo employed by most ancient farmers indrylandsmdashbuilding walls to trap soil and divert or stem water flow building channels(including underground in the case of the Turkmenistan qanats [Chapter 6] and Libyan foggaras [Chapter 9]) to divert water and so on The third is the remarkable variability inthe overall systems that were put together from such building blocks and the way theyinvariably reflect detailed local knowledge of topography weather patterns and so onancient farmers knew from observation exactly how and where the water would flowafter a storm and so knew how best to manage that flow to suit their purposes
The fourth critical finding from our survey though is that the diversity we can observe in the archaeology of dryland farming systems is in no sense just a straightforward matterof commonsense observation by ancient farmers of what was the lsquobest fitrsquo to particular environmental or economic circumstances We can see today how dryland communitiesattempt to manage themselves and their habitats within the context of a whole nexus ofattitudes beliefs as well as economic social geographical educational agricultural andtechnological processes and whilst many of these details will elude archaeologists giventhe nature of our evidence the case studies illustrate how people took choices and notalways the right ones within a complex mix of factors including perceptions of risk theneed or desire for economic advantage and the institutional and regional context inwhich they were operating As Widgren comments (Chapter 14) models of agrarian development too often assume even developments of farming systems in response toparticular environmental social or economic pressures but the archaeological recordemphasizes above all the unevenness of development in both space and time We do notsee the kind of evolutionary development of land use systems so often assumed for thepast for example from simple to complex or from extensive to intensivemdashthey were lsquoformed and changed within specific place-bound social historical and ecological contextsrsquo (Widgren p 262)
One key influence on the character and scale of an irrigation system was notsurprisingly related to the extent to which the agricultural product was to serve only thelocal community Examples of such systems lsquoislandsrsquo of relatively intense landscape development are described for societies varying widely in time place and socialcomplexity for example in the North American Southwest in prehistoric times (MinnisChapter 15) the Achaemenid or Parthian periods in Turkmenistan (Chapter 6) the Libyan oases (Mattingly Chapter 9) and the Tricastin region of southern France (van derLeeuw Chapter 18) in the Roman empire and at various localities in sub-Saharan Africa in recent centuries (Sutton Chapter 11 Soper Chapter 12 Widgren Chapter 14) Widgren illustrates how both hierarchies and the absence of hierarchies can be associated
The archaeology of drylands 8
with labour-intensive agriculture and how both market orientation and subsistencefarming can be connected with labour-intensive farming
These landscapes and irrigation systems stand in contrast not only to the examples described in Israel (Rosen Chapter 3) Jordan (Barker Chapter 4) Syria (Newson Chapter 5) and northwest Libya (Gilbertson et al Chapter 8) of archaeological landscapes that were once very productive if peripheral parts of the imperial economiescharacterized by the large-scale import and export of goods products and information tothe core areas of the Mediterranean but also to the somewhat similar relationship ofhighland Mexico to the rest of the Inca state (Parsons and Darling Chapter 16) lsquoPatchinessrsquo in distribution has also been identified in the studies by Jones and Crook (Chapter 17) of the Swiss bissesmdashcanal systems that have tapped and redistributed water within the surprisingly dry environment of the Canton of Valais for over a millenniumbut that remain a relatively unknown but vital component of the economy of one theworldrsquos richest and technologically sophisticated countries Clearly archaeological investigation of lsquomarginalrsquo landscapes has to engage with the need for complex inter- and intra-regional articulations of explanations of cause and effect
All archaeologists also have to recognize the commonplace dictum that lsquothe past is a foreign countryrsquo things were thought and done differently there Then as now it seemsthat many individuals organizations and polities have behaved in relation to theirsituation and their environment in manners that do not appear rational to the modernexternal observer or to those with the wisdom of hindsight Then as now people madepoor decisions foolish decisions self-interested decisions carried out actions that they ortheir neighbours had cause to regret or more generally they misunderstood their landand situation One striking example of long-term devastation caused by the economic needs of the Roman empire was the pollution of the Wadi Faynan in Jordan by copperand lead mining (Barker Chapter 4) but it is important that we do not dismiss such actions as the exclusive domain of market-driven economies (like the profligacy of Turkmenistan irrigation farmers once they lost their sense of ownership of the system inthe Soviet period Chapter 6) as Minnis (Chapter 15) comments in the case of North America indigenous subsistence foragers and farmers equally have not always beenenvironmentally lsquocorrectrsquo lsquosoundrsquo or lsquoneutralrsquo
A good example of the importance of perception affecting decision making by drylandfarmers responding to adverse conditions occurred during the great drought that affectedthe wheat-arid far north frontier of South Australia between 1881 and 1884 the climatic effects of which were documented in remarkable detail by a sophisticated network ofinstrumental records maintained by amongst others telegraph operators (Bureau ofMeteorology nd) The well-known account of this episode by Meinig (1962) basedupon parliamentary records and newspapers reveals that the human and economicimpacts of the first two yearsrsquo drought were devastating Many wheat farmers weresustained by their belief that the lsquorain followed the ploughrsquo more ploughing and tilling they thought would release further soil moisture into the atmosphere and so break thedrought Others farmers desperate to maintain overall production totals and to pay theirmortgages to the Government for their newly acquired lands (pastureland hitherto)ploughed and planted ever larger areas of bush believing that minimal returns from vastareas would compensate for the poor yields per acre Both strategies of course made
Living at the margin themes in the archaeology of drylands 9
things far worsemdashas the drought grew worse and human distress grew ever moreprofound ever larger areas of land were being ploughed and farmed It is salutary forarchaeologists who rarely have access to such sophisticated documents or precisechronologies to reflect on the complete reversal in present thinking of the nature ofhumanmdashenvironment interactions that underpinned these farmersrsquo behaviour just over a century ago
The case studies demonstrate repeatedly that drought however pernicious and sustained is not necessarily the sole cause of lsquoabandonment phasesrsquo identified in the archaeological record of drylands Chapter 10 by Butler and DrsquoAndrea for example shows that episodes of famine and distress (like indeed those of success and prosperity)cannot be explained in drylands by consideration of one factor even such vital factors asdrought or flood Discussing the Northern Highlands of Ethiopia a region almostsynonymous with drought and famine for most readers they emphasize the potency forunderstanding famine in the area of the following sometimes archaeologically invisiblefactors human smallpox cattle rinderpest plagues of predators such as locusts ants andarmy worms conflict and social and political circumstances In fact drought aloneseldom causes famine The sustainability of many dryland communities pastoral(Kinahan Chapter 13) as well as agricultural is underpinned especially by the flexibility of traditional practices their capacity to avoid to mitigate and to create buffers againstrisk and adversity more generally their ability to organize themselves effectively indrought-prone habitats and in the last resort their willingness to relocate (Mortimore 1998122) We can see from the archaeological record that systems without suchflexibility did not have the necessary resilience for longterm survival as in the case of theproductive but short-lived systems of cash-crop farming that Romanized Libyansdeveloped in the Tripolitanian Pre-desert to supply the local military and coastal urban markets (Gilbertson et al Chapter 8) or the intensively irrigated fields built to feed the large industrial workforce of Roman miners in the Wadi Faynan in Jordan (BarkerChapter 4) or the massive centralist-administered irrigation systems of Soviet Turkmenistan (Nesbitt and OrsquoHara Chapter 6)
ARCHAEOLOGY AND DESERTIFICATION
During the last few decades many of the worldrsquos drylandsmdashthe hotter drylands especially but not exclusivelymdashhave been viewed as threatened by lsquodesertificationrsquo The term was coined by Aubreville (1949) in a report on the vegetation of Africa and itsmeaning has developed through time Thomas and Middleton (19949ndash10) defined it as lsquoland degradation in arid semiarid and dry sub-humid areas resulting mainly from adverse human impactrsquo and though some authors have also used the term to refer to landdegradation caused by a sustained aridification of climate most prefer to use the term torefer to the effects of human actions though climate and people can clearly work intandem to produce deterioration in dryland environments as may be the case in thecontext of global warming today (Barrow 1995 Millington and Pye 1994 SpellmanChapter 2) The key ideas focus upon significant and long-term degradation producing a loss of potential in biological soil and water resources Manifestations of humanly
The archaeology of drylands 10
caused degradation may include decreased vegetation cover timber loss salinizationreduced water supplies lower crop yields outbreaks of disease accelerated erosion ofsoils and dust storms and induced regional climatic change all encapsulated in thepopular metaphor of the conversion of pastoral or arable lands within drylands intodesolate and sterile desert
The Green agendas of recent decades have rightly and repeatedly focused on dryland ecosystems and the sometimes appalling consequences of human impacts upon them(Fantechi and Margaris 1986) often induced or certainly exacerbated by top-down programmes of economic development (IFAD 1992) Beaumont (1993474) concludedhis book with a bleak prediction of the inevitability of this process for the worldrsquos poorest nations lsquoin certain cases land degradation may be a sacrifice which has to be paid inorder that local populations can survive future drought or faminersquo According to Tolba and el-Kholy (1992134) the current rate of desertification is about 60000 km2 per year amounting to 011 per cent per year of the total area of dryland On this calculationdesertification today threatens no less than 70 per cent of the worldrsquos drylands which represents over 25 per cent of the worldrsquos land surface Grainger (1990) and Spooner(1989) argued that such desertification can be recognized in Australia North Americaand South Africa in the first half of the twentieth century and that it was likely to havebeen a factor in antiquity too
There are however conflicting views about the extent of desertification today Thomas and Middleton (1994) for example questioned whether desertification in recentdecades is actually a global problem of such vast dimensions as opposed to localmanifestations of local problems of smaller significance There are also examples ofspirals of land degradation being reversed by indigenous technological adaptationsworking in combination with population growth and market opportunities The Machakosdistrict of Kenya was considered an environmental disaster in the 1930s because ofmassive soil erosion and famine but by 1990 terrace construction had protected arableland farmed and protected trees provided sufficient fuel-wood and agricultural production per person and per hectare had increased sustaining a population five timeslarger than that of the 1930s (Tiffen et al 1994) Deforestation and massive erosion on the Yatenga plateau in Burkino Faso were exacerbated by mechanization programmesfunded by ill-judged development aid but the reinstatement of traditional systems of terrace building stopped erosion and doubled crop yields (Lean 1994) Mortimore(1998149ndash56) described an examples of c 150 years of sustainable intensification bysmallholders in Kano Nigeria in the context of population growth and monetization
Archaeological remains in many drylands have been grist to the mill of thedesertification debate For example Hughes with Thirgood (198260 74) wrote that
in the more arid regions forests that formerly moderated the climate and equalized the water supply were stripped away permitting the desert to advance The image of ruined cities in North Africa from which olive oil and timber were exported in ancient times but which were buried beneath desert sand epitomizes the environmental factor in the decline of civilizationhellip Roman dams and canals stand in dry wadis today as witness to the fact that the destruction of the vegetation and consequent desiccation have changed the
Living at the margin themes in the archaeology of drylands 11
environment
Terms such as excessive land use population pressure loss of biological diversity andvegetation cover mis-use of water accelerated soil erosion and ever-larger human needsall recur in archaeological explanations (Gilbertson et al Chapter 8) often in associationwith reference to times of perceived political unrest military invasion conflict anddrought
In his review of the historical likelihood of humanly induced desertification Spooner(1989) cautioned against the simplistic tendency to assume that desertification faminedrought and poverty will inevitably be found together and several case studies in thisbook support those who argue for the complexity of desertification processes today Forexample according to Barker et al (1996) and Gilbertson et al (Chapter 8 this volume)the vast Romano-Libyan and Islamic settlements and farms of the Libyan Pre-desert at theedge of the northern Sahara seem to have neither produced nor experienced the kind ofself-induced environmental degradation described by Hughes with Thirgood (1982)Indeed the increase in human population numbers farming intensity and landmanagement probably promoted a greener more diverse and infinitely richer andproductive environment than has occurred since the great aridification in climate thatafflicted the region some 4500 years ago There are in fact good reasons to suspect thatRomano-Libyan farmers did have significant and deleterious impacts upon their aridenvironment but there are few reasons to believe that catastrophic long-term climaticchange short-term catastrophic drought or anthropogenically induced environmentaldegradation played a central role in the progressive abandonment of these settlementsmdashaprocess that is still underway after nearly 1500 years
Rosen (Chapter 3) also suggests that substantial cultural changes in the ancient Negevdesert are not best explained by climatic catastrophe invasion or the inabilities of peopleto manage their desert environment rather periods of cultural florescence were related toincreased economic and social input from or integration with the Mediterranean lsquocorearearsquo with desert pastoralism strongly geared to active markets in the settled zone andlikely to be sorely afflicted by the latterrsquos collapse Yet in the adjacent deserts of southernJordan the Wadi Faynan Landscape Survey (with many of the same members as theUNESCO Libyan Valleys Survey and using similar methodologies) has foundconvincing evidence for dramatic humanly-induced land degradation in the wake ofagricultural and industrial intensification in the context of Roman imperialism In theSaharan Fezzan Garamantian development of foggara irrigation systems may have beena key factor leading to the decline of their civilization as a result of over-extraction from anon-renewable groundwater source (Mattingly Chapter 9) Ballais (Chapter 7) argues thatincreases in soil erosion in the eastern Maghreb in classical antiquity reflected specificcombinations of climatic change and human activities and affected parts of the landscapein different ways In the Roman imperial centuries increased intensity of rains or theannual amount of precipitation badly affected regions already made vulnerable byvegetation degradation or unwise cultivation systems whereas the irrigated zones andterraced mountains were more resilient In the middle and lower Rhocircne valley in southernFrance episodes of climatic change are out of step with archaeologically visible episodesof human impact in the Tricastin region here accelerated erosion can be tied clearly with
The archaeology of drylands 12
the lack of maintenance of Roman drainage systems (van der Leeuw Chapter 18) The role of pastoralism in desertification like dryland farming is much debated It has
often been asserted to be particularly pernicious the prime cause of a legacy ofapparently exhausted depleted and deserted drylands today The archaeological literatureon drylands is replete with references to the likelihood of lsquoover-grazingrsquo lsquoexcessive grazingrsquo and so on implying that there is today and that there was in times past someknowable and realizable intensity of pastoral activity in drylandsmdasha lsquocarrying capacityrsquo which if exceeded must have had dire consequences for the pastoralists and theirhabitats However the core of this viewmdashthat such a carrying capacity existsmdashis now being challenged by ecologists who consider that lsquoboom-or-bustrsquo models of animal population numbers may be more appropriate (Thomas and Middleton 1994) There maynever have been sufficient time for any medium- or long-term balance to be struck between livestock numbers and arid environments because arid lands are too variable intheir production of foragemdashthis variability itself a consequence of precipitation which is variable in space and time (Holling 1973 Noy-Meir 1978 Olsvig-Whittaker et al1993 Scoones 1995 Walker et al 1981) Typically drought is likely to have reduced animal numbers drastically before irreparable damage was done to pastures (Noy-Meir 1978) Indeed grazing drylands pastures may have had some overall beneficial impacts(Warren and Khogali 1992) whilst Olsvig-Whittaker et al (1993) argue that many dryland pastures may in some sense be lsquoadaptedrsquo to grazing stress and that pastoral disturbance could be regarded as a natural component of many arid environments Inbrief not all dryland environments are as fragile as some popular literature suggests(Thomas and Allison 1993)
Parallel arguments about the likely complexity of past relations between pastoralism and ecological change are put forward on the basis of archaeological evidence byGilbertson (1996) and Kinahan (Chapter 13) In the case of the Maghreb Ballais(Chapter 7) also concludes that periods of conquest often assumed in this region to beperiods of environmental devastation wrought by nomadic pastoralists were probably infact characterized by less arable land a progressive development of lsquonaturalrsquo vegetation and pastures and so less soil erosion
Hollingrsquos (1973) ecological view that arid lands are lsquonon-equilibrium but persistentrsquo may have utility for many archaeologists working in drylands not least because it servesas a disincentive to extrapolate from the local diagnosis of an ancient cause and effect tothe inference of causality at regional or global scales The possibility of non-linear relationships within and between environmental processes and human activities must alsobe considered Relatively minor changes in the human or biophysical environment can inprinciple set in train self-sustaining sequences of events and processes that can cause theenvironment to transform from one state to another with cause and effect entangled Indrylands today relationships between individuals communities institutions and thelandscapes with which they interact are clearly neither simple nor linear in form (Ellis1995 Phillips 1993 Chapter 10) The principal argument of the case studies in thisvolume is that the same was certainly the case in the past even though the nature ofarchaeological and palaeoecological evidence is such that it may sometimes be difficultor impossible to identify the key players critical species or ideas the nature ofunderlying trends or pre-disposing factors the agencies of stability and the triggers of
Living at the margin themes in the archaeology of drylands 13
change
CONCLUSION
The archaeology of drylands is one of the richest bodies of archaeological data andfrequently the most visible for archaeological surveyors though it is also often amongstthe most vulnerable to destruction by development programmes far from the eye ofnational archaeological services Moreover the deflated landscapes of most drylandsfrequently pose daunting challenges of archaeological analysis given the paucity of deepstratigraphies and organic remains though in some cases both the latter are present inrich abundance palaeoecologists often face similar challenges requiring similarpersistence and ingenuity in response However the practical (and often logistical)difficulties of dryland archaeology should not dissuade us from attempting to understandits significance whilst the details of structure and agency in past dryland settlement willoften be problematical to determine a better understanding of the complexity of peoplersquos interactions with dryland environments must surely underpin the desertification debate
In many parts of the world too investigating how past societies lived in drylands is critical for understanding not just how and why they lived as they did and desertificationtheory but also in the case of ancient water-management systems the extent to which the latter could or should be rebuilt to the advantage of local communities and theirecosystems today (eg Barker et al 1996) Similar arguments apply to pastoraldevelopment programmes (Kinahan Chapter 13) We need a sophisticated understanding of the environmental and social contexts of ancient dryland farmers and herders detailedknowledge of modern dryland ecologies and sympathetic awareness of issues such as theownership empowerment and organization of local technologies by indigenous peoples(Cullis and Pacey 1992 Reij 1991 and Gilbertson et al [Chapter 8] Minnis [Chapter 15] and Jones and Crook [Chapter 17])
Finally as Steve Rosen also points out in Chapter 3 (pp 57ndash8) the advancement of understanding of past dryland societies and landscapes through the combination of goodarchaeological science and social archaeology is critical most of all to combat thepoliticization of much past theorizing on these matters Relations between the desert andthe sown underpin many origin myths of ethnicity and the lsquobiographyrsquo of arid lands has frequently been rewritten to changing political agendas The role of desert pastoralists inancient Palestine as told to us through the Old and New Testaments is an obvious case inpoint where it is repeatedly represented as a moral force for good in the history of theIsraelites the preferment of the shepherd Abel over his brother Cain the farmer thesubstitution of the ram for Abrahamrsquos son Isaac the commandment that lists the ox andthe ass before the wife the parable of the sheep and goats and the lost sheep and the roleof Christ himself as the lamb of God the Lord our Shepherd of Psalm 23 A pastoralideology with numerous parallels to the Biblical stories underpinned the origin myth ofthe Incas and their sense of their right to rule subject peoples (Brotherston 1989) Yet inboth the Near East and North Africa simplistic notions of Islamic pastoralist invaders asthe prime causes of environmental and cultural decline have stemmed primarily frompolitical agendas (Rosen Chapter 3 Ballais Chapter 7) and one of the planks of modern
The archaeology of drylands 14
Zionism has often been the contrast between the lsquogreening of the desertrsquo of the kibbutz movement (never mind its long-term implications for the River Jordan) and the depiction of recent bedouin pastoralisrn as inefficient and environmentally destructive The drylandfarming systems of Native American peoples have been variously portrayed asenvironmentally destructive or in sympathy with the landscape according to changingcolonial and post-colonial perspectives (Minnis Chapter 15) At the root of the 1990s massacres in Rwanda was the lsquoTutsisrsquo and Hutusrsquo belief that they are derived respectivelyfrom Nilotic cattle-herders and Bantu farmersmdasha note left with a group of Europeantourists killed in Bwindi National Park in Uganda by Hutu guerillas in 1999 read inbroken French lsquohere is the fate of all the Anglo-Saxons who betray us to the Nilotics against the Bantu cultivatorsrsquo (Hannan 1999)mdashbut in fact there is very little to distinguish the two groups and the NiloticBantu dichotomy is almost certainly amistaken creation of nineteenth- and twentieth-century scholarship (Hall 1987 1996) Asdryland peoples face the uncertainties of the twenty-first century understanding the richness diversity and above all the complexity of the archaeology of their antecedentshas never been more urgent
REFERENCES
Aubreville A (1949) Climats Foragravets et Deacutesertification de lrsquoAfrique Tropicale Paris Societeacute drsquoEditions Geacuteographiques Maritimes et Coloniales
Barich BE (1987) Archaeology and Environment in the Libyan Sahara TheExcavations in the Tadrart Acacus 1978ndash1983 Oxford British Archaeological Reports International Series 368
Barich BE (1998) People Water and Grain The Beginnings of Domestication in the Sahara and the Nile Valley Rome lsquoLrsquoErmarsquo di Bretschneider
Barker G Gilbertson D Jones B and Mattingly D (1996) Farming the Desert The UNESCO Libyan Valleys Archaeological Survey Volume One Synthesis VolumeTwo Gazetteer and Pottery Paris UNESCO Publishing
Barrow CJ (1995) Developing the Environment Problems and Management London Longman
Beaumont P (1993) Drylands Environmental Management and Development London Routledge
Brotherston G (1989) Andean pastoralisrn and Inca ideology In JClutton Brock (ed)The Walking Larder 240ndash55 London Unwin Hyman
Bureau of Meteorology Commonwealth of Australia (nd) Results of Rainfall Observations Made in South Australia and Northern Territory 1839ndash1950 Melbourne Bureau of Meteorology Commonwealth of Australia
Cremaschi M and di Lernia S (1998) (eds) Wadi Teshuinat Palaeoenvironment and Prehistory in South-western Fezzan (Libyan Sahara) Florence Insegna del Giglio
Cullis A and Pacey A (1992) A Development Dialogue Rainwater Harvesting in Turkana London Intermediate Technology Publications
Ellis J (1995) Climatic variability and complex ecosystem dynamics implications forpastoral development In IScoones (ed) Living with Uncertainty New Directions in
Living at the margin themes in the archaeology of drylands 15
Pastoral Development in Africa 37ndash46 London International Institute for theEnvironment and Development
Fantechi R and Margaris NS (1986) (eds) Desertification in Europe Proceedings of the Information Symposium in the EEC Programme on Climatology Held in MytileneGreece 15ndash18 April 1984 Dordrecht DReidel Publishing Company
Fritts HC (1991) Reconstructing Large-scale Climatic Patterns from Tree-Ring Data Tucson University of Arizona Press
Gilbertson DD (1996) Explanations environment as agency In GBarker DGilbertson BJones and DMattingly Farming the Desert The UNESCO LibyanValleys Archaeological Survey Volume One Synthesis 291ndash317 Paris UNESCO Publishing
Grainger A (1990) The Threatening Desert London Earthscan Hall M (1987) The Changing Past Farmers Kings and Traders in Southern Africa
Cape Town David Philip Hall M (1996) Archaeology Africa Cape Town David Philip Hannan L (1999) Tourist massacre in Bwindi National Forest Uganda The
Independent 3 March 1999 Holling CS (1973) Resilience and stability of ecological systems Annual Review of
Ecology and Systematics 41ndash23 Houghton J (1997) Global Warming The Complete Briefing Cambridge Cambridge
University Press second edition Hughes JD with Thirgood JV (1982) Deforestation erosion and forest management in
ancient Greece and Rome Journal of Forest History 2660ndash75 IFAD (International Fund for Agricultural Development) (1992) Soil and Water
Conservation in Sub-Saharan Africa Towards Sustainable Production by the RuralPoor Amsterdam Free University of Amsterdam Centre for Development Cooperation Services
Lamb HF Gasse F Benkaddour A el-Hamouti N van der Kaars S Perkins WT Pearce NJ and Roberts CN (1995) Relations between century-scale Holocene arid events in tropical and temperate zones Nature 373134ndash7
Lean G (1994) How stones can hold back the Sahara The Independent on Sunday 16 October 199416
Meinig DW (1962) On The Margins of the Good Earth The South Australian WheatFrontier 1869ndash1884 Adelaide Rigby Limited
Millington AC and Pye K (1994) (eds) Environmental Change in Drylands Beogeographical and Geomorphological Perspectives Chichester John Wiley and Sons
Mortimore M (1998) Roots in the African Dust Sustaining the Drylands Cambridge Cambridge University Press
Nicholson SE (1994) Rainfall fluctuations in Africa and their relationship to pastconditions over the continent The Holocene 4(2)121ndash31
Noy-Meir I (1978) Grazing and production in seasonal pastures analysis of a simplemodel Journal of Applied Ecology 15809ndash35
Olsvig-Whittaker LS Hosten PE Marcus I and Schochat E (1993) Influence of grazing on sand field vegetation in the Negev desert Journal of Arid Environments
The archaeology of drylands 16
2481ndash93 Phillips JD (1993) Biophysical feedbacks and the risks of desertification Annals of the
Association of American Geographers 83(4)630ndash40 Reij C (1991) Indigenous Soil and Water Conservation in Africa London International
Institute for the Environment and Development Gatekeeper Series no SA27 Scoones I (1995) (ed) Living With Uncertainty New Directions in Pastoral
Development in Africa London International Institute for the Environment andDevelopment
Spooner B (1989) Desertification the historical significance In RHuss-Ashmore and SHKatz (eds) African Food Systems in Crisis Part One Micro-Perspectives 111ndash62 New York Gordon and Breach
Sutton JEG (1977) The African Aqualithic Antiquity 5125ndash34 Thomas DSG and Allison RJ (1993) (eds) Landscape Sensitivity Chichester John
Wiley and Sons Thomas DSG and Middleton NJ (1994) Desertification Exploding the Myth
Chichester John Wiley and Sons Tiffen M Mortimore M and Gichuki F (1994) More People Less Erosion
Environmental Recovery in Kenya Chichester John Wiley and Sons Tolba MK and el-Kholy OA (1992) (eds) The World Environment 1972ndash1992
London Chapman Hall Walker BH Ludwig B Holling CS and Peterman RM (1981) Stability of semi-
arid savannah grazing systems Journal of Ecology 69473ndash98 Warren A and Khogali M (1992) Assessment of Desertification and Drought in the
Sudano-Sahelian Region 1985ndash1991 London International Institute for Environmentand Development
Wendorf F and Schild R (1980) Prehistory of the Eastern Sahara New York Academic Press
Wendorf F Schild R and Close AE (1984) (eds) Cattle-Keepers of the Eastern Sahara The Neolithic of Bir Kiseiba Dallas Southern Methodist University
Wendorf F Schild R and Close AE (1989) (eds) The Prehistory of the Wadi Kubbaniya Dallas Southern Methodist University two volumes
Living at the margin themes in the archaeology of drylands 17
2 The dynamic climatology of drylands
GREG SPELLMAN
DEFINING DRYLANDS
Surprisingly given that the critical and unifying variable for dryland environments is ashortage of water on a seasonal or longer-term basis there has been a long-standing difficulty in determining their geographical extent (Beaumont 1989 Wallen 1967)though it is generally estimated that hyper-arid arid and semi-arid lands in total cover a third of the Earthrsquos land surface (UNEP 1992 see Fig 11) The absence of significant moisture is manifest in the characteristics of the soils vegetation and topographyConsequently Oliver (1973) and Nir (1974) have suggested ways of identifying aridlands by a variety of non-climatic criteria Straightforward classical approaches create regionalizations using isopleths of climatic elements with respect to associations withvegetation and agricultural conditions such as the 250 mm rainfall limit as the aridboundary (Oliver 1981) In contrast indexing methods delimit regions with differinglevels of aridity by the application of objective standard formulas
The best-known classical method is that of Koppen (1931) who defined drylandregions in terms of an annual precipitation and temperature index Assuming a meanannual temperature of 18degC his formula gives a maximum precipitation of 640 mm for semi-aridity with summer rainfall and 360 mm with winter rainfall and the calculation that drylands occupy about 26 per cent of the total Earth surface with the desert regioncovering 12 per cent and semi-desert and steppes the other 14 per cent The system wascriticized by Mather (1974) for failing to consider water supply and having no physicalmeaning or indication of the atmospheric processes involved Water-balance models were developed independently by Penman (1948) and Thornthwaite (1948) Penmanrsquos model is more sophisticated and considers turbulent transfer and energy balance approachesThornthwaitersquos model considers the energy balance alone using P the mean annual precipitation (mm) and a calculation of Pe the mean annual potential evaporation (mm)in the calculation of a moisture index (Im) resulting in Im=100[PPeminus1] In this system arid regions have an index value of under minus667 whereas a semi-arid region is defined where Im lies between minus333 and minus667 The method was criticized by Wallen (1967) fortending to over-estimate water supplies Other water-balance methods are reviewed by Jones (1997)
The definitive map of the spatial distribution of dryland areas produced by UNEP (1992 see Fig 11) divides the globe on degrees of bioclimatic aridity using the values ofthe ratio PPET that is P=the mean annual precipitation (mm) and PET=the mean annual potential evapotranspiration (mm) as calculated by Penmanrsquos formula Three categories are relevant here hyperaridity where the PETP ratio is less than 005 aridity from 005
to 020 and semi-aridity from 020 to 045 Some classifications of drylands also include the dry sub-humid regions (045ltPPETlt065) (Le Houerou 1996) The hyper-arid zone is characterized by true desert climates where precipitation is extremely low andirregular in occurrence Perennial vegetation is almost totally absent and neither pastoralnor arable farming is possible using rainfall The arid zone receives annual rainfall totalsof 80ndash350 mm with inter-annual rainfall of 50ndash100 per cent (Beaumont 1989) Scattered vegetation allows low-intensity grazing but rain-fed agriculture is unlikely The semi-arid region with precipitation totals of 200ndash700 mm is dominated by grassland and scrub providing relatively good grazing Table 21 gives the regional distribution for theclasses types of drylands and Table 22 gives some examples of each regime
These methods give precise definitions of the desert boundary but discrepancies occur estimates of the worldrsquos dryland areas based on climatic classifications range from 26 to 36 per cent the greatest difference being with respect to the location of hyper-arid areas (Table 23) A figure of 43 per cent was obtained by McGinnies (1988) using moistureshortage of moisture in the main criterion Yair and Berkowicz (1989) have suggestedthat aridity should be redefined to include the sensitivity of an area to low rainfall byassessing the intensity and duration of rainfall soil salinity and the ratio of soil cover tobare rock
The following dryland types can be identified with reference to rainfall and thermal regimes two rainy seasons (eg Venezuela) winter rainfall (eg the North Africancoast) summer rainfall (eg the Sahel) almost rainless (eg the Sahara) and fog andmist (eg the Atacama desert) Seasonality in rainfall reflects the dominant rain-producing mechanism summer rains are
Table 21 Regional distribution of world drylands (103km2) (After Le Houerou 1996)
Zone AfricaAsiaAustralasiaEurope North America
South America
Total
Hyper-arid 6720 2773 0 0 31 257 9781 75 Arid 5035 6257 3030 110 815 445 15692 121 Semi-arid 5138 6934 3090 1052 4194 2645 23053 177 Dry sub-humid
2687 352 513 1835 2315 2070 12947 99
Table 22 Rainfall regimes at selected dryland stations (Data from Pearce and Smith 1984)
Place Altitude (m) Location Annual Total (mm) J F M A M JJ A S O N D Fiya Chad 225 18deg0primeN 18 0 0 0 0 0 0 0 18 0 0 0 0 19deg10primeE Khartoum 390 15deg37primeN 157 0 0 0 0 25 7 53 71 18 5 0 0 Sudan 32deg33primeE
The dynamic climatology of drylands 19
brought by the Inter Tropical Convergence Zone (ITCZ) so regions with summer rainfalllie on equatorward-margins of drylands whereas regions with winter rainfall under theinfluence of mid-latitude disturbances are on the poleward sides Dryland types based ontemperature classifications are tropical deserts exhibiting little change in monthlytemperatures (eg Somalia) subtropical deserts experiencing considerable temperaturechanges (eg the Thar and Australian deserts) temperate drylands with cold winters (egdrylands in Iran Syria and Mongolia) and cold highland areas (eg Tibet) Yet anotherscheme is that of Thomas (1989) hot arid lands (coldest month temperature 20ndash30degC) drylands with mild winters (coldest months 10ndash20degC) drylands with cool winters (coldest months 0ndash10degC) and drylands with cold winters (coldest month less than 0degC) Despite these varieties of definition however the overall purpose of drylandclassification is the same to identify their global significance to examine the processesthat operate to create them and to assess whether any major changes are occurringAgnew and Anderson (1992) remark that there is a grave danger that arid lands aretreated by water resource managers as homogeneous entities with similar environmentsand similar problems when this is clearly not the case
CLIMATIC CHARACTERISTICS
Precipitation
By definition all dryland areas receive low annual precipitation and in most drylandareas as rainfall amounts diminish there is a corresponding increase in variability andunreliability (Le Houerou 1996) Mean values therefore do not adequately describe thetrue nature of the precipitation regime because annual totals will show significant year-
Kashgar 1309 39deg24PrimeN 86 15 3 13 5 8 5 10 8 3 3 5 8 China 76deg07primeE Amman 777 31deg57primeN 278 69 74 31 15 5 0 0 0 0 5 33 46 Jordan 35deg57primeE Lima Peru 120 12deg05primeS 43 3 0 0 0 5 5 8 8 8 3 3 0 77deg03primeW
Table 23 Estimates of the land area of arid lands using the climate classifications of Koppen (1931) Thornthwaite (1948) Meigs (1953) and UNESCO (1977)
Koppen Thornthwaite Meigs UNESCOHyper-arid 120 153 205 195 Semi-arid 143 152 158 133 Total 263 305 363 328
The archaeology of drylands 20
to-year departures from long-term norms Nir (1974) for instance mentions a rain eventfrequency of once every eight years at certain sites in the Sahara and once every eighteenyears in Peru The interquartile or 10ndash90 percentile ranges are more useful indicators(Beaumont 1989) The variability of rainfall in arid areas is greater than that of temperateregimes because of the character of the measure used the coefficient of variation (COV)calculated by the division of the standard deviation by the mean Areas with low rainfallwill inevitably record the highest variation even though the magnitude of that variabilityaway from the mean is smaller The variability in absolute terms may not be much greaterthan that of temperate regions (Cooke and Warren 1973) but for areas with low rainfalleven small variations are extremely significant (Agnew and Anderson 1992)
When rainfall events do occur in dryland areas it is often when rainbearing frontalsystems or tropical cyclones penetrate the region Incursions are therefore more frequentat the margins of dryland areas The usual mechanism in poleward areas is the southwardmovement of cold lsquoupper lowsrsquomdashareas of cold air in the upper atmosphere that have beencut off from the prevailing westerly circulation under conditions of low zonal flow in themid-latitude index cycle (Barry and Chorley 1998) On the equatorward margins theremnants of tropical cyclones and the seasonal advance of the ITCZ are important givingvery localized short-lived and often high-intensity rainfall events (Fig 13) Examples in Algeria listed by Barry and Chorley (1998) include 87 mm in three minutes (El Golea)385 mm in 25 minutes (Beni Abes) and 46 mm in 63 minutesmdashthough such catastrophic events are not a universal characteristic of drylands (Gordon and Lockwood 1970) Whatis certainly typical is the highly localized nature of rainfall (Beaumont et al 1988 Sharon 1972 1981)
Synoptic climatological methods have long demonstrated their validity for the analysis of regional rainfall variability (Barry and Perry 1973 Sweeney and OrsquoHare 1992) and to model regional scenarios of climate change (Wilby and Wigley 1997) A weather-type indexing method originally developed in an investigation of rainfall variability in Egyptby El Dessouky and Jenkinson (1975) has been adapted for investigating the role ofatmospheric circulation pattern on rainfall in dry areas of Spain surface index values canbe correlated with rainfall amounts and statistically significant associations have beenidentified (Spellman 2000) There are therefore distinct opportunities for the analysis ofhistoric rainfall events and drought
Drought
The World Meteorological Organization (1975) defines drought as ldquoa deficit of rainfall in respect to the long term mean affecting a large area for one or several seasons or yearsthat drastically reduces primary production in natural ecosystems and rain-fed agriculturerdquo Drought is commonly defined in terms of its impacts rather than its causeshence the terms lsquoagricultural droughtrsquo and lsquohydrological droughtrsquo have been proposed (Smith 1992) Drought and aridity are not the same thing aridity refers to a negativeratio between mean annual rainfall and mean annual potential evapotranspiration Thedegree of aridity is inversely related to the magnitude of this ratio but drought is more orless related to aridity because arid regions experience frequent droughts Drought impactsare worst in dryland areas because the low mean annual variability of rainfall is
The dynamic climatology of drylands 21
associated with high variability and drought duration is greatermdashthe drought in the Sahel began in the late 1960s and continues rainfall still not reaching the 1931ndash1960 mean (Hulme and Kelly 1993 Morel 1992 Nicholson 1993 Nicholson et al 1988)
Temperature
It is far harder to generalize about the thermal regimes of dryland areas Annualtemperature ranges are greatly affected by altitude and the distance from the sea but ingeneral high summer temperatures as a consequence of high radiation loads are commonto all regions (eg Fig 21) In the Sahara maximum average daily temperatures of more than 45degC are recorded in the interior and July temperatures of 375degC are recorded elsewhere except in the highlands
Figure 21 Thermal regimes in two dryland locations Aswan Egypt (112m 24ordm 02prime N 32deg 53prime E) and Jacobabad Pakistan (57m 28deg 17prime N 68deg 29prime E)
Relative humidity
Atmospheric moisture content is typically very low above dryland areas (Table 24) dryness is a response partly to the lack of local evapotranspiration and partly to the lackof horizontal moisture advection Coastal drylands have high humiditiesmdash60 per cent in parts of Western Australia for example compared with under 20 per cent 150 km inlandContrasting humidity regimes are shown in Figure 22 Walvis Bay Namibia in coastal drylands has high humidity In Saleh in the interior of Algeria has similarly low rainfallbut a much drier atmosphere Damascus Syria and Timbuktu Mali respectively northand south of the subtropical anticyclones have annual regimes that mirror each otherdepending on the timing of the rainy season
The archaeology of drylands 22
Wind
Persistent strong winds are common in drylands often a consequence of extensive flatareas with little vegetation cover to disturb air movement in the boundary layer Wind is amajor agent of erosion lifting significant quantities of dust from the dry soil hazyatmosphere with low visibilities is commonplace Wind has an important indirect climaticrole because the amount of dust influences the surface energy balance aiding the processof desertification (Le Houerou et al 1993)
ATMOSPHERIC PROCESSES CAUSING ARIDITY
Condensation of moisture in clouds occurs when moist air is cooled to the point wherebysaturation is reached This occurs through ascent mixing radiation-cooling or contact-cooling with a colder underlying surface The clouds that form must then grow to asufficient depth in order that drops of water can grow to a size to overcome air resistanceand fall to the ground The
Table 24 Mean relative humidity at various isobaric levels for radiosonde stations in the Sahara and the Arabian peninsula (After Lockwood 1974)
Isobaric level January July Fort Trinquet (25deg14primeN 11deg35primeW 360m) 850 33 16 700 22 22 500 15 31 Aulef-el-Arab (27deg04primeN 1deg06primeE 275m) 850 28 14 700 22 17 500 15 21 Habbaniya (33deg22primeN 43deg34primeE 45m) 850 53 15 700 33 18 500 26 7
The dynamic climatology of drylands 23
Figure 22 Mean monthly relative humidity at four locations with their mean annual rainfall (mm) shown in parentheses
atmospheric processes that promote aridity are thus those that (1) result in a completelack of atmospheric moisture or (2) inhibit cooling of the air through the prevention ofconvection or the creation of inversions or (3) reduce humidity by warming theatmosphere Climatological processes that cause aridity operate at both global andregional scales Thompson (1975) outlines four main processes that help to explain thedistribution of arid lands
The first and most important (Hills 1966) is atmospheric subsidence on the poleward side of the subtropical anticyclones Aridity results as descending air is slowlycompressed and subsequently adiabatically warmed leading to a dry stable atmosphere(Fig 21) Subsidence within the anticyclones does not extend right to the surface since normally the warm dry subsiding air is insulated from the surface by a shallow layer ofrelatively cool air The properties of this boundary layer can be completely different tothat of the sinking air and are usually maintained by a source outside that of the mainanticyclone If the air forming this surface layer originated over the sea it may be moistand contain layer cloud which can result in light rain or drizzle Sinking air will alsoprevent significant depth of thermal convection despite high radiation receipt andsubsequent strong surface heating under clear skies Dryland areas are centred beneaththe subtropical anticyclones in both hemispheres
Wind direction is also important air flowing over the interior of a continent has areduced opportunity to absorb moisture at its base so strong stability and low humiditieswill develop in the lower levels In the northern hemisphere dry northeasterly winds (thereturning flow of the Hadley Cell circulation) contribute to much of the aridity of
The archaeology of drylands 24
Southwest Asia and the Middle East The third factor is topography natural obstaclesacross the path of prevailing winds can cause aridity on their leeward side Thus as moistair is forced to rise over a mountain range (Fig 23) air cools at the dry adiabatic lapse rate (A to B) until saturation is reached and then at the moist adiabatic lapse rate (B to C)until the cloud top in the lee of the mountain range the descending air warms at the dryadiabatic lapse rate (C to D) and will be warmer than the ascending air at eachcorresponding altitude The air stream will arrive at the other side as a dry desiccatingwind as for example occurs on the leeside of the Sierra Nevada in North America or theAndes in South America
The fourth process is cold ocean currents Onshore winds that pass over cool equatorward-flowing ocean currents close to the shore will be rapidly cooled in the lowerlayers (up to 500 m) This induces atmospheric stability which then reduces the potentialfor rainfall production by promoting thin extensive sheets of stratiform cloud cover andpersistent coastal mists and fogs At higher altitudes the air will be warm thus creating astrong inversion that further prevents convection Examples where this effect is importantinclude the Atacama and Namib deserts under the influence respectively of the Peru andBenguela currents
Figure 23 The rainshadow effect leading to aridity Source After Agnew and Anderson 1992
SURFACE ATMOSPHERIC CIRCULATION
The low latitudes are dominated by the meridional circulation of the Hadley Cells athermally driven rising limb of air in the equatorial zone a poleward-moving flow in the
The dynamic climatology of drylands 25
upper atmosphere a sinking limb in the region of the subtropics and a returning tradewind flow at the surface that converges with corresponding winds from the oppositehemisphere at the ITCZ (Fig 24) On the poleward side of the upper atmosphere abovethe return branch of the cell is the subtropical jet a relatively narrow band of high-velocity westerly winds encircling the Earth The Hadley Cells exhibit marked seasonalvariation in intensity geographical extent and latitudinal position
Subtropical anticyclones
Between about 20deg to 40deg mean surface pressure patterns are dominated by adiscontinuous belt of subtropical high (STH) pressure areas broadly elliptical in shapeand oriented in an east-west direction On average they dominate the ocean basins inthese latitudes The geographical positions of the centres of the subtropical highsfluctuate In winter in the southern hemisphere they intensify and spread over theadjacent continental areas and an almost closed belt of high pressure can be formed Insummer thermally produced low-pressure centres over land masses (Australia southern Africa) disrupt the pattern In the northern hemisphere higher central pressure isexhibited in summer which is a time when the STHs also show their greatest extent
The high-pressure centres display a regular movement During the winter season STH centres exhibit equatorwards movement which is reversed in summer A change of onlyone third of a degree of latitude (about 35 km) in the position of the Atlantic high (almostunobservable) causes a one degree
Figure 24 The Hadley Cell circulation of the tropical northern hemisphere
Source After Musk 1988
change in latitude in the position of the ITCZ with an immense effect on rainfall in theSahel (Oliver 1981) For reasons that are unclear (Hastenrath 1985) the STH centresalso migrate longitudinallymdashin winter in the northern hemisphere all subtropical highsare centred over the eastern regions of their respective ocean basins whereas duringsummer they migrate to the west
The archaeology of drylands 26
Changes in the intensity of STHs (as measured by sea level pressure) are clearlydisplayed in the southern hemisphere The South Pacific anticyclone tends to be strongestin the southern hemisphere spring Jones (1991) has shown that the centre of the SouthPacific anticyclone has declined in strength over the period 1951ndash1985 yet the northern flanks have strengthened In the northern hemisphere temporal variations in the STHintensity have also been identified the North Atlantic anticyclone for instance showed asignificant increase in surface pressure between 1946 and 1987 (Inoue and Bigg 1995)
Subtropical highs are generally asymmetrical in structure with highest pressures in the east at the Earthrsquos surface and maximum pressure to the west at altitude Consequently the circulation around the centre is not parallel to the Earthrsquos surface but slopes gently towards the west with subsidence dominating the eastern half and rising air currentsmore frequent in the west Thus western air masses are more unstable and humidwhereas in the east conditions are generally cloud-free or if clouds are present they display limited vertical development like thin stratocumulus
Origins of the subtropical anticyclones
Classical dynamic explanations of the origins of the subtropical anticyclones attributetheir existence to the lsquopiling uprsquo or convergence of the poleward-flowing upper air lsquoanti-tradesrsquo at about 20deg inducing a downward movement of air and high pressure at thesurface Alternatively the main cause may be the movement of polar air (McIlveen1992) As a result of changes in the Coriolis Force with latitude anticyclonic cells nearthe polar front have a tendency to move equatorwards while low-pressure centres generally migrate towards the poles (Rossby 1947) These travelling cold anticyclonesfrequently rejuvenate the subtropical highs Polar outbreaks would therefore prefer theeastern parts of the ocean basins where cold ocean currents prevail and where frictionalong the continental coasts gives a strong meridional influence on these movementsPulses in the intensity of subtropical highs on a daily time scale might be explained bythis idea (McGregor and Nieuwolt 1998) In addition the interaction between cold polaroutbreaks and surface ocean currents maintains the observed higher pressure over theeastern oceans at low levels and the stronger development of STHs over the southernoceans Thermal explanations have also been proposed involving cooling in the upper airand cooling at the Earthrsquos surface Upper air cooling will occur when air in the upper poleward-moving branches of the Hadley Cells loses heat by long-wave radiation to space The air thus becomes progressively denser subsides and leads to high pressure atall levels Cooling at the Earthrsquos surface is seen as a response to cold ocean currents andthe cool continental land areas in winter These features may correlate with the cellularpattern of the subtropical highs and their extension over continents
An explanation of the existence of subtropical highs can also be found in theconsideration of both thermal and dynamic mechanisms According to McIlveen (1992)the Coriolis effect may impose dynamic constraints on the flow of air in anticyclonicsystems once upper level convergence occurs the Coriolis Force prohibits the outflow ofair at surface leading to atmospheric mass build-up in the anticyclonic centres with the warming effect of air subsidence reducing the vertical pressure gradient so that airpressure falls more slowly with increasing altitude Isobaric surfaces therefore tend to
The dynamic climatology of drylands 27
lsquodomersquo rising to greater heights than in the surrounding air which produces deep andwarm anticyclonic systems
Trade winds
Between the subtropical highs and the ITCZ the low-level circulation of the atmosphere is dominated by the persistent easterly winds known as the lsquotrade windsrsquo These have a distinct three-layer structure the heights of which increase towards the equator (Fig 25) The height and intensity of the inversion layer will have a strong influence onprecipitation mechanisms These features are generally dependent on latitude or distancefrom the STH centre yet lower more intense inversions (and subsequent weakerconvection) are associated with the east side of ocean basins on the coast of West Africafor example the intensity can be between 5ndash8degC markedly reducing rainfall potential
Figure 25 The structure of the trade wind atmosphere
The Inter-tropical Convergence Zone
At the equator flank of dryland areas in regions classified as semi-arid rainfall is governed by the seasonal fluctuations in the position of the Inter-tropical Convergence Zone (ITCZ) This is commonly perceived as a belt of low pressure encircling the globewhere the two Hadley circulations meet but Waliser and Gautier (1993) have identifiedseven separate ITCZ zones (Table 25) differentiated by structure and behaviour In the northern hemisphere dry conditions are associated with hot continental tropical airwhich moves in behind the ITCZ as it migrates southwards during the winter In Africaat its most southerly extent in January or February the ITCZ lies at about 8deg north of the equator in the west but about 15ndash20deg south of it in the east This is the dry season of the north The ITCZ moves north during the northern summer but the extent of theprogression (up to about 20degN) shows considerable year-to-year variability with latitudinal departures of up to 6deg (800 km) for some regions and up to 2deg for the global
The archaeology of drylands 28
average These departures which can last from 3 to 18 months may produce lengtheneddrought periods
UPPER AIR CIRCULATION
Explanations of the spatial variability of precipitation conditions can be afforded byreference to upper atmospheric flow particularly the position and intensity of thesubtropical jet stream During the winter dry season when anticyclonic conditions prevailover North Africa the jet stream becomes convergent towards the equator and produces adownward shift of air to feed high pressure at the surface In the summer months upperair divergence results from the convergence of southwesterly and northeasterly winds atthe surface This tends to draw in moist air and increase the likelihood of precipitationLow rainfall totals in dryland areas are explained by weak easterly jet streams associatedwith weaker circulation in the middle and upper troposphere
Subtropical Westerly Jet
Classical models commonly portray the upper air poleward-moving section of the Hadley Cell as a meridional flow (the lsquoanti-tradesrsquo) yet in reality this will be strongly redirected by the Coriolis Force as soon as it moves away from the equator resulting in a narrowband of high-velocity westerly winds known as the Subtropical Westerly Jet which isfound on average at around 30deg from the equator in both hemispheres Palmen andNewton (1969) describe the SWJ as a persistent long-wave pattern encircling the globe with wave troughs at 20degW 150degW and 90degE and wave crests at 70degW 40degE and 150degE Maximum wind speeds of up to 100 m per second are found in the vicinity of the wavecrests
The mean position of the SWJ and the year-to-year variability of the crests and troughs influence precipitation patterns in the low latitude regions Flow reaches its maximumintensity in the winter months when the pole-to-equator thermal gradient is greatest the core moves towards the poles as the Hadley circulation strengthens At 200 mb the SWJwill lie over the poleward flanks of the STHs If individual high-pressure cells contract away from one another as meanders develop in the jet between them the troughs canextend southwards to interact with low-level (850 mb) underlying tropical easterlies (Fig
Table 25 Seven ITCZ zones (after Waliser and Gautier 1993) Zone Longitude limits (deg)Africa 10ndash40E Indian 60ndash100E West Pacific 100ndash150E Central Pacific 160Endash160W East Pacific 100ndash140W South America 45ndash75W Atlantic 10ndash40W
The dynamic climatology of drylands 29
26) In the central parts of the Sahara rainfall occurs under the variable northward penetration of the West African monsoon trough which allows tongues of moistsouthwesterly air to travel comparatively far north producing short-lived low-pressure centres Low-pressure centres then move north along the meander trough though they are often lsquorained outrsquo when they reach the central Sahara Related to the SWJ but far less common is the southward movement of Mediterranean cold fronts Barry and Chorley(1998) noted such an event in December 1976 in southern Mauretania which yielded 40mm of precipitation
Tropical Easterly Jet
The Tropical Easterly Jet extends from Southeast Asia (80degE) to North Africa (50degE) at approximately latitude 15degN It spirals out clockwise from the subtropical high pressure centres and flows in the northern summer months (June to September) because it isrelated to the seasonal heating cycle (Hastenrath 1985) The strongest intensity is atabout 15 km altitude where maximum speeds are in the region of 40 m per second thereis a second weaker easterly flow at about 5 km The TEJ owes its existence to the strongsurface heating in summer over the land masses in Africa and Asia where very intense
Figure 26 The interaction between the subtropical westerly flow and the tropical easterlies leading to the creation of Saharan depressions which move eastwards along the trough axis
Source After Barry and Chorley 1998
The archaeology of drylands 30
heat lows promote the ascent of air to the upper atmosphere Mass convergence in theupper high-pressure system is so intense that pressure surfaces bulge upwards creatingan atmospheric thickness difference between the subtropics and the equatorial and midlatitudes resulting in a reversal in the normal equator-to-subtropics temperature gradient with warmer temperatures recorded in the subtropical upper atmosphere (McGregor andNieuwolt 1998) Convergence in the jet over Africa induces subsidence over the Saheland may be responsible for preventing the advancement of the West African monsoonrains Generally rainfall is greatest north of the jet entrance in the southern Asian regionand south of the jet exit in the West African region
West African Mid-Tropospheric Jet
The West African Mid-Tropospheric Jet is associated with rainfall patterns in African drylands and arises in response to mid-tropospheric temperature gradients between the warm Sahara desert area and the cool waters of the Gulf of Guinea off West Africa It hasits core at 15degN at about 4500 m (600 mb) and is located in a region south of the centralarea of anticyclonic outflow Maximum intensity occurs in the northern winter whenhemispheric temperature gradients are steepest flow can reach 10 m per second Rainfalloccurs on the equator side of the jet above the Saharan heat low
East African Low-Level Jet
The East African Low-Level (850 mb) Jet has a wandering parabolic course over thewestern Indian Ocean (Fig 27) It exists in all months but its
The dynamic climatology of drylands 31
Figure 27 The monthly progression of the East African Low-Level Jet Core
Source After McGregor and Niewolt 1998
greatest development is related to the onset of the African-Asian monsoon circulation In winter it is confined to the southern hemisphere but it is an integral part of the northernsummer monsoon circulation in the African-Indian area Maximum coolness moistureand cloudiness coincide with the jet core and its eastern regions over the coast of easternAfrica where maximum ascent of air occurs (Kamara 1986) Minimum cloudinessoccurs above the jet core and to the west in the direction of the footslopes of the EastAfrican plateau where the air is descending creating the warmest driest and most stableair (Findlater 1972) The core of this jet occurs at about 1500m where velocities can
The archaeology of drylands 32
reach 25ndash50 m per second Branches of the jet can penetrate inland over eastern Africathrough topographic breaks in the East African plateau It has been related to rainfalloccurrences in the northern parts of Ethiopia and the tracks of so-called Sudano-Sahelian depressions (Fig 28)
PREVAILING WINDS AND MOUNTAIN BARRIERS
Where topographical factors are added to those caused by the general circulation aridityis greatly increased and it is in these areas that the most severe deserts are foundMoisture available for precipitation is trapped in a shallow layer beneath the subtropicalinversion The depth of this moist layer varies but if a mountain barrier projectsthroughout this moist layer it interrupts the surface flow and the surface moist layer willnot penetrate behind the mountain range Even if the range does not completely block themoist layer the reduction in moisture advection to the lee of the range can still besubstantial of the order of 60ndash70 per cent (Lockwood 1974) Dryness can further beenhanced by subsidence of air near the inversion down the lee slopes Such mountain-enclosed inland basins can be extremely arid Death Valley in California is a primeexample
OCEAN TEMPERATURES
Ocean temperature has a considerable influence on climate particularly in coastalregions cool ocean currents moving towards the equator stabilize the atmosphere andreduce atmospheric instability When cold water along the equator is well developed theair above will be too cold even to take part in the ascending motion of the Hadley Cellcirculation Along the coast of Peru the surface moist layer is less than 800 m deep andnormally only drizzle will fall from a deck of stratus The coasts of South America andSouthwest Africa are sheltered respectively by the Andes and Namib escarpment fromthe dynamically stable easterly trades allowing shallow tongues of cold air to roll in fromthe west These are capped by strong inversions at c600ndash1500 m which reinforce the trade wind inversions precluding the development of intense convective cells exceptwhere orographic ascent occurs Precipitation from fog may also result from oceancurrents When rain does fall it is on those rare occasions when large-scale pressure changes prevent sea breezes and fog
DESERTIFICATION
Desertification has been defined in various ways recently (McGregor and Nieuwolt1998) as the process by which dryland conditions are brought into
The dynamic climatology of drylands 33
Figure 28 The tracks of Sudano-Saharan depressions over the Sahara
Source After Barry and Chorley 1998
areas where such conditions did not previously exist According to Le Houerou (1996) ifhyper-arid zones are excluded (as not susceptible to further desertification) 38 per cent of drylands can be described as desertified which is 16 per cent of the overall land area (LeHouerou 1996 Table 26) However it remains debatable how extensive desertification is or how fast it is proceeding (Thomas and Middleton 1994) Two major factors areinvolved in the desertification process (though their relative magnitudes are unknown)human activities and drought as a consequence of climatic variability
Observing that desertification occurred in the Sahel during the 1950s and early 1960sin spite of the fact that rainfall was well above the long-term average Le Houerou (1996146) concluded that lsquodesertification may therefore result from land abuse alonersquo Most meteorological models for dryland expansion or the occurrence of episodes ofdrought point ultimately to local changes to the surface energy balance or to large-scale shifts in atmospheric circulation lsquoHuman impactrsquo theories generally focus on areas withsparse vegetation which will commonly have surface air temperatures that are lower thantheir surroundings due to the increased amounts of surface reflectivity of solar radiation(Otterman 1974) Charney (1975) suggested a mechanism (lsquobiogeophysical feedbackrsquo) whereby over-grazing of desert margins can increase surface albedo decreasing the total energy absorbed at surface and reducing thermal convection thereby enhancing stabilityand reducing rainfall potential this in turn provides a positive feedback because lowermoisture availability leads to even less surface vegetation amounts Since much drylandrainfall comes from re-evaporated rainfall not from advected moisture from elsewhere
The archaeology of drylands 34
declining soil moisture may intensify drought conditions (Hulme 1989 Laval 1986)though the theory is disputed (Courel et al 1984 Idso 1977 Williams and Balling1994) A large artificial body of water such as Lake Nasser in Egypt does not increaserainfall in the Nubian desert despite its low albedo which extends over an area of 5000km2 and which contrasts with the very high reflectivity of the surrounding desert (LeHouerou 1996)
Bryson and Murray (1977) suggested that surface desiccation would lead to largeamounts of soil particles being entrained and then lifted aloft by the wind increasing theatmospheric albedo cooling the air aloft and causing it to subside warm adiabaticallyand form an inversion hence preventing convection and cloud development Theysuggested that this was illustrated by the Rajputana Desert on the borders of India andPakistan where extreme dustiness stifles rain processes even though atmospherichumidity is as high as in humid tropical forests
Large-scale climatic explanations for very dry episodes have recently focused onteleconnections Of importance are the possible impacts of anomalous patterns inenvironmental variables particularly sea surface temperature anomalies (SSTAs) whichinfluence the flux of moisture and sensible heat at the ocean-atmosphere interface at locations geographically remote from the region under investigation One well-known example is the association between dry episodes in the subtropics in the mid-twentieth century and El Nintildeo (ENSO) events Attempts have been made to link SSTAs in the tropical Atlantic to rainfall in the Sahel Owen and Ward (1989) have linked recurringSSTA patterns to notably wet and dry conditions in sub-Saharan Africa Another example of teleconnections is suggested by Gray (1990) who identified a positive associationbetween rainfall patterns in West Africa and the frequency of intense hurricanes reachingthe Atlantic coast of the United States During the period of drought in the western Sahel(1979ndash1987) there was a mean annual incidence of only fifteen hurricane days in the Atlantic basin compared with thirty per year in the wetter phase (1947 to 1969)
ENSO events in the Pacific have been seen to influence some drought events for
Table 26 The extent and severity of desertification (after Le Houerou 1996) Region Light Moderate Strong Severe Total area area area area area Africa 1180 9 1272 10 707 50 35 02 12860 560 Asia 1567 9 1701 10 430 30 5 01 16718 420 Australasia 836 13 24 4 11 02 4 01 6633 320 North America 134 2 588 8 73 01 0 00 7324 500 South America 418 8 311 6 62 02 0 00 5160 290 Total 4273 8 4703 9 1301 25 75 01 51691 397 Area desertified in 103km2 area desertified as of total drylands (where drylands = arid+semi arid+dry subhumid) Total drylands =percentage of desertified areas in the non hyper-arid drylands
The dynamic climatology of drylands 35
instance the strong 1982ndash1983 ENSO showed good correlations with drought inAustralia (Nicholls 1987) Indonesia (Malingreau 1987) and western South America(Serra 1987) In other areas the relationship was dubious or had very low statisticalsignificance for example northeast Brazil (Gasques and Magalhes 1987) and southernAfrica (Nicholson et al 1988) No relation exists between the present twenty-five years of drought in the Sahel and ENSO (Glantz 1987) although there is a clear link betweenthe drought and positive SSTAs in the Gulf of Guinea which are in turn related to theBenguela current There seems to be a South Atlantic Oscillation (SAO) comparable toENSO with many similarities between the Humboldt and Benguela currents theirupwellings and the generation of coastal deserts
In the Mediterranean basin the history of drought does not seem to be related to ENSO events ENSO events occur at regular intervals of about 64 years yet Mediterraneandroughts are totally acyclical and unpredictable especially in North Africa and the NearEast (Le Houerou 1996) A considerable amount of work has been carried out on thissubject (eg Folland et al 1986 Kane 1999 Kiladis and Diaz 1989) Trenberth (1993)describes El Nintildeo as having lsquodifferent flavoursrsquo Consequently finer classifications havebeen attempted Kane (1999) for instance has identified lsquounambiguousrsquo ENSO events in which the Tahiti-Darwin sea level pressure minima occur in the middle of the calendar year It is these events that have more impact on drought conditions elsewhere
Some General Circulation Models (GCMs) have predicted a slight increase in rainfall variability others a decrease some indicate an increase in winter rain and a decrease insummer precipitation others forecast the opposite (Williams and Balling 1994)Commonly the resolution for rainfall predictions is very coarse Le Houerou (1996)concludes that in view of the fact that there have been no significant observed trends inrainfall in any dryland area no change must be assumed for the not too distant future Incontrast to the predictions about rainfall however GCMs agree (at a 50 per centconfidence level) that the twenty-first century is likely to be characterized by an increasein temperature of 2ndash3degC in the subtropics and 1ndash2degC in the tropics Statistical analysis oftemperature and mean annual evapotraspiration (PET) shows that each degree oftemperature corresponds to 72 mm of PET a year using the Penman equation (LeHouerou 1996) A temperature rise of 1ndash3degC would therefore correspond to a PET increase of 72ndash232 mm a yearmdasha significant increase in climatic aridity Furthermore effects on the movement of the ITCZ and patterns in the westerlies will have an impacton the regime in semi-arid regions at the desert margin This increase in aridity can only enhance the current expansion of the dryland areas
CONCLUSION
This climatologically-based analysis has emphasized that drylands should not be seen as homogeneous entities with broadly similar environments Many different types ofdrylands can be recognized The reasons for the shortage of precipitation are many andcomplex The physical processes inherent in the maintenance of drylands involvesynergies and subtleties at a variety of time and spatial scales Similarly it is clear thatdifficult problems remain to be resolved before the magnitude and significance of human-
The archaeology of drylands 36
environmentmdashclimate interactions in drylands today can be fully elucidated let alone those in the distant past
REFERENCES
Agnew C and Anderson E (1992) Water Resources in the Arid Realm London Routledge
Barry RG and Chorley RJ (1998) Atmosphere Weather and Climate London Routledge
Barry RG and Perry RJ (1973) Synoptic Climatology Methods and Applications London Methuen
Beaumont P (1989) Drylands Environmental Management and Development London Routledge
Beaumont P Blake GH and Wagstaff JM (1988) The Middle East A Geographical Study London Fulton
Bryson RA and Murray TJ (1977) Climates of Hunger Mankind and the Worldrsquos Changing Weather Madison University of Wisconsin Press
Bullock P and Le Houreou P (1996) Land degradation and desertification In Climate Change 1995 Impacts Adaptations and Mitigations of Climate Change Scientific andTechnical Analysis 171ndash90 Cambridge Cambridge University PressIntergovernmental Panel of Climate Change
Charney J (1975) Dynamics of deserts and drought in the Sahel Quarterly Journal of The Royal Meteorological Society 101193ndash202
Cooke RU and Warren A (1973) Geomorphology in Deserts London Batsford Courel MF Kandel RS and Rasool SI (1984) Surface albedo and the Sahel drought
Nature 307528ndash31 El Dessouky TM and Jenkinson AF (1975) An Objective Daily Catalogue of Surface
Pressure Flow and Vorticity Indices for Egypt and its Use in Monthly RainfallForecasting Bracknell Meteorological Office Synoptic Climatology Branch Memorandum 46
Findlater J (1972) Aerial explorations of the low level cross equatorial current overeastern Africa Quarterly Journal of the Royal Meteorological Society 98274ndash89
Folland CKPalmer TN and Parker DE (1986) Sahel rainfall and worldwide seasurface temperatures Nature 320602ndash7
Gasques JG and Magalhes AR (1987) Climatic anomalies and their impact in Brazilduring the 1982ndash83 ENSO event In MGlantz RKatz and MKrenz (eds) The Societal Impacts Associated with the 1982ndash83 Worldwide Climate Anomalies 30ndash6 Boulder CO National Center for Atmospheric Research
Glantz M (1987) Impacts of the 1982ndash83 climate anomalies in the West African Sahel In MGlantz RKatz and MKrenz (eds) The Societal Impacts Associated with the 1982ndash83 Worldwide Climate Anomalies 62ndash4 Boulder CO National Center forAtmospheric Research
Gordon AH and Lockwood JG (1970) Maximum one day falls of precipitation inTehran Weather 252ndash8
The dynamic climatology of drylands 37
Gray WM (1990) Strong association between West African rainfall and US landfall ofintense hurricanes Science 2491251ndash6
Hastenrath S (1985) Climate and Circulation of the Tropics Dordrecht DReidel Hills ES (1966) Arid Lands London Methuen Hulme M (1989) Is environmental degradation causing drought in the Sahel An
assessment from recent empirical research Geography 7438ndash46 Hulme M and Kelly M (1993) Exploring links between desertification and climate
change Environment 354ndash11 39ndash45 Idso SB (1977) A note on some recently proposed mechanisms of the genesis of
deserts Quarterly Journal of the Royal Meteorological Society 103369ndash70 Inoue M and Bigg GR (1995) Trends in wind and sea level pressure in the tropical
Pacific Ocean for the period 1950ndash1979 International Journal of Climatology 15 35ndash52
Jones JAA (1997) Global Hydrology Processes Resources and Environmental Management Harlow Longman
Jones PD (1991) Southern hemisphere sea level pressure data an analysis andreconstruction back to 1951 and 1911 International Journal of Climatology 11 585ndash608
Kamara SI (1986) The origins and types of rainfall in West Africa Weather 41 48ndash56 Kane RP (1999) Rainfall extremes in some selected parts of central and South America
ENSO and other relationships re-examined International Journal of Climatology19423ndash55
Kiladis GN and Diaz HF (1989) Global climatic anomalies associated with extremesof the Southern Oscillation Journal of Climate 21069ndash90
Kodama Y (1992) Large scale common features of subtropical precipitation zones (theBaiu Front The South Pacific Convergence Zone the South Atlantic Convergencezone) Part 1mdashCharacteristics of the Subtropical frontal zones Journal of the Meteorological Society of Japan 70813ndash35
Kodama Y (1993) Large scale common features of subtropical precipitation zones (theBaiu Front The South Pacific Convergence Zone the South Atlantic Convergencezone) Part IImdashConditions for generating the subtropical convergence zones Journal of the Meteorological Society of Japan 71581ndash610
Koppen W (1931) Die Klimate der Erde Berlin Lamb HH (1982) Climate History and the Modern World London Routledge Laval K (1986) General circulation model experiments with surface albedo change
Climatic Change 991ndash102 Le Houerou HN (1977) Biological recovery vs desertization Economic Geography
53413ndash20 Le Houerou HN (1996) Climate change drought and desertification Journal of Arid
Environments 34133ndash85 Le Houerou HN Popov GF and See L (1993) Agrobioclimatic Classification of
Africa Rome Food and Agriculture Organization Agrometeorology Series Working Paper No 6
Lockwood JG (1974) World Climatology An Environmental Approach London Edward Arnold
The archaeology of drylands 38
Malingreau VP (1987) The 1982ndash83 drought in Indonesia Assessment and monitoring In MGlantz RKatz and MKrenz (eds) The Societal Impacts Associated with the 1982ndash83 Worldwide Climate Anomalies 11ndash18 Boulder CO National Center forAtmospheric Research
Mather JR (1974) Climatology Fundamentals and Applications New York McGraw-Hill New York
McGinnies WG (1988) Climatic and biological conditions of arid lands a comparisonIn EEWhitehead CFHutchinson BNTimmerman and RGVardy (eds) Arid Lands Today and Tomorrow 61ndash8 Boulder CO Westview Press
McGregor GR and Nieuwolt S (1998) Tropical Climatology Chichester John Wiley and Sons second edition
McIlveen R (1992) Fundamentals of Weather and Climate London Chapman Hall Meigs P (1953) World distribution of arid and semiarid homoclimates UNESCO Arid
Zone Program 1203ndash10 Middleton NJ (1991) Desertification Oxford Oxford University Press Morel R (1992) Atlas Agroclimatique de Pays de la Zone de CILSS Niamey
AGRHYMET Musk LF (1988) Weather Systems Cambridge Cambridge University Press Nicholls N (1987) The El NineSouthern Oscillation phenomenon In MGlantz RKatz
and MKrenz (eds) The Societal Impacts Associated with the 1982ndash83 Worldwide Climate Anomalies 2ndash10 Boulder CO National Center for Atmospheric Research
Nicholson SE (1993) An overview of African rainfall fluctuations of the last decadesJournal of Climate 61463ndash6
Nicholson SE Jeeyong K and Hoopingarner J (1988) Atlas of African Rainfall and its Annual Variability Tallahassee Florida State University
Nir D (1974) The Semi-Arid World London Longman Oliver JE (1973) Climate and Manrsquos Environment Chichester John Wiley and Sons Oliver JE (1981) Climatology Selected Applications London Edward Arnold Otterman J (1974) Baring high albedo soils by over-grazing Science 86531ndash3 Owen JA and Ward MN (1989) Forecasting Sahel rainfall Weather 4457ndash64 Palmen E and Newton CW (1969) Atmospheric Circulation Systems New York
Academic Press Pearce EA and Smith CG (1984) World Weather Guide London Hutchinson Penman H (1948) Natural evaporation from open water bare soil and grass
Proceedings of the Royal Society A193120ndash45 Rossby CG (1947) On the general circulation of the atmosphere in the middle latitudes
Bulletin of the American Meteorological Society 28255ndash80 Serra RB (1987) Impact of the 1982ndash83 ENSO on the southeastern Pacific fisheries
with emphasis on Chilean fisheries In MGlantz RKatz and MKrenz (eds) The Societal Impacts Associated with the 1982ndash83 Worldwide Climate Anomalies 24ndash9 Boulder CO National Center for Atmospheric Research
Sharon D (1972) The spottiness of rainfall in a desert area Journal of Hydrology 17 161ndash75
Sharon D (1981) The distribution in space of local rainfall in the Namib desert Journal of Climatology 169ndash75
The dynamic climatology of drylands 39
Smith K (1992) Environmental Hazards Assessing Risk and Reducing Disaster London Routledge
Soliman KH (1953) Rainfall over Egypt Quarterly Journal of the Royal Meteorology79389ndash401
Spellman G (2000) An objective weather type method for the Iberian peninsulaWeather (in press)
Sweeney JC and OrsquoHare GP (1992) Geographical variations in the precipitationyields and circulation types in Britain and Ireland Transactions of the Institute of British Geographers 17448ndash63
Thomas DG (1989) (ed) Arid Zone Geomorphology London Bellhaven Press Thomas DG and Middleton NJ (1994) Desertification Exploding the Myth
Chichester John Wiley and Sons Thompson RD (1975) The Climatology of the Arid World Reading University of
Reading Department of Geography Paper No 35 Thornthwaite CW (1948) An approach towards a rational classification of climate
Geographical Review 3855ndash94 Trenberth KE (1993) The different flavours of El Nintildeo Proceedings of the 18th Annual
Climate Diagnostics Workshop 50ndash3 Boulder CO National Center for AtmosphericResearch
UNEP (1992) World Atlas of Desertification Nairobi UNEP and London Edward Arnold
UNESCO (1977) Map of the World Distribution of Arid Regions Man and Biosphere Paris Technical Note 7
Waliser DE and Gautier C (1993) A satellite-derived climatology of the ITCZ Journal of Climate 6 2162ndash74
Wallen CC (1967) Aridity definitions and their applicability Geografiska Annaler 49a 367ndash84
Wilby RL and Wigley TML (1997) Downscaling general circulation model output areview of methods and limitations Progress in Physical Geography 21 530ndash48
Williams MAJ and Balling RC (1994) Interactions of Desertification and Climate Geneva World Meteorological Organization
World Meteorological Organization (1975) Drought in Agriculture Technical Note No 138 Geneva World Meteorological Organization
Yair A and Berkowicz SM (1989) Climatic and non-climatic controls of aridity the case of the northern Negev of Israel Catena Supplement 14 Arid and Semi Arid Environments
The archaeology of drylands 40
Part II SOUTHWEST AND CENTRAL
ASIA
3 The decline of desert agriculture a view from the
classical period Negev STEVEN AROSEN
INTRODUCTION
The presence of sophisticated large-scale agricultural systems dating to classical timesin the arid regions of the central Negev southern Jordan and Sinai has long served bothto illustrate the ingenuity of the ancient peoples of the region and as an inspiration tomodern peoples as to the potential of wise exploitation of the desert Archaeologicalsurvey has demonstrated that agriculture was practised throughout the Irano-Turanian desert steppe zone in areas that today receive as little as 75 mm average annual rainfall(compare Evenari et al 198232 fig 13 to Kedar 1967) Virtually every wadi worthy of the name shows terrace systems for the damming of flash-floods and their exploitation for farming (Fig 31)
The amazing efficacy of these systems has been repeatedly demonstrated Both texts (Bruins 198687 Kraemer 1958 Document 82 Mayerson 1960224ndash69) and experimental archaeology (Evenari et al 1982191ndash219) have indicated that yields from the desert zone using run-off water catchment systems could in fact approximate thoseof the Mediterranean zone (Bruins 198687 Evenari et al 1982191ndash219) Excavations and surveys have revealed the existence of large and numerous wine presses (Mazor1981 Rubin 199654 Shershefski 1991198ndash200 Fig 32) suggesting industriallevel production of grapes and wine The reconstruction and operation of some of thesesystems over several decades have demonstrated that in some ways they constitute anagricultural regime more resistant to drought than their counterparts in the better-watered areas farther north Finally in the central Negev there were six towns which togetherwith their village and homestead hinterland comprised an urban system proper with apopulation of over 20000 people whose subsistence was based on this agriculturalregime (Broshi 1979 Elliot 1982103ndash14 Shershefski 1991200ndash14 Fig 34)
In the light of the impressive nature of these systems their decline is all the more marked By the tenth or eleventh centuries AD the entire settlement system of the centralNegev had been abandoned All previously occupied
Figure 31 Terraced dam system in the upper reaches of Nahal Nizzana in the central Negev
Note The terrace dams are marked by the lines of vegetation across the wadi bed the system of dams extends for several km along this stretch of the wadi Photograph SRosen
sites including towns villages farmsteads and even nomadic encampments had beendeserted and there is no evidence for any alternative settlements either permanent ornomadic (for example Avni 1996 Nachlieli 1992 Rosen 1987a Rosen and Avni1993) The desert had reverted to desert
The stark contrast between the rich archaeological remains and the contemporary desolation has struck every traveller through the region (Fig 33) and there has been no shortage of attempts to explain this apparent lsquotriumph of the desertrsquo Two general factors have been suggested as primary causes for Negev desertification (1) the Moslem or Arabconquests and the ensuing destruction of Byzantine civilization (for exampleLowdermilk 1945136 Negev 198815 Palmer 1872243 Reifenberg 195598Sharon 1969) and (2) climatic deterioration rendering habitation impossible due toshifting sands increased erosion and reduced water for agriculture (for exampleHuntington 1911 Issar 1995 Issar and Govrin 1991) Additional subfactors haveincluded the negative effects of over-grazing by the flocks and herds of bedouin(Reifenberg 195598) the destructive effects of earthquakes (Fabian 1994) andincreased marauding by nomads (Sharon 1976 cf also Lowdermilk 1945129)
Critical examination of these factors in the light of recent intensive archaeologicalresearch carried out in the Negev indicates that each of these explanations isfundamentally flawed as a prime mover in the desertification
The decline of desert agriculture a view from the classical period Negev 45
Figure 32 The wine press at Shivta (Subeita) Note This is a relatively small press located in one of the central squares of the town The actual pressing floor is the square area in the background while the collecting and settling vat is in the foreground An intermediate settling or filtering area is poorly preserved located to the left of the vat Photograph SRosen
of the Negev although each plays a role within a larger perspective The key issue rarelydiscussed in reviews of the decline of classical civilization in the Negev is that periods ofcultural florescence can usually be tied to increased economic and social input from orintegration with the Mediterranean core area The collapse of the economic core willinevitably result in the collapse of its dependants unless alternative economic paths areavailable
ARCHAEOLOGICAL OVERVIEW
The central Negev in the sixth century AD the Byzantine period in local terms was thewell-integrated frontier province of Palestina Tertia of the Late Roman empire (Mayerson 1994 Rubin 1997 Shershefski 1991 also see Isaac 1992) Although thelucrative trade route of the Nabatean period had long since been eclipsed by alternativetrade systems (Crone 1987 Negev 1988) the province functioned both as a strategicsouthern buffer zone protecting the Levantine heartland (Gihon 1980 Mayerson 19861990 also Isaac 1992 for a differing view) and as a gateway to both the holy pilgrimagedestinations of the Sinai and to the mineral-rich desert regions farther east and south (Mayerson 1982 1983)
The archaeology of drylands 46
Figure 33 Palmerrsquos pen-and-ink sketch of the Byzantine town of Shivta (Esbeita or Subeita) in the central Negev showing the rich archaeological remains amidst the desert environment
Source After Palmer 1872314
Archaeologically the region is marked by two complementary settlement systems(Avni 1996 Elliott 1982 Haiman 1995a Mayerson 1989 Negev 1988 Rubin 1990Rosen 1987b Rosen and Avni 1993 Shersehfski 1991 Fig 34) First in the north and in the higher mountains both better watered than areas farther south large towns such asAvdat (Fig 35) supported by intensive run-off agricultural systems (Fig 36) evolved out of the Limes Palestina and the preceding Nabatean caravanserai over the course of several centuries By the sixth century AD the six towns of Elusa (modern Haluza)Ruheiba (Rehovot) Subeita (Shivta) Nessana (Nizzana) Oboda (Avdat) and Mampsis(Mamshit or Kurnub) represent the expansion of Byzantine society and economy deepinto the desert The design and construction of these towns are dominated by anarchitecture whose roots are undeniably in the Mediterranean zone with little adjustmentfor local conditions (Shershefski 1991228) excepting the use of local raw materials(Negev 1980) Christianity is the only religion represented at these sites in this pre-Islamic period and classic basilica-style churches are present in the plural at each townThe wealth of the towns is especially evident in these churches which showed suchfeatures as wall facings and furniture of marble imported from Anatolia elaboratemosaics and vaults of large wooden beams imported from the Mediterranean zone(Negev 1974)
The decline of desert agriculture a view from the classical period Negev 47
Figure 34 Map of the general settlement system of the central Negev during the Late Byzantine and Early Islamic periods
Key urban zone with agricultural support = village and farmstead agricultural hinterland pastoral nomadic region lacking evidence for agricultural exploitation | | | agro-pastoral region showing combination of pastoral sites with agricultural exploitation The major cities were Elusa (modern Haluza) Ruheiba (Rehovot) Subeita (Shivta) Nessana (Nizzana) Oboda (Avdat) and Mampsis (Mamshit or Kurnub) For detailed discussion see Rosen and Avni (1993)
The archaeology of drylands 48
Figure 35 View of the Byzantine town of Avdat (looking north) Note The left edge of the cliff shows the remains of churches and a late Nabatean temple and a Byzantine fortress are located to the right of this The domestic quarter is located on the slopes and to the right (foreground) Photograph SRosen
Although defensively postured defence does not seem to have been a primaryconsideration in the settlements Aside from the isolated nature of many of the villagesand farmsteads only Mampsis shows a circumference wall although Avdat shows afortification wall on one side of the settlement Neither is especially massive Both Avdatand Nessana show internal forts indicating military presence Subeita presents a limitednumber of access gates to the town but these gates are in fact breaks in the continuum ofattached structures and not the gates of a city wall (Shershefski 1991184ndash8)
The agricultural systems surrounding these towns both those in direct association with the towns and those that were part of the village-farmstead hinterland are perhaps the most impressive evidence of the wealth and long-term stability of the Byzantine regime (Bruins 1986 Evenari et al 1982 Kedar 1967 Mayerson 1960) Vast areas of both wadi floodplain and upper alluvial terraces show elaborate systems of terraced damsdrainage channels sluice gates and support walls Hill slopes are covered with tuleiliot el anabmdashrows of stone mounds and stone linesmdashwhose function was presumably connected to either ground clearance for run-off enhancement or some other form of agricultural activity (Evenari et al 1982127ndash47) Calculations based on aerial photography pedestrian survey and farm reconstruction demonstrate that the averageratio of drainage catchment to farmed area was
The decline of desert agriculture a view from the classical period Negev 49
Figure 36 Elaborate raised field and dam system on Nahal La van just south of Shivta (Subeita)
Note Notice the wadi bed to the right of the fields water flow was drained onto the raised fields several km upstream Photograph SRosen
approximately 211 so that with run-off estimated at 15 per cent of actual rainfall an average annual rainfall of 100 mm could be transformed to an effective annual rainfall forthe farmed fields of more than 400 mm (Evenari et al 198295ndash119) Not only is this more than sufficient for growing barley and wheat (the basic cereal staples of the period)but it sufficed for growing grapes and olives as well The presence of olive and winepresses at each town sometimes at an industrial scale demonstrates clearly the practiceof arboriculture and viticulture dates figs and even pomegranates were also grown(Mayerson 1960 Mazor 1981 Rubin 1996) Rubin (1996) characterizes this system asthe adoption of the Mediterranean agricultural system into the Negev
The second system which is less well documented than Palestina Tertia is that of the pastoral hinterland located in the deserts beyond the village-farming hinterland (Avni 1996 Haiman 1995a Rosen 1987b 1994 Rosen and Avni 1993) Aside from thesignificantly lower rainfall associated with these southern areas the region is marked bythe general scarcity of agricultural remains and the presence of the larger-scale pastoral encampments The remains of pastoral encampments are found throughout the desert andsteppe zones but the larger aggregate camps are located only south of the agriculturalareas These camps are obviously smaller than the Byzantine towns and villages but they also differ in their basic architecture and organization In essence the structuresrevealed at such encampments are to be interpreted as ephemeral tent bases or in somecases as hut foundations that carried brush or tent superstructures The settlements alignalong secondary and tertiary drainages in patterns dictated by topography Analyses of
The archaeology of drylands 50
material culture also support the interpretation of these settlements as basicallypastoralnomadic (Rosen and Avni 199762ndash81) and textual references (Mayerson 1989)accord well with this A key point in analyses of these pastoral systems is their essentialdependence on the settled system to the north for both their subsistence and their materialculture The markets of the settled zone were a sine qua non for pastoral existence in the desert (cf Khazanov 1984) Relations between the desert and the sown while perhapsoccasionally tense must have been essentially stable for the nomadic system to havethrived
In summary when the agricultural exploitation of the desert was at its peak in the classical period the region had been well integrated into the Roman-Byzantine (and later Ummayad) empire (Rubin 1996 1997) That integration in essence established theeconomic and social stability that enabled the desert to bloom
THE ISLAMIC CONQUESTS AS CAUSE FOR DESERTIFICATION
The battle of Gaza in AD 633ndash4 marks the beginning of the political end of the Byzantine empire in the Negev Although the events leading up to that battle and thecauses behind the Byzantine collapse have been much discussed and are beyond thescope of this paper in terms of desertification several important points require attention
Archaeologically there is no evidence for the destruction or violent conquest of any ofthe Negev towns (per contra Negev 198815) In fact the processes of urban declineseem to have been initiated well before the Islamic period Mampsis (Negev 198815)does not appear to show an occupation in the seventh century at all Avdat showsevidence for a major earthquake at the beginning of the seventh century after which thecity seems to have been abandoned for two centuries and eventually reoccupied duringthe Islamic period (Fabian 1994) Significantly an earlier fourth- or fifth-century earthquake resulted in repairs and lsquoretro-fittingrsquo of various structures against furtherearthquake damage Nessana shows continued occupation at least into the late seventhcentury and probably well into the eighth both in the archaeology (Shershefski 1991550) and in the archives recovered from the site (Kraemer 1958213) with little obviousdisruption although a clear decline can be traced At Subeita the presence of a mosquewedged into an open space next to a church (Baly 1935 Segal 1983 Shershefski199174) indicates both clear continuity of occupation well into the eighth century and itscontemporaneity with at least one church on the site indicating the peaceful coexistence of the two religions during the Ummayad period Ruheiba (Shershefski 199195) alsoseems to show continued occupation into the early Islamic period Recent excavations atElusa have not revealed any evidence for Islamic occupation but nor is there anyevidence for destruction (although it must be admitted that the areas excavated are stillquite limited) The excavator (Goldfus pers comm) suggests an abandonment prior to the Islamic period The city of Beer-Sheba (Figueras 1979) in the northern Negev seems to show archaeological decline as well although again with no evidence for eitherabandonment or destruction
In this context it is important to recognize that the first decades of the seventh century were catastrophic for the Byzantine empire as a result of its long wars with the
The decline of desert agriculture a view from the classical period Negev 51
Sassanids Although it is unlikely that the Persian armies that devastated the Levantactually came as far south as the central Negev the havoc wreaked on the Mediterraneanheartland could not but have been felt on the periphery as well
On the other hand in spite of the decline marked in the cities the Ummayad and earlyAbassid periods seem to show a rural florescence The central village and satellite farmsat Sede Boqer (Nevo 1985 1991 Fig 37) are the best example of this phenomenonAnother example is the farmstead at Nahal Mitan (Haiman 1995b) Avni (1994) hasindicated the presence of at least thirteen mosques in the Negev highlands in this periodsome of
Figure 37 The early Islamic village of Sede Boqer in the central Negev
Note The site is surrounded by numerous agricultural terraced field systems not pictured Photograph SRosen
which are clearly associated with farming settlements and others with pastoralencampments Finally a series of large homesteads that were colonized during theUmmayad and early Abassid periods has been excavated recently around the outskirts ofBeer-Sheba (for example Bar-Ziv and Katz 1993 Gilead et al 1993 Katz 1993 Katz and May 1996 Nachshoni et al 1993 Negev 1993)
Evidence from the nomadic periphery also shows continuity with little evidence for destruction or invasion Pastoral settlements dating to the eighth and perhaps ninthcenturies AD have been excavated in the southern central Negev (Rosen and Avni 1997)Some of these in the higher areas seem to show the adoption of floodwater farming intothe pastoral subsistence system (Rosen and Avni 1993) The continued import and use of
The archaeology of drylands 52
typologically Byzantine ceramics (and other elements of material culture) from thesettled regions into the pastoral sites demonstrate underlying economic continuitiesThere was no break in relations between the nomads and the farmers in the transition tothe Ummayad administration Importantly there is no incursion of nomadic settlementtypes into the agricultural zones in this period Although erosion is a dominant feature inthe desert landscapes today it cannot be linked to the over-grazing that is often tied to such pastoral incursions since there is no evidence of such incursions
In short the Islamic conquestsmdasha problematic concept in itself for the Negevmdashdid not bring any desertification Whilst the late Byzantine period saw an urban decline in theNegev the early Islamic period seems by contrast to have seen a rural renaissance
CLIMATIC DETERIORATION AS CAUSE FOR DESERTIFICATION
Establishing climatic change as a prime factor in cultural transformation requires threedistinct steps First one must establish the reality of the climatic change itself Secondthe suggested climatic change must be correlated chronologically with the culturaltransformation Third a reasonable scenario or mechanism for causality must beestablished beyond the mere fact of correlation it is not enough to establish a climaticchange indicate a contemporaneity with a cultural change and then claim a causal link
There are several lines of evidence suggesting a change in climate some time following the classical period settlements The most obvious of these is the deposition of extensiveterraces sometime during the classical period (Bruins 1986189 Goldberg 1994)followed by their erosion and wadi downcutting (Ben-David 1997 Bruins 1986189 Reifenberg 1955) It is clear that there has been landscape degradation but it is not cleareither when this degradation occurred or whether it was the result of climatic changes orof other factors such as microtectonics or human intervention Certainly acceleratederosion can be expected if terrace systems are not maintained (cf Butzer 1974) and some of the gullying that can be seen in the Negev today is the undoubted result ofbreached dam systems and not climatic change
One possible indication of a climatic component is the existence of post-classical downcutting in areas where agriculture or its abandonment can be discounted as afactor The pastoral encampment of Nahal Oded (Fig 34) south of the Ramon Crater shows two post-classical wadi channels one a modern one and the other an earlier and somewhat higher one that cuts several eighth-century structures located on higher alluvial terraces In the absence of any agriculture in the area Ben-David (1997) suggests that these downcutting events reflect episodes of extreme aridity both of which post-date the Ummayad-period occupation of the site
The infiltration and movement of sand dunes blocking drainages and burying settlements have also been suggested as reflective of climatic deterioration Issar (1995)claims that the burial of the Byzantine towns of Elusa and Ruheiba beginning cAD 800 is the result of an increased supply of Nile sands on the Levantine littoral to be correlatedwith increased monsoon rains in East Africa
Dead Sea water levels as established from salt cave evolution and analysis of coresediments have also been used to reconstruct climatic sequences (Frumkin et al 1991
The decline of desert agriculture a view from the classical period Negev 53
1994 Geyh 1994 Issar 1995 Neev and Emery 199562) Summarized briefly higherDead Sea levels are evident during the first two centuries AD (the early Roman period)indicating greater humidity This period was followed by a warmer more arid period inthe middle of the first millennium BC that was not ameliorated until the beginning of thesecond millennium AD
Analyses of oxygen isotope ratios from cave speleothems and marine molluscs (Gatand Magaritz 1980 Geyh 1994) show high 18O ratio peaks of c2300 BP and c1500 BP indicating cooler temperatures (and presumably higher humidity) with cooler (andmoister) periods between and following These analyses accord well with the studies ofthe Dead Sea water level Of further interest is the apparently significantly warmer (anddrier) period prior to 2300 BP so that although it was not especially cool or moist on anyabsolute scale the c2300 BP episode is a relatively significant amelioration Laterepisodes do not approach this first in the scale of change
Given the above data from different sets of evidence it is hard to argue that climateremained stable during the first millennium AD (per contra Rubin 1989) The next issues are whether the climatic fluctuations outlined above do indeed correspond withand can explain the cultural and physical desertification of the Negev
The weakest link in the argument is that of dating since shifts of a few hundred years quite within the range of radiocarbon errors given problems of fractionation intrusionand so on significantly affect historical interpretation (Gat and Magaritz 1980)However given current dating of the climatic events it is hard to reconcile them with thedesertification of the Negev Thus the Nabatean and early Roman periods in the finalcenturies BC and first two centuries AD when agriculture was incipient at best (Bruins1986189 Mayerson 1963) seem to have been at a climatic optimum whereas thecultural peak in the succeeding Byzantine period seems to have been climatically dryThe Byzantine collapse and rise of the early Islamic empire seem to have been eitherstable climatically or marked by only minor fluctuations Although sand dunes did indeedbury those cities built in the dune areas Goldfus (pers comm) suggests that Elusa was in fact abandoned relatively early prior to the eighth-century dune invasions claimed by Issar (1995) Notably Avdat Subeita and Mampsis were not affected by dunes at all It isimportant to stress here that the gradual abandonment of the Byzantine cities is notequivalent to either the abandonment of the Negev or desertification for as indicatedearlier there is a significant Early Islamic agricultural presence in the Negev at least until the ninth or tenth centuries AD The final abandonment of the central Negevprobably in the tenth or perhaps even eleventh centuries AD may in fact even beassociated with the beginning of climatic amelioration In short climatic change does notadequately explain the decline of classical civilization in the desert or the reversion of thedesert to desert
THE RISE OF THE DESERT
To understand the rise of the desert we must understand first its domestication Theessence of the classical period lsquoGreen Revolutionrsquo in the Negev was the transplantation of a Mediterranean-zone agricultural complex into the arid zone This complex in the
The archaeology of drylands 54
Mediterranean zone consists of cereal (wheat and barley) farming the cultivation of fruitcrops including grapes olives figs and dates and animal husbandry based especially onsheep goat and cattle with significantly less emphasis on pig Landscape managementin the form of hill slope terracing and various forms of irrigation is integral to thecomplex as well (Grigg 1974123ndash8 132ndash4 Stager 1985 compare also with Braudel 197256 59 423) Despite claims concerning the inappropriateness and instability ofMediterranean-zone farming systems in the New World and other non-Mediterranean environments (see Butzer 1996) the expansion of the Mediterranean zone into thedesert in terms of culture society and subsistence in fact proved a remarkably stablephenomenon enduring for at least half a millennium The stability of this system is evenmore marked given the political perturbations that occurred during this periodperturbations that included the rise and decline of urban centres the rise of Christianitythe collapse of Byzantine administration and the rise of Islam
Two points are particularly relevant for comprehending the success of the Mediterranean system The first is the integration of the desert economy both in terms oftrade and subsistence into the larger state Even beyond the fact of active imperialsubsidy that the desert settlement system was well embedded in the classical world is reflected in virtually all aspects of material culture economy and society The second isthat the Mediterranean economy itself should be seen as a flexible strategy fluctuatingbetween emphasis on cash crops and subsistence staples depending on the historical andeconomic contexts Within the Mediterranean zone during periods of social collapse thecomplex shifts towards subsistence mode whereas during times of economic prosperitycash crops play a larger role (Stager 1985)
In the desert zone the subsistence mode may be insufficient by itself especially given large urban populations that were at least partially supported by trade Even withoutreference to climatic change the desert environment exerts pressures on settlementsystems not felt in better watered areas Thus regardless of the effectiveness of run-off irrigation systems agriculture in the Negev must have required significantly more labourinput than further north for example in the construction and constant maintenance ofterrace dam systems Subsistence is more difficult in a desert and therefore the raison drsquoecirctre of permanent settlement in the Negev has always been its integration with someother core region The decline of the core-region economy resulted in a reversion towardsthe subsistence end of the Mediterranean complex spectrum one that may not have beensustainable in the desert at the high population levels there of village and urban society
The Mediterranean complex continued well into the early Islamic period In this context it is important to understand that the early Islamic horizon in the Negev in spiteof its rural character still shows a high degree of social and economic integration withthe Mediterranean core area This is most obvious in the material culture continuitiesbetween the core and the periphery However it is especially impressive in theideological integration such that Negev rock inscriptions from this period follow verystandardized Islamic formulae (Sharon 1990) burials are typically Moslem (Rosen andAvni 199713) and mosques follow standard definitions (Avni 1994) It was in the laterAbbassid period following the political and economic shift of the Caliphate to Baghdadthat the Levant itself declined With that decline the means for the integration of thedesert and the sown were no longer available and the entire desert system both settled
The decline of desert agriculture a view from the classical period Negev 55
and nomadic was abandoned Glantz (1994) defines desertification as the creation of unproductive desertlike
landscape in a place where none had existed in the recent past In the sense that thecentral Negev reverted from being a productive and integrated component of aMediterranean state system to its original desert state the processes reviewed here areindeed those of desertification
FINAL NOTE
The history of research on the rise of Near Eastern deserts is one inextricably tied to thepolitical and ideological struggles of the region For example the nineteenth-century British Orientalist Edward Palmer (1872241ndash3) viewed the decline of civilization and the rise of the desert as the result of invasion and indigence on the part of the localinhabitants Ellsworth Huntingdonrsquos (1911) environmental determinism in which heclaimed that settlement and the rise and decline of civilization were dictated by thecarrying capacity of a region in turn determined by climate and environment was inantithesis to attitudes like Palmerrsquos it was adopted as state policy by the British Foreign Office in its administration of Palestine and used as a rationale for limiting Jewishimmigration In response Zionist ideologues such as Ben-Gurion and Ben-Zvi (1979 [1918]) claimed that the decline of Palestine and the rise of the desert were the result ofnegligent administration discounting the role of climatic change (Troen 1989) IndeedBen-Gurion (1961) idealized the rebirth of the desert As a part of the scientific background to the Zionist vision of the blooming desert the role of the black goat as afactor in the reduction of vegetation and in the consequent rise in erosion has often beenstressed (Orni and Efrat 1980470 Reifenberg 195598 see also Kohler-Rollefson 1992 for a claim for destructive over-grazing in the Neolithic) thus legitimising expropriation of bedouin grazing lands In response some scholars have deniedtraditional pastoral nomadism and grazing as a significant factor in landscape degradation(Thomas and Middleton 199413 67ndash73)
Desertification is the result of a complex chain of causality On its most simple level itis clear that land degradation is the result of physical processes However these physicalprocesses are often set in motion by human activities (Glantz 1994) such that the issuesare social and historical (Blaikie and Brookfield 1987) The historical causalities are alsocomplex it is accepted knowledge that over-grazing by pastoralists causes erosion (Orniand Efrat 1974470 Reifenberg 195598) but Danin (198317) notes that lsquoduring the few years that several Negev and Sinai areas were closed to bedouin and their domesticanimals no substantial changes in the list of species and plant communities could bediscernedrsquo As noted above land degradation as a consequence of over-grazing may be only the latest stage in desertification In the circum-Mediterranean region the steppe zones inhabited by bedouin in the nineteenth and early twentieth centuries were almostall exploited for agriculture during the classical period and subsequently abandoned tobe exploited by pastoralists only later The simplistic notions of Islamic invasion orclimatic catastrophe as prime causes in the decline of the Negev in fact mask politicalagendas It is the historical complexities in all their richness and texture that need to be
The archaeology of drylands 56
addressed before we can critically understand desertification as a social phenomenon
ACKNOWLEDGEMENTS
I am grateful to Graeme Barker and David Gilbertson for the opportunity to participate inthe symposium on the Archaeology of Drylands at the Fourth World Archaeological Congress and to the other participants for their stimulating and eye-opening papers Haim Goldfus was good enough to read an early draft of this paper and make valuablecomments Arlene Miller Rosen shared her knowledge of climate and climate-change freely and happily The photographs were developed from slides by Alter Fogel and themap was drafted by Patrice Kaminsky both of Ben-Gurion University
REFERENCES
Avni G (1994) Early mosques in the Negev highlandsmdashnew archaeological evidence on Islamic penetration of southern Palestine Bulletin of the American Schools of Oriental Research 29483ndash100
Avni G (1996) Nomads Farmers and Town-Dwellers Jerusalem Israel Antiquities Authority
Baly C (1935) Sbaita Palestine Exploration Fund Quarterly Statement 62171ndash81 Bar-Ziv H and Katz O (1993) Beer ShevamdashNahal Ashan Archaeological News 100
100ndash01 (In Hebrew) Ben-David R (1997) The geology and geomorphology of the Nahal lsquoOded site area In
SARosen and GAvni The lsquoOded Sites Investigations of Two Early Islamic Pastoral Camps South of the Ramon Crater 109ndash18 Beersheva Ben-Gurion University
Press Beersheva XI Studies by the Department of Bible and Ancient Near East Ben-Gurion D (1961) Introduction In YMorris Masters of the Desert 11ndash20 New
York Putnams Ben-Gurion D and Ben-Zvi Y (1979 [1918]) Eretz Israel in Past and Present
Jerusalem Yad Ben Zvi Press (Translated from Yiddish to Hebrew by D Niv) Blaikie P and Brookfield H (1987) Defining and debating the problem In PBlaikie
and HBrookfield (eds) Land Degradation and Society 1ndash7 London Methuen Braudel F (1972) The Mediterranean and the Mediterranean World in the Age of Philip
II New York Harper Broshi M (1979) The population of western Palestine in the Roman-Byzantine period
Bulletin of the American Schools of Oriental Research 2361ndash10 Bruins HJ (1986) Desert Environment and Agriculture in the Central Negev and
Kadesh Barnea during Historical Times Nijkirk Netherlands Midbar Foundation Butzer KW (1974) Accelerated soil erosion a problem of man-land relationships In
IRManners and MWMikesell (eds) Perspectives on Environment 57ndash77 Washington DC Association of American Geographers
Butzer KW (1996) Ecology in the long view settlement histories agrosystemicstrategies and ecological performance Journal of Field Archaeology 23141ndash150
The decline of desert agriculture a view from the classical period Negev 57
Crone P (1987) Meccan Trade and the Rise of Islam Princeton Princeton University Press
Danin A (1983) Desert Vegetation of Israel and Sinai Jerusalem Cana Elliott Jack D Jr (1982) The Elusa Oikoumene A Geographical Analysis of Ancient
Desert Ecosystem Based on Archaeological Evironmental Ethnographic and HistoricData Mississippi State Mississippi State University Cobb Institute of ArchaeologyOccasional Papers 82ndash01
Evenari M Shanan L and Tadmor N (1982) The Negev The Challenge of a Desert Cambridge Mass Harvard University Press
Fabian P (1994) New evidence for earthquakes and their influence on the urbandevelopment of Avdat Paper presented at the 20th Archaeological Congress inJerusalem Israel
Figueras P (1979) The Roman-Byzantine period In YGrades and EStern (eds) Beersheva 39ndash52 Jerusalem Keter (In Hebrew)
Frumkin A Magaritz M Carmi I and Zak I (1991) The Holocene climatic record ofthe salt caves of Mount Sedom Israel The Holocene 1191ndash200
Frumkin A Carmi I Zak I and Magaritz M (1994) Middle Holocene environmentalchange determined from the salt caves of Mount Sedom Israel In O Bar-Yosef and RKra (eds) Late Quaternary Chronology and Paleoclimates of the EasternMediterranean 315ndash22 Tucson University of Arizona Radiocarbon Press
Gat JR and Magaritz M (1980) Climatic variations in the eastern Mediterranean seaarea Naturwissenschaften 6780ndash7
Geyh MA (1994) The paleohydrology of the eastern Mediterranean In OBar-Yosef and RKra (eds) Late Quaternary Chronology and Paleoclimates of the EasternMediterranean 131ndash45 Tuscon University of Arizona Radiocarbon Press
Gihon M (1980) Research on the Limes Palaestina a stocktaking In WSHanson and LJFKeppie (eds) Roman Frontier Studies 1919 843ndash64 Oxford British Archaeological Reports International Series 71
Gilead I Rosen SA and Fabian P (1993) Horvat Beter (Bersquoer Matar) 1990ndash1991 Archaeological News 9988ndash89 (In Hebrew)
Glantz MH (1994) Drought desertification and food production In MHGlantz (ed)Drought Follows the Plough 7ndash32 Cambridge Cambridge University Press
Goldberg P (1994) Interpreting late Quaternary continental sequences in Israel InOBar-Yosef and RKra (eds) Late Quaternary Chronology and Paleoclimates of the Eastern Mediterranean 89ndash102 Tucson University of Arizona Radiocarbon Press
Grigg DB (1974) The Agricultural Systems of the World An Evolutionary Approach Cambridge Cambridge University Press
Haiman M (1995a) Agriculture and nomad-state relations in the Negev desert in theByzantine and early Islamic periods Bulletin of the American Schools of Oriental Research 29729ndash54
Haiman M (1995b) An early Islamic period farm at Nahal Mitnan Atiqot 261ndash13 Huntington E (1911) Palestine and Its Transformation Boston Houghton amp Mifflin Isaac B (1992) The Limits of Empire Oxford Clarendon Press Issar A (1995) Climatic change and history of the Middle East American Scientist
83350ndash5
The archaeology of drylands 58
Issar A and Govrin Y (1991) Climatic changes and the desertification of the Negev atthe end of the Byzantine period Katedra 6167ndash81 (In Hebrew)
Katz O (1993) Beer ShevamdashNahal Bekarsquo 1 Archaeological News 9987ndash8 (In Hebrew)
Katz O and May V (1996) Beer Sheva Ramot B Archaeological News 106162ndash4 (In Hebrew)
Kedar Y (1967) Ancient Agriculture in the Negev Highlands Jerusalem Bialik Institute
Khazanov AM (1984) Nomads and the Outside World Cambridge Cambridge University Press
Kohler-Rollefson I (1992) A model for the development of nomadic pastoralism on the Jordanian steppe In OBar-Yosef and AMKhazanov (eds) Pastoralism in the Levant Archaeological Materials in Anthropological Perspectives 11ndash18 Madison Prehistory Press
Kraemer CJ (1958) Non-Literary Papyri Excavations at Nessana Volume III Princeton Princeton University Press
Lowdermilk WC (1945) Palestine Land of Promise London Gollancz Mayerson P (1960) The Ancient Agricultural Regime of Nessana and the Central Negev
Excavations at Nessana Volume I London Colt Institute Mayerson P (1963) The desert of southern Palestine according to Byzantine sources
Proceedings of the American Philosophical Society 107160ndash72 Mayerson P (1982) The pilgrim routes to Mount Sinai and the Armenians Israel
Exploration Journal 3244ndash57 Mayerson P (1983) The city of Elusa in the literary sources of the fourth-sixth centuries
Israel Exploration Journal 33247ndash53 Mayerson P (1986) The Saracens and the Limes Bulletin of the American Schools of
Oriental Research 26235ndash47 Mayerson P (1989) Saracens and Romans micro-macro relationships Bulletin of the
American Schools of Oriental Research 27471ndash9 Mayerson P (1990) Toward a comparative study of a frontier Israel Exploration
Journal 40267ndash79 Mayerson P (1994) Monks Martyrs Soldiers and Saracens Papers on the Near East in
Late Antiquity ( 1962ndash1993 ) Jerusalem Israel Exploration Society Mazor G (1981) The wine presses of the Negev Qadmoniot 53ndash5451ndash60 (In
Hebrew) Nachlieli D (1992) The Negev Highlands and the Arava During the Early Arab Period
Tel Aviv University unpublished MA thesis Nachshoni P Ustinov Y and Bar-Ziv H (1993) Beer ShevamdashNahal Kovshim
Archaeological News 9984ndash5 (In Hebrew) Neev D and Emery KO (1995) The Destruction of Sodom Gomorrah and Jericho
Oxford Oxford University Press Negev A (1974) The churches in the central Negev an archaeological survey Revue
Biblique 81400ndash22 Negev A (1980) House and city planning in the ancient Negev and Provincia Arabia In
GGolany (ed) Housing in Arid Lands Design and Planning 3ndash32 New York John
The decline of desert agriculture a view from the classical period Negev 59
Wiley and Sons Negev A (1988) The Nabatean Cities in the Negev Jerusalem Ariel 62ndash63 Negev N (1993) Beer ShevamdashKiryat HaUniversita (Mizrach) Archaeological News
9985ndash6 (In Hebrew) Nevo YD (1985) Sede Boqer and the Central Negev in the 7ndash8th Centuries AD
Jerusalem Israel Publication Services Nevo YD (1991) Pagans and Herders Jerusalem Israel Publication Services Orni E and Efrat E (1980) Geography of Israel Jerusalem Israel University Press Palmer EH (1872) The Desert of the Exodus New York Harper and Bros Reifenberg A (1955) The Struggle Between the Desert and the Sown Jersualem The
Jewish Agency Rosen SA (1987a) Demographic trends in the Negev highlands preliminary results
from the Emergency Survey Bulletin of the American Schools of Oriental Research26645ndash58
Rosen SA (1987b) Byzantine nomadism in the Negev results from the EmergencySurvey Journal of Field Archaeology 1429ndash42
Rosen SA (1994) The nomadic periphery archaeology of pastoralists in the southcentral Negev during late antiquity Aram 6295ndash309
Rosen SA and Avni G (1993) The edge of empire the archaeology of pastoral nomadsin the southern Negev highlands in late antiquity Biblical Archaeologist 56 189ndash99
Rosen SA and Avni G (1997) The lsquoOded Sites Investigations of Two Early Islamic Pastoral Camps South of the Ramon Crater Beersheva Ben-Gurion University Press Beersheva XI Studies by the Department of Bible and Ancient Near East
Rubin R (1989) The debate over climatic changes in the Negev fourth-seventh centuries CE Palestine Exploration Quarterly 12171ndash8
Rubin R (1990) The Negev as Settled Land Jerusalem Hebrew University Press Rubin R (1996) Urbanization settlement and agriculture in the Negev desertmdashthe
impact of the Roman-Byzantine empire on the frontier Zeitschrift des Deutschen Palstini-Vereins 11249ndash60
Rubin R (1997) The Romanization of the Negev Israel geographical and culturalchanges in the desert frontier in late antiquity Journal of Historical Geography 23 267ndash83
Segal A (1983) The Byzantine City of Shivta (Esbeita) Negev Desert Israel Oxford British Archaeological Reports International Series 179
Sharon M (1969) The history of Palestine from the Arab conquest until the Crusades(633ndash1099) In MAvi-Yonah (ed) A History of the Holy Land 185ndash220 London Weidenfeld amp Nicolson
Sharon M (1976) Processes of destruction and nomadization in Eretz Israel underIslamic rule (633ndash1517 CE) In MSharon (ed) Notes and Studies on the History of theHoly Land Under Islamic Rule 7ndash32 Jerusalem Yad Ben Zvi Press
Sharon M (1990) Arabic rock inscriptions from the Negev In MSharon and MHalloun(eds) Supplement to the Map of Har Nafha (196) 9ndash35 Jerusalem Israel Antiquities Authority
Shershefski J (1991) Byzantine Urban Settlements in the Negev Desert Beersheva Ben-Gurion University Press Beersheva V Studies by the Department of Bible and
The archaeology of drylands 60
Ancient Near East Stager LE (1985) The first fruits of civilization In JNTubb (ed) Palestine in the
Bronze and Iron Ages Papers in Honor of Olga Tufnell 172ndash88 London University of London Institute of Archaeology
Thomas DSG and Middleton NJ (1994) Desertification Exploding the Myth New York John Wiley and Sons
Troen I (1989) Calculating the economic absorptive capacity of Palestine a study of thepolitical uses of scientific research Contemporary Jewry 1019ndash38
The decline of desert agriculture a view from the classical period Negev 61
4 Farmers herders and miners in the Wadi Faynan
southern Jordan a 10000-year landscape archaeology
GRAEME BARKER
INTRODUCTION
The Wadi Faynan Landscape Survey is a study of the landscape evolution of WadiFaynan in southern Jordan from prehistoric times to the present day as a contribution tothe issue that is the central theme of this volume the importance of providing a long-term archaeological perspective on how people have lived in arid lands How did past societiesin marginal environments learn to cope with risk What solutions did they develop andhow successful were they Why did they take the choices they took To what extent didtheir actions affect their landscape and for good or ill The rationale of the project hasbeen to bring together an inter-disciplinary team of archaeologists geographers and environmental biologists in the investigation of the landscape history of the chosen studyarea within a single integrated research framework (The Acknowledgements at the endof the chapter detail the numerous colleagues working in the project whose researches aresummarized in these pages)
The Wadi Faynan is situated about 40 km from Petra the world-famous capital of the Nabatean kingdom that flourished in the last few centuries BC before the Romanconquest of the region (Fig 41) The catchment of the wadi forms a transect about five km wide running for some 15 km westwards from the rim of the Jordanian plateauc1500 m above sea level to the floor of the Wadi Arabah rift valley at about sea levelThe main wadi today is a bleak landscape arid and largely denuded of vegetation (Figs42 45) though where they cut through the plateau escarpment the channels of the threemain feeder tributaries (the Dana Ghuwayr and Shayqar) are in places well watered andcomparatively well vegetated from ground springs The Wadi Faynan today (part of theDana Nature Reserve of Jordanrsquos Royal Society for the Conservation of Nature) is usedmainly by nomadic bedouin herders but is also well known for its abundantarchaeological remains
The principal archaeological monuments of the Wadi Faynan long known to early travellers are the Khirbet Faynan (Fig 42) a major settlement of
Figure 41 The location of Wadi Faynan within its region
Farmers herders and miners in the Wadi Faynan southern Jordan 63
Figure 42 Looking northeast across part of the ancient field system to Khirbet Faynan (the prominent hill in the right middle ground) thought to be the ancient settlement of Phaino mentioned by classical writers
Photograph GBarker
Nabatean Roman and late Roman (Byzantine) date located at the head of the WadiFaynan near the confluence of the three main tributaries and nearby an aqueductreservoir and water mill of RomanByzantine date To the west of this complex is asubstantial (c5 km long) field system of rubble walls its surface pottery indicatingprimary use contemporary with that of the Khirbet Faynan settlement (Fig 43) Before our project began in 1996 reconnaissance surveys had also located a variety ofprehistoric sites both in the main wadi and in its tributaries some of which are beingexcavated by other teams Wadi Faynan and its environs are also characterized by richmineral deposits and from the work especially of the Bochum Mining Museum thehistory of copper exploitation here is comparatively well documented (Hauptmann 19891992 Hauptmann and Weisgerber 1987 Hauptmann et al 1992) Although Faynan copper was used by neolithic and chalcolithic societies the first intensive exploitationseems to date to the Early Bronze Age c3500ndash1900 BC There was a second significant episode in the first part of the first millennium BC the Edomite Iron Age Copper wasthen extracted on a major scale in Nabatean and especially Roman and Byzantine timesit is generally agreed that Khirbet Faynan must be the settlement of Phaino mentioned by classical writers as the place to which prisoners such as Christians from Palestine andEgypt were transported in the third and early fourth centuries AD to work the coppermines it controlled
The archaeology of drylands 64
Figure 43 The survey area of the Wadi Faynan Landscape Survey showing the ancient field systems and the archaeological sites recorded up to 1999
Note The topography shown is from a photogrammetric map the boundaries of which do not extend as far as the boundaries of the survey area
The Wadi Faynan seemed therefore a particularly attractive location for investigating the lsquoarchaeological historyrsquo of interactions between a desertic landscape and its human inhabitants given the rich archaeological record that appeared to be prima facie evidence for episodes of intensive settlement and sedentary farming in the past that were verydifferent from settlement and land use today
METHODOLOGIES
The project began in 1996 and the fieldwork ends in 2000 ongoing results have beenreported in annual papers in Levant (Barker et al 1997 1998 1999 2000)
Geomorphological mapping and palaeoecological analysis of exposures and coredsediments are establishing an environmental sequence for the past 200000 years with aparticular focus on the past 10000 years The resulting event sequence is being datedvariously by radiocarbon and Optical Spin Luminescence (OSL) dates and bystratigraphic association with dated archaeological sites The changes that are beingobserved in the palaeoenvironmental sequence can be linked with increasing confidencevariously to natural processes of change such as climatic shifts and phases of tectonicactivity and to cultural processes of change such as arable pastoral and industrialactivities Geochemical analysis of sediments using EDMA (Energy Dispersive X-Ray Microanalysis) is also being used to monitor the bioaccumulation of heavy metals as a
Farmers herders and miners in the Wadi Faynan southern Jordan 65
result of metalliferous pollution providing an invaluable independent indicator of thechanging scales technologies and environmental impacts of mining and smeltingactivities to compare with the Bochum teamrsquos studies of the mining and smelting sites
In the first three seasons of the project the programme of archaeological fieldwork concentrated on the detailed exploration of the relict field system (termed WF4 in thesurvey catalogue) that covers the lower terraces on the southern side of the present-day wadi channel (Figs 42 and 43) This involved the verification on the ground of walls shown on an earlier photogrammetric survey the systematic collection and counting ofartefacts in each individually numbered field of the system (c 900 fields in total) and the detailed recording of constructional details of wall types and of other structures within oradjacent to the fields A series of smaller satellite field systems on the northern side of themain channel has also been recorded in the same way These studies combined with testexcavations and identification of the prolific lithic and pottery collections haveconfirmed that the main phase of construction and use of the field systems is broadlylsquoclassicalrsquo (NabateanRomanByzantinemdasha period of some 1000 years) but have also established that the evidence is at the same time a complex palimpsest of reuse andadaptation of agricultural activities and systems of land management spanning the past6000 years or so
The focus of the fieldwork has now shifted to frame these data within the broader landscape context of the study area defined for the archaeological investigation shown inFigure 42 which measures just over 30 km2 Extensive field walking and recording of the block of terrain around the ancient field systems in 1999 were facilitated by usinghand-held Garmin 12 GPS units within a grid based on UTM (Universal TransverseMercator) coordinates This survey located over 1000 lsquositesrsquo varying from lithic scatters to settlement structures and enclosures dating to all archaeological periods fromprehistoric times to the recent past The investigation of a representative sample of thesesites to attempt to refine our understanding of their chronological and functionalpatterning formed the primary focus of the archaeological fieldwork in the final (2000)season
The archaeological survey is drawing on the results of another component of the projectmdasha programme of ethnoarchaeological research (Fig 44) This involves elucidating how farmers (fellaheen) and pastoralists (bedouin) exploit the landscape of the study area and adjacent zones of the Wadi Arabah and plateau today and how theyhave done so in the recent past Within the
The archaeology of drylands 66
Figure 44 Ethnoarchaeological survey the typical site of a winter bedouin tent (beit sharsquoar) in Wadi Faynan
Note Gullies to direct rainwater away from the cleared menrsquos and womenrsquos sections are clearly visible The menrsquos hearth is under the tent poles to the left of the photographic scale The womenrsquos hearth is in the far left-hand corner of the cleared area where there are fire-blackened stones To the rear there is a thick dark accumulation of animal dung where the goats were housed at night Scale 1 m Photograph CPalmer
study area planning recently abandoned bedouin structures and analysing their floorsediments combined with interviewing the families who used the structures isestablishing archaeological signatures of seasonality and different age and gender groupsto inform our interpretations of the settlement archaeology
The final major component of the projectrsquos methodology is the development of a Geographical Information System integrating all the above data This is attempting torefine further our understanding of changing relationships between arable pastoral andindustrial activities between the lsquoeconomicrsquo lsquosocialrsquo and lsquoritualrsquo landscapes that are being defined and between all these cultural activities and the development of the naturallandscape which is the core issue of the projectrsquos research agenda
THE NEOLITHIC EARLY FARMING c9500ndash4000 BC
The transition from the Pleistocene to the Holocene or Postglacial the modern climaticera occurred at approximately 9500 BC Our geomorphological studies have established
Farmers herders and miners in the Wadi Faynan southern Jordan 67
that as elsewhere in the region the Pleistocene late glacial environment was cold anddry We then have strong and widespread evidence from sediment sequences and thefauna and flora within them that the early Holocene was characterized by a significantlywetter environment than today which lasted until about 6000 years ago
The Near East has some of the earliest evidence in the world for agriculture which can be recognized very soon after the beginning of the Holocenemdashthe culmination of trends in settlement subsistence and social change that can be observed amongst the Natufianpeoples of the late Pleistocene following the peak of glacial conditions c20000 years ago (Sherratt 1997) After about 15000 years ago there was a sudden dramatic warmingand the Natufians were able to develop semi-permanent settlements by lakes and springsin the Jordan valley (the lsquoLevantine corridorrsquo) Excavations show that Natufian settlements such as Jericho and Abu Hureyra were sustained by a combination of fishingfowling hunting (especially gazelle) collecting forest foods in the valley and gatheringwild cereals and other grasses on the steppelands above (Hillman 1996) With the returnto cold and dry conditions termed the Younger Dryas (11000ndash9500 BC) the steppelands returned to being a resource-poor environment and lake levels shrank Natufians responded in various ways (Bar-Yosef and Belfer-Cohen 1992) some moved others diversified their subsistence but in the Levantine corridor in particular the signs are thatpeople began to concentrate even more on spring-side locations and on collecting cereals perhaps engaging in activities that can be regarded as incipient horticulture So far wehave only lithic evidence (flint implements found on the surface) for the presence ofNatufian hunter-fisher-gatherers in the Wadi Faynan but significantly most of it has been found in the upper tributaries near the springs (Fig 45)
With the beginning of the Holocene c9500 BC there was a sudden return to warmerconditions coinciding with the first generally accepted evidence in the Levantine corridorthat the main settlements (Pre-Pottery Neolithic A or PPNA) were being sustained at least in part by the cultivation of plants though wild seeds and fruits continued to be staplefoods augmented by fishing fowling and hunting a variety of game (Bar-Yosef and Kislev 1989 Byrd 1992) Sedentary mixed farming in which goat herding wascombined with cereal cultivation then developed throughout the Near East about 1000years later (the Pre-Pottery Neolithic B or PPNB c8500ndash6500 BC) coinciding with major changes in architecture (the appearance of substantial square or rectangulardwellings rather than the circular or oval rubble shelters of Natufian and PPNA sites)PPNA and PPNB sites were invariably located by springs presumably because the latterprovided naturally irrigated land for cereal fields (Bar-Yosef 1995)
This transformation is exemplified in the development of the well-known PPNA and PPNB settlement of Beidha (Byrd 1994 Kirkbride 1966) on the plateau near Petra butis also clear from current investigations of PPNA and PPNB settlements in the WadiFaynan A PPNA site of simple rubble shelters and pits has evidence for mixed huntinggathering and plant cultivation (Finlayson and Mithen 1998) The inhabitants of asubstantial PPNB settlement of well-built stone houses were mixed farmers growingwheat barley and legumes and herding domestic sheep and goats (Simmons and al-Najjar 1996) The two sites are only 100 m apart in the Wadi Ghuwayr at the junctionbetween the mountains and the main wadi by the spring where Natufians also camped(Fig 45 upper photograph) We have also found traces of similar settlements in the
The archaeology of drylands 68
upper Wadi Dana by the main spring there Although these first agricultural communitiesclearly preferred the well-watered upper tributaries for their primary settlements otherlithic scatters indicate that they also used the main wadi presumably for hunting andherding
By the sixth and fifth millennia BC the zone of principal arable settlement hadexpanded out into the main wadi Excavations a few years ago revealed a lateneolithicearly chalcolithic settlement of simple rectangular drystone houses at Tell WadiFaynan 1 km west of Khirbet Faynan (al Najjar et al 1990 Fig 45 lower photograph) Our geomorphological investigations show that when these people settled at Tell WadiFaynan the climate was significantly wetter than either before or afterwards there was amore or less perennial stream by the sitemdashthe archaeological sediments contain for example frustules of the diatom Navicula a freshwater organism and the pottery andmortar contain fragments of reeds and grass as well as straw The likelihood is that theprimary farming zone was able to expand from the tributaries to the main wadi floor atthis time because the climatic amelioration allowed farmers to exploit the seasonalfloodwaters of the main wadi for their crops with methods akin to those used by theearlier neolithic farmers in the upper tributaries Presumably they sowed their crops oneither side of what was then the Faynan stream after spring floods had soaked the soilson either side of its course
Farmers herders and miners in the Wadi Faynan southern Jordan 69
Figure 45 The settlement locations of the first farmers in the Wadi Faynan
Note (above) Looking east from near Khirbet Faynan up the Wadi Ghuwayr the PPNB settlement was on the low terrace immediately above and the PPNA settlement on the low terrace immediately to the right of the wadi channel where it issues from the hills at the centre of the picture (the spring is behind the PPNB settlement) (below) looking west from Khirbet Faynan towards the Wadi Arabah the late neolithic settlement of Tell Wadi Faynan is the prominent cliff at the channel edge in the distance on the right-hand side of the photograph when it was occupied there was a perennial stream flowing down the wadi Photographs GBarker
The archaeology of drylands 70
THE BRONZE AGE c4000ndash1200 BC THE BEGINNINGS OF METALLURGY AND DRY FARMING
Aridification began to develop in the fifth millennium BC leading to the development ofa relatively steppic landscape by the fourth and third millennia BC the period of theEarly Bronze Age This was a period of immense social change in the Near Eastcharacterized by the development of metallurgy long-distance trade networks and in some regions complex polities with quasi-urban settlements (Finkelstein 1995 Gophna 1995)
Nuggets of surface copper were collected by neolithic (and probably earlier) people inWadi Faynan presumably for ornamental purposes and Faynan copper was exploited bychalcolithic people and traded with the surrounding region However the first clearevidence for the systematic mining of the copper ores and their processing at Faynansettlements is in the Early Bronze Age (Adams and Genz 1995 Hauptmann 1989Wright et al 1998) This was the context for the emergence of local elites who controlledcopper production and the exchange to other regions of smelted copper ores and finishedartefacts The research by the Bochum Mining Museum suggests that at first ores visibleat the surface were mined by open-cast methods and then smelted in simple crucibles in the settlements but as demand increased deeper ores were mined by galleries and thensmelted in smelting ovens located on the windward side of ridges near the settlementsOur geochemical analyses of sediments at Tell Wadi Faynan indicate that these smeltingactivities caused small-scale localized pollution
The primary settlement zone shifted during this period into the main wadi and expanded throughout it The survey has revealed a series of discrete zones of bronze agesettlements associated with field systems both on the southern side of the wadi within thearea demarcated by the later classical field system and in the small tributary wadis on thenorthern side One zone of the classical field system encompasses the most substantial ofthese settlements where excavations by Dr Karen Wright have revealed evidence forirregularly built drystone structures together with enclosures middens pits and storagebins and evidence of working smelted copper into ingots and finished artefacts (Wrightet al 1998) In the area of this settlement we have been able to recognize a series of boulder walls within and underlying the later field system that appear to be vestiges ofbronze age structures and field boundaries some of the latter terraced (Fig 46) They are associated with circular or oval cisterns 30ndash50 cm deep fed by short feeder walls The northern settlement zones include sequences of roughly built terrace walls and checkdams built across the shallow floors of tributary wadis with pottery in associatedsediment sequences indicating a bronze age date (Fig 47) In the hills to the north the 1999 survey found a series of small sites with crudely built one- and two-roomed structures with circular enclosures presumed to be pastoral encampments (they haveanalogies with the pastoral encampments of the Negev Chapter 3) and a few larger more complex
Farmers herders and miners in the Wadi Faynan southern Jordan 71
Figure 46 Part of the Wadi Faynan field system WF4 showing (above) the early bronze age and (below) the classical landscape in unit WF413
settlements near ridge-top spreads of bronze age slag and furnace waste presumed tohave hadmdashat least in partmdashan industrial function
The indications are therefore that early bronze age settlement in Wadi Faynan was characterized by three different archaeological complexes linked to three overlapping butdiverse economic orientations agricultural pastoral and metallurgical Whereas neolithicfarmers in the Wadi Faynan were able to exploit well-watered locations bronze age
The archaeology of drylands 72
farmers were having to develop
Figure 47 A field system on the northern side of the Wadi Faynan
Note The outer diversion walls and many of the field walls in the centre very probably date to the Iron Age although potsherds in sediment exposures indicate that some of the simple check dams at points A and B are bronze age
strategies for coping with the more arid environments evidenced for the fourth and thirdmillennia BC such as building walls to collect and trap seasonal floodwaters in storagecisterns and in terraced fields laid out along the direction of water flow If correctly dated(and the dating is still tenuous) this will be the earliest evidence for floodwater farmingyet found in the Near East making this another indicator of the social and economictransformations that characterized this phase of settlement in the region A degree ofpastoral specialization may have been another way in which bronze age societies wereable to respond to aridity whilst also being like metallurgy an indicator of complexeconomic structures of production and exchange What is also interesting is that we haveevidence for strong soil erosion through the second and first millennia BC andpalynological indicators that this reflects the impact of human activities on the landscapesuch as clearance of fuelwood for smelting and intensification in systems of cultivationand herding rather than climatic change
What happened in the Wadi Faynan during the second millennium BC (the LaterBronze Age) remains unclear as throughout the entire region The end of the millenniumwas marked by the widespread disintegration of urban polities throughout the EastMediterranean and Levant There is some evidence for climatic deterioration this is the
Farmers herders and miners in the Wadi Faynan southern Jordan 73
period of the Thera or Santorini eruption in the Aegean Seamdashand Egyptian scribes make references to raiding by lsquoSea Peoplesrsquo so there has been a lively debate aboutwhether lsquoexternalrsquo environmental or cultural factors such as these caused economic collapse or whether (more likely) they exacerbated internal processes of social changethat were already in train In the Levant it has commonly been argued that societiesturned to nomadic pastoralism at this time (Finkelstein 1995 LaBianca 1990) thoughconvincing archaeological signatures for such pastoralism are unclear (Bar-Yosef and Khazanov 1992) The debate is further obscured by Biblical archaeologists andhistorians looking for the nomadic peoples of Old Testament origin myths In theethnohistorical record furthermore specialized pastoralism is a highly complexeconomic system that invariably operates not in isolation but in close relationship withadjacent agricultural and urban systems as it may have operated in articulation witharable and metallurgical activities at Faynan in the Early Bronze Age It is thereforedifficult to separate absence of settlement evidence from evidence of settlement absencein the Wadi Faynan at this time but it does seem likely that smaller-scale systems of mixed cultivation and herding characterized life in the wadi during the secondmillennium BC
THE IRON AGE c1200ndash300 BC TRANSFORMATIONS IN SETTLEMENT FARMING AND MINING
The early first millennium BC saw the emergence of iron age states west of the Jordanand tribal kingdoms in the Jordan valley and steppeland to its east Ammon in northern Jordan Moab in central Jordan and Edom in southern Jordan (LaBianca and Younker1995) The State of Edom in which Wadi Faynan was situated was the least denselysettled of the three Although historians have accepted Josephusrsquo statement that these kingdoms were destroyed by the Neo-Babylonians in the sixth century BC archaeological surveys in fact indicate continuity in rural settlement for example aroundMadama in Ammon (LaBianca 1990)
In the Wadi Faynan the iron age landscape was quite different from that of the Bronze Age dominated by a single substantial settlement site WF424 in our survey record builtimmediately below its successor Khirbet Faynan at the strategic centre of the Faynanregion at the point where the three major tributary wadis come together to form the mainchannel of the Faynan We have also found zones of iron age settlement along thesouthern margins of the field system and on the northern side of the wadi The evidencesuggests therefore that the Edomite settlement system consisted of a few large anddiscrete habitation units probably organized hierarchically with WF424 the dominantsite Recent work in the neighbouring Wadi Fidan indicates that there may have beenother more ephemeral settlement forms as well (Levy et al 1999) By this period deep ores were being mined extensively in the hills and then smelted at settlements such asWF424 where we found thick deposits of slags charcoal-rich unlike the bronze age slags suggesting experimentation with new technologies to deal with the far largerquantities of ore being processed for the Edomite economy Geochemical analysesconfirm the increasing scale of smelting pollution
The archaeology of drylands 74
WF424 was associated with a field system of boulder-built walls often set orthostatically and also with substantial boundary walls built upstream of these fieldsalong the junction between the hill and the wadi floor These boundary walls collectedwater from the surrounding slopes and channelled it to exit sluices above the terracedfields so that maximum water flow could be directed down the central part of the fieldsystem Similar boundary walls enclosed iron age field systems on the northern side ofthe Wadi Faynan and in part at least they had a water-diversion function (Fig 47) We cannot be sure of the dating of these boundary walls but the fact that we have only foundthem enclosing field systems with significant iron age material and the constructionalsimilarities between the walls and the fields they enclose suggest an iron age date If thisdating is correct it implies that whereas bronze age farmers built simple terrace walls atright angles across wadi beds to check floodwater flow and to try to spread it laterallyover surrounding fields together with small catchments to collect water in cisterns ironage farmers in Faynan had learned to construct substantial and rather sophisticated wallsto divert the flow of floodwaters sometimes hundreds of metres from their natural line sothat far greater quantities of water could be collected and sent down a field system thanwas possible with bronze age technology
NABATEAN SETTLEMENT c300ndash63 BC
The Nabatean state with its capital Petra developed at the time that Romersquos power was expanding across the eastern Mediterranean in the last three centuries BC and flourisheduntil Palestine was annexed by Rome in 63 BC The Nabatean settlement system in thewadi like that of the Iron Age was dominated by one central settlement the communityat Khirbet Faynan This settlement presumably controlled copper production which onthe evidence of both mining archaeology and sediment geochemistry increaseddramatically in scale in the later centuries of the first millennium BC The landscape nowconsisted of a series of adjacent settlement units organized in some kind of hierarchicalrelationship with respect to Khirbet Faynan a series of large farmsteads of broadlyNabatean date on the southern slopes overlooked the classical field system excavationsby Wright et al (1998) found small buildings of Nabatean date in the area of majorbronze age settlement within the classical field system (Fig 46) and our survey has identified a variety of structures with similar pottery elsewhere within the field system
The technology of floodwater farming was further refined by Nabatean farmers The particular focus of their wall-building activities was a series of small tributary wadis thatrun parallel to the main wadi along its southern side (Fig 48) Water was dammed as it issued from the adjacent hills diverted westwards by boulder walls along the contour ofthe slope and then channelled through simple sluices (gaps) and spillways (steppedstructures) onto terraced fields below Nabatean technology on the southern slopes mayalso have included channels formed by parallel walls that fed water directly into the fieldson either side through sluice gaps
Farmers herders and miners in the Wadi Faynan southern Jordan 75
THE ROMAN IMPERIAL LANDSCAPE 63 BC-cAD 600
With the Roman annexation of Palestine Faynan became one of the principal suppliers ofcopper and lead for its eastern empire with its extraction probably state-organized A garrison was located at Khirbet Faynan the Phaino of the classical sources and the surrounding hills were honeycombed with deep mine shafts There are references toChristian slaves being sent to work the mines as punishment (according to some sourcesthey were crippled to prevent their escape by blinding having a hand removed or havingtheir Achilles tendon severed) though most of the mining and smelting was probablydone by a workforce paid well for skilled and dangerous work The quantities of orebeing smelted by the Phaino labour gangs have left kilometre-long spreads of tap slag on the ridges above Khirbet Faynan The EDMA geochemical studies of the heavy metalsfound in a 2500-year long sequence of sediments behind a barrage at Khirbet Faynanindicate extraordinary levels of air pollution in Roman and Byzantine times with levelsseveral times lethal in terms of modern pollution criteria (Fig 49)
Figure 48 A field map of part of the field system WF4 after ground verification by the survey teams
Note The photograph of Khirbet Faynan shown as Figure 42 is looking across the fields mapped as Units 4 3 and 2 in this plan The upper (southern) part of this field system was probably laid out by Nabatean farmers who diverted water as it flowed out of tributary wadis onto the upper terraces for example diverting water at point F from its channel F-G along the wall F-H and then through sluices onto the terraced fields below The entire system was managed as an integral unit by Roman farmers who also built the mill complex
The archaeology of drylands 76
The satellite farms were abandoned leaving Khirbet Faynan as the single dominant settlement Our studies of field layout and construction and of the surface materialindicate that the entire agricultural landscape was now managed as a more or less integralunit or estate Systems of long parallel walls were built to divert water from the mainwadi into adjacent fields on low-elevation terrain and from the southern tributary wadis (Fig 410) Further down the tributary wadis similar channels were built at c45 degrees to water flow to collect any water that had bypassed the higher diversion walls or drainedback into the wadis from the higher terraced fields to force it once more onto adjacentcultivable land The effectiveness of the system is in part explained by the uniformly lowlevels of water infiltration we have found at sample sites from the upper slopes to thelowest fields but organizational factors were also important The field evidence supportsthe hypothesis of cooperation between areas of the field system fed by the parallelchannels rather than farmers with land upslope having exclusive access to thefloodwaters of particular wadis at the expense of other farmers with land
Figure 49 The distribution of copper (in parts per millionmdashppm) through the sediments that accumulated behind the Khirbet Faynan barrage
Note The sequence extends from c2500 years ago (far right) to the present day (far left)
further down the direction of flow the internal linkages between the system imply thatwater resources were shared down the length of the field system The construction ofmajor parallel channels to carry floodwater through the system demonstrates the sameengineering skills in moving water relatively long distances over gentle gradients as aredisplayed by the Roman engineers who designed the rock-cut feeder channel that brought water several kilometres from the Wadi Ghuwayr spring to the aqueduct feeding thereservoir and ore-crushing mill near Khirbet Faynan
Farmers herders and miners in the Wadi Faynan southern Jordan 77
This imperial landscape was highly organized with large-scale industrial processing sustained by an integrated agricultural and hydraulic system But it was also increasinglybarren fast eroding and grossly polluted Charcoal samples from the smelting sitesstudied by the German team show that whereas Nabatean miners cut local firewood fortheir smelting activities by the Roman period timber was having to be brought downfrom the plateau because local supplies had been exhausted (Engel 1993 Hauptmann1992) Our pollen evidence indicates the same process of humanly induceddesertification by the time of Christ the landscape consisted of very degraded steppelandand this degradation then accelerated significantly through the first millennium AD Bythe end of the Roman period the steppic component of the pollen diagrams collapsesevidence of olive cultivation disappears and signs of cereal cultivation drasticallyreducemdashthe flora at this time was analogous to the modern pollen rain in the Dead SeaWe have also found widespread evidence that Roman farmers were trying to combat theeffects of wadi-downcutting in
The archaeology of drylands 78
Figure 410 Section through a Roman-period water conduit channel visible on the surface as two parallel walls between the ranging pole in the foreground and the right-hand tree in the distance
Note The section shows that the parallel walls on the surface overlie buried walls of an ancient conduit filled with water-lain sediments that contained Roman pottery Photograph GBarker
their alterations to the floodwater farming systems though these were ultimatelyineffective (the main wadi now flows at least 5 m below the parallel channel systems thatdiverted the Faynan floodwaters into the ancient fields)
Farmers herders and miners in the Wadi Faynan southern Jordan 79
THE POST-BYZANTINE LANDSCAPE
The desertic environment has persisted to the present day though there is evidence for anepisode of even greater aridity in the period cAD 1600ndash1850 The nature of the Islamic and later settlement systems following the eventual abandonment of Khirbet Faynanremains unclear but it is at least evident from our survey work so far that it was notcharacterized by a renewal of substantial settlement within the field system zone andthere are indications of ephemeral settlements on the surrounding slopes akin to the sitesof recently abandoned Bedouin encampments Small-scale increases in smelting pollutants in levels above the post-Byzantine collapse (such as the peak at c90 cm depth in Figure 49) indicate a revival of industrial activity at some time probably in the early second millennium AD (radiocarbon dates are awaited) and the range of pollutantssuggest the reworking of Roman and Byzantine slag deposits rather than renewed miningon any scale The likelihood is that the degraded landscapes of the post-Byzantine period have for the greater part supported only systems of land use like those of the bedouin inthe region today
CONCLUSION
As described above we are beginning to detect oscillations in environmental changesettlement forms and agricultural and industrial activity over the past 10000 years If wecan understand how they do and do not inter-relate we should be able to write alandscape history in the Braudelian sense of complex interactions between short-term processes medium-term processes and the longue dureacutee that can provide a significant archaeological contribution to the desertification debate
In terms of environmental change after the wetter phase of the early Holocene we can discern a principal trend of progressive aridification and degradation culminating inextremely degraded landscapes by the mid-first millennium AD However it is clear that the trend was not constant in its progression and that it contains oscillations In terms ofland use from the Late Neolithic onwards a number of increasingly sophisticated systemsof water control can be discerned but again it is clear that there is no simple progressionin land use from simple to complex but rather oscillations between the two Anothercomplex non-linear sequence is emerging regarding the impact of people on landscape The expansion of farming down the wadi in the Later Neolithic appears to have had nosignificant impact on the landscape but it is possible that the erosion we can detectduring and after the Early Bronze Age whilst probably mainly a response to aridificationpartly reflects poor land management techniques Given the evidence of the geochemistryfor the beginning of smelting pollution at this time wood cutting for metallurgicalprocessing may also have been a factor However that may be it is clear that the demandsof Nabatean and in particular RomanByzantine mining in parallel with the intensiveagricultural practices developed to feed the workforce had an ultimately devastatingimpact on the landscape Whether or not climatic change was also a factor it is clear that
The archaeology of drylands 80
large-scale stripping of the landscape of vegetation made it extremely vulnerable toerosional forces
The geochemical evidence also demonstrates that the effects of Roman and Byzantine mining and smelting are still felt todaymdashthe milk urine and faeces of the bedouinsrsquo goats today have significant levels of heavy metals from grazing the polluted ground andcereal growth is also badly affected around the smelting sites Does the extraordinarydensity of Roman and Byzantine potsherds carpeting the field system indicate large-scale manuring in an attempt to deal with falling cereal yields Certainly there is a strongpossibility (currently being tested by skeletal analysis) that the health of the Roman andByzantine population in particular was directly affected from inhalation skincontamination and bioaccumulation of polluted animal and plant foods Whilst thecollapse of intensive farming and mining in Late RomanByzantine times no doubt in partreflects changing economic relations between Faynan and the wider world it seemsinescapable that the activities of these farmers and miners had a profound impact on theirlandscape (which has still not recovered) and probably directly on their own well-being
The project has succeeded so far in establishing the principal components of the environmental and settlement sequences in the Wadi Faynan What we can tell so far oftheir inter-relationships indicates something of the potential complexity of the interplaybetween long-term medium-term and short-term processes that is likely to emerge as the project develops However the richness of the data also suggests that especially throughGeographical Information Systems (GIS) we should be able to integrate our findings onhow the landscape has changed and the role of people in this to investigate also howdifferent populations in the past perceived their changing landscape and their place withinit It should be possible for example to model the spatial characteristics of air pollutionat different distances from the major smelting settlements and its different effects onsurrounding populations There is much still to be learned about the extent to whichfarmers using the field system operated independently or collaborated and in the lattercase whether they did so by cooperation or coercion There is also the question of howthe sacred and the secular related to one another in different periods There is intriguingevidence for example that whereas bronze age people kept their cemeteries and fieldsapart classical farmers deliberately constructed water diversion systems so that theyincorporated pre-existing burial cairns at key nodal points Reconstructing landscapehistories needs natural scientists to analyse changing forms of landscape andarchaeologists to analyse changing settlement morphologies and systems We hope thatthe effective partnership of disciplines in our project will allow us ultimately not just todescribe these two landscape histories for Faynan but also to understand theirinteractions looking at the perceptions and choices that underpinned human actions inthis landscape and shaped the latter with ultimately devastating consequences
ACKNOWLEDGEMENTS
This chapter represents the work of the Wadi Faynan Landscape Survey team especiallyRAdams (University of Bristol prehistoric pottery analysis) OCreighton (University ofExeter archaeological survey) PDaly (University of Oxford archaeological survey and
Farmers herders and miners in the Wadi Faynan southern Jordan 81
GIS analysis) DDGilbertson (University of Bournemouth geomorphology) JPGrattan(University of Wales Aberystwyth geomorphology geochemistry) COHunt(University of Huddersfield palynology) DJMattingly (University of Leicesterarchaeological survey) SJMcLaren (University of Leicester geomorphology)HMohammed (University of Benghazi Libya palynology) PNewson (University ofLeicester archaeological survey and GIS analysis) CPalmer (University of Leicesterethnoarchaeology) HParton (British Institute at Amman for Archaeology and HistoryRoman and post-Roman pottery analysis) FBPyatt (Nottingham Trent University environmental biology geochemistry) TEGReynolds (Cambridgeshire CountyCouncil lithic analysis) HSmith (University of Bournemouth ethnoarchaeologyenvironmental archaeology) RTomber (Museum of London Roman and post-Roman pottery analysis) and ATruscott (University of Wales Aberystwyth OSL dating)mdashtogether with the archaeological field team that has done most of the foot work Gratefulthanks are also due especially to the Arts and Humanities Research Board the Councilfor British Research in the Levant the Natural Environment Research Council theSociety of Antiquaries of London and the Universities of Leicester Huddersfield andAberystwyth for funding the project and to the Jordanian Department of AntiquitiesCBRLrsquos British Institute at Amman for Archaeology and History and the Royal Societyfor the Conservation of Nature for other essential support
REFERENCES
Adams R and Genz H (1995) Excavations at Wadi Fidan 4 a chalcolithic villagecomplex in the copper ore district of Feinan southern Jordan Palestine Exploration Quarterly 1278ndash20
Barker G Gilbertson D Jones B and Mattingly D (1996) Farming the Desert The UNESCO Libyan Valleys Archaeological Survey Volume One Synthesis Paris UNESCO London Society for Libyan Studies Tripoli Department of Antiquities
Barker G Creighton OH Gilbertson DD Hunt CO Mattingly DJ McLaren SJand Thomas DC (1997) The Wadi Faynan Project southern Jordan a preliminaryreport on geomorphology and landscape archaeology Levant 29 19ndash40
Barker G Adams R Creighton OH Gilbertson DD Grattan JP Hunt COMattingly DJ McLaren SJ Mohammed HA Newson P Reynolds TEG andThomas DC (1998) Environment and land use in the Wadi Faynan southern Jordanthe second season of geoarchaeology and landscape archaeology (1997) Levant 305ndash26
Barker G Adams R Creighton OH Crook D Gilbertson DD Grattan JP HuntCO Mattingly DJ McLaren SJ Mohammed HA Newson P Palmer C PyattFB Reynolds TEG and Tomber R (1999) Environment and land use in the WadiFaynan southern Jordan the third season of geoarchaeology and landscapearchaeology (1997) Levant 31255ndash92
Barker G Adams R Creighton OH Daly P Gilbertson DD Grattan JP HuntCO Mattingly DJ McLaren SJ Newson P Palmer C Pyatt FB ReynoldsTEG Smith H Tomber R and Truscott AJ (2000) Archaeology and
The archaeology of drylands 82
desertification in the Wadi Faynan the fourth (1999) season of the Wadi FaynanLandscape Survey Levant 3227ndash52
Bar-Yosef O (1995) Earliest food producersmdashpre-pottery neolithic (8000ndash5500 BC) In TLevy (ed) The Archaeology of Society in the Holy Land 190ndash204 London Leicester University Press
Bar-Yosef O and Belfer-Cohen A (1992) From foraging to farming in theMediterranean Levant In ABGebauer and TDPrice (eds) Transitions to Agriculture in Prehistory 21ndash48 Madison Prehistory Press Monographs in World Archaeology4
Bar-Yosef O and Khazanov A (1992) (eds) Pastoralism in the Levant Archaeological Materials in Anthropological Perspectives Madison Prehistory Press Monographs inWorld Archaeology 10
Bar-Yosef O and Kislev ME (1989) The pre-pottery neolithic B period in eastern Jordan Paleacuteorient 15 (2)150ndash6
Byrd B (1992) The dispersal of food production across the Levant In ABGebauer andTDPrice (eds) Transitions to Agriculture in Prehistory 49ndash61 Madison Prehistory Press Monographs in World Archaeology 4
Byrd D (1994) Public and private domestic and corporate the emergence of thesouthwest Asian village American Antiquity 59 (4)639ndash66
Engel T (1993) Charcoal remains from an iron age copper smelting slag heap at FeinanWadi Arabah (Jordan) Vegetation History and Archaeobotany 2205ndash11
Finkelstein I (1995) Living on the Fringe The Archaeology and History of the NegevSinai and Neighbouring Regions in the Bronze and Iron Ages Sheffield Sheffield University Monographs in Mediterranean Archaeology 6
Finlayson B and Mithen S (1998) The Dana-Faynan (South Jordan) EpipalaeolithicProject report on reconnaissance survey 14ndash22 April 1996 Levant 30 27ndash32
Gophna R (1995) Early bronze age Canaan some spatial and demographic observationsIn TLevy (ed) The Archaeology of Society in the Holy Land 269ndash80 London Leicester University Press
Hauptmann A (1989) The earliest periods of copper metallurgy in Feinan Jordan InAHauptmann EPernicka and GAWagner (eds) Archaemetallurgie det Alten WelttOld World Archaeometallurgy 119ndash36 Bochum Deutsches Bergbau-Museum Der Anschnitt Beiheft 7
Hauptmann A (1992) FeinanWadi Feinan American Journal of Archaeology 96510ndash12
Hauptmann A and Weisgerber G (1987) Archaeometallurgical and mining-archaeological investigations in the area of Fainan Wadi lsquoArabah (Jordan) ADAJ31419ndash37
Hauptmann A Begemann F Heitkemper E Pernicka E and Schmitt-Strecker S (1992) Early copper produced at Feinan Wadi Araba Jordan the composition of oresand copper Archaeomaterials 61ndash33
Hillman GC (1996) Late Pleistocene changes in wild plant-foods available to hunter-gatherers of the northern Fertile Crescent possible preludes to cultivation In DRHarris (ed) The Origins and Spread of Agriculture and Pastoralism in Eurasia 159ndash203 London UCL Press
Farmers herders and miners in the Wadi Faynan southern Jordan 83
Kirkbride D (1966) Five seasons at the pre-pottery neolithic village of Beidha in Jordan Palestine Exploration Quarterly 988ndash72
LaBianca OslashS (1990) Hesban 1 Sedentarization and Nomadization Berrien Springs (MI) Andrews University Press
LaBianca OslashS and Younker RW (1995) The kingdoms of Ammon Moab and Edomthe archaeology of society in late bronze ageiron age Transjordan (ca1400ndash500 BCE) In TLevy (ed) The Archaeology of Society in the Holy Land 399ndash415 London Leicester University Press
Levy T Adams R and Shafiq R (1999) The Jebel Hamrat Fidan Project excavationsat the Wadi Fidan 40 cemetery Jordan (1997) Levant 31293ndash308
al-Najjar M Abu Dayyeh A es-S Suleiman E Weisgerber G and Hauptmann A(1990) Tell Wadi Feinan a new pottery neolithic tell in southern Jordan Annual of the Department of Antiquities of Jordan 3427ndash56
Sherratt AG (1997) Climatic cycles and behavioural revolutions the emergence ofmodern humans and the beginning of farming Antiquity 71271ndash87
Simmons AH and al-Najjar M (1996) Test excavations at Ghwair I a neolithic settlement in the Wadi Feinan ACOR Newsletter 827ndash8
Wright K Najjar M Last J Moloney N Flender M Gower J Jackson NKennedy A and Shafiq R (1998) The Wadi Faynan Fourth and Third MillenniaProject 1997 report on the first season of test excavations at Wadi Faynan 100 Levant3033ndash60
The archaeology of drylands 84
5 Differing strategies for water supply and farming
in the Syrian Black Desert PAUL NEWSON
INTRODUCTION
Increasing evidence has been gathered through the twentieth century of largescalesettlement across the high plateau of the Jebel al-Arab in Syria part of the so-called Hauran during the Roman period between the first and the third centuries AD (Tate199755 Fig 51) This region is within the 200 mm rainfall isohyet which is theaccepted limit for dry farming without irrigation Especially intriguing however isevidence for apparently permanent settlements of the same antiquity on the desert plateaubeyond in the region long called the al-Harra (lsquoBurnt Landrsquo in Arabic) by the bedouin and the Black Desert by European travellers (Braemer et al 1996b1) The water management features associated with some of these sites provide the focus of thischapter A number of reasons to explain the development of permanent settlement in theHarra in the Roman period has been suggested ranging from significant changes inclimate (that is increased rainfall compared with today) to the sedentarization ofpreviously nomadic tribes Certainly the new pattern of settlement did not endure by theLate Byzantine and Early Islamic periods (the seventh to ninth centuries AD) most of thesettlements had dwindled drastically in size or had been abandoned
The region where these ancient remains are to be encountered is within the Syrian desert (the Badiyat al-Sham)mdasha huge area of arid plateau (Hamada) forming the northern part of the Arabian steppe bounded by more fertile regions to the west (the Mediterraneanlittoral) the north (the Taurus foothills) and the east (the River Euphrates) and by theNafudh desert to the south The average annual rainfall within this region declines withlatitude from c150 mm in the north to less than 50 mm in the south The topography of the region is dominated by an extensive area of lava flows and basalt rocks up to 100 kmwide and reaching 250 km in length towards the southeast making it difficult to traverse(Helms 198119) The lava emanated from a number of fissures in at least six successiveflows over a relatively short period of time These flows solidified into layers of hardbasalt
Figure 51 The Hauran and the Harra regions of Syria and other regions and sites discussed in Chapter 5
each on average 30 m thick The resulting flattish plateau is broken by a number offissure cones and dissected by a series of wadis generally flowing in an easterly directionThe wadis radiate out from the high relief of the Jebel al-Arab cutting across the lava flows and have long formed the main lines of communication across this difficult area
The archaeology of drylands 86
In a quite waterless region they also provide the main access points to water so naturally form the main foci for settlement
The Black Desert has experienced limited permanent settlement in certain periods linked to the utilization of water-harvesting techniques that have been characterized by varying degrees of sophistication Some of the earliest systems have been dated to thethird millennium BC notably at Jawa on the southeastern edge of the Jabel al-Arab (Helms 1981) and on a smaller scale at Khirbet el-Umbashi to the northeast (Braemer et al 1996a) However this chapter concentrates on the methods of water management thatcurrent evidence suggests were used at dryland settlements in the Harra during theRoman period taking three sites as case studies where I have conducted fieldwork Thefirst of these sites ad-Diyatheh is located on the western edge of the Harra near the steepdescent from the Jebel Al-Arab The second site al-Namara is located in the middle of the northern part of the Harra at the confluence of the Wadi Sham and a tributary wadiThe third site Qasr Burqursquo is on the eastern edge of the Harra near its junction with the Ruhba the large fertile alluvial plain that in season lsquohas provided highly-prized grazing for nomads from time immemorialrsquo (Braemer et al 1996b1)
FARMING THE BLACK DESERT THREE CASE STUDIES
Ad-Diyatheh
This site comprises a number of connected elements based around the well-preserved remains of a small Roman fort (Fig 52) Ad-Diyatheh lies at the junction between thesettled Hauran in the west and the Hamada steppelands to its east This demarcation linealso echoes the important 100 mm precipitation isohyet which follows the relief edge ofthe Jebel al-Arab The site is also situated along the edge of the Wadi Sham which is one of a number of wadis whose floodwaters have etched deep-sided valleys into the Jebel as they flow eastwards into the Harra In relation to this wadi the site of the fort wascarefully chosen being a small flat plateau above the north bank of the wadi withcommanding views across the Harra
Clustered around the fort and along the edge of the north bank of the wadi are theremains of c100 stone-built structures all of a similar construction and with ahomogeneous layout Many of these houses are in a very good state of preservation somehave underground rooms with extensions excavated into the wadi bank It has beenconvincingly suggested that this collection of structures forms a village ofcontemporaneous construction (Villeneuve 1986) Apart from a walled meeting orfunction area no public building of any other sort has been identified and there seems tobe no indication of a stone-lined birket or reservoir (Sadler 1993) The latter is contrary to normal expectations as ancient settlements of the Jebel region invariably have suchreservoirs However in the wadi bed south of the village and fort and continuing up ontothe right bank can be seen the entrances to some
Differing strategies for water supply and farming 87
Figure 52 Plan of the Roman-period settlement of Ad-Diyatheh and its water channels and field systems
Key Stippling marks areas of ancient fields Source Adapted from Sadler 1993
stone-lined wells which are still utilized tapping into the underground water flow of the Wadi Sham In addition there are the extensive remains of stone-cleared fields and walls (Fig 53)
The construction of the village houses on two floors with the lower floor in some buildings containing evidence for stone troughs suggests an economy based on cattlerearing as in the areas of the limestone massif in northern Syria (Tate 1992 Villeneuve1986) As the French survey of the village was coming to a close it was realized that theagricultural operation was on a very large scale Initially it had seemed like a simplediversion of wadi floodwater onto the Harra plain below the fort and village but furtherinvestigation revealed the remains of a complex floodwater farming system together withother significant features such as watermills built onto leats extending out onto the Harraplain on the northern side of the wadi course
Sadler (1993) outlined the main features of these field systems in the following terms (Fig 52) At a point 300 m downstream of the village occurs the first major diversion across the wadi bed and there are at least two other such diversions situated a further 2and 3 km downstream (Sadler 1993428) The first is by far the best preservedconstituting a long low barrage of medium-sized stones crossing the wadi bed at anoblique angle to the flow The barrage leads directly into a long straight canalconstructed on a shallow gradient which would have allowed the momentum ofwaterflow to overcome the height difference between the Harra plain and the wadi bedseveral
The archaeology of drylands 88
Figure 53 The Roman-period settlement of Ad-Diyatheh and its channel walls in the Wadi Sham viewed from the west
Photograph PNewson
metres below Although the networks have been eroded somewhat by the action offloodwaters and neglect the course of the main canal can be identified as running moreor less in a parallel direction to the main wadi As it approaches the plain a succession oftributary canals leads off it to feed water into the irrigated zones This process is repeatedwith the other diversion barrages and their associated primary secondary and tertiarycanals further downstream Two secondary canals in the first network feed their own self-contained network of smaller-sized tertiary channelways or canals These smaller tertiary canals (up to several hundred metres in length) lead off from the secondary canals to feeda number of rock-cleared fields scattered at intervals across the gently sloping plateau
The second of these secondary canals is the longest (around 3 km) and serves a larger but more dispersed field system The canal heads in a northeasterly direction almost asfar as the next wadi coming down from the Jebel the Wadi Gharaz Running from thiscanal in an easterly direction and at regular intervals along its course is a succession ofparallel tertiary canals up to 800 m in length feeding a large number of rock-cleared spaces that constitute the fields of this sub-network It has been calculated that the total area that could be irrigated by this first barrage and its associated system of primarysecondary and tertiary canals amounts to around 1200 ha (Sadler 1993431) At the endof each sub-network there is evidence for the collection of excess water into a small canal which appears to carry this along to the following network No water was wastedor allowed to erode the canals or fields by ponding
Around the slopes to the north of the village along the edges of a shallow valleyleading off from the Jebel evidence for another water-capture strategy has been recorded
Differing strategies for water supply and farming 89
This valley is at too high an altitude to be irrigated by the canals and received too littlefloodwater to allow cereals to be cultivated Nonetheless there are remains of stone-cleared fields and the vestiges of long low parallel walls some leading down the slopesat regular intervals and others at lower levels forming low terraces (Sadler 1993433)This complex appears to represent a different strategy of water management concernedwith collecting surface run-off water from higher up in the valley and leading it in a controlled way down to a terrace system where water and water-borne sediment could be captured and controlled Similar systems dating to the Roman period have beendocumented in comparable dryland environments such as the Negev desert in Israel(Evenari et al 1982) and the pre-desert plateau of Tripolitania in northwest Libya(Barker et al 1996 and see also this volume Chapter 8) The fact that much of the central part of this system at Ad-Diyatheh has been destroyed by erosion since it was abandoned testifies to the extent to which this concentration of run-off was successful when it was constructed and maintained (Sadler 1993434)
The watermills are another striking illustration of the overall success of this scheme of floodwater farming in the Black Desert These were presumably utilized for grinding thewinter cereals that were grown here on a large scale in Roman times and are stillcultivated by the local bedouin when the conditions allow The remains of eight suchmills have been located so far All display a similar construction and are usually locatedalong short mill-race canal sections leading off from the secondary canals The position of the first two mills immediately below the village settlement at ad-Diyatheh and the similarities in construction between these and the village houses and other mills that canbe found situated amongst the canal networks of the Harra plateau have led Sadler(1993435) to suggest that the field system and the village must be contemporary From acursory assessment of the pottery from the remains of the fort and village Villeneuve(1986713) concluded that the main phase of occupation lasted from the late third centuryAD to the fifth and perhaps even into the seventh centuries but this date range is by nomeans certain
Al-Namara
The site of al-Namara (Fig 54) lies some 60 km east of ad-Diyatheh fully on the Harra plateau at the confluence of the Wadi Sham and a small tributary called the Wadi SaadThe name al-Namara refers to the large basin etched into the plateau by the confluence ofthese two wadis at the basinrsquos centre is an lsquoislandrsquo of resistant rock the remains of a volcanic (basalt) plug Water flowing down from the Jebel in the spring floods poolsnaturally at points in the wadi beds around this island and lasts well into the summerTherefore the combination of ample water in an otherwise arid region and the strategicvantage point of the island has long attracted local pastoralists as well as the attention ofregimes trying to control them On the flat top of the island or lsquocitadelrsquo are the few remains of reused structures of the Roman army and a later Arab occupation whilst in thesurrounding basin and beyond are the extensive remains of water catchment systems andencampments (Figs 55 and 56) Little archaeological investigation has taken place at the site apart from an initial assessment and topographical plan of the citadel and itsimmediate environs within a 2times1 km area completed while work was being done to
The archaeology of drylands 90
construct a modern reservoir on the wadi course at the eastern end of the site (Braemer et al 1996b)
The Wadi Sham has eroded the Harra plain to an average depth of 15 m and an average width of 200 m At some points as a result the wadi floor has widened to form smallalluvial plains 500ndash1000 m wide Al-Namara is situated in an S-bend of the wadi where the erosive action of the floodwaters has created a small plain measuring some 800 mlong by 400 m wide The island intrudes into this plain rising to around 10 m in heightThe wadi has to curve its course round the island to the south as a result of which thesouth bank of the curve has gentle slopes whereas the north bank forms a cliff some 2ndash8 m high Downstream the banks lower to around 1 m in height and the distance betweenthem widens into a plain up to 1 km wide The northern terrace of the plain hassubstantial evidence for simple irrigation systems up to 1 km long whereas the southernside has more limited evidence (Braemer et al 1996b4)
Differing strategies for water supply and farming 91
Figure 54 Plan of the main irrigated zone along the Wadi Sham by the Roman-period settlement of al-Namara
Source After Braemer et al 1996b
The archaeology of drylands 92
Figure 55 View of the main irrigated zone along the Wadi Sham at al-Namara looking north from the lsquocitadelrsquo
Photograph PNewson
Figure 56 Canal 3 at al-Namara viewed from the east Photograph PNewson
Differing strategies for water supply and farming 93
The remains of around five diversion barrages have been located and these are associated with canals leading from them The four barrages on the Wadi Sham are of asimilar construction consisting of a line of large othostatic boulders positioned on anoutcrop of hard basalt placed at an oblique angle to the direction of flow of wadifloodwaters The boulders were bound together with smaller stones placed around themThe low dams thus formed obstructed the wadi course only partially in order to capturewater in a controlled manner without being liable to being destroyed by the force of theflood (The same thing can be observed at ad-Diyatheh) Three of the barrages in fact only partially cross the full width of the wadi course (Braemer et al 1996b9) The canals that lead off from these barrages seem to have served two purposes The twocanals positioned upstream from the citadel are fairly short in length (around 800 m long)and seem to have been built to capture water from the wadi and bring it to the basin southof the island where the water naturally pools Downstream of the island two barragescapture water from the wadi and their associated canals lead it to areas where fields havebeen cleared of stones and laid out for irrigation
The first canal (Canal 3 on Figure 54) lies on the northern terrace of the wadi and isaround 25 km long The length of the other canal (Canal 4) cannot be measuredaccurately because it has been partially obscured by the new reservoir The area irrigatedby the first canal lies immediately adjacent to the wadi forming a small terrace up to 100m in width and roughly 2 km long This canal follows a similar pattern in its constructionto all the other canals with a channel way 1ndash2 m wide and a wall 05ndash1 m high of medium-sized boulders edging the downslope of the channel which follows the contourof the terrace bank (Fig 56) The wall was made waterproof with the packing of smaller stones and earth in the gaps between the larger stones The floodwaters flowed behindthis low wall and were let into the fields below at certain points by the means of simplespillways At the end of the main canal is a small secondary stretch which stops theremaining water from entering back into the wadi redirecting it back into the irrigatedarea There are two further canals along the course of the Wadi Saad these are short inlength (250 m and 800 m respectively) and seem to have been constructed to capturewater to irrigate small stone-cleared fields immediately adjacent to this wadi
Qasr Burqursquo
The site of Qasr Burqursquo lies some 100 km east of ad-Diyatheh and some 70 km southeast of al-Namara At this location stone structures have been built in a natural shallow basinwhere water from surface and subsurface run-off from the surrounding slopes naturally ponds (Betts et al 19906 Fig 57) The most significant structure is a tall rectangularstone tower rising to a height of c5 m (Fig 58) the study of its arches and floor levels suggested that it may have reached a height of 13 m (Betts et al 199116) The tower is surounded by a series of rooms forming a rough square which
The archaeology of drylands 94
Figure 57 The ancient reservoir at Qasr Burqursquo Source After Kennedy and Riley 1990
Figure 58 Air photograph of Qasr Burqursquo and its reservoir looking southeast
Photograph BBewley
Differing strategies for water supply and farming 95
appears to have been constructed at a later date An analysis of the architectural details ofthe tower and surrounding structures along with the evidence from two burialinscriptions written in Greek led Svend Helms to suggest the site to be a monasticfoundation dating to the Late Antique and Early Islamic periods (Betts et al 199116) Potsherds found within the structure for the most part date to the Late Antique and EarlyIslamic periods and initially it was thought that the tower was probably constructed atthis time as the residence for a recluse (Gaube 1974) However earlier Roman potteryhas also been found in the area (Betts and Helms 19898 Betts et al 199122) and there is a strong possibility that the Roman army established themselves at the site earlierincluding perhaps constructing the tower (Kennedy and Riley 1990) Later this towercould have served as the focus for a monastery that grew up around it and changed itsfunction to that of a watch tower and possible refuge from hostile nomads Alternativelythe tower and enclosure buildings may have had a military purpose for the Arabgovernment in the sixth-century AD Ghassanid period (Betts et al 199117)
Whatever the function of these buildings it was vital for the occupants to secure anadequate all-year-round supply of water Natural ponding of flood-and rainwater occurs at this point due to a bed of more resistant basalt rock crossing the wadi valleyPresumably the quantity of water collected behind this natural barrier was not sufficientto sustain an adequate population throughout the year so the volume of ponded water was enlarged and secured by the building of a dam downstream of the buildings on topof the low ridge that was responsible for the natural pooling of floodwater within thewadi bed (Fig 57) This dam the lower courses of which still survive was composed oflarge-sized roughly dressed basalt boulders laid in a series of stretcher courses cappedby one of headers Two such walls were built about 10 m apart following the top of theridge the space between probably being filled with earth and rubble Along the sidefacing the water is evidence for plastering which together with the form of wallconstruction indicates that the dam is probably of a similar date to the tower and itssurrounding enclosure (Betts et al 199112) The edges of the reservoir are lined withlow roughly-coursed stone walls from which two short stone staircases lead down to thereservoir one on each side of it
No evidence has been found for a sluice gate of any type within the wall of the damwhich suggests that the water was not used to irrigate a network of fields in the manner ofthe impressive dam and field network further to the north at Qasr el-Gherbi near Palmyra (Schlumberger 1986) However one of the lsquoroomsrsquo of the enclosure (room 11) has been tentatively identified as a windmill (Betts et al 199117) an interpretation that if correct clearly implies that cereals were being grown in the vicinity of the settlementAround the dammed lake and its associated buildings are huge numbers of encampmentsburial cairns and corral remains along with scatters of artefacts from all periods in amanner similar to that at al-Namara Given the location of the settlement and its verylarge facility for water storage the likelihood is that the main function of Qasr Burqursquo was as an oasis settlement serving a series of northmdashsouth and eastmdashwest desert routes (Betts et al 19906)
No evidence for field systems has been found by the recent survey However in thenearby more fertile and stone-free Ruhba there are lsquofields of barley planted for animal fodder and watered by flood irrigationrsquo (Lancaster 1981 Helms 1989) beside the fixed
The archaeology of drylands 96
bedouin encampments of Feytha arRisha al-Fawq and ar-Risha al-Taht It is likely that any permanent population living at Qasr Burqursquo whether Roman or Arab was served byequivalent field systems to those situated in the Ruhba
DISCUSSION
On first consideration all three sites exhibit an attempt to impose Roman control on theperipheral regions adjacent to the settled provinces of Arabia and Syria The exercising ofcontrol would have been specifically aimed at the local populations of pastoralists thetraffic of caravans and the activities of occasional groups of bandits (Isaac 1990) Thistook the form of strategically-placed military structures at all three sites where watercould be collected and which would have attracted the travellers and pastoralists of theregion whose movements could thus be more easily monitored At present it is almost impossible to say at which sites settlement of local people preceded or post-dated the building of the Roman structures What is certain is that at all three sites substantialsettlements did occur At ad-Diyatheh this took the form of a village with substantialpermanent stone buildings At al-Namara and Qasr Burqursquo the settlements mostly consist of stone-cleared corrals for temporary structures such as tents but this does notnecessarily imply that the settlements were not of a long-term nature they could have been permanent on a year-to-year basis with families periodically changing the locationof their campsites or they could have been occupied for substantial periods of the year bydifferent family groups of nomadic pastoralists
All the sites exhibit measures for the control and use of limited supplies of water forboth drinking and agricultural purposes but the methods utilized were on different levelsand scales At ad-Diyatheh varied and sophisticated techniques were employed forfloodwater farming The systems for floodwater farming at al-Namara are simpler and smaller At Qasr Burqursquo even though floodwater farming systems have not yet been identified cereal cultivation is implied by the probable presence of a mill and the sitecertainly displays an impressive scale of planning for the storage of much largerquantities of water than at the other sites
These substantial differences in floodwater farming techniques and strategies for water storage imply that the inhabitants of these sites whether Roman indigenous bedouin orboth were dealing with different issues in building the settlements and their associatedstructures Local environmental considerations were undoubtedly one important factoraffecting strategies for water management At ad-Diyatheh for example there would have been a greater quantity of water moving at a higher speed of flow than at the othertwo sites requiring substantial and carefully designed structures to allow some of thiswater to be brought under control The topography at ad-Diyatheh at the juncture of the Jebel slope and the gently descending alluvial fan below it allowed also for theconstruction of a large extensive network of irrigation channels and a large dispersedspread of fields The speed of waterflow would have decreased at al-Namara 60 km downstream of ad-Diyatheh and diverting the waterflow would have been especiallyimpeded by the fact that at this point the wadi bed lies up to 15 m below much of thesurrounding plateau so irrigation is really effective only along areas of adjacent
Differing strategies for water supply and farming 97
floodplain On the other hand Qasr Burqursquo lies in a natural dip within the surroundingplateau with the majority of water collecting here through groundwater seepage(Lancaster and Lancaster 1999132)
A second area of differences though one harder to identify with the archaeologicalevidence available relates to the social and economic conditions of the inhabitants of thethree settlements At ad-Diyatheh the field system can be directly related to the substantial permanent village at the site whose inhabitants relied to a great extent uponthe cereals they grew judging by the number of mills In addition construction of the dwellings echoes in form structures found on the Jebel and other regions of Syria in theprovision of accommodation for livestock below the family rooms (Villeneuve 1986)The reliance on crops for subsistence would have been much less for the pastoralpopulations postulated for al-Namara and Qasr Burqursquo and so the need for large farming systems proportionally less though the surrounding Harra could have supported onlylow-density populations
At both al-Namara and Qasr Burqursquo it may well have been only the residents of thesubstantial permanent structures be they soldiers or monks who attempted to grow cropson any scale The pastoral populations at these settlements may have grown cereals onany scale only when favourable climatic conditions allowed so would not have beeninclined to spend valuable time in investing in the substantial infrastructure requiredbeforehand using the locations simply as convenient watering places (Macdonald 1993)Caution is needed in advancing such an hypothesis however for the modern bedouin ofthe region in fact try to irrigate areas when settled in a particular location over a period oftime (Lancaster and Lancaster 1999132ndash66) It is probably reasonable to assumethough that the occupants of the substantial structures were generally more reliant onprovisions provided from the immediate area given that pastoralists would have had thecapacity to move to new areas when they had exhausted supplies at any particularlocation
It may simply be the case that people with different traditions and cultures were responsible at various periods of time for constructing the different systems of watercontrol and floodwater farming at all three sites The people at ad-Diyatheh may have come from more settled areas on the Jebel though the poor construction of the houses atad-Diyatheh compared with the more refined architecture of the adjacent Jebel villagessuggests that ad-Diyatheh was built by lsquosedentarized nomadsrsquo (Villeneuve 1986710) The small field system at al-Namara could have been constructed either by a small military garrison or by semi-sedentary pastoralists The water storage system at QasrBurqursquo may well be a later enhancement of a much older catchment system
Whatever the origins of these systems though the water management techniquesemployed in the Roman period at both ad-Diyatheh and al-Namara reveal an agricultural culture not extensively practised in the region since the Bronze Age a fact stronglysuggesting that these new approaches to agriculture came as a direct consequence ofpolitical and social changes brought about by the imposition of Roman hegemony Theextent to which these new agricultural regimes were imposed created and operated by theRoman army or reflect a variety of complex responses by indigenous populations to theopportunities of Romanization still remains unclear However as the precedingdiscussion has shown the likelihood is that the archaeology of the Syrian Black Desert
The archaeology of drylands 98
reflects both the external imposition of and indigenous adaptions to the forces of Romanimperialism
ACKNOWLEDGEMENT
I would like to thank Bob Bewley for permission to reproduce as Figure 58 his airphotograph of Qasr Burqursquo and its reservoir from his Aerial Archaeology in Jordan Project
REFERENCES
Barker G Gilbertson D Jones B and Mattingly D (1996) Farming the Desert The UNESCO Libyan Valleys Archaeological Survey Volume One Synthesis Paris UNESCO London Society for Libyan Studies Tripoli Department of Antiquities
Betts A (1993) The BurqursquoRuweishid Project preliminary report on the 1991 fieldSeason Levant 251ndash11
Betts AVG and Helms SW (1989) A water harvesting and storage system at Ibn al-Ghazzi in eastern Jordan Levant 213ndash11
Betts A Helms S Lancaster W Jones E Lupton L Martin L and Matsaert F(1990) The BurqursquoRuweishid Project preliminary report on the 1988 field season Levant 221ndash20
Betts A Helms S Lancaster W and Lancaster F (1991) The BurqursquoRuweishid Project preliminary report on the 1989 field Season Levant 237ndash28
Braemer F Echallier J-C and Taraqji A (1996a) Khirbet el Umbashi (Syrie) Rapport preacuteliminaire sur les campagnes 1993 et 1994 Syria 73117ndash29
Braemer F Echallier J-C Hatoum H and Macdonald MCA (1996b)Archaeological and Epigraphic Rescue Survey at Al-Nemara Report on the First Season Sept-Oct 1996 Damascus Department of Antiquities and Museums of Syriaunpublished report
Evenari M Shanan L and Tadmor N (1971) The Negev The Challenge of a DesertCambridge Mass Harvard University Press
Evenari M Shanan L and Tadmor N (1982) The Negev The Challenge of a DesertCambridge Mass Harvard University Press second edition
Gaube H (1974) An examination of the ruins of Qasr Burqursquo Annual of the Department of Antiquities of Jordan 1993ndash100
Helms SW (1981) Jawa Lost city of the Black Desert London Methuen Helms SW (1989) Jawa at the beginning of the Middle Bronze Age Levant 21 141ndash68 Isaac B (1990) The Limits of Empire The Roman Army in the East Oxford Oxford
University Press Kennedy D and Riley D (1990) Romersquos Desert Frontier From the Air London
Batsford Lancaster W (1981) The Ruwala Bedouin Today Cambridge Cambridge University
Press
Differing strategies for water supply and farming 99
Lancaster W and Lancaster F (1999) People Land and Water in the Arab Middle EastEnvironments and Landscapes in the Bilacircd ash-Shacircm Studies in Environmental Anthropology Volume 2 Amsterdam Harwood Academic Publishers
Macdonald MCA (1993) Nomads and the Hawran in the Late Hellenistic and Romanperiods a reassessment of the epigraphic evidence Syria 70303ndash413
Sadler S (1993) Le terroir agricole de Diyateh lrsquoirrigation comme condition drsquoexistance de ce terroir In BGeyer (ed) Techniques et Pratiques Hydro-Agricoles Traditionnelles en Domaine Irrigueacute Approche Pluridisciplinaire des Modes de Culture avant la Motorisation en Syrie (Actes du Colloque de Damas 27 juin-1 juillet 1987) 421ndash51 Paris PGeuthner Bibliothegraveque Archeacuteologique et Historique 136
Schlumberger D (1986) Qasr el-Heir el-Gharbi Paris PGeuthner Bibliothegraveque Archeacuteologique et Historique 120
Tate G (1992) Les Campagnes de la Syrie du Nord di IIe au VIIe Siegravecle Un Exemple drsquoExpansion Demographique et Economique a la Fin de lrsquoAntiquiteacute Paris PGeuthner Bibliothegraveque Archeacuteologique et Historique 133
Tate G (1997) The Syrian countryside during the Roman era In SEAlcock (ed) The Early Roman Empire in the East 55ndash70 Oxford Oxbow Books
Villeneuve F (1986) Ad-Diyatheh village et castellum romains et byzantins a lrsquoest du Jebel Druze (Syrie) In PWMFreeman and DLKennedy (eds) Defence of the Roman and Byzantine East Volume II 697ndash715 Oxford British Archaeological Reports International Series 297
The archaeology of drylands 100
6 Irrigation agriculture in Central Asia a long-
term perspective from Turkmenistan MARK NESBITT AND SARAH OrsquoHARA
INTRODUCTION
Agriculture in the newly independent republics of the former Soviet Central Asia isalmost entirely dependent on irrigation Consequently access to water is essential and ithas long played an important role in the social environmental economic and politicalsituation of the region Today as in the past agriculture represents the single mostimportant economic activity throughout the region and currently over 40 per cent of thepopulation is employed in the commercial agricultural sector with the vast majority ofCentral Asians either partially or wholly dependent on subsistence agriculture Theagricultural sector throughout Central Asia however is under threat because of the rapiddeterioration in the water distribution and irrigation since the collapse of the SovietUnion (OrsquoHara in press)
Central Asia boasts a long history of irrigated agriculture but the exploitation of theregionrsquos water resources and the expansion of the irrigation network peaked during the latter part of the Soviet era During this period huge water diversion and irrigationprojects were constructed to satisfy Moscowrsquos continual demands for cotton In order to maximize agricultural output water was taken from areas of surplus to those of deficitoften involving transfers over considerable distances and in some case from otherrepublics Today however this huge highly integrated network serves five independentstates each following its own agenda for reform The implications for the regionrsquos water resources are immense and it is becoming increasingly difficult to reach a consensus onhow the water distribution and irrigation system should be managed and maintained(Bedford 1996 OrsquoHara in press) Further complicating the matter is the fact that CentralAsiarsquos irrigation zones are plagued by secondary salinization and high water tables (OrsquoHara 1997 Smith 1992) and it is evident that these large-scale Soviet-built systems are environmentally unsustainable The situation is not likely to improve and indeedcould be exacerbated by changing land and agricultural policies coupled with anincreased demand for water as population rises Should the system fail the consequences would be enormous and could ultimately undermine regional security The question ofsustainable irrigation is therefore urgent Given that Central Asia not only has a longhistory of irrigated agriculture but has witnessed the rise and fall of a number of majorempires over the last few thousand years it may well be that lessons can be learned fromthe past An assessment of former irrigation and water management practices mayhighlight whether sustainable irrigation is a feasible option and if so how it might be
achieved Here we review the history of settlement agriculture and irrigation over some 8000
years in southern Turkmenistan (Fig 61) A large body of archaeological evidence isavailable for this region much of it resulting from the establishment of the SouthTurkmenistan Multi-Disciplinary Archaeological Expedition (YuTAKE) in 1946 Many of its publications were not widely distributed even within the former Soviet Union butwe have been able to draw on a wide range of useful syntheses published in westernjournals A more recent phase of fieldwork involving a number of international researchteams has resulted in a series of renewed excavations at several important sites includingJeitun Anau Gonur Depe and Merv Although many of these projects are ongoingimportant papers pertaining to the area have emerged (Harris et al 1993 1996 Herrmann 1997 Herrmann et al 1998 Hiebert 1994) providing valuable informationon changes in environment and society over this period Historical sources are moreproblematic Although literate civilizations have existed in the region since theAchaemenid period there is no systematic body of texts comparable to the clay tablets ofMesopotamia For the medieval period we are largely dependent on short descriptions inaccounts by Arab or Chinese travellers or Arab historians Some Sasanian records havesurvived through their use by the Arab historians Prior to this we are again dependent onbrief travellersrsquo accounts and histories compiled far away to the west in classical Greece and Rome Our understanding of the political dynamics underlying the increasingly well-documented settlement archaeology is therefore currently less sophisticated than in theNear East proper
ENVIRONMENT
Turkmenistan covers an area of 480000 km2 90 per cent of which is covered by thevirtually uninhabited Kara Kum Desert (Babaev 1996) Most of Turkmenistan compriseslowlands with mountains being confined to the southern and western parts of thecountry It lies within the temperate desert zone (Babaev 1994) and has a markedcontinental climate (Orlovsky 1994) Precipitation mainly falls as snow or rain in winterwith almost none in the agriculturally-active summer months of June through to September Average annual precipitation varies from 90 mm in Dashouz to nearly 400mm in the southwest highlands of the Kopet Dagh but in much of the country it
The archaeology of drylands 102
Figure 61 Turkmenistan showing locations mentioned in Chapter 6
is less than 200 mm per annum and therefore insufficient for dryland agriculture Averagetemperatures are high varying from 12 to 18degC The coldest months are December toFebruary with temperatures frequently falling below 0degC and the hottest months June to August when temperatures often exceed 45degC Potential evaporation rates vary accordingly from 1ndash2 mm per day in winter to 10ndash15 mm per day in summer Total annual potential evaporation rates are of the order of 2500ndash3000 mm which are far higher than the precipitation rates
The hydrological network is weakly developed and all major sources of water rise outside the countryrsquos borders (Fig 61) The headwaters of the Amu Darya the largest river in Central Asia are in the Pamirs and the river flows through a number of countriesbefore discharging into the Aral Sea It displays two periods of peak discharge oneduring the spring associated with snow melt the other later in the summer when ice meltincreases its flow The other main rivers all rise in the mountains to the south the Atrekflowing into the Caspian Sea and the Murgab and Tejen draining into the Kara KumDesert Although small when compared with the Amu Darya they are an importantsource of water and have long been used by people occupying the region Fed by winterrains and snowmelt they have only one period of peak discharge during the spring Inaddition to these rivers there is a number of smaller intermittent rivers and springs mostof which cease to flow during the summer Today as in the past human settlement inTurkmenistan is concentrated in two zones the piedmont at the foot of the steep slopes ofthe Kopet Dagh mountains and the desert oases
Irrigation agriculture in Central Asia 103
THE NEOLITHIC EARLY FARMING IN THE PIEDMONT
The beginning of agriculture in Turkmenistan is best documented by the importantexcavations at Jeitun (Djeitun) located some 25 km north of Ashgabat in the piedmontbetween the Kopet Dagh mountains and the Kara Kum Desert First discovered in theearly 1950s Jeitun has been the subject of a number of detailed excavations that haveproduced one of the bestknown archaeological sequences in Central Asia Todayfourteen such Jeitun culture sites have been identified across southwestern Turkmenistan
Jeitun comprises some thirty excavated small rectangular mudbrick houses located on the distal reach of the Kara Su (Black Water) a small ephemeral stream that rises in theKopet Dagh and discharges into the Kara Kum Desert The settlement covers less than07 ha and estimates of the cultivated land surrounding it in the neolithic period arebetween 15 and 33 ha First excavated by Masson in the 1950s and 1960s the site wasdated to c6000 BC on the basis of ceramic assemblages (Masson and Sarianidi 1972) This date was later confirmed by a series of eleven radiocarbon dates from the British-RussianmdashTurkmen excavations of 1989ndash94 which indicates that the site was occupiedbetween c6300 and 5600 cal BC (Harris et al 1993 1996)
Recovery of animal bones and charred plant remains from these new excavations hasallowed a reassessment of the sitersquos subsistence base The results confirm earlier evidence for a primarily agricultural system of subsistence based on cereals and domesticsheep and goat augmented by hunting primarily of gazelle The cereals are dominatedby einkorn wheat (Triticum monococcum) with small amounts of emmer (Tdicoccum)and naked and hulled forms of barley (Hordeum sativum) Other artefacts from the site point to the importance of cereal cultivation for the inhabitants of Jeitun with sickleblades accounting for 37 per cent of all tools found in Massonrsquos early investigation of the site (Masson and Sarianidi 1972) In addition to cereal cultivation the inhabitants ofJeitun herded goats and sheep faunal analysis shows that although raised primarily formeat these animals could also have been an important source of milk wool hair andskins The dominance of domesticated plants and animals from the very bottom of theJeitun sequence together with the absence of wild progenitors of wheat and sheep inCentral Asia supports the view that agriculture and its attendant domesticated species didnot evolve independently in the region but rather reached it from the Fertile Crescent ofsouthwest Asia via the Zagros mountains of Iran (Harris and Gosden 1996)
Jeitun is often cited as one of the oldest known sites of irrigation in the world (Dukhovny 1995 Harris et al 1993 Lisitsina 1984) There is however some difference in opinion as to how crops were irrigated at this time Lisitsina (1981) forexample assumed that cultivation at Jeitun was entirely dependent on run-off from the Kopet Dag with Lewis (1966) suggesting that Jeitunrsquos location on the distal reaches of the Kara Su was due to the fact that neolithic farmers were better able to control andmanipulate flows in this part of the river system However Kohl (1981) argued thatJeitun was in fact located on the distal reaches of the Tejen Delta which at this timedischarged further into the Kara Kum Desert than today The presence of many seeds ofthe weeds club-rush (Scirpus maritimus) and goat-face grass (Aegilops tauschii) in
The archaeology of drylands 104
association with the charred cereal remains led Harris et al (1993) to conclude that cereals were being grown in areas with high water table and high salinity (lsquotakyrsrsquo) rather than on stream sides irrigated by less saline floodwaters Takyrs are highlyimpermeable almost flat clay surfaces that retain water and are of considerableimportance to communities living in the desert today All these different scenarios wouldlargely draw on naturally irrigated land with only relatively small-scale channels or embanking required No definite evidence of such features has been discovered thoughthis may in part reflect subsequent processes of deposition or erosion
The belief that early agriculturists at Jeitun irrigated their fields however is based largely on the assumption that the climate during the Neolithic was similar to todayThere is some evidence to suggest that this assumption may be unfounded for the base ofthe dunes overlying fluvial deposits at Jeitun has yielded an OSL date of c4500ndash5000 years BP so it is possible that arid conditions similar to today developed somewhat laterthan previously thought (just as wetter environments characterized the early Holocene inthe Levant Chapter 4) Further support for this hypothesis is provided by a recentanalysis of plant remains from the site suggesting that cultivation may have beenpossible without irrigation (MCharles pers comm) Despite this uncertainty though it is evident that there is a long history of agriculture in this region and that by the fifthmillennium BC agricultural settlements were spread along the piedmont from KyzalArvat in the west to Tejen in the east
ENEOLITHIC TO IRON AGE PIEDMONT SETTLEMENT AND EXPANSION TO THE OASES
The establishment of agricultural communities such as Jeitun in the Neolithic wasfollowed by several millennia of continuing settlement largely in the piedmont of theKopet Dagh during the Eneolithic or Chalcolithic (c4800ndash3000 BC) Early Bronze Age (c3000ndash2500 BC) and Middle Bronze Age (c2500ndash1900 BC) Sites grew significantly in sizemdasheneolithic settlements such as Altyn Depe Anau and Namazga cover up to 25 hamdashand there were two key changes in settlement pattern the expansion into the Geoksyuroasis in the Eneolithic and the emergence of state-level urbanism in the piedmont zone in the terminal Early Bronze Age
The Geoksyur oasis (Fig 61) is situated on the Tejen River delta and unlike otheroases in the region is contiguous with the foothills of the Kopet Dagh Nine prehistoric sites comprising large widely scattered mudbrick houses have been found in the oasisdating from the earliest Eneolithic (Kohl 1984) but the oasis appears to have beenabandoned by the end of the Eneolithic Earlier eneolithic settlements are at the end ofbranches of the river delta suggesting that a modified form of floodwater irrigation waspractised Later in the Eneolithic well-developed artificial irrigation systems aredocumented for the first time in Turkmenistan (Namazga III period c3500ndash3000 BC) Aerial photographs and excavations have shown that land around the site of Geoksur Iwas irrigated by three parallel canals each up to 3 km long and 5 m wide possiblyirrigating an area of about 50 ha by means of small aryks (irrigation canals) branching off and leading to fields (Lisitsina 1969) The water flow into secondary canals was
Irrigation agriculture in Central Asia 105
controlled by inlet structures where they joined the main canals (Lisitsina 1981) In the piedmont proper the last part of the Early Bronze Age witnessed a
transformation of settlements with the appearance of specialized production areasfortification walls around settlements increased status differentiation in burials andevidence of much interaction between settlements throughout the Kopet Dag region allconsistent with a state-level society (Hiebert 1994) These trends continued into theMiddle Bronze Age and by its terminal phase (2200ndash1900 BC) the foothills contained a number of very large sites such as Namazga (50 ha) and Altyn Depe (25 ha) This periodof expansion came to an end in the Late Bronze Age The settlements at Anau andNamazga for example were considerably smaller now covering only a few hectaresRelatively little is known about agriculture in the piedmont zone in the Eneolithic andBronze Age Irrigation canals have not been located in the piedmont but this may reflectdeposition and erosion in this geomorphologically-active zone The presence of bread wheat and six-row hulled barley in lencolithic samples from Anau dated to c4500ndash3000 BC has been cited as possible evidence for irrigation (Miller 1999) but both cereals weregrown in many regions of the Old World without irrigation (Maier 1996)
Paralleling the decline of settlement on the northern piedmont was the spread of irrigation to the lower reaches of the Murgab river at the end of the Middle Bronze Agealthough this occurred while some sites such as Altyn Depe were still very large Anumber of factors has been cited for this shift in agricultural settlement Masson (1957)for example suggested that a rise in population stretched resources to the extent thatpeople were forced to migrate whilst some authors have highlighted the potentialimpacts of climate change Lewis (1966) argued that there is no evidence of a major shiftin climate during this period but as mentioned above evidence is emerging for a shift todrier conditions c5000ndash4500 years ago coinciding with the rise of agriculture in theMurgab oasis It is possible therefore that conditions became sufficiently dry toprecipitate change
The bronze age settlements of the Merv oasis covered an area of 100 km northmdashsouth by 50 km eastmdashwest which is almost five times larger than the later medieval andclassical oasis to the south Hiebertrsquos recent re-analysis of the ceramic chronology and survey data suggests that the colonization of the oasis was rapid (Hiebert 1994) Thesites cluster in lsquomicro-oasesrsquo forming linear patterns that presumably followed old river branches (Fig 62) Settlements are characterized by large fortified building complexes with intervening fields which as Hiebert points out typify Central Asian oasisarchitecture of the time Initially settlements were located on the northern margins of theoasis with the system expanding southwards some 400 years later (Hiebertrsquos Gonur Period 3) Initial settlement was at the northern fringe of the oasis because large-scale canals were not used Instead fields were irrigated by ditches carrying water from thesmaller streams into which the Murgab river split near the edge of the delta As Bader et al (1996) comments settlers from the Kopet Dagh would already have been familiar with the technology of using streams of the piedmont
Archaeobotanical analysis indicates that over time greater numbers of plants andanimals were domesticated By the Bronze Age the variety of crops grown had increasedsignificantly compared with in the Neolithic samples from the middle bronze age site ofGonur Depe in the Merv oasis for instance are dominated by hulled and naked barley
The archaeology of drylands 106
with free-threshing wheat lentils peas chickpeas and grape also present (Miller 1993
Figure 62 Bronze and iron age settlement in the Merv oasis (adapted from Hiebert 1994)
Irrigation agriculture in Central Asia 107
Moore et al 1994) These finds are consistent with those from the neighbouring Geoksyur oasis (Lisitsina 1969 Lisitsina and Prishchepenko 1976) and are typical ofbronze age settlements throughout the Near East Late bronze age samples from TahirbajTepe in the Merv oasis were also dominated by hulled barley but add broomcorn millet(Panicum miliaceum) to the repertoire of crops (Nesbitt 1994)
Iron age settlements in the Kopet Dagh foothills are widely distributed and often continue on the same sites as bronze age settlements but are smaller and marked by lessmaterial complexity (Kohl 1984) In the Merv oasis iron age sites are concentrated infour lsquomicro-oasesrsquo The northernmost two Takhirbai and Togolok contain bronze agesettlements while the southernmost two Yaz depe and Aravali represent newoccupation thus forming part of the pattern of southward movement of settlements thatcontinues until the Achaemenid period (Bader et al 1996 see below) This shift in settlements is most plausibly explained by increased extraction of water upstream bysettlements using more sophisticated canal systems collecting water near the head of thedelta However early sites in the upper part of the oasis may have been masked byalluvial deposition accounting in part for this pattern
ACHAEMENID TO MEDIEVAL URBAN SOCIETIES
The Achaemenid period (530ndash330 BC) marked two important transitions for the Merv oasis it was the first of several periods when Merv came under the control of an empirebased to the south and for the first time a series of urban centres emerged in the oasisFrom this time onwards Merv was also militarily important as a frontier city at thenortheastern part of firstly the Achaemenid and later the Seleucid (330ndash140 BC) Parthian (140 BCndashAD 220) and Sasanian (AD 220ndash651) empires Surveys of the magnificentruins of Mervrsquos urban centre show a steady increase in its size The earliest city Erk Kala had walls enclosing an area of 20 ha It later became the citadel of the adjoiningSeleucid city of Gyaur Kala (400 ha) (Fig 62) which continued to be occupied for a period of over a millennium even after the construction of the nearby city of Sultan Kalain the eighth century AD Survey work in rural areas in the north of the oasis confirmsthis basic pattern of expansion with increasing residential areas from Achaemenid toSeljuk times (Bader et al 199394 Gubaev et al 1998) At its greatest extent the oasis covered c700 km2 The area cultivated appears to have fluctuated with a decrease inrural settlement in the Hellenistic period and a marked increase in cultivation andprobably the first construction of a large central dam and canal network in the Parthianperiod
Although written sources state that Merv was destroyed by Mongol invasions in AD1221ndash2 there is archaeological evidence for a substantial post-Seljuk occupation and in the early fifteenth century a new much smaller city was built by the TimuridsNotwithstanding this the oasis declined in importance and the Timurid city wasabandoned by the nineteenth century Overall therefore changes in settlement patternsuggest three key phases in the occupation of the Merv oasis the initial colonization by dispersed but numerous bronze age settlements c2200 BC urban development in theAchaemenid period c600 BC and the gradual abandonment of intensive settlement in
The archaeology of drylands 108
most of the oasis in the centuries after the Mongol invasions of AD 1221ndash2 The large-scale sampling of contexts carried out at the city of Merv by the
International Merv Project (Boardman 1997 Nesbitt 1994) has provided some the bestarchaeobotanical evidence for the Sasanian period It indicates that during the LateSasanian period (the fifth to seventh centuries AD) cereals consisted as before ofabundant hulled barley and free-threshing wheat and rarer broomcorn millet lentils(chickpea seems to disappear after the Bronze Age) very abundant cotton seed and awide range of fruits and vegetables including cucumbermelon grape almond peach andnuts Two changes are apparent in comparison with the Bronze Age first an increase incrop diversity particularly in the fruits and second and more importantly the addition ofcotton which is a source of both textile and oil and like millet a crop that expands thegrowing season through the summer after the wheat and barley harvest Overall therange of crops seems similar to that mentioned in Islamic times in the tenth centuryMervrsquos famously soft cotton textiles were exported as far as Africa and Spain and thereare thirteenth-century references to Mervrsquos fine grapes and other fruits (Serjeant 1972)
Of the range of crops grown in the Sasanian period barley and cotton are relativelytolerant of soil salinity (though not of course of heavily salinized soils) bread wheat ismoderately tolerant melon and grape are moderately sensitive and almond and peach aresensitive (Maas 1987) That crops sensitive to salinity are present throughout the lateSasanian sequence from Merv and that the full suite of crops is present in broadly similarquantities throughout the sequence strongly suggest that irrigation agriculture wassustained through this period without occurrence of catastrophic salinization This isconsistent with evidence for unbroken intensive settlement in the oasis from the Parthianto Seljuk periods Although salinization has often been viewed as an inherent andimminent threat in ancient irrigation systems in the Near East particularly inMesopotamia (Jacobsen and Adams 1958) there is increasing evidence that soilmanagement practices that avert salinization (and which are ethnographicallydocumented in Mesopotamia) were applied effectively in the past (Powell 1985)
The success of Merv and other settlements in the region depended to a large extent on how water resources were managed There were two major technological innovationsduring the urban period In the foothill zone the qanat was introduced in the first millennium BC Like the foggaras of the Sahara (Chapter 9) this system allows groundwater to be tapped by underground tunnels cut into the foothills and is mostwidely used in the highland area of Iran Qanats are difficult to date directly butassociations with sites suggest that they became widespread in the highlands of Iran andneighbouring areas at this time In the Merv oasis there is indirect evidence from survey work of large state-sponsored irrigation works in the Parthian and Sasanian periods (Bader et al 199394 1996 Gubaev et al 1998) like the contemporary transformationsoccurring further south in Susiana (Wenke 1975ndash76) and Mesopotamia (Adams 1981)Major changes in irrigation technology in the Merv oasis are therefore later (if thedating is correct) than the first urbanization at Erk Kala and Gyaur Kala but coincide withthe increase in the population of the oasis in the Parthian and Sasanian periods
In contrast with the bronze age canals these later irrigation systems are difficult to investigate because they have been largely destroyed by twentieth-century agriculture but the medieval irrigation system of the tenthtwelfth centuriesmdashduring which time the
Irrigation agriculture in Central Asia 109
oasis nourishedmdashmay give a good parallel The Arab historians and geographers such asMuqaddasi Al-Biruni and Yakut provide valuable accounts of water distribution and irrigation systems (see Bartold 1914 1928 and Le Strange 1905 for translations anddiscussions of their works) It is evident from these that the administration of scarcewater resources was central to the way in which the social and political hierarchy ofsettlements operated water was viewed as a lsquogift from Godrsquo that could not be owned or controlled by an individual The city of Merv had access to only one source of watermdashthe Murgab river which rises in the Afghan mountains and drains northward into the KaraKum Desert The riverrsquos annual discharge is about 12 km3 which is approximately 5 per cent of the total amount of water available for use in Turkmenistan at present (OrsquoHara 1997)
The oasis was renowned for its productivity not only producing enough food to feed its large population but exporting produce to adjacent areas (Herrmann and Petersen1997) The regionrsquos agricultural success was in part due to the land and watermanagement strategies of the time Land for example was divided into small plots thatwere intensively cultivated receiving water on a regular basis The amount of landcultivated in any given year depended on water availability Muqaddasi writing in thetenth century AD described how a depth gauge situated at the Razik Dam to the south ofthe city was used to determine whether there would be a surplus or deficit of water thatyear If the level reached the 60th point water would be plentiful that year and the orderwould be given to increase the amount of land cultivated whereas in years of low wateravailability the area was reduced and only the best lands were cultivated The dam wasextremely important and was in effect the only water storage facility for the city Itsmaintenance was assured by 400 divers employed around the clock each diver having todeliver a specified amount of wood and mud to the dam each day (Bartold 1914) Yakutwho resided in the city at its zenith in the early thirteenth century provides furtherdetails He described how water gauges were installed at the head of every canalthroughout the city The whole system was headed by the mirab bashi (chief water master) and hourly reports on the water level in the main canal were passed to his officeso that he could decide which off-takes were to be opened or closed The system was managed by elected senior officials and maintained by over 12000 workers paid by the water users who were also expected to take part in major construction schemes and in theannual maintenance programme
EXPANDING THE IRRIGATION SYSTEM THE TSARIST AND SOVIET PERIODS
When Central Asia finally came under Tsarist control in the late 1880s the newadministration attempted to introduce reforms in the irrigation sector These failedhowever and the authorities declared that irrigation would be run lsquoby customrsquo Notwithstanding this a number of subtle changes was made most important irrigationofficials became part of the Tsarist civil service and as such were no longer controlled bywater users This act severed the link between water users and providers so effectivelyundermining the traditional system of water management State salaries for officials were
The archaeology of drylands 110
low and there was no longer any incentive to control the system The situation wasexacerbated by the imposition of irrigation officials unaccustomed to the traditionalmethod of management resulting in increased problems within the system whichbecame subject to corruption and abuse by the wealthy and more powerful water users
More significant than Tsarist interventions in water management however were thechanges in agricultural policies The authorities in Moscow keen to end their reliance onAmerica for cotton (particularly following the American Civil War when supplies almostceased) recognized that Central Asia had the potential to become a major cotton growingregion in fact the main factor behind initiatives to increase the amount of land irrigatedwas cotton production (Lipovsky 1995) The subtle but nonetheless important changesin water management coupled with increased demand and use of water appear to havecaused widespread land degradation In the Merv oasis for example the irrigationnetwork was expanded by some 33000 ha (Zaharchenko 1990) but poor management ofthe system caused local water tables to rise resulting in salinization and widespreadsurface ponding that not only degraded the soils but also led to outbreaks of malaria(Pierce 1960)
The Bolshevik Revolution and the subsequent emergence of the Soviet Union heralded a period of radical change in the way water was managed in Central Asia In 1923 theSoviet administration decreed that water management was to be taken lsquoout of the hands of traditional elders and councils with whom it residedrsquo (Black et al 1991) and like land was to become a common resource for the benefit of all Various governmentbodies were established to be responsible for the development of a regional watermanagement strategy that would allow centrally-determined production targets to be met With cotton production the priority for Moscow huge sums of money were invested inthe region in the development of massive highly integrated systems of water distributionand irrigation (Micklin 1991) Land was irrigated no longer by a single local source as in the past but by water often piped over considerable distances the Kara Kum Canalfor example considered to be one of the engineering feats of the Soviet era now transfersin excess of 129 km3 of water from the Amu Darya along its 1400 km length every yearirrigating an area of c1 million ha (Hannan and OrsquoHara 1998)
There has been much criticism of the management and maintenance of Soviet irrigation systems and the inefficiency of water use (eg Micklin 1991) Losses occurredthroughout the system with problems of seepage and evaporation from the manythousands of kilometres of unlined irrigation canals creating huge problems withwaterlogging and soil salinization Within a few years of the Kara Kum Canal beingconstructed for instance the water table in the Merv region had risen over 20 m(Kornilov and Timoshinka 1975) and vast tracts of land had become salinized (OrsquoHara 1997) Water use at the field level also rose as field size increased to accommodateincreasingly bigger agricultural machinery not only increasing the amount of time that ittook to water fields but also causing the traditional practice of night-time watering to be replaced by daily and often continuous twenty-four-hour irrigation Yet despite an emphasis on the need to modernize the agricultural sector furrow irrigation continued todominate with large and poorly levelled fields creating huge problems for irrigatorsMoreover unlike in the past access to water was not a problem with diversion schemesbringing what to many seemed an infinite supply of free water people who had long
Irrigation agriculture in Central Asia 111
viewed water as a scarce commodity forgot its worth and wasted it Further exacerbating the situation was the fact that government agencies rather than
individual water users were responsible for amongst other things maintaining theirrigation infrastructure dredging canals and ensuring that the drainage system was cleanAt the farm level maintenance became the responsibility of a few collective workers Inall cases the bulk of the work was done using heavy equipment Consequently waterusers had little if anything to do with the management or maintenance of the waterdistribution and irrigation system Despite Soviet successes in expanding the irrigationnetwork and increasing agricultural output the systems they built were (and still are)inflexible and highly inefficient By the 1980s agricultural land in Turkmenistan wasbeing abandoned at a rate of over 50000 ha per annum (Zaharchenko 1994) cleartestimony to the fact that this huge irrigation system is not sustainable
CONCLUSION
In Tables 61 and 62 we summarize the major trends in settlement and agriculture insouthern Turkmenistan It is evident that there is a strong correlation between the degreeof urbanization and population size (themselves correlates of centralized political control)and the sophistication of irrigation technology The range of crops likewise increasesthrough time Although
Table 61 Simplified chronological chart of prehistoric settlement in Turkmenistan
Archaeological period and date (cal BCAD)
Settlement Irrigation systems Crops
Neolithic 6300ndash4800 BC
Small farming villages on piedmont of Kopet Dagh Key site Jeitun
Crops cultivated in areas of high water-table possibly also simple diversion of streams
Main crop einkorn also emmer hulled and naked barley
Eneolithic (Chalcolithic) 4800ndash3000 BC
Larger complex settlements to 15 ha with shrines and fortifications in Middle Eneolithic (4000ndash3500 BC) spread of settlements to Geoksyur oasis in Middle Eneolithic but abandonment of oasis by EBA Key sites Altyn-depe Anau Geoksyur Namazga (phases IndashIII)
Simple irrigation assumed for piedmont and early occupation of Geoksyur large irrigation canals identified in Geoksyur oasis in late Eneolithic (Namazga phase III)
Hulled barley free-threshing wheat
Early Bronze Sites to 25 ha restricted to Irrigation assumed Hulled barley
The archaeology of drylands 112
Age (EBA) 3000ndash2500 BC
piedmont zone Key sites Altyn-depe Namazga (IV)
for large settlements in piedmont but no direct evidence
free-threshing wheat grape
Middle Bronze Age (MBA) 2500ndash1900 BC
Sites to 50 ha monumental architecture Abandonment of piedmont sites at end of MBA Major fortified settlements appear in Merv oasis in terminal period (2200ndash1900 BC) Key sites Altyn-depe Namazga (V) Gonur depe
Smaller-scale irrigation at northern fringe of Merv oasis
Main crop hulled barley also free-threshing wheat lentil pea chickpea grape
Late Bronze Age (LBA) 1900ndash1500 BC
More dispersed smaller sites (to 2 ha) in piedmont Period of Bactrian-Margiana Archaeological Complex abundant large sites in Merv oasis Key sites Namazga (VI) Gonur depe
Sophisticated canal irrigation in Merv oasis using water from main channels of rivers
Early Iron Age 1500ndash550 BC
Abundant settlements (to 15 ha) in piedmont and oases Key sites Tahirbaj Yaz-depe (Merv oasis)
Introduction of qanat(kiariz) irrigation to piedmont In Merv oasis settlement continues to shift to south
Broomcorn millet
Table 62 Simplified chronological chart of settlement in the Merv oasis in the historic period
Historical period and date (cal BCAD)
Settlement Irrigation systems Crops
Achaemenid 550ndash330 BC
Founding of Achaemenid city at Erk Kala c500 BC dispersed settlement centred on large buildings throughout the oasis and continuing to Seljuk period
Seleucid (Hellenistic) 330ndash140 BC
Construction of the city of Antiochia (Gyaur Kala) incorporating Erk Kala as citadel
Marked reduction in rural settlement
Parthian New settlements in north Expansion of Main crops
Irrigation agriculture in Central Asia 113
140 BCndashAD 220 of the oasis fortifications on perimeter such as Gobekli
cultivated area possible construction of central Murghab dam and extensive canal network
cotton hulled barley free-threshing wheat also lentil grape almond
Sasanian AD 220ndash651
Peak of settlement in the Merv oasis much continuity with Parthian period Possible construction of Wall of Antiochus (usually dated to the Hellenistic period) which marks northern limit of most post-bronze age settlement
Cultivated area stable
Main crops cotton hulled barley free-threshing wheat also lentil pea melon grape almond peach broomcorn millet
Umayyad Abbasid 8thndash9th centuries Samanid Seljuk and Post Seljuk 11thndash13th centuries Mongol conquest 1221
Continuity in settlement after Arab conquest of AD 651 Merv is capital of Seljuk empire in 11th and 12th centuries Sultan Kala established 9th century fortified 12th century Devastating conquest of city by Mongols but archaeological evidence suggests some post-conquest occupation
Continuity in area cultivated until Mongol conquest results in destruction of dam system Abundant textual evidence for function of irrigation system in 10thndash13th centuries
Historical period and date (cal BCAD)
Settlement Irrigation systems Crops
Timurid 14thndash15th centuries Safavid 1502ndash1736
New city built 1409 at Abdullah Khan Kala but decline continues
Central dam rebuilt in Timurid period destroyed in war of 1727
Turkmen 18thndash19th centuries
Dispersed settlement with semi-independent landlords
Irrigation system functioning but small-scale cultivation
Russian conquest 1890 Soviet Union 1919ndash1991 Republic of Turkmenistan 1991ndash
Establishment of modern settlement at Mary planned villages and communal farms
Introduction of large-scale irrigation systems for cotton Karakum canal
American cotton species
The archaeology of drylands 114
we would hesitate to identify simple cause and effect it would appear that increasedpopulation was linked through a complex sequence of interactions with the expansion ofirrigated agriculture and increased centralization of authority The increase in populationnot only required an increase in the amount of land irrigated but also provided theworkforce necessary for this expansion to take place Irrigation flourished during periodsof political stability often when a single polity ruled over the area and declined inperiods of invasion or unstable internal political conditions
The decline of Merv can clearly be traced to the Mongol destruction of AD 1221ndash2 The Mongols took advantage of the fact that Merv like most other settlementsthroughout Central Asia was reliant on a single water source In their rapid conquest ofthe region the Mongols frequently forced communities to capitulate by disrupting watersupplies and damaging irrigation structures and all they needed to do at Merv was todestroy the main dam that controlled water in the oasis Whilst the city was in partrebuilt the irrigation systems were never fully reconstructed until the region once againcame under the influence of another empiremdashthat of the Soviets
Significantly the widespread environmental degradation that plagues Soviet-built irrigation systems in the region does not appear to have been a major problem in the pastsuggesting that sustainable irrigation in Turkmenistan is not only feasible but has beenthe norm Traditional irrigation systems were generally localized and often dependent ona single water supply that was not only limited but also liable to fluctuate considerablyfrom year to year Water management required considerable skill hence the mirab bashiresponsible for highly important and often contentious decisions on water allocation anddistribution was one of the most senior officials in central governmentmdashindeed the success of many political officials often hinged on their skill at managing local waterresources Yet whilst water was managed centrally all water users were responsible forthe upkeep of the system with those gaining more being expected to contribute moreThe fact that individuals could benefit as a result of their efforts gave all users a vestedinterest in ensuring that the irrigation network was maintained and that water was usedefficiently The Soviet system effectively broke this link with the system managedcentrally but from afar Together with the collectivization of land the imposition ofcentral planning meant that benefits were no longer linked to duty water users had no sayin how the system was managed nor were they responsible for its maintenance Theestablishment of myriad agencies to oversee different parts of the network resulted inunnecessary bureaucracy and waste In sum traditional irrigation and water distributionsystems tended to be small highly productive well managed extremely efficient andsustainable over the long term In contrast Soviet-built systems are huge inefficient inflexible poorly managed and for the most part unsustainable
The decline in the water distribution and irrigation network since the break-up of the Soviet Union is thus unsurprising What remains to be seen however is how this declinewill be managed and what can be done to ensure the future sustainability ofTurkmenistanrsquos (and indeed Central Asiarsquos) water distribution and irrigation networkThe Central Asian Republics have inherited a Soviet-built system and must learn to work
present throughout oasis completed 1967
Irrigation agriculture in Central Asia 115
with the system and resources that are available While we cannot revert to the pastCentral Asiarsquos water managers would do well to look to the past for some of the answersto their current and future problems
REFERENCES
Adams RM (1981) Heartland of Cities Surveys of Ancient Settlement and Land Use onthe Central Floodplain of the Euphrates Chicago University of Chicago Press
Babaev AG (1994) Landscapes of Turkmenistan In VFet and KIAtamuradov (eds)Biogeography and Ecology of Turkmenistan 5ndash22 Dordrecht Kluwer
Babaev AG (1996) Problems of Aridland Development Moscow Moscow University Press
Bader A Gaibov V and Koshelenko G (199394) The northern periphery of the Mervoasismdashfrom the Achaemenid period to the Mongol conquest Silk Road Art and Archaeology 351ndash70
Bader AN Gaibov VA and Koselenko GA (1995) Walls of Margiana In AInvernizzi (ed) In the Land of the Gryphons Papers on Central Asian Archaeology in Antiquity 39ndash50 Florence Le Lettere
Bader A Gaibov V Gubaev A and Koshelenko G (1996) The oasis of Merv thedynamics of its settling and irrigation Ancient Civilizations from Scythia to Siberia349ndash60
Bartold VV (1914) K istorii Orosheniya Turkestana St Petersburg Selrsquoskago Vestnika (In Russian)
Bartold VV (1928) Turkestan down to the Mongol Invasion London Luzac second edition
Bedford DP (1996) International water management in the Aral Sea basin Water International 2163ndash9
Black C Dupree L Endicott-West E Naby E Matuszewski DC and WaldronAN (1991) The Modernization of Inner Asia Armonk NY MESharpe
Boardman S (1997) Plant use in the Merv oasis Iran 3529ndash31 Dukhovny V (1995) Civilisation and Water Resources Management in Central Asia
Tashkent World Bank Gubaev A Koshelenko G and Tosi M (1998) (eds) The Archaeological Map of the
Murghab Delta Preliminary Reports 1990ndash95 Rome ISIAO Hannan T and OrsquoHara SL (1998) Managing Turkmenistanrsquos Kara Kum canal
problems and prospects Post-Soviet Geography and Economics 39225ndash35 Harris DR and Gosden C (1996) The beginnings of agriculture in western Central
Asia In DRHarris (ed) The Origins and Spread of Agriculture and Pastoralism in Eurasia 370ndash89 London UCL Press
Harris DR Masson VM Berezin YE Charles MP Gosden C Hillman GCKasparov AK Korobkova GF Kurbansakhatov K Legge AJ and Limbrey S(1993) Investigating early agriculture in Central Asia new research at JeitunTurkmenistan Antiquity 67324ndash38
Harris DR Gosden C and Charles MP (1996) Jeitun recent excavations at an early
The archaeology of drylands 116
Neolithic site in southern Turkmenistan Proceedings of the Prehistoric Society 62423ndash42
Herrmann G (1997) Early and medieval Merv a tale of three cities Proceedings of the British Academy 941ndash43
Herrmann G and Petersen A (1997) The Ancient Cities of Merv TurkmenistanLondon University College London Institute of Archaeology International MervProject
Herrmann G Kurbansakhatov K and Simpson SJ (1998) The International MervProject Preliminary report on the sixth season (1997) Iran 3653ndash75
Hiebert FT (1994) Origins of the Bronze Age Oasis Civilization in Central AsiaCambridge MA Harvard University Peabody Museum of Archaeology andEthnology American School of Prehistoric Research Bulletin 42
Jacobsen T and Adams RM (1958) Salt and silt in ancient Mesopotamian agricultureScience 1281251ndash8
Kohl PL (1984) Central Asia Palaeolithic Beginnings to the Iron Age Paris Editions Recherche sur les Civilisations Synthegravese 14
Kornilov BA and Timoshinka VA (1975) The impact of the Kara Kum canal on theenvironment Soviet Geography 15308ndash14
Le Strange G (1905) The Lands of the Eastern Caliphate Cambridge Cambridge University Press
Lewis RA (1966) Early irrigation in West Turkestan Annals of the Association of American Geographers 56467ndash91
Lipovsky I (1995) The central Asian cotton epic Central Asian Survey 14529ndash42 Lisitsina GN (1969) The earliest irrigation in Turkmenia Antiquity 43279ndash88 Lisitsina GN (1981) The history of irrigation agriculture in southern Turkmenia In PL
Kohl (ed) The Bronze Age Civilization of Central Asia 350ndash8 Armonk NY MESharpe
Lisitsina GN and Prishchepenko LV (1976) The significance of paleoethnobotanicalremains for the reconstruction of early farming in the arid regions of the USSR Folia Quaternaria 4783ndash8
Maas EV (1987) Salt tolerance of plants In BRChristie (ed) CRC Handbook of Plant Science in Agriculture Volume II 57ndash71 Boca Raton FL CRC Press
Maier U (1996) Morphological studies of free-threshing wheat ears from a Neolithic sitein southwest Germany and the history of the naked wheats Vegetation History and Archaeobotany 539ndash55
Masson VM (1957) Jeitun and Kara-tepe Sovetskaya Arkheologiya 1143ndash60 (In Russian)
Masson VM and Sarianidi VI (1972) Central Asia Turkmenia before the Achaemenids London Thames amp Hudson
Micklin PP (1991) The Water Management Crisis in Soviet Central Asia Pittsburgh University of Pittsburgh Center for European and Russian Studies Carl Beck Papers inRussian and East European Studies
Miller NF (1993) Preliminary archaeobotanical results from the 1989 excavation at thecentral Asian site of Gonur Depe Turkmenistan International Association for the Study of the Cultures of Central Asia Information Bulletin 19149ndash63
Irrigation agriculture in Central Asia 117
Miller NF (1999) Agricultural development in western Central Asia in the Chalcolithicand Bronze Ages Vegetation History and Archaeobotany 813ndash19
Moore KM Miller NF Hiebert FT and Meadow RH (1994) Agriculture andherding in the early oasis settlements of the Oxus Civilization Antiquity 68 418ndash27
Nesbitt M (1994) Archaeobotanical research in the Merv Oasis Iran 3271ndash3 OrsquoHara SL (1997) Irrigation and land degradation implications for agriculture in
Turkmenistan central Asia Journal of Arid Environments 37165ndash79 OrsquoHara SL (in press) Central Asiarsquos water resources contemporary and future
management issues International Journal of Water Resources Development Orlovsky NS (1994) Climate of Turkmenistan In FFet and KIAtamuradov (eds)
Biogeography and Ecology of Turkmenistan 23ndash48 Dordrecht Kluwer Academic Press Monographicae Biologicae 72
Pierce RA (1960) Russian Central Asia 1867ndash1917 A Study in Colonial RuleBerkeley University of California Press
Powell MA (1985) Salt seed and yields in Sumerian agriculture A critique of thetheory of progressive salinization Zeitschrift fuumlr Assyriologie und VorderasiatischeArchaumlologie 757ndash38
Serjeant RB (1972) Islamic Textiles Material for a History up to the Mongol ConquestBeirut Librairie du Liban
Smith DR (1992) Salinization in Uzbekistan Post-Soviet Geography 3321ndash33 Wenke RJ (1975ndash76) Imperial investments and agricultural developments in Parthian
and Sasanian Khuzestan 150 BC to AD 640 Mesopotamia 10ndash1131ndash221 Zaharchenko BT (1990) Voda v Turmenskoy Zhizni (Water in Turkmen Life)
Ashgabat (In Russian) Zaharchenko BT (1994) A Brief History of the Construction of the Niyazov Kara Kum
Canal Ashgabad (In Russian)
The archaeology of drylands 118
Part III SAHARA AND SAHEL
7 Conquests and land degradation in the eastern
Maghreb during classical antiquity and the Middle Ages
JEAN-LOUIS BALLAIS
INTRODUCTION
Ibn Khaldounrsquos description of the eleventh-century Arab invaders of the Maghreb (Ibn Khaldoun 1968) in which he likened them to plagues of locusts is well known Sincethen many historians have tended to believe that the Arab invaders were responsible forthe land degradation that is so visible today in many parts of north Africa More thantwenty years ago after French decolonization controversy was particularly strong(Berque 1970 1972 Cahen 1968 Idris 1968a 1968b Poncet 1967 1968) thoughinterest has since decreased Since this period however advances in geoarcheologicalresearch allow a reassessment of the role of Arab nomads in land degradation in northAfrica the basis for which was more ideological than factual The purpose of this chapteris to discuss the geoarchaeological record of the eastern Maghreb (Fig 71) and to compare it with the historical record of conquests invasions and occupations in order toassess the respective roles of climate and people in shaping this regionrsquos landscape in classical antiquity and the Middle Ages
THE ARRIVAL OF THE PHOENICIANS
The Phoenicians settled on the eastern coasts of present-day Tunisia between the eleventh and ninth centuries BC (Decret 1977) This period coincides with the beginning of theLate Holocene a phase characterized throughout the region by a trend to aridificationfollowing the Middle Holocene climatic optimum (Ballais 1991a) This aridification wasthe reason for the reappearance of aeolian deflation on the great lsquochottsrsquo or lsquosebkhasrsquo (salt flats) of southern Tunisia as well as the incision of the lower prehistoric holoceneterrace especially in the present-day semi-arid subzone (Ballais and Benazzouz 1994Table 71) The phase is contemporaneous with the erosional crisis at the transitionbetween the Bronze and Iron Ages in the northern Mediterranean
Figure 71 The eastern Maghreb showing locations mentioned in Chapter 7
Key 1 Ksar Rhilane 2 chott Rharsa 3 Ksar Rheriss 4 Wadi Keacutebir-Miliane 5 Henchir Rayada 6 Wadi Cheacuteria-Mezeraa 7 Wadi el Akarit 8 Haiumldra Stippling denotes major areas of sand dunes horizontal dashed lines denote salt flats (lsquosebkhasrsquo or lsquochottsrsquo)
Table 71 Morphoclimatic evolution in the eastern Maghreb during the later Holocene dates in radiocarbon years BP
Isotopic chronology Morphogenesis Shifting sands incision610 +minus 110 Terrace aggradation Incision 1350 +minus 70 Terrace aggradation 1470 +minus 190 Terrace aggradation 1730 +minus 185 Pedogenesis 2380 +minus 155 Flood deposits 2420 +minus 70 Gyttja deposits 2590 +minus 90 Pedogenesis Deflation
The archaeology of drylands 122
(Jorda et al 1993) The characteristics of this sediment morphogenesis in the eastern Maghreb as well as those of the associated pollen (Brun 1989) indicate the occurrenceof strong floods due to intense rainfalls the latter probably highly concentrated andepisodic Thus the climate was probably characterized by greater seasonality than beforeand a hydric balance less favourable than today As far as we can tell Phoeniciancolonization in the eastern Maghrebmdashfor long limited to a few coastal sitesmdashdoes not appear to have been responsible for this morphogenic crisis
THE ROMAN CONQUEST AND COLONIZATION
From the Punic Wars up to the Vandal conquest during almost six centuries threemorphogenic phases can be discerned in the region the end of an aggradation period anincision period and a second alluvial aggradation period (Table 71)
The end of an aggradation period
Between c2400 and 2200 BP aggradation became widespread once more The evidenceincludes washed sand deposits with Helicella molluscs in dunes on the eastern margin of the Grand Erg Oriental sand sea at Ksar Rhilane in the Saharan subzone (Fig 71 site 1 Fig 72) gyttja accumulation in the Rharsa chott in the arid subzone (Ballais 1992 Fig 71 site 2) and fine alluvium along the watercourses flowing down from the Nemencha Mountains in the semi-arid subzone (Ballais and Benazzouz 1994)
Today at Ksar Rhilane sands are blown by the wind and rillwash and sheetwash never occur In the Rharsa chott the principal deposit is sodium chloride In the NemenchaMountains the fine alluvium has been organized in continuous beds by slow streams in alarge channel The mean annual accumulation rate was 14ndash22 mm which is very close to the rate calculated for the low prehistoric Holocene terrace (Ballais 1991b) The fewpollen grains taken from those deposits show that a woodland could have colonized theslopes All these characteristics are compatible with a more positive hydric balance in thisphase than either during the previous climatic phase or today especially in the arid andSaharan subzones In comparison with the climatic optimum period (Ballais 1991a) thedegradation of climate in the semi-arid subzone is shown by a probable increase insummer evaporation though winters remained sufficiently cool to permit the growth ofCedrus on the summits of Nemencha 1800 m above sea level
During this period Roman penetration of the interior seems to have been very limitedno evidence for Roman agricultural works has yet been observed
Incision 3680 +minus 160 Terrace aggradation 4220 +minus 50 Terrace aggradation
Conquests and land degradation in the eastern Maghreb 123
An incision period
No deposit has been identified for the period from c2200 BP to c1650 BP (the second and third centuries AD) but there is widespread evidence for
Figure 72 Flood deposits at Ksar Rhilane (Tunisia) bedded silts and sands (dated to 2380 +minus 155 BP) with ripple-marks in the upper part with the present-day dunes behind
Photograph J-LBallais
watercourse downcutting This period was thus characterized by streams having thecapability to incise their channels and to transport their alluvial sediments up to the baselevels It is possible precisely to measure neither the scale of the incision nor its annualrate owing to the lack of chronological markers Nevertheless it should be rememberedthat the mean annual rate of downcutting of the low prehistoric Holocene terrace was 12mm between 40003500 BP and 17001600 BP (Ballais 1991b) Thus during this periodslopes furnished very few colluvial sediments indicating either that they were wellprotected by rather dense vegetation or that the agricultural techniques
did not produce soil erosion The beginning of this period was marked by the almostcontinuous expansion of Roman colonization both westwards and southwards (Feacutevrier 1989) Even though they cannot be dated precisely numerous dams were built at aboutthis time on the watercourses of the Nemencha and Auregraves mountains (Ballais 1976 Birebent 1964 Leveau 1974ndash75) and in southern Tunisia (Ballais 1990 Trousset 1974) Sometimes their construction can be shown to be associated with the constructionof the Roman frontier works (limes) at the boundary of the arid and Saharan subzones
The archaeology of drylands 124
(Baradez 1949 Trousset 1974) Presumably it was the combination of favourableconditions of climate vegetation and soils and well-organized agricultural activities that was responsible for such limited soil erosion through the second and third centuriesAD
An aggradation period
Evidence for an aggradation terrace dating to historic times is widespread along most ofthe watercourses from the north to the south and from the humid subzone to the upperSaharan subzone The exception is the lower Saharan subzone so far the historicaggradation terrace unlike the late prehistoric terrace has not been identified in the farsouth of Tunisia As with the late prehistoric terrace the later feature developed for themost part in those catchments formed in loose erodible rocks The sediments aregenerally fine in texture beige in colour in the south and greyer in the north This terracealso has an extensive surface area particularly in the north and forms a sediment unitthat varies in thickness from 1 m (along small watercourses) to 5ndash6 m (along major rivers)
On the largest rivers detailed studies of sedimentation patterns show variations withlatitude Thus in the Wadi Leben terrace at Ksar Rheriss towards 35degN in the arid subzone (Ballais 1991b Fig 71 site 3) desiccation cracks appear in thin beds formedby washed clay and have been filled by sand during later flooding These patterns whichare characteristic of intermittently-flowing wadis disappear in the more humid subzoneswith the exception of the Keacutebir-Miliane wadi (Fig 71 site 4) which is today a perennial stream Conversely further to the north the terraced sediments of many watercoursescontain dark hydromorphic silt-like facies rich in organic matter or in manganese oxideIn eighteen different cases it is possible to correlate the presence of these facies with theperennial nature of the watercourse or inversely their absence with the intermittentnature of the watercourse In three other cases this correlation is not apparent Thesediments nearly always contain fragments of Roman or even earlier pottery (more thantwenty recorded examples have been noted) and often charcoal hearths or other artefactsThey cover the base of Roman bridges or aqueduct piles (Fig 73) or fill in low-volume dams built during the second and third centuries AD In another case along the Wadi esSgniffa a whole ancient settlement is covered by alluvial sediments There are still veryfew isotopic dates for the very low main historic Holocene terrace but two examples ofdated sediments are in the Wadi Cheacuteria-Mezeraa (Fig 71 site 6) which was radiocarbon-dated using samples of land molluscs to 1370 +minus 70 years BP (Fig 74) and in the Wadi el Akarit (Fig 71 site 7) where the terrace is dated to 1470 +minus 190 years BP (Ballais 1995 Fig 75) The rate at which the low terrace sediments of the historicperiod were accumulating became considerable at the eighteen locations examined theaverage reached 74 mm per year which is five times the rate of accumulation of thelower (prehistoric) Holocene terrace suggesting that the geosystems in which the historicterrace formed differed significantly from those when the earlier terrace accumulated
It is of interest to note that this historic aggradation started only when the area in question with the exception of the Sahara was occupied by sedentary populations as aresult of a lengthy process of political and economic
Conquests and land degradation in the eastern Maghreb 125
Figure 73 The modern aqueduct crossing Wadi Bou Jbib Carthage built on the piles of the Hadrianic aqueduct
Note The right-hand pile is partly covered by grey silts belonging to the historic-period Holocene terrace Photograph J-LBallais
evolution Generally speaking the northeast of Tunisia near Carthage was cultivatedfrom perhaps the fifth century BC onwards and the cereal plains in the northwest ofTunisia and some parts of eastern Algeria from the third and second centuries BC Thesteppes of Algeria Tunisia and Libya were cultivated from the first and second centuriesAD and the borders of the Sahara during the second century AD at the time of theconstruction of the Severan limes (Trousset 1986) If the observed variation in alluvial deposits was linked solely to the political and economic development of this vast area wemight expect to find the imbalance of the geosystems resulting from this developmentoccurring at the same time in all places with the threshold producing the change fromincision to aggradation taking place at the same time in the north and south after eightcenturies of sedentary settlement in the former region and after a few decades of suchsettlement in the latter However this coincidence though not impossible seems highlyunlikely particularly if the tremendous differences in mean annual rainfall between thenorth and the south are taken into account In addition it has now been shown thatcomparable terrace deposits were deposited throughout the Mediterranean particularlytowards the end of classical antiquity and in the early Middle Ages (Ballais and Crambes1992 Bruumlckner 1986 Vita-Finzi 1969) The extensive occurrence of this feature can best be explained by an
The archaeology of drylands 126
Figure 74 Holocene terrace of Wadi Cheacuteria-Mezeraa (Algeria) Note The stratigraphy visible to the left of the standing figure is from bottom to top grey silts bedded pebbles (Early Holocene) beige-grey silts (dated to 1370 +minus 70 BP) and a Roman drainage ditch filled with pebbles Photograph J-LBallais
underlying climatic shift across the region In view of the characteristics of the alluvialdeposits it seems likely that they originated from erosion of soils particularly those thatdeveloped during the Holocene Climatic Optimum and at the beginning of the Romanperiod
However land use systems in classical antiquity probably exacerbated these erosionaltrends Presumably the spread of cultivation and ploughing destroyed a large part of thevegetation on the watersheds reducing the cohesion of soils and superficial formations Itthen required only a small change in rainfall characteristics perhaps in the annual totalor at least in intensity and periodicity to produce considerable soil erosion and the startif not the return of water in the stream channels and increased discharge though thisincrease was not enough to carry the large sediment load from the slopes to the baselevels
The absence of the very low historic terrace in the lower Saharan subzone is probably due to the lack of agriculture on the watersheds For these very recent periods it isdifficult to compare the climatic situation with that of Sahara However the fact thatAcacia and Tamarix could be found in central Serir Tibesti at around 1700 BP and 1400BP indicates that at least in tropical Sahara the mean annual rainfall was more than thepresent day rainfall of 5 mm (Pachur 1974)
Conquests and land degradation in the eastern Maghreb 127
Figure 75 Holocene terraces in the Wadi el Akarit (Tunisia) A very low terrace of post-Islamic date (610 +minus 110 BP) is fitted into a gypseous terrace of late prehistoric date
Photograph J-LBallais
THE VANDAL CONQUEST AND OCCUPATION
The Vandals are not one of the beloved peoples of eastern Maghreb history Too often itseems that their brief passage in the fifth and sixth centuries AD has no longer left traces(Courtois 1955) In fact despite new studies (Modeacuteran 1988) the Vandal period remains badly known for two main reasons The first one was the lack of interest ofFrench archeologists of the colonial period in the post-Roman civilizations (Feacutevrier 1989) The second one is a consequence of the first one the destruction of the Vandalsites established on top of Roman towns As a result neither the limits of Vandal territory(in the present day Constantinois for example) nor their modes of soil occupation andland use are well known There is insufficient evidence on which to base any detaileddiscussion of climate soils vegetation and peoplersquos possible effects on the landscape at this time We can note however that alluvium continued to accumulate as the very lowhistoric terrace (Table 71)
BYZANTINE CONQUEST AND COLONIZATION
The Byzantines have been less neglected by French archeologists and historians of thecolonial period because they presented themselves as the legitimate heirs of the Roman
The archaeology of drylands 128
emperors (Ducellier 1988) They occupied the eastern Maghreb from the sixth to theseventh centuries AD As for the Vandal phase there is great uncertainty regardingmodes of rural settlement and land use Most information is available for the wars againstBerbers (Modeacuteran 1989) In particular most of the numerous small forts still visibletoday were attributed to Byzantine colonization but it now appears that some are laterthan the Arab conquest (Mahjoubi 1978) According to isotopic datings the end of theaggradation of the very low historic alluvial terrace coincides more or less with theByzantine period This can be confirmed at Haiumldra (Ammaedara Fig 71 site 8) where the foundations of a Byzantine bridge constructed at the same time as the sixth-century fort (Baratte and Duval 1974) were dug into the alluvial deposit (Ballais 1991c)
ARAB CONQUEST AND COLONIZATION
Seventy years were necessary for successive waves of Arab forces to conquer the easternMaghreb during the second part of the seventh century AD (Marccedilais 1946) presaging the high Islamic period which was a time of general economic prosperity (Vanacker1973) The last wave of Arab invaders the well-known Hilalian arrived in the middle ofthe eleventh century They are described as nomadic shepherds coming from UpperEgypt the lsquoplague of locustsrsquo in Ibn Khaldounrsquos memorable phrase who lsquopushed their flocks into the middle of the fields devastated the gardens stripped and ill-treated the country persons plundered the hamletsrsquo (Marccedilais 1946) In theory the consequences ofsuch devastation would have been so disastrous that they would have provokedcatastrophic and long-lived decline in the Tunisian economy (Al-Idrisi 1983 Ibn Khaldoun 1968 Marccedilais 1946 Vanacker 1973)
Following the aggradation of the very low historic terrace the general trend for watercourses in the study area was vertical incision (Table 71) with two to three interruptions The main interruption can be seen in the very low post-Islamic terrace which as far as we know today is little represented presumably because of its smallsize With only one exception (Wadi Keacutebir-Miliane) this terrace covers very small areas in particular in convex meander lobes and its height above the major bed rarely exceeds2 m Occasionally it appears as a rocky terrace that was breached in the previous build-up elsewhere the facies can sometimes be compared to that of the previous terracethough it is sometimes considerably coarser at least at the base The age of the terrace isstill rather uncertain because appropriate means of dating are not available but atHenchir Rayada (Fig 71 site 5) it contains Islamic pottery from the tentheleventhcenturies and in Wadi el Akarit (Fig 71 site 7) it was dated by radiocarbon using collagen to 610 +minus 110 years BP (Fig 75) This terrace is thus much younger than theperiod of the presumed Hilalian invasions
As for the previous terrace the widespread presence of a terrace of the same age can beseen throughout the Mediterranean basin (Ballais and Crambes 1992 Vita-Finzi 1969) Moreover in contrast to the final years of classical antiquity (the third to the fifthcenturies AD) population was probably low at the beginning of the Hafside period inTunisia during the twelfth century AD This period marked the end of the medievalclimatic optimum and the beginning of the Little Ice Age in Europe (Lamb 1977 Le
Conquests and land degradation in the eastern Maghreb 129
Roy-Ladurie 1967) In Morocco the existence of a Little Ice Age is controversial (El Bouch and Ballais 1997 Lamb et al 1989 Stockton et al 1985) Once again it seems most likely that such a widespread terrace resulted from climatic causes but furtherstudies will be required to clarify this point
This conclusion emphasizes the extreme ideological character of the theories regarding the impact of nomadic shepherds on the landscape at the time of the Arab invasions Evenif those shepherds did in fact cut down trees to any extent their main effect would havebeen to substitute pastures for cultivated fields The consequences would have been asfollows an increase in the rate of vegetation cover over the soil a consequent decrease inthe direct splash impact of rain drops on the soil and thus of pluvial erosion and sheeterosion and an increase in the lsquoroughnessrsquo of the terrain and thus a diminution in winderosion Within such a model it is necessary to moderate the intensity of such processesaccording to the climatic subzones and different grazing intensities but it seems realisticto suppose that in general such a move from arable to pastoral land use is likely toproduce less rather than more soil erosion
CONCLUSION
In the eastern Maghreb during classical antiquity and the Middle Ages fluctuations ingeosystems and in particular increases in soil erosion can be seen to have reflectedspecific combinations of climatic change and human activities A climatic fluctuation thatincreases the intensity of rains or the annual amount of precipitation affects slopes onlyif they have been made vulnerable by vegetation degradation or by cultivation systemsthat have not been designed to counteract erosion In other words phases of massiveagricultural colonization and phases of extension of the cultivated surface are veryfavourable to such erosion This was the case in the study area as in many parts of theMediterranean during the Roman period On the other hand periods of conquestgenerally seem to have been characterized by a contraction of the cultivated surface and aprogressive development of lsquonaturalrsquo vegetation or of pastures that limited soil erosion This may have been the situation in the case of the nomadic Hilalian shepherds of theArab conquest However there were exceptions to these trends in particular in theirrigated zones and in the terraced mountains
REFERENCES
Ballais J-L (1976) Morphogenegravese holocene dans la region de Cheacuteria (Nementchas-Algeacuterie) Actes du Symposium sur les Versants en Pays Meacutediterraneacuteens 127ndash30 Aix-en-Provence CEGERM 5
Ballais J-L (1990) Terrasses de culture et jessours du Maghreb oriental Meacutediterraneacutee3451ndash3
Ballais J-L (1991a) Evolution holocene de la Tunisie preacutesaharienne et saharienne Meacutediterraneacutee 431ndash8
Ballais J-L (1991b) Vitesse drsquoaccumulation et drsquoentaille des terrasses alluviales
The archaeology of drylands 130
holocegravenes et historiques au Maghreb oriental Physio-Geacuteo 22ndash2389ndash94 Ballais J-L (1991c) Les terrasses historiques de Tunisie Zeitschrift fur Geomorphologie
Suppl Bd 83221ndash6 Ballais J-L (1992) Le climat au Maghreb oriental apports de la geacuteomorphologie et de
la geacuteochimie Les Nouvelles de lrsquoArcheacuteologie 5027ndash31 Ballais J-L (1995) Alluvial Holocene terraces in eastern Maghreb climate and
anthropogenic controls In JLewin MGMacklin and JCWoodward (eds)Mediterranean Quaternary River Environments 183ndash94 Rotterdam Balkema
Ballais J-L and Benazzouz MT (1994) Donneacutees nouvelles sur la morphogenegravese et les paleacuteoenvironnements tardiglaciaires et holocegravenes dans la valleacutee de lrsquooued Cheacuteria-Mezeraa (Nemencha Algeacuterie orientale) Meacutediterraneacutee 3459ndash71
Ballais J-L and Crambes A (1992) Morphogenegravese holocene geacuteosystegravemes et anthropisation sur la montagne Sainte-Victoire Meacutediterraneacutee 1229ndash41
Baradez J (1949) Vue Aeacuterienne de lrsquoOrganisation Romaine dans le Sud Algeacuterien Fossatum Africae Paris Arts et Meacutetiers Graphiques
Baratte F and Duval F (1974) Les Ruines drsquoAmmaedara-Haiumldra Tunis Socieacuteteacute Tunisienne de Diffusion
Berque J (1970) Les Hilaliens repentis ou lrsquoAlgeacuterie rurale au XVIe s drsquoapregraves un manuscrit jurisprudentiel Annales Economic Socieacuteteacute Civilisation 51325ndash53
Berque J (1972) Du nouveau sur les Banucirc Hilacircl Studia Islamica 3699ndash113 Birebent J (1964) Aquae Romanae Recherches drsquoHydraulique Romaine dans lrsquoEst
Algeacuterien Alger Baconnier fregraveres Bruumlckner H (1986) Manrsquos impact on the evolution of the physical environment in the
Mediterranean region in historical times GeoJournal 13 (1)7ndash17 Brun A (1989) Microflores et paleacuteoveacutegeacutetations en Afrique du Nord depuis 30 000 ans
Bulletin de la Socieacuteteacute Geacuteologique de France 8(1)25ndash33 Cahen C (1968) Quelques mots sur les Hilaliens et le nomadisme Journal of Economic
and Social History of the Orient 11 (1)130ndash2 Courtois C (1955) Les Vandales et lrsquoAfrique Paris Arts et Meacutetiers Graphiques Decret F (1977) Carthage ou lrsquoEmpire de la Mer Paris Editions du Seuil Diehl C (1896) LrsquoAfrique Byzantine Histoire de la Domination Byzantine en Afrique
(533ndash709) Paris Imprimerie Nationale Dore JN and van der Veen M (1986) ULVS XV radiocarbon dates from the Libyan
Valleys Survey Libyan Studies 1765ndash8 Ducellier A (1988) Les Byzantins Histoire et Culture Paris Editions du Seuil El Bouch A and Ballais J-L (1997) Travertinisation deacutetritisme et anthropisation a Fegraves
(Maroc) Wuumlrzburger Geographische Arbeiten 92213ndash24 Fegravevrier P-A (1989) Approches du Maghreb Romain Aix-en-Provence Edisud Ibn Khaldoun A (1968) Muqqadima Beirut Commission Libanaise pour la Traduction
des Chefs drsquoOeuvre translated by VMonteil Idris HR (1968a) Lrsquoinvasion hilalienne et ses consequences Cahiers de Civilisation
Meacutedieacutevale 3353ndash71 Idris HR (1968b) De la reacutealiteacute de la catastrophe hilalienne Annales Economic Socieacuteteacute
Civilisation 23390ndash6 Al-Idrisi A (1983) Le Maghrib au 6e siegravecle de lrsquoHeacutegire Paris Publisud translated by
Conquests and land degradation in the eastern Maghreb 131
Hadj Sadok Jorda M Parron C Provansal M and Roux M (1993) Erosion et deacutetritisme holocene
en Basse Provence calcaire Lrsquoimpact de lrsquoanthropisation Travaux du Centre Camille Jullian 14225ndash33
Lamb HF Eicher U and Switsur VR (1989) An 18000-year record of vegetation lake-level and climatic change from Tigalmamine Middle Atlas Morocco Journal of Biogeography 1665ndash74
Lamb HH (1977) Climate Present Past and Future London Methuen Le Roy-Ladurie E (1967) Histoire du Climat depuis lrsquoAn Mil Paris Flammarion Leveau P (1974ndash75) Une valleacutee agricole des Neacutemenchas dans lrsquoAntiquiteacute romaine
lrsquooued Hallail entre Djeurf et Aiumln Mdila Bulletin Archeacuteologique du Comiteacute des Travaux Historiques et Scientifiques 10ndash11b103ndash21
Mahjoubi A (1978) Recherches drsquoHistoire et drsquoArcheacuteologie agrave Henchir El-Faouar (Tunisie) Tunis Publications de lrsquoUniversiteacute de Tunis
Marccedilais G (1946) La Berbeacuterie Musulmane et lrsquoOrient au Moyen Age Paris Aubier Modeacuteran Y (1988) Les premiers raids des tribus sahariennes en Afrique et la Johannide
de Corippus Histoire et Archeacuteologie de lrsquoAfrique du Nord 2479ndash90 Modeacuteran Y (1989) La deacutecouverte des Maures Reacuteflexions sur la lsquoreconquecirctersquo byzantine
de lrsquoAfrique en 533 Cahiers de Tunisie 43155ndash6 211ndash38 Pachur HJ (1974) Geomorphologische Untersuchungen im Raum des Serir Tibesti
(Zentrasahara) Berliner Geographische Abhandlungen 176ndash58 Poncet J (1967) Le mythe de la lsquocatastrophersquo hilalienne Annales Economie Socieacuteteacute
Civilisation 23660ndash2 Poncet J (1968) Encore agrave propos des Hilacircliens la mise au point de RIdris Annales
Economie Socieacuteteacute Civilisation 23660ndash2 Stockton CW et al (1985) Long-Term Reconstruction of Drought in Morocco Tucson
University of Arizona Press Trousset P (1974) Recherches sur le Limes Tripolitanus du Chott el Djerid agrave la
Frontiegravere Tuniso-Libyenne Paris CNRS Trousset P (1986) Limes et frontiegravere climatique Histoire et Archeacuteologie de lrsquoAfrique du
Nord 55ndash84 Paris CTHS Vanacker C (1973) Geographic eacuteconomique de lrsquoAfrique du Nord selon les auteurs
arabes du IXe au milieu du XIIe siegravecle Annales Economie Socieacuteteacute Civilisation 3 659ndash80
Vita-Finzi C (1969) The Mediterranean Valleys Geological Changes in HistoricalTimes Cambridge Cambridge University Press
The archaeology of drylands 132
8 Success longevity and failure of arid-land
agriculture RomanomdashLibyan floodwater farming in the Tripolitanian pre-desert
DAVID GILBERTSON CHRIS HUNT AND GAVIN GILLMORE
INTRODUCTION
This chapter examines the successes and failures of RomanomdashLibyan and later floodwater farmers in the Tripolitanian pre-desert in Libya (Fig 81) This vast region of rocky plateaux and incised wadis lies between the higher better watered Gebel Nafusaand the Mediterranean coastlands to the north and the desert of the Hamada al-Hamra to the south
Critical to floodwater farming in this region were complex networks of walls that wereused to manage occasional storm-water to sustain agricultural settlement with manyimpacts on soils geomorphology and biogeography (Fig 82) The vast scale of the ancient settlement stands in stark contrast to the depopulated modern landscape As longago as 1857 the explorer Heinrich Barth recorded that the landscape displayed a lsquosea-like level of desolationrsquo (1857125) Today the region remains empty and inhospitableexcept for a few pastoralists with mixed herds of sheep and goats The pastoralists exploitbore water rare springs and small wells In the hotter and drier parts of the year herdsmay be taken north to the better watered and cooler Gebel The modern towns of BeniUlid and Mizda are the only significant settlements in the region The area around BeniUlid is a dense mixture of modern development and remains of ancient buildingsevidencing substantial occupation from RomanomdashLibyan times to as recently as only 400 years ago (Gilbertson and Hunt 1988) The Wadi Merdum through Beni Ulid is alsoone of the last if not the last wadis in this region where active floodwater farmingcontinues Date palms figs plums and ancient olive trees can still be found growingalong 4ndash5 km of the modern wadi floor (Goodchild and Ward-Perkins 1949) together with eucalyptus whilst the sheltered floors of small side-wadis sometimes yield a crop of barley
Figure 81 Tripolitania northwest Libya showing the principal landforms and settlements and the location of the UNESCO Libyan Valleys Survey
Key The 200 mm 100 mm and 20 mm isohyets are shown as dashed lines the contours are in metres
Figure 82 Romano-Libyan floodwater farming in the Wadi Gobbeen in the Tripolitanian pre-desert
Note A wadi-edge diversion wall is visible in the right foreground and a series of crosswadi walls down the wadi in the distance with the fortified farm (gasr) on the horizon on the right Photograph GBarker
The archaeology of drylands 134
ROMANO-LIBYAN AND LATER SETTLEMENT
The character of ancient settlement and land use in the Tripolitanian pre-desert was summarized in the two-volume report of the UNESCO Libyan Valleys Survey (ULVS)(Barker et al 1996a 1996b) The survey was a combined enterprise between the Department of Antiquities in Tripoli and a group of archaeologists and geographers fromthe Universities of Leicester Huddersfield Manchester and Sheffield in England Themain archaeological evidence is presented in a Gazetteer based upon 2437 site records(Barker et al 1996b) Many individual sites are themselves complex for example asingle entry deals with the complex networks of hundreds of substantial wadi walls thatwere mapped over 10 km of the Wadi Umm el-Kharab (Barker et al 1996a) The scale and significance of the past occupation of the predesert are clear from a cursoryexamination of the distribution of large open (that is unenclosed or undefended) farmsand farmsteads some built in the Opus Africanum style attributed to the first to the third centuries AD and the imposing enclosed and possibly fortified barn-like gsur (Fig 83) The
Success longevity and failure of arid-land agriculture 135
Figure 83 Simplified distribution of Opus Africanum and other early Romano-Libyan farms and farmsteads (above) and fortified farms (gsur) (below) in the UNESCO Libyan Valleys Survey study area
Source After Barker et al 1996162 165
The archaeology of drylands 136
latter are mainly attributed to the third to fourth centuries AD though many examplesalso date to the Islamic period and at some gsur significant activity continued until the sixteenth century AD (Gilbertson and Hunt 1988)
These substantial settlement remains appear to have been built by local neo-Punic-speaking Macae tribespeople here referred to as Romano-Libyan (Mattingly 1989) Initially transhumant pastoralists they took advantage of the extension of Romaninfluence into Tripolitania rapidly to develop a robust long-lasting mixed farming economy and a substantial increase in population in a desert environment broadly similarto that of today The ancient farming was more than self-sufficient producing surpluses of olive oil perhaps other tree crops grapes and cereals Stock keeping no doubtcontinued through the RomanomdashLibyan period and it is transhumant pastoralism thatcharacterizes the human geography of the modern landscape A substantial trade tookplace between the pre-desert the Mediterranean coast and beyond significant quantities of olive oil and perhaps cereal crops were sent north to the coastal cities whence somewere exported to the wider Roman world Products produced by better-watered regions were imported into the pre-desert including even lsquoexoticrsquo foodstuffs such as deep-water sea fish
The density of RomanomdashLibyan and later settlement in the pre-desert varied significantly through both space and time (Flower and Mattingly 1995 Mattingly withFlower 1996) A dramatic transformation of pre-desert settlement with a rise in population to about 20000 in perhaps 2000 farms occurred in the study region duringthe first century AD The longevity of this occupation is rarely securely known Theestimated 1000 gsur built from the third century AD were perhaps fortified farms Thesemust have provided massive and secure storage of crops and food as buffers againstsequences of adverse drought years By the sixth to seventh centuries AD the entirenetwork had probably become consolidated into major lsquoagricultural estatesrsquo controlled by powerful rich local elites or warlords (Mattingly 1996) These lsquoestatesrsquo are manifest in the changing density of both walls and gsur with no obvious linkages to topographic or hydrological features (Gilbertson and Chisholm 1996 Mattingly 1996) Curiously thisintensification and reordering of settlement in the pre-desert were associated with a decreasing import of goods from the Mediterranean countries and decreasing quantitiesof olive oil sent to the coast (Mattingly 1996) GIS-based analyses suggest both a general lsquothinningrsquo of settlement and a slight northward and westward shift from the early extensive phase through this late RomanomdashLibyan period followed by a notable shiftnorthwards and a trend towards clustering of settlement during the Islamic period(Mattingly with Flower 1996) Eventually the only remaining large settlement withsignificant modern floodwater farming was Beni Ulid on the northern edge of the pre-desert
The information that is now available to explore these ideas is vastly superior to that compiled before the ULVS survey It is nevertheless limited in scope and reliability andis often incapable of sustaining sophisticated theoretical enquiry For example thelogistical reality of access by off-road vehicle in the difficult terrain and generally arduous circumstances forced a concentration of surveys upon the more accessible wadisat the expense of other wadis and the vast plateaux between them This survey pattern islikely to have under-represented features such as ancient farms in basins or on ancient
Success longevity and failure of arid-land agriculture 137
route-ways on the plateaux as found for example near the Wadi Umm el-Kharab (Barker et al 1991) or detected through Landsat images and field survey (Dorsett et al1984) Detailed surveys on foot were carried out only in the Wadi Umm el-Kharab and in small sectors of the Wadis el-Amud Gobbeen Mansur and Mimoun (Barker et al1996a) probably locating only 70ndash80 per cent of the walls initially present (Barker withGilbertson 1996) Some wadi-floor walls were totally buried by sediment and detected only in gully exposures or as lines of bushes (Gilbertson et al 1994) Elsewhere our understanding is mostly based upon field sketches and simple field maps (Gilbertson et al 1984) as a consequence of the lack of air photographs and appropriate base maps forfield workers in this remote and politically sensitive area
COMPETING EXPLANATIONS
Reasons for the initiation growth stability and eventual decline of settlement andfarming in this pre-desert region remain remarkably unclear Some possible explanationsare set out below
First we consider the possible explanationsmdashsingly or in combinationmdashfor the abandonment of Romano-Libyan settlements and farms in the Tripolitanian pre-desert
Human processes
bull Political social and economic changes at the coast and in the wider Roman empire brought about the loss of the market for pre-desert produce
bull Political social and economic changes at the coast and in the wider Roman empire brought about the loss of the lsquoRomanrsquo technocrats who made the system work
bull The arduous life did not give sufficient rewards to the farmers bull It was too unrewarding and unexciting for young peoplemdashthe opportunities and
good life in the coastal cities were too attractive bull Insecuritymdashraids and menace from the desert tribes to the south could no longer be
managed bull The first and second Arab invasions prompted the abandonment of the farms bull The quelling of nomadism by the development of farmssettlements and the
provision of better water supplies in cisterns caused unacceptable albeit localized environmental degradation through over-grazing loss of pastures excessive soil loss and so on
bull The demands of the imperial economy and imperial attitudes undermined the ideology attitudes self-sufficiency and ultimately the vigour agricultural success and subsistence basis of the indigenous population at the floodwater farms
bull The settlers were pioneers not developers and were followed by parasitic professionals who failed to support the development process the professionals so drained the economic basis of the region that the economy failed
bull The region was drained by the activities of the equivalent of itinerant lsquocarpetbaggersrsquo who drained this region of its vitality and wealth
bull The question is based upon a misinterpretation the people were never fully
The archaeology of drylands 138
Natural processes
lsquoInduced environmental changersquo
Now we turn to the possible explanations for the expansion of these same settlements
sedentary and failed to return rather than leftmdashit was a threat to or loss of their mobility that was critical
bull A disease epidemic removed the capacity of people crops or livestock to continue in this demanding desert environment
bull The lsquolong-termrsquo climate became too arid bull There was one or more relatively short but pernicious droughts that lasted too long
for people unable to import food or water in adequate quantities to sustain themselves and their plants and animals through such adverse times
bull The inherent instability of the biophysical systems in dryland environments led to the growth of mutually reinforcing links between any of many possible cause-and-effect relationships which resulted in the non-reversible growth and persistence of an originally minor human or environmental disturbance and subsequent desertification or non-sustainable intensification of grazing
bull There was a local version of the lsquoCharney Effectrsquo there was an increased exposure of the soil and rock at ground surface as a result of more intensive and widespread livestock and arable farming in the pre-desert which eventually brought about a downward spiral resulting in progressive desiccation
bull Intensive and widespread livestock and arable farming raised such large quantities of dust into the atmosphere that a regional climatic change was induced resulting in greater aridity
bull Accelerated soil erosion made arable and pastoral farming too difficult on the plateaux
bull Goats and sheep lsquoravagedrsquo the pastures bull Excessive trapping of water in soil produced soil salinization in plateau-basins and
on wadi floors bull Excessive removal of vegetation led to the salinization of soil the reduction of
evapotranspiration caused a greater deposition of salts as a result of the induced (periodic) soil water-logging
bull Excessive cropping caused the loss of soil fertility and unexpected crop failure bull The supply of timber for fuel and other domestic purposes was effectively
exhausted bull The inherent instability of the biophysical systems in dryland environments led to
the growth of mutually reinforcing links between any of many possible cause-and-effect relationships which resulted in the non-reversible growth and persistence of an originally-minor environmental disturbance and subsequent desertification or non-sustainable intensification of grazing
Success longevity and failure of arid-land agriculture 139
Human processes
Environmental processes
Several explanations appear to be contradictory some are counter-intuitive whilst others probably exceed the strength of the present evidence Several possible explanationscannot be investigated with the methods currently available Indeed many explanationswould be difficult to explore even in arid lands that have been the subject of sustained
bull There was a demand from the coastal cities and wider Roman economy for olive oil grapes wheat barley dates figs pistachios and animal products which was met by indigenous people andor settlers
bull The southern lsquofrontierrsquo of the Roman empire was secured by a lsquodefence in depthrsquo made up of farming communities and some military installations
bull Driven by people who may have had (variously) particular types of ideology or belief misunderstanding adventure anticipation of quick or long-term profit a need to escape from the confines of contemporaneous life optimism a pioneer culture military imperatives a need for new lands and so on the lsquofrontierrsquo moved south bringing with it farmers and settlers
bull The arrival of Roman lsquoknow-howrsquo and lsquocan-dorsquo
bull The climate was lsquowetterrsquo (caused by cloud cover in greater quantity differently distributed more reliable andor more frequent) so agriculture prospered and settlement extended deeper into the pre-desert
bull The climate worsened and the settlers or indigenous people caught between the desert and hostile neighbours were obliged to develop intensive agriculture in the wadis
bull lsquoPioneerrsquo or lsquoeccentricrsquo people had experimented with small-scale cultivation (perhaps experimental in outlook) using ideas from indigenous people or elsewhere and started a lsquofashionrsquo that was thought worthy or otherwise good for personal development
bull The farms were started and maintained as a tax avoidance or tax mitigation measure bull The effect of the water-harvesting and the planting of tree and other crops as part of
floodwater farming was to so change the nature of the relationships between climate soil and vegetation that the pre-desert became transformed by many other occupants who created through their type of land use a biologically-productive as well as more wooded and economically-productive environment
bull Relatively minor small-scale developments associated with water-harvesting and plant production produced a series of biophysical feedbacks between the various components of the pre-desert environmental system These proceeded to reinforce each other eventually leading to the transformation of the pre-desert from one stable state characterized by relatively low biological productivity to a different stable state characterized by a much higher level of biological activity (and soil developments precipitation and so on) which appeared to encourage the extension andor the intensification of farming developments
The archaeology of drylands 140
intensive and extensive systematic studymdashwhich is far from the case here Through thefilter of nearly 2000 years it is difficult to separate cause and effect and lsquotriggersrsquo from pre-disposing or maintaining factors as well as to disentangle feedback effects and synergies Key events will rarely have occurred in isolation Numerous combinations ofprocesses will have operated at different times leading to a variety of outcomes Themost critical information and key ideas are set out below together with new evidenceproduced since the publication of the UNESCO survey in 1996
AGRICULTURE
The products of RomanomdashLibyan agriculture in the pre-desert are summarized in Table 81 (Gilbertson et al 1994 van der Veen et al 1996) They are essentially similar to theproducts of modern intensive mixed farming in many Mediterranean lands including theLibyan coastal plain and its hinterlands One noteworthy archaeozoological pattern wasthe increasing proportion of wild animals consumed with increasing distance south intothe more arid parts of the pre-desert Overall for most people meat was likely to have been a luxury item Of at least equal importance were the wool hair milk labour andmanure that domestic animals produced
There were many similarities to but also some differences from the more lsquonormalrsquo agricultural economy of the Mediterranean lands to the north For example the cropsproduced are similar to those of rain-fed agriculture to the north but the details of theagricultural practices used must have been notably different since precipitation in thepre-desert is both minimal and unreliablemdashless than 25ndash100 mm a year with substantial variability in both time and space In common with ancient arid-land farming in many other deserts agriculture and settlement in this arid land were dependent upon
Table 81 Farm products of the Tripolitanian pre-desert first to fifteenth centuries AD
Farm products Centuries ADPLANT CROPS 1ndash5 10ndash16 Hordeum vulgare (hulled six-rowed barley) + + Triticum (wheat) + + Pisum sativum (field pea) + + Lens culinaris (lentil) + + Other pulses + + Ficus carica (fig) + + Vitis vinifera (grape) + + Phoenix dactylifera (date) + + Olea europea (olive) + + Prunus amygdalus (almond) + + Pistacia atlantica (wild pistacia) + ANIMAL PRODUCTS Sheepgoat + +
Success longevity and failure of arid-land agriculture 141
harvesting rainwater with particular emphasis upon managing overland flow andcontrolling floodwater in the wadi floorsmdashpractices originally termed lsquofloodwater farmingrsquo by Bryan (1929) in his studies in the American Southwest
Unfortunately the palaeoeconomic analyses that underpin these ideas derive from aminute subset of sites in the study region Nevertheless the ULVS survey design didattempt to ensure that material was available from the primary types of agriculturalbuildings located during the project The studies of seeds and animal bones that underpinpresent understanding derive from excavations at middens or in buildings at only fouropen farms three of which were associated with olive presses and six gsur one of which was associated with an olive press
ENVIRONMENT
Knowledge of the detailed environmental history of the pre-desert from the mid-Holocene to the period of instrumental records is vital to understanding its human historyUnfortunately such understanding is often rudimentary
Palaeoenvironment
Only four palaeoenvironmental studies have been reported previously Three of thesesuggested the cultivation of olives in the RomanomdashLibyan period A study of the sedimentary fill of a karstic plateau-basin north of Beni Ulid indicated the presence of awetter climate during the early Holocene with shallow semi-permanent lakes surrounded by a grassy steppe perhaps with some scrub or trees in what are nowadays dry basins(Gilbertson et al 1994 Gilbertson et al 1994) Aridification took place from 4000 to5000 years ago creating an environment essentially similar to the modern arid steppe Astudy of cave deposits near Beni Ulid indicated the essential similarity of Romano-Libyan and modern conditions (Gale et al 1993) A third study analyzed pollen fromsediments infiltrated into a RomanomdashLibyan cross-wadi wall in the Wadi Mansur (Hunt et al 1986) suggesting a degraded steppe flora very similar to that of the modern Wadi
Gazelle + + Bovid + + Pig + Canid + + Camel + + Harerabbit + + Equid + + Antelope + + The present archaeobotanical evidence suggests that there were no fundamental differences between the agricultural economies of the RomanomdashLibyan Late Antique and Islamic periods In general hunted as opposed to herded animals became increasingly more important further south into the desert
The archaeology of drylands 142
Mansur and the cultivation of cereals The fourth study was a multi-disciplinary assessment of sediments infilling the conduit
that fed water to gasr Mm10 in the Wadi Mimoun (Hunt et al 1987) These deposits probably date from the abandonment of the gasr in the late Romano-Libyan period A landscape of steppe and scrub was suggested more biodiverse and perhaps wetter thanoccurs today Cereals were cultivated The high frequency of charcoal recovered suggestsburning nearby Interestingly pods of Medicago sp were also excavated from these deposits This plant is native to the pre-desert and is of considerable interest Its seeds were also recovered from RomanomdashLibyan deposits further south at Ghirza (van der Veen et al 1996 Fig 84) During the nineteenth century in South Australia a system of mediccereal rotation was developed by dryland farmers to improve nitrogen levels intheir soilsmdasha system still used today (Chatterton and Chatterton 1984) Medic-enriched grasses are sown and allowed to flower and produce seeds in the first season cereals aregrown in the second seeds from the first medic pasture then germinate to create anotherpasture in the third season The adoption of this lsquoAustralianrsquo system in parts of the Tripolitanian and Cyrenaican Gebel in the 1970s and 1980s increased cereal yields by asmuch as 50 per cent and allowed stocking rates to rise dramatically Chatterton andChatterton (1984) argued that if Romano-Libyan farmers had left land fallow for two or three years between cereal crops the resulting substantial medic pasture would haveimproved soil fertility and grazing Such a scenario is probable because in many areasrainfall would not have been sufficient every year to justify planting a cereal crop Overtime the ancient farmers may well have noticed the benefit of this type of crop rotation
A new palaeoecological study (Hunt unpublished data) is reported in outline hereThis is a palynological study of modern and Romano-Libyan coprolites from the middens and room fills of the farmstead Lm4 at Wadi el-Amud in the south of the pre-desert (Gilbertson and Hunt 1990) The modern samples reflect an extremely degradedenvironment with low local
Success longevity and failure of arid-land agriculture 143
Figure 84 A RomanomdashLibyan fortified farm (gasr) and its satellite buildings at Ghirza
Kite photograph GDBJones
pollen productivities and the local flora dominated by drought-resistant species In contrast the samples from contexts dating to the RomanomdashLibyan and Arab periods contain pollen of grasses and a diverse steppic flora with abundant pollen of cereals andolives reflecting crop plants Critically also the taphonomic patterns suggest that animalswere fed on monoculture cropsmdashgrasses cereal waste and chenopods Evidence from
The archaeology of drylands 144
Lm4 had previously suggested the stalling of animals at this site (Gilbertson and Hunt1990) This is a very different type of husbandry pattern than occurs today in thepredesert where goats forage widely and indiscriminately
Palaeoclimatology
Palaeoclimatic evidence from Morocco and the Nile basin suggests severe prolongedLate Holocene drought events The Late Holocene was notably drier than much of theEarly and Middle Holocene (Hassan 1981 1996 1997 1998 Lamb et al 1994 1995) Significant arid phases were identified for 4600ndash4000 BP 3800ndash3600 BP 2500 BP 2000 BP 1500 BP and approximately 700 BP (radiocarbon years) The flood record of theRiver Nile is especially interesting for the last 1500 years indicating low to very lowflows from AD 760 to 1070 with especially low flows between AD 930 and 1070 andbetween AD 1180 and 1350 (Hassan 1981)
Parallel evidence has not been found in the Tripolitanian pre-desert mainly because deposits suitable for investigation are rare and these phases may not be resolved at thestudied sites The interior of Libya is heterogeneous and environmentally complex andclimate changes occurring elsewhere in North Africa may not necessarily havemanifested themselves there in quite the same way Gilbertson and Hunt (1996) andNicholson (1989 1994) describe the regional climatology The quantity variability andreliability of precipitation are not well known In general annual precipitation averagesbelow 100 mm north of Beni Ulid to less than 25 mm in the south Thus Wadi Umm el-Kharab and Wadi el-Amud can be anticipated to have an unreliable and variableprecipitation regime averaging about 30 mm a year Nowadays drought may occur inmany consecutive years A yearrsquos rain may fall in just one or two very intense and localized rainstorms with adjacent areas remaining completely dry
In other arid regions of north Africa it is known that lsquodesert farmingrsquo was not sustained by harvesting rainwater or floodwater Rather it was supported by a reliableunderground water supply perhaps a spring as at Lemasba Algeria (Shaw 1982) oroases as in the Fezzan (Mattingly this volume Chapter 9 van der Veen 1992) Spring-fed oases supporting ancient agriculture are known at Gheriat el-Gharbia in the study region
Geomorphology
Understanding of the regional geomorphology is summarized in Anketell et al 1995 Gale et al (1993) Gilbertson et al (1993) Gilbertson and Hunt (1988 1996a) and Hunt et al (1986) Plateau-basins near Beni Ulid contain a well-developed palaeosol indicative of former wetter conditions presumably dating to the Early Holocene Screes and alluvialfans may well have developed on several occasions during the Holocene Overall in theperiod during and after extensive farming it is evident that there were several episodes ofslope erosion fluvial aggradation incision and aeolian reworking Anthropogenicdepositsmdashlarge middens layers of ash or dungmdashoccur notably at the olive farm at Wadiel-Amud (Gilbertson and Hunt 1990)
Success longevity and failure of arid-land agriculture 145
Erosion deposition and floodwater farming on wadi floors
Of particular interest are changes to the intensity of run-off and patterns of erosion and deposition resulting from floodwater-based agriculture on wadi floors During the passage of a storm pulse the roots of modern olives and date palms bind the floodplainsediments whilst intervening gullies can be scoured over 1 m deep leaving the trees oneroded pinnacles Elsewhere modern barley is successfully grown on the level surface ofrecently deposited flood-loams Cross-wadi walls promote fine-grained sedimentation and resultant increases in soil moisture seed catch and shrub growth on their upstreamside Downstream from walls waterfall-effects during flood promote gullies Later long after the walls are over-topped subsequent subsurface flow may promote temporarysprings sapping and piping These observations led to the development of a spatial modelto explain agricultural practice on wadi floors (Fig 85) The model also predicts where browse and shelter would have been available for stockmdashand thus an immediate source of the manure necessary to sustain intensive cereal cultivation (Barker et al 1996a Chatterton and Chatterton 1984)
Figure 85 Model of RomanomdashLibyan agriculture Note Zone A is a zone of deposition in quiet water upstream of a wadi wall used for cereal cultivation Zone B is the zone of turbulence downstream from a wadi wall where tree crops were grown Source After Gilbertson et al 1984
The alluvial and biological materials on the wadi floors are mobile and frequently reworked by wind rain storm and occasionally by burrowing or grazing animalsSubsurface processes are less securely known Field and laboratory evidence indicatesthat near-surface water is sometimes saline and it is not unreasonable to question whethersoil salinization may have been locally important in the past especially given the
The archaeology of drylands 146
deliberate introduction of large quantities of water on to wadi floors At present there isno evidence for large-scale salinization of wadi floors in RomanomdashLibyan or more recent times (Gilbertson 1996 Gilbertson et al 1993)
WALLS AND WALL NETWORKS IN RELATION TO RUN-OFF AND FLOODWATER FARMING
The spread of desert walls
The immense numbers of walls are one of the most important signs of the ancientfarming in the pre-desert (Fig 86) The vital role of walls in facilitating ancient farmingin drylands by trapping water and sediment has been recognized by numerous
Figure 86 Walls in the desert wall systems in the Wadi Mimoun near Gasr Lebr (Mm10)
Source After Hunt et al 1987
Success longevity and failure of arid-land agriculture 147
archaeologists and earth scientists (see Gilbertson 1986 Pacey with Cullis 1986)Brushwood fences were used in addition to stone walls to divert floodwaters at times ofrain-storm in the American Southwest (Bryan 1929 Nabhan 1986a 1986b Nabhan andSheridan 1977 and see Minnis Chapter 15) However fences for water control have notbeen detected in the pre-desert although thorn scrub is still in widespread use to corralanimals
The antiquity of walls
Inevitably the age relationships of walls are often unknown They are inferred fromnearby archaeological features whose antiquity was typically determined from associatedceramicsmdasha necessary first assumption but one that may often be incorrect It is alsoclear that walls are likely to have been frequently rebuilt reused repaired repositioned orreformed In some areas most walls were perhaps associated with the second majorphase of Romano-Libyan settlement characterized by the construction of gsurElsewhere many walls relate to the open farms of the earlier RomanomdashLibyan settlement phase Some may even be older and associated with the modest numbers oflater prehistoric settlements other walls may post-date the RomanomdashLibyan period as demonstrated by Hunt et al (1986) in the case of the Wadi Mansur
Design principles
In nearly every case studied the position of a plateau or wadi-side wall was apparently intended to maximize the quantitymdashand perhaps the ratemdashat which run-off was delivered to the wadi floor Often water was led from the plateau or hill slopes into cisterns manywith sediment traps or into caves at the wadi edge On the wadi floor the primaryobjective was apparently to capture floodwater causing it to sink into the long-term storage provided by the wadi-floor alluvium Occasionally water was conducted intocisterns adjacent to ancient settlements Many cisterns remain in use today or at leastthey still function Erosion appears to have been understood and managed by theRomano-Libyan farmers Numerous wadi-floor walls contain lsquodrop structuresrsquo reinforced gaps through a boulder-built wall often leading onto a stone-reinforced area immediately downstream They appear to be devices to avoid walls being overwhelmedand breached during flood peaks the reinforced surfaces downstream prevent scour andgully erosion These features still appear to be operating effectively with few displayingevidence of damage
A substantial literature describes the role of wall-managed floodwater for contemporary dryland management and developmentmdashnotably improving subsistence farming or as a means of soil reclamation (for example Evenari et al 1971 1982 Pacey with Cullis 1986 Reij 1991 van der Wal and Zaal 1990 and references therein) Eventhough the wall systems of the Libyan pre-desert were originally constructed two millennia ago the robustness of the technology is evident since they continue to harvestand channel run-off and storm water with marked ecological and biogeographic consequences (Gilbertson et al 1994)
The archaeology of drylands 148
Walls and risk management
Floodwater farming in Romano-Libyan times had to cover large areas and to some extent to be opportunistic it had to cope with the patchiness and unreliability of desertrainfall Substantial permanent investment of time and human energy in the constructionand maintenance of walls would have been necessary People had to sow seed on wettedsoil and may have travelled to wadis that had received run-off from localized rainfall events Major confluences or positions down-wadi must have been more reliable places to grow cereals especially in times of general drought A balance had to be struck in suchlocations between the opportunity to use the more frequent and larger run-off events and the risks posed by sequences of floods which would have eroded seed sown after earlierfloods
WALL FUNCTIONS
Six hypotheses have been proposed that singly or in combination might explain thefunction of the wall systems observed in the pre-desert (Gilbertson et al 1984) Walls may
These functions are not necessarily exclusive The same wall may have had one or moreuses when first constructed and later may have acquired or lost other roles
Walls whose primary purpose was clearly to delineate ownership have not been foundin the pre-desert Detailed surveys of wall distributions in the Wadi Umm al-Kharab indicate that cross-wadi walls were grouped and associated with different communities atvarious points along the length of the wadi (Gilbertson and Chisholm 1996) It is notknown how widespread this practice was Overall most walls appear related tohydrologicalgeomorphic factors whilst an absence of walls may reflect their deep burial(Gilbertson et al 1984) It is quite possible that the delineation of land ownership tenureor managementmdashif it was marked on the groundmdashoperated within the hydrological constraints of the wall systems The archaeological consequence is that it is very difficultto distinguish factors such as past community ownership or social groupings from thepresent information on wall networks
bull capture store and redistribute surface water for human and animal consumption and irrigation
bull control fluvial erosion sediment entrainment transport and deposition bull control the movements of animals either acting as pens and enclosures for
domesticated herds or by excluding animals from cultivated areas or by controlling wild animals during hunts
bull delineate areas of different land use bull represent the by-products of stone clearance to ease cultivation bull define parcels of land owned or controlled by different individuals or groups
Success longevity and failure of arid-land agriculture 149
ENVIRONMENTAL INFLUENCES
Within the long periods in which the wall systems were in use many droughts and otherenvironmental vicissitudes must have occurred Indeed the evidence describedpreviously from the Nile basin and Morocco suggests that droughts of 10ndash50 yearsrsquo duration probably occurred several times during the last two millennia Large parts of thepre-desert are likely to have been abandoned at times of prolonged drought once buffer stocks and imported feed were exhausted Nevertheless no direct evidence for temporaryabandonment yet existsmdashat present both the environmental and archaeological data aretoo coarse to distinguish such events Similarly there are no clear indications that thedevelopment and major shifts in the form and distribution of RomanomdashLibyan settlement or its subsequent decline were associated with some form of climatic changeor fluctuation Nevertheless the existing palaeoecological and palaeoclimatic informationfrom the study area suggests that the climate during much of the period of RomanomdashLibyan settlement was not dissimilar to that which prevails today though vegetation wasgenerally less degraded
At present the essential robustness and the long-term duration of floodwater farming in the pre-desert as well as the available palaeoecological evidence and modernecological theory provide no support for many possible explanations of region-wide changes in settlement or movement out of the Libyan pre-desert (Gilbertson 1996 see above) In brief the widely argued litany of anthropogenic agencies of desertificationdoes not seem to have played a central role in transforming the widely farmed and settledRomano-Libyan pre-desert into the modern arid wilderness The possible significance of disease and synergistic or feedback effects though remains completely unknown
HUMAN AGENCIES
As a result of the analysis described previously broad-scale interpretations of the ancient settlement and farming in the pre-desert must focus upon human agencies of change the outcomes of developments in the economic military political psychological and socialworlds (Barker et al 1996a Mattingly 1996) In brief the prime factor encouraging the Macae tribal pastoralists to become sedentary floodwater farmers appears to have beenregional stabilization resulting from the expansion of Roman influence into TripolitaniaEffective incorporation into the wider imperial economy produced different patterns ofland use greater stability access and trade with the vast new market and a majorincrease in population There are no grounds for believing that a widespread militarycolonization by soldier-farmers (limitanei) or a frontier army ever played a significantrole as was once suspected (Goodchild and Ward-Perkins 1949) Neither is there any indication that the expertise ideas or technology of floodwater farming were introducedfrom outside Probably the desert dwellers developed these approaches indigenously Asdemonstrated in many chapters in this volume the essence of this technology wasrepeatedly invented in antiquity in very different places
The archaeology of drylands 150
It remains unproven whether the replacement of open undefended farms by gsurshould be related to a greater sense of insecurity or whether the rich and powerful of thetime adopted these imposing enclosed structures to follow fashion as a display ofprestige or because the shade size and airiness of such buildings were well adapted tothe rigours of desert life (Fig 84) The progressive abandonment of settlement and farms in the southern part of the pre-desert region is perhaps best attributed to the widerpolitical and economic changes throughout the Mediterranean at the end of the Romanera with the development of smaller more regionalized group identities (Mattingly1996) There are no grounds for suspecting that the arrival of the Arab armies in the AD640s and later brought about major changes in pre-desert settlement or farming Floodwater farming was to continue at smaller scales for another thousand years indeedit continues to be practised today in the region
THE DECLINE OF FLOODWATER FARMING
It is clear that dramatic explanations of the abandonment of the floodwater farmingsystems as the result of climatic economic or political change are not congruent with thehistory of the pre-desert as presently understood It is also clear that ecologicaldegradation of the landscape for example at the Lm4 farm post-dated the end of RomanomdashLibyan floodwater farming Floodwater farming seems to have come to an endgradually on a piecemeal basis in some areas though there may have been rapid earlyretreats from the southernmost outposts such as Wadi el-Amud as these became uneconomic with the collapse of long-distance trade networks in late RomanomdashLibyan times It is clear that partial use of systems such as at Mm10 and Lm4 continued afterformal use of the RomanomdashLibyan buildings ended People continued to grow cereal crops and keep stalled animals though they often no longer lived in the RomanomdashLibyan buildings At this stage the landscape still had a distinctly steppic aspect
The population of the pre-desert was never very large For maximum efficiencylabour-intensive maintenance of the wall networks is essential One might envisage that as Roman influence waned and the political landscape became unstable intensiveinvestment in farming complexes became a risky strategy People began to readoptlsquobedouinrsquo ways of life which are flexible and in many ways less arduous than living infixed settlements in this region As people abandoned buildings for tents a transhumantlifestyle became possible and people started to move to where rain had fallen mostintensively each year to grow their crops Because of increased mobility it was no longerimportant that walls were rigorously maintained Systems would be abandoned as they became inefficient The end of animal stalling as seen at Lm4 would have placedadditional stress on the landscape because grazing removed the steppe vegetation and ledto the modern pre-desert ecology Rapid alluviation events in the medieval or early post-medieval periods may perhaps have been linked with landscape degradation of this type(Barker with Gilbertson 1996 Gilbertson and Hunt 1996a)
Success longevity and failure of arid-land agriculture 151
CONCLUSION
The systems of floodwater farming developed by Romano-Libyan farmers in the Tripolitanian pre-desert seem to have been lsquosustainablersquo in the true sense of the word They have persisted in some localities for two millennia surviving the fall of empiresmajor economic catastrophes climate fluctuations and changes and other adversitiesMuch of the resilience of the floodwater farming systems is clearly the result of theexploitation of detailed (if informal) understanding of run-off and fluvial processes and local geomorphological conditions by the local population from the Macae tribespeople onwards together with their engineering skills and their capacity to take advantage ofpatchy and unreliable storms Their farming systems seem to have been well adjusted tolocal conditions Details are however still sparse The hypothesis of ageomorpbiologically-adjusted polyculture with tree crops in erosive areas and grain crops under-planted with medic pasture in depositional areas is plausible but unproven The possibility that RomanomdashLibyan farmers stalled their stock is significant becauseanimals kept this way would be less able to de-vegetate and thus degrade the landscape The end of floodwater farming seems to have been piecemeal and gradual and not linkedto most of the cited lsquopush-factorsrsquo such as the fall of Rome which were relatively rapidIt may be that the bedouin lifestyle simply became more attractive to the small populationof the Tripolitanian pre-desert
REFERENCES
Anketell MJ Ghellali SM Gilbertson DD and Hunt CO (1995) Quaternaryfloodplain and wadi floor infill deposits in northeastern Libya and their implicationsfor landscape development In JLewin MMacklin and JMWoodward (eds)Quaternary Mediterranean River Environments 231ndash44 Amsterdam AA Balkema
Barker G with Gilbertson DD (1996) Farming the desert retrospect and prospect InGBarker DDGilbertson GDBJones and DJMattingly Farming the Desert The UNESCO Libyan Valleys Archaeological Survey Volume 1 Synthesis 343ndash63 Paris UNESCO Publishing
Barker G and Gilbertson DD with Hunt CO and Mattingly DJ (1996) RomanomdashLibyan agriculture integrated models In GBarker DDGilbertson GDBJones andDJMattingly Farming the Desert The UNESCO Libyan Valleys Archaeological Survey Volume 1 Synthesis 265ndash90 Paris UNESCO Publishing
Barker G Gilbertson DD Jones GDB and Welsby DA (1991) The UNESCOLibyan Valleys Survey XXIII the 1989 season Libyan Studies 2231ndash60
Barker G Gilbertson DD Jones GDB and Mattingly DJ (1996a) Farming the Desert The UNESCO Libyan Valleys Archaeological Survey Volume 1 SynthesisParis UNESCO Publishing
Barker G Gilbertson DD Jones GDB and Mattingly DJ (1996b) Farming the Desert The UNESCO Libyan Valleys Archaeological Survey Volume 2 Gazetteer and
The archaeology of drylands 152
Pottery Paris UNESCO Publishing Barth H (1857) Travels and Discoveries in North and Central Africa London
Longman Bryan RK (1929) Floodwater farming Geographical Review 19 (3)444ndash56 Chatterton BA and Chatterton L (1984) Medicagomdashits possible role in Romanomdash
Libyan dry farming and its positive role in modern dry farming Libyan Studies15157ndash60
Dorsett JE Gilbertson DD Hunt CO and Barker G (1984) The UNESCO LibyanValleys Survey IX image analysis of Landsat data and its application to environmentaland archaeological Surveys Libyan Studies 1571ndash80
Evenari M Shanan L and Tadmor N (1971) The Negev The Challenge of a DesertCambridge Mass Harvard University Press
Evenari M Shanan L and Tadmor N (1982) The Negev The Challenge of a DesertCambridge Mass Harvard University Press second edition
Flower CP and Mattingly DJ (1995) ULVS XXVII mapping and spatial analysis ofthe Libyan Valleys Data using GIS Libyan Studies 2649ndash78
Gale SJ Gilbertson DD Hoare PG Hunt CO Jenkinson RDS Lamble APOrsquoToole C van der Veen M and Yates G (1993) Late Holocene environmental change in the Libyan pre-desert Journal of Arid Environments 241ndash15
Gilbertson DD (1986) (ed) Run-off Farming in Rural Arid Lands Theme Volume 6 (1 and 2) of Applied Geography
Gilbertson DD (1996) Explanations environment as agency In GBarker DDGilbertson GDBJones and DJMattingly Farming the Desert The UNESCOLibyan Valleys Archaeological Survey Volume 1 Synthesis 291ndash318 Paris UNESCO Publishing
Gilbertson DD and Chisholm NT (1996) Manipulating the desert environmentancient walls floodwater farming and territoriality in the Tripolitanian pre-desert of Libya Libyan Studies 2717ndash52
Gilbertson DD and Hunt CO (1988) The UNESCO Libyan Valleys Survey XIX theCenozoic geomorphology of the Wadi Merdum Beni Ulid in the Libyan pre-desert Libyan Studies 1995ndash121
Gilbertson DD and Hunt CO (1990) The UNESCO Libyan Valleys Survey XXIgeomorphological studies of the Romano-Libyan Farm its floodwater control structures and weathered building stone at site Lm4 at the confluence of the Wadi elAmud and the Wadi Umm el Bagul in the Libyan pre-desert Libyan Studies 21 25ndash42
Gilbertson DD and Hunt CO (1996a) Quaternary geomorphology and palaeoecologyIn GBarker DDGilbertson GDBJones and DJMattingly Farming the Desert The UNESCO Libyan Valleys Archaeological Survey Volume 1 Synthesis 49ndash82 Paris UNESCO Publishing
Gilbertson DD and Hunt CO (1996b) RomanomdashLibyan agriculture walls and floodwater farming In GBarker DDGilbertson GDBJones and DJ MattinglyFarming the Desert The UNESCO Libyan Valleys Archaeological Survey Volume 1Synthesis 191ndash216 Paris UNESCO Publishing
Gilbertson DD Hayes PP Hunt CO and Barker G (1984) The UNESCO LibyanValleys Survey VII a classification and functional analysis of ancient irrigation and
Success longevity and failure of arid-land agriculture 153
wall systems in the Libyan pre-desert Libyan Studies 1545ndash70 Gilbertson DD Hunt CO and Fieller NRJ (1993) ULVS XXVI sedimento-logical
and palynological studies of Holocene environmental changes from a plateau basininfill sequence at Grerat Drsquonar Salem near Beni Ulid in the Tripolitanian pre-desert Libyan Studies 241ndash19
Gilbertson DD Hunt CO Fieller NRJ and Barker G (1994) The environmentalconsequences and context of ancient floodwater farming in the Tripolitanian pre-desert In ACMillington and KEPye (eds) Environmental Change and GeomorphicProcesses in Arid Lands 229ndash51 Chichester John Wiley and Sons
Goodchild RG and Ward-Perkins JB (1949) The Limes Tripolitanus in the light of recent discoveries Journal of Roman Studies 3981ndash95
Hassan FA (1981) Historical floods and their implications for climatic change Science2121142ndash5
Hassan FA (1996) Abrupt Holocene climatic events in Africa In GPeti and R Soper(eds) Aspects of African Archaeology 83ndash9 Harare University of ZimbabwePublications
Hassan FA (1997) Holocene palaeoclimates of Africa African Archaeological Review14 (4)213ndash30
Hassan FA (1998) The archaeology of North Africa at Kiekrz 1997 African Archaeology Review 15 (1)85ndash93
Hunt CO Mattingly DJ Gilbertson DD Barker G Dore JN Burns JRFleming AM and van der Veen M (1986) The UNESCO Libyan Valleys SurveyXIII interdisciplinary approaches to ancient farming in the Wadi Mansur TripolitaniaLibyan Studies 177ndash47
Hunt CO Gilbertson DD van de Veen M Jenkinson RDS Yates G andBuckland PC (1987) The UNESCO Libyan Valleys Survey XVII the palaeoecologyand agriculture of the abandonment Phase at Gasr Mm10 Wadi Mimoun in theTripolitanian pre-desert Libyan Studies 181ndash14
Lamb HF Duigan CA Gee JHR Keits K Lister G Maxted RW Merzouk ANiessen F Tahri M Whittington RJ and Zeroual A (1994) Lacustrinesedimentation in a high altitude semi-arid environment the palaeo-limnological record of Lake Isli High Atlas Morocco In ACMillington and KEPye (eds)Environmental Change and Geomorphic Processes in Arid Lands 229ndash51 Chichester John Wiley and Sons
Lamb HH Gasse F Benkaddour A El Hamouti N van der Kaar S Perkins WTPearce NJ and Roberts CN (1995) Relations between century-scale Holocene arid intervals in tropical and temperate zones Nature 373134ndash7
Mattingly DJ (1989) Farmers and frontiers exploiting and defending the countryside ofRoman Tripolitania Libyan Studies 20135ndash53
Mattingly DJ (1996) Explanation people as agency In GBarker DDGilbertsonGDBJones and DJMattingly Farming the Desert The UNESCO Libyan ValleysArchaeological Survey Volume 1 Synthesis 319ndash42 Paris UNESCO Publishing
Mattingly DJ with Flower C (1996) RomanomdashLibyan settlement site distribution and trends In GBarker DDGilbertson GDBJones and DJMattingly Farming the Desert The UNESCO Libyan Valleys Archaeological Survey Volume 1 Synthesis
The archaeology of drylands 154
159ndash90 Paris UNESCO Publishing Nabhan GP (1986a) Papago Indian desert agriculture and water control in the Sonoran
Desert 1697ndash1934 Applied Geography 66 (1)42ndash3 Nabhan GP (1986b) lsquoAk-cintilde lsquoarroyo-mouthrsquo and the environmental setting of Papago
Indian fields in the Sonoran Desert Applied Geography 6 (1)61ndash75 Nabhan GP and Sheridan TE (1977) Living fencerows of the Rio San Miguel
Sonora Mexico traditional technology of floodplain management Human Ecology597ndash111
Nicholson SE (1989) Long term changes in African rainfall Weather 4446ndash56 Nicholson SE (1994) Rainfall fluctuations in Africa and their relationship to past
conditions over the continent The Holocene 4(2)121ndash31 Pacey A with Cullis A (1986) Rainwater Harvesting The Collection of Rainfall and
Run-off in Rural Areas London Intermediate Technology Productions Reij C (1991) Indigenous Soil and Water Conservation in Africa London International
Institute for the Environment and Development Gatekeeper Series no SA27 Shaw BD (1982) Lamasba an ancient irrigation community Antiauiteacutes Africaines 18
61ndash103 van der Veen M (1992) Garamantian agriculture the plant remains from Zinchecra
Fezzan Libyan Studies 237ndash39 van der Veen M Grant A and Barker G (1996) RomanomdashLibyan agriculture crops
and animals In GBarker DDGilbertson GDBJones and DJMattingly Farming the Desert The UNESCO Libyan Valleys Archaeological Survey Volume 1 Synthesis227ndash64 Paris UNESCO Publishing
van der Wal A and Zaal F (1990) Bibliography on Indigenous Soil and Water Conservation with Special Reference to Africa Amsterdam Vreije Universiteit Center for Development Cooperation Services
Success longevity and failure of arid-land agriculture 155
9 Twelve thousand years of human adaptation in
Fezzan (Libyan Sahara) DAVID MATTINGLY
INTRODUCTION
With a few notable exceptions (Bousquet 1996 Nesson et al 1973 Trousset 1986) the archaeology and long-term history of the Saharan oases remain poorly documented Inmany cases pioneer studies have not been followed up in recent decades (Ball andBeadnell 1903 Fakhry 1974 RSGI 1937 Scarin 1934 1937) Yet there isundoubtedly much to learn from the manner in which desert people have exploitedresources and mastered the limitations of their environment A better understanding ofhuman adaptation to the desert environment has clear relevance for modern concernsabout the sustainability of oasis farming
To illustrate this theme this chapter will focus on the Fezzan Project which I direct and which has completed four seasons of work (1997ndash2000) The project is investigating the archaeology of a region of the Libyan Sahara c1000 km south of Tripoli (Fig 91) and follows on from earlier British work carried out by the late Charles Daniels Hisexploration and excavations from 1958 to 1977 accumulated a vast dossier of informationon one of the most important Saharan peoples of classical antiquitymdashthe Garamantes (Daniels 1969 1970 1971 1989 cf also Pace et al 1951) The full publication of his work is being undertaken in parallel with the renewed work (Edwards et al 1999)
The Garamantes were the dominant power in the Libyan Sahara from c500 BC to cAD 500 and at the height of their influence they controlled a vast desert territory ofc250000 km2 at times threatening both the Romanized cities of the Mediterranean coast and the sub-Saharan populations of Chad and Niger Liverani (1999) for example describes Garamantian forts on the routes south of Ghat itself 300 km southwest of theGaramantian capital They were several times defeated by Roman armies sent againstthem but their territory was never annexed to the Roman empire and for much of theRoman period they seem to have thrived on a combination of oasis agriculture and trade(Mattingly 199533ndash7 68ndash77 on relations with Rome)
Figure 91 Map showing the location of the Fezzan and the area of most detailed survey around Germa
The renewed fieldwork has aimed to amplify this picture by setting the Garamantes in a longer-term framework of human lifeways in the regionmdashbroadly focusing on the Holocene but with backward glances at the very different Pleistocene environment(Mattingly et al 1997 1998a 1998b 1999a 1999b) At the heart of the project is aconcern with human interaction with the environment and a study of how this has variedover an extended period of time and changing conditions The Fezzan Project then hasrelevance to wider debates than parochial Libyan ones though the evolution of Libyanculture and society in the desert is of major importance in its own right (Bates 1914Brett and Fentress 1996 Camps 1980) The transition to farming the emergence ofsocial complexity and the formation of a distinctive Saharan culture were all achieved ina region undergoing massive climatic degradation and desiccation
DESERT LANDSCAPES
The Libyan Desert is an area almost the size of India (Bagnall 1935 map 1) but with atiny populationmdashLibya itself has under five million people most concentrated in thecoastal cities The desert is an extraordinarily difficult environment to live well in and asin most desert regions water is a critically scarce resource in large parts of the countryMost of Fezzan has negligible annual rainfall and today depends entirely on subterraneanfossil water sources for sustaining its human population its livestock and its areas ofcultivation In this respect the region has very different characteristics to the Libyan pre-desert zone between Fezzan and the Mediterranean coast studied in an earlier project
Twelve thousand years of human adaptation in Fezzan 157
(Barker et al 1996 Gilbertson et al Chapter 8) where careful harvesting of the limited rainfall proved the key to past exploitation The story of farming the desert in Fezzan isone of people evolving strategies to utilize more effectively the huge groundwaterresources at the limited number of locations where they are relatively accessible at orclose to the surface Over the last few millennia the groundwater levels appear to havefallen significantly leading to small lakes drying up spring-lines ceasing irrigation systems being abandoned wells being deepened and so on As described below thestrategies evolved to tap into the groundwater were at certain times highly sophisticatedand in all probability highly organized within the society The peak population level(before the modern era) seems to have been reached in the Garamantian period when thematerial prosperity of the region also reached its apogee
The landscapes of Fezzan are very variablemdasha mixture of great sand seas gravel andboulder-strewn wastelands and hyper-arid rock plateaux (Fig 91) The project focuses on a long depression aligned east-west called the Wadi el-Agial (also known as al-Hayat) though it is not in fact a true dry river (the normal meaning of wadi) The el-Agial depression contains a chain of small oases over a length of about 150 km drawing on aseries of aquifers The traditional pattern of cultivation involves scattered palm groves with intensively irrigated plots of wheat barley and sorghum Duveyrier (1864147ndash216 439) and Lyon (1821270ndash78) both give good accounts of Fezzanese plants and cultivation systems in the nineteenth century There are no perennial springs here todaythough there are hints in the landscape that at some point in the past there was an activespring-line along the south side of the valley Since the invention of diesel and electrical pumps extraction of water from the aquifers has accelerated greatly and has caused waterlevels in many wells to fall by up to 100 m in the last century As we shall see there areimportant lessons for the present to be learnt from the past history of human activity andover-exploitation of this resource
A vast sand sea (ergedeyen) rises on the northern side of the oases Although the scale of the dunes is forbidding and the crossing of them can be perilous (Denham andClapperton 1826177ndash85) water was once more abundant within the sands and even today there is a number of small relict lakes which sustain small stands of date palms Inthe neolithic and classical periods there were undoubtedly more of these lakesfacilitating travel across and life within the sands
The south side of the Wadi el-Agial is dominated by a sheer cliff-like escarpment behind which extends a great sandstone plateau (hamada) turned black by desert varnish and dissected by deep gorge-like wadis running off to the south and southeast Some ofthe strata in this formation are fine-grained silicified sand- and mud-stones which were exploited extensively as a source for stone tools in the palaeolithic period and to a lesserextent in the mesolithic and neolithic periods There are no perennial water sources onthe plateau itself the main aquifer lies deep beneath it but this is the one part of theFezzan to receive rain with any regularity at certain times of year pools of water can befound in the wadi beds along with some rough grazing Engraved rock art dating to boththe later prehistoric and historic periods is abundant in many of the hamada wadisrepresenting seasonal exploitation of this forbidding landscape by hunters and mobilepastoralists (Lutz and Lutz 1995) Even at the peak periods of oasis cultivation it is clearthat pastoral groups operated alongside cultivators in exploiting the potential of the desert
The archaeology of drylands 158
landscapes (Nicolaisen and Nicolaisen 1997 UNESCO Nomades 1962) The consensus scientific view is that the Saharan region in the middle palaeolithic
period (say 250000ndash40000 BP) was very much wetter than today with an abundance of vegetation and wildlife flourishing around vast inland lakes (Petit-Maire 1982 Petit-Maire et al 1980 Ziegert 1995) Thereafter in the late Pleistocene and early Holoceneconditions became subject to a series of dramatic swings from wetter to drier conditionsand back again (Fig 92) The neolithic period (c6000ndash1000 BC) was generally one of worsening conditions (Lutz and Lutz 1995 Petit-Maire 1988 Shaw 1976) and by about 3000 BC it is likely that the Saharan climate was much as today (Cremaschi 1998)However subterranean water sources based on the huge
Figure 92 The major climatic fluctuations of the Holocene in the Libyan Sahara
Note The lower part of the diagram shows possible phases of Saharan rock art Source After Lutz and Lutz 1995
Continental Intercalate aquifer system (Edmunds and Wright 1979 Zaluski and Sadek1980) may have been more abundant and more readily accessible at that time (moresprings small lakes and shallow aquifers) with wild fauna more diverse as a resultNeolithic rock engravings show that at the start of this period the Sahara supported alarge and rich wild fauna including species like the crocodile which require permanentwater but that this was crucially changed with increasing aridity leading to extinction ofmany species north of the Sahara and to major changes in the lifestyle of the survivinghuman groups (Barker 1989 Encyclopeacutedie 1997 Le Quellec 1987 Lutz and Lutz 1995 Mori 1969 1988) In the rock art we see evidence of the domestication of animalsand an increasing emphasis on fertility and ritualmdashperhaps reflecting the social stress caused by environmental change (Encyclopeacutedie 19972791ndash96 2800ndash2 Lutz and Lutz 1995145ndash65 169ndash75) The same trends towards domestication of animals are evidentalso in the few well-excavated neolithic rock shelters (Barich 1987 Cremaschi and di Lernia 1998)
Twelve thousand years of human adaptation in Fezzan 159
AIMS OF THE FEZZAN PROJECT
The project addresses a number of key questions
The project geomorphologists have been studying the regional hydrology using field dataand remote sensing techniques to map gypsum formations which are indicative ofancient springs and small dried-up lakes sampling palaeo-lake sediments cross-sectioning spring mounds and assessing dune morphology (Mattingly et al 1998b117ndash22 1999b129ndash31) A series of possible prehistoric lake sediments has been identified at various points in the landscape and dating of these is a priority of continuing work toestablish whether they are of Pleistocene or Holocene date Laboratory techniques being used in support of the programme of palaeoenvironmental reconstruction include stableisotope analysis particle size distributions and mineral magnetic analysis of the putativelake deposits to characterize them combined with uranium-thorium and optically stimulated luminescence techniques of dating The uranium-thorium dating is being used on both Melanoides tubercolota shells found in association with dark organic lake-edge deposits and on gypsum crystals from a line of defunct springs at the foot of theescarpment With these methodologies we are hoping to be able to track through well-dated contexts the shrinkage and disappearance of the lakes with the onset of desiccationperhaps accompanied by the drying up of the palaeo-spring-line at the foot of the escarpment
The primary archaeological component of the project is the excavation of a site within the major ancient urban centre of the region Old Germa (or Garama as it was known inantiquity) (Fig 93) This is a still-standing medieval caravan town controlling one of the larger and more fertile oases of the el-Agial and situated on a trans-Saharan trade route There is a complex stratigraphy of a sequence of earlier cities superimposed one onanother to a depth of 4ndash5 m Some earlier clearance excavation (Ayoub 1967a) hasrevealed a group of Garamantian buildings at the core of the site Unlike most of the laterstructures these have stone walls and reflect the power and wealth of the site in itsheyday in the period between the first and fourth centuries AD The origins of the
bull the transition to farming in the Saharan region and in particular the origins of agriculture
bull the diffusion or invention of farminghydraulic technology and the spread of different cultivated plants
bull the response of human populations to the climatic and environmental changes bull the origins of urbanization in the Sahara and its evolution over time bull the construction of identity through material culture bull inter-regional contact across the Sahara (trade) bull processes of desiccation (and desertification) in the northern Sahara bull the recognition of palaeo-hydrological features in the landscape (spring-lines lakes
marsh) bull the dating of changes in the hydrology bull the identification and dating of evidence of climatic and environmental change
The archaeology of drylands 160
settlement go back until at least the fifth century BC again with a sequence of mud-brick buildings
Figure 93 The settlement of Germa (ancient Garama) the capital of the Garamantes
Photograph DMattingly
The current excavations here are designed to refine knowledge of this long urban sequence producing a series of time-slices illustrating the entire history of this remarkable site The material culture revealed demonstrates clear change over timemdashsome phases are unmistakably more impoverished than others For instance much of themedieval and early modern periods is characterized by relatively low numbers ofimported goods despite the existence of trans-Saharan trade at this time In the Garamantian period by contrast an abundance of wine and olive oil amphorae ceramicfinewares and glass ware was imported from the Roman world (Fontana 1995)
Systematic sieving of deposits is recovering a sample of the plant fragments and animal bones present in each phase The good preservation of many plant fragments indesiccated form is significant because it means that a reasonably full range of cultivatedand weed species is represented in the samples This gives useful information about thelocal environment the cultivated plants all require irrigation and the weeds often reflectthe arid and salty background conditions We are also identifying a series of significantbotanical horizons including a lsquomaize horizonrsquo representing the coming of New World crops and a lsquosorghum horizonrsquo representing northward transference from the Sahel orSudan probably within the Garamantian period This study builds on earlier work by vander Veen (1992) and is extending our knowledge of changing patterns of plant cultivationback from the present to c900 BC (for the broader North African context see van der Veen 1995 1999) Analysis of the faunal remains is also indicating change over time
Twelve thousand years of human adaptation in Fezzan 161
with the bones providing important information about not only what stock was raised butalso the age at which animals were slaughtered and the butchery techniques used
In order to provide a wider context for the picture of life in the town the excavation is complemented by fieldwalking and by more extensive survey in the Germa region Thisaims to build upon Danielsrsquo earlier survey which was successful in locating a large number of cemeteries of Garamantian date particularly in the form of cairn graves alongthe foot of the hamada escarpment He was less successful in locating Garamantiansettlements though he was aware of a few village-like sites in the oasis and a number ofhillforts along the escarpment Systematic fieldwalking has now revealed that theGaramantian settlement pattern was far denser than previously suspected with numeroussatellite villages all around Old Germa (Fig 91) The fieldwalking essentially logs thedensity of archaeological material (humanly-made or imported goodsmdashlithics ceramics ostrich shell beads and so on) and isolates significant concentrations of such material aslsquositesrsquo Topographic survey of a selection of these sites has added structural detail andconfirms that we are dealing with settlements and not simply rubbish disposal Thesurvey complements the evidence of a series of excavations by Daniels on additionalGaramantian sites which we are also preparing for publication Zinchecra an earlyGaramantian hillfort and cemetery (Daniels 1968) Saniat Ben Howedi a rich Romanperiod cemetery (Ayoub 1968 Daniels 1989) and Saniat Gebril an oasis village
Our fieldwork has discovered sites of many different phases of activity not simply theGaramantian phase On the hamada to the south of Germa we have recorded a series ofimportant lithic scatters of palaeolithic date comprising tools such as 100000-year old handaxes together with chunky waste flakes and chippings at the locations where toolswere produced In our 1999 season a series of neolithic occupation sites was alsoidentified close to the edge of the sand sea to the north of Germa These sites yieldingextremely finely worked lithics early pottery grindstones ostrich eggshell fragments andbeads were probably occupied in the last few millennia BC when climatic conditionswere rapidly deteriorating Their inhabitants seem to have exploited a shallow and nowvanished lake site
A gazetteer of ancient sites throughout the el-Agial is being compiled combining boththe Danielsrsquo material and the new work Transcription of a series of air photographs takenin the 1950s and 1960s is revealing a wealth of information now destroyed by moderndevelopment This work complements a programme of remote sensing using modernsatellite imagery comparison of the satellite imagery and the air photographs hasrevealed the extent to which deep-bore artesian wells have expanded the area under cultivation in the last twenty years but at the cost of dramatically lowering the regionalwater table
The extension of the cultivated area and the growth of modern villages that has accompanied it have particularly affected the preservation of one of the most importantand enigmatic classes of monumentmdashthe foggaras These are underground irrigation canals similar to the Persian qanat or the Arabian falaj or aflaj which tapped into an aquifer below the foot of the escarpment and led flowing water out into the oasis proper(Bousquet 1996 Goblot 1979 Klitzsch and Baird 1969 Locirc 1953 1954 Mattingly et al 1998a 190ndash2 1998b137ndash42 1999b139ndash42 Nesson et al 1973) They are readily identifiable at the surface where traces survive from the regularly spaced vertical shafts
The archaeology of drylands 162
that were dug to facilitate construction and maintenance of the channels though theymust have added hugely to the labour involved The shafts can be up to 20 m deepgradually diminishing in depth until the channel emerges at the surface (Fig 94) The available dating evidence indicates that the foggara system was introduced to Fezzanduring the Garamantian period with their use probably extending into the early Islamicperiod It is clear that these structures were a key to ancient irrigation in the regionthough evidently they have been dry for many centuries now There are many hundredsof these structures visible on the air photographs most being at least several km inlength The labour involved in their construction and maintenance was on a significantscale (Mattingly et al 1999b140ndash1)
HISTORICAL RECONSTRUCTION
What does all this new evidence add up to It is clear that there has been dramatic changein environment climate and human activity over time What
Figure 94 Schematic cross-section of a foggara tapping into water-bearing strata below the escarpment and leading flowing water along a tunnel to the oasis zone in the valley floor
follows is a very simplified and provisional analysis with suitable disclaimers attachedThe reconstruction proposed at this stage is essentially a series of models designed forfurther testing and elaboration The clear trend running through though is one of anoverall decrease in water availability over time Climatic change and the onset ofdesertification have reduced rainfall to negligible levels and caused old surface watersources such as lakes and springs largely to dry up
During the Upper Pleistocene the region is known to have been very different from thedesert environment it has become The hamada plateau is assumed to have been well-vegetated savanna with abundant rainfall supporting a large range of animals and hunter-gatherer human groupsmdashwhose tool assemblages occur in profusion across its surface It is generally agreed that last phase of the Pleistocenemdashthe period c40000ndash10000 BPmdashwas one of high aridity in North Africa reflected in the Fezzan in a dearth of evidence for
Twelve thousand years of human adaptation in Fezzan 163
the characteristic lithic assemblages of this phase The reappearance of substantial humanpopulations in the Early Holocene after 10000 BP can be related to a new period ofincreased rainfall (Fig 95) The landscape was well vegetated in this phase supporting a wide range of wild animals which were initially exploited through hunting especially onthe plateau and wadis of the hamada However in successive phases of further climaticchangemdashwhether major oscillations as indicated on Figure 92 or a more step-like progression towards acute aridificationmdashhuman settlement became increasingly focusedon locations where water was to be found at shallow depth Thus many sites presumablyseasonal camp-sites have been located in the el-Agial depression and around small lakeson the edge of the sand sea Because the mesolithic and neolithic phases were far from auniform period climatically it is necessary to undertake more work on the phasing ofsites of these periods through further analysis of tool types rock art phases and the
Figure 95 Model of the neolithic landscape around Germa with settlement and activity (stock raising and later cultivation) based around perennial water sources (lakes and springs) in the valley and the edge of the sand sea
The archaeology of drylands 164
evolution of pottery use In this respect the research by Cremaschi and di Lernia (1998)marks an important advance For the moment our model of later prehistoric settlement(Fig 95) includes sites from both wetter and drier phases of the Early Holocene
The huge climatic fluctuations of this period form a backdrop to the transition to farming here The domestication of animals can be traced both in the rock art (which canbe dated only in relative terms at present) and from some of the excavated rock sheltersThe exploitation of plant resources is most clearly signalled by the abundant grindstonesat the neolithic campsites by the lakes and water sources What is particularly interestingabout this transition in Libya is that it seems to arise as a response to adversity rather thanto opportunity people turned to stock raising and cultivation here during the fifth andfourth millennia BC when a dramatic change in the availability of water made a hunter-gatherer existence increasingly more precarious (Barker 1989 1996) The Fezzan Projectwill hopefully make an important contribution to these debates It is likely though hasnot yet been demonstrated that neolithic farmers grew their crops in small patches of soilnaturally irrigated by higher groundwater levels in contrast with the floodwater farmingsystems developed on the northern margins of the Sahara fringe by Romano-Libyan farmers (Barker et al 1996 see also Chapter 8)
With pastoralism and small-scale cultivation established there is then little evidence for significant change in subsistence through the third and second millennia BC Theperiod of the Garamantes however (between 900 BC and AD 500) marked a dramaticdevelopment in farming technologies and systems associated with transformations incultural complexity These transformations included
The Garamantes represent in part a continuation of the local neolithic tradition as is clearfrom lithic and ceramic finds at their early settlements But they probably comprised agreat confederation of tribes and there are indications that some elements may havemigrated from oases further east nearer Egypt bringing with them knowledge ofimproved technology for oasis cultivationmdashnotably the foggara There are clear parallels
bull the rise of a major polity and civilization in the Sahara (Daniels 1970 Ruprechtsberger 1997)
bull the development of urbanism (Daniels 1971262ndash5) bull the evolution of a hierarchical and probably slave-using society (Daniels 197027ndash
35) bull the adoption of a written script for the Libyan language (Daniels 1975) bull the further development of agriculture to encompass a range of Mediterranean and
desert crops that require intensive irrigation (cereals grapes olives dates) (Daniels 198956ndash58)
bull the introduction of the horse the camel and wheeled transport to the Sahara (Camps 1989)
bull the creation of trade and political relations that extended north to the Mediterranean east to Egypt and south to sub-Saharan Africa (Bovill 19681ndash44 Fontana 1995) and
bull a massive demographic expansion to a level that was probably not equalled again until the last forty yearsmdashDaniels (198949) estimated that there were at least 120000 Garamantian burials in the el-Agial alone
Twelve thousand years of human adaptation in Fezzan 165
for instance between the Libyan tribesmen on Egyptian reliefs and in rock art of southernLibya and Algeria (Lutz and Lutz 1995140ndash1 Ruprechtsberger 199766ndash9) Most of the early Garamantian settlements currently known are situated along the edge of theescarpment many in defensible positions such as the classic hillfort site of ZinchecraBotanical remains from sites like Zinchecra dating to the first half of the first millenniumBC demonstrate that irrigated cultivation had begun by that early date (Daniels 198956ndash8 van der Veen 1992)
The main phase of occupation at Zinchecra ended around 500 BC (van der Veen 199212ndash13) at which point it appears that an urban site originated at Germa Over timeGarama emerged as the Garamantian capital and in the Roman period was adorned withsubstantial public buildings and temples utilizing stone on a scale and with a quality ofdressing not previously witnessed Since there is no evidence to suggest a Romanoccupation of Fezzan these must be the result of contact diplomacy and trade betweenthe Roman empire and the Garamantian kingdom Garamantian culture nowhere betterillustrated than in its extraordinary funerary architecture was extremely eclecticmdashthough the variety of tomb types in contemporary use may also reflect the maintenance ofdiscrete tribal identities within the structure of the polity (Ayoub 1967b 1968 Daniels1971265ndash8 el-Rashedy 1988 Ruprechtsberger 199751ndash65)
The evolved settlement pattern (Fig 96) reflects the increasing localization of farmingactivity in the oases along the base of the depression In addition to the large urban centreat Garama there were regularly-spaced village settlements all along the valley to match the extensive evidence of cemeteries along the foot of the escarpment (tens of thousandsof graves have been recorded as noted above) Hundreds of foggaras facilitated the large-scale and extensive cultivation of the valley-floor oasis area A crucial question we arestill seeking an answer to is why these systems were abandoned perhaps it was becauseof falling water levels in the aquifer The settlement density the number and scale of thecemeteries and the foggara systems all combine to highlight the Garamantian period asone of peak population and oasis cultivation
Garamantian civilization was thus the result of raised population levels in the northernSahara following the development of advanced irrigation systems The concentration oftens of thousands of people in the largest of these oases allowed them to dominate a largeexpanse of the Saharamdashraiding and trading in equal measure to all points of the compassClassical sources speak of the Garamantes hunting the troglodytae and lsquoEthiopiansrsquo which gives a strong hint of slave raiding against neighbouring peoples (Herodotus4183 cf Tacitus Hist 450) Quite apart from the possibility of selling-on such captives north across the Sahara the intensive irrigated cultivation and the dangerous task offoggara construction could have absorbed large numbers of slaves The evidence for theexistence of trans-Saharan trade at this date is partial at best but the large quantities ofRoman trade goods found at Garamantian sites and in their burials indicate thatsomething of value must have been passing the other way Apart from slaves it ispossible that the Garamantes also traded in salt gold semi-precious stones and natron (the latter used in glass making) (Bovill 1968) The funerary evidence indicates theemergence of a social hierarchy with a prominent elite order enjoying significantlygreater wealth than the majority of the population who were still buried in relativelysimple cairn graves
The archaeology of drylands 166
In the early Islamic period some at least of the Garamantian villages appear to have continued and may have been embellished with castle-like structures (gsur see Ruprechtsberger 199777ndash81 for examples) built of mud brick
Figure 96 Model of the evolved Garamantian landscape around Germa with its extensive irrigation systems urban centre and satellite villages and numerous cemeteries
Over time however the number of villages seems to have declined markedly perhapslinked to a shift from foggara to well irrigation (Fig 97) The problem with irrigation based on wells is that water must be mechanically raised by bucket before being fed intoirrigation canals with the result that in general each well can irrigate only a limited areaof fields around it The late medieval and early modern pattern is thus of small clumps ofpalms and cultivated fields clustered around many scattered wells in contrast with theevidently more extensive areas that appear to have been cultivated whilst the foggaraswere operating
Twelve thousand years of human adaptation in Fezzan 167
Garama was displaced as the regional capital by sites further east and south (MurzukTraghen Zuila) but its substantial walls and kasbah guaranteed it a role in the politicsand warfare of the period Nonetheless when the earliest
Figure 97 Model of the medieval landscape around Germa showing shrinkage of the cultivated area and demographic decline after the failure of the foggara systems and the refocusing of agriculture around wells in the valley centre
European travellers penetrated into the Sahara in the late eighteenth and early nineteenthcenturies they found the Wadi el-Agial a desperately impoverished region with many ofits villages underpopulated and crumbling and the bulk of its agricultural productiontaken as taxes and rents by absentee sheiks and Turkish officials (Barth 1857143ndash9 Bruce-Lockhart and Wright 1999 Denham and Clapperton 1826169ndash77)
Only in the last forty years have modern artesian wells reversed the trend of declineand revived the population and agricultural productivity However this has been at a cost
The archaeology of drylands 168
to the aquifer levels which have already fallen significantly In the long term (Fig 98) it is possible that agriculture will be forced to contract around a limited group ofagricultural settlements with very deep bore artesian wells serving clusters of individualirrigated crop
Figure 98 Model of hypothetical future direction of settlement and farming in Fezzan with the concentration of population around a series of agricultural settlements irrigating large circular fields with very deep artesian wells
circles each of c300 m diameter This system developed elsewhere in Libya forexploiting fossil water supplies deep below the Sahara (cf Allan 1979) The FezzanProject cannot offer solutions to the problem of where water is to come from next but ithas graphically illustrated the human consequences of past changes in water availabilityin the desert Whilst we may take pride in human ingenuity in finding ways to live in thedesert we may also reflect on the environmental costs that such lsquomasteryrsquo brings in its wake Garamantian development of the foggara irrigation systems may in the long termhave been a key factor leading to the decline of their civilization as a result of over-extraction from a non-renewable groundwater source
Twelve thousand years of human adaptation in Fezzan 169
ACKNOWLEDGEMENTS
The Fezzan Project is sponsored by the Society for Libyan Studies the Arts andHumanities Research Board the British Academy and the University of Leicester The final publication of the earlier work by Daniels is supported by a major grant from theLeverhulme Trust This chapter was written during the tenure of a Research Readershipaward from the British Academy The project involves the work of many individualswho are thanked as a group here but whose contributions are clearly acknowledged inthe multi-authored interim reports Thanks are also due to DMiles-Williams for Figure 92 and LFarr for Figures 93ndash97
REFERENCES
Allan JA (1979) Managing agricultural resources in Libya recent experience Libyan Studies 1017ndash28
Ayoub MS (1967a) Excavations in Germa between 1962 and 1966 Tripoli Ministry of Education
Ayoub MS (1967b) The Royal cemetery at Germa A preliminary report Libya Antiqua3ndash4213ndash19
Ayoub MS (1968) The Cemetery ofSaniat Ben Howedy Tripoli Ministry of Education Bagnall RA (1935) Libyan Sands Travel in a Dead World London Hodder amp
Stoughton Ball J and Beadnell HJL (1903) Banana Oasis Its Topography and Geology Cairo
National Printing Department Barich BE (1987) (ed) Archaeology and Environment in the Libyan Sahara The
Excavations in the Tadrart Acacus 1978ndash1983 Oxford British Archaeological Reports International Series 368
Barker GW (1989) From classification to interpretation Libyan prehistory 1969ndash1989 Libyan Studies 2031ndash43
Barker G (1996) Prehistoric settlement In GBarker DGilbertson BJones andDMattingly Farming the Desert The UNESCO Libyan Valleys ArchaeologicalSurvey Volume 1 Synthesis 83ndash109 Paris UNESCO London Society for Libyan Studies
Barker G Gilbertson D Jones B and Mattingly D (1996) Farming the Desert The UNESCO Libyan Valleys Archaeological Survey Volume 1 Synthesis Volume 2Gazetteer and Pottery Paris UNESCO London Society for Libyan Studies
Barth H (1857) Travels and Discoveries in North and Central Africa London Longman Brown amp Green (reprint Longman 1965)
Bates O (1914) The Eastern Libyans London Frank Cass (reprint 1970) Bousquet B (1996) Tell-Douch et sa Region Geographic drsquoune Limite de Milieu agrave une
Frontiegravere drsquoEmpire Cairo Institut Franccedilaise drsquoArcheacuteologie Orientale Bovill EW (1968) The Golden Trade of the Moors Oxford Oxford University Press
The archaeology of drylands 170
(second edition) Brett M and Fentress E (1996) The Berbers Oxford Blackwell Bruce-Lockhart J and Wright J (1999) Difficult and Dangerous Roads Hugh
Clappertonrsquos Travels in Sahara and Fezzan (1822ndash1825) London Sickle Moon Press Camps G (1980) Berbegraveres Aux Marges de lrsquoHistoire Toulouse Hespeacuterides Camps G (1989) Les chars sahariens Images drsquoune societeacute aristocratique Antiquiteacutes
Africaines 2511ndash40 Churcher CS and Mills AJ (1997) Reports from the Survey of Dakhleh Oasis Western
Desert of Egypt 1977ndash1997 Oxford Oxford Archaeological Monographs Cremaschi M (1998) Late quaternary geological evidence for environmental change in
south-western Fezzan (Libyan Sahara) In MCremaschi and Sdi Lernia (eds) Wadi Teshuinat Palaeoenvironment and Prehistory in South-western Fezzan (Libyan Sahara) 13ndash47 Florence Insegna del Giglio
Cremaschi M and di Lernia S (1998) (eds) Wadi Teshuinat Palaeoenvironment and Prehistory in South-western Fezzan (Libyan Sahara) Florence Insegna del Giglio
Daniels CM (1968) Garamantian excavations Zinchecra 1965ndash1967 Libya Antiqua5113ndash94
Daniels CM (1969) The Garamantes In WHKanes (ed) Geology Archaeology and Prehistory of the Southwestern Fezzan Libya 31ndash52 Castelfranco-Veneto Petroleum Exploration Society of Libya
Daniels CM (1970) The Garamantes of Southern Libya London Oleander Daniels CM (1971) The Garamantes of Fezzan In FFGadallah (ed) Libya in History
Proceedings of a Conference Held at the Faculty of Arts University of Libya1968261ndash85 Benghazi University of Libya
Daniels CM (1975) An ancient people of the Libyan Sahara In JBynon and T Bynon(eds) Hamito-Semitica 249ndash65 The Hague Mouton
Daniels CM (1989) Excavation and fieldwork amongst the Garamantes Libyan Studies2045ndash61
Denham D and Clapperton H (1826) Narration of Travels and Discoveries in Northernand Central Africa in the Years 1822ndash1824 London John Murray (reprinted 1965 as Missions to the Niger III Haklyt Society second series CXIX edited by EW Bovill)
Duveyrier H (1864) Les Touaregs du Nord Paris Challamel Aineacute Edmunds WM and Wright EP (1979) Groundwater recharge and palaeoclimate in the
Sirte and Kufra basins Journal of Hydrology 11971ndash87 Edwards D Hawthorne J Dore J and Mattingly DJ (1999 The Garamantes of
Fezzan revisited publishing the CMDanielsrsquo archive Libyan Studies 30109ndash27 el-Rashedy F (1988) Les pratiques funeacuteraires des Garamantes et leurs relations avec
celles drsquoautres peuples drsquoAfrique du Nord In Libya Antiqua Histoire Geacuteneacuterale de lrsquoAfrique Etudes et Documents III 85ndash114 Paris UNESCO
Encyclopeacutedie (1997) Encyclopeacutedie Berbegravere fasc xviii sv Fezzan 2777ndash817 Aix-en-Provence Edisud
Fakhry A (1974) The Oases of Egypt II Bahariyah and Farfara Oases Cairo American University of Cairo Press
Fontana S (1995) I manufatti romani nei corredi funerari del Fezzan Testimonianza deicommerci e della cultura dei Garamanti (IndashIII secd C) In Productions et
Twelve thousand years of human adaptation in Fezzan 171
Exportations Africaines Actualiteacutes Archeacuteologiques 405ndash20 Paris VI Colloque International sur lrsquoHistoire et lrsquoArcheacuteologie de lrsquoAfrique du Nord
Goblot H (1979) Les Qanats une Technique drsquoAcquisition de lrsquoEau Paris and New York Mouton eacutediteur Industrie et Artisanat 9
Kanes WH (1969) (ed) Geology Archaeology and Prehistory of the Southwestern Fezzan Libya Castelfranco-Veneto Petroleum Exploration Society of Libya
Klitzsch E and Baird DW (1969) Stratigraphy and palaeohydrology of the Germa(Jarma) area southwest Libya In WHKanes (ed) Geology Archaeology and Prehistory of the Southwestern Fezzan Libya 67ndash80 Castelfranco-Veneto Petroleum Exploration Society of Libya
Le Quellec J-L (1987) LrsquoArt Rupestre du Fezzan Septentrional (Libye) Widyan Zredaet Tarut (Wadi esh-Shati) Oxford British Archaeological Reports International Series 365
Liverani M (1999) Ultime scoperte nella terra dei Garamanti Archeo Attualitagrave del Passato 15830ndash9
Locirc Capitaine (1953) Les foggaras du Tidikelt Travaux de lrsquoInstitut de Recherches Sahariennes 10139ndash79
Locirc Capitaine (1954) Les foggaras du Tidikelt Travaux de lrsquoInstitut de Recherches Sahariennes 11 49ndash77
Lutz R and Lutz G (1995) The Secret of the Sahara Innsbruck Golf Verlag Mattingly DJ (1995) Tripolitania London Batsford Mattingly DJ al-Mashai M Balcombe P Chapman S Coddington H Davison J
Kenyon D Wilson AI and Witcher R (1997) The Fezzan Project 1997methodologies and results of the first season Libyan Studies 2811ndash25
Mattingly DJ al-Mashai M Balcombe P Chapman S Coddington H Davison J Kenyon D Wilson AI and Witcher R (1998a) The Fezzan Project I researchgoals methodologies and results of the 1997 season Libya Antiqua ns 3 175ndash99
Mattingly DJ al-Mashai M Aburgheba H Balcombe P Eastaugh E Gillings M Leone A McLaren S Owen P Pelling R Reynolds T Stirling L Thomas DWatson D Wilson AI and White K (1998b) The Fezzan Project 1998 preliminaryreport on the second season of work Libyan Studies 29115ndash44
Mattingly DJ al-Mashai M Aburgheba H Balcombe P Eastaugh E Gillings M Leone A McLaren S Owen P Pelling R Reynolds T Stirling L Thomas DWatson D Wilson AI and White K (1999a) The Fezzan Project II preliminaryreport on the 1998 season Libya Antiqua ns 463ndash93
Mattingly DJ al-Mashai M Balcombe P Drake N Knight S McLaren S Pelling R Reynolds T Thomas D Wilson A and White K (1999b) The FezzanProject 1999 preliminary report on the third season of work Libyan Studies 30129ndash45
Mori F (1969) Prehistoric cultures in Tadrart Acacus Libyan Sahara In WHKanes(ed) Geology Archaeology and Prehistory of the Southwestern Fezzan Libya 21ndash30 Castelfranco-Veneto Petroleum Exploration Society of Libya
Mori F (1988) Lrsquoart rupestre preacutehistorique dans le Sahara libyen comme aboutissementdrsquoun long processsus bioculturel In Libya Antiqua Histoire Geacuteneacuterale de lrsquoAfrique Etudes et Documents III 157ndash63 Paris UNESCO
The archaeology of drylands 172
Muzzolini A (1991) Proposals for updating the rock-drawing sequence of the Acacus (Libya) Libyan Studies 227ndash30
Nesson C Rouvillois-Brigol M and Vallet J (1973) Oasis du Sahara Algeacuterien Paris Institut Geacuteographique National
Nicolaisen J and Nicolaisen I (1997) The Pastoral Touareg Ecology Culture andSociety London Thames amp Hudson two volumes
Pace P Sergi S and Caputo G (1951) Scavi Sahariani Monumenti Antichi 41 150ndash549
Petit-Maire N (1982) (ed) Le Shati Lac Pleistocene au Fezzan Paris CNRS Petit-Maire N (1988) Climatic change and man in the Sahara In JBower and D Lubell
(eds) Prehistoric Cultures and Environments in the Late Quaternary of Africa 19ndash42 Oxford British Archaeological Reports International Series 405
Petit-Maire N Delibrias G and Gaven C (1980) Pleistocene lakes in the Shati area Fezzan (27deg30primeN) In MSarnthein ESeibold and PRognon (eds) Sahara and the Surrounding Seas 289ndash93 Rotterdam Balkema Palaeoecology of Africa 12
RSGI (1937)=Real Societagrave Geographica Italiana Il Sahara Italiano Fezzan e Oasi diGat Rome Societagrave Italiana Arti Grafiche
Ruprechtsberger EM (1997) Die Garamanten Geschichte und kultur eines Libyschen Volkes in der Sahara Mainz Verlag P von Zabern
Scarin E (1934) Le Oasi del Fezzan Bologna Zanichelli two volumes Scarin E (1937) Le Oasi Cirenaiche del 29deg Parallelo Florence Sansoni Shaw BD (1976) Climate environment and prehistory in the Sahara World
Archaeology 82133ndash49 Trousset P (1986) Les oasis preacutesahariennes dans lrsquoantiquiteacute partage de lrsquoeau et division
du temps Antiquiteacutes Africaines 22161ndash91 UNESCO Nomades (1962) = Recherches sur la Zone Aride XIX Nomades et Nomadisme
au Sahara Paris UNESCO Publications van der Veen M (1992) Garamantian agriculture the plant remains from Zinchecra
Fezzan Libyan Studies 237ndash39 van der Veen M (1995) Ancient agriculture in Libya a review of the evidence Acta
Palaeobotanica 35 (1)85ndash98 van der Veen M (1999) (ed) The Exploitation of Plant Resources in Ancient Africa
New York Plenum Zaluski M and Sadek KE (1980) Hydrogeology of mesozoic aquifers in the western
part of Wadi al Ajal Symposium on the Geology of Libya 2635ndash42 Ziegert H (1995) Das neue Bild des Umenschen UniHH Forschung Beitrage aus der
Universitaumlt Hamburg 309ndash15
Twelve thousand years of human adaptation in Fezzan 173
10 Farming and famine subsistence strategies in
Highland Ethiopia ANN BUTLER AND A CATHERINE DrsquoANDREA
INTRODUCTION
The Highlands of Ethiopia have an environment that is governed by the high altitude andwithin relatively low longitudes they have a temperate climate This supports a widerange of crops which include both indigenous African domesticates and the cool-season grain crops developed in and introduced in antiquity from southwest Asia Cultivation israinfed and the technology is largely ox-plough The region therefore presents an ideal situation for the study of traditional dryland agriculture and provides an opportunity tounderstand some of the rationale that underlies these farming practices The results of anew ethnoarchaeological study in the Ethiopian Highlands are presented here integratedwith some published accounts of traditional agriculture in the region
FIELDWORK
Ethnobotanical studies were carried out between 1996 and 1998 in the EthiopianHighlands (Tigrai province) about 2000 m above sea level in the mid-altitude region This is the agro-climatic zone known as dry woina dega (Bekele-Tesemma 19936) The average temperature range is between 5 and 40degC (Gebremedin and Haile 1997) Fieldwork was concentrated on the northern edge of the Giba plateau in the Endertaadministrative region (woreda) and village group (tabia) of Mahabere Genet about 15 km northwest of the provincial capital Mekelle Adi Ainawalid a village (kushet) of 180 households was selected for a detailed study (Fig 101) Supplementary records were made both at further kushets within the same tabia and also at others within the woredaof Entalo-Wajeret near Adi Gudem about 30 km south of Mekelle Farming practices were observed and farmers were interviewed between May and June and during the mainharvest time between November and December (Butler in press Butler et al 1999 DrsquoAndrea et al 1997 1999)
Figure 101 Map of Ethiopia showing Adi Ainawalid in Tigrai province
During the political upheavals in Ethiopia between 1973 and the early 1990s there waslarge-scale compulsory resettlement Individually owned and managed farmland wastaken and incorporated into large co-operatives By 1996 these had been dispersed land holdings had only recently been reallocated to individual farmers and traditional farmingpractices had been resumed However several families at Adi Ainawalid were spared theturmoil throughout the conflict they occupied their original family homes and they nowretain some land farmed by their grandparents
ENVIRONMENT AGRARIAN SYSTEMS AND CROPS IN THE NORTHERN HIGHLANDS
Subsistence is largely vegetarian and depends on the household production of graincrops supplemented by a few resources such as salt and oil bought from the regionalmarket in Mekelle Land holdings are based on units (tsumdi) representing the area of land that can be ploughed by one ox-team in a day which is estimated at a quarter of a hectare (Adebo 199348 Konde 1993 18) Individual holdings are very small rangingfrom one to eight tsumdi commonly consisting of at least two plots usually one adjacent
Farming and famine subsistence strategies in Highland Ethiopia 175
to the dwelling and the other(s) up to an hourrsquos walk away towards the perimeter of thevillage This is similar to the situation described for other areas in Ethiopia (Tsegaye1997) It was noted that some families in the kushets investigated are newly settled and have missed an allocation of land They have had either to rent land from a householdlacking the means to work it or to find an alternative to farming as a means of livelihoodPlot boundaries may be defined by stones shallow drainage channels or spinybrushwood
The houses are round or rectilinear built of local stone and with thatched earthen or wooden roofs A separate kitchen building is common as well as an enclosed area foranimals within the surrounding stone-walled compound (DrsquoAndrea et al 1997 Fig 102) Livestock is also kept inside houses especially donkeys horses and calves toprotect them from predation by hyenas Gardens are common in larger maturecompounds but are rarely found with small houses The soils are largely derived fromlimestones weathered to vertisols and cambisols which are clays and sandy clays and inthis region they are typically stony thin and eroded (Hunting Technical Services Ltd1973ndash4 Mitiku Haile pers comm) The natural vegetation in the region is described as Acacia savannah (Bekele-Tesemma 19936) but today there are few trees
The action of heavy rains and trampling by livestock on the treeless and uncultivated soils tend to cause surface crusting this restricts penetration by
Figure 102 Residential compound near fields west end of Adi Ainawalid facing southwest November 1997
Photograph CDrsquoAndrea
water and promotes run-off (Butzer 1981) Attempts are made to catch and retainrainwater in clay-lined artificial ponds which are used mainly for watering livestock and for washing The main anti-erosion strategy in the region is the use of soil-retentive
The archaeology of drylands 176
terracing low walls are constructed on the surrounding slopes using stones dug out fromlocal outcrops by the farmers during the slacker farming periods or in food-for-work programmes To further reduce erosion and conserve rainfall a tree-planting scheme has recently been undertaken on the slopes around Adi Ainawalid using native species ofgenera such as Acacia Mill and Erythrina L In the settlement area small stands ofEucalyptus species have been introduced mainly for shade and fuel There are occasionalcompounds with small mainly leguminous trees planted for shade and supplementarylivestock feed Until recently water for household use has had to be carried from theriver up to two hoursrsquo walk away often twice daily but at Adi Ainawalid three wells arenow available for use Irrigation is rare and confined to plots near the river which arerented for the cash-cropping of introduced crops such as tomatoes potatoes and maize
Livestock
In Tigrai oxen play a central role in the household economics (Bauer 1975 McCann199548ndash56) Their availability is essential to cultivation although donkeys and morerarely camels and mules also supply labour A 2-year-old ox takes a yearrsquos training and can give up to five yearsrsquo work (Spiess 1994) During the study period of those farmers questioned at Adi Ainawalid only about one third owned an ox which is a similar findingto other surveys in the region (eg FAO 1986) Animals are commonly loaned to makeup a ploughing team and for threshing when up to eight or more oxen may be used totrample the yield from a single harvest Also a man with a team will plough land forothers for a payment of half the harvested crop Occasionally donkeys mules or mixedteams may be used for ploughing
The number of animals is restricted by a shortage of feed This is most scarce just prior to the heaviest ploughing season thus the oxen tend to be undernourished and least fitwhen their labour is most in demand (Konde 199370) Cattle small ruminants equidsand the few camels graze field edges and stubble Feed crops are not cultivated nor island set aside for hay To conserve the pasture and reduce erosion from over-grazing the availability of the communal village grazing lands is carefully restricted to certain periodsand to particular animals mainly oxen and cattle By-products of cultivation such as weeds and crop-processing residues and food-preparation residues are the mostimportant feed resources the choicest of which are fed to the oxen Tree and shrubvegetation provides useful supplementary fodder
Dung is the most valuable animal by-product It is used primarily for fuel and also for fertilizers and as the raw material to make various household features and effects such ashouse floors storage-jar lids and pot stands Skins are an important source of income and are used for example to store honey and grain and to make baby carriers and as mats
Plant resources
The wide range of grain crops that is cultivated includes species of indigenous Africandomesticates members of the assemblage of southwest Asian founder crops and someintroductions from the New World (Table 101) Farmers possess much detailed local knowledge of the habit and environmental requirements of the different varieties of each
Farming and famine subsistence strategies in Highland Ethiopia 177
cultigen which is a skill widely recognized (Tsegaye 1997 Worede and Mekbib 1993)This rural knowledge has been a very important factor in the successful reinstatement oftraditional farming systems following nearly twenty years of agricultural changes
Each year the choice of crop and the selection of the particular variety or mixture of species are based primarily on an estimate of which is most likely to succeed best in aparticular field under the environmental conditions anticipated during the followingseason A range of traits in each taxon is desirable such as both early and late maturingvarieties (Gebremedin and Haile 1997) Also crop types are selected for variables suchas the colour of resulting food products and the baking quality and storage quality ofgrains are more important than the grain yield and the size (Haile 1995 Webb and vonBraun 1994) To extend the range of crop types available to individual farmers locallydeveloped populations of grains are exchanged and further varieties or species may bepurchased from the regional market These measures help to perpetuate recognized land-races (Worede and Mekbib 1993)
To intensify the yield produced from the small land-holdings and to spread risk mixtures of different species are inter-cropped At Adi Ainawalid a wheat and barley mixture (hanfetse) and mixtures of wheat species are common (Fig 103) In other areas mixtures such as pea and faba bean (ater-abie) or sorghum with chickpea are sownWhen possible the varieties of the crops are chosen for their synchroneity ofdevelopment for example a hanfetse of shahan wheat and burguda barley can be sown harvested and processed as a single crop The grains may then be treated as a singleresource and prepared for food as one or the constituent grain types may be separated inthe home When a single species is planted several varieties may be mixed Theproportion of different grains in the mixture at harvest differs from that sown so themixtures have to be reassembled each sowing season Following periods of drought cropfailures can result in a reduction in the number of crop species and varieties harvested aswell as the total yield
In the few house gardens chilli (berbere) garlic (tarsquoeda shigurtee) onions (shigurtee)basil (seseg) and other spice plants and herbs may be cultivated These provide important nutritional components and when available are always added to the staple carbohydratefoods As an example of the exotic drought-tolerant New World species that have been introduced prickly pear Opuntia ficus-indica Mill (beles) is commonly planted as hedging the leaves also being a valuable source of fodder and the fruits a human food
Wild plant resources are collected mainly for medicinal and other non-food uses For example grasses such as Hyperrhenia hirta (L) Stapf (sarsquori awald)
Table 101 Crops cultivated at Adi Ainawalid Tigrai Ethiopia Crop species Common name and varieties Local name Sorghum bicolor (L) Moench sorghum (5 varieties) mashella Eragrostis tef (Zucc) Trotter teff (red white) taffTriticum turgidum conv durum (Desf) MacKey
durum (black with hexaploid characters)
tselimoi
Triticum aestivum subsp vulgare (Vill) MacKey
bread wheat (shahan Canada wheat)
sindai
The archaeology of drylands 178
Figure 103 Intercropped bread and durum wheats near Mai Kayeh Tigrai November 1997
Photograph CDrsquoAndrea
Hrufa (Nees) Stapf and Eleusine floccifera (Forssk) Spreng (rigaha) are gathered to weave into baskets and the labiate Otostegia integnfolia Benth (chirsquoindogwee) has insecticidal properties the juices being smeared onto livestock to prevent damage to theirhides
Cultivation systems
In Ethiopia as a whole the environmental factor that has the greatest control over thefarming schedule is said to be rainfall (McCann 199528ndash31) which characteristically is bimodal The small spring rains (belg) and the main summer rains (kremt) support two cropping seasons However throughout the year there can be unpredictable rains
Hordeum vulgare L barley (burguda sarsquosaa) segemTriticum L spHordeum L sp wheatbarley intercrop mixture hanfetse Eleusine coracana (L) Gaertn finger millet (black red
white) dagousha
Lens culinaris sspculinaris Medikus lentil bersheem Lathy rus sativus L grasspea gwayya
seberoCicer arietinum L chickpea shimbra Zea mays L maize (arigo beraho) efoonTrigonella foenum-graecum L fenugreek abakaLinum usitatissimum L linseed indata
Farming and famine subsistence strategies in Highland Ethiopia 179
sometimes with heavy hailstorms damaging to crops Cyclical episodes of low rainfalloccur associated with the El Nintildeo Southerly Oscillation (ENSO) with frequenciesvarying between three and fifteen years (Bekele 1997 Wolde-Georgis 1997 and see Chapter 2) There is also evidence that for several decades the basic annual rainfall hasbeen decreasing Although a belg season still occurs in the central Highlands in the provinces of Wello and Shewa (Rahmato 199154) and even in parts of the northernHighlands (Emergencies Unit for Ethiopia 1997) over the past thirty years the totalannual rainfall in much of the northern province of Tigrai has been drastically reducedand the spring rains have virtually ceased (Emergencies Unit for Ethiopia 1994) Thissituation has largely restricted the cropping periods to a single season (meher) associated with the kremt (Adebo 199375 Konde 199379) Today despite an annual rainfall of between 450 and 900 mm (Butzer 1981) Tigrai is known as one of the most drought-prone regions of Ethiopia (Webb and von Braun 1994)
Thus in southeastern Tigrai the land is tilled for a single growing season The production from the small plots is optimized by measures such as inter-cropping double cropping the ploughing-up of headland and land rental so that at any one time the maximum area is cultivated Soil preparation begins in late winter or early spring withploughing The seed bed for cereals usually receives several ploughings but the plots forpulses may be ploughed only once Stones and tree stumps are retained in the soil toreduce erosion from wind and water Grain crops are usually broadcast sown betweenMay and Julymdasha period associated with the start of the big rains which are concentrated between June and September Cereal crops attract the priority of farming input but evenfor this crop category it appears minimal Cereal fields may be manured with dung fromgrazing ruminants although government supplies of chemical fertilizers are sometimesavailable their application is usually precluded by their expense Pulses are very rarelyfertilized Weeding also is rare but is more common for cereals Cereal plots may beploughed again to aerate the soil and facilitate drainage and reduce the weeds The mainharvesting season falls between October and December Crop plants are commonlyuprooted individually by hand (Fig 104) or they may be either uprooted or cut bysickle Weeds may also be uprooted and harvested separately for feed Unpalatable orspiny weeds remain standing in the fields Cultivation of the different crops is commonlystaggered across the growing season to increase the breadth of the harvest season therebypreventing an excessive concentration of labour and resources at one period Fingermillet sorghum maize and lentil are usually sown earlier than wheats barleys andgrasspea chickpea is often the latest crop (DrsquoAndrea et al 1997)
Grain separation takes place on threshing floors of compacted soil constructed at theedges of fields or within the settlement area as described in detail by DrsquoAndrea and others (1997 1999) The harvested crop is carried to the edge of the floor where it ispiled to dry or threshed immediately A threshing team of up to eight oxen is drivenaround the floor over the crop and the crop fractions are sorted with forks and brushes(Fig 105) Winnowing is a complex set of operations involving the use of several implements and it results in a cleaned pile of grain and the crop residues which arenormally amalgamated for feed (Fig 106) The threshing floor is swept clean ready for the next crop and the separated crop fractions are carried to the house
Attempts are made to maximize the yield of crop varieties that are prone to shatter by
The archaeology of drylands 180
harvesting them either when they are still slightly under-ripe or else early in the day prior to the highest temperatures Crop losses from the predation of birds and rodents areminimized by child scarers or scare-crows stationed in the fields Crop rotation is used tobreak cycles of plant parasites such as Striga asiatica (L) Kunze (selemi) on cereals particularly sorghum and Orobanche minor Sm (mrsquoandat tali) on legumes Barley and
Figure 104 Harvesting grasspea by hand uprooting Adi Ainawalid November 1996
Photograph AButler
tselimoi wheat are said to be the cereals most resistant to Striga Crop rotation usually consists of three to four years of cereals planted to every one of pulses Pulses especiallychickpeas and flax are planted to reinvigorate plots depleted of nutrients Because of theshortage of land fallowing is unusual A fallowed plot will often signal a shortage ofoxen or human labour (Adebo 199383 Konde 199380)
Farming and famine subsistence strategies in Highland Ethiopia 181
Individual families normally provide the labour to work their own land with assistance when required from the extended family or from near neighbours Sometimes labour willbe hired Ploughing is undertaken by men and older boys In cases where no male familymember is available to work the land it may be rented out in return for half the yield atharvest It was reported that although not unknown it was very rare for a woman toplough The whole family may be engaged in weeding and harvesting Men
Figure 105 First threshing of teff Adi Ainawalid November 1997
Photograph CDrsquoAndrea
The archaeology of drylands 182
Figure 106 Winnowing teff Adi Ainawalid November 1996 Photograph AButler
and children perform the threshing and winnowing stages of crop processing Women areconcerned with small-scale winnowing and fine grain cleaning within the household andthe preparation of food Importantly they also play a significant role in discussions on theannual farming schedule and on crop and seed selection
Grain storage
Storage is overseen by the women (Tsegaye 1997) Cleaned grain is stored inside thehouses in clay or bamboo vessels about 1 m tall and sealed with dung (Fig 107) Fumigants or insecticides are not added but it is believed that the dung acts as an insectrepellant Small crop yields may be kept in skin bags or sacks Unthreshed crops arestacked in the house compounds as are the threshing residues for animal feed Cats arekept to deter rodents
Farming and famine subsistence strategies in Highland Ethiopia 183
Figure 107 Grain storage jars Adi Ainawalid November 1996 Photograph AButler
The long-term storage of grains in clay-lined underground pits is said to be a practice in highland regions where the soil is dry The pits are concealed to minimize the loss ofgrain through plunder Sorghum is known to have survived such emergency storage for atleast five years (McCann 199567ndash8 Rahmato 199131 Worede and Mekbib 1993)However owing to its secret nature this storage system was not investigated during thisstudy
Farming and fuel
In the Ethiopian Highlands the paucity of trees is believed to be long-standing and is increasing (Staringhl 1993) Historical descriptions give varying accounts of the vegetation
The archaeology of drylands 184
of Tigrai and the nineteenth-century illustrations of Salt (1814) show moderate tree cover rather than dense woodland near Mekelle In the early 1900s it was estimated thatthere was about 40 per cent tree cover in the country as a whole but by the late 1980s thishad fallen to 56 per cent (FAO 1986) In Tigrai the demands for fuelwood and fortimber for manufacturing appear to have reduced the number of mature trees mainly tothe wooded conservation areas around churches or to single specimens used ascommunity assembly points This loss has been accelerating due partly to continuinghousehold demands and also to the need to supplement income outside farming the saleof firewood has been a traditional supplementary source of income until the recent pastNowadays in order to protect the remaining trees gathering timber for fuel is licensed(Derege Asefa pers comm) Wood continues to be the raw material for house supportsand farming implements such as plough beams yokes and winnowing forks Up to 55 percent of the fuel resources are provided by alternatives to timber (World Bank 1984) andare mainly farming by-products Dung is perhaps the most valued It accumulates inresidential compounds where livestock are penned overnight and is collected by childrenfrom the grazing areas it is then spread on walls to dry and be stacked The culms ofsorghum and other vegetable material are also important fuel resources
Thus fuel is sparingly used and dried grasses are the usual kindling For each cooking episode small fires are lighted individually within the stoves The latter are usuallypermanent fixtures of clay and stone constructed inside the kitchen building but smallportable stoves are also used (DrsquoAndrea et al 1999)
CROP DIVERSITY
Many Ethiopian crops are noted for an impressive diversity of form under environmentalconditions that were described by Vavilov (1935347) as relatively uniform within thehigh altitude This morphological diversity incorporates traits adapted to various stressconditions and it seems to be maintained by both environmental and human agenciesRecent studies of Ethiopian wheats have shown that crop varieties with deeply pigmentedblack and purple grains appear to be adapted to high altitudes (Tesemma 1991)Interestingly at Adi Ainawalid a type of black wheat (tselimoi) is grown which has been identified as a hybrid form of durum with hexaploid characteristics (Gordon Hillmanpers comm) Temperature and drought stress are important factors that affect variables such as plant height the protein content of grains and the timing of heading(Annicchiarico et al 1995) Many varieties have been developed that are pigmented and also of low height and early maturation these desirable traits confer resistance to fallingin wet or windy weather (lodging) (Bejiga et al 1996 Belay et al 1995) The selection of similar traits is seen in other crops such as barley (Demissie and Bjornstad 1996) andpea (Govorov 1930) Many of these crop types tend to be low-yielding (Tesemma 1991 Tsegaye 1997) and are officially regarded as having low industrial quality but becauseof their performance under potentially stressful environmental conditions they continue tobe selected by farmers (Belay et al 1995)
However while a general diversity of crop varieties appears to be continuallymaintained by farmersrsquo selection the cultivation of some species of food plants is
Farming and famine subsistence strategies in Highland Ethiopia 185
becoming restricted due to the changing pattern of climate and land shortages This isaffecting the long-term local availability of some crops
Following a succession of bad years there has been a long-term reduction in the range of grains planted in the study area Table 102 lists the crops that are no longer found at Adi Ainawalid or indeed in the tabia of Mahabere Genet as a whole However they arefamiliar to many farmers of Adi Ainawalid and they have been grown locally in the pastThese crops are still grown in areas of higher rainfall particularly further south in Tigraiand occasionally they can be found in the regional market in Mekelle As opposed to theblack wheat (tselimoi) mentioned above lsquoclassicrsquo durum wheat recently still said to beone of the major cereals in Ethiopia (Engels and Hawkes 1991) is an uncommon crop inthe study area one farmer in the kushet of Adi Akel immediately adjacent to AdiAinawalid had twice obtained some durum grain but this had produced only sterileplants All the rare grains are valued as traditional resources with special propertiesdurum makes a heavy solid loaf for sustaining field lunches at harvest-time emmer wheat is made into a nutritious and easily digested gruel for invalids and babies peas andfaba beans although expensive are still regularly bought in for meals on
Holy Days throughout the year The Ethiopian pea is particularly sought after and theaddition of even a few seeds to a festival legume dish is held to enrich the celebrations
CULTIVATION UNDER DROUGHT CONDITIONS
The annual precipitation in Tigrai is well within the levels generally considered to beenough to support dryland agriculture (Butzer 1981) At the same time however thisrainfall is acknowledged to be insufficient (Emergencies Unit for Ethiopia 1994 1997)The discrepancy can be explained by changes in pattern the expected rains of the springand summer monsoon seasons have diminished but the total annual rainfall may beaugmented to near-normal levels by unpredictable showers often of heavy hailstorms destructive to crops The opportunistic use of watered soil is a feature of Ethiopianhighland agriculture
The study periods between 1996 and 1997 fell within an episode of low precipitation
Table 102 Crops no longer cultivated at Adi Ainawalid Tigrai Ethiopia Crop species Common name Local
name Triticum turgidum spp diccocum (Schrank) Thell
Emmer wheat ares
Tturgidum conv durum (Desf) MacKey Durum wheat kinkinai Pisum sativum ssp abyssinicum ABr Ethiopian pea dekoko Psativum ssp sativum var arvense L Field pea aterVicia faba L Faba bean abiePisum sativum LVicia faba L Peafaba bean
intercrop aterabie
The archaeology of drylands 186
and it was possible to observe the effects of water-stress on farming Rainfall in the region of Mekelle was generally reduced scattered in its distribution and of spatiallydiffering amounts The reservoir at Adi Ainwalid was dry in both years although the oneshared by nearby kushets appeared to have retained some water In 1997 some farmers in adjacent settlements appeared to be producing yields of most grain crops yet the farmersof Adi Ainawalid although they had cultivated a restricted range of the most drought-tolerant crop species experienced severe crop failure with very reduced harvestsFollowing a very dry summer there were outbreaks of heavy rain in November thatcaused lodging and ruined many of the surviving small harvests At the adjacent kushet of Adi Akel where the soils appear to be deeper and more water-retentive harvests seemed to be less affected and following the unseasonal rains some farmers ploughed for extraend-of-season crops of barley and chickpeas which would have been ready forharvesting at the end of the following spring In other highland regions when the summerrains are especially heavy and flooding occurs double-cropping is common Short-season species such as grasspea can be sown and harvested on the semi-waterlogged fields prior to the main growing season on the drained land (Abate Tedla pers comm)
Rainfall above 200 mm is not officially classified as a drought (World Bank 1984) yetin Tigrai a chronic food shortage prevails Between 1988 and 1992 three-quarters of the families produced insufficient food (Holt and Lawrence 199326ndash31) and in 1997 the regional food production was deficient by about 20 per cent (Gebremedin and Haile1997) This is in contrast with previous times documentation from the sixteenth centuryfor example describes southern Tigrai as a land of great abundance of production withgreat yields of cereals and pulses (Alvares 152031) The current shortfall is thought tobe less a reflection of the demands of an increasing population than the result of the dual effects of the changing pattern of climate and the political circumstances (Pankhurst1992318 Zewde 1991195ndash6)
On the local level in 1996 crop production throughout Tigrai was officially reportedby the aid agencies as being poor (Ahrens and Spiess 1997) and Enderta was singled outas being in particular need of food aid (Emergencies Unit for Ethiopia 1997) At AdiAinawalid some supplies of grain had been distributed although towards the end of 1997food aid had been discontinued These food shortages prompted surveys in the Highlandsthat have highlighted a number of farming problems with priorities that vary slightly atdifferent seasons and in different areas lack of rain small land holdings insufficientoxen lack of human labour shortage of fuel high market prices shortage of livestockfeed livestock disease and unclean drinking water (Adebo 199370ndash89 Konde 199376ndash86) Most of these problems are familiar to the farmers of Adi Ainawalid
CONCLUDING REMARKS
The northern Highlands of Ethiopia are a region commonly regarded as being associatedwith both drought and famine It is now believed that drought alone seldom causesfamine more often a combination of factors is involved These include epidemics ofhuman and veterinary diseases such as smallpox and cattle rinderpest plagues on cropsof insect predators such as locusts ants and army worms and social and political
Farming and famine subsistence strategies in Highland Ethiopia 187
circumstances (McCann 199589ndash91) The recent situations of conflict appear to underliemany of the nutritional difficulties seen in the region today
Sustainable agriculture in Ethiopia is characterized both by maintained traditional practices and flexibility Farming as described prior to the revolution of the early 1970s(eg Westphal 1975 Simoons 1960) was abruptly and severely altered followingresettlement and lsquovillagizationrsquo until the early 1990s (Rahmato 1985 Rock 1994)Families were split up relocated in regions of unfamilar ecology and expected tocultivate alien (to them) crop assemblages on collective farms Now reinstated traditionalfarming is demonstrating in the selected crops and technology an essential stability thathas been successfully supported by long-term rural knowledge
The farming strategies devised and developed in the Ethiopian Highlands as witnessed during this study include mechanisms for survival during periods of climatic stress Cropgermplasm is carefully conserved to allow the best selection of crop types to suit theagricultural situation Within a wide repertoire agrarian systems are flexible andencorporate strategies to maximize production under whatever conditions pertainMechanisms have been developed to minimize erosion water is reserved at run-off soil-water is exploited grazing is controlled and alternative sources of fuel have been foundIt appears that periods of low rainfall and crop failure can be endured for several years bythe careful storage of grain during good years Social networks of exchange and the sharing of resources such as oxen are essential mechanisms to bridge periods ofshortage These strategies appear to be able to promote survival for at least two to threeyears of high aridity
When droughts last longer than a few years or when epidemics of disease or pests orbouts of conflict or political unrest are superimposed upon drought periods then itappears that the social and farming mechanisms described above may be insufficient toavert severe food shortage A better understanding of climatic perturbations could speedthe implementation of measures to lessen future threats of famine (Wolde-Georgis 1997)
ACKNOWLEDGEMENTS
The fieldwork was funded by the Social Science and Humanities Research Council ofCanada (Grant No 410ndash96ndash1520) and supported with the valuable assistance of Dr Mitiku Haile Dean of Mekelle University College (MUC) We are grateful to theCommittee for Research and Conservation of Cultural Heritage (CRCCH) Addis Ababaand the Tigrai Bureau of Culture Tourism and Information Mekelle for grantingpermission to undertake the study Field assistance was provided by Shewiaye BelayZelealem Tesfay Derege Asefa and Alemtsehay Tsegay of MUC Figure 101 was drawn by Shannon Wood of Simon Fraser University (SFU) Useful comments on the text byDiane Lyons (SFU) are acknowledged Our deepest thanks go to the kind generous andpatient farmers of Adi Ainawalid and neighbouring settlements in south-central Tigrai
The archaeology of drylands 188
REFERENCES
Adebo S (1993) Report of Diagnostic Survey of Debre Tabia in Enderta Wereda Addis Ababa FARMAfrica
Ahrens J and Spiess H (1997) Field Trip to Amhara and Tigray Regions 515ndash6196Situation Report Emergencies Unit for Ethiopia (UNDP-EUE) (httpwwwsasupenneduAfrican_StudiesEUEamhra696htm)
Alvares F (1520) The Prester John of the Indies Volume I Cambridge Cambridge University Press translation by CFBuckingham and GWBHuntingford (eds) 1961edition Hakluyut Society vols 114ndash15
Annicchiarico P Pecetti L and Damania AB (1995) Relationships betweenphenotype variation and climatic factors at collecting sites in durum wheat landracesHereditas 122163ndash7
Bauer DF (1975) For want of an oxhellip land capital and social stratification in Tigre InHGMarcus (ed) Proceedings of the First United States Conference on Ethiopian Studies 235ndash48 East Lansing MI Michigan State University African Studies Center
Bejiga G Tsegaye S Tullu A and Erskine W (1996) Quantitative evaluation ofEthiopian landraces of lentil (Lens culinaris) Genetic Resources and Crop Evolution43 293ndash301
Bekele F (1997) Ethiopian use of ENSO information in its seasonal forecast Internet Journal for African Studies 2
Bekele-Tesemma A (1993) Useful Trees and Shrubs for Ethiopia Nairobi Regional Soil Conservation Unit Swedish International Development Authority
Belay G Tesemma T Bechere E and Mitiku D (1995) Natural and human selectionfor purple-grain tetraploid wheats in the Ethiopian highlands Genetic Resources and Crop Evolution 42387ndash91
Butler EA (in press) Sustainable agriculture in a harsh environment an Ethiopianperspective In FHassan (ed) Drought Food and Culture Food Security in Africarsquos Later Prehistory New York Plenum Publishing Corporation
Butler EA Tesfay Z DrsquoAndrea AC and Lyons DE (1999) The ethnobotany of Lathyrus sativus L in Highland Ethiopia In M van der Veen (ed) The Exploitation of Plant Resources in Ancient Africa 123ndash36 New York Plenum Publishing Corporation
Butzer KW (1981) Rise and fall of Axum a geoarchaeological approach American Antiquity 46471ndash95
DrsquoAndrea AC Haile M Butler EA and Lyons DE (1997) Ethnoarchaeologicalresearch in the Ethiopian highlands Nyame Akuma 4719ndash26
DrsquoAndrea AC Lyons D Haile M and Butler A (1999) Ethnoarchaeologicalapproaches to the study of prehistoric agriculture in the Ethiopian highlands In M vander Veen (ed) The Exploitation of Plant Resources in Ancient Africa 101ndash22 New York Plenum Publishing Corporation
Demissie A and Bjornstad A (1996) Phenotypic diversity of Ethiopian barleys inrelation to geographical regions altitudinal range and agro-ecological zones as an aid
Farming and famine subsistence strategies in Highland Ethiopia 189
to germplasm collection and comservation strategy Hereditas 12417ndash29 Emergencies Unit for Ethiopia (1994) Situation Report for Region 1 (Tigray)
Emergencies Unit for Ethiopia (UN-EUE) (httpwwwsasupenneduAfrican_StudiesEUEtigray0494html)
Emergencies Unit for Ethiopia (1997) Field Trip to Amhara and Tigray National Regional States Emergencies Unit for Ethiopia (UN-EUE) Development Programme (httpwwwsasupenneduAfrican_StudiesEUEnorth0296html)
Engels JMM and Hawkes JG (1991) The Ethiopian gene centre and its geneticdiversity In JMMEngels JGHawkes and MWorede (eds) Plant Genetic Resources of Ethiopia 23ndash41 Cambridge Cambridge University Press
FAO (1986) Ethiopia Economic Analysis of Land Use Technical Report 8 Rome FAO Gebremedin B and Haile M (1997) Food Security and Dryland Agriculture the Case
of Tigray Utvikingsfundet (the Development Fund) (httpwwwu-fondetnoengelsktemakonf1-3html)
Govorov LI (1930) The peas of Abyssinia A contribution to the problem of the originof cultivated peas Essay II Bulletin of Applied Botany Genetics and Plant Breeding (Leningrad) 24399ndash431
Haile M (1995) Indigenous knowledge and agricultural practices in Central TigrayUnpublished paper presented at Rural Development Workshop Mekelle Tigray
Holt J and Lawrence M (1993) Making Ends Meet A Survey of the Food Economy of the Ethiopian North-East Highlands London Save the Children UK
Hunting Technical Services Ltd (1973ndash4) Tigray Rural Development Studies Map of Landforms in Mekelle District Gradients Soil Depth and Soil Types (1ndash6) London Ministry of Overseas Development
Konde A (1993) Report of Diagnostic Survey of Debre Medhanit Tabia in Dedebama Derga-Agen Wereda Addis Ababa FARMAfrica
McCann JC (1995) People of the Plow Madison Wisconsin University of Wisconsin Press
Pankhurst RA (1992) A Social History of Ethiopia Trenton New Jersey The Red Sea Press
Rahmato D (1985) Agrarian Reform in Ethiopia Trenton New Jersey The Red Sea Press
Rahmato D (1991) Famine and Social Strategies Uppsala Scandinavian Institute of African Studies
Rock MJ (1994) Famine and Food Insecurity in Ethiopia A Critical Assessment of theNotion of ldquoCoping Strategiesrdquo University of Leeds unpublished PhD thesis
Salt H (1814) A Voyage to Abyssinia and Travels in the Interior of that CountryLondon FC and JRivington
Simoons FJ (1960) Northwest Ethiopia Madison University of Wisconsin Press Spiess H (1994) Report on Drought Animals under Drought Conditions Emergencies
Unit for Ethiopia httpwwwsasupenneduAfrican_StudiesEUEdrought 0794html Staringhl M (1993) Foreward In ABekele-Tesemma (ed) Useful Trees and Shrubs for
Ethiopia vii Nairobi Regional Soil Conservation Unit Swedish International Development Authority
Tesemma T (1991) Improvement of indigenous durum wheat landraces in Ethiopia In
The archaeology of drylands 190
JMMEngels JGHawkes and MWorede (eds) Plant Genetic Resources of Ethiopia288ndash95 Cambridge Cambridge University Press
Tsegaye B (1997) The significance of biodiversity for sustaining agricultural productionand role of women in the traditional sector Agriculture Ecosystems and Environment62215ndash27
Vavilov NI (1935) The phyto-geographical basis of plant breeding Theoretical Bases for Plant Breeding Moscow 1 Reprinted and translated in DLoumlve (1992) Origin and Geography of Cultivated Plants 316ndash66 Cambridge Cambridge University Press
Webb P and von Braun J (1994) Famine and Food Security in Ethiopia Chichester John Wiley and Sons
Westphal E (1975) Agricultural Systems in Ethiopia Wageningen Centre for Agricultural Publishing and Documentation (PUDOC)
Wolde-Georgis T (1997) El Nintildeo and drought early warning in Ethiopia Internet Journal for African Studies 2
Worede M and Hailu Mekbib H (1993) Linking genetic resource conservation tofarmers in Ethiopia In Wde Boef Kojo Amanor and KWellard (eds) Cultivating Knowledge 78ndash84 London Intermediate Technology Publications
World Bank (1984) Ethiopia Recent Economic Developments and Future ProspectsWashington DC World Bank
Zewde B (1991) A History of Modem Ethiopia 1855ndash1974 London James Curry
Farming and famine subsistence strategies in Highland Ethiopia 191
Part IV EASTERN AND SOUTHERN
AFRICA
11 Engaruka farming by irrigation in Maasailand
cAD 1400ndash1700 JOHN EGSUTTON
INTRODUCTION THE RIFT VALLEY AND CRATER HIGHLANDS OF NORTHERN TANZANIA
The equatorial highlands of East Africa are bisected by the north-south trough of the Rift Valley They contain marked variations in altitude precipitation and vegetation as wellas in their exploitation in recent centuries by hunters herders and cultivators Thecontrasts are especially sudden and striking at Engaruka situated at the foot of the east-facing Rift wall at three degrees south (Fig 111) At an altitude of 1000 m (which is low for this interior region) and with unreliable and variable rainfall estimated at not morethan 400 mm in an average year it is a relatively hot dry and dusty place with highevapotranspiration Despite the attraction of a permanent supply of clear water in theEngaruka river no cultivators would ever have contemplated settling here by relying onthe rain alone for their crops
Immediately behind Engaruka the escarpment rises to 2000 m above sea level and towering above that are the Crater Highlands with the wide dome of Lolmalasin reachingto some 3500 m These highlands catch two or three times the rainfall of the Rift floortheir vegetation ranges from montane forests to open grasslands The latter have in recenttimes supported wild herbivores and pastoral communities with cattle sheep and goatsSince the seventeenth or eighteenth centuries AD these pastoralists have been MaasaiHowever the history of pastoralism here stretches back some 3000 years during whichtime successive groups of which the Maasai are the most recent have replaced orassimilated those who preceded them (Sutton 1993) Although these cool highlands havenot attracted agricultural settlement the run-off which descends the escarpment indeeply cut gorges has been essential for that of Engaruka at its foot
These gorges of different sizes spectacularly incise the escarpment face at Engaruka along a stretch of 9 km (Figs 112 and 113) Nowadays only one of them carries waterpermanently and this is the Engaruka river itself (no 2 on Figure 112) This is a fast but shallow stream usually 3ndash4 m wide
Figure 111 The Rift Valley and Crater Highlands of northern Tanzania showing Engaruka and related sites
The archaeology of drylands 196
Figure 112 Engaruka and the Rift Valley escarpment showing the main river (2) and seasonal streams (1 4 and 5) the area of ancient fields as surveyed (stippled) the artery canals as traced (broken lines) and villages (black circles)
Engaruka farming by irrigation in Maasailand cAD 1400ndash1700 197
Figure 113 The Engaruka escarpment from the east with the gorge of Engaruka river (central)
Photograph JEGSutton
as it descends the rocky scree it is easily fordable except during spates following stormson and behind the escarpment With its speed compensating for its small dimensions itsdischarge into the plain is considerable The other streams (nos 1 4 and 5) flowseasonally or in the case of certain escarpment gullies and clefts (notably no 3) openvery occasionally after exceptional rain Of the seasonal streams the Makuyuni (no 4) onthe northern side is the most reliable in good years flowing for six to nine months andvery occasionally lasting throughout That at the south end Olemelepo (no 1) may carrynearly as much water overall but is extremely temperamental liable to open in spate andthen to fail equally suddenly
The effects of extreme spates occurring at intervals over many millennia (of the Pleistocene presumably as well as the Holocene) are clear from the sizes of the outwashfans These consist of soil mixed with water-worn lava boulders of all sizes which haveaccumulated immediately below the points where the gorges of the seasonal as well asthe main river open onto the escarpment foot As a result the streams enter the plain at acommanding level on the crests of these fans the land falling away not only towards theRift floor but also on either side of the stream beds This situation continues in the caseof the main river for a distance of nearly 2 km downstream of the gorge
Not surprisingly this source of permanent water with its rapid descent andadvantageous level has been exploited for irrigated agriculture but at two separateperiods The second of these persists the present community of Engaruka continues toexpand and enlarge its cultivation area This dates from the 1890s when a few farmersfrom different parts of what was then German East Africa became established a short
The archaeology of drylands 198
distance down the river (that is largely below the area of the archaeological field remainsof the earlier period) and began cultivating the soft soil with the help of furrows takenoff both banks This community has been reinvigorated by new settlers on occasions inthe 1920s and 1940s during the British mandateship of Tanganyika and again in the1970s within the Tanzanian lsquovillagizationrsquo (ujamaa) movement The latter involved the incorporation of numbers of Maasai who previously herded in the surrounding plain sothat the character of what used to be called the lsquoSwahilirsquo village of Engaruka has altered In a series of good years some of these farmers cultivate in the Makuyuni basin (on thenorth side) too by relying on a combination of rainfall and water furrowed from thatstream (no 4) while its flow lasts Others farm in the Olemelepo basin (to the south) byusing a long cross-valley furrow (following close to the line of an ancient one) taken offthe Engaruka gorge
Before the 1890s however Engaruka was according to available reports desertedexcept for some pastoral Maasai whose cattle and goats grazed and browsed the sparsepasture of the Rift floor within reach of the river that being the only permanent water inthe district Information about previous inhabitants gleaned from local Maasai early in thetwentieth century is vague and is probably not genuine tradition so much as guessesoffered in response to direct questions about the lsquoruinsrsquo (for discussion see Fosbrooke 1938 Sassoon 196680ndash81 Sutton 197867ndash68) This negative reaction indicates thatthe place was deserted before the nineteenth century at latest The recent and existingcultivating community does not appear to be descended in any way from the earlierirrigation farmers who lived here between approximately the fifteenth and theseventeenth centuries The evidence for that settlement and its fields is exclusivelyarchaeological
THE ANCIENT FIELDS AND IRRIGATION SYSTEM
These earlier fields and irrigation worksmdashwhich cover some 2000 ha at the base of theescarpment around the foothills and into the plain (Fig 112)mdashare distinguished from the modern ones by their use of stone for dividing and levelling the plots for irrigation(Figs 114 and 115) by means of revetments and mild terracing and for lining andembanking the artery canals (Figs 116 and 117) and feeder furrows Equally distinctiveamong these ancient fields are two other types of stone features The first of theseconsists of numerous square or angular cairns standing up to 2 m high with rubble coresretained by drystone casings of larger boulders (Fig 118) They are interpreted as stone clearance devices necessitated by the thinness of soil and abundance of surface stonethese conditions being doubtless exacerbated by intensive cultivation with irrigation overa considerable period Secondly there are round enclosures up to 10 m across consistingof thick stone walls
Engaruka farming by irrigation in Maasailand cAD 1400ndash1700 199
Figure 114 Engaruka south fields stone field divisions and feeder furrow
Photograph JEGSutton
Figure 115 Grid of feeder furrows and levelled field plots below the intermediate north gorge (no 3 on Figure 112)
Photograph JEGSutton
The archaeology of drylands 200
Figure 116 The support for the lsquogreat northernrsquo canal running along the escarpment foot
Photograph JEGSutton
Engaruka farming by irrigation in Maasailand cAD 1400ndash1700 201
Figure 117 The embanked causeway of the lsquogreat northernrsquo canal
Photograph JEGSutton
of similar construction faced both outside and in with a narrow entrance gap Mostprobably these were built not to contain houses but for cattle The latter would have beenvalued for providing manure for the fields as well as milk and meat and would haveneeded to be stall-fed owing to the intense cultivation all around and the lack of pasturein the vicinity for much of the year
This suggestion of stall feeding and manuring at Engaruka is deduced from the evidence for cattle keeping obtained when excavating rubbish deposits in the villages
The archaeology of drylands 202
(Thorp 1986)mdashthese villages being situated on the escarpment scree above the levelattainable by channelled watermdashconsidered alongside examples of certain recent and existing compact and integrated agricultural
Figure 118 An angular cairn (later colonized by termites) in the fields on the south side of Engaruka river its rubble core revealed by breakage in the faced casing (on right)
Photograph JEGSutton
systems in Africa (Sutton 1986 1989) Of relevance also is the archaeological exampleof the Nyanga terraced fields and connected stone-walled farmsteads with sunken stock-
Engaruka farming by irrigation in Maasailand cAD 1400ndash1700 203
pens in Zimbabwe (Sutton 1988 and Soper this volume Chapter 12) Similarly information about the crops that were cultivated on the Engaruka fields is in large partdeduced from ethnographic examples of communities currently cultivating in some caseswith irrigation at comparable altitudes with medium to low rainfall in this interior regionof East Africa
Of particular value in this exercise are the Sonjo villages and irrigated basins 100 kmor so to the north close to the TanzaniamdashKenya border Here sorghummdashan ancient African grain and the principal crop of the savanna regions across the continentthroughout the Iron Agemdashmaintains (despite the progressive popularity of maize in thetwentieth century) its dominant position with varieties selected and developed locallyboth for withstanding droughts and for tolerating waterlogging and irrigation (Adams et al 1994) Confirmation that sorghum was grown at Engaruka is attested from excavations in the villages where charred seeds from the hearths and granaries have beenrecovered Other crops suggested by the examples of Sonjo and drier parts of the RiftValley generally would be finger-millet (eleusine) and varieties of pulses The latter in rotation with grains can provide valuable nutrition both for the soil and for the farmingcommunity Finger-millet while less tolerant than sorghum of heavy irrigation (and perhaps less productive in grain harvested per hectare) has the advantage of ripening ona low rainfall and also of long storage qualities Probably it would have been sown totake advantage of the rains in the main and if the crop were successful stored in theroofs or homestead granaries as reserve against a famine year It would also have beenvalued for beer as would surplus sorghum too doubtless improved by honey obtained inthe forests above Engaruka
Much but not all of Engarukarsquos ancient field area lies closer to the escarpment thanthat now settled and cultivated so that it survives most unusually as an expanse offossilized fields and irrigation devices Despite the effects of subsequent erosion inplaces with gullies damaging and destroying features on cutting through the loose andstony soil the upper part of the field system is preserved in a remarkably pristine stateVisibility depends on the season and the amount of grass and greenery on the trees andbushes Moreover there has been a noticeable increase of thornbush over the last fortyyears so that certain photographsmdashnotably Figure 115 taken in 1971mdashcannot be repeated This vegetational change apparently relates to a reduction of grazing by variouswild herbivores and also by Maasai cattle as well as to a cessation of burning as theagricultural population of the new Engaruka villages has increased
The lower part of the old field area has been subject to the opposite experiencemdashthat of redeposition of soil eroded from the upper partmdashso that the stone features there tend tobe obscured But the typical field divisions can be seen in gully sides and the irrigationgrid pattern is very clear from air photographs taken in the 1960s before the recentexpanse of bush It is on this relatively level soil with less surface stones that the presentinhabitants have chosen to cultivate avoiding not surprisingly the stone-strewn terrain much of it bereft of soil closer to the escarpment and on the outwash fan
Stonework was used for dividing and terracing the fields for lining the canals and feeder furrows for the stone clearance devices and stock enclosures in the fields and alsofor terracing and revetting the numerous homestead platforms in seven large villages that are situated immediately above the top canals (Fig 112) The sheer density of these
The archaeology of drylands 204
remains over such a wide extent combined with complete abandonment of this system atleast two more probably three centuries ago makes old Engaruka unique as anarchaeological field system and one that can be mapped and studied on the ground (Formore detailed description see Sutton 1998 for discussion of particular features seeSutton 1978 and 1986 the villages and excavations undertaken in them are furtherdescribed by Sassoon 1966 and 1967 and by Robertshaw 1986) This intense use ofstone which in older ethno-historical literature of eastern Africa (such as Murdock 1959) was labelled lsquomegalithicrsquo is as explained partly attributable to the ubiquity of the surface gravel and boulders that needed to be moved if one was to put the land to any useThe obvious solution was to utilize these stones in the field divisions and terraces and inthe furrow and canal sides with any remaining excess being piled in the enclosure wallsand especially the cairns which were built as neatly and vertically as possible in order tominimize the waste of cultivable ground As pressure on resources of both soil and waterincreased in this isolated situationmdashone circumscribed by the limits to which irrigated water could be led by gravitymdashit appears that this commitment to stonework (which is explicable in the first place in functional and environmental terms) developed into acultural attachment if not a hallmark of the old Engaruka community
However before becoming unduly enthralled by the stone lsquoruinsrsquo of Engaruka and the accidents of survivalmdashlsquothe tyranny of the monumentsrsquo in Ian Farringtonrsquos phrasemdashit is encumbent to consider Engaruka in its regional ethnographic and historical contextThere is in fact nothing very unusual about irrigated agriculture with lsquoindigenous rootsrsquo in the precolonial past along the western wall of the Rift in northern Tanzania and Kenyathere are examples from four degrees south through the celebrated instances of Sonjo(Adams et al 1994) Baringo (Anderson 1989) and Marakwet (Hennings 1951 Soper 1983 Watson et al 1998) to two degrees north or again to the east of the Rift in the highlands of northeastern Tanzania and the Kenya border notably Pare TaitaKilimanjaro and Mount Meru (For a survey of these remains see Sutton 1973 19841989 Widgren and Sutton 1999 and see also Widgren this volume Chapter 14) Among these communities numerous varieties of field systems are found prepared fordifferent crops and combinations and depending in greater or lesser measure on artificialirrigation of the lsquohill furrowrsquo sort (Adams 1989) that is by constructing small gravity-fed canals off springs or mountain streams But since these present and recent fields areobliged for obvious reasons (the water sources and the basic hydraulics of gravity-fed furrows) to use much the same irrigable land as the older ones the latter are notrecognizable as such on the landscape Or rather where there is a strong suggestion ofcontinuity of settlement and of cultivation dependent on irrigation over a long period onemay as a historian have to be content with regarding the ancient and the existing fieldsas all one At best therefore in these favoured areas of concentrated agriculturalsettlement one rarely gains anything more than an impression of the cultivation andirrigation system that operated in the past and one cannot discern the ancient fields asphysical units
Engarukamdashfor the dual reasons of its being a deserted site and a conspicuous onebecause of its stoneworkmdashis different therefore and extremely valuable as a researchresource being (with a few minor related sites in the district) a rare example of anarchaeological field and irrigation system that can be studied directly on the ground It
Engaruka farming by irrigation in Maasailand cAD 1400ndash1700 205
differs from many of the existing East African examples moreover in the degree of itsdependence on irrigation Few if any of those cited are in quite so arid terrain and somelike those on the southern and eastern slopes of Kilimanjaro and other mountains receivea high rainfall adequate to support forest if cultivation is restricted In a number of thesecases therefore a fair amount of cultivation is possible without any artificial irrigationand in places the latter option may be barely activated in average years necessary thoughit may be for survival through droughts and bad runs But the more important point is thatirrigationmdashor the ability to turn to itmdashhas become an essential element in theseagricultural systems because of the success of the latter over time and the size to whichthe communities have grown This has necessitated more production per hectare than isafforded by the rain alone (at least in years of low rainfall) and therefore the extension ofplanting assisted by irrigation devices into the dry season and the adoption of specialcrops varieties or combinations to suit the complex regime In this way devices that mayat first have been considered optional or supplementary in difficult years would in timehave become permanent and essential complements to developing agricultural systemsand the communities dependent on them
A further factor in the nineteenth century at particular favoured locations adjoining dryplains such as Taveta South Pare and Baringo was the supplying of trading caravansThis required the production of more than a normal surplus or at least facilities forgrowing a fast second crop to restock the granaries Although trade-routes and transport methods have changed in the twentieth century new opportunities have arisen forproduction for local and more distant markets On the slopes of Kilimanjaro and the hillsof North Pare for instance the production on smallholdings of coffee for exportinterplanting it with subsistence food crops has encouraged farmers to maintain theirrigation systems to ensure watering around the year And when coffee prices are low achannelled water supply is still valued for domestic needs in a highly populated ruralarea with stalled cattle kept by these means in some locations
Besides its excessively stony terrain it is doubtless Engarukarsquos extreme situationmdashan impossible one in fact in the eyes of most cultivatorsmdashthat explains the exquisite layout and detail of its fields and irrigation system over so wide an area Moreover as arguedbelow this situation is demonstrably harsher now than it was when Engaruka was firstinhabited which further explains why most of the ancient field area has remaineduntouched since it was abandoned The first settlers doubtless began on a small scale withrudimentary irrigation works probably on the easier terrain some way downstream of thegorge In time however as the community increased in numbers on the success of acultivation system evolved to handle the peculiarities of the location it would have beenobliged to expand its cultivation area into the more broken and stony ground closer to theescarpment and eventually as high onto the scree as could be reached by waterchannelled from the gorges of the Engaruka river and the seasonal streams
The top canals were accordingly led from the highest practicable points in the gorges that is at the vertices of the outwashes and carried along the rocky escarpment base at themaximum level attainable while permitting a gravity-induced flow The actual take-off works on the sides of the stream beds have of course not survived the spates since thetime of their abandonment but one must imagine improvised structures of stone andtrash as in existing irrigation systems requiring annual maintenance if not complete
The archaeology of drylands 206
rebuilding The positions of these take-offs can be estimated fairly accurately by tracingthe visible upper stretches of the artery canals back to source The stone linings andembanking of these upper artery canals are preserved quite spectacularly in somestretches along the base of the escarpment scree (Fig 116) around the small hills on the edge of the plain and through the whole area of ancient field remains (Fig 112) Every effort was made to keep these as close to the horizontal as was practicable for purposesof controlling the flow and preventing undue scouring of the canal beds and breaches inthe furrow walls and equally important to increase the area of the plain and of the sidesof hills standing in it which could be reached by the furrowed water Despite theunavoidable rapid descent or small cascade here and there on the steep and rockyescarpment for most of their lengths these canals fall at angles less steep than 120 and inplaces as gently as 1100 The existing examples of Marakwet and Sonjo illustrate howsuch engineering and levelling perfection can be achieved through a combination ofexperience and trial-and-error
The longest of these artery canals is that running northwards from the gorge of the Engaruka river it measures 1ndash2 m wide between its stone edges although since therewas no laid bottom the actual water flow over the gravel and silt bed would doubtlesshave been narrower It is traceable up to 3 km from take-off with divisions at various points Along the canalrsquos upper stretch as it descends the escarpment (at a relatively steepangle between 115 and 120 Adams 1986) its lower side is substantially supported bygravel embanking When it reaches the foot of the escarpment scree it swings at a rightangle towards a hill standing separately in the plain But in order to maintain as muchheight as possible the canal is carried across this narrow valley on an embankedcauseway (an aqueduct in effect) up to 3 m high (Fig 117) By these means it achieves an advantage when it reaches the hill where it divides to run as contour furrowsconstructed round each side It appears from the levels of the latter that the effect of theembanking was augmented by a wooden scaffolding device to carry the water higher still on hollowed logs
There are other instances of stretches of embanked canal in the field system with suggestions of wooden superstructures or at least split and hollowed logs to carry thewater over the porous gravel These belong to an evolved stage when the highest arterycanals were constructed and clearly represent an effort to gain the maximum advantagefrom the available water regardless of the correspondingly intensive demands this placedon the hydraulic ingenuity of the community and the sheer labour required forconstruction and constant maintenance of the works It appears moreover that largeareas of the fields were relaid to accord with these long embanked canals this isindicated where series of older field divisions and feeder furrows are superimposed by orincorporated into later grids with variant alignments
AN INTEGRATED AND CIRCUMSCRIBED SYSTEM UNDER STRAIN
At some point the limits of feasible improvements and of the communityrsquos technological resource would have been reached Since it was not possible to increase the area ofreliable cultivation beyond that to which water could be carried by gravity through the
Engaruka farming by irrigation in Maasailand cAD 1400ndash1700 207
highest and longest canals a crisis must have been faced as the population attained themaximum that this finite amount of land and water could feed despite all the complexityand ingenuity of the irrigation devices and other specialized elements of this integratedsystem In fact these doubtless together with the operation of the most productive croprotations available (sorghum and varieties of pulses in particular) combined withratooning of the sorghum to obtain a supplementary harvest and the application of cattlemanure to maintain fertility and improve yields may be seen more as reactions to thelimitations of the situation rather than methods devised to achieve agricultural efficiencyand increased productivity for their own sake Equally likely in so intricate andspecialized a set of arrangements there would have been a danger of trying to intensifytoo far in reaction to stress in particular shortening the fallow would have exacerbatedsoil-exhaustion and erosion Indeed despite all the effort of levelling and terracing to counteract these tendencies the upper fields became denuded through heavy wateringthe field divisions and furrows standing remarkably prominently here while as notedthe lower ones are overlaid with redeposited soil with the old field lines there beingvisible only in recent gully sides
At the same time this population was having to contend with hydrological declinewith less water flowing off the escarpment This process is strikingly illustrated by theexistence of canals leading off the gorges of the seasonal and occasional streams whoseflows are now far too inadequate to reward such labour It is not necessary to concludefrom this that those streams were perennial at the time when the canals were constructed together with the laying out of grids of levelled fields irrigated from them It is clearnevertheless that they must have enjoyed longer flows than now with sufficient volumesof water in their catchments to ensure their persisting some time after a period of rainThis argument applies especially to Olemelepo (no 1 on Figure 112) where any attempt to reopen the canals that led from that gorge and to reactivate the archaeological fieldsthat cover its outwash fan would be pointless now Even more striking is the case of theintermediate north gorge (no 3)mdasha narrow cleft in the escarpment from which waterissues in occasional years and then only for a few days following exceptional stormswhen irrigation would be least needed But at the time when the ancient settlementflourished it was found worthwhile to construct short canals along the foot of theescarpment on either side of this gorge to irrigate a grid of fields (Fig 115) on this small outwash (being too high for watering from the long northbound canal led through thevalley below from the main river) Immediately above those canals were built the twonorth-most villagesmdasha further indication that there must have been a natural flow fromthis cleft for at least a few months of the year
Similarly at Makuyuni (no 4) the positions of the top canals and the large area of fields on the outwash served by these seem to require more water and a longer season ofreliable flow than obtains now By implication too the volume of the Engaruka riveritself would have been greater then so that when the declining trend set in it may havebecome insufficient to irrigate adequately the whole basin dependent on its canalsAlternatively as the performances of the seasonal streams and of the fields on theiroutwash fans became increasingly unreliable the need would have arisen to channelwater as broadly as possible from both sides of the main river This was effected throughthe long cross-valley canals with take-offs at the gorge opening that were designed to
The archaeology of drylands 208
deliver water from the main river into the middle and lower parts of the Olemelepo andMakuyuni basins at seasons when those rivers had driedmdashprojects that required relaying of feeder furrows and field grids as already noted From the plan (Fig 112) these alternative sources of water and routes for channelling it may lend an impression ofwonderful flexibility In practice however there would usually have been little choicethe complex arrangements being dictated by the sheer necessity of carrying water to aslarge an area of fields as could possibly be reached from the main river In years of lowrainfall and therefore heavy dependence on irrigation this may have exhausted the wholevolume of the riverrsquos flow at certain seasons (as can happen nowadays although the areaof existing cultivation downriver is not as extensive as that of the ancient settlements attheir prime)
This hydrological decline must have been relative because had Engaruka been much wetter at the time of the first settlement (about the fourteenth century apparently) theneed to irrigate or at least to devise such elaborate arrangements would not have arisenThat notwithstanding the archaeological evidencemdashthe configuration of fields canals and villagesmdashdemonstrates clearly enough a change in the performances of theescarpment streams with their discharges now being definitely less than they were 500years ago Equally clearly these changes occurred or at least began during the life of theold settlementmdashthat is by the seventeenth century at latestmdashsuggesting that together with the strains imposed by the circumscribed situation they constituted the main factorin the collapse and desertion of Engaruka no later than the eighteenth century The datingis not perfectly precise being based on a number of radiocarbon results from excavationsin the villages on the generally late iron age affinities of the pottery and other artefactsas well as on the lack of any clear local memory about the former inhabitants (Sutton1998)
It is presumed that the decline in the flows of the escarpment streamsmdashso crucial at Engarukamdashwas due in large part to a drier climatic trend in the region at large That implies that there would have been less rain falling at Engaruka itself as well as in thehighlands behind it where the escarpment streams rose with the effect that dependenceon the latter would have been increasing just as their flows were declining Any suchclimatic trend ought to be detectable in the archaeological and geomorphological recordof the broader region and especially in lake deposits although no consistent body ofevidence can be cited at the present stage of research In the broader region of thehighlands and Rift Valley in Kenya as well as northern Tanzania there are individualinstances of springs that are now dry and of late iron age settlements in marginal areasthat are now deserted but these have not yet been dated precisely enough forchronological comparison Nevertheless the essential question and the approximatedating are inescapably posed by the Engaruka experience
The further question is whether the decline in the escarpment streams may have been due in greater or lesser measure to very local factors in particular environmentaldamage caused by a cultivating community of several thousand people over a period ofthree or more centuries Wood requirementsmdashfor fuel for fencing the villages and forhouse building and equally for the scaffolding and hollowed logs employed for carryingstretches of the canalsmdashwould have placed substantial demands on the forest resources on and above the escarpment Arguably such deforestation could have affected surface
Engaruka farming by irrigation in Maasailand cAD 1400ndash1700 209
moisture and the aquifers in the mountains and therefore the flows of the escarpmentstreams rendering them more liable to sudden spates and equally sudden failureHowever the scale of the change suggests that a human factor of this sort can be onlypartly responsible and that a decrease in rainfall must have occurred beginning in aboutthe sixteenth century The size of that decrease in the highlands need not necessarily havebeen so substantial but merely enough to diminish significantly the normal discharges ofthe streams descending the escarpment and their reliability for regular irrigation at thebottom
Whatever the cause or combination of factors Engaruka was comprehensivelyabandoned at some point around AD 1700 or perhaps somewhat after on the (admittedly imprecise) dating indicators available The state of the abandoned field and irrigationsystem especially its upper part and of the numerous homestead platforms in the sevenvillages immediately above the top canals lends an impression of sudden desertion ratherthan slow decline and piecemeal abandonment That is not easily testable and theimpression may be illusory but it invites one to speculate on other causes of desertioneven catastrophic ones The possibilities are legion Among those that have beensuggested are a violent earthquake along the Rift fault arguably upsetting the flows ofthe mountain streams and the canal take-offs an unusually heavy eruption of the nearbyvolcano Oldonyo Lengai coating the fields with sulphurous ash or a devastating attackby expanding Maasai pastoralists (or by Tatoga before them) anxious to secure the waterof the Engaruka river and the adjacent grazing There is no direct evidence to support anyof these speculations and the last would appear unlikely since any pastoral group in thearea would have been outnumbered and would also have benefited from exchange ofproducts with an agricultural community in its midstmdashassuming that the cattle that the latter kept for essential manuring as well as milking did not provoke insuperable jealousy
More likely the central cause of the collapse of old Engaruka and its highly specializedand integrated irrigation agriculture was inherent in that system which being physicallycircumscribed by the lie of the land and the volume of water in the escarpment streams(the latter moreover declining) could not in the long run cope with the strains itinevitably generated in particular its own demographic success While this generalexplanation does not rule out other possible contributory factors it is supported by theexistence of several lesser sites in the district situated similarly by streams (or springs)issuing from the escarpment base with stone-lined canals and field divisions identical tothose of Engaruka (see Figure 111) These obviously belonged to the same cultural group and ethnicity and were presumably abandoned about the same time as was themain settlement and cultivated area at Engaruka (although it is quite possible that some ofthese outlying sites on the far side of the Crater Highlands by Lake Eyasi and in thenortherly direction above Lake Natron may have been abandoned earlier or contrarilyhave lingered on a little later) Whatever the exact chronology this general phenomenonof abandonment of settlements over a radius of some 60 km suggests an inherent andunderlying factor rather than a catastrophic event
The fact of desertion in whatever manner it is to be explained should not however be interpreted as failure in an historical sense A system so accomplished and complex asthat of Engaruka which evolved over two or three centuries was surely a story ofsuccess in adjusting so effectively to the peculiarities of its own special environment It is
The archaeology of drylands 210
more important to understand how it worked and succeeded than to worry about why itcollapsed eventually or again (the common antiquarian reaction to stone ruins) wherethat population lsquowentrsquo The answer to that last question is that the degree of specializationand the various details of the settlement and its agricultural system had become so specific culturally as well as functionally to the situation and community of Engarukathat they could not be transplanted in other words this ethnicity would have expired asthose villages and their fields had to be abandoned Remnants presumably took refugeand became assimilated among other peoples of the region thereby losing their Engarukaidentity
One of these may have been Sonjomdasha group of compact villages each situated by a spring or river above Lake Natron 100 km or so to the north (Adams et al 1994) It thus forms an lsquoislandrsquo of Bantu-speaking cultivators surrounded by Maasai pastures the latterconsisting of poor scrubland in the Rift to the east but also the extensive plateaugrasslands of Serengeti to the west Each of the main existing Sonjo villagesmdashas well as some that have been abandoned for a whilemdashhas relied on a basin of irrigated fields Sorghum has been the principal crop together with some finger-millet and distinctive varieties of beans There is also a fair amount of rainfed cultivation in most years withthe normal rainfall being slightly higher than at Engaruka so that Sonjorsquos dependence on its irrigation works is not as extreme as was that of Engaruka Cattle manure is notapplied as fertilizer to the fields and Sonjo as far back as reliable information goes havenot kept cattle for fear it is said of provoking neighbouring Maasai to raid (This is tooverlook some experimentsmdashand mixed experiences in contending with both cattlediseases and neighbouring Maasaimdashin the 1970s and 1980s) There is moreover littleuse of stonework in the field divisions and canal banks as is so distinctive at Engarukathe basic reason being the relative paucity of surface stone in the alluvial soils of theSonjo basins
Stone is however used extensively in the concentrated villages of Sonjo which are situated on hills above the fields in particular for terracing and revetting the homesteadplatforms (Fig 119) and also for public areas and some rather peculiar open-air fireplaces Fairly close parallels for these have been revealed by excavations in theEngaruka villages (Fosbrooke 1938 Sassoon 1967) Pending excavations of old Sonjosites it is not known how far back these features may be dated there but it seems likelythat some of the Sonjo villages were settled before the collapse of Engaruka PossiblySonjo and Engaruka constituted the northern and southern wings respectively of a singlecultural group of which only the one has persisted to the present In that case a detailedarchaeological study of Sonjo should be revealing about the regional history This studyshould compare the existing Sonjo settlements that is the lsquotraditionalrsquo ones destroyed in 1975 during the Tanzanian governmentrsquos lsquovillagizationrsquo (ujamaa) campaign and those deserted at unspecified dates in the nineteenth or preceding centuries Such an exerciseshould be expected to carry the sequence back towards the time of Engaruka and thusillustrate the connection Even if it transpires that the Sonjo people are not relatedlinguistically or in a direct cultural sense to those formerly inhabiting Engaruka (seeNurse and Rottland 1993) a fuller study of their villages both existing and desertedshould contribute to an understanding of
Engaruka farming by irrigation in Maasailand cAD 1400ndash1700 211
Figure 119 Sonjo wooden house with thatch dome on stone-revetted platform-terrace in Oldonyo Sambu village (Kura)
Photograph JEGSutton
specializedmdashor what are commonly called lsquointensiversquo (Sutton 1984 Widgren this volume Chapter 14 Widgren and Sutton 1999)mdashagricultural practices in isolated situations as raised by the archaeological record at Engaruka
ACKNOWLEDGEMENT
The authorrsquos continuing field research on the agricultural settlement of East Africa is supported by an emeritus fellowship awarded by the Leverhulme Trust held at the BritishInstitute in Eastern Africa and the Pitt Rivers Museum in Oxford
REFERENCES
Adams WM (1986) Observations on the Engaruka irrigation furrows and riverdischarges Pp 49ndash51 in JEGSutton The irrigation and manuring of the Engarukafield system Azania 2126ndash51
Adams WM (1989) Definition and development in African indigenous irrigationAzania 2421ndash7
Adams WM Potkanski T and Sutton JEG (1994) Indigenous farmer-managed
The archaeology of drylands 212
irrigation in Sonjo Geographical Journal 16017ndash32 Anderson DM (1989) Agriculture and irrigation technology at Lake Baringo in the
nineteenth century Azania 2484ndash97 Fosbrooke HA (1938) Rift Valley ruins Tanganyika Notes and Records 658ndash60 Hennings RO (1951) African Morning London Chatto amp Windus Murdock GM (1959) Africa Its Peoples and their Culture History New York
McGraw-Hill Nurse D and Rottland F (1993) The history of Sonjo and Engaruka a linguistsrsquo view
Azania 281ndash5 Robertshaw P (1986) Engaruka revisited excavations of 1982 Azania 211ndash26 Sassoon H (1966) Engaruka excavations during 1964 Azania 179ndash99 Sassoon H (1967) New views on Engaruka Journal of African History 8201ndash17 Soper RC (1983) A survey of the irrigation systems of the Marakwet In BE Kipkorir
RCSoper and JWSsenyonga (eds) Kerio Valley Past Present and Future 75ndash95 Nairobi University of Nairobi Institute of African Studies
Sutton JEG (1973) The Archaeology of the Western Highlands of Kenya Nairobi British Institute in Eastern Africa Memoir 3
Sutton JEG (1978) Engaruka and its waters Azania 1337ndash70 Sutton JEG (1984) Irrigation and soil-conservation in African agricultural history
Journal of African History 2525ndash41 Sutton JEG (1985) Irrigation and terracing in African agricultural history
intensification specialisation or over-specialisation In ISFarrington (ed) Prehistoric Intensive Agriculture in the Tropics 737ndash64 Oxford British Archaeological ReportsInternational Series 232 Volume 2
Sutton JEG (1986) The irrigation and manuring of the Engaruka field system Azania2126ndash51
Sutton JEG (1988) More on the Nyanga terraces the case for cattle manureZimbabwean Prehistory 2021ndash4
Sutton JEG (1989) Towards a history of cultivating the fields Azania 2498ndash112 Sutton JEG (1993) Becoming Maasailand In TSpear and RWaller (eds) Being
Maasai Ethnicity and Identity in East Africa 38ndash60 London James Currey Sutton JEG (1998) Engaruka irrigation agriculture in the northern Tanzanian Rift
Valley before the Maasai era Azania 331ndash37 Thorp C (1986) Engaruka faunal remains Pp 21ndash26 in PRobertshaw Engaruka
revisited excavations of 1982 Azania 211ndash26 Watson EE Adams WM and Mutiso SK (1998) Indigenous irrigation agriculture
and development Marakwet Kenya Geographical Journal 16467ndash84 Widgren M and Sutton JEG (1999) (eds) Islands of Intensive Agriculture in the East
African Rift and Highlands a 500-year Perspective Stockholm Stockholm University Department of Human Geography working paper 43
Engaruka farming by irrigation in Maasailand cAD 1400ndash1700 213
12 The agricultural landscape of the Nyanga area of
Zimbabwe ROBERT SOPER
INTRODUCTION
With 750ndash1200 m of rainfall a year Nyanga in the eastern highlands of Zimbabwe (Fig 121) cannot pretend to be a dryland or even semi-arid environment but it can be regarded as marginal in some other respects Furthermore its well-preserved field systems and evidence for water management practices represent parallel responses tomany of the questions addressed in this volume even if overall aridity was not theprimary driving compulsion The landscape of Nyanga and adjacent areas to the west isindelibly printed with the lsquolandesque capitalrsquo remains of past agricultural activities These take the form of stone-faced terraces and lowland cultivation ridges together withassociated stone-built settlement structures in all covering around 7000 km2 (Soper 1996) Whilst the agricultural features themselves are difficult to date the settlement sitesrange from about AD 1400 to 1900 with the earlier sites having no direct association asyet with the agricultural features
The area south of Nyanga town consists of a broad dissected plateau at around 1800 m above sea level falling relatively gently to the southwest to the main watershed betweenthe Zambezi and SabiLimpopo catchments To the east it rises to Mount Nyangani atnearly 2600 m beyond which are steep mountains and valleys into Mozambique Forabout 60 km north of Nyanga the highlands narrow progressively to a high ridge ataround 2000 m with higher peaks and with steep escarpments to east and west To thewest of this ridge granite inselbergs form often substantial hills rising from a base levelof around 1200 m while dolerite sills and dikes form lesser features The highland rangeextends northwards at a lower level for another 20ndash30 km while the surrounding lowlands decline to around 900 m The underlying geology consists of various granitesoverlain by sedimentary rocks and dolerites that cap the highlands
Drainage radiates from Mount Nyangani into major rivers such as the Gairezi andNyangombe to the north-northeast and north-northwest and
Figure 121 Location of the Nyanga area Zimbabwe
the Pungwe to the south Annual rainfall is almost entirely between November andMarch with the average ranging from c750 mm in the northern lowlands to 1200 mm ormore in the highlands Annual variation may be as much as +minus50 per cent
THE AGRICULTURAL LANDSCAPE OF NYANGA
Terraces
Stone-faced terraces cover large areas of the highland escarpments and the slopes of foothills and detached hills and ridges mainly to the west Some slopes have ranges of upto 100 terraces (Fig 122) The altitudinal range is from about 900 m in the northern lowlands to around 1700 m on the escarpments and in the highlands with very littleabove this level which is about the upper limit for the cultivation of traditional graincrops at the present day
Study of aerial photographs has identified a minimum area of 22000 ha of terracing excluding sporadic outlying occurrences Distribution favours dolerite soils and rocks Onthe geological map sheet covering 2750 km2 within which the main concentrations ofterracing occur (Stocklmayer 1978)
The agricultural landscape of the Nyanga area of Zimbabwe 215
Figure 122 Terraced hillsides in the Nyanga lowlands Photograph RSoper
over 19000 ha of terracing have been plotted of which 42 per cent are on dolerite 57 percent on granite and less than 1 per cent on sedimentary rocks (not well represented in thisarea) However 26 per cent of the dolerites below 1675 m are terraced as against only55 per cent of the granites and most of the latter are adjacent to dolerite ocurrences Thedolerites weather to red clay loams or sandy clay loams of greater fertility than the sandygranite soils but are heavily leached on the highland plateaux The younger slope soilshave more inherent fertility but are often thin and very stony so that terracing isnecessary to clear the stones and concentrate the soil for cultivation Terracing alsoprovides fairly level surfaces protects against erosion and impedes drainage to allowwater percolation
Terrace surfaces are generally narrow commonly between 15 and 3 m except on very gentle slopes where they may be up to 10 m wide Fall between terraces is normallybetween 25 and 80 cm except on the steepest slopes Slopes of up to 30 degrees wereregularly terraced in some cases up to 40 degrees Construction varies with geologytopography and the amount of stone to be disposed of and possibly also with datethough the latter remains to be established The best terraces have substantial wallsaround a metre in thickness with a double facing of large stones and a fill of smallerstones A low lip is usually present but the wall may rise a metre or more above theupper terrace surface where there was a large amount of stone to clear Such terraces
The archaeology of drylands 216
now have a more or less horizontal profile which could be the result of soil movementsince abandonment Terraces are not precisely levelled on the contour allowing forlongitudinal drainage so that it could not have been intended to flood them eitherartificially or by rainfall Stone-lined drains carried excess run-off down slope while in some cases upstanding walls were pierced by drain holes
This type of terrace is generally found on dolerite but occasionally also on granite The soil is often shallow from less than 20 cm up to a maximum of 50ndash60 cm against the lower wall face It is relatively stoneless so that it must have been worked over toremove even the smallest stones during construction The substratum is of denselypacked stones in a red clay matrix in the case of dolerite or more or less decomposedrock on granite
In granite areas with less stone and on sedimentary argillites in the northern part of the complex terraces are generally lower and the stonework appears to consist of no morethan a simple revetment while terrace profiles are sloping with gradients of up to 15 oreven 20 degrees being recorded The only excavated transect showed numerous stonesremaining in the soil This type may represent the rapid exploitation of less favourablebut still fertile soils and it is not known if it is contemporary with the former type
Most of the terraces do not appear to have been irrigated There are a few cases whereold water furrows do traverse ranges of terraces and they may well have been used forirrigating those below but no distribution channels have been observed and settlementsites also appear to have been served In the case of the detached hills to the west manyof which also have extensive terracing gravity irrigation would not have been feasible
The chronological range of terrace building is uncertain but probably spanned at least the seventeenth to early ninetenth centuries Dating and associations are discussed belowunder landscape development The only direct radiocarbon date for a terrace (Pta-7601) is 200 +minus50 BP calibrated to anywhere between 1618 and 1878 at one sigma This datewas obtained from tiny disseminated charcoal fragments in soil of a second phase ofterrace construction in a granite area adjoining a stone enclosure A total of 537 sherdsmostly small and worn was also obtained from some 3 m3 of soil The sherds and charcoal may derive from manuring with domestic refuse from the enclosure but thelatter appears to date from the later nineteenth century and there are some differences inthe pottery so they may well derive from an earlier site perhaps contemporary with theearlier terrace phase In either case the date gives only a maximum age which is not veryuseful in view of the wide calibration bracket
Cultivation ridges
The second notable feature of the old agricultural landscape comprises extensivenetworks of ridges and ditches on the lower less stony slopes below the escarpments and extending some 60 km to the west No quantification of these has been attempted but thetotal area must equal or exceed that of the terracing In the terminology of Denevan andTurner (1974) these are long flat-topped linear ridges Some especially in wetter situations tend to be more cambered due to the greater height and somewhat closerspacing needed for effective drainage The features are parallel or sub-parallel linear banks usually 7ndash10 m wide between ditches up to a metre or so deep They often run for
The agricultural landscape of the Nyanga area of Zimbabwe 217
several hundred metres with a more or less shallow longitudinal gradient These occurboth in areas of impeded drainage (termed lsquovleisrsquo) and on the valley sides or interfluves
An example may be described at the base of the main escarpment near Maristvalesome 40 km north of Nyanga town Here there is a broad bay in the escarpment about 2km wide between high projecting spurs and a series of streams converges across thepiedmont slope Virtually the whole of the interfluves and most of the stream valleys arescored with ridges and ditches covering around 1000 ha The central interfluve (Fig 123) provides an area some 1750 m long and around 500 m wide with a longitudinal fallof c60 m and a maximum lateral height of c10ndash12 m Almost all of this is occupied byridges except for a stonier crest towards the upper end which is terraced and a fewminor areas of outcropping rock with stone enclosures The ridges trend longitudinallydown the interfluve with a broadly parallel alignment sometimes rather braided Asimilar pattern is seen on the other interfluves At the head of this interfluve at the base ofthe escarpment is a furrow take-off from a small stream There appears to be no main feeder furrow from this but water could be directed down any of the ditches or to theoccupation sites Towards the lower end of the interfluve a furrow did carry water from adeep set of ditches diagonally across the ridges probably to a stone enclosure on thecrest Soils here are silty sands over a sheet of consolidated rounded quartz gravel
Other occurrences are in more specifically waterlogged areas An example is aregularly waterlogged perched vlei on the piedmont slope a few km south of the above site Here there is a dendritic pattern of banks and ditches for maximum drainage and asection showed a metre or so of mottled sandy clay loam overlying dense black clay Theclay loam must derive by erosion from the terraced area immediately above perhapsbefore terracing anchored the soil or perhaps from inefficient use of the terraces It wasthen re-exploited after deposition
Ridge size patterns and orientation to slope appear to vary even within a singlelocalized drainage basin and must represent a flexible system of balancing drainage andwater retention under varying conditions of soil slope rainfall and seasonal water tableIn the first case described the primary purpose would seem to be controlled drainageraising the cultivation beds above any actual or potential waterlogging without removingrainfall too directly If necessary supplementary water could have been introduced to theditches though the water available at present would seem
The archaeology of drylands 218
Figure 123 Vertical aerial photograph of cultivation ridges crossed by an old trackway and a water furrow
Photograph Harare Office of the Surveyor General
inadequate for any extensive irrigation In wetter areas drainage could be more direct Denevan and Turner (1974) review the advantages of raised beds in general Relevant
points here may be control of erosion provision of drier cultivation conditions wherethere is permanent or periodic inundation or waterlogging but with some water retentionin the ditches still available to crop roots wide beds reducing the ditch area aeration ofthe soil and modification of microclimate if there is danger of frost To these could beadded the variation of moisture availability across the ridge and ditch appropriate todifferent crops Moisture-loving traditional root plants such as Colocasia (taro) and Zantedeschia (calla lily) would be appropriate for the wetter ditches while sorghummillets and legumes could grow on the ridges with Plectranthus (lsquoLivingstone potatorsquo) perhaps somewhere in between
The ridging systems remain to be dated since none of the related stone enclosures mentioned above has been excavated They cannot be later than nineteenth century andthey are different from recent mihomba cultivation ridges which are shorter narrower
The agricultural landscape of the Nyanga area of Zimbabwe 219
straighter and generally restricted to waterlogged environments such as wet stream banksThe intimate relationship to terraces in the Maristvale area suggests contemporaneitywith at least some terracing It seems likely though that the large labour demands forconstructing and operating both systems simultaneously on a large scale would have beenbeyond the capacity of individual communities
Water management
Water is a critical resource in African agriculture generally and its management infavourable conditions can provide insurance against bad rainfall years and extended dryperiods within a normal wet season as well as giving the potential to extend the growingseason before or after The possibility of supplementary water supply to some ridgesystems has been mentioned above as has the general lack of evidence for widespreadterrace irrigation Terraces and cultivation ridges even if not directly irrigated reflectwater management by controlled drainage to provide good infiltration
Permanent streams are common in the Nyanga highlands and descending the escarpments the potential of these was clearly appreciated because numerous old furrowshave been observed mainly in the highlands where they are better preserved by perennialgrass cover A tentative classification of these furrows can be suggested
Type 1 is the commonest and most widespread in the highlands and would have serveddomestic requirements livestock and homestead gardens Some could be diverted toflush out stone-lined pits used for livestock and provision was often made to impoundthe resultant slurry Only a few cases of Type 2 have been recorded both in the highlandsand on the lower escarpment slopes while the only case of Type 3 known to date is thatdescribed above Type 4 appears to be restricted to a limited area centred on the northernpart of Nyanga National Park and must have been for irrigation of unterraced fields sinceno terraces are associated below them and only very rarely are settlement sites servedType 5 is thought to belong to the colonial period The others belong at least to the nineteenth century and probably earlier while some examples of Type 1 are likely to beassociated with seventeenth-century sites
Authorship
The authorship of the agricultural works can almost certainly be attributed to theancestors of the present indigenous inhabitants (that is before the relocation ofpopulations consequent on colonial land policies) These are the Unyama people for the
1 small furrows of varying gradient and length associated with occupation sites 2 generally well-graded furrows on relatively narrow revetted shelves traversing
ranges of terraces probably used for irrigating those below but also often serving occupation sites
3 furrows assocated with ridging systems 4 well-graded furrows involving more or less massive earthen banks with potentially
irrigable land below sometimes with recognisable branch furrows or ditches and 5 furrows without major banks or stone work
The archaeology of drylands 220
area north of Nyanga town the Manyika to the south and the Maungwe west of theNyangombe river Genealogies and traditions of the chiefly families (Beach 1995) goback well into the eighteenth century at least and more in the case of the Manyika and itis surprising that more oral traditions have not survived on the construction and use ofterraces and ridges It would seem that knowledge and use of these specializedagricultural techniques were common to a number of political and dialect groupings innortheastern Zimbabwe and should not be attributed to a single group
AGRICULTURAL SYSTEMS
The agricultural systems of the terrace builders integrated crops and animals Cattle werealmost certainly penned in a sunken stone-lined pit or small stone enclosure within thehomestead In the case of pits roofed or tunnel-entrance passages would have admittedonly dwarf cattle bones of which have been recovered from the only site with good bonepreservation (Plug et al 1997) The small enclosures in the northern part of the area however have open entrances and could have accommodated larger beasts Pits andinternal enclosures in the lowlands are relatively small with an internal diameternormally around 3 m and a depth or height of around 12ndash15 m Fairly small cattle holdings are thus indicated Seasonal permanent stall-feeding has been suggested by Sutton (1988) as practised for instance in parts of Nigeria and Ethiopia (Hallpike 1970Netting 1968) but the heightdepth rules out any substantial accumulation of manure in situ Pits in the highlands are larger and deeper usually 5ndash9 m in diameter and 180ndash3 m deep More cattle are thus indicated above the terrace zone where the depth could haveaccommodated the accumulation of manure but was more likely for protection from thecold winds of winter Goats and possibly sheep were kept in the houses many of whichhave a low dividing wall with one half paved with stones (Soper 1996)
Pits and internal enclosures rarely contain any deposits beyond leaf mould and a little silting and no dung heaps or other substantial middens have been found Dung was thusregularly removed and must have been used for manure with some possibly being driedfor fuel in the highlands where wood may have been at a premium Pits were providedwith drains and in many cases in the highlands would have been flushed out with water from furrows Again in the highlands small dams were often built below the homesteadto catch the slurry or ditches were dug to channel it to small hollows Many such pitstructures have radial walls which are thought to have sheltered gardens on which theslurry could have been used Where no furrow was available as more particularly in thelowlands dung must have been removed by hand and any flushing have relied onrainwater goat dung must similarly have been removed by hand from the housesDomestic refuse was doubtless added to the manure
It is unlikely that there would have been sufficient manure to fertilize the full range of cultivated land On general ethnographic analogies one would expect it to have beenused on homestead gardens irrigated where practicable and on terraced or other plots inthe vicinity but rarely on more outlying fields Results of phosphate analysis from thearchaeological contexts are ambiguous regarding the extent of manuring
Terracing per se could thus be considered a specialized technique implying only
The agricultural landscape of the Nyanga area of Zimbabwe 221
relative lsquointensificationrsquo but a higher degree of the latter may be postulated in an inner zone around the homesteads probably dependent on available water supply Thecultivation of outlying terraces even on the more fertile dolerite soils would have beenless sustainable and a continuous process of terrace building can be envisaged witholder terraces being fallowed or abandoned as fertility declined
The lack of excavation of settlement sites associated with cultivation ridge systems inhibits any conclusions on their use as yet
The range of crops and cultivation methods might be expected to have varied over the altitudinal and rainfall range of the complex Summers (1958) identified seeds from Ziwaruins in the lowlands at around 1300 m These comprised mainly traditional grains andlegumes including Sorghum Pennisetum (bullrush millet) Eleusine (finger millet) Vigna unguiculata (cow pea) Vigna subterranea Ricinus and perhaps Citrullus part of a maize cob was also found but in a surface context Seeds recovered by flotation in thepresent research have not added any cultivars to this list Enquiries about traditional cropsadd a number of important root crops Plectranthus esculenta (lsquoLivingstone potatorsquo) Colocasia (taro) and probably Zantedeschia (calla lily) as well as pumpkins andcucumbers and several semi-wild fruits leaf plants and oil-seed plants as well as numerous wild fruits and other plants were also harvested The traditional varieties ofColocasia and also Zantedeschia are toxic without extended boiling Traditional cropping practices commonly involved interplanting of grains legumes and cucurbits It isprobable that outlying terraces were devoted mainly to grain staples but predation bywild animals and birds could have been a problem Gardens and in-fields were probably used more for vegetables roots and legumes here a more intimate familiarity with soildepth and quality would have enabled more attention being given to the individualrequirements of different plants
LANDSCAPE DEVELOPMENT
The processes and sequence of landscape modification are not yet well understood but itis unlikely that the terracing and ridging were the work of a large and dense populationover a relatively short time period The Unyama people within whose territory thegreatest concentrations of terracing occur were not sufficiently important to attract anyattention or record from the Portuguese who interacted closely with the Mutapa state tothe west and also with the Manyika immediately to the south The settlement patternrepresents loosely dispersed homesteads in village groupings Although the stone-built homesteads are very numerous and may be locally concentrated especially in lowlanddolerite areas none appears to represent prolonged occupation and there are very fewstratified sites or substantial middens We must therefore see terrace construction as anongoing process over many generations among the communities of a fairly limitedoverall population There is some indication of the reoccupation of homesteadssuggesting that whole settlements and their fields may have been fallowed and resettledtaking advantage of the established capital infrastructure
The limited number of dated sites enables only a tentative interpretation of the process of development of the complex with some notable lacunae that may be real or only
The archaeology of drylands 222
apparent Thirty radiocarbon dates are now available all from the northern area from justsouth of Nyanga town All those later than about 400 BP have very wide calibrationranges
The earliest dated stone ruins are all in the highlands By the fourteenth and fifteenth centuries relatively extensive sites were occupied on the highest peaks and ridges ataltitudes over 2000 m followed in the seventeenth century by pit structures nowcompletely ruined at slightly lower altitudes These sites are all above the level ofterracing so there is no direct association Later highland pit structures tend to be loweragain and are relatively well preserved some with surviving dhaka (clay) walls Their construction must have continued well into the nineteenth century and there may havebeen a hiatus in highland occupation from the earlier ruined pits perhaps occasioned bythe second severe phase of the lsquoLittle Ice Agersquo (Tyson and Lindesay 1992) Occupants ofthese sites must have been responsible for the terracing on the western escarpmentsFurther south the banked furrows of the National Park area with their implied irrigationof unterraced fields are a local perhaps relatively late development probably also thework of pit-structure occupants living more or less closely above them
In the lowlands most of the dated sites are stone enclosures within the Ziwa ruins National Monument and range between 140 and 220 BP calibrating anywhere betweenthe second half of the seventeenth century and the early or even late nineteenth centuryEarlier sites may exist here or elsewhere in the lowlands but have not been dated orperhaps not recognized if not built in stone so it is not known if there was any occupationcontemporary with the earlier highland sites The extensive terracing of the Ziwa area with which the stone enclosures are associated can probably be bracketed between theseventeenth and early nineteenth centuries most of the western lowlands between theescarpment and the Nyangombe river were depopulated by the end of the nineteenthcentury when the first European travellers passed through Further north there is adifferent type of homestead design with small well-built central livestock enclosures Three dates from here are recent at 100 BP or less but a couple of dates probably fromsecondary contexts (including the terrace date quoted above) suggest occupationcontemporary with Ziwa
In general one may suggest a continuous process of terrace construction with new terraces being built as older ones declined in fertility and were abandoned Terracingwould have concentrated initially on dolerite soils and then spread to adjacent graniteareas Ultimately terraceable land may have run out and the fertility of homestead plotsproved unsustainable resulting in piecemeal or wholesale removals to new sites In thisway the impressive agricultural landscape we now see could have been created with arelatively low overall population density
The position of the cultivation ridges in this development is uncertain pending the dating of associated settlement sites It may be assumed that some wetter lands werealways exploited by terrace builders where conveniently available and that ridgepractitioners resorted to terracing of suitable land within their ambit but a large-scale simultaneous use of both terraces and ridges by the same communities seems unlikely interms of labour requirements Either lowland practice switched from a concentration onterracing to one on ridging (or vice versa) or each local community emphasized one orthe other system according to the type of land available While little direct research has
The agricultural landscape of the Nyanga area of Zimbabwe 223
been done in areas to the west and southwest it may be noted that the ridging systemscontinue to the RusapeHeadlands area but terracing becomes more sporadic probablyconcentrated mainly on the limited dolerite occurrences Each system was probably aparallel exploitation by related communities
DISCUSSION
Although the resources of Nyanga would have been less critical than in the more aridsituations that are the focus of most studies in this volume the tactics of soil and watermanagement show many parallels The agricultural systems represent a range ofspecialized responses tailored to the potentialities of the local environment The relativefertility of the younger slope soils was clearly appreciated and their potential was realizedby the development of appropriate terracing technology for stone clearance soilconservation and control of drainage The cultivation ridges enabled the exploitation ofthe more leached and often waterlogged valley soils The alternative options of ridgingand terracing complemented each other and provided a risk strategy for coping with short-term climatic fluctuations emphasizing either terrace cultivation in wet years or thevalley soils in dry years This might be within a single community where both resourceswere available or by reciprocal co-operation between local communities A similar co-operative relationship may be envisaged between highland and lowland communitieswith greater concentration on cattle and cultivation respectively Water resources wereexploited by furrow technology for domestic convenience and garden irrigation to extendthe growing season and lessen the effects of dry spells Integration with livestockmanagement produced manure to extend the fertility span of at least part of the cultivatedland and to maximize returns from labour investment
Exotic items are extremely rare or absent apart from a few glass beads This and the lack of Portuguese references to the area indicate little participation in trading networksand little differentiation in relative wealth Production for basic subsistence is thusindicated Design and construction of homesteads appear to go beyond purely functionalnecessities reflecting no great economic stress while ample storage facilities show anadequate level of food production Some sites such as lsquofortsrsquo with evidence of regular occupation suggest some degree of local authority but no marked social stratificationEthnographic parallels for terracing and irrigation in East Africa in general areconsistently associated with acephalous kin-based social organization within which the agricultural systems are integrated for land allocation labour mobilization and thesettlement of disputes (Haringkansson 1989) Something similar may be suggested here the various chiefships within which the complex fell are unlikely to have had any significantfunction in directing subsistence activities or extracting undue tribute Terrace building asan on-going piecemeal process is feasible within the labour resources of a family group perhaps assisted by mutual working parties within the local community Most of thewater furrows would also be within the capabilities of the family with the exception ofType 4 which must have required community co-operation for the substantial earth movement involved
The stimulus to the very labour-intensive cultivation practices would not seem to
The archaeology of drylands 224
derive directly from serious environmental constraints While political constraints areuncertain for the earlier centuries they do not appear to have been particularly pressingfor the eighteenth century defensive structures indicate the need for temporary refugeprobably in response to more or less local raiding but lowland settlement at least wouldhave been vulnerable to any consistent outside threats For explanation one may perhapslook more to the opportunities offered by local circumstances as suggested by Brookfield(1986) whereby innovations adopted for the exploitation of particular niches in this casethe fertility of dolerite slope soils offered a lsquoquantum leaprsquo in productivity Although overall population was low initial relative local density induced by the preference for thedolerite areas provided the necessary labour resources and would have been enhanced bythe resultant success
Reasons may be suggested for the decline and abandonment of the systems but theyremain to be tested Declining fertility in the long term may have reduced the populationbelow a critical level Drastic drought could have been a factor for instance acatastrophic drought occurred in the lower Zambezi and coastal area in the 1820s thoughit did not necessarily affect the Nyanga region In Unyama at least persistent strugglesfor the chiefship between two factions from the late eighteenth century contributed to thedepopulation of large areas of the western lowlands by the time of European penetrationin the 1890s but this should not have affected areas in the neighbouring chiefdoms Thesystems had already survived the last drier cold phase of the Little Ice Agemdashmay indeed even have been a response to itmdashand perhaps the subsequent climatic amelioration from the first half of the nineteenth century made them unnecessary
ACKNOWLEDGEMENTS
The research on which this chapter is based was carried out under a joint project of theBritish Institute in Eastern Africa and the History Department University of Zimbabwein close co-operation with the National Museums and Monuments of ZimbabweGratitude is acknowledged to these institutions and to various agencies of the governmentof Zimbabwe for facilitating the work Particular thanks are due to John Sutton forinitiating the project and advising on all stages of the research
REFERENCES
Beach D (1995) Archaeology and History in Nyanga Zimbabwe Harare University of Zimbabwe unpublished seminar paper
Brookfield HC (1986) Intensification intensified Archaeology in Oceania 21 3177ndash80
Denevan W and Turner B (1974) Forms functions and associations of raised fields inthe Old World Tropics Journal of Tropical Geography 3924ndash33
Haringkansson T (1989) Social and political aspects of intensive agriculture in East Africa some models from cultural anthropology Azania 2412ndash20
Hallpike CR (1970) Konso agriculture Journal of Ethiopian Studies 8 131ndash43
The agricultural landscape of the Nyanga area of Zimbabwe 225