Ancient Earthquakes

Ancient Earthquakes

edited by Manuel Sintubin Department of Earth and Environmental Sciences Katholieke Universiteit Leuven Celestijnenlaan 200E, B-3001 Leuven Belgium Iain S. Stewart School of Geography, Earth and Environmental Sciences University of Plymouth Room 109, Fitzroy, Drake Circus Plymouth, Devon, PL4 8AA UK Tina M. Niemi Department of Geosciences University of Missouri–Kansas City 5100 Rockhill Road Kansas City, Missouri 64110 USA Erhan Altunel Department of Geological Engineering Eskişehir Osmangazi University 26480 Eskişehir Turkey

Special Paper 471 3300 Penrose Place, P.O. Box 9140

Boulder, Colorado 80301-9140, USA

2010

Copyright © 2010, The Geological Society of America (GSA), Inc. All rights reserved. GSA grants permission to individual scientists to make unlimited photocopies of one or more items from this volume for noncommercial purposes advancing science or education, including classroom use. For permission to make photocopies of any item in this volume for other noncommercial, nonprofit purposes, contact The Geological Society of America. Written permission is required from GSA for all other forms of capture or reproduction of any item in the volume including, but not limited to, all types of electronic or digital scanning or other digital or manual transformation of articles or any portion thereof, such as abstracts, into computer-readable and/or transmittable form for personal or corporate use, either noncommercial or commercial, for-profit or otherwise. Send permission requests to GSA Copyright Permissions, 3300 Penrose Place, P.O. Box 9140, Boulder, Colorado 80301-9140, USA. GSA provides this and other forums for the presentation of diverse opinions and positions by scientists worldwide, regardless of their race, citizenship, gender, religion, or political viewpoint. Opinions presented in this publication do not reflect official positions of the Society. Copyright is not claimed on any material prepared wholly by government employees within the scope of their employment. Published by The Geological Society of America, Inc. 3300 Penrose Place, P.O. Box 9140, Boulder, Colorado 80301-9140, USA www.geosociety.org Printed in U.S.A. GSA Books Science Editors: Marion E. Bickford and Donald I. Siegel Library of Congress Cataloging-in-Publication Data Ancient earthquakes / edited by Manuel Sintubin … [et al.]. p. cm. — (Special paper ; 471) Includes bibliographical references. ISBN 978-0-8137-2471-3 (pbk.) 1. Paleoseismology. I. Sintubin, M. QE539.2.P34A53 2010 551.22—dc22 2010036726 Cover, front: At Cape Sounion, the southernmost tip of the Attica peninsula (Greece), a temple is dedicated to Poseidon, the “Earth-Shaker,” god of earthquakes. The displaced drums of the multidrum columns are typically believed to be evidence of ancient earthquakes. The temple was built around 440 B.C. (photograph courtesy of K. Reicherter, Rheinisch-Westfälische Technische Hochschule Aachen, Germany). Back (top to bottom): Examples of earthquake-induced damage. Al Harif aqueduct, Syria (see Sbeinati et al., this volume, Chapter 20) (courtesy of M. Meghraoui, Institut de Physique du Globe, Strasbourg, France); Cnidus, Turkey; Hierapolis, Turkey; Sagalassos, Turkey; Petra, Jordan; Baelo Claudia, Spain (see Grützner et al., this volume, Chapter 12) (photographs courtesy of M. Sintubin, Katholieke Universiteit Leuven, Belgium).

10 9 8 7 6 5 4 3 2 1

Contents

Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . v Manuel Sintubin, Iain S. Stewart, Tina M. Niemi, and Erhan Altunel Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xiii Understanding Earthquakes in the Ancient World 1. Dynamic landscapes and human evolution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 Geoffrey C.P. King and Geoffrey N. Bailey 2. Tectonic environments of ancient civilizations: Opportunities for archaeoseismological and anthropological studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 Eric R. Force and Bruce G. McFadgen Historical Earthquakes and Their Societal Impact 3. The door knockers of Mansurah: Strong shaking in a region of low perceived seismic risk, Sindh, Pakistan . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29 Roger Bilham and Sarosh Lodi 4. San Antonio de Mucuñó, Mérida Andes, Venezuela: Relocation of a doctrine town following the 1674 earthquake . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39 Jaime Laffaille, Franck Audemard M., and Miguel Alvarado 5. New interpretations of the social and material impacts of the 1812 earthquake in Caracas, Venezuela . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47 Rogelio Altez 6. The impact of the 1157 and 1170 Syrian earthquakes on Crusader–Muslim politics and military affairs. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59 Kate Raphael 7. Western Crete: From Captain Spratt to modern archaeoseismology . . . . . . . . . . . . . . . . . . . . . . . 67 Manolis I. Stefanakis 8. Earthquake archaeology in Japan: An overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81 Gina L. Barnes

iii

iv

Contents Commentaries and Perspectives on Archaeoseismological Research 9. Historical earthquake catalogues and archaeological data: Achieving synthesis without circular reasoning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97 John D. Rucker and Tina M. Niemi 10. Historical earthquakes in Srinagar, Kashmir: Clues from the Shiva Temple at Pandrethan . . . 107 Roger Bilham, Bikram Singh Bali, M. Ismail Bhat, and Susan Hough 11. Earthquakes and civilizations of the Indus Valley: A challenge for archaeoseismology . . . . . . . 119 Robert L. Kovach, Kelly Grijalva, and Amos Nur 12. Comparing semiquantitative logic trees for archaeoseismology and paleoseismology: The Baelo Claudia (southern Spain) case study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129 Christoph Grützner, Klaus Reicherter, and Pablo G. Silva 13. Long-term effect of seismic activities on archaeological remains: A test study from Zakynthos, Greece . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145 Melek Tendürüs, Gert Jan van Wijngaarden, and Henk Kars 14. Assessment of seismically induced damage using LIDAR: The ancient city of Pınara (SW Turkey) as a case study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157 Barış Yerli, Johan ten Veen, Manuel Sintubin, Volkan Karabacak, C. Çağlar Yalçıner, and Erhan Altunel Practices in Archaeoseismology 15. Ancient earthquakes from archaeoseismic evidence during the Visigothic and Islamic periods in the archaeological site of “Tolmo de Minateda” (SE Spain). . . . . . . . . . . . . . . . . . . . 171 M.A. Rodríguez-Pascua, P.G. Silva, V.H. Garduño-Monroy, R. Pérez-López, I. Israde-Alcántara, J.L. Giner-Robles, J.L. Bischoff, and J.P. Calvo 16. Distinguishing damages from two earthquakes—Archaeoseismology of a Crusader castle (Al-Marqab citadel, Syria) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 185 Miklós Kázmér and Balázs Major 17. Archaeoseismological studies at the temple of Amenhotep III, Luxor, Egypt . . . . . . . . . . . . . . . 199 Arkadi Karakhanyan, Ara Avagyan, and Hourig Sourouzian 18. Archaeological evidence for Roman-age faulting in central-northern Sicily: Possible effects of coseismic deformation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 223 Giovanni Barreca, Maria Serafina Barbano, Serafina Carbone, and Carmelo Monaco 19. Faulting of the Roman aqueduct of Venafrum (southern Italy): Methods of investigation, results, and seismotectonic implications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 233 Paolo A.C. Galli, Alessandro Giocoli, Jose A. Naso, Sabatino Piscitelli, Enzo Rizzo, Stefania Capini, and Luigi Scaroina 20. Timing of earthquake ruptures at the Al Harif Roman aqueduct (Dead Sea fault, Syria) from archaeoseismology and paleoseismology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 243 Mohamed Reda Sbeinati, Mustapha Meghraoui, Ghada Suleyman, Francisco Gomez, Pieter Grootes, Marie-Josée Nadeau, Haithem Al Najjar, and Riad Al-Ghazzi 21. Offset archaeological relics in the western part of the Büyük Menderes graben (western Turkey) and their tectonic implications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 269 Önder Yönlü, Erhan Altunel, Volkan Karabacak, Serdar Akyüz, and Çağlar Yalçıner

The Geological Society of America Special Paper 471 2010

Preface Manuel Sintubin Department of Earth and Environmental Sciences, Katholieke Universiteit Leuven, Celestijnenlaan 200E, B-3001 Leuven, Belgium Iain S. Stewart School of Geography, Earth and Environmental Sciences, University of Plymouth, Room 109, Fitzroy, Drake Circus, Plymouth, Devon, PL4 8AA, UK Tina M. Niemi Department of Geosciences, University of Missouri–Kansas City, 5100 Rockhill Road, Kansas City, Missouri 64110, USA Erhan Altunel Department of Geological Engineering, Eskişehir Osmangazi University, 26480 Eskişehir, Turkey

Damaging earthquakes typically recur at intervals of centuries to millennia, but the instrumental record extends for no more than a century. To reduce the hazards from earthquakes and prepare proper mitigation plans we need a longer record of earthquakes than can be provided instrumentally. Looking for ancient earthquakes may be the key to the puzzle. We define ancient earthquakes primarily as pre-instrumental earthquakes that can only be identified through indirect evidence in the archaeological or geological record. While the latter is the subject of paleoseismology, the former is the subject of archaeoseismology. Earthquakes that are documented in the historical record (historical seismicity) may be included if they left marks in the archaeological or geological record. The problem seismic-hazard practitioners now face is that the instrumental record is too short and the historical record too incomplete. Historical catalogues record only a tiny proportion of the major earthquakes that have struck a region over centuries and millennia (cf. Ambraseys et al., 2002). That missing population of earthquakes clearly tempers reliable seismic-hazard assessments. The archaeological record, however, can bolster and augment that historical archive. What’s more, in extending the earthquake record beyond written sources, archaeoseismology serves as a bridge between instrumental and historical seismology, on the one hand, and paleoseismology and earthquake geology, on the other hand. Only the integration of all potential evidence of ancient earthquakes will enable a better understanding of the complex earthquake history of a region. Archaeoseismology has the potential to be a legitimate and complementary source of seismic-hazard information. Archaeoseismology thus aims at studying ancient earthquakes through indicators left in the archaeological record, such as destruction layers, structural damage to man-made constructions, cultural piercing features, indications of repairs, abandonment, cultural changes, etc. Archaeology can be used in three ways to help confront the seismic-hazard threat. First, where archaeological relics are displaced they can be used to find active faults, show in which direction faults slipped during the earthquake(s), and establish comparative fault slip rates. Second, archaeological information can date episodes of faulting and shaking. Third, we can search for ancient signs of seismic damage, often related to the ground shaking. The obvious difficulty with

Sintubin, M., Stewart, I.S., Niemi, T.M., and Altunel, E., 2010, Preface, in Sintubin, M., Stewart, I.S., Niemi, T.M., and Altunel, E., eds., Ancient Earthquakes: Geological Society of America Special Paper 471, p. v–xi, doi: 10.1130/2010.2471(00). For permission to copy, contact [email protected]. © 2010 The Geological Society of America. All rights reserved.

v

vi

Preface the last approach is that it is hard to distinguish between damage caused by an earthquake and that caused by another destructive event, such as war or the natural failure of foundations. Typologies of earthquakecharacteristic damage have been proposed but rarely have they been subjected to a critical and systematic analysis. Consequently, these archaeoseismological indicators are accepted by some earthquake scientists and rejected by others. In this volume we collected a series of case studies convincingly illustrating the different ways that the archaeological record can serve seismic-hazard studies. Moreover, the difficulties archaeoseismology face are discussed extensively. This volume is the first publication in the framework of the International Geoscience Programme IGCP 567 “Earthquake Archaeology: Archaeoseismology along the Alpine-Himalayan Seismic Zone” (ees.kuleuven.be/igcp567/). The key element of the program is our contention that archaeological earthquake evidence can make a valuable contribution to long-term seismic-hazard assessment in earthquakeprone regions where there is a long and lasting cultural heritage. Archaeological evidence indeed has the potential to determine earthquake activity over millennial time spans, especially where integrated with historical documents and geological evidence, as demonstrated in several papers in this volume. Archaeoseismology’s greatest challenge—and its foremost attraction—is to integrate the principles and practices of a wide range of sciences, from history, anthropology, archaeology, and sociology, over geology, geomorphology, geophysics, and seismology, to architecture and structural engineering. IGCP 567 seeks to establish an inclusive framework in which such a multidisciplinary approach can take root. It aims to encourage transparent discussions between this wide range of specialists, to promote a common vocabulary and understanding of purpose, and to foster a standardized methodology based on the consensus of those active in the field of archaeoseismology. Innovative archaeoseismological research has emerged in many parts of the globe, but its roots lie in the Eastern Mediterranean and the Middle East. The majority of methodological developments has indeed been grafted on archaeological investigations in this earthquake-prone region atop the Alpine collision zone between Africa and Eurasia, and depended strongly on identifying structural damage to buildings and other cultural remains at specific sites. Therefore, the program’s second objective is to extend the geographical provenance of archaeoseismological studies beyond its traditional territory, eventually pursuing a common archaeoseismological knowledge platform and a shared protocol. Considering the wide range of disciplines in which the authors are active, and the geographical distribution of cases presented, the main objectives of IGCP 567 are clearly reflected in this volume. In this respect, Special Paper 471 forms a new contribution to the ongoing process of refining our research strategies and practices in archaeoseismology. It visualizes the significant progress that persistent research efforts during the last two decades has made possible. This volume thus frames in a series of publications, each reflecting the gradual evolution toward an ever increasing multidisciplinary approach (cf. Caputo and Pavlides, 2008; Galadini et al., 2006; Stiros and Jones, 1996).

Preface UNDERSTANDING EARTHQUAKES IN THE ANCIENT WORLD The first two papers in Ancient Earthquakes reflect on the way tectonically active environments may have influenced man’s own prehistory. These thought-provoking papers give a wider, more theoretical context to a volume that is primarily focused on the effects of single earthquake events. In the paper by King and Bailey the authors argue that the geological instability in tectonically active regions may very well be an environmental driver for human development by creating more stable environmental conditions for sustained human settlement. This paper does not focus on the occasional disruptive earthquake events but on the modifications of regional topography resulting from the cumulative effect of earthquake activity over centuries to millennia. The refreshing hypothesis is convincingly illustrated by the early hominid settings in different parts of the East African Rift. This idea is further developed in the second paper. Force and McFadgen see a robust spatial correlation between ancient civilizations and active plate boundaries in the Alpine-Himalayan seismic zone, in particular with respect to the outline of trading routes. They also point out that a correlation can be supposed between increased seismicity and accelerated cultural change, which may define a new research question challenging practitioners of archaeoseismology. HISTORICAL EARTHQUAKES AND THEIR SOCIETAL IMPACT In this section four papers focus on historically known earthquakes to look for clues to assess their societal impact. These papers show that each earthquake, independent of its magnitude and frequency, provokes different societal responses, largely depending on the political, social, and economic context. The latter will eventually determine whether or not an earthquake disaster leads to the decline of a society. Earthquakes in themselves are incapable of causing the collapse of a community, let alone a civilization. These historical papers clearly show that any reference to neocatastrophism (e.g., Evans, 1928; Marinatos, 1939; Nur, 2008; Schaeffer, 1948) that has classically plagued archaeoseismology (cf. Ambraseys, 2005; Kovach and Nur, 2006) should be omitted. Bilham and Lodi argue that the discovery of decorated door knockers beneath a collapsed wall in Mansurah, the eighth-century capital of the Sindh province in Pakistan, supports an earthquake hypothesis to explain the destruction and abandonment of the capital in ca. 980 A.D. Such indications for strong shaking in a region of low perceived seismicity make a case for establishing a record of long-term seismicity using the 5 millennia of archaeological remains in Pakistan. In the paper by Laffaille et al. the impact of a historical earthquake in 1674 in northern Venezuela to a seventeenth-century Spanish settlement is analyzed. By balancing historical data with geological, geomorphological, and paleoseismological data, earthquake effects on a microscale are demonstrated. Earthquaketriggered landslides and mudflows eventually caused a relocation of the settlement over a few hundred meters. Altez, on the other hand, presents a detailed historical analysis of the very destructive earthquake that struck Caracas, Venezuela, on 26 March 1812. Altez uses documentary evidence to show that this earthquake had a significant impact on the Venezuelan society, primarily because the earthquake struck in a context of political, social, and economic turmoil. In this vulnerable historical context the quality of the build environment has deteriorated, biasing post-factum intensity estimates. Also the number of earthquake victims was overestimated because of the inclusion of victims of societal turmoil (e.g., war, famine). This social and political context is even more obvious in the historical assessment by Raphael of two devastating earthquakes that struck the Levant in a time span of less than 20 years in the twelfth century. Raphael nicely illustrates that the political and military balance between the Crusader Kingdom of Jerusalem and the Muslim Sultanate of Syria played a crucial role in the way both earthquakes influenced the regional political and military affairs. From these historical-anthropological approaches a clear lesson can be learned by earth scientists primarily focusing on the physical aspects of earthquake events. An earthquake disaster always seems to have a significant social component to it. The historical context should therefore be taken into account before any conclusions are made with respect to the societal effects of ancient earthquakes that are evidenced in the archaeological record. This section is rounded off with two papers reflecting on the history of archaeoseismology. The paper by Stefanakis is paying tribute to Captain T.A.B. Spratt (1811–1888) as the nineteenth-century forerunner of

vii

viii

Preface archaeoseismology in western Crete, Greece. His observations of coastal uplift in western Crete, as well as in the ancient harbors of Phalasarna and Kissamos, are still in line with the current knowledge with respect to the seismotectonics of western Crete. The paper by Barnes illustrates how the development of an archaeoseismological methodology largely depends on the particular regional context. Japanese earthquake archaeology differs greatly from the standard Mediterranean archaeoseismology, primarily because of the nature of the evidence. While in the Mediterranean approach focus has been put on the structural damage to buildings and other archaeological remains, secondary phenomena, such as liquefaction, landslides, and surface cracking, within a more territorial context are the main features used in the Japanese approach. COMMENTARIES AND PERSPECTIVES ON ARCHAEOSEISMOLOGICAL RESEARCH The first paper in this section on archaeoseismological research, by Rucker and Niemi, comments on one of the major problems that is encountered in archaeoseismology. When confronting historical earthquake catalogues with archaeological and geological evidence—all inherently incomplete—it should be clear that an ancient earthquake is not necessarily present in all records. A forced correlation will inevitably lead to circular reasoning. Rucker and Niemi state that the latter can only be avoided by independent supporting evidence and a clear assessment of the level of uncertainty of an earthquake hypothesis. Only then can archaeological earthquake evidence—objectively revealing unknown ancient earthquakes—provide useful data for seismic hazard assessment and mitigation. Also the paper by Bilham et al. holds a cautionary commentary on earthquake hypotheses that are postulated solely based on structural damage to buildings. By means of the fortuitous existence of photographs taken of the same place from the same viewpoint over time, the authors were able to use the tenth-century Shiva Temple at Pandrethan as a strong-motion seismometer in an attempt to elucidate the earthquake history of the Kashmir valley in the past millennium. Their analysis shows that an earthquake hypothesis cannot be retained to explain the damage recorded on the series of photographs of the temple taken at different times during the past 200 years. The paper by Kovach et al. focuses on the Harappan civilization in the Indus Valley region (Pakistan, India). Their survey clearly shows that this part of the Alpine-Himalayan seismic zone has great potential for archaeoseismology. But the archaeoseismological research is confronted with the challenges to incorporate secondary earthquake effects, such as changes in the fluvial systems and coastal elevation. The three following papers in this section present the results of different novel approaches, clearly offering new perspectives for future archaeoseismological research. These papers frame in our collective search for shared protocols and standardized methodologies. Grützner et al. use their extensive archaeological, geomorphological, and geological database in and around the archaeological site of Baelo Claudia—a Roman city along the southwestern Iberian coast—to compare the logic tree approach for paleoseismology, developed by Atakan et al. (2000), and for archaeoseismology, designed by Sintubin and Stewart (2008) (the latter is also applied in the paper by RodríguezPascua et al.). Both approaches try to express semiquantitatively the level of confidence with respect to an earthquake hypothesis. This comparison reveals a number of strengths and weaknesses of both approaches. In their paper Tendürüs et al. tackle another interesting problem of archaeoseismology. It concerns the question of whether or not the spatial distribution of archaeological remains and their preservation conditions can be correlated to seismic activity. The Tendürüs et al. modeling approach calculates the cumulative effect of continuing seismic activity for a period of 100 years and expresses it as cumulative peak ground acceleration. With respect to the build environment the focus is shifted from the effects of single earthquake events to a continuing deterioration. The model is applied to the island of Zakynthos in western Greece. It results in a promising correlation between the distribution and preservation of archaeological remains and seismic activity. Finally, Yerli et al. explore the applicability of ground-based LIDAR (Light Detection and Ranging) as a new technique in our pursuit of a more quantitative archaeoseismology. They have chosen the Roman theater in the ancient city of Pınara in southwestern Turkey as a case study. Ancient earthquakes have indeed been evidenced at this archaeological site. The detailed LIDAR mapping reveals a particular damage pattern (e.g., back-tilting of seating rows), that would otherwise pass by unnoticed. Yerli et al. suggest that the Roman theater may have recorded fault-block rotation related to activity during Roman times on a nearby fault.

Preface PRACTICES IN ARCHAEOSEISMOLOGY This section consists of seven papers illustrating the diversity of practices in archaeoseismology. Each archaeological site has unique characteristics, related to both the archaeological and seismotectonic context, so that a specific approach is required each time. These papers illustrate how archaeological sites may serve as seismoscopes recording ancient—to date unknown—earthquakes. The first three cases concern archaeological evidence for damage caused by earthquake-related ground shaking. The paper by Rodríguez-Pascua et al. focuses on the ancient settlement of El Tolmo de Minateda in southeastern Spain. This site has been occupied for ca. 3800 years, from the Late Bronze age to the Visigoth and Islamic periods. This site is situated in an active intraplate region characterized by a very low frequency of major seismic events. Archaeological sites are thus a primary tool to discover major ancient earthquakes, contributing to a correct assessment of the seismic hazard in the region. The integration of archaeological evidence with unique geomorphological evidence—rock falls including Visigoth carved tombs—reveals at least two major seismic events in the seventh to tenth century A.D. Kázmér and Major discuss in their paper earthquake hypotheses for typical structural damage and repairs in the well-preserved thirteenth-century fortification of the Al-Marqab citadel in Syria. They conclude that the damages observed can be associated with major ground shaking caused by at least two separate earthquake events. The major damage is attributed to the 1202 earthquake, suggesting a local intensity of VIII–IX at the site of the citadel. The latter is higher than assumed before, inferring an increase of calculated magnitude for that particular historical earthquake. In the paper by Karakhanyan et al. the results of an extensive study of the damage patterns on the worldfamous Colossi of Memnon in the temple of Amenhotep III at Luxor in Egypt are presented. An earthquake hypothesis is developed based primarily on structural damage characteristics, but furthermore supported by evidence of massive liquefaction exposed by paleoseismological trenching. The authors’ findings are seemingly in contradiction to the lack of any clear earthquake account in 3500 years of papyri and epigraphic sources, with the exception of the 27 B.C. earthquake described by Strabo. The earthquake evidence cannot, however, be correlated with Strabo’s earthquake, but is indicative of a major earthquake that struck the region of Luxor between 1200 and 900 B.C. The widespread destruction of the temple suggests a nearby—to date unknown—active fault, capable of producing major—potentially hazardous—earthquakes. The following four papers deal with displaced archaeological remains that serve as cultural piercing points from which to derive fault displacement, identify individual surface-rupturing seismic events, and estimate long-term slip rates. The cases illustrated in the paper by Barreca et al. are found in Late Roman sites in central-northern Sicily, Italy, a region characterized by moderate instrumental and historical seismicity. The archaeological evidence consists of a ruptured votive niche, the prevalence of pottery pieces and coins dated to the fourth century, a sudden decrease in human activity by the end of the fourth century, and an atypical Late Roman grave. This evidence is correlated to the historical 361 A.D. earthquake in central Sicily. Barreca et al. consider their results as a first step in better constraining the epicentral area of this historical earthquake. Aqueducts are very interesting man-made structures when it comes to identifying capable faults and deriving slip rates, primarily because of their linear nature extending over long distances across a seismic landscape and their particular slope, guaranteeing sufficient hydrological head. The use of aqueducts as archaeoseismological tools is convincingly demonstrated in the following two papers. Galli et al. present the results of a geological, geophysical, and geodetic survey of a first-century B.C. aqueduct in southern Italy. The cumulative displacement infers at least two surface-rupturing seismic events after the construction of the aqueduct. No historical record exists for either event. The Al Harif Roman aqueduct in Syria is a unique archaeoseismological site, because it crosses the Dead Sea transform fault—i.e., the plate boundary between the African and Arabian plate. The aqueduct exhibits a left-lateral offset of more than 13 m since Roman times. Moreover, calcareous deposits associated with the overflowing of carbonate-saturated waters can be used as an extra dating tool in the reconstruction of the earthquake history. Sbeinati et al. combine the extensive archaeological and paleoseismological evidence with the specific dating evidence from the tufa deposits to reconstruct the succession of surfacerupturing events between 63 B.C. and 1170 A.D., the last faulting event affecting this segment of the Dead Sea transform.

ix

x

Preface Finally, the paper by Yönlü et al. discusses displaced relics in the ancient city of Priene and the deformation of an Ottoman bridge, both situated along the northern margin of the active Büyük Menderes graben in southwestern Turkey. Detailed field observations and LIDAR mapping show a significant right-lateral offset on top of the expected normal component of the boundary fault system of the graben. This volume gives a nice sample of our ongoing efforts with respect to the search for ancient earthquakes. It shows the diversity of approaches and the wide range of disciplines involved, but also the potential of archaeoseismology with respect to a better understanding of the earthquake history in many places around the world that are threatened by seismic hazards. We hope this volume offers a taste of the complexity with which archaeoseismologists are confronted, from interpreting structural damage patterns and secondary earthquake phenomena on archaeological sites to assessing the historical, political, and social context in which the ancient earthquake occurred. In this respect, we hope to have been able to arouse an interest for archaeoseismology in the broader community of earth scientists, seismologists, historians, and archaeologists, one of the primary aims of the International Geoscience Programme IGCP 567. 13 September 2010

REFERENCES CITED Altez, R., 2010, this volume, New interpretations of the social and material impacts of the 1812 earthquake in Caracas, Venezuela, in Sintubin, M., Stewart, I.S., Niemi, T.M., and Altunel, E., eds., Ancient Earthquakes: Geological Society of America Special Paper 471, doi: 10.1130/2010.2471(05). Ambraseys, N.N., 2005, Archaeoseismology and neocatastrophism: Seismological Research Letters, v. 76, no. 5, p. 560–564, doi:10.1785/gssrl.76.5.560. Ambraseys, N.N., Jackson, J.A., and Melville, C.P., 2002, Historical Seismicity and Tectonics: The Case of the Eastern Mediterranean and the Middle East, in Lee, W.H.K., Kanamori, H., Jennings, P.C., and Kisslinger, C., eds., International Handbook of Earthquake & Engineering Seismology: International Geophysics Series 81A: Academic Press, Amsterdam, p. 747–763. Atakan, K., Midzi, V., Moreno Toiran, B., Vanneste, K., Camelbeeck, T., and Meghraoui, M., 2000, Seismic hazard in regions of present day low seismic activity: Uncertainties in the paleoseismic investigations along the Bree Fault Scarp (Roer Graben, Belgium): Soil Dynamics and Earthquake Engineering, v. 20, p. 415–427, doi: 10.1016/S0267 -7261(00)00081-6. Barnes, G.L., 2010, this volume, Earthquake archaeology in Japan: An overview, in Sintubin, M., Stewart, I.S., Niemi, T.M., and Altunel, E., eds., Ancient Earthquakes: Geological Society of America Special Paper 471, doi: 10.1130/2010.2471(08). Barreca, G., Barbano, M.S., Carbone, S., and Monaco, C., 2010, this volume, Archaeological evidence for Romanage faulting in central-northern Sicily: Possible effects of coseismic deformation, in Sintubin, M., Stewart, I.S., Niemi, T.M., and Altunel, E., eds., Ancient Earthquakes: Geological Society of America Special Paper 471, doi: 10.1130/2010.2471(18). Bilham, R., and Lodi, S., 2010, this volume, The door knockers of Mansurah: Strong shaking in a region of low perceived seismic risk, Sindh, Pakistan, in Sintubin, M., Stewart, I.S., Niemi, T.M., and Altunel, E., eds., Ancient Earthquakes: Geological Society of America Special Paper 471, doi: 10.1130/2010.2471(03). Bilham, R., Singh, B., Bhat, I., and Hough, S., 2010, this volume, Historical earthquakes in Srinagar, Kashmir: Clues from the Shiva Temple at Pandrethan, in Sintubin, M.,

Stewart, I.S., Niemi, T.M., and Altunel, E., eds., Ancient Earthquakes: Geological Society of America Special Paper 471, doi: 10.1130/2010.2471(10). Caputo, R., and Pavlides, S.B., 2008, Earthquake Geology: Methods and Applications: Tectonophysics, v. 453, 296 p. Evans, A., 1928, The Palace of Minos, part II: London, 844 p. Force, E.R., and McFadgen, B.G., 2010, this volume, Tectonic environments of ancient civilizations: Opportunities for archaeoseismological and anthropological studies, in Sintubin, M., Stewart, I.S., Niemi, T.M., and Altunel, E., eds., Ancient Earthquakes: Geological Society of America Special Paper 471, doi: 10.1130/2010.2471(02). Galadini, F., Hinzen, K.-G., and Stiros, S.C., 2006, Archaeoseismology at the beginning of the 21st century: Journal of Seismology, v. 10. Galli, P.A.C., Giocoli, A., Naso, J.A., Piscitelli, S., Rizzo, E., Capini, S., and Scaroina, L., 2010, this volume, Faulting of the Roman aqueduct of Venafrum (southern Italy): Methods of investigation, results, and seismotectonic implications, in Sintubin, M., Stewart, I.S., Niemi, T.M., and Altunel, E., eds., Ancient Earthquakes: Geological Society of America Special Paper 471, doi: 10.1130/2010.2471(19). Grützner, C., Reicherter, K., and Silva, P.G., 2010, this volume, Comparing semiquantitative logic trees for archaeoseismology and paleoseismology: The Baelo Claudia (southern Spain) case study, in Sintubin, M., Stewart, I.S., Niemi, T.M., and Altunel, E., eds., Ancient Earthquakes: Geological Society of America Special Paper 471, doi: 10.1130/2010.2471(12). Karakhanyan, A., Avagyan, A., and Sourouzian, H., 2010, this volume, Archaeoseismological studies at the temple of Amenhotep III, Luxor, Egypt, in Sintubin, M., Stewart, I.S., Niemi, T.M., and Altunel, E., eds., Ancient Earthquakes: Geological Society of America Special Paper 471, doi: 10.1130/2010.2471(17). Kázmér, M., and Major, B., 2010, this volume, Distinguishing damages from two earthquakes—Archaeoseismology of a Crusader castle (Al-Marqab citadel, Syria), in Sintubin, M., Stewart, I.S., Niemi, T.M., and Altunel, E., eds., Ancient Earthquakes: Geological Society of America Special Paper 471, doi: 10.1130/2010.2471(16). King, G.C.P., and Bailey, G.N., 2010, this volume, Dynamic landscapes and human evolution, in Sintubin, M., Stewart, I.S., Niemi, T.M., and Altunel, E., eds., Ancient Earthquakes: Geological Society of America Special Paper 471, doi: 10.1130/2010.2471(01).

Preface Kovach, R.L., and Nur, A., 2006, Earthquakes and archeology: Neocatastrophism or science?: Eos (Transactions, American Geophysical Union), v. 87, p. 317–318. Kovach, R.L., Grijalva, K., and Nur, A., 2010, this volume, Earthquakes and civilizations of the Indus Valley: A challenge for archaeoseismology, in Sintubin, M., Stewart, I.S., Niemi, T.M., and Altunel, E., eds., Ancient Earthquakes: Geological Society of America Special Paper 471, doi: 10.1130/2010.2471(11). Laffaille, J., Audemard M., F., and Alvarado, M., 2010, this volume, San Antonio de Mucuñó, Mérida Andes, Venezuela: Relocation of a doctrine town following the 1674 earthquake, in Sintubin, M., Stewart, I.S., Niemi, T.M., and Altunel, E., eds., Ancient Earthquakes: Geological Society of America Special Paper 471, doi: 10.1130/2010.2471(04). Marinatos, S., 1939, The volcanic destruction of Minoan Crete: Antiquity, v. 13, p. 415–439. Nur, A., 2008, Apocalypse: Earthquakes, Archaeology, and the Wrath of God: Princeton, New Jersey, Princeton University Press, 309 p. Raphael, K., 2010, this volume, The impact of the 1157 and 1170 Syrian earthquakes on Crusader-Muslim politics and military affairs, in Sintubin, M., Stewart, I.S., Niemi, T.M., and Altunel, E., eds., Ancient Earthquakes: Geological Society of America Special Paper 471, doi: 10.1130/2010.2471(06). Rodríguez-Pascua, M.A., Silva, P.G., Garduño-Monroy, V.H., Pérez-López, R., Israde-Alcántara, I., Giner-Robles, J.L., Bischoff, J.L., and Calvo, J.P., 2010, this volume, Ancient earthquakes from archaeoseismic evidence during the Visigothic and Islamic periods in the archaeological site of “Tolmo de Minateda” (SE Spain), in Sintubin, M., Stewart, I.S., Niemi, T.M., and Altunel, E., eds., Ancient Earthquakes: Geological Society of America Special Paper 471, doi: 10.1130/2010.2471(15). Rucker, J.D., and Niemi, T.M., 2010, this volume, Historical earthquake catalogues and archaeological data: Achieving synthesis without circular reasoning, in Sintubin, M., Stewart, I.S., Niemi, T.M., and Altunel, E., eds., Ancient Earthquakes: Geological Society of America Special Paper 471, doi: 10.1130/2010.2471(09). Sbeinati, M.R., Meghraoui, M., Suleyman, G., Gomez, F., Grootes, P., Nadeau, M.-J., Al Najjar, H., and Al-Ghazzi, R., 2010, this volume, Timing of earthquake ruptures at the

xi

Al Harif Roman aqueduct (Dead Sea fault, Syria) from archaeoseismology and paleoseismology, in Sintubin, M., Stewart, I.S., Niemi, T.M., and Altunel, E., eds., Ancient Earthquakes: Geological Society of America Special Paper 471, doi: 10.1130/2010.2471(20). Schaeffer, C.F.A., 1948, Stratigraphie Comparée et Chronologie de l’Asie Occidentale: London, Oxford University Press. Sintubin, M., and Stewart, I.S., 2008, A logical methodology for archaeoseismology: A proof of concept at the archaeological site of Sagalassos, southwest Turkey: Bulletin of the Seismological Society of America, v. 98, no. 5, p. 2209– 2230, doi:10.1785/0120070178. Stefanakis, M.I., 2010, this volume, Western Crete: From Captain Spratt to modern archaeoseismology, in Sintubin, M., Stewart, I.S., Niemi, T.M., and Altunel, E., eds., Ancient Earthquakes: Geological Society of America Special Paper 471, doi: 10.1130/2010.2471(07). Stiros, S.C., and Jones, R.E., 1996, Archaeoseismology, in Whitbread, I.K., ed., Fitch Laboratory Occasional Paper: Athens, Institute of Geology & Mineral Exploration & The British School at Athens, p. 268. Tendürüs, M., van Wijngaarden, G.J., and Kars, H., 2010, this volume, Long-term effect of seismic activities on archaeological remains: A test study from Zakynthos, Greece, in Sintubin, M., Stewart, I.S., Niemi, T.M., and Altunel, E., eds., Ancient Earthquakes: Geological Society of America Special Paper 471, doi: 10.1130/2010.2471(13). Yerli, B., ten Veen, J., Sintubin, M., Karabacak, V., Yalçıner, C.Ç., and Altunel, E., 2010, this volume, Assessment of seismically induced damage using LIDAR: The ancient city of Pınara (SW Turkey) as a case study, in Sintubin, M., Stewart, I.S., Niemi, T.M., and Altunel, E., eds., Ancient Earthquakes: Geological Society of America Special Paper 471, doi: 10.1130/2010.2471(14). Yönlü, Ö., Altunel, E., Karabacak, V., Akyüz, S., and Yalçıner, Ç., 2010, this volume, Offset archaeological relics in the western part of the Büyük Menderes graben (western Turkey) and their tectonic implications, in Sintubin, M., Stewart, I.S., Niemi, T.M., and Altunel, E., eds., Ancient Earthquakes: Geological Society of America Special Paper 471, doi: 10.1130/2010.2471(21). MANUSCRIPT ACCEPTED BY THE SOCIETY 19 MAY 2010

Printed in the USA

Acknowledgments

The editors acknowledge the reviewers for their contribution to the quality of this volume. The reviewers were: A. Agnon (Hebrew University, Israel), H.S. Akyiiz (Istanbul Technical University, Turkey), N. Ambra­ seys (Imperial College London, UK), K. Atakan (University of Bergen, Norway), F. Audemard (Fundacion Venezolana de Investigaciones Sismologicas, Venezuela), B. Batten (Oberlin University, Tokyo, Japan), C. Beck (Universite de Savoie, France), Z. <;:akir (Istanbul Technical University, Turkey), R. Caputo (Uni­ versity of Ferrara, Italy), M. Cisternas (Universidad Catolica de Valparaiso, Chile), M. Daeron (LSCE, France), K. Decker (University of Vienna, Austria), J. Ebel (Boston College, USA), L. Guerrieri (APAT, Italy), E. Guidoboni (Instituto Nazionale di Geofisica e Vulcanologia, Italy), B. Helly (Maison de L'Orient Mediterraneen, Lyon, France), V. Hopgood (Greenhead College, UK), A. Rubert-Ferrari (Universite de Liege, Belgium), R. Jones (University of Glasgow, UK), V. Karakbacak (Eski§ehir Osmangazi Univer­ sity, Eski§ehir, Turkey), G. King (Institute de Physique du Globe, Paris, France), A. Korjenkov (National Academy of Sciences, Kyrgyzstan), I. Koukouvelas (University of Patras, Greece), R. Kovach (Stanford University, USA), M. Le Beon (Institute de Physique du Globe, Paris, France), J.N. Malik (Indian Insti­ tute of Technology, Kanpur, India), S. Marco ( Tel Aviv University, Israel), M. Meghraoui (Institute de Physique du Globe, Strasbourg, France), A. Michetti (Universita dell'Insubria, Como, Italy), J. Moody (University of Texas, USA), M. Mukul (Centre for Mathematical Modelling and Computer Simulation, Bangalore, India), R.M.W. Musson (British Geological Survey, UK), K. Reicherter (RWTH Aachen, Ger­ many), A. Salamon (Geological Survey, Israel), P.G. Silva (Universidad de Salamanca, Spain), S. Stiros (University of Patras, Greece), C. Vita-Finzi (National History Museum, London, UK), R. Yeats (Oregon State University, USA), and a number of anonymous reviewers.

xiii

The Geological Society of America Special Paper 471 2010

Dynamic landscapes and human evolution Geoffrey C.P. King* Laboratoire Tectonique, Institut de Physique du Globe Paris, 4 place Jussieu, 75252 Paris, France Geoffrey N. Bailey* Department of Archaeology, University of York, The King’s Manor, York, YO1 7EP, UK

ABSTRACT This paper discusses the relationship between dynamic landscape change resulting from tectonic activity and patterns of human land use and human development. Archaeological studies of human settlement in its wider landscape setting usually focus on climate change as the principal environmental driver of change in the physical features of the landscape, even on the longer time scales of early human evolution. Tectonic processes are usually assumed to operate too slowly to be of any significance except as the source of occasional disruptive events, or at best to have some indirect effect on climate change as a result of long-term regional uplift. Herein, examples are shown from Europe and Africa to illustrate the ways in which changes of significance to human settlement can occur at a range of geographical scales and on time scales that range from lifetimes to tens of millennia. We emphasize that these changes are not always or necessarily destructive in their impact but can also create and sustain attractive conditions for human settlement and that these conditions have exercised powerful selection pressures on human development.

periods of the archaeological record where the effect of earthquakes and volcanoes is most easily visible in terms of its impact on the built environment, and hence on the negative or destructive consequences of such activity. For the earlier periods of prehistory, the consequences of seismic activity rarely register in the archaeological or geoarchaeological record, except as occasional dramatic effects like tsunamis (e.g., Bondevik et al., 1997) or comparable catastrophic events such as major volcanic eruptions (e.g., Ambrose, 1998). Otherwise, reconstructions of regional settlement patterns and land use have tended to proceed on the assumption of an essentially static physical land surface, which changes only as a result of climatically induced variations in erosion, sedimentation, vegetation, and water supply. In this view, tectonics are relegated to the role of an episodic source of disruption or to a time scale of operation considered too large or too far back in time to be of relevance to human activity and human

INTRODUCTION Growth in our understanding of active tectonics and tectonic geomorphology during the past 25 yr has made it clear that tectonic processes happen at a variety of scales relevant to human history. These effects range from an individual earthquake at the shortest end of the spectrum to modifications of regional topography resulting from the cumulative effect of earthquake activity over centuries and millennia, and on a time scale of tens to hundreds or thousands of millennia, to more substantial modifications resulting in the creation of major mountain ranges and rifts. Interest in archaeoseismology naturally focuses on individual earthquakes or volcanic eruptions as the most obviously visible expression of tectonic processes on the human scale, on recent *E-mails: King: [email protected]; Bailey: [email protected].

King, G.C.P., and Bailey, G.N., 2010, Dynamic landscapes and human evolution, in Sintubin, M., Stewart, I.S., Niemi, T.M., and Altunel, E., eds., Ancient Earthquakes: Geological Society of America Special Paper 471, p. 1–19, doi: 10.1130/2010.2471(01). For permission to copy, contact [email protected]. © 2010 The Geological Society of America. All rights reserved.

1

2

King and Bailey

history or prehistory. Our aims in this paper are to show how processes of faulting and folding can modify landscape features to produce conditions inherently attractive to human settlement, even over short time spans or in conditions when rates of activity are modest, to review a range of examples from Europe and Africa that illustrate these effects, and to highlight their significance in the broader patterns of human evolution and dispersal. BACKGROUND El Asnam Earthquake A key source of evidence for changing views of earthquake processes and active tectonics is the 1980 El Asnam earthquake (Ms ~7.2) in Algeria (King and Vita-Finzi, 1981; King and Yielding, 1984; Philip and Megharoui, 1983). During the earthquake, the Sera el Maarouf ridge lifted ~5 m as a result of coseismic anticlinal folding (Fig. 1). This impeded the flow of the Chelif River, resulting in the development of a lake upstream of the fault/ fold front that could be identified from old maps as having existed

A

in the past. While the event was the result of motion on a buried reverse fault, reverse ruptures were only modestly expressed at the surface. On the other hand, normal surface breaks were extensive as a consequence of flexure due to the folding. These observations have had far reaching effects on our understanding. First, they demonstrated that motion on small faults such as those visible on the surface of the Ser el Maarouf ridge may not represent a “regional stress field” and that folds can be active and are evidence for seismic hazard (Stein and King, 1984). The seismogram of the event was shown to be the result of a series of subevents that could be correlated in space with the location of fault fold segments that had moved at the time of the earthquake; a direct correlation between the nature of fault slip at the time of the event and mappable surface features was demonstrated. Second, the distribution of these features also showed that the longerterm morphology was a cumulative consequence of earthquakes similar to that in 1980 repeating every few hundred years. Estimates of the rate of folding could be made from the presence of ceramics and radiocarbon dates in uplifted terraces of Vita-Finzi’s Younger Fill (Vita-Finzi, 1969), and could also be estimated from

C New Lake New Lake

α

αʹ

D M

B α

til

ti st lted er ian gra ar vels tif ac ts

ou

te

d

gr

av

els

αʹ

Figure 1. Features associated with the 1981 El Asnam earthquake. (A) A map of the region where the Chelif River cuts through the Sera el Maarouf anticinal ridge. A new lake and old maps show that a marshy region existed in the past. Reverse faulting surface breaks could be identified and mapped, but subparallel normal faulting was more extensive. (B) A section though the anticline from α to αʹ as indicated in A. The extensional normal faulting resulted from the flexure associated with the folding. Tilted gravels outcrop extensively along the front of the anticline. (C) A view of the new lake and the Sera el Maarouf ridge. (D) Tilted gravels. The gravels rotated by slip along bedding planes (bookshelf faulting). This can be seen most clearly at the surface.

Dynamic landscapes and human evolution the presence of Mousterian artifacts in tilted gravels along the anticline front. Vita-Finzi (1969), working some years earlier in the same region, had demonstrated the effect of climatic controls on the behavior of the Chelif River sediments. Between about A.D. 500 and A.D. 1500, the period of Vita-Finzi’s Younger Fill, the river was aggrading, producing a fertile sediment-filled valley with a high water table. At the end of this episode, the river began to incise through the earlier sediments, the water table dropped, and the fertility of the valley reduced dramatically in consequence. However, this transition did not occur uniformly throughout the river basin. On the contrary, the river on the upstream side of the El Asnam anticline remained continuously depositional and continuously fertile. This example demonstrates how localized folding and faulting transverse to the main axis of drainage can impede water flow, sustain local conditions of fertility, and moderate the impact of climatically induced changes in hydrology.

20°E

,

It was appreciated retrospectively that the extensive evidence of human activity in the earthquake region in Mousterian times was not simply a convenient source of evidence for dating earthquake repeat times. It provided evidence for the attractions of a fertile and well-watered basin for human settlement in an otherwise relatively arid region as a consequence of the repeated uplift of the anticline, the damming back of water and sediments, and the long-term maintenance of lacustrine or marshy conditions (Megraoui and Doumaz, 1996) favorable to plant and animal life and human subsistence. Paleolithic Landscapes of Epirus Since the 1960s, archaeological projects centered on the Epirus region of northwest Greece (Fig. 2) have been dedicated to developing methods of analyzing, at a regional scale, the relationship between archaeological sites occupied during the Paleolithic

$

21°E

.

$ / % $

1

N

3

(

Klithi

=$*25,

40°N

5

. <

5

(

$

3

,

5

Grava

Kastritsa 8 6

Chiliomodi Chiliomodi

,RQLDQ6HD

Morfi

Karvounari Asprochaliko Kokkinopilos

OPEN AIR SITE

Stefani

ROCK SHELTER ³PHWHUV

39°N

³PHWHUV ³PHWHUV RYHUPHWHUV





*XOIRI$UWD

NLORPHWHUV

Figure 2. The Epirus region of northwest Greece. Sites showing substantial evidence of Paleolithic human occupation are named. Circles indicate open-air sites, and triangles indicate rock-shelter sites with substantial deposits containing both stone artifacts and bone material. The Asprochaliko and Kokkinopilos sites are discussed in the text and referred to in Figures 3 and 4. The present coastline is shown and the region of seafloor that would have been land when sea level was 100 m lower is indicated in white.

4

King and Bailey

period and their local and regional environmental setting (Bailey, 1997). Mapping of Paleolithic rock shelters and open-air sites in relation to topographic features and Quaternary deposits, and interpretations of landscape and vegetational change, was central to the original research strategy. Major deposition and erosion of Quaternary sediments had already been documented in an earlier phase of investigation (Dakaris et al., 1964; Higgs and Vita-Finzi, 1966; Higgs et al., 1967), including the famous “red beds,” many of which contained Paleolithic artifacts, and seemed to refer to an earlier land surface that had been modified subsequently by geomorphological change. The most famous of these, the site of Kokkinopilos (literally red clay), with its Paleolithic artifacts eroding out of deep gullies in massive beds of red sediment (Fig. 3), had generated considerable interest and controversy about how and when such a thick deposit of fine-grained sediments had come to be deposited on an elevated interfluve, whether by the action of wind or water, and the timing and causes of the subsequent erosion. Tectonics had not yet entered the picture, in retrospect, a surprising omission given the seismic activity of the region (King et al., 1983), but once tectonic influence was appreciated, it not only suggested a solution to the Kokkinopilos puzzle (Fig. 4), but it indicated that other Paleolithic sites in the region were typically

A

Kokkinopilos

B

associated with tectonic controls similar to the El Asnam region of the Chelif River (King and Bailey, 1985). The rock shelter site of Asprochaliko (Fig. 3B), for example, is located in a narrow gorge between the edge of an extensive plain to the south and a smaller sedimenting valley to the north. With the longest Paleolithic sequence of the region, extending over some 100,000 yr, and containing Mousterian, Upper Paleolithic, and probably later activity, Asprochaliko was clearly a place of repeated occupation, suggesting a preferred location with enduring attractions for human visitation and activity over a long period of time. One obvious attraction of the site lies in its location in a narrow gorge close to a major animal migration route between an extensive area of coastal lowland and the regional hinterland. A tectonic perspective highlights this feature, because repeated earthquake activity not only sustained and renewed the fertility of fault-bounded basins, it also accentuated and maintained topographic barriers that could be used to tactical advantage in the monitoring, trapping, and control of mobile prey species. Extension of this theme more widely in the region demonstrated a significant relationship between archaeological sites and the juxtaposition of topographic barriers and fertile basins, both at a local scale and at a wider regional

Asprochaliko

C

Asprochaliko

Kokkinopilos

beds

Figure 3. Photographs of the Asprochaliko and Kokkinopilos site regions. (A) The red beds that form the Kokkinopilos deposit. They can be seen to be gently folded. The limestone is folded, but because it lacks clear bedding, this is not easily appreciated at this scale (see King et al., 1993). (B) The Asprochaliko rock shelter. (C) An overview showing the Kokkinopilos beds and the Louros River valley. The anticlinal uplift axis is approximately in the same direction as the photograph (King et al., 1993).

Dynamic landscapes and human evolution

Arta plain deposition

A

5

C

deposition Uplift n u tti

w n -c

fo

g

Asprochaliko

do

approximate fold axis

ed ld

ds be

d rie bu ult fa

Kokkinopilos beds Uplift

5 kilometers

Present-day deposition

Elevation (m)

Ancient uplifted deposits

D

B 160 240 deposition 80

River Louros Up l

ift down cut tin g

deposition

Figure 4. Changes to the region around the Asprochaliko and Kokkinopilos sites due to tectonic activity (anticlinal folding) over ~100 k.y. (A) Map of the site region identifying the Kokkinopilos open-air and Asprochaliko rock-shelter (gray symbol) sites. Regions presently experiencing deposition are indicated in green, and uplifted sediments including the Kokkinopilos beds are shown in orange. (B) The river profile is shown and can be seen to be disturbed by ongoing tectonic activity in a similar way to that found for the Chelif River associated with repeating earthquakes of El Asnam type. Uplifted terraces and terrace fragments can be traced up to the level of the present Kokkonopilos beds, which in the past were at the same level as the current Arta plain. (C) A simplified figure showing the present relation among the Kokkinopilos beds, the Asprochaliko rock shelter, and the present river cut into a narrow gorge. The deformation process appears to be associated with buried faulting indicated on the figure. (D) The environment of the site when it was occupied.

scale, showing how the Paleolithic inhabitants of the region benefited from a tectonically active environment and exploited its features (Bailey et al., 1993; King et al., 1994, 1997). The fact that repeated seismic activity is not uniformly beneficial is demonstrated by the Kokkinopilos site, where a long-standing basin of deposition was eventually transformed into an elevated and eroded badlands landscape. General Principles The Epirus example highlights three long-term advantages of a seismically and tectonically active landscape. The first is the rejuvenation of basins with sedimentation and well-watered fertility as originally identified in the El Asnam region. The second is the concomitant creation and accentuation of barriers that offer tactical advantage to human predators dependent on mobile and elusive prey. A third advantage is ongoing creation of a landscape that is not only topographically complex, but one that is often ecologically complex also, offering a variety of resources that may further enhance the attractions for human settlement.

The Epirus and El Asnam examples come from tectonic regions that are undergoing compression near plate boundaries, with reverse faulting and progressive uplift. However, similar features are present in normal faulting and strike-slip environments, and, in these cases, volcanic activity can add an additional contribution to topographic complexity and soil fertility. In both cases, faulting or folding create localized basins of sediment accumulation with high water tables and good water supplies, alternating with a complex uplifted and folded topography made up of barriers and enclosures formed by fault scarps and lava flows that can be used to gain protection and tactical advantage in the control of the wider landscape (Fig. 5). As long as tectonic activity persists over time, these features are regularly rejuvenated. However, if activity ceases, then the topographic complexity in time becomes smoothed by erosion, barriers are removed or reduced, and the water table drops, making water less easily available and more vulnerable to variations in precipitation resulting from climate change. Regions of little or no tectonic activity are thus more vulnerable to the vagaries of climate change (Fig. 6), suggesting the paradox that the geological instability of a tectonically

A

river terraces

wet and marshy, subsidence

folding, dry, uplifted gorge

lake

ditions wet con ditions dry con

Typical horizontal scale 5 - 25 km

B

gorge dry, uplifted flank

lake river terraces

wet and marshy, subsidence

ditions wet connditions dry co

Figure 5. Schematic illustration of the effects of tectonic activity creating landscape features illustrated by examples in this paper. (A) Features typical of contractional (reverse-faulting) environments, where the surface expression of faulting appears in the form of folds as well as faults that cut the surface. (B) Features typical of extensional (normal-faulting) environments. These produce similar features as in A, but the faults usually fully cut the surface.

A

wet conditions dry conditions

B

wet conditions dry conditions

Figure 6. Schematic illustration of effects of changing water table in flat landscapes. (A) In regions of low relief, pans, water holes, and lakes are sensitive to changes of climate and can become dry if the water table drops. (B) Rivers in regions of low relief are also sensitive to changes of water table resulting from climate change and can be perennial or completely dry. If there is a reliable source of water from a high rainfall headwater region, such rivers can be more reliable.

Dynamic landscapes and human evolution active landscape may actually create more stable conditions for long-term human settlement than an environment that is geologically inactive. AFRICA The African Rift The African Rift is one of the largest and longest-lived onland tectonic structures in the world, and the fact that it is also home to some of the most concentrated evidence of archaeological sites and human fossils relating to the earliest stages in human evolution suggests a prima facie case for the investigation of tectonic factors (Fig. 7). However, there has been almost no consideration of the impact of tectonics on evolutionary processes, except in terms of their indirect influence on regional climate change, or the possible effect of geological instability and rifting in preserving archaeological and fossil materials in rapidly accumulating sediments and then exposing them to discovery through rifting and erosion. The consideration of external forcing factors on human evolution has almost invariably focused on climate and vegetational change (Maslin and Christensen, 2007) and more rarely on the potential effect of tectonic or volcanic catastrophes. From our initial studies in Europe and the Mediterranean, it was a small step to the hypothesis that the advantageous features of dynamic landscapes molded by active tectonics were uniquely well developed in an environment such as the African Rift, and might therefore have played a significant selective role in shaping the human evolutionary trajectory (Bailey et al., 2000, 2010; King and Bailey, 2006). The African Rift creates landscapes of complex topography par excellence. The rifting process itself results in the progressive rejuvenation of the rift floor and uplift of the rift flanks and high levels of earthquake and volcanic activity, resulting in vertical fault scarps and lava flows. Uplift and volcanic activity generate high volumes of erodible material, and internally draining basins trap eroded sediments and water to create fertile lake basins. Fault scarps and lava flows afford safe areas for the protection of young and a tactical advantage in the avoidance of predators or the capture of animal prey. Such features offer a niche for an unspecialized but increasingly intelligent omnivore, providing potential selective advantage in the development of features such as bipedalism, which facilitates scrambling and climbing over barriers that deter most quadruped mammals, an omnivorous diet in which meat-eating assumes greater prominence, increased brain size, and the delayed onset of maturity. In short, we argue that a tectonically active environment acted as an evolutionary pacemaker in developing human advantage in the competition for resources against faster moving or more specialized species. One of the difficulties of pursuing this hypothesis is that levels of landscape change were so rapid and so dramatic in many parts of the African Rift that it is often quite difficult if not impossible to reconstruct the topography as it existed in the vicinity of fossil and archaeological locations that were formed far back in

7

time, as, for example, with the early sites in Ethiopia, such as those of the Awash River Valley in the Afar (region A of Fig. 8; Johanson et al., 1982; White et al., 2003; Semaw et al., 2003). Despite the many reconstructions of vegetation and environmental conditions in the immediate vicinity of these sites (De Heinzelin et al., 1999; Johanson et al., 1982; Kalb et al., 1982a, 1982b; Radosevich et al., 1992; Semaw et al., 2005; WoldeGabriel et al., 1994, 2001), it is not possible to identify the wider configuration of environmental conditions over a larger area except in very general terms, or their relationship to topographic variables, because too much has changed in the past 2 to 3 m.y. At the time when these early sites were occupied, their location was close to the point where the Awash River entered the active rift floor. Today, they are far from the active rift as a result of progressive uplift. Conditions analogous to those of Pliocene-Pleistocene times can be found at the present day in the area where the Awash River emerges onto the currently active rift floor (region B of Fig. 8). A suitable analogue lies in the area of the Karub volcano in the active rift (Figs. 9 and 10). This region includes a range of environments: an annual lake (shaded gray in Fig. 9B), wetland and swampland (associated with the Awash River) that is now flood-controlled for agriculture (dark green in Fig. 9B), and many smaller zones of grassland currently exploited by modern shepherds (light green in Fig. 9B). The Afar region has in the past been more humid (Gasse, 2001), with a freshwater or slightly brackish lake at >6 ka, that may have been continuous with, or linked to, the present-day lakes Adobada and Abbe. The smaller area in the vicinity of the Gablaytu volcano (Fig. 10) is dissected by active faults that create vertical cliffs, enclosed fertile valleys, and blocky lava flows. The latter are exactly the sorts of features that would have provided some measure of protection against cursorial carnivores and secure areas for vulnerable hominin young. In the locations where Pliocene-Pleistocene fossils are found (region A of Fig. 8), important details of landscape features have now been eroded and smoothed over time, but these details are still clear and uneroded in the analogue are of region B. Features shown there are very common when a rift is active and disappear when activity ceases. The young geological processes along the active rift zone, well illustrated by the case study around the Karub volcano, demonstrate the fragmented nature of this environment and highlight its implications for human development. However, the rate of geological change is so fast that equivalent parts of the rift when early hominids were present are now located some distance away from the active zone, and the evidence of tectonism, and the key structures that once accommodated it, is muted or obscured. Ironically, we need to shift attention to South Africa, where equally abundant finds of early fossils and archaeological materials occur, but where the rate of geological action is less intense, and there is a better opportunity to capture the association between early human sites and tectonic structures. Although it has long been inferred that this southern region is devoid of neotectonic or modern-day seismic activity, this is not the case. Though less pronounced than in highly active settings, the tell-tale signs of

8

King and Bailey

Volcanoes active in the last 10,000 yr Hadar Middle Awash Melka Kunture

Historically active volcanoes Major hominin sites Omo West Turkana Lothagam

Lainyamok Olduvai Gorge ~ 500 km

Fejej East Turkana Lake Baringo Sites

Peninj Laetoli

Uraha Kanapoi

Makapansgat Sterkfontein region

Border Cave

Taung Florisbad Hofmeyer Boomplaas Figure 7. Major sites in Africa.

recent tectonism can still be found in the landscape, as is evident from the following case studies. South Africa South Africa has a history of early hominin occupation that extends back into the Pliocene, with a number of famous fos-

sil finds including Australopithecines and specimens of early Homo, notably in the Cradle of Humankind region north of Johannesburg, at sites such as Makapansgat, Sterkfontein, and Taung, together with later sites of middle and late Pleistocene ages (region A of Fig. 8; Mitchell, 2002; Partridge, 2000). Border Cave is closely associated with the southern end of the African Rift, and Hofmeyer Site shows some evidence of tectonic

Dynamic landscapes and human evolution

9

Re d Se a

12°30’N of Djibouti Gulf n e d A Addis Ababa Dire Dawa

12°N

B Rive

r Aw ash

Mille

11°30’N

Lake Gamarri

River

A

Lake Abbe

River Awash

11°N

Figure 8. The Afar region of Ethiopia. The area shaded in light brown is the currently active African Rift. Rectangle A shows the region of the fossil sites of the Middle Awash River. It is located on the now-uplifted flank of the active rift, but at the time when the sites were being formed, it was located in the active rift margin (see lower inset). Rectangle B is the region chosen as our analogue in the present active rift margin (Ayele et al., 2007; Manighetti et al., 1998, 2001; Rowland et al., 2007).

ca. 5 Ma Rift opening Present

40°30’E

0

41°E

activity in the region, but only one river-abraded fossil has been found there. In mining regions, earthquakes can be frequent and large enough to be destructive, but it is commonly assumed that this is due to mining operations, and that in their absence activity would be minimal. Otherwise, there is little reported evidence of tectonic activity, or none that has been recorded. However, the South African seismic network shows a steady level of activity in areas that have no mining activity, and destructive events have been recorded prior to the use of instrumental records (Saunders

20

40 km

41°30’E

et al., 2008). As discussed later herein, major earthquakes have left their mark on the geomorphology of regions where many of the South African sites occur, notably at Makapansgat and Taung, although this has not previously been documented. Ongoing tectonic activity is less clear for Sterkfontein although important indicators are described here and are subject to an ongoing study (Dirks et al., 2010). Similarly, Florisbad is located in an area of subdued topography lacking the topographic features typical of active tectonics, although the hot springs that emerge at the

10

King and Bailey

A

Figure 9. The analogue region situated in the active rift margin (B in Fig. 8). The features found in this region are characteristic of the active rift and could be illustrated in other places. (A) Landsat thematic mapper image (bands 2, 4, and 7 are red, green, and blue, respectively) draped over exaggerated SRTM 3 digital elevation data to give a three-dimensional (3-D) effect of the area in the vicinity of Karub volcano. (B) Interpretation of A showing salient landscape features: a wetland where the Awash River enters the plain (dark green), smaller plains (light green), and a region that hosts an occasional lake under present climatic conditions and a permanent lake prior to ca. 6 ka (Gasse, 2001). A white dashed line outlines the area shown in Figure 10.

foreground scale

~5 km N

B agriculture and swamp land

occasional lake Karub volcano

earlier lake > 6 ka

grassland

site are believed to be related to tectonic activity, and the wider region is seismically active (Douglas, 2006; Kuman et al., 1999). The later site of Boomplaas clearly takes advantage of a favorable environment created by ongoing activity. In other words, all inland sites are associated in some way with tectonically created and maintained features. For the present, coastal sites are excluded because of the complications associated with sea-level variation and coastal change, and we turn to a closer examination of some of the key inland sites next. Makapan Valley The Makapan Valley contains a series of caves dating to various periods within the Pliocene-Pleistocene and into historic

times, such as the Buffalo Cave, Cave of Hearths, and the famous Makapansgat Limeworks (Fig. 7). From the Limeworks Cave Member 3 and 4 breccias (dated at 3.2–2.7 Ma), 27 fossil specimens of Australopithecus africanus have been recovered (Tobias, 2000). Hominins and cercopithecine monkeys were accumulated by various predators, including hyenas, birds of prey, and carnivores (Reed, 1997). The later deposit, Member 5 contains a small assemblage of mammalian fossils but no hominin specimens. A three-dimensional view highlights the topographic complexity of the Makapan Valley region (Fig. 11A), and faults and other geomorphological features are interpreted in Figures 11B and 11C, with close-up detail shown in Figure 12. In the south (upper part of Figs. 11A and 11B), there is a large plain, now

Dynamic landscapes and human evolution

11

C

A

D

foreground scale

~500 m

B

ve wi ry r th ece crater lake ve nt rti fa ca ul D l s ts ca C rp s photographs E F

narrowing valley

smaller plain

E

Figure 10. Close-up of the region indicated in Figure 9. (A) Modified Google image showing a crater lake, volcanic lavas, and fault scarps in the vicinity of the Gablaytu volcano. (B) Interpretation of features shown in A, highlighting the Gablaytu crater lake, a narrow valley confined by lava flows and faults, and associated plains of varying size. (C–F) Photographs of features indicated in B, showing, respectively, major fault scarps (C), crater lake (D), and valleys bounded by relatively impassable lava flows and fault scarps (E) and (F).

F

large plain

exploited for agriculture, which is down-dropped with respect to the mountains in the foreground by an active fault system identified as fault Alpha (α). A contemporary feature of this plain is the Nylsvlei wetland, which has long been considered to be possibly related to continued tectonic activity (McCarthy and Hancox, 2000; Wagner, 1927). The active fault scarp probably associated with the creation and maintenance of the wetland forms a geomorphic feature shown in Figure 11C. A second fault, Beta (β), can be identified with the River Nyl running close to its base. The asymmetry of the valley could indicate continued tectonic activity or simply a shift of the river course to the west by sediment that reaches the valley from the east (left). A third fault, Gamma (γ), passes close to the site. It shows unequivocal evidence for ongoing activity (Fig. 12) and is responsible for uplift and consequent down-cutting and sedimentation close to the site, i.e., conditions corresponding to the model outlined in Figure 5B.

The immediate vicinity of the site is characterized by a gorge and associated steep cliffs and rough terrain; these features would have afforded important opportunities for safety and security. The river and the fertile, well-watered, sedimentary plains would have offered good foraging areas nearby. If, as Anton et al. (2002) argued, the foraging range of australopithecines was as small as 38 ha (essentially the area within ~500 m of a given point in the landscape), then this localized combination of rough terrain and productive resources would have been a key feature of their local environment. A viable breeding population would of course require a larger territory, so that the wider area around the Makapan Valley would also be important to long-term viability. The wider area reveals a combination of smaller plains, large open plains, and wetlands within ranging distance of the valley itself. In addition, a variety of valley conditions ranging from dry to marshy would have offered a range of habitats. These features are consistent with on-site indicators suggesting high biodiversity

12

King and Bailey

A Site

~ 5 km

N

B

Nylsvlei wet land large plain

several fault branches down dropped

Site

fault α

A

up down

fault γ back-til te

d

medium plain small plain 1

small plain 2

C

do wn

up

fault β

Figure 11. Region around the Makapan Valley, Northern Province, South Africa. (A) Landsat thematic mapper image (bands 2, 4, and 7 are red, green, and blue, respectively) draped over exaggerated digital elevation data to give a three-dimensional (3-D) effect. A vertical exaggeration of between 5 and 10 times the standard elevation is used for oblique images, which enhances the visual interpretation of high-lying versus low-lying regions. The effect of this exaggeration gives an impression similar to that of a land-based observer viewing the topography. (B) Faults α, β, and γ are identified. Recent activity of α is partly demonstrated by the presence of the Nylsvlei wetland, which results from a perturbed river (McCarthy et al., 2004). (C) A fault scarp typical of repeated earthquakes is associated with fault α. Fault β may be active, since the river is displaced to the west side of the valley, but river displacement could also be due to sediment sources coming from the east. Fault γ close to the site is clearly active (see caption to Fig. 12).

24.8°S, 28.43°E fault scarp

fault scarp

and diverse habitat conditions within close range of the site, including both wetland (C3) and dry-land (C4) indicators (Cadman and Rayner, 1989; McKee, 1999; Reed, 1997; Sponheimer and Lee-Thorp, 1999; Vrba, 1982). Taung The lime-mining quarry at Taung was the site of the discovery of the type specimen of Australopithecus africanus in 1924 (Dart, 1925), though dating based on faunal correlations suggests that it is younger than other australopithecine sites, with an age of ca. 2.6–2.4 Ma (McKee, 1993; Partridge, 2000; Tobias, 2000).

Much of the original hominin-bearing tufa deposit was destroyed by mining processes before systematic excavation could be undertaken, so that on-site paleoenvironmental data are lacking, and interpretation depends solely on interpretation of landscape features (Fig. 13). Two faults on either side of the Taung region create a rift valley (graben) with uplifted, drier flanks on each side and a downdropped, sedimented plain in the center (Fig. 13). As can be seen in the foreground, this fertile, sedimented plain is today being used for agriculture. The faulting has down-dropped the valley, causing rivers to cut into the uplifted valley sides. These faults

Dynamic landscapes and human evolution

13

A

Site

lt

α fau

small plain 2

small plain 1 foreground scale

~1 km

B

View from site to small plain 1

C

View from site up valley

crosscut earlier geological structures and are not controlled by them. Two rivers cut into the rift flanks, with downcutting in the vicinity of the Taung site. All these features indicate essentially similar conditions to those in the Makapan Valley, corresponding to the model outlined in Figure 5B, with varied habitats near the site and a complex topography affording opportunities for protection in the immediate vicinity and for monitoring of resources in the wider landscape. Sterkfontein The deposits at Sterkfontein make up seven members, of which two are well studied: Member 4 (ca. 2.8–2.4 Ma), with Australopithecus africanus fossils, and Member 5 (ca. 2.5– 1.4 Ma), with a succession of Oldowan and Acheulean stone-tool industries as well as two later species of hominins, namely, early Homo and Paranthropus (Kuman and Clarke, 2000). Faunal remains indicate a mixture of grassland and woodland species in Member 4, and fossilized wood fragments indicate the presence of gallery forest and tropical understory shrubs in the near vicinity of the site (Bamford, 1999). In Member 5 times, all faunal

Figure 12. Close-up of Makapan Valley. (A) Closer view of the topography modified from Google Earth image. Fault γ is indicated by arrows. Red arrows indicate steepening of the base of the slope. A yellow arrow indicates a “wine glass,” a valley that narrows toward the fault. A blue arrow indicates a spur that has been truncated by the fault. Together, these features are unambiguous evidence of active faulting. Small fertile plains result from sediment back-filling of earlier features as a result of the tectonically modified drainage. (B) Sediment-filled valley resulting from down-dropping on fault γ. (C) Steep topography in the Makapan Valley.

indicators suggest generally more open conditions but with some persisting woodland. The topographic setting of the site (Figs. 14A and 14B) shows evidence for faulting that crosscuts the mapped geological structures and has disturbed the profile of the adjacent river to create areas of sedimentation and downcutting of several meters, most probably due to continued movement (as in Fig. 5B). However, there are no clear earthquake fault scarps as at Makapansgat and Taung, so that continuing activity cannot be unequivocally established as yet. In the region to the north, the rivers are deeply incised, again suggesting activity. As at Makapansgat, these features are consistent with the presence of environmental signals in the on-site evidence indicating a combination of open grassland and more wooded habitats. A project is currently under way to improve the seismic network coverage in the area and to measure erosion and river downcutting rates using cosmogenic dating. Boomplaas Boomplaas cave has an important sequence of deposits dating from ca. 70 ka onward and includes stone tools from the

Up U p

Up

Site

Down

activ

e

cu down

tting

ps car lt s Fau

sedimented plain

~10 km

N

Figure 13. Region around the Taung Valley. Landsat thematic mapper image (bands 2, 4, and 7 are red, green, and blue, respectively) draped over exaggerated digital elevation data to give a three-dimensional (3-D) effect. An earthquake fault scarp on the east side of the valley is indicated by white arrows. The fault crosscuts geological features and varies in altitude and thus cannot be a river terrace or other erosional or depositional feature. Uplifting of the eastern flank has caused rivers to incise, and the down-dropped valley has been filled with sediment. Faulting may also be associated with the western side of the valley, causing the incised valley that hosted the breccia with the hominin fossil. The down-dropped sedimented basin is today used for agriculture. The presumed faults at depth are shown as gray dashed lines.

A Up Down

50 km B

B

10 km gorge

Elevation (m)

A

1450

Sterkfontein

deposition

1400

undis 1350

downcutting

turbe

d pro

gorge riv

file

er pro file

~ 25 km A

B

Figure 14. Evidence of tectonic activity in the Sterkfontein region. (A) Shaded relief map of the Sterkfontein “Cradle of Humankind” area (yellow rectangle) and the region to the east. White arrows indicate an east-west fault. This substantial feature extends for >150 km, offsets the morphology (up to the north, down to the south), and crosscuts earlier geological structures. The shaded map uses SRTM 3 data with the light source located at N45°W at an elevation of 5°. (B) Landsat thematic mapper image (bands 2, 4, and 7 are red, green, and blue, respectively) with relief shading based on digital elevation model (DEM; ~10 m resolution) derived from Stereo SPOT images. The fault identified in A passes to the north of the Sterkfontein site (dashed gray line) and apparently controls the position of the Blaaubank River, which follows the northern side of the valley. As it flows to the east, the river passes through a gorge (several meters deep). The inset shows a profile of the river (from SPOT DEM), which is downcutting in the gorge and aggrading above it. Such a profile is commonly associated with active faulting.

Dynamic landscapes and human evolution Middle Stone Age (MSA) industries of Still Bay and the succeeding Howiesons Poort, Later Stone Age material, and evidence of sheep pastoralism (Deacon, 1995; Deacon et al., 1978; Henshilwood, 2010). The cave itself has an open aspect overlooking a valley filled with sediment, and Deacon (1979) suggested that the site was well placed to intercept migratory animals. From a topographic perspective, the site is located on the boundary between a down-dropped valley and an uplifted scarp (Fig. 15). The valley area has been down-dropped by the fault to the north, and this has caused valleys in the earlier topography to become partly buried, resulting locally in highly fertile sediment-filled valleys. The cave site itself was formed in the earlier topography and is

15

not a consequence of the system that is now active. Fault scarps like those shown in Figure 15 can be found elsewhere in the cape region and may well partly control coastal sites. DISCUSSION Our review of key South African fossil and archaeological sites shows that there is considerable evidence for dynamic landscape changes resulting from tectonic activity over the time span during which the sites were formed, and that this evidence has not previously been recognized. The changes are not as rapid or as dramatic as in the African Rift, but they are changes that

A Up Down

fault

Site Sedimented Cango valley Downcutting

0

kilometers 5

10

B

Site

Sedimented valley

Figure 15. Evidence for tectonic activity in the Boomplaas region. (A) Landsat thematic mapper image (bands 2, 4, and 7 are red, green, and blue, respectively). White arrows indicate two faults. Yellow arrows indicate the sediment-filled Cango valley. Red arrows indicate downcutting. (B) View of cave site showing the Cango valley in foreground. The valley is subject to intermittent flooding and provides rich farmland. Earthquakes are commonly felt in this region.

16

King and Bailey

would have been advantageous for human occupation. They would have created repeated disturbance of drainage networks to sustain localized wetlands with fertile resources and water supplies, and formed or maintained complex local topography affording diversity of resource zones in close proximity and tactical advantages in hiding from predators or accessing mobile prey. The relatively subdued nature of tectonic activity compared to the Ethiopian Rift, which at first sight appears to pose a difficulty for the hypothesis of a close association between tectonically disturbed landscapes and early hominin settlement, turns out to be an advantage, allowing a direct evaluation of the relationship between tectonic structures and human activity without resorting to an analog approach. Three issues for discussion arise from this review. A first issue is the problem of bias in the distribution of known archaeological and fossil sites resulting from differential preservation and visibility of material. Geological processes in tectonically active areas tend to generate higher rates of sedimentation, which bury and therefore preserve fossil and archaeological material, and high rates of subsequent disruption and erosion, which then expose to view deposits formed at a much earlier period in time. Thus, it can be argued that the correlation between early archaeological and fossil sites and areas of tectonic activity is purely coincidental and is the result of greater visibility and exposure of material in such environments compared with environments where tectonic activity is very low or nonexistent. A similar argument is sometimes made about caves and rock shelters. These tend to be easily visible targets for archaeological investigations and often provide a protective environment for the accumulation of sediments and the preservation of archaeological material in stratified sequences. Hence, the argument can be proposed that the distribution of prehistoric settlement patterns derived from cave and rock-shelter deposits is unrepresentative and biased by conditions of geological visibility, a proposition pertinent to our South African region, where many key sites are in caves. Taken to its logical conclusion, this argument would require us to suppose that the distribution of prehistoric archaeological and fossil materials in space and time is solely a function of geological processes affecting visibility, that tectonically active areas of Africa have no particular significance for human evolution, and that actually other regions of Africa or elsewhere were equally important, if not more so, despite the absence of evidence in favor of such a proposition. Such an argument would be simplistic. We doubt that the concentration of finds in East and South Africa is wholly unrepresentative or can tell us nothing about the environmental conditions in which early human populations prospered. Other areas with early sites are sometimes claimed to lack the tectonic activity or the topographic features to which we have drawn the readers’ attention here. A notable case in point is the early finds of hominin fossils in the Chad region (Brunet et al., 1995). However this region, though far from the East African Rift, is one of the most tectonically active areas of sub-Saharan Africa outside the rift (Burke, 1996).

Factors of differential visibility resulting from geological processes are not trivial, and we do not discount them. However, they can be addressed in a variety of ways. In the case of caves and rock shelters, not all that were available for use contain evidence of human activity. Of those that do, some clearly show evidence of more activity than others. Some regions with available caves and rock shelters clearly show greater concentrations of evidence than others. Open-air sites can be targeted to provide a control and are often found once they are sought out. A similar approach can be employed in relation to tectonic factors. Moreover, it is not necessarily the case that archaeological materials from very early periods will be invisible in areas that are not tectonically active because these areas are likely to have smoothed surfaces that are subject neither to accumulation of obscuring sediment nor to erosion, and artifacts once deposited are likely to remain in place for many tens or hundreds of millennia. Granted, surface artifacts are more difficult to date than stratified material and likely to comprise only stone tools, but if such areas were attractive at an early period, we might expect to find concentrations of distinctive stone tools characteristic of Lower Paleolithic industries. Many such areas exist in the African Rift sensu lato, including the now-uplifted flanks of the rift, which would have been available for occupation at a relatively early stage in the Pleistocene, but little archaeological evidence of human activity was recorded in such areas until much later periods of human development and, in many cases, not until the expansion of pastoralist societies in the Holocene. A second issue concerns the variability in rates of landscape change resulting from different levels of tectonic activity and the long-term evolutionary implications of variable activity. The South African region as a whole clearly differs in its general rate of activity compared with the most active parts of the East African Rift. Even within South Africa, there appears to be variation between the sites and regions we have discussed, and there are additional sites that we have not included in our review where tectonically informed studies have yet to be carried out. Nevertheless, at a general level our results suggest that even quite modest rates of tectonic activity are likely to generate the sorts of topographic features we have described, and therefore are likely to be advantageous for human settlement, even in regions with few or no earthquakes, or relatively small ones within the lifetime of a human individual. A critical variable in this equation between rates of tectonic activity and creation of rough landscapes is the rate of erosion. In regions with relatively soft rocks and active forces of erosion, modest rates of tectonic activity may be insufficient to offset the smoothing effects of erosion, and the long-term trend will be toward a flat topography lacking the advantages for human activity that we have described. Conversely, areas with very hard rocks may preserve the rough and complex topographic features created by occasional tectonic disruption for longer periods despite generally low rates of tectonic activity. This is certainly a contributing factor to the topography of the areas we have described in South Africa, where the rock formations are metamorphosed and

Dynamic landscapes and human evolution mostly were formed earlier than the Cenozoic (Hartzer, 1998). These rocks are very hard and resistant to erosion, and they tend to maintain vertical cliffs and fissures rather than degrading to the rounded and flattened features that result from erosion of softer rock formations. There is, then, a continuum of topographic conditions. At one extreme, regions have so little tectonic activity, and rock formations liable to erosion, such that the resulting topography is likely to be generally smooth. Such plains environments offer little diversity of resources or complexity of topography offering tactical advantage, except at their margins or in localized areas where erosion has created some topographic relief. From this perspective, it is no surprise that extensive plains environments such as the Asian steppes or the Great Plains of North America show limited evidence of human occupation in prehistory until the adoption of the horse as a riding animal. Equally at the other extreme, very active tectonics may have consequences that are as much destructive as constructive, at least at a local or subregional scale, resulting in geographical displacement of favorable areas, and destruction of once-fertile basins. The implications of such variation for evolutionary trends, particularly with respect to the general pattern of hominin evolution, remain to be worked out, but it is worth noting that Reynolds (2007) related the higher rate of species turnover amongst large mammals in East Africa in comparison with South Africa to the higher rates of tectonic activity in the East African Rift. The selective impact of dynamic topography may take different forms, depending on the scale of the landscape changes involved. One possibility is that populations of large mammals are isolated by insurmountable topographic barriers, resulting in genetic divergence and speciation. Another possibility is that rough topography, by sustaining favorable environmental conditions during climatic downturns, helps to maintain higher levels of population than would otherwise have been the case, and hence maintains a larger pool of genetic variability that can provide the basis for later adaptation and evolutionary change. Finally, a tectonically active and complex topography may select for greater locomotory and cognitive adaptability able to cope with changeable environmental conditions, for example, bipedal movement suitable for moving through rough terrain and climbing rock barriers, and cognitive abilities able to take tactical advantage of complex topography in tracking mobile animal prey or avoiding predators. A critical factor in exploring further these possibilities is careful reconstruction of topographic conditions at a variety of scales and, in particular, better dating of rates and periodicities of tectonic movement, earthquake repeat times, and volcanic eruptions. Satellite imagery along with a new generation of cosmogenic dating techniques will play an important role in conjunction with existing methods of dating, mapping, and stratigraphic interpretation. Without a tectonically informed reconstruction of local landscape conditions as they existed during the periods in question, claims that very early archaeological sites or hominin fossils typically occur in regions lacking rough topography or tectonic

17

activity cannot be sustained. Issues of differential preservation cannot be discounted and almost certainly add an extra layer of variability that needs to be addressed alongside other factors. More studies are needed at a regional scale of the type that we have described here, involving a systematic program of dating of geological surfaces and deposits alongside systematic surveys for archaeological and fossil sites. ACKNOWLEDGMENTS This article is a contribution to the United Nations Educational, Scientific and Cultural Organization–funded International Geoscience Programme IGCP 567, “Earthquake Archaeology: Archaeoseismology along the Alpine-Himalayan Seismic Zone.” We acknowledge funding from NERC, UK (grant NE/ A516937/1) as part of its EFCHED (Environmental Factors in Human Evolution and Dispersal Programme). Funding for the South African work came from the France–South Africa !Khur project. It is IPGP (Institute de Physique du Globe de Paris) contribution number 3034. Satellite data © CNES 2007, distribution Spot Image S.A., was used for some figures. REFERENCES CITED Ambrose, S.H., 1998, Late Pleistocene human population bottlenecks, volcanic winter and differentiation of modern humans: Journal of Human Evolution, v. 34, p. 623–651, doi: 10.1006/jhev.1998.0219. Anton, S.C., Leonard, W.R., and Robertson, M.L., 2002, An ecomorphological model of the initial hominid dispersal from Africa: Journal of Human Evolution, v. 43, p. 773–785, doi: 10.1006/jhev.2002.0602. Ayele, A., Jacques, E., Kassim, M., Kidane, T., Omar, A., Tait, S., Nercessian, A., de Chabalier, J.-B., and King, G.C.P., 2007, The volcano-seismic crisis in Afar, Ethiopia, starting September 2005: Earth and Planetary Science Letters, v. 255, p. 177–187, doi: 10.1016/j.epsl.2006.12.014. Bailey, G.N., 1997, Klithi: Palaeolithic Settlement and Quaternary Landscapes in Northwest Greece: Cambridge, UK, McDonald Institute Monographs. Bailey, G.N., King, G.C.P., and Sturdy, D., 1993, Active tectonics and land-use strategies: A Paleolithic example from northwest Greece: Antiquity, v. 67, p. 292–312. Bailey, G.N., King, G.C.P., and Manighetti, I., 2000, Tectonics, volcanism, landscape structure and human evolution in the African Rift, in Bailey, G., Charles, R., and Winder, N., eds., Human Ecodynamics: Proceedings of the Association for Environmental Archaeology Conference 1998: Oxford, Oxbow, p. 31–46. Bailey, G.N., Reynolds, S., and King, G.C.P., 2010, Landscapes of human evolution: Models and methods of tectonic geomorphology and the reconstruction of hominin landscapes: Journal of Human Evolution, doi: 10.1016/j.jhevol.2010.01.004 (in press). Bamford, M., 1999, Pliocene fossil woods from early hominid cave deposit, Sterkfontain, South Africa: South African Journal of Science, v. 95, no. 5, p. 231–237. Berger, L.R., de Ruiter, D.J., Churchill, S.E., Schmid, P., Carlson, K.J., Dirks, P.H.G.M., and Kibii, J.M., 2010, Australopithecus sediba: A new species of Homo-like Australopith from South Africa: Science, v. 328, no. 5975, p. 195–204, doi: 10.1126/science.1184944. Bondevik, S., Svendsen, J.I., Johnsen, G., Mangerud, J., and Kaland, P.E., 1997, The Storegga tsunami along the Norwegian coast, its age and runup: Boreas, v. 26, p. 29–53. Brunet, M., Beauvilain, A., Coppens, Y., Heintz, E., Moutaye, A.H.E., and Pilbeam, D., 1995, The first australopithecine 2,500 kilometres west of the Rift Valley (Chad): Nature, v. 378, p. 273–275, doi: 10.1038/378273a0. Burke, K., 1996, The African plate: South African Journal of Geology, v. 99, no. 4, p. 341–409.

18

King and Bailey

Cadman, A., and Rayner, R.J., 1989, Climatic change and the appearance of Australopithecus africanus in the Makapansgat sediments: Journal of Human Evolution, v. 18, no. 2, p. 107–113, doi: 10.1016/0047-2484(89)90065-1. Dakaris, S.I., Higgs, E.S., and Hey, R.W., 1964, The climate, environment and industries of Stone Age Greece, Part I: Proceedings of the Prehistoric Society, v. 30, p. 199–244. Dart, R.A., 1925, Australopithecus africanus: The ape-man of South Africa: Nature, v. 115, p. 195–197, doi: 10.1038/115195a0. Deacon, H.J., 1979, Excavations at Boomplaas Cave: A sequence through the Upper Pleistocene and Holocene in South Africa: World Archaeology, v. 10, no. 3, p. 241–257. Deacon, H.J., 1995, Two late Pleistocene-Holocene archaeological depositories from the southern Cape, South Africa: South African Archaeological Bulletin, v. 50, no. 162, p. 121–131, doi: 10.2307/3889061. Deacon, H.J., Deacon, J., Brooker, M., and Wilson, M.L., 1978, The evidence for herding at Boomplaas Cave in the southern Cape, South Africa: South African Archaeological Bulletin, v. 33, p. 39–65, doi: 10.2307/3888251. de Heinzelin, J., Clark, J.D., White, T., Hart, W., Renne, P., WoldeGabriel, G., Beyene, Y., and Vrba, E., 1999, Environment and behavior of 2.5million-year-old Bouri hominids: Science, v. 284, p. 625–629, doi: 10.1126/ science.284.5414.625. Dirks, P.H.G.M., Kibii, J.M., Kuhn, B.F., Steininger, C., Churchill, S.E., Kramers, J.D., Pickering, R., Farber, D.L., Mériaux, A.-S., Herries, A.I.R., King, G.C.P., and Berger, L.R., 2010, Geological setting and age of Australopithecus sediba from southern Africa: Science, v. 328, no. 5975, p. 205–208, doi: 10.1126/science.1184950. Douglas, R.M., 2006, Formation of the Florisbad spring and fossil site—An alternative hypothesis: Journal of Archaeological Science, v. 33, p. 696– 706, doi: 10.1016/j.jas.2005.10.005. Gasse, F., 2001, Paleoclimate: Hydrological changes in Africa: Science, v. 292, no. 5525, p. 2259–2260, doi: 10.1126/science.1061940. Hartzer, F.J., 1998, A Stratigraphic Table of the SADC Countries: Praetoria, Council for Geoscience (South Africa), published on behalf of SADC by the Council for Geoscience, 3 sheets. Henshilwood, C.S., 2010, The ‘Upper Paleolithic’ of southern Africa: The Still Bay and Howiesons Poort, in Reynolds, S.C., and Gallagher, A., eds., African Genesis: Perspectives on Hominin Evolution: Johannesburg, Witwatersrand University Press (in press). Higgs, E.S., and Vita-Finzi, C., 1966, The climate, environment and industries of Stone Age Greece, Part II: Proceedings of the Prehistoric Society, v. 32, p. 1–29. Higgs, E.S., Vita-Finzi, C., Harris, D.R., and Fagg, E.E., 1967, The climate, environment and industries of Stone Age Greece. Part III: Proceedings of the Prehistoric Society, v. 33, p. 1–29. Johanson, D., Taieb, M., and Coppens, Y., 1982, Pliocene hominids from the Hadar Formation, Ethiopia (1973–1977): Stratigraphic, chronologic and palaeoenvironmental contexts, with notes on hominid morphology and systematics: American Journal of Physical Anthropology, v. 57, p. 373– 402, doi: 10.1002/ajpa.1330570402. Kalb, J., Oswald, E.B., Tebedge, S., Mebrate, A., Tola, E., and Peak, D., 1982a, Fossil mammals and artefacts from the Middle Awash Valley, Ethiopia: Nature, v. 298, p. 17–25, doi: 10.1038/298017a0. Kalb, J.E., Jolly, C.J., Mebrate, A., Tebedge, S., Smart, C., Oswald, E.B., Cramer, D., Whitehead, P., Wood, C.B., Conroy, C.G., Adefris, T., Sperling, L., and Kana, B., 1982b, Fossil mammals and artefacts from the Middle Awash Valley, Ethiopia: Nature, v. 298, p. 25–29, doi: 10.1038/298025a0. King, G.C.P., and Bailey, G.N., 1985, The palaeoenvironment of some archaeological sites in Greece: The influence of accumulated uplift in a seismically active region: Proceedings of the Prehistory Society, v. 51, p. 273–282. King, G.C.P., and Bailey, G.N., 2006, Tectonics and human evolution: Antiquity, v. 80, p. 265–286. King, G.C.P., and Vita-Finzi, C., 1981, Active folding in the Algerian earthquake of 10 October 1980: Nature, v. 292, no. 5818, p. 22–26, doi: 10 .1038/292022a0. King, G.C.P., and Yielding, G., 1984, The evolution of a thrust fault system: Processes of rupture initiation, propagation and termination in the 1980 El Asnam (Algeria) earthquake: Geophysical Journal of the Royal Astronomical Society, v. 73, p. 605–639. King, G.C.P., Tselentis, A., Gomberg, J., Molnar, P., Roecker, S.W., Sinvhal, H., Soufleris, C., and Stock, J.M., 1983, Microearthquake seismicity and active tectonics of northwestern Greece: Earth and Planetary Science Letters, v. 66, p. 279–288, doi: 10.1016/0012-821X(83)90141-3.

King, G.C.P., Sturdy, D., and Whitney, J., 1993, The landscape geometry and active tectonics of N.W. Greece: Geological Society of America Bulletin, v. 105, p. 137–161, doi: 10.1130/0016-7606(1993)105<0137:TLGAAT >2.3.CO;2. King, G.C.P., Bailey, G., and Sturdy, D., 1994, Active tectonics and human survival strategies: Journal of Geophysical Research, v. 99, no. B10, p. 20,063–20,078, doi: 10.1029/94JB00280. King, G.C.P., Sturdy, D., and Bailey, G.N., 1997, The tectonic background to the Epirus landscape, in Bailey, G.N., ed., Klithi: Paleolithic Settlement and Quaternary Landscapes in Northwest Greece: Cambridge, UK, McDonald Institute Monographs, p. 541–558. Kuman, K., and Clarke, R.J., 2000, Stratigraphy, artefact industries and hominid associations for Sterkfontein, Member 5: Journal of Human Evolution, v. 38, no. 6, p. 827–847, doi: 10.1006/jhev.1999.0392. Kuman, K., Inbar, M., and Clarke, R.J., 1999, Palaeoenvironments and cultural sequence of the Florisbad Middle Stone Age hominid site, South Africa: Journal of Archaeological Science, v. 26, p. 1409–1425, doi: 10.1006/jasc .1999.0439. Manighetti, I., Tapponnier, P., Gillot, P.Y., Courtillot, V., Jacques, E., Ruegg, J.-C., and King, G.C.P., 1998, Propagation of rifting along the ArabiaSomalia plate boundary into Afar: Journal of Geophysical Research, v. 103, p. 4947–4974, doi: 10.1029/97JB02758. Manighetti, I., Tapponnier, P., Courtillot, V., Gallet, Y., Jacques, E., and Gillot, P.Y., 2001, Strain transfer between disconnected, propagating rifts in Afar: Journal of Geophysical Research, v. 106, p. 13,613–13,665, doi: 10.1029/2000JB900454. Maslin, M.A., and Christensen, B., 2007, Tectonics, orbital forcing, global climate change, and human evolution in Africa: Introduction to the African paleoclimate special volume: Journal of Human Evolution, v. 53, no. 5, p. 443–464, doi: 10.1016/j.jhevol.2007.06.005. McCarthy, T.S., and Hancox, P.J., 2000, Wetlands, in Partridge, T.C., and Maud, R.R., eds., The Cenozoic of Southern Africa: Oxford, Oxford University Press, p. 145–161. McCarthy, T.S., Tooth, S., Brandt, D., Marren, P., and Hancox, P.J., 2004, On the origin of the Nylsvlei wetland, in Geoscience Africa 2004 Conference Extended Abstracts: Johannesburg, University of Witwatersrand. McKee, J.K., 1993, The faunal age of the Taung hominid deposit: Journal of Human Evolution, v. 25, no. 5, p. 363–367, doi: 10.1006/jhev.1993.1055. McKee, J.K., 1999, The autocatalytic nature of hominid evolution in African Plio-Pleistocene environments, in Bromage, T.G., and Schrenk, F., eds., African Biogeography, Climate Change and Human Evolution: Oxford, Oxford University Press, p. 115–137. Meghraoui, M., and Doumaz, F., 1996, Earthquake-induced flooding and paleoseismicity of the El Asnam (Algeria) fault-related fold: Journal of Geophysical Research, v. 101, p. 17,617–17,644. Mitchell, P., 2002, The Archaeology of Southern Africa: Cambridge, UK, Cambridge University Press, 515 p. Partridge, T.C., 2000, Hominid-bearing cave and tufa deposits, in Partridge, T.C., and Maud, R.R., eds., The Cenozoic of Southern Africa: Oxford, Oxford University Press, p. 100–125. Philip, H., and Meghraoui, M., 1983, Structural analysis and interpretation of the surface deformation of the El Asnam earthquake of October 10, 1980: Tectonics, v. 2, no. 1, p. 17–49, doi: 10.1029/TC002i001p00017. Radosevich, S.C., Retallack, G.J., and Taieb, M., 1992, Reassessment of the paleoenvironment and preservation of hominid fossils from Hadar, Ethiopia: American Journal of Physical Anthropology, v. 87, p. 15–27, doi: 10.1002/ajpa.1330870103. Reed, K.E., 1997, Early hominid evolution and ecological change through the African Plio-Pleistocene: Journal of Human Evolution, v. 32, no. 2–3, p. 289–322, doi: 10.1006/jhev.1996.0106. Reynolds, S.C., 2007, Mammalian body size changes and Plio-Pleistocene environmental shifts: Implications for understanding hominin evolution in eastern and southern Africa: Journal of Human Evolution, v. 53, no. 5, p. 528–548. Rowland, J.V., Baker, E., Ebinger, C.J., Keir, D., Kidane, T., Biggs, J., Hayward, N., and Wright, T.J., 2007, Fault growth at a nascent slow-spreading ridge: 2005 Dabbahu rifting episode, Afar: Geophysical Journal International, v. 171, no. 3, p. 1–21, doi: 10.1111/j.1365-246X.2007.03584.x. Saunders, I., Brandt, M., Steyn, J., Roblin, D., and Kijko, A., 2008, The South African National Seismograph Network: Seismological Research Letters, v. 79, no. 2, p. 203–209, doi: 10.1785/gssrl.79.2.203. Semaw, S., Rogers, M.J., Qaude, J., Renne, P.R., Butler, R.F., DominguezRodrigo, M., Stout, D., Hart, W.S., Pickering, T., and Simpson, S.W.,

Dynamic landscapes and human evolution 2003, 2.6-million-year-old stone tools and associated bones from OGS-6 and OGS-7, Gona, Afar, Ethiopia: Journal of Human Evolution, v. 45, no. 2, p. 169–177, doi: 10.1016/S0047-2484(03)00093-9. Semaw, S., Simpson, S.W., Quade, J., Renne, P.R., Butler, R.F., McIntosh, W.C., Levin, N., Dominguez-Rodrigo, M., and Rogers, M.J., 2005, Early Pliocene hominids from Gona, Ethiopia: Nature, v. 433, no. 7023, p. 301– 305, doi: 10.1038/nature03177. Sponheimer, M., and Lee-Thorp, J.A., 1999, Isotopic evidence for the diet of an early hominid, Australopithecus africanus: Science, v. 283, no. 5400, p. 368, doi: 10.1126/science.283.5400.368. Stein, R., and King, G.C.P., 1984, Seismic potential revealed by surface folding: The 1983 Coalinga, California, Earthquake: Science, v. 224, p. 869– 872, doi: 10.1126/science.224.4651.869. Tobias, P.V., 2000, The fossil hominids, in Partridge, T.C., and Maud, R.R., eds., The Cenozoic of Southern Africa: Oxford, Oxford University Press, p. 252–276. Vita-Finzi, C., 1969, The Mediterranean Valleys: Cambridge, UK, Cambridge University Press, 140 p. Vrba, E.S., 1982, Biostratigraphy and chronology, based particularly on Bovidae, of southern hominid associated assemblages, in de Lumley, H., and

19

de Lumley, M.-A., eds., Makapansgat, Sterkfontein, Taung, Kromdraai, Swartkrans; also Elandsfontein (Saldanha), Broken Hill (now Kabwe) and Cave of Hearths: Premier Congrès International de Paléontologie Humaine, p. 707–752. Wagner, P.A., 1927, The geology of the northeastern part of the Springbok Flats and surrounding country: An explanation of Sheet 17 (Springbok Flats): Geological Survey of South Africa. White, T.D., Asfaw, B., DeGusta, D., Gilbert, H., Richards, G.D., Suwa, G., and Howell, F.C., 2003, Pleistocene Homo sapiens from Middle Awash, Ethiopia: Nature, v. 423, no. 6941, p. 742–747. WoldeGabriel, G., White, T.D., Suwa, G., Renne, P.R., de Heinzelin, J., Hart, W.K., and Heiken, G., 1994, Ecological and temporal placement of early Pliocene hominids at Aramis, Ethiopia: Nature, v. 371, p. 330–333, doi: 10.1038/371330a0. WoldeGabriel, G., Haile-Selassie, Y., Renne, P.R., Hart, W.K., Ambrose, S.H., Asfaw, B., Heiken, G., and White, T., 2001, Geology and palaeontology of the late Miocene Middle Awash Valley, Afar rift, Ethiopia: Nature, v. 412, p. 175–178, doi: 10.1038/35084058. MANUSCRIPT ACCEPTED BY THE SOCIETY 19 MAY 2010

Printed in the USA

The Geological Society of America Special Paper 471 2010

Tectonic environments of ancient civilizations: Opportunities for archaeoseismological and anthropological studies Eric R. Force* Department of Geosciences, University of Arizona, Tucson, Arizona 85721, USA Bruce G. McFadgen* School of Maori Studies, Victoria University of Wellington, Wellington 6140, New Zealand

ABSTRACT The close spatial relation between ancient civilizations and active tectonic boundaries is robust in the Eastern Hemisphere but counterintuitive given the seismic disadvantages it implies. Explanations for the observation remain debatable, and no single explanation seems sufficient. Some possibly important factors are unrelated to seismicity, e.g., the influence of tectonism on local water resources and on resource diversity. When examined on finer spatial scales, the relation is still robust. A quantifiable influence of tectonism on civilization locations even along Mediterranean shores is suggested by their distribution. The stronger links of tectonism with derivative civilizations suggest a role of ancient trade connections. Several clues point to cultural response as an important ingredient in the dynamics resulting in the spatial relation. These are: correlation between static character and location of civilizations relative to tectonic locus; archaeologic and historic records of accelerated cultural (especially religious) change following tectonic events; and evidence that the spatial relation evolves through time via trade goods and routes. Archaeoseismology is in a key position to provide additional clues to this paradoxical relation in at least three ways: (1) providing detail on evolving societal response; (2) determining the most pertinent tectonic styles; and (3) determining the role of seismicity in Neolithic cultures that eventually became civilizations.

INTRODUCTION

ture in this direction presently supports (Bailey et al., 1993; King et al., 1994; Trifonov and Karakhanian, 2004; Force, 2008). The tectonic boundaries that belong in this pattern are mostly convergent to transcurrent ones, associated with the southern margin of the Eurasian plate. The civilizations are those conventionally regarded as the greatest of antiquity (they could be defined as those that score eight or higher in the ten original criteria of Childe, 1950). It is worth recalling that conventional criteria for greatness in a civilization have valued cities, monumental (preservable) architecture, and (preserved and deciphered)

The papers in this volume offer abundant evidence of the destructive seismic environment of many ancient sites. Yet, the highest civilizations to which these sites belong are clustered along the very seismogenic tectonic boundaries that give rise to the documented destruction (Fig. 1). This counterintuitive pattern requires further attention, much more than a slender litera*E-mails: [email protected]; [email protected].

Force, E.R., and McFadgen, B.G., 2010, Tectonic environments of ancient civilizations: Opportunities for archaeoseismological and anthropological studies, in Sintubin, M., Stewart, I.S., Niemi, T.M., and Altunel, E., eds., Ancient Earthquakes: Geological Society of America Special Paper 471, p. 21–28, doi: 10.1130/2010.2471(02). For permission to copy, contact [email protected]. © 2010 The Geological Society of America. All rights reserved.

21

22

Force and McFadgen

writing—and anthropologists these days are equally interested in other aspects of cultural complexity. However, it is this particular assemblage of traits that corresponds so closely with tectonic boundaries, regardless of the semantic label that is attached. Quantification of the relation was attempted for the Eastern Hemisphere by Force (2008) based on probabilistic comparison to random distribution. Measurements were taken from originating sites (to avoid the problem of imperial sprawl) of 13 civilizations to the nearest tectonic plate boundary as conventionally mapped, and these resulted in an average distance of 75 km, with two prominent exceptions (Egypt and China; the influence of tectonism on development of the latter is appreciable but not treated here). This distance can be converted to a polygon averaging 150 km wide (civilizations could be on either side of a tectonic boundary) along the total on-land length of the boundaries, and the probability was calculated for 11 of 13 civilizations finding themselves in this tectonically defined polygon, assuming random distribution. These probabilities were calculated for two different assumptions of available land areas, and both the included civilizations and tectonic boundary locations (especially where plate boundaries are partitioned) were varied to provide a sensitivity analysis. The calculations were necessarily approximate, but the calculated probabilities are so miniscule that random distribution can be rejected. The conclusion that ancient civilizations seem to be preferentially located near active tectonic boundaries seems robust. The observed pattern is a simple one, but many mechanisms and dynamics seem possible. All of them are inherently untestable in real time, and this limits the rigor with which linkages can be proven. However, we can constrain the factors that are most consistent with our information. It is possible that the importance

of tectonism acts via some other variable that is more obviously required for civilization to become established and thrive, such as climate, soils, water, and transport potential. Volcanism appears not to be of systematic importance, since few of the ancient civilizations are near Holocene volcanic centers (except in the Western Hemisphere, an intriguing relation outside the purview of this volume). Quaternary volcanism, however, did provide important soil and building material assets in Italy. The civilizations differ so much from each other in their apparent environmental characters that it is difficult to say which of them might be related to tectonism (Table 1). Climatecontrolled vegetation varies greatly and may have changed since antiquity, but latitudes ranging from ~27° to 42° have not (the difference between Tampa and Boston). Relation to type of water resource varies similarly. The relation to slightly varying styles of active tectonism is more commonly shared. So far, it is not possible to say for certain whether seismic activity per se is important in the relation, or whether some other aspect of position near an active tectonic boundary is involved. For example, the relation shown in Figure 1 is clearly more closely related to plate boundaries than to seismic risk, especially in Iran, Tibet, western China, etc. One positive effect that is independent of seismicity is local water resources, which can be enhanced by active tectonism. If fractures in fault gouge are abundant and randomly oriented, one set will be held open by active stress, permitting greater permeability of active faults than old inactive ones (Hickman et al., in Force, 2008). Along tectonic boundaries, the prevalence of malarial paleopathology in those Neolithic villages that were involved in transitions to civilization (locations from Maisels, 1999) is consistent with this being an important factor.

TABLE 1. VARIATIONS AMONG RIVERBORNE WATER SUPPLY, CLIMATIC VARIABLES, AND TECTONIC MICRO-ENVIRONMENT OF ORIGINAL SITES OF SOME ANCIENT CIVILIZATIONS Civilization Rivers Climate Tectonic microLatitude Vegetation environment Roman

M

41°50′N

1

Etruscan

S

42°10′N

1

E+S

Gree k

S

1

E > S + Tr

1

E > S + Tr

E+S

Mycenaean

S

38°N 37°50′N

Minoan

S

34°40′N

1

E+S

S&M

31°50′–33°20′N

1, 2

Tr > E

SW Asian Assyrian

L

36°30′N

2

Th + Tr

Mesopotamian

L

2

Th + Tr

Persian

S

31°N 30°–32°20′N

2, 3, 4

Th + Tr

Indus

L

27°20′N

5, 7

Th + Tr

Aryan India

L

29°30′N

5, 6

Th

Egyptian

L

29°50′N

7, 8

N.A.

Chinese

L

34°40′N

9

N.A.

Notes: Rivers: S—small, M—mid-size, and L—large. Vegetation: 1—Mediterranean scrub; 2—short-grass steppe; 3— conifer forest; 4—mountain vegetation; 5—dry tropical forest; 6—dry tropical scrub; 7—desert; 8—floodplain; 9—mixed and broadleaf forest. Tectonic environment: E—extensional; S—subduction-related; Tr—transcurrent; Th—thrust. N.A.— not applicable.

Tectonic environments of ancient civilizations

23

N Eurasian plate

2 1

H-T plate 3

4

13

7 5 6

9 8

12

11 10

Arabian plate

African plate

I n d o - Australian plate 0

1000

2000

km

H-T = Hellenic-Turkish plate

Figure 1. Locations of originating sites of 13 prominent ancient civilizations relative to various aspects of the southern boundary of the Eurasian plate (after Force, 2008). Civilizations (and sites) shown are 1—Roman (Rome), 2—Etruscan (Tarquinii-Veii), 3—Greek (Corinth) and Mycenaean (Mycenae), 4—Minoan (Knossos-Phaestos), 5 and 6—West Asian (Tyre and Jerusalem), 7—Assyrian (Ninevah), 8—Mesopotamian (Ur-Uruk), 9—Persian (Susa-Pasargadae), 10—Indus (Mohenjodaro), 11—Aryan India (Hastinapura), 12—Egyptian (Memphis), and 13—Chinese (Zhengzhou).

Another proposed factor independent of seismicity is the juxtaposition of different soils and topographies along active faults, especially transcurrent faults (cf. Bailey et al., 1993), providing a greater diversity of resources in the cultural equivalent of a biological edge effect. Trifonov and Karakhanian (2004) suggested the presence of anomalous geochemical fields along active faults, a possibility that seemed remote until de Boer et al. (2001) documented an example at Delphi. ADDITIONAL QUANTIFIABLE LINKS The relation between ancient civilizations and tectonic boundaries seems apparent not only at hemispherical scale but also at finer scales, for example, along the shores of the Medi-

terranean. These shores were certainly important to emerging civilizations there, and some would say this factor predominates. However, if the distribution of the African-Eurasian plate boundary is compared to ancient sites of civilization along opposing shores, a probable influence of tectonism shows through. The tectonic boundary broadly follows the south shore in the western Mediterranean but the northern and north-insular shores in the eastern part (Fig. 1). Ancient sites of civilizations (using a looser definition to allow more cases) follow the same path (Carthage, Syracuse, Rome, Tarquinii-Veii, Corinth, Mycenae, and Knossos-Phaistos). This relation itself looks persuasive, but an additional quantitative test would complement it. We can compare distances of tectonic boundaries versus seashores (of the time, where this is known) for the 11 originating sites of Figure 1.

24

Force and McFadgen

The average seashore distance is ~1.48 times greater than the tectonic-boundary distance, implying a probability of random distribution that is very small but ~75 times greater than that for tectonic boundaries. Even along Mediterranean and Near Eastern shores, civilizations apparently located their originating sites near active tectonic boundaries. Does the relation result in finer-scale vignettes? Perhaps a tectonic influence suggests itself in the Hellenic realm (Fig. 2). Mycenaean and later Greek civilization nucleated in a zone of distributed deformation along the on-land prolongation of the North Anatolian fault and its offsets along the Gulf of Corinth and other extensional structures. Where the tectonic zone passes northward into the Aegean, the spread of each civilization stalled, at Iolkos for the Mycenaeans and on Euboea (Evia) for the Greeks, though earlier, simpler Neolithic villages continued across this boundary into more quiescent parts of Thessaly. Force (2008) attempted to constrain the number and structure of possible solutions by comparing subsets of both the civilizations and their pertinent geologic environments. He showed that the closest relation is between tectonic boundaries and those ancient civilizations generally called derivative, i.e., those that evolved under some influence from more senior civilizations. This relation of civilization subsets suggests an influence of trade (however accomplished).

The propagation of trade routes, from more advanced settlements in the Near East to sites that eventually became great ancient civilizations, mimics tectonic boundaries more closely than one would expect given the availability of other routes. Figure 3 shows such propagation from Bronze Age through early Iron Age times in the Eastern Mediterranean area. Perhaps most impressive is the replication of the “draped” shape of the plate boundary between Cyprus and Crete by trade routes, based on the distribution of Bronze age stone anchors. Other trade routes existed also, of course, extending, for example, to Maikop, Danubian, Hallstatt, Scythian, etc. (cultures in seismically more quiescent north-central Eurasia), but these routes did not produce civilizations (as defined here) until long after the period of antiquity, if ever. If one measures the length of trade routes propagating toward eventual ancient civilizations as shown in Figure 3, ~79% of this length is within 100 km of an active plate boundary. (Routes destined for other sites of production or resource exploitation such as Kanesh do not count in this calculation.) A probability analysis was not attempted, because the appropriate structure for it is unclear, but perhaps it is unnecessary because its conclusion is intuitively obvious. Trade routes to eventual civilizations tended for some reason to follow tectonic boundaries, perhaps because each incremental stage in trade-route propagation was to

N 100 km

X X X

Aegean Sea Plate boundaries including margins of distributed deformation Mycenaean sites damaged 1250 or 1200 B.C. X

Early Geometric sites (ca. 900 B.C.)

Figure 2. Sites of Mycenaean palaces destroyed by earthquakes in 1200 and 1250 B.C. (from Kilian [1996] and other sources) relative to approximate boundaries of distributed deformation along the projection of the North Anatolian fault in the Hellenic realm (from McClusky et al., 2000). The distribution of destroyed Mycenaean palaces, with only a few exceptions, includes all the palaces of that civilization, suggesting not only ancient tectonic activity along this structural trend, but also the localization of palaces along it. The major early Geometric Age sites shown (from Coldstream, 2003) follow the same trend ~300 yr later.

Tectonic environments of ancient civilizations

25

NORTH

BLACK SEA EURASIAN PLATE 42 º

TO MAGNA GRAECIA

40 º

TH ANATOLIAN NOR FAU

KA

TROY

38 º

HELLENIC

AND

NE

LT

SH

TURKISH PLATES

AS

36 º

34 º

SU

R

ARABIAN PLATE

20 º E

U

AFRICAN PLATE 30 º

25 º

40 º

TO INDUS

32 º N

R

MEDITERRANEAN SEA

45 º

ROUTES OF “OLD - ASSYRIAN MERCHANT” (2200 - 1900 B.C.) SUBSEQUENT ADDITIONAL ROUTES OF GEOMETRIC GREECE (900 - 700 B.C.) 0

200

400

600

800

1000 km

Figure 3. Evolution of trade routes in the Near East and Eastern Mediterranean from Bronze Age through early Iron Age times, superposed on plate boundaries. Trade in tin and copper from 2200 to 1900 B.C. is from Kuhrt (1998) and that for Geometric Greece (900–700 B.C.) is from Coldstream (2003), using distribution of late Bronze Age stone anchors in the marine realm from McCaslin (1980, route 2).

a settlement where an expectation of change produced more receptivity to “civilized” goods. This trade-route clue and two others suggest the importance of long-term cultural response in the observed spatial relation. One of these is a relation between cultural character of civilizations and their distance to active tectonic boundaries. Those farthest from these boundaries tended to endure longer times with essentially the same character (Force, 2008, Fig. 2 therein), i.e., they were more static (and perhaps used their building material resources more slowly). A third clue suggesting the part played by cultural response is the link with cultural change, especially religious change, corresponding to tectonic events. Earthquakes in Greece and Cyprus in the fourth century A.D. correspond to changes in predominant religion in Corinth and in Kourion, respectively (Rothaus, 1996; Soren and James, 1988). An earthquake in Sparta ca. 464 B.C. provided the opportunity for a revolution (noted by Thucydides; see de Boer and Sanders, 2005). Earthquakes in Gortyn on Crete separate its Roman-era history into phases (DiVita, 1996). An earlier earthquake in southern Crete separates the Bronze Age archaeology of Kommos into

different modes pre– and post–1700 B.C. (Shaw, 2006). Many other examples (reviewed by de Boer and Sanders, 2005; Nur, 2008) provide abundant evidence of societal change catalyzed by earthquakes. Thus, we have some clues about the structure of a relation between ancient civilizations and active tectonism, but many questions remain, and explanations remain unclear. It therefore seems appropriate to turn to the potential contribution that archaeoseismology can make in explicating the true nature of the relation. This discipline (in cooperation with other archaeologists) is in a key position to provide clues that can be supplied no other way. There are at least three possible avenues of investigation that seem promising; they will now be discussed. ARCHAEOSEISMOLOGY IN DELINEATING AN EVOLUTION OF SOCIETAL RESPONSES Archaeoseismologists have concentrated until now on providing evidence of seismic destruction at ancient sites, and have been very successful at many of them despite many doubts

26

Force and McFadgen

among earlier archaeologists (reviewed by Nur, 2008). Progress in this direction will undoubtedly continue, and it could be harnessed in new ways. First of these is to determine the progression of societal response following destructive events. The work of many archaeologists has shown that habitation layers below and above seismic destruction horizons are different. A variety of causes have been attributed. An especially intriguing example to archaeoseismologists is the increasing sophistication of antiseismic devices through antiquity (cf. Stiros, 1996). We have seen examples of cultural discontinuities that apparently correspond with horizons giving evidence of tectonic activity. Perhaps most remarkable in this regard is the work of the late Klaus Kilian at Mycenaean Tiryns; he showed that archaeologic evidence of earthquakes corresponds temporally in at least three horizons with the emergence of newly dominant pottery styles. He concluded “earthquakes marked the beginning of a new phase and were related to, or even responsible for, changes in the organization and planning of the site” (Kilian, 1996, p. 67). Discontinuities of this sort can be modeled in a number of ways. One end member is represented by a combination of archaeologic and historic evidence about the 1855 earthquake of magnitude 8.2 in the Wellington area of New Zealand, which emptied nearly all kitchen cupboards and smashed the ceramics they contained (Grapes, 2000). The destroyed ceramic population was a mixture of modes that went back several decades, whereas their replacements would have been chosen from what was in vogue at that time (McFadgen and Clough, 2009). The same principle could apply to the architecture of the containing structures and even town layout, as was better demonstrated by the 1848 Marlborough earthquake (Grapes, 2000) and 1931 Napier earthquake of New Zealand. This type of case amplifies change that has already taken place, but in itself would generally not stimulate evolutionary change. To some extent, change is only stylistic. At another extreme, there is cultural change that involves innovation, values, and/or cultural evolution. For example, the 1755 “Lisbon” earthquake set in motion important philosophical changes throughout the Age of Enlightenment world (de Boer and Sanders, 2005). We have seen examples that involve religious change. This cultural-discontinuity model too has historical support. Rozario (2007) reviews evidence of accelerated cultural change initiated by disasters, including its treatment in the economic and psychological literatures. A third end member can be change in habitation patterns as a result of changes in the physical environment of the site, from sea level to landslides, that result from earthquakes (cf. McFadgen, 2007). For all three models, tectonism disrupts societal inertia and thus forces the pace of change. These three end-member models apply in some respects to cultures independent of their complexity. Indeed, it might be useful to better understand the cultural response to tectonism of many different types of cultures, not just the very complex ancient ones addressed here. There is some interest in the anthropological community for this question (Eiselt, 2009).

If a net long-term effect of tectonic activity in antiquity was an evolution toward civilization, we would like to know how the observed shorter-term changes might link up to contribute to this evolution. This will require the attention of archaeoseismologists in conjunction with archaeologist colleagues to determine the responses that are immediate, in closest association with earthquake damage, and those evolutionary changes that occur in response. It would be quite instructive to know at Tiryns, for example, whether characteristic types of pottery or any other cultural material closely followed seismic events but preceded the establishment of longer-lived styles. It would also be interesting to know whether any of these changes themselves form a pattern across a succession of seismic events. ARCHAEOSEISMOLOGY IN ESTABLISHING TYPICAL SEISMIC STYLES OF ANCIENT CIVILIZATIONS A tendency of ancient civilizations of the old world to be located along plate boundaries of transcurrent to convergent type seems clear (Fig. 1). For a number of these civilizations, however, the most pertinent tectonic micro-environment involves extensional faulting in association with such boundaries, as, for example, extension above subduction zones (Table 1). This implies that particular tectonic styles, including seismic styles, are associated with ancient civilizations, and because societal response of some sort must be involved, one would suspect that the styles in question must have characteristic recurrence intervals and/or intensities of events. Recurrence intervals less than human lifetimes (even those of antiquity) seem typical of the pertinent tectonic structures, and the close proximity of originating sites to those structures (Force, 2008) shows that these recurrence intervals did indeed affect civilizing societies. In Assyrian Ninevah, for example, typical recurrence intervals were of sufficient importance and regularity that scribes recorded them as being 21 yr, a figure similar to that derived by Kilian (1996) for Tiryns. Much more could be known about the relations among recurrence, intensity, and the development of civilized society. For example, are there optima in recurrence or intensity that accelerate this development relative to locales with too much activity (leading in extreme cases to abandonment) or too little? Archaeoseismology can potentially establish the chronologies and intensity records upon which the answers must be based. ARCHAEOSEISMOLOGY OF NEOLITHIC SITES In the area of this study, almost all the great ancient civilizations grew from indigenous roots (Aryan India being the one exception), though all those called derivative did so with inputs from adjacent established civilizations. These inputs could have spread to indigenous societies in many directions, but they seem to have preferentially spread instead along tectonic boundaries. It is therefore the influence of tectonism on indigenous prehistoric

Tectonic environments of ancient civilizations societies that needs attention, and the links that lead to civilization must be sought among them. In the region of this volume, this means Neolithic cultures, and some later cultures that retained Neolithic lifestyles as metal trade goods began to arrive (reviewed by Renfrew, 1972; Redford, 1992; Whittle, 1996, for different regions). The case for tectonic influence for some Neolithic (and “Chalcolithic”) societies is stronger than that for the civilizations that evolved from them because those precursors were closer to the tectonic boundaries. This is especially the case for the complex pre–Bronze Age cultures of Mesopotamia and the IndusSaraswati area, as located by Maisels (1999). The first indications of advanced culture arose in the foothills east of the Tigris (at sites like Choga Mami; Oates and Oates, 1976) and west of the Indus (Mehrgarh and Nausharo; Kenoyer, 2000) near the thrust components of partitioned plate boundaries. Some of these in Mesopotamia and the Indus are the Neolithic sites, from which come the previously mentioned evidence of abundantly productive springs associated with active faults. Only when the need for increased irrigation scale arose did the nascent civilizations move away from those boundaries. For the formative stages of these cultures, we of course have no written records; archaeology alone can supply reliable information, and archaeoseismology alone can supply information on the influence of tectonism on these societies. This can then be contrasted with the development of coeval societies in quiescent settings. Since nominally Neolithic societies have survived into historic times in many parts of the world, a complementary approach would be to study effects of tectonism on their lifestyles and history. The processes that influenced cultural change in these societies might contribute to understanding the processes of cultural change that led to civilization elsewhere. McFadgen (2007), for example, showed ties between tectonism and cultural change based on precontact archaeology for the New Zealand Maori. CONCLUSIONS The dynamics of long-term change in cultures subject to active tectonism are ambiguous, but, in contrast, the eventual result seems fairly clear in the area treated in this volume— those cultures that, in addition to having a range of environmental advantages, were subjected to particular varieties of tectonic activity tended to become exceedingly complex (“great ancient civilizations” and some particularly precocious Neolithic precursors). Though the observed spatial relation of ancient civilizations with active tectonism associated with the southern boundary of the Eurasian plate is most apparent at hemispheric scale, it seems valid in some areas at finer scales also. A previously documented closer relation to tectonic boundaries of derivative compared to primary civilizations suggests a relation to trade, and comparison of tectonic boundaries with trade-route propagation

27

toward emerging sites of civilization shows some startling resemblances. Two other lines of evidence also suggest that long-term cultural evolution may be directly related to tectonic activity— correlation between static character and location of civilizations relative to tectonic locus, and archaeologic and historic records of accelerated cultural (especially religious) change following tectonic events. If this is the case, anthropological comparisons of “tectonic” and “nontectonic” societies of different sorts seem promising. The existence of any spatial relation between ancient civilizations and active tectonism raises many questions, and most of them remain unresolved. Archaeoseismology is in a unique position to provide important clues in at least three areas: 1. Previous work with seismic destruction horizons has shown corresponding discontinuities in ceramic and/or architectural remains; these tectonic events have forced the pace of change but may or may not reflect cultural discontinuities. Input from the archaeoseismologist is required to differentiate first responses from subsequent ones in order to establish an evolution of changes, and to see if these link up from one event to the next to delineate any kind of long-term influence. 2. The style of societal response may vary with the typical recurrence interval between destructive events and with their intensities. Such variations are implicit in the relation of ancient civilizations to specific tectonic environments—especially extensional faulting within convergent to transcurrent boundaries— and in the typical distances of originating sites to seismogenic structures. The archaeoseismologist is indispensible in compiling the chronologic and intensity records required to establish such links. 3. The region treated in this volume is blessed with ancient historical records in addition to archaeoseismologic evidence. However, the relation probably precedes the historic period; the relation behaves as if certain Neolithic cultures were somehow favored by tectonic factors to become civilizations with histories. The archaeoseismologist is a necessary source of relevant information on the ways in which tectonism influenced Neolithic cultures. ACKNOWLEDGMENTS Force thanks all those who responded to his former website. As presented here, he included thoughts contributed by John Dohrenwend, Daniel Kent, Barbara Mills, Arda Ozacar, Evelyn Roeloffs, David Soren, and Jane Force. McFadgen thanks members of the New Zealand Archaeological Association at the 2009 annual conference for their responses to some of the ideas presented here, and Anne French for useful discussion. We appreciate reviews by Claudio Vita-Finzi and Victoria Hopgood (Buck). Logistic support was provided by Jim Bliss, Jennifer Dodge, and John Birmingham. This article is a contribution to the International Geoscience Programme IGCP 567, “Earthquake Archaeology: Archaeoseismology along the Alpine-Himalayan Seismic Zone.”

28

Force and McFadgen

REFERENCES CITED Bailey, G., King, G., and Sturdy, D., 1993, Active tectonics and land-use strategies: A Paleolithic example from northwest Greece: Antiquity, v. 67, p. 292–303. Childe, V.G., 1950, The urban revolution: The Town Planning Review, v. 21, p. 3–17. Coldstream, J.N., 2003, Geometric Greece (2nd ed.): London, Routledge, 453 p. de Boer, J.Z., and Sanders, D.T., 2005, Earthquakes in Human History: Princeton, New Jersey, Princeton University Press, 278 p. de Boer, J.Z., Hale, J.R., and Chanton, J., 2001, New evidence for the geological origins of the ancient Delphic oracle: Geology, v. 29, p. 707–710, doi: 10.1130/0091-7613(2001)029<0707:NEFTGO>2.0.CO;2. DiVita, A., 1996, Earthquakes and civil life at Gortyn (Crete), in Stiros, S., and Jones, R.E., eds., Archaeoseismology: Fitch Laboratory Occasional Paper 7 (British School at Athens), p. 45–50. Eiselt, B.S., 2009, Americanist archaeologies: 2008 in review: American Anthropologist, v. 111, p. 137–145, doi: 10.1111/j.1548-1433.2009.01106.x. Force, E.R., 2008, Tectonic environments of ancient civilizations in the Eastern Hemisphere: Geoarchaeology, v. 23, p. 644–653, doi: 10.1002/gea.20235. Grapes, R., 2000, Magnitude Eight Plus—New Zealand’s Biggest Earthquake: Wellington, Victoria University Press, 208 p. Kenoyer, J.M., 2000, Early developments of art, symbol, and technology in the Indus Valley tradition: Indian Archaeological Studies, v. 22, p. 1–18. Kilian, K., 1996, Earthquakes and archaeological context at 13th century BC Tiryns, in Stiros, S., and Jones, R.E., eds., Archaeoseismology: Fitch Laboratory Occasional Paper 7 (British School at Athens), p. 63–68. King, G., Bailey, G., and Sturdy, D., 1994, Active tectonics and human survival strategies: Journal of Geophysical Research–Solid Earth, v. 99, no. B10, p. 20,063–20,078. Kuhrt, A., 1998, The Old Assyrian merchants, in Parkins, H., and Smith, C., eds., Trade, Traders, and the Ancient City: London, Routledge, p. 16–30. Maisels, C.K., 1999, Early Civilizations of the Old World: London, Routledge, 504 p. McCaslin, D.E., 1980, Stone anchors in antiquity: Coastal settlements and maritime trade-routes in the Eastern Mediterranean ca. 1600–1050 BC: Studies in Mediterranean Archaeology, Volume 61: Goteborg, Astroms, 145 p. McClusky, S., Balassanian, S., Barka, A., Demir, C., Ergintav, S., Georgiev, I., Gurkan, O., Hamburger, M., Hurst, K., Kahle, H., Kastens, K., Kekelidze, G., King, R., Kotzev, V., Lenk, O., Mahmoud, S., Mishin, A., Nadariya,

M., Ouzounis, A., Paradissis, D., Peter, Y., Prilepin, M., Reilinger, R., Sanli, I., Seeger, H., Tealeb, A., Toksöz, M.N., and Veis, G., 2000, Global Positioning System constraints on plate kinematics and dynamics in the Eastern Mediterranean and Caucasus: Journal of Geophysical Research, v. 105, no. B3, p. 5695–5719. McFadgen, B., 2007, Hostile Shores—Catastrophic Events in Prehistoric New Zealand and Their Impact on Maori Coastal Communities: Auckland, Auckland University Press, 298 p. McFadgen, B., and Clough, R., 2009, Foreshock-Aftershock: The Archaeology of Chews Lane, Wellington: Auckland, Clough and Associates Report 6, 106 p. Nur, A., with Burgess, D., 2008, Apocalypse: Earthquakes, Archaeology, and the Wrath of God: Princeton, New Jersey, Princeton University Press, 309 p. Oates, J., and Oates, D., 1976, Early irrigation agriculture in Mesopotamia, in Sieveking, G.G., Longworth, I.H., and Wilson, K.E., eds., Problems in Economic and Social Archaeology: London, Duckworth, p. 109–134. Redford, D., 1992, Egypt, Canaan, and Israel in Ancient Times: Princeton, New Jersey, Princeton University Press, 512 p. Renfrew, C., 1972, The Emergence of Civilization; The Cyclades and the Aegean in the Third Millennium BC: London, Methuen, 595 p. Rothaus, R.M., 1996, Earthquakes and temples in Late Antique Corinth, in Stiros, S., and Jones, R.E., eds., Archaeoseismology: Fitch Laboratory Occasional Paper 7 (British School at Athens), p. 105–112. Rozario, K., 2007, The Culture of Calamity: Disaster and the Making of Modern America: Chicago, University of Chicago Press, 313 p. Shaw, J.W., 2006, Kommos: A Minoan Harbor Town and Greek Sanctuary in Southern Crete: Athens, American School of Classical Studies, 171 p. Soren, D., and James, J., 1988, Kourion—The Search for a Lost Roman City: New York, Anchor, 223 p. Stiros, S.C., 1996, Identification of earthquakes from archaeological data: Methodology, criteria, and limitations, in Stiros, S., and Jones, R.E., eds., Archaeoseismology: Fitch Laboratory Occasional Paper 7 (British School at Athens), p. 129–152. Trifonov, V.G., and Karakhanian, A.S., 2004, Active faulting and human environment: Tectonophysics, v. 380, p. 287–294, doi: 10.1016/j.tecto.2003 .09.025. Whittle, A., 1996, Europe in the Neolithic: Cambridge, UK, Cambridge University Press, 460 p. MANUSCRIPT ACCEPTED BY THE SOCIETY 19 MAY 2010

Printed in the USA

The Geological Society of America Special Paper 471 2010

The door knockers of Mansurah: Strong shaking in a region of low perceived seismic risk, Sindh, Pakistan Roger Bilham Cooperative Institute for Research in Environmental Science and Department of Geological Sciences, University of Colorado, Boulder, Colorado 80309-0399, USA Sarosh Lodi Department of Civil Engineering, NED University of Engineering and Technology, Karachi 75270, Pakistan

ABSTRACT Mansurah, the eighth-century Arabic capital of Sindh province, Pakistan, flourished for a mere 200 yr. Its destruction by an earthquake ca. 980 A.D. was first proposed by archaeologists who reported the discovery of crushed skeletons amid dateable coins found among its rubble. An abrupt natural death to the city was challenged by others who noted that the absence of wood or valuables was consistent with the city being sacked and systematically looted. The recent discovery of four decorated door knockers beneath the collapsed walls of one of the largest structures in Mansurah, however, reopens the case for an earthquake, since an invading army would almost certainly have removed them as booty. We suggest that an earthquake not only destroyed the city and its suburbs (intensity ≈ VIII), but resulted in postseismic avulsion of the river on which its citizens depended for agriculture, sanitation, and trade. Since natural levees have been observed in India to collapse in intensity VII shaking, it is unnecessary to invoke coseismic uplift as a requirement for upstream river avulsion. The absence in the past two centuries of large earthquakes in the region has resulted in central Sindh being depicted as a region of low seismic hazard, yet in 1668, in the same province, an earthquake destroyed nearby Samawani and also initiated avulsion of the Indus. A case can be made for reevaluating the five millennia of archaeological ruins in Pakistan to establish a long-term view of seismicity unavailable from the short instrumental record.

INTRODUCTION

historical cities that had been described by Arabic and Mughal historians. Bellasis used the name Brahminabad, but later writers have referred to the ruins as Mansura, al-Mansurah, Bhramanabad, and Bhamanabad, and Cousens (1929) lists a dozen more. It is now known that the ruins described by Bellasis were the ruins of al-Mansurah, the capital of the Arabic province of Sindh established ca. 734 A.D. It is less accepted that Mansurah was constructed on the ruins of the earlier Hindu city of

The ruins at Mansurah (25.882°N, 68.777°E) were brought to the attention of the archaeological community in a series of reports published in Bombay and London magazines in the mid-nineteenth century (Bellasis 1857a, 1857b; Sykes, 1857a, 1857b). At the time of writing, and for many years thereafter, the numerous ruined sites in the region were identified with different

Bilham, R., and Lodi, S., 2010, The door knockers of Mansurah: Strong shaking in a region of low perceived seismic risk, Sindh, Pakistan, in Sintubin, M., Stewart, I.S., Niemi, T.M., and Altunel, E., eds., Ancient Earthquakes: Geological Society of America Special Paper 471, p. 29–37, doi: 10.1130/2010.2471(03). For permission to copy, contact [email protected]. © 2010 The Geological Society of America. All rights reserved.

29

30

Bilham and Lodi

Bhamanabad (Farooq, 1986). The evolving discussion concerning the nomenclature of the site and nearby urban centers can be followed in Elliot (1867), Cunningham (1871), Haig (1874, 1894), Raverty (1893), Cousens (1905, 1929), Panhwar (1983a, 1983b), Wheeler (1992), Hodīvālā (1939), Farooq (1986), and most recently Khan (1990). To avoid confusion, and since the identification of the city with earlier ones is irrelevant to the present discussion, we shall refer to the ruins as Mansurah. The vast area of rubble (2 km × 1.5 km) and the apparently orderly layout of the city led Bellasis to characterize Mansurah as the “Pompeii of the East,” a city frozen in time with many of its citizens interred within its ruins. The ruins are located 30 km east of the present path of the Indus, 60 km NNE of Hyderabad and 16 km ESE of Shahdadpur, Pakistan (Fig. 1). They consist of heaps of fragmented bricks separated by rubble-clogged ancient streets in orthogonal rows on a gently undulating area elevated above the meander of an abandoned river (Figs. 2 and 3). Mansurah was a planned city. Some of the streets were 65 m wide and were paved with bricks usually laid on edge (longest and narrowest dimension down), and underlain with wastewater drains (Farooq, 1986). In places, wastewater and sewage were led via conduits to terracotta-lined soakaway pits (Khan, 1990). The ruins are encircled by the remains of a 3-m-wide fortified wall with a perimeter length of 6.4 km, which some accounts (Elliot, 1867) describe as being surmounted by 1200 bastions— clearly an exaggeration if these ornamented the perimeter of the main city. A more conservative estimate is provided by Abul Fazl writing in the sixteenth century (Jarrett, 1891), who enumerated the number of ruined bastions as 140, spaced ~50 m apart (Cunningham, 1871). This very closely matches the measured perim-

eter, which would have required a 45 m mean spacing between 140 bastions. Khan (1990) describes the bastions as semicircular and spaced at 33 m intervals with remnants found at a height of 3.5 m. The remains of the bastions are now indistinguishable from the irregular mounds that characterize the level of the city. The several Arabic accounts that describe the founding of the city describe it as an island surrounded by a branch of the Mihran (the Indus). It was noted for its verdure and cleanliness, although some accounts complain of abundant fleas (Elliot, 1867). In the nineteenth century, at the time of the annexation of Sindh to India, the region was arid, with water found only in wells. Twentieth-century irrigation fed by the nearby Jamaro canal has now ponded parts of the abandoned river, which is distinctly concave where it abuts the eastern walls of the city. Elsewhere the ancient path of the river is obscured by agriculture. The description of the city as occupying an island is consistent with its current elevation 5–10 m above the shore of the present water in the river. The straightness of the SW wall and its continuation to a contiguous ruined settlement to the SE are suggestive of the existence of an artificial moat to the SW rather than a river meander. The path of the Jamaro canal was excavated with a 5° inflection to avoid crossing the Mansurah ruins (Fig. 1). Dominating the ruins, there is a single masonry tower known as the Thul (Fig. 4) that afforded subsurface access to a well (Cousens, 1929). The bricks of the town consist of thin fired clay bricks cemented with mud, but in the Thul, these bricks were cemented above ground with a gypsum-based cement. The foundations of the Thul used much larger bricks. Few intact bricks remain in the town, presumably because they have been scavenged for construction elsewhere. Pottery shards and glazed

2 km

Shahdadpur Mansurah Fig. 3 Fig. 2 Fig. 4

Mansurah Berani Tando Alam

10 km Figure 1. Google map view of Mansurah and its suburbs. A canal passes SW and a drain runs NE of the ruins, which show as light gray. An oxbow lake has now occupied the formerly abandoned river to the east of the city. Arrows to the right show photo angles of Figures 2–4.

Strong shaking in Mansurah, a region of low perceived seismic risk in Sindh, Pakistan

31

earthenware fragments are abundant. Coins found among the ruins in the past 150 yr are mostly copper, with some silver coins. A single gold coin has been reported, the location of which is now unknown (Khan, 1990). Even the soils and dust of the ruins have been removed by local farmers who have found their composition desirable as a fertilizer (Cousens, 1929). EVIDENCE FOR AND AGAINST AN EARTHQUAKE According to Bellasis (1857a, 1857b) and Sykes (1857), Mansurah was destroyed by an earthquake sometime after 975 A.D. Their evidence comes from the large number of skeletons discovered during excavations in doorways and room corners, and from the dates of coins scattered throughout the ruins: Figure 2. Close-up of pottery and brick shards with limestone fragments typical of the Mansurah ruins. The muzzle to butt length of the gun is ~80 cm.

Figure 3. A view from the eastern side of the Mansurah ruins near the perimeter wall facing north showing the abandoned river course, now occupied by a lake.

The human bones were chiefly found in doorways, as if the people had been attempting to escape, and others in the corner of rooms. Many of the skeletons were in sufficiently perfect state to show the position the body had assumed; some were upright, some recumbent with their faces down, and some crouched in a sitting posture. One in particular I remember finding in a doorway; the man had evidently been rushing out of his house, when a mass of brickwork had in its fall crushed him to the ground, and there his bones were lying extended full length and the face downwards. (Bellasis, 1857a, p. 417)

The coins provide the latest date for the layer of widespread collapse of structures in the city. Bellasis in his excavation found thousands of badly corroded specimens that were passed (in a 14 kg bag) to experts for identification (Thomas, 1858). The reports on these focused on the earliest coins indicative of minting in Mansurah (e.g., A.D. 750), and dwelled little on the details of coins of younger vintage, and hence the date of the inferred earthquake is poorly bracketed. The latest coins suggest internment after A.D. 975, but the date could be earlier according to Haig (1874, p. 287), who noted that Mansurah was in ruins “when

Figure 4. The Thul at Mansurah in 1897 (left from Cousens, 1929) and in 2008 (right). The integrity of this 12-m-high structure, which is bonded with a tough mortar, suggests that in the past millennium, the maximum shaking intensity in the region cannot have much exceeded Mercalli intensity X. The faced reentrant corner remains approximately vertical. Cousens excavated through the foundation of the Thul to virgin soils at a depth of 5 m below ground level. The 3-m-thick walls enclosed a 2.2 m central spiral staircase leading to a well within.

32

Bilham and Lodi

Biladuri wrote his Futuh-as-Sindh—perhaps about 870–880 A.D.” It is not certain on what authority Haig formed this conclusion because Biladuri died ca. 892 A.D. never having visited Sindh. Biladuri may have been confused by the identity of Mansurah and Brahmanabad mentioned above. Moreover, Elliot’s (1867, p. 122) translation of the Arab conquest indicates that ruination was a consequence of the battle for the city ca. 664 A.D.: “Then Muhammad, son of Kasim, went to old Brahmanabad, two parasangs from Mansurah, which town did not then exist, its site being a forest. The remnant of the army of Dahir rallied at Brahmanabad and resistance being made, Muhammad was obliged to resort to force, when 8, or as some say 26 thousand men were put to the sword. He left a prefect there. The place is now in ruins.” Returning to the ca. 980 A.D. damage following the establishment of the Arab capital, Cunningham (1871) also favored the destruction of the city by an earthquake. He suggested that the sack of the city by an aggressor was unlikely because of the absence of reports of charred timber, pillage being invariably preceded or followed by arson in medieval warfare. He found compelling evidence in the disposition of crushed skeletons, especially a quote by Richardson cited in Bellasis (1857a, p. 423), who described a brick “which entered corner-ways into a skull, and which, when taken out had a portion of the bone adhering to it.” His unstated implication is that a brick, particularly the lightweight Arabic tabular brick, would not have been a weapon of choice for an invading army. Cousens (1929), however, who substantiated the absence of a burn layer, dismissed earthquake damage. He pointed to a slow decline in the city that may have persisted to the thirteenth century before its abandonment. He found the absence of any gold coins or other items of value to support an alternate interpretation— warfare followed by systematic looting. He invoked the scattered skeletons as characteristic of a massacre. Citizens found in doorways, rooms, and streets were killed and interred by invaders intent on hurriedly dismantling a city. Cousens (1929, p. 71) adds, “Walls were thrown down in order to get at the door and window frames and roof timbers; and being brick ornamented with mud, were easily overturned with this rough treatment.” Panhwar (1983b) also attributed the destruction of the city to a punitive army, citing Farrucki’s account of Mahmud of Gazni sacking the city in 1025. It is possible that the city had by then had recovered economically from earthquake damage, sufficiently for survivors to have attracted a punitive attack, but had not yet reconstructed its defensive walls, rendering it an easy target for Mahmud’s army. Although the historical evidence for or against an earthquake remains inconclusive, there is no disagreement on two issues: that around the end of the tenth century, Mansurah underwent a catastrophic change—widespread ruin, followed by a significant decline in the availability of water. The ultimate abandonment of the city is linked, by all, to a change in the course of the river. This chronological sequence of damage followed by a change in hydrology is deduced from archaeological excavations. Cousens’ excavations in 1897 and 1909 revealed three layers of stratig-

raphy at the site: a pre-Muslim layer of occupation, a layer of orderly city structures (Mansurah), and finally a layer of disordered structures with numerous cylindrically lined wells that cut through the two lower levels, suggesting the drying up of the river and the need to tap groundwater supplies. Cousens theorized that following the sack of the city, the survivors reconstructed the city from the ruins of Mansurah and needed to sink wells to access drinking water. Excavations by the Department of Archaeology, Pakistan, have added details to this layered chronology (Farooq, 1986; Khan, 1990), and although the earlier confusion between ancient Brahmanabad and Medieval Mansurah remains unsettled, the excavations agree on the impoverishment of the most recent structures, and the apparent dependence of the later citizens on well water. The village that now occupies the westernmost part of the site also uses well water. Previous historians have considered the timing of the destruction of the city and the drying up of the river to be a coincidence. Certainly, had the events occurred in reverse order, we should agree with them. However, the historical record of large earthquakes contains many examples of rivers for which courses have changed following a nearby earthquake. A well-known case is the permanent shift in the geometry of Mississippi following the 1811–1812 New Madrid earthquakes (Johnston and Schweig, 1996). Other examples are described by Schumm (2005). Assuming that the river near Mansurah was confined within banks and levees not more than 3 m above the normal level of the river, an earthquake with uplift exceeding 3 m would be required to divert the course of the river. This would require a substantial earthquake (e.g., M >7) and the development of a surface uplift feature for which there is no evidence. We note, however, that strong shaking from a deep earthquake (associated with minimal or zero local uplift) could also trigger avulsion. To have affected the flow near Mansurah, avulsion of the river must have occurred upstream. During the 1811 New Madrid (Schumm et al., 2000) and 1897 Shillong earthquakes (Oldham, 1899), the most striking earth movements near rivers involved the collapse of river banks. In the New Madrid earthquake, the banks of the river plunged into the river, creating waves that further undermined the shaken banks, and briefly raised the bed level, impeding channel flow. Whether or not the natural levees that confined the river north of Mansurah would open to permit flooding of the hinterland would depend to some degree on their width, but fissuring and slumping would no doubt have weakened their ability to confine the river in a flood. Overbank flooding may not have been immediate, but in the absence of a reliable levee, extreme flood levels would no longer be a requirement. Flooding could therefore have occurred during the first heavy monsoon. Once the river breached through the banks of the river upstream, avulsion and the abandonment of the former channel would have been almost inevitable. We thus find the account of city collapse followed by a change in the course of the river, typical of earthquakes elsewhere, to represent a neglected consideration relevant to the earthquakerelated destruction of Mansurah. The 1668 Samawani earthquake

Strong shaking in Mansurah, a region of low perceived seismic risk in Sindh, Pakistan (Ambraseys, 2004), seven centuries later but in the same Mughal province of Nasirpur, appears also to have initiated avulsion of the Indus (Bilham et al., 2007). River shift was not immediate. In the half century following the 1668 earthquake, the river slowly shifted its course westward, eventually finding a path to the west of Hyderabad. Cities were hastily founded and abandoned in a half-century-long attempt to keep up with its evolving path, eventually leading to the establishment of Hyderabad in 1768 (Haig, 1894). THE DOOR KNOCKERS OF MANSURAH The collapsed remains of three large structures were exhumed in the post-1966 excavations: a grand mosque (50 m × 80 m with 2-m-thick walls), and two nearby civic buildings with slightly smaller scales (Farooq, 1986; Khan, 1990). These ancillary buildings are interpreted to represent administrative assembly buildings or schools having roofs supported by 1.8-m-square brick pillars. Deep beneath the prodigious quantities of rubble from one of these enormous buildings, four ∼50 kg bronze door knockers were found (Fig. 5). Their improbable survival owes everything to the depth of their burial. The size and quality of these decorative knockers—50 cm wide, 80 cm high, and 10 cm deep—made them a significant feature of this civic structure (Khan, 1990), and it is doubtful that a pillaging army, had they seen them, would have failed to regard them as a valuable souvenir. The knockers were fastened by 2-cm-diameter bolts or nails to what may have been the largest wooden doors in the city; the removal of this material elsewhere, Cousens (1929) argued, was the reason for the collapse of the structures in which they were embedded. We consider it doubtful that an invading army would have the resources or inclination needed to demolish these three large civic structures. To demolish them without first removing these attractive souvenirs appears to us even less likely. The survival of these remarkable fittings is presumably because they were buried by the collapse of the large building of which they were part. The pile of rubble was presumably too daunting to be removed by survivors or subsequent scavengers. It is difficult to escape the conclusion that the door knockers of

33

Mansurah were buried by the collapse of the entrance of a civic structure in an earthquake. DISCUSSION From the foregoing arguments, we present evidence that favors the destruction of tenth-century Mansurah by an earthquake of sufficient severity to tumble the 2-m-thick walls of the huge central Mosque, and other equally substantial civic structures. Much of the city collapsed with these larger structures, but survivors and scavengers were able to remove most of what was of value from the smaller rubble piles of private dwellings and shops of the commercial districts. The Thul, a tower assembled from numerous courses of bricks bound together with a strong cement fared better in the earthquake and has survived ten centuries of weathering. It has also survived determined efforts to dismantle it for its bricks. From these observations, we conclude that the brick-andmud structures of the city were destroyed by at least MSK intensity VII, and probably intensity VIII shaking. The conclusion is similar to that formed by Quittmeyer et al. (1979), who indicated a solitary intensity VIII on their Figure 2. The survival of the Thul (Fig. 4) provides an upper limit to accelerations since its construction ca. 800 A.D. We estimate that it could probably survive intensity X but that higher intensities would have disrupted the structure. The foundations of the tower itself have not been exposed to soil liquefaction, because, despite the fact that its foundations extended to sediments near the medieval water table, its few remaining finished walls appear approximately vertical. The archaeological excavations reveal no obvious examples of liquefaction or warping of the drainage courses within the city. No excavations have been undertaken in the lower ground adjoining the city where liquefaction would have been more likely. Subsequent avulsion of the river is also consistent with intensity VIII shaking. A single breach in a levee upstream caused by bank collapse would be sufficient to facilitate a shift in the river in flood. The intensity of shaking required to do this can be inferred from bank integrity at the time of the Mw = 8.1 1897

Figure 5. The bronze door knockers of Mansurah (Khan, 1990). The diameter of each 1-cm-thick circular plate is 56 cm, with raised relief of 17 cm and an average weight of 53 kg. The Sufic characters engraved on the outer annulus are from the Qu’ran and include the name of the Habbari ruler Abdullah.

50 cm

34

Bilham and Lodi

Great Assam earthquake. Bank collapse was noted by LaTouche in areas assigned MSK intensity VII or greater on the Brahmaputra and Surma River systems (Bilham, 2008). Thus, one would not need to invoke vertical deformation of the form recorded in the 1819 Allah Bund earthquake to restrict or divert flow of the river, nor would avulsion need to have occurred immediately after the earthquake. It may have occurred in the first severe monsoon floods a year or more later. The ensuing shift in the river would have made the site less attractive to survivors, both from the resulting restricted availability of water, but presumably also because the city had lost much of the river trade that supported its former prosperity.

Mach 1931

29

4.6

R.

4.4 4.8 4.4

us

Sibi 4.6

5.9

5.24.5

Ind

3.9 4.7

The causal faults of the ca. 980 Mansurah and Samawani 1668 earthquakes are currently unknown, and few clues to their present-day activity are available from recent microseismicity in the region (Fig. 6). The sediments of the Indus are known to be faulted at depth where these have been subjected to seismic prospecting, but none is known to cut the surface (Nakata et al., 1991). Several lineations have been mapped by Kazmi (1979) near Sibi, and to the SE of the river by Kazmi and Rana (1983) from Landsat imagery. Figure 7 shows an enhanced view of digital elevations in the Indus floodplain surrounding Mansurah, illuminated at an angle to minimize the artifacts of seams in the SRTM image. Undulations with wavelengths of 10–30 km, and

4.2

4.9

Muree 1852

4.8

4.6 4.5 4.8

Sukkur

4.4

H J-K

100 km

igh

3.9 4.3

5.44.9

Fig. 7

4.4 4.7

4.9

Jaisalmer 5.5

4.7

Nawabshah

5.2

4.9 4.5

Mansurah 980

4.4

Samawani 1668

Thar Desert

Hyderabad Karachi

Umarkot

4.8 4.9

Bhanbore

h ug Thatta s Tro Badin du r In e Low 5 5.1

4.7 4.7

4.6

4.4

Allah Bund 1819 4.1

24°N

Lakhpat 67°E

Bhuj 2001 71

Figure 6. Basement faulting between 25.7°N and 27.8°N is inferred from aeromagnetic contours of Zaigham and Mallick (1999). The two parallel lines between 68.3°E and 70.3°E indicate the approximate location of an inferred ancient rift system, and the shaded area delineates an area of shallow subsurface bedrock identified as the Jacobabad-Khairpur (J-K) High (Kazmi and Rana, 1982) that outcrops near Sukkur. Tick marks indicate down. Numbers indicate location and magnitude of instrumental epicenters (omitting the Bhuj 2001 aftershocks) from Villasenor and Engdahl (2007). Less accurate locations for M>3 events in central Sindh from the International Seismological Centre catalog are shown as open circles. Abandoned river channels following the 980 A.D. and 1668 A.D. earthquakes are shown as heavy dashed lines. Place names are in small italics; significant earthquakes are named with dates (Ambraseys and Bilham, 2003).

Strong shaking in Mansurah, a region of low perceived seismic risk in Sindh, Pakistan

35

27 +18 m

50 km

75

12

Jamaro Canal

m

14

/k

m

m

16

57

Figure 7. Shuttle Radar Tomography Mission digital elevation image of the Indus floodplain with smoothed 2 m contours referred to an arbitrary datum to the SE. Illumination is from N68E at an elevation angle of 5° (arrow). A dune field occupies the NE quadrant of the map. The 8 m contour indicates the approximate axis of a reduction in the mean down-valley slope from 75 mm/km in the 100 km north of Mansurah, to 57 mm/yr in the 100 km SE of Mansurah. Circles outline the pre-1668 bed of the Indus, which appears to follow a gentle dome in the topography. Numerous other abandoned paths are evident.

10

m /k

m

8

m

6

Mansurah 4 2 0

l Hyderabad

25°N 68

Indus

pre-1668 Indus

several lineations, are evident in the data, but the most prominent features of note are a subtle reduction of down-valley slope from 75 mm/km NW of Mansurah to 57 mm/km to the south, and gentle dome following the pre-1668 course of the Indus. It is tempting to conclude that the dome-like ridge is causal to the 1668 avulsion, since it is subparallel to the fold belt west of Hyderabad; however, it is possibly a sedimentary feature. The seismic significance of the down-valley change in gradient is unclear; however, the avulsion of the river in the tenth century and seventeenth centuries near this location is suggestive that block tilting may be active. The sense of slope change is, however, opposite to the sense of basement uplift described by Zaigham and Mallick (1999), suggesting that if the morphology is related to tectonics, a reversal of the geological sense of slip is now occurring. The principal seismicity in the surrounding region follows the fold belts along the Kirthar fold-and-thrust belt to the west, and the Sulaiman Range to the north (Fig. 6). Significant 6 < Mw < 7.6 earthquakes have occurred historically north of 28°N (Quittmeyer et al., 1979; Ambraseys and Bilham, 2003), and south of 25°N. Relatively minor seismicity occurs to the east, most notably a shallow Mw = 5.5 dextral earthquake that produced a surface rupture near Jaisalmer in 1992. The region near Mansurah shows little recent microseismicity.

70°E

With the exception of the uplift accompanying the Allah Bund earthquake, geological mapping reveals no recent faults that have disturbed the surface of the Indus sediments. However, Zaigham and Mallick (1999) using aeromagnetic methods identified offset structural features at 5–9 km depth beneath the cover of thin-skinned tectonics in the west that they projected eastward into the Indus basin. These trends suggest that Samawani and Mansurah both may be located above a buried horst 5 km beneath the sedimentary cover. The block on which they are apparently located is elevated 2 km above the contiguous structural block to the north. Strike-slip geological offsets are also observed on these structures, and Zaigham and Mallick (1999) argued that recent seismicity west of Nawabshah indicates that they are presently active. They proposed further that seismicity throughout the region is attributable to reactivation of faults related to an ancient allochogen along the western margin of India. The significance of a buried, poorly defined rift system in the Indus basin on assessing seismic hazards in the region is also discussed by Sawar (2004); however, the lack of clarity of structural features (Kazmi, 1979; Kazmi and Rana, 1982) beneath the Indus provides considerable room for speculation. Spatial variations in river sinuosity support the notion of present-day vertical tectonics in Sindh (Schumm et al., 2000;

36

Bilham and Lodi

Schumm, 2005). Jorgensen et al. (1993) demonstrated that the geometry of the Indus, and its evolving path, responds to geomorphic segmentation of the valley apparently related to tectonic forcing. Basin and domal structures recognized in geological investigations appear to be mimicked by present-day vertical displacements inferred from lateral variations in river sinuosity. The structural trends mapped by Zaigham and Mallick (1999) are orthogonal to the thin-skinned thrusts of the Kirthar range but may be understood in the context of the shear stresses of India’s convergence with Asia, which would tend to activate block rotation in the region. In order to account for the 1819 Allah Bund and 2001 Bhuj Mw ≥7.6 earthquakes Stein et al. (2002) proposed that the Sindh region has been fragmented and is converging with India as a result of this collision. If this is occurring, it must be doing so at rates of less than 1.5 mm/yr, the current rate of global positioning system (GPS) motion we have measured between Karachi and the Indian plate. Seismic Hazards in Sindh The occurrence of two damaging earthquakes NE of Hyderabad in the past millennium suggests that the structures beneath the Indus sediments must be considered active. The Mansurah earthquake resulted in intensity VIII, but it is uncertain how large were the intensities in the Samawani 1668 earthquake. The absence of damage reports from other towns suggests that both earthquakes may have been relatively modest (Mw <6.5). The assignment of any magnitude to an earthquake with information from one location is entirely speculative, and this magnitude should not be used for estimates of seismic risk. The absence of significant microseismicity, and the apparent absence of significant strain rates indicated in preliminary GPS results in the past 3 yr provide few clues on which to improve our knowledge of the source regions of the two events. The large number of ruins in the region of Mansurah, however, may in the future provide additional data on the scale of damage. The same can be said of the Samawani earthquake, although here we know of many more towns that existed in the seventeenth century that were apparently undamaged in 1668. Despite our lack of precision (no epicenter or magnitude has been identified for either earthquake), the two damaged towns lie in the current zone 2A of Pakistan’s tentative 2006 zonation map (cf. Bilham et al., 2007). In contrast, the GSHAP hazard map (Giardini et al., 1999) shows a lobe of higher shaking potential at the latitude of Mansurah, presumably accommodating Bellasis’s (1857a, 1857b) accounts and Quittmeyer et al.’s (1979, Fig. 2) earlier evaluation of intensity VIII damage at Mansurah. CONCLUSIONS The ruins of the eighth-century Arabic capital of Mansurah are comparable in scale to those of Taxila or Mohenjodaro, yet Mansurah flourished as a capital for less than 200 yr, and it had been completely abandoned by 1300. Coins indicating a promi-

nent layer of destruction prior to its final abandonment are dated some time near the end of tenth century. The recent discovery of a set of attractive bronze door knockers that were buried deeply beneath the ruins of one of the three largest civic structures in the ruined city provides compelling evidence that an earthquake was the cause of the destruction of the city. Had Mansurah been destroyed by an enemy, or simply abandoned due to river avulsion, these, like more accessible treasures and construction materials that have long since been removed, would have been stolen. Moreover, the deliberate destruction of these massive buildings by the inferred conquerors would have required considerable effort. We estimate that shaking intensities of at least VII and probably VIII caused the collapse of most of the dwellings and shops in this city of bricks and dried mud. Weak evidence for maximum shaking intensity X is inferred from the presence of a surviving tower that incorporated tough and tenacious mortar in its construction. We find no clear evidence for localized vertical crustal deformation near, or upstream from, Mansurah, although we note that the region lies close to a change in mean valley gradient. The valley 100 km south of Mansurah slopes ~25% less than the valley to the north. A change in the course of the river occurred following the inferred earthquake, possibly due to a change in valley slope, but more probably because of bank collapse and avulsion during a subsequent flood. From evidence elsewhere in India, we deduce that shaking intensity VII is sufficient to promote the collapse of river banks. Upstream collapse of a river bank may have created a gap in the natural levee that lined the river upstream, catalyzing avulsion. The inferred decline and eventual abandonment of Mansurah is attributable to the drying up of the river near the town. We note that a similar fate befell Samawani following the 1668 earthquake (Bilham et al., 2007). The location of Samawani has been lost since it was described by Hodīvālā (1939), but it is believed to have been located on or near a former bank of the Indus near the present town of Nasirpur, within the administrative district in which it was located in Mughal times (Sarkar, 1947, 1978; Habib, 1982). The river near Samawani shifted westward over the next century, eventually to a location west of the current city of Hyderabad, the foundation of which in 1768 is indirectly attributable to the 1668 earthquake. It has not been possible to identify a causal fault for either the Mansurah or Samawani earthquakes, and although several basement structures are known, microseismicity in the past 40 yr is unable to clarify which of these structures are active, or to confirm the boundaries of east-trending structural elements that have been established beneath the Kirthar range. Preliminary GPS data from Pakistan indicate that deformation in the region occurs at less than 2 mm/yr between Karachi and the Indian plate, suggesting low rates of seismic productivity, with a return time for large earthquakes that may be many hundreds of years. Despite the apparent seismic quiescence of the southern Indus plains, the occurrence of two damaging earthquakes in the past millennium suggests that seismic hazards in the region may be

Strong shaking in Mansurah, a region of low perceived seismic risk in Sindh, Pakistan underestimated. A systematic reevaluation of the five millennia of archaeological sites in Pakistan for their possible ruin by damaging earthquakes would be of immense benefit to supplement the relatively short instrumental record. ACKNOWLEDGMENTS This study was made possible by grant EAR-0739081 from the National Science Foundation. Walter Szeliga processed the global positioning system (GPS) data from Karachi reported in this paper. We thank our Sindh police escort for the use of the firearm used as scale in Figure 2. REFERENCES CITED Ambraseys, N.N., 2004, Three little known earthquakes in India: Current Science, v. 86, no. 4, p. 506–508. Ambraseys, N., and Bilham, R., 2003, Earthquakes and crustal deformation in northern Baluchistan: Bulletin of the Seismological Society of America, v. 93, no. 4, p. 1573–1600, doi: 10.1785/0120020038. Bellasis, A.F., 1857a, An account of the ancient and ruined city of Brahminabad, in Sind: Journal of the Bombay Branch of the Royal Asiatic Society, v. 5, p. 413–425. Bellasis, A.F., 1857b, Further observations of the ruined city of Brahminabad, in Sind: Journal of the Bombay Branch of the Royal Asiatic Society, v. 5, p. 467–477. Bilham, R., 2008, Tom LaTouche and the Great Assam Earthquake of 12 June 1897; letters from the epicenter: Seismological Research Letters, v. 79, no. 3, p. 426–437, doi: 10.1785/gssrl.79.3.426. Bilham, R.S., Lodi, S., Hough, S., Bukhary, S., Khan, A.M., and Rafeeqi, S.F.A., 2007, Seismic hazard in Karachi, Pakistan: Uncertain past, uncertain future: Seismological Research Letters, v. 78, no. 6, p. 601–631. Cousens, H., 1905, Conservation of ancient monuments in the Bombay Presidency, in Scott, R., ed., Journal of the Bombay Branch of the Royal Asiatic Society, Extra Number, Centenary Memorial Volume: London, Kegan Paul, Trench, Trübner & Co., p. 149–162. Cousens, H., 1929, The Antiquities of Sind, with Historical Outline: Oxford, Archaeological Survey of India, Imperial Series 46, 184 p. Cunningham, A., 1871, The Ancient Geography of India: London, Trubner and Co., 589 p. Elliot, H.M., 1867, A History of India by Its Own Historians: The Muhammadan Period, Volume 1: London, Trubner, 543 p. Farooq, A.A., 1986, Excavations at Mansurah (13th Season) 1974–1986: Pakistan Archaeology, v. 10–22, 35 p. Giardini, D., Grunthal, G., Shedlock, K., and Zheng, P., 1999, The GSHAP Global Seismic Hazard Map: Annali di Geofisica, v. 42, p. 1225–1230. Habib, I., 1982, An Atlas of the Mughal Empire: Political and Economic Maps with Detailed Notes, Bibliography and Index: Oxford, Oxford University Press, 105 p. Haig, M.R., 1874, On the sites of Brahmanabad and Mansurah in Sindh; with others of less note in their vicinity: Journal of the Royal Asiatic Society, v. 16, p. 281–294. Haig, M.R., 1894, The Indus Delta Country; A Memoir Chiefly on Its Ancient Geography and History: London, Kegan Paul, Trench & Trübner, 148 p. Hodīvālā, S.H., 1939, Studies in Indo-Muslim History: A Critical Commentary on Elliot and Dowson’s History of India as Told by Its Own Historians, Volume 1: Bombay, Kokil and Co., 712 p. Jarrett, H.S., 1891, The A’in-i Akbari by Abu’l-Fazl Allami, Volume 2; Translated from the original Persian by Col. H.S. Jarrett, Asiatic Society of Bengal: Calcutta, Baptist Mission Press, 432 p. Johnston, A.C., and Schweig, E.S., 1996, The enigma of the New Madrid earthquakes of 1811–1812: Annual Review of Earth and Planetary Sciences, v. 24, p. 339–384, doi: 10.1146/annurev.earth.24.1.339.

37

Jorgensen, D.W., Harvey, M.D., Schumm, S.A., and Flam, L., 1993, Morphology and dynamics of the Indus River: Implications for the Mohenjo Daro site, in Shroder, J.F., ed., Himalayas to the Sea: Geology, Geomorphology and the Quaternary: London and New York, Routledge, 429 p. Kazmi, A.H., 1979, Active faults of Pakistan, in Farah, A., and DeJong, K.A., eds., Geodynamics of Pakistan: Quetta, Geological Survey of Pakistan, p. 285–294. Kazmi, A.H., and Rana, R.A., 1982, Tectonic Map of Pakistan: Quetta, Geological Survey of Pakistan, scale 1:2,000,000, 1 sheet. Khan, A.N., 1990, Al Mansurah—A Forgotten Arab Metropolis in Pakistan: Department of Archaeology and Museums, Government of Pakistan, Museum and Monuments Series 2, 112 p. Nakata, T., Tsutsumi, H., Khan, S.H., and Lawrence, R.D., 1991, Active faults of Pakistan: Hiroshima, Japan, Research Center for Regional Geography, Hiroshima University Special Publication 21, 141 p. Oldham, R.D., 1899, Report on the Great Earthquake of 12th June 1897: Memoirs of the Geological Survey India, v. 29, 379 p. Panhwar, M.H., 1983a, Chronological Dictionary of Sind: Jamshoro, Pakistan, Institute of Sindhology, University of Sind, 203 p. Panhwar, M.H., 1983b, Locations of Bhramanka, Patala, Demetrias, Minnaggara, Brahmano, Brahmanva, Brahmanabad, Daluri and Mansurah: Sindh Quarterly, v. 11, no. 2, p. 15–22. Quittmeyer, R.C., Farah, A., and Jacob, K.H., 1979, The seismicity of Pakistan and its relation to surface faults, in Farah, A., and DeJong, K.A., eds., Geodynamics of Pakistan: Quetta, Geological Survey of Pakistan, p. 271–284. Raverty, H.G., 1893, The Mihran of Sind, and its tributaries: Journal of the Asiatic Society of Bengal, v. 61, p. 155–297. Sarkar, J., 1947, Maasir-i-Alamgiri—An English Translation of Saqi Must’ad Khan’s History of the Emperor Aurangzib: Lahore, Suhail Academy, Reprint 1981, 350 p. Sarkar, J., 1978, The A’in-i Akbari by Abu’l-Fazl Allami; Translated from the Original Persian by Col. H.S. Jarrett, Corrected and Further Annotated by Jadunnath Sarkar (2nd edition reprint): New Delhi, Oriental Books Reprint Corp., 236 p. Sawar, G., 2004, Earthquakes and the neo-tectonic framework of the KutchHyderabad-Karachi triple junction area, Indo-Pakistan: Pakistan Journal of Hydrocarbon Research, v. 14, p. 35–40. Schumm, S.A., 2005, River Variability and Complexity: Cambridge, UK, Cambridge University Press, 220 p. Schumm, S.A., Dumont, J.F., and Holbrook, J.M., 2000, Active Tectonics and Alluvial Rivers: Cambridge, UK, Cambridge University Press, 276 p. Stein, S., Sella, G.F., and Okal, E.A., 2002, The January 26, 2001, Bhuj earthquake and the diffuse western boundary of the Indian plate, in Stein, S., and Freymueller, J., eds., Plate Boundary Zones: Washington, D.C., American Geophysical Union, Geodynamics Series, v. 30, p. 243–264. Sykes, W.H., 1857a, Relics from the buried city of Brahmunabad: Illustrated London News, 846, 21 February 1857, p. 166–167. Sykes, W.H., 1857b, The ancient and ruined city of Brahmunabad in Sind: Illustrated London News, 847, 28 February 1857, p. 187–189. Thomas, E., 1858, Essays on Indian Antiquities, History, Numismatics and Palæographic, of the Late James Prinsep, to Which Are Added Useful Tables Illustrative of Indian History, Chronology, Modern Coinages, Weights Measures etc., Volume 2: London, John Murray, 120 p. Villasenor, A., and Engdahl, E.R., 2007, Systematic relocation of early instrumental seismicity: Earthquakes in the international seismological summary for 1960–1963: Bulletin of the Seismological Society of America, v. 97, p. 1820–1832, doi: 10.1785/0120060118. Wheeler, R.E.M., 1992, Five Thousand Years of Pakistan: An Archaeological Outline: Karachi, Pakistan, Royal Book Company, 149 p. Zaigham, N.A., and Mallick, K.A., 1999, Prospect of hydrocarbon associated with fossil-rift structures of the southern Indus Basin, Pakistan: American Association of Petroleum Geologists Bulletin, v. 84, no. 11, p. 1833–1848.

MANUSCRIPT ACCEPTED BY THE SOCIETY 19 MAY 2010

Printed in the USA

The Geological Society of America Special Paper 471 2010

San Antonio de Mucuñó, Mérida Andes, Venezuela: Relocation of a doctrine town following the 1674 earthquake Jaime Laffaille* Facultad de Ciencias, Departamento de Física, Universidad de los Andes, FUNDAPRIS, La Hechicera, Mérida, Edo Mérida, Venezuela, 5101 Franck Audemard M.* Fundación Venezolana de Investigaciones Sismológicas, El Llanito Caracas, Venezuela, El Marqués 1070 Miguel Alvarado* Facultad de Ingeniería, Escuela de Ciencias Geológicas, Universidad de los Andes, La Hechicera, Mérida, Edo Mérida, Venezuela, 5101

ABSTRACT The prime cause of the relocation of one of the first villages founded in Venezuela by Spaniards in the early seventeenth century was likely motivated by earthquakes. San Antonio de Mucuñó, located in the Merida Andes ~200 km south-southeast of Maracaibo, was subjected to the effects of landslides triggered by a series of seismic events that took place in and around the year 1674. Historical documents, the geological and seismo-tectonic setting, and paleoseismic data support the conclusion that the earthquakes of 1674 occurred on the nearby, seismically active Bocono fault.

operating system whereby the land was acquired by the conquerors, and, in exchange, the natives would till the land as a reward offered by the Crown for the investments and risks of conquest. Furthermore, the Indians would receive cultural and religious training. According to some historical documents, the town of Mucuñó was founded in the year 1620 in the Acequias Valley (“Valle de las Acequias”) on the Chaquentá flat-top mountain (“Mesa de Chaquentá,” Fig. 1) by order of the General Visitor of Mérida, Pamplona and Tunja Provinces, Alonso Vásquez de Cisneros. The general did not only specify the encomiendas that would meet to form this town, but also defined its spatial array (Clarac, 1990), and thus established its social structure. Among the factors that influenced the foundation of Mucuñó, the convenience of gathering as many encomiendas into a few

INTRODUCTION In the Mérida Andes (Venezuela), ~30 km to the south of the city of Merida, on a hill on the east margin of the Mucusás River, there lie the ruins of a small settlement named San Antonio de Mucuñó (Fig. 1). These ruins, almost all made of footcompacted earth walls, are placed in such a way that they appear to correspond to two different villages, only ~100 m apart. Two large churches and associated plaza or squares, as well as other constructional features such as the urban grid and housing, are still easily recognized at the site (Fig. 2). This settlement was one of the named “doctrine towns,” which were founded by Spaniards in colonial times in order to gather the natives into several encomiendas for conversion to Christianity. An encomienda is an

*E-mails: Laffaille—[email protected]; Audemard—[email protected]; Alvarado—[email protected]. Laffaille, J., Audemard M., F., and Alvarado, M., 2010, San Antonio de Mucuñó, Mérida Andes, Venezuela: Relocation of a doctrine town following the 1674 earthquake, in Sintubin, M., Stewart, I.S., Niemi, T.M., and Altunel, E., eds., Ancient Earthquakes: Geological Society of America Special Paper 471, p. 39–46, doi: 10.1130/2010.2471(04). For permission to copy, contact [email protected]. © 2010 The Geological Society of America. All rights reserved.

39

40

Laffaille et al.

Figure 1. Relative location of the Boconó fault (Boc-b in Fig. 7) and the pull-apart basins of Lagunillas and Apartaderos town (indicated by white dashes), with respect to the village of Mucuñó. White arrows indicate the wave propagation direction, assuming that the epicentral region of one of the main events of the 1673–1674 A.D. sequence could be hypothetically located near the middle of segment Boc-b. Boxes in upper left and lower right show the location of the study area (white star) and some reference sites, respectively.

Figure 2. Ground view of the site where the ruins of the two towns of San Antonio de Mucuñó are located. To the left, the location of the church corresponding to the first foundation is indicated. To the right of the picture, the place of the church of the second foundation is indicated (the linear distance between the two churches is ~240 m). In Figure 4, we have indicated the direction along which this view was taken.

towns along the basin of Nuestra Señora River (22 encomiendas) as possible stands out. This plan facilitated the conversion of natives and allowed the Spanish Crown to exert the politicaleconomic control on the population, as well as to control the doctrinaires and the heads of each encomienda (“encomenderos”; Clarac, 1990). Acosta (1982) pointed out that another factor was the excellent natural conditions for agricultural production, since this land was good for cultivation of wheat (which led Merida to become one of the main production areas of this cereal), corn, celery and, in general, those vegetables that natives usually consumed. However, in a series of documents and letters dating from

1672 archived at the Merida State Historical Files (“Archivo Histórico del Estado Mérida”; Nadal and Villafañe, 2004), the residents of Mucuñó claimed that new land had to be assigned for the town relocation. This request was supported by two major arguments. These included the ongoing water shortage for land irrigation and the threat to the residents’ safety due to frequent “bolcanes.” Bolcán is the word the natives of the Mérida Andes use for a mudflow and/or debris flow that are well known for their great destructive power. It is equivalent to the Peruvian term “Huayco.” The bolcanes problem was severe due to the intense rains of the year 1672 that were so strong as to activate several

San Antonio de Mucuñó, Mérida Andes, Venezuela: Relocation following the 1674 earthquake of these “bolcanes,” which could leave the villagers isolated (Clarac, 1990). Several authors have suggested (e.g., Acosta, 1982; Centeno, 1940; Clarac, 1990) that this town was abandoned and later relocated due to earthquakes, attributing this to diverse events of the seventeenth century (including earthquakes that happened at far distant places, such as the Plamplona 1644 earthquake, Colombia, or the 1684 event of Cumaná, eastern Venezuela). However, none of the available historical documents up to 1672 mentions earthquakes as a plausible threat to the stability of the town. There are no reliable arguments, therefore, to consider that earthquakes prior to this date were directly related to the town relocation. As for the case of the 1684 earthquake, which occurred at a distance greater than a 1000 km, there is no indication so far that this event had effects on the Venezuela Andes or was even felt in the area. In this paper, we present evidence that the relocation of San Antonio de Mucuñó was motivated not only by the shaking of the 1674 seismic events, but also to the reactivation of slope movements of the sediments on which the original settlement was founded.

41

two decades. Documents up to 21 May 1692 testify in favor of demands for town relocation (Nadal and Villafañe, 2004). These documents are signed by tribe chiefs (“caciques”), notaries, foremen, captains, and the Christian friars of the town, among others. In the 1692 documents, the urgent need of moving the town was proposed due to frequent earthquakes, rigorous winters, and the appearance and reactivation of large ground cracks and fissures affecting a significant part of the town. Figure 3 depicts the church of the first village. Its dimensions are nearly 50 m in length and close to 18 m in width (Fig. 3A). Large buttresses seem to have been added to reinforce the walls (Figs. 3A and 3C), possibly after the mentioned earthquakes. Figure 3B illustrates the height of the walls (which reach up to 5 m) and thickness (which is more than 1 m in some places). In the Depository of Notary Protocols of the General Archives of the State of Mérida (Fondo de Protocolos Notariales, Archivos Generales del Estado Mérida, 1692, in Nadal and Villafañe, 2004, p. 3), there is a document in which the doctrine friar of San Antonio de Mucuñó, Don Francisco de Eusa, reports to the Lieutenant of Governor of the Merida Province with the following:

ANALYSES OF HISTORICAL DOCUMENTS As early as 1672, the relocation of Mucuñó was first requested but approved neither that year nor in the following

I certify that the mentioned town is very dangerous and with very clear risk, threatening to fall completely apart. Since the past tremors everything has split into pieces by deep cracks; one of which is crossing the

A

B

C

Figure 3. (A) General view of the church of the first village, with a length of nearly 50 m and a width close to 18 m. Notice the large buttresses to strengthen walls (A and C). (B) Photo depicting the height of the walls (near 5 m high) and thickness (over 1 m in some places). (C) Close-up of the buttresses.

42

Laffaille et al.

church. This crack has “deepened” (activated and became deeper) with the tremors of last year. Another crack crosses through the backyard of my house, as well as across the church, thus leaving the floor of this church and its walls heavily damaged. In addition to this, the typical heavy rains of every year rainy season and flush floods, combined to the steep slopes on which the town is settled, deepens, or forms new, gullies, resulting in more frequent “bolcanes” [debris flows]. In turn, this impedes any access to or communication with the town, inducing the progressive abandonment of the village by the Indians. In the same way, the boys and young girls … fear to perish buried by a debris flow. (translation by the authors)

Some geomorphologic evidence corroborating this description was identified during fieldwork (Fig. 4). Another series of documents in the same archive (Fondo de Protocolos Notariales, Archivos Generales del Estado Mérida, 1692, in Nadal and Villafañe, 2004) helps us to understand why the two towns were founded one next to the other (Fig. 2) against the wishes of the natives. This place adjacent to the old town was not pleasant for the Mucuñoes—they complained about it to their local authorities, tribe chiefs (“caciques”), captains, and captain governor, pointing out that they had proposed a better place in the lands of Captain Alonso de Toro Holguin. This proposal was rejected by Don Francisco de Eusa, who forced them to occupy the place he chose. This sort of behavior by some of the doctrine friars and Spanish authorities was considered necessary because natives used different pretexts to escape from the encomiendas and go back to their original settlements or moorlands (Clarac, 1990). This was detrimental to the economy (i.e., land cultivation and harvesting) and to the doctrine and dominance imposed by the Spaniards through the doctrine towns and the encomiendas system.

A

THE 1674 EARTHQUAKES AND RELOCATION OF SAN ANTONIO DE MUCUÑÓ The documents and testimonies described in the previous section led us to hypothesize that a succession of strong tremors occurred between 1672 and 1692, and the associated effects of these earthquakes on the land where the town was founded triggered the need to relocate Mucuñó. Careful revision and interpretation of existing catalogues of historical seismicity (e.g., Centeno, 1940; Grases et al., 1999) establishes that Don Francisco de Eusa refers to the earthquakes of 1674 when stating in 1692: “Since the past tremors everything has split into pieces by deep cracks; one of which is crossing the church. This crack has ‘deepened’ (activated and became deeper) with the tremors of last year ....” It is then clear that he blames some events taking place in 1691 (the year before the letter) but also other prior earthquakes felt in the region (i.e., the crack existed prior to the 1691 events). The previous events can be interpreted as those of 1674, since the aforementioned catalogues do not report other events in the time period under consideration. In agreement with Palme and Altez (2002), the 1674 earthquakes affected the entire western region of Venezuela. A sequence of several seismic events beginning on 8 December 1673, and lasting until 23 January 1674, triggered large and numerous mass movements in many catchment areas, damming several valleys of the northern flank of the Mérida Andes that later breached catastrophically, unleashing mud and debris flows. According to the same authors, between 16 and 18 January 1674, the inhabitants felt more than 30 tremors. The strongest of them was the 3:30 p.m. earthquake on the 16 January 1674. San Antonio de Mucuñó lies inside the VII–VIII isoseismal zone proposed by Palme and Altez (2002) for this major

B

C D

Figure 4. (A) Two of the major scarps of the landslide (white dashed lines) and some of the ruins of the affected houses. Note that the three dotted lines parallel to the sliding mass surfaces present an angular variation with respect to the horizontal reference (the distance between the white dashed lines is ~20 m). (B) Panoramic view of the ruins from the northwest, where only one of the major slide scarps has been highlighted (the scarp is ~500 m long). In Figure 5, we have indicated the direction along which these two views were taken. (C–D) Field evidence of hydrometeorological events named “bolcanes” (debris flows) by Mucuñoes: (C) ruins of Mucuño next to a gully along which these flows transited (these ruins are ~1.8 m tall), and (D) debris flows cutting off access to the town of Mucuñó (location shown in Fig. 5; the width of this channel is, on average, about 4 m), as mentioned in historical accounts.

San Antonio de Mucuñó, Mérida Andes, Venezuela: Relocation following the 1674 earthquake event. This sequence of earthquakes very likely continued up to 1675, based on interpretation of historical documents (Palme and Altez, 2002). Mucuñoes refer to the events that damaged their town as follows: “they were very strong and continuous, thus damaging the ground and opening large cracks.” This description is comparable to what happened in 1674 in other regions of the Venezuelan Andes and supports the hypothesis that seismically induced shaking generated or reactivated a local mass movement that affected San Antonio de Mucuñó. The interpretation of aerial photographs of the area (Mission 010480 from 1980), in combination with field work, confirms the presence of a landslide at the Mucuñó site (Figs. 4A, 4B, and 5). Figure 4A shows two of the major scarps of the landslide and some of the ruins of the affected houses. From the ground level and visible in the background of Figure 4A, we can safely deduce that this slide was rotational in nature. A panoramic view from the northwest (Fig. 4B) clearly shows the arcuate shape of the main head scarp of the slide that lies tangent to the town ruins. From our interpretation of the historical accounts, we propose that this slide was very likely retrogradational, implying that the slide head scarp moved uphill, thus enlarging the slide stepwise through time. In this way, the first town of Mucuñó was progressively affected by large open cracks associated with incorporation of new rotated slope mass to the slide. The step-by-step process was probably due to recurrent natural processes such as heavy rains and/or strong local ground motion from earthquakes (at least in the 1674 and 1691 events). Geomorphologic photo interpretation (Fig. 5) depicts the general shape of the slide in relation to the two different Mucunó villages. Geologically, the flat-top hill or mesa of Chaquentá (Fig. 1) where both Mucuñó villages were founded is composed of a sequence of metamorphic rocks

43

including slates, phyllites, schists, fine-grained gneisses, massive silicate rocks, and amphibolites. Foliation of these rocks dips south to southwest, thus favoring the formation of dip-slip landslides (Fig. 5). MUCUÑÓ LANDSLIDE AND ITS TRIGGERING Since the pioneering work of Rod (1956), the Boconó fault is acknowledged as the most active tectonic feature in the Venezuelan Andes region with a significant seismogenic capacity. The Boconó fault trends NE-SW over a distance of ~600 km from the Táchira depression to the town of Morón on the Caribbean coast (Fig. 1). The fault extends along the Venezuelan Andes slightly oblique to its longitudinal axis (Figs. 1 and 6) and displays rightlateral kinematic motion with an average slip rate of 9–10 mm/yr in the Apartaderos region (Audemard and Audemard, 2002). This major tectonic structure has been divided into five segments (Boc-a to Boc-e; Fig. 6) based essentially on geometric criteria (Audemard et al., 2000). Audemard (2009, personal commun.), utilizing data from 11 paleoseismic trenches, confirmed that the geometric segments were also likely seismogenic segments. The most important historical earthquake for the city of Mérida and surroundings, where the ruins of the Mucuñó village are located, is the earthquake sequence of 1673–1674. The 1674 earthquake, or sequence of seismic events, is clearly linked to rupture of the Boc-b segment of the Boconó fault (Fig. 6), which strikes toward the northeast from the Lagunillas pull-apart basin to the Apartaderos region (Fig. 1). However, it is evident that the VIII isoseismal zone of the 1674 earthquake extends farther to the northeast and partially covers the Boc-c segment of the same fault. These data suggest that the coseismic rupture(s) of these

Landslide scarp Ruins Alluvial fans Channel water Mérida’s old route Water furrow

0

Figure 5. Air photograph of area of interest (mission 010480, scale 1:35,000; courtesy of Instituto Geográfico Venezolano Simón Bolívar [IGVSB]). The geological photo-interpretation shows a series of crowns corresponding to a mass movement affecting the constructions of the first village of Mucuñó (church 1), motivating the town relocation toward church 2. The legend is provided on the left, as well as a photograph of fresh in situ foliated rocks observed at point A, shown on aerial photo.

44

Laffaille et al.

Figure 6. Relative location of all the paleoseismic explorations performed along the different segments of the Boconó fault (Boc-a to Boc-e), as well as the most important historical earthquakes in the central and southern regions of the Mérida Andes, each one represented by their macroseismic epicenter (stars) and VIII isoseismal value (after Audemard, 2008). The region covered in this figure is approximately the same as for the lower right box in Figure 1.

events parallel the longest SW-NE axis of the isoseismal line that extends ~130 km (Audemard, 2009, personal commun.). This fault rupture length indicates a magnitude of M 7.4–7.5. A similar magnitude estimation was presented by Palme et al. (2005), who used the Bakun and Wentworth (1997) method (M is equivalent to seismic moment magnitude Mw because Palme et al. [2005] used this magnitude type for calibration). The earthquake cluster that occurred between 1673 and 1674 likely had more than one main shock (Palme and Altez, 2002). Based on the identification of a historical earthquake after 1660 in the Morro de Los Hoyos excavation on segment Boc-b (Audemard et al., 1999; T3 in Fig. 6) and in the Mesa del Caballo trench on section Boc-c (Audemard et al., 2008; T5 in Fig. 6), Audemard et al. (2008) proposed that this series of seismic events ruptured both splays of the Boconó fault bounding the Apartaderos basin. Rupture of the Boc-b segment of the Boconó fault in the 1674 earthquake was has also been confirmed in a paleoseismic trench at the Quinanoque site (Alvarado, 2008). Audemard et al. (2008) also suggested that one segment could have triggered rupture on the contiguous fault segment in a matter of days, weeks, or months. This potential rupture sequence implies that the geometric complexity is not a permanent barrier to rupture propagation but should be considered as a leaky barrier (Audemard, 2009, personal commun.). This double rupture is also supported by the very widespread landslide distribution on the northern flank of the Mérida Andes and the reported building damages between the towns of Mérida and Trujillo, which are separated by 120 km (Fig. 1). This has probably led to the larger magnitude estimate of M 7.4–7.5 (Palme et al., 2005). In this alternative interpretation, that there was more than one large

earthquake, the magnitude of these events would be closer to the M 6.8 value determined by Palme and Altez (2002) from analyses of reported damage in historical records. In the region of the town of Lagunillas (state of Mérida) and to the north of the ruins of San Antonio de Mucuñó, the Boconó fault has two traces (Boc-a and Boc-b; Fig. 7) that overlap for >10 km in a transtensional relay. This geometry is responsible for the pull-apart basin of Lagunillas that contains the Urao lake (Alvarado, 2008). Fault rupture data were obtained from two paleoseismic sites, the Quinanoque trench on the Boc-b fault segment west of the Urao lake (Figs. 7 and 8) and Pantaleta trench on Boc-a fault segment (Alvarado, 2008; Audemard, 2009, personal commun.). These studies found that the Boc-b and not the Boc-a fault segment ruptured in the 1674 earthquake. The ruins of San Antonio de Mucuñó lie less than 20 km southeast of the Lagunillas pull-apart basin and the active trace of the Boconó fault (section Boc-b; Fig. 1) that ruptured in the 1674 earthquake. Consequently, it is likely that the 1674 earthquake sequence triggered widespread landsliding both on the distant northern flank of the Mérida Andes (Palme and Altez, 2002) and the southwest-facing slope of the Chaquentá flat-top hill where the first village of Mucuñó is situated. DISCUSSION The 1674 seismic events represent a significant earthquake sequence in the seismic history of Venezuela because of both their physical characteristics (multiple premonitory events, hundreds of aftershocks, and more than one main shock) and their macroseismic effects, many of which were associated mainly with

San Antonio de Mucuñó, Mérida Andes, Venezuela: Relocation following the 1674 earthquake

45

Figure 7. Relative location of the Pantaleta and Quinanoque trenches across the northern and southern Boconó fault traces, respectively, in the region of Lagunillas pull-apart basin (Fig. 1; after Alvarado, 2008).

Geomorphology: FT—fault trench; PR—pressure ridge; SR—shutter ridge; SP—sag pond; LV—lineal valley; OD—offset drainage; TH—saddle; BE—bench; Stratigraphy: Q1—Pleistocene; Q2—Middle Pleistocene; Q4—Holocene; 500 m Graphic scale 0 N2-Q—Pilocene-Pleistocene.

—fault scarp 1 Km

Figure 8. Log of the west wall of the Quinanoque trench, where all 14C ages are reported. The youngest recorded deformation is associated with the 1674 earthquake (among organic samples 31 and 32). Details on the chronologic difficulties shown by dated horizons are provided in Alvarado (2008) and Audemard (2008).

coseismic geomorphic events (large and widespread mass wasting in this case). These two aspects should be related in some way. For example, a series of foreshocks of significant magnitude might make the ground prone to large slope failure during the main earthquakes. These factors thus notably complicate the analyses of seismically induced features and damage in the assessment of magnitude of historical earthquakes (e.g., Palme and Altez, 2002). Paleoseismic investigations have provided essential information for the determination of the main earthquake sources along the Boconó fault. In addition, these studies have shed light on the possible rupture mechanism by which two different contiguous seismogenic segments of the Boconó fault (Boc-b and Boc-c) broke together but not necessarily simultaneously (Alvarado, 2008; Audemard et al., 2008). From both geomorphologic interpretation of aerial photographs and fieldwork at the Mucuño town site, it is clear that the first village was deeply affected by a rotational slide that crosscuts the urban grid. Furthermore, the slide has likely been reactivated and grown stepwise as deduced from the historical accounts. The mass movement first destroyed the borders of the village and then the inner parts of it. The landslide triggering mechanism was probably not a continuous or permanent source, but rather an intermittent trigger such as heavy rains or earthquake ground motion. It is plausible that the construction

of the town and human alteration of the general landscape and neighboring slopes helped to destabilize the land and produce conditions favoring mass movements (Laffaille et al., 2002). This issue might be of major significance when evaluating the potential triggering agents. Taking into account that the earthquakes of 1673–1674 have been correlated to rupture of Boc-b and Boc-c segments of the Boconó fault based on four paleoseismic trenches (two in the Lagunillas pull-apart basin and the two in Apartaderos pull-apart basin; Audemard et al., 1999, 2008; Alvarado, 2008; Audemard, 2008; Audemard, 2009, personal commun.), we postulate that the landslide that caused the abandonment of the first location of the village of Mucuñó was due to seismic shaking in an earthquake that occurred sometime between 1620 (when the village was first founded) and 1692 (when the relocation actually occurred). More precisely, historical documents indicate that damage to the village was triggered in an event prior to the 1691 shock that led to the final request for relocation. During this time span (1620–1591), the most likely earthquake in the Venezuelan catalogue of historical seismicity that would be close to the Mucuñó site and strong enough to trigger mass wasting is the earthquakes sequence of 1673–1674. The mass wasting of 1673–1674 was likely favored by rainy conditions along the northern flank of the Mérida Andes. In

46

Laffaille et al.

contrast, the Lagunillas region is characterized by semiarid to arid conditions year-round due to orographic effects that alter the wind patterns and cloud formation. The slide-dammed lakes that formed along the northern flank were suddenly drained, implying enough running water to fill them and later breach the earth dams. These catastrophic failures produced very large debris flows that were responsible for substantial losses in agricultural products within a short time after the earthquakes. Nevertheless, the heavy rains in Chaquentá area alone cannot be completely ruled out as a triggering mechanism for the slide. However, historical accounts reported the formation of debris flows (“bolcanes”) around the town at the time of the opening of the wide cracks, such as the one shown in Figures 4C and 4D. The distance between the first town of Mucuñó and the Boc-b rupture segment of the Boconó fault during the 1673–1674 earthquake sequence is ~20 km. It is much closer than the distance to the larger and wetter mass movements in the northern flank. The smaller size of the landslide at the Mucuñó site compared to those on the northern flank of the Mérida Andes might be better attributed to the dry conditions prevailing in the Las Acequias valley where the ruins of Mucuñó are located rather than to its proximity to the epicenter. Assuming that the epicentral region of one of the main events of the 1673–1674 earthquake sequence was located near the middle of the Boc-b segment, the wave propagation direction (white arrows in Fig. 1) might have impacted almost perpendicular to the SW-facing hillside of the Chaquentá mesa and Mucuñó site. The foliation dip direction and the steepness of the slope, among other factors, probably facilitated downslope movement. Additional geologic, photo interpretation, and detailed fieldwork on the Boc-b fault segment, as well as an exhaustive search of new and relevant historical data, would help to further resolve the triggering and timing of the mass-wasting events in relationship to earthquake faulting. This study combined detailed geomorphologic studies of mass-wasting processes and paleoseismic interpretation of the timing of faulting together with the evaluation and interpretation of historical accounts, not only in terms of past seismic activity and the associated phenomena, but also in terms of social sciences (history, economy, anthropology, sociology, theology, urbanism, among others). The use of this type multidisciplinary approach has allowed us to better interpret the historical data in relationship to earthquake rupture models and ultimately to provide better earthquake source parameters for seismic hazard analyses. ACKNOWLEDGMENTS Our gratitude goes to the Venezuelan Foundation for Seismological Research (FUNVISIS) and to the Foundation for Prevention of the Seismic Risk of the State Merida (FUNDAPRIS) for their unconditional support to this investigation. This research is a contribution to project FONACIT 2001002492 and 2002000478. This article is a contribution to the United Nations Educational, Scientific and Cultural Organization (UNESCO)–funded Inter-

national Geoscience Programme (IGCP) 567, “Earthquake Archaeology: Archaeoseismology along the Alpine-Himalayan Seismic Zone.” REFERENCES CITED Acosta, L., 1982, San Antonio de Mucuño. Formación de un Pueblo Indígena de Encomienda y de Doctrina en el Valle de Acequias, 1558–1620 [tesis de grado]: Mérida, Venezuela, Escuela de Historia, Universidad de Los Andes, 160 p. Alvarado, M., 2008, Caracterización Neotectónica de la Cuenca La González, Estado Mérida, Venezuela [M.Sc. thesis]: Mérida, Universidad Central de Venezuela, 220 p. Audemard, F.A., 2005, Paleoseismology in Venezuela: Objectives, methods, applications, limitations and perspectives: Tectonophysics, v. 408, p. 29–61. Audemard M., F.A., Pantosti, D., Machette, M., Costa, C., Okumura, K., Cowan, H., Diederix, H., and Ferrer, C., 1999, Trench investigation along the Merida section of the Boconó fault (central Venezuelan Andes), Venezuela, in Pavlides, S., Pantosti, D., and Peizhen, Z., eds., Earthquakes, Paleoseismology and Active Tectonics: Selected Papers to the 29th General Assembly of the Association of Seismology and Physics of the Earth’s Interior (IASPEI), Thessaloniki, Greece, August 1997: Tectonophysics, v. 308, p. 1–21, doi: 10.1016/S0040-1951(99)00085-2. Audemard, F.A., Machette, M., Cox, J., Dart, R., and Haller, K., 2000, Map and Database of Quaternary Faults in Venezuela and Its Offshore Regions: U.S. Geological Survey Open-File Report 00-0018, including map at scale 1:2,000,000 and 78 p. text. Audemard M., F.A., Ollarves, R., Betchtold, M., Díaz, G., Beck, C., Carrillo, E., Pantosti, D., and Diederix, H., 2008, Trench investigation on the main strand of the Boconó fault in its central section, at Mesa del Caballo, Mérida Andes, Venezuela: Tectonophysics, v. 459, p. 38–53, doi: 10.1016/j.tecto.2007.08.020. Audemard, F.E., and Audemard, F.A., 2002, Structure of the Mérida Andes, Venezuela: Relations with the South America–Caribbean geodynamic interaction: Tectonophysics, v. 345, no. 1–4, p. 299–327, doi: 10.1016/ S0040-1951(01)00218-9. Bakun, W., and Wentworth, C., 1997, Estimating earthquake location and magnitude from seismic intensity data: Bulletin of the Seismological Society of America, v. 87, p. 1502–1521. Centeno, M., 1940, Estudios Sismológicos: Caracas, Venezuela, Litografía El Comercio, 365 p. Clarac, J., 1990, Etnohistoria de San Antonio de Mucuñó: Mérida, Venezuela, Universidad de Los Andes, Boletín Antropológico 20, 120 p. Grases, J., Altez, R., and Lugo, M., 1999, Catálogo de Sismos Sentidos o Destructores. Venezuela 1530–1998: Caracas, Academia de Ciencias Físicas, Matemáticas y Naturales, Facultad de Ingeniería, Universidad Central de Venezuela, Editorial Innovación Tecnológica, 654 p. Laffaille, J., Ferrer, C., and Rengifo, M., 2002, San Antonio de Mucuñó: Evidencias históricas de la actividad antrópica como detonante de amenazas naturales: Caracas, Venezuela, Memorias del las III Jornadas de Sismología Histórica, FUNVISIS, Serie Técnica 1-2002, p. 219–221. Nadal, A., and Villafañe, M., 2004, Mudanza del pueblo de San Antonio de Mucuño para otro sitio más apropiado en tierras de la Encomienda del Capitán Alonso de Toro Holguín, en el Valle de Acequias, 1692. Procesos Históricos: Revista de Historia, Arte y Ciencias Sociales, no. 5: Mérida, Venezuela, Universidad de Los Andes, 125 p. Palme, C., and Altez, R., 2002, Los terremotos de 1673 y 1674 en los Andes Venezolanos: Interciencia, v. 27, no. 5, p. 220–226. Palme, C., Morandi, M., and Choy, J., 2005, Re-evaluación de las intensidades de los grandes sismos históricos de la región de la cordillera de Mérida utilizando el método de Bakun and Wentworth: Revista Geográfica Venezolana, Número Especial, v. 2005, p. 233–253. Rod, E., 1956, Earthquakes of Venezuela related to strike slip faults?: American Association of Petroleum Geologists Bulletin, v. 40, p. 2509–2512.

MANUSCRIPT ACCEPTED BY THE SOCIETY 19 MAY 2010

Printed in the USA

The Geological Society of America Special Paper 471 2010

New interpretations of the social and material impacts of the 1812 earthquake in Caracas, Venezuela Rogelio Altez* School of Anthropology, Universidad Central de Venezuela, Caracas 1040, Venezuela, and Venezuelan Society of Geosciences History, Caracas 1070, Venezuela

ABSTRACT This work sheds light on one of the most important earthquakes in Venezuelan history. At 16:07 on Holy Thursday, 26 March 1812, Caracas and the surrounding province of Venezuela suffered a very destructive earthquake. The earthquake occurred at a time of great political, economic, and social upheaval, with the beginning of the republican revolution and the Spanish royalist military response. Within this historical context of conflict, documentary information may be biased and subjective. This chapter is a methodological and epistemological analysis of the 1812 earthquake damage from letters and manuscripts and an interpretation of the social impact of the earthquake within ideological, subjective, and political context. The widespread destruction of the city of Caracas was heterogeneous in its distribution. Damage was determined largely by the differences in the construction style and quality and by the maintenance status of the building. Based on analyses of three funeral books from the era, the number of earthquake victims in Caracas in 1812 may have been close to 2000. This value is lower than regional estimates of the death toll.

INTRODUCTION

Although it is generally accepted that the Caribbean plate moves eastward with respect to South America, this plate boundary is not a simple dextral type (Soulas, 1986; Beltrán, 1994; Audemard and Audemard, 2002). Instead, it is a broad active deformation zone resulting from a long-lasting oblique-collision process (Audemard, 1993, 1998). Nevertheless, a large portion of this right-lateral motion seems to take place along the dextral Boconó–San Sebastian–El Pilar fault system (Schubert, 1984; Soulas, 1986; Audemard and Singer, 1996). Seismicity of the fault zones aligned along the southern boundary of the Caribbean plate is known from historical earthquakes (Centeno Graü, 1940; Grases, 1990; Grases et al., 1999). Moreover, the seismicity of northern Venezuela (over the Boconó–San Sebastian–El Pilar fault system) suggests that margin deformation does not occur along a single fault zone (Audemard and Singer, 1996). Other

This study is a qualitative analysis of the documentary evidence of the 26 March 1812 earthquake in Venezuela and the associated seismic damage in the city of Caracas. The earthquake ruptured the San Sebastian system fault, one of the many structures along the transform plate boundary that separates South America from the Caribbean plate (Fig. 1). According to Grases and Rodriguez (2001), the magnitude of the 1812 earthquake is estimated to have been between M 6.9 and 7.2. This event is a good example of the expected seismic hazard from the interconnection between southern boundary of the Caribbean plate and the other system faults related to the South American plate. *[email protected]

Altez, R., 2010, New interpretations of the social and material impacts of the 1812 earthquake in Caracas, Venezuela, in Sintubin, M., Stewart, I.S., Niemi, T.M., and Altunel, E., eds., Ancient Earthquakes: Geological Society of America Special Paper 471, p. 47–58, doi: 10.1130/2010.2471(05). For permission to copy, contact [email protected]. © 2010 The Geological Society of America. All rights reserved.

47

48

Altez

Figure 1. Diagram of plate boundaries of Venezuela (after Audemard et al., 2000).

small earthquakes can be localized on minor faults that everywhere cross the principal systems (Audemard et al., 2005). The fundamental aim of this investigation is to estimate the earthquake intensity, earthquake fatalities, and potential microzonation of damage (e.g., Altez and Laffaille, 2006). This historiographic approach to assessing damage in the 1812 earthquake (see Altez, 1998, 2005a, 2006) contrasts with previous studies that have focused solely on assigning earthquake intensities, drawing isoseismals, or exclusively on the style of structural damage (as an example of that, see FUNVISIS, 1997). As stated by Mocquet (2005, p. 130): “The assessment of macroseismic intensities from historical reports requires one to combine information on vulnerability and damages. Similar earthquakes occurring at different epochs can produce different effects, depending on the demography and local life conditions at the time of occurrence.” Critical analyses of historical information pertaining to earthquakes, i.e., the field of historical seismology, utilize methodological approaches and knowledge across the disciplines of seismology, history, anthropology, and earthquake-resistance engineering (Guidoboni and Ferrari, 2000). Three fundamental variables should be understood and analyzed in order to sufficiently reconstruct damage from a seismic event. The most important of these variables is the historical context in which the seismic phenomena took place and the information and direct testimonies concerning the impacts and

effects that are recorded. Understanding this historical context leads us to comprehend the information that was elaborated on and discussed at that moment in history. This has been called a “critical joint” (Olson and Grawonski, 2003) or “disaster juncture” (Altez et al., 2005). Emerging from complex circumstances at a time of military conflict, the 1812 earthquake is recorded with different perspectives from the initial experience and later interpretations of the event. A second variable in understanding this historic seismic event lies in the material context of the built environment. Observations of the constructive typology and construction conditions at the time of the earthquake are important parameters for assessing the intensity of damage (Altez, 2005a; Yamazaki et al., 2005). Damage in the earthquake is quantified on the European Macroseismic Intensity Scale (EMS, 1998), ranging from I through XII. In assessing the specific construction damage, the “Classification of damage to masonry buildings” (also from EMS, 1998), which ranges from I to V, is particular useful. Lessons from the social impacts of this disaster also are learned from the extent of material destruction and the subsequent consequences felt in the population. Finally, the number of earthquake fatalities is an important variable that is often open for interpretation. The number of deaths in the earthquake provides a good estimation of the severity of the event. To calculate the number of deaths in the earthquake, I extrapolate data from multiple sources. The results of these analyses of the historical

The social and material impacts of the 1812 earthquake in Caracas, Venezuela text for seismic intensities and the number of fatalities seem to be far from classic studies on the 1812 Caracas earthquake and provide a new interpretation on this catastrophic phenomenon in the history of Venezuela. HISTORICAL CONTEXT OF VENEZUELA IN 1811–1812 The political events in South America in the early nineteenth century would transform American society forever. Venezuela, as well as most of the Latin American continent, was on a tortuous road from 300 yr of colonial rule toward independence. In July 1811, Venezuela declared independence. This caused an acceleration of the breakup of the institutional and administrative bonds with the Spanish crown, which was then in crisis due to the Napoleonic invasions. In March 1812, before the earthquake, a small royal army started out from the city of Coro, 500 km to the west of Caracas, heading toward the main city with the objective of recapturing the rebellious counties (for a general approach to the historical context, see Lynch, 1985, or Guerra, 1992). The earthquake of 1812 occurred during the Holy Week between Palm Sunday and Easter, with all of its Christian celebrations, magnificence, and ritual ceremonies. The 26 March 1812 was Holy or Maundy Thursday. The religious mandates for the day were rigorously fulfilled, even more in those days, when faith still had an indisputable nature. Most of the population of the region was concentrated in the city of Caracas (Fig. 2), the province capital (territory) of the General Captainship (jurisdic-

49

tion) and of the Archdiocese (major ecclesiastical authority) of Venezuela. Caracas was also the administrative, institutional, and political center. Thus, Holy Thursday included the presence in the capital city of the most prominent civil, military, and ecclesiastical authorities. In the middle of this religious ceremony, and the threat from the royal invasion and political tension, a dreadful earthquake struck the city at 4:07 in the afternoon. Because the earthquake occurred at this turning point in history between colonial rule and independence, the extent of the damage in the event has become an interpretive problem complicated by heroic and nationalistic reconstructions of history. The fall of the patriots’ ambitions, the solid discourse of the Christian faith, the royal manipulation, the denial of the independence movement, and the exaggeration of the context narrators are all factors that have colluded in the overestimation of the importance and magnitude of the tragedy. The critical conditions imposed by the war, including destruction, abandonment, and migration, contributed to the hopeless loss of information. Files were destroyed in the collapse of buildings, lootings occurred on both sides, paper was used for fuel in bonfires that protected against the cold night in the outdoors, and any type of paper was used to ram the gunpowder for combat; all this combined to produce the information loss that is sorely missed today (for a detailed description on the matter, see Altez, 2006). The 1812 Venezuelan earthquake is categorized by all historians and investigators as one of the largest in the country’s history. Information written about this earthquake was motivated by many reasons. This has potentially contaminated the context

Figure 2. Caracas and its faults (detail from Audemard et al., 2000).

50

Altez

of the point of view of the chronicler, suggesting bias in opposing or supporting one side of the nationalism movement and the war of independence. MATERIAL CONTEXT: THE CITY OF CARACAS IN 1812 From an urban point of view, in 1812 Caracas was not a major city compared to other capitals of the Spanish colonial world. More impressive economic and architectural development and higher population densities were found in cities such as Mexico, Lima, and Bogotá. However, Caracas had shown signs of institutional growth since the late eighteenth century.

The appointment of the General Captainship in 1777, the Royal Audience in 1786, the Royal Consulate in 1793, and the Archdiocese in 1804, granted to the capital of the Province of Venezuela considerable status. By 1810, Caracas was the center of political power and economic growth. However, these aspects had not yet decisively impacted the city’s architecture. Urban development in Caracas at the time of the 1812 earthquake lagged behind its other American contemporaries (Fig. 3). Basically, in early nineteenth-century Caracas, the public buildings were old and ill-maintained (Table 1). Most documentation from Caracas and the region, written by the priests narrating the churches’ condition, denounced the dramatic situation of their structures. They describe cracks, roof problems, and infestations

Figure 3. Map of Caracas near 1810, by Mendoza Solar (1910).

The social and material impacts of the 1812 earthquake in Caracas, Venezuela of ants and termites. On 25 August 1812, the Archbishop of Caracas, Narciso Coll y Prat, when speaking about conditions of the cathedral prior to the earthquake, stated that the church was “old, incompetent, disproportionate…” (Coll y Prat, 1812). In 1812, the only structures in the city of considerable height were the churches, based on the rationale that no construction could show more height than “the house of God.” The only tall structures in Caracas in 1812 were the towers of the churches (Grases et al., 1999). Thus, the remaining constructions were of lower and smaller size, as seen on the perspective view of Caracas in Figure 4. Most houses were of typical colonial architecture, with a single story and a central patio. Building materials varied between adobe, bahareque, and masonry. The testimonies of travelers and visitors at the beginning of the nineteenth century attest to the fact that Caracas hardly had any two-storied houses (e.g., Ker Porter,

51

1997; Duane, 1968). In this sense, when speaking of “buildings,” one should emphasize both the small size and scale. Constructions in the city can be classified into three groups: churches (high buildings with towers), public administration buildings (generally two-storied and spacious houses), and living quarters or housing (low structures built in three different styles). The role of construction type was a determining factor in whether the building collapsed in the earthquake. As deduced from the historic documents, the materials used for the housings and buildings construction were not able to withstand the seismic shaking. According to Pedro Cunill Graü (1987), the housings of the poorest members of society were made of bahareque (stick interwoven with canes and mud) and were roofed with straw or palms. Constructions in the middle sector of the society (merchants and artisans) had adobe, walls reinforced by thin trunks to hold roofs of straw or tiles. The houses of the richest levels of

TABLE 1. AGE OF SOME OF THE CONSTRUCTION OF MAJOR STRUCTURES IN CARACAS AT THE TIME OF THE 26 MARCH 1812, VENEZUELAN EARTHQUAKE Construction Approximate year of construction Cathedral Church Built as parish church toward the second half of the sixteenth century. Remodeled in 1636 as a cathedral and as the seat of the Diocese of Venezuela. Rebuilt after the 1641 earthquake. Altagracia Church 1656 La Pastora Church 1745 San Pablo Church 1580 La Merced Church 1638 San Mauricio Church 1570. Rebuilt in 1667, after the damages received in the 1641 earthquake. San Jacinto Church Late sixteenth century San Lázaro Hospital Construction initiated in 1759. La Candelaria Church 1708 Santa Rosalía Church Built as church in 1732, before it was a chapel. La Concepción Convent 1619 Carmelitas Nuns Convent 1739 Seminar School 1675. Recently repaired, with construction works that lasted from 1809 through 1811.

Figure 4. Detail of the painting titled Nuestra Señora de Caracas, painted in 1766 (artist unknown) and currently located in the Consejo Municipal of Caracas. Note the height of constructions of the time. Church towers and some few two-storied buildings are the only tall buildings and are concentrated around downtown.

52

Altez

the community were built with adobe and stone walls with brick facades, wooden barred windows, and tiled roofs. The houses of the upper classes generally had very elevated roofs that made them more fragile during the earthquake (Ker Porter, 1997). Civic and religious constructions were mostly built of wooden frames with stone facades brick arches, tiled roofs, and whitewashed walls. According to Alejandro Ibarra (1862), the quality of the period constructions was highly responsible for the earthquake damages: “The limited depth of the bases or lack thereof, materials of diverse density and shapes and even of inappropriate nature, which were indiscriminately used to make the main walls of even the larger buildings: their lack of thickness, the use of poor quality mortar and other mistakes related to the art of building, were causes that undoubtedly contributed to increase the devastation of the 1812 earthquake” (p. 2). The conditions pointed out by Ibarra (1862) are of great value, since he was a pioneer in the investigation of the effects of the 1812 Venezuelan earthquake closer to the date of the earthquake and prior to the massive transformations of the city in the twentieth and twenty-first centuries. EARTHQUAKE DAMAGE The earthquake effects of 26 March 1812 were devastating in Caracas. As all documentation points out, Caracas was in ruins (Delpeche, 1813; Díaz, 1829; Palacio Fajardo, 1817). Descriptions near the date and detached from all political or romantic ideas are very eloquent. In the first days after the disaster, the Cabildo (the city council) meetings reassured the citizens, calling on them to make efforts to better the conditions of the city. The city council ordered cleanup of the debris in the streets, thus allowing access to Caracas for food deliveries from outside of the city (Actas del Cabildo, 1972). These first orders exemplified the critical necessity of provisioning the city. The roads drew a perimeter around the city, leading to the Capuchinos square, where provisions would be received. The types of structural damage in the earthquake were very diverse and largely determined by the type of construction, which in turn was based on social class. According to Cunill Graü (1987), the underprivileged classes formed suburbs in the periphery of the city, while the most notable and powerful citizens resided in blocks near the city center. The same author states that the housing of the poorest suffered less damage in the earthquake because they were built of bahareque (more earthquake-resistant material in his opinion). It can be inferred that there would be fewer victims in these houses due to light cane roofs that would be less fatal than the collapse of a tile roof. However, testimony of the city council and the orders of Spanish government given in August (Monteverde, 1812; Actas de Cabildo, 1972) indicate that debris was abundant throughout the city. The house construction types of the middle and upper classes were vulnerable, with very heavy and high roofs of tiles propped up by weak thin trunks. It is assumed that these houses caused the largest quantity of deaths and destruction. In the burial

certificates that could be found (Book of Burials, La Candelaria Parish, number 7, 1806–1817; Book of Burials, Adults, San Pablo Parish, 1808–1812; Book of Burials, Chacao Parish, number 2, 1797–1821), most of the deceased registered by the priests in charge died in their houses. There was a high death toll of maids, children, and slaves because families were at the celebrations in the cathedral at the time of the earthquake. The damage in the more important buildings seems to have been related to the quality of their construction and materials. Ibarra (1862, p. 2–3) stated:

An example of all this is provided in the temple of Altagracia whose walls especially the South one and surely that of the North, also built as just said, came to earth and with them all the temple; at the same time its beautiful front that is of good construction resisted the shakes of the tremor and the shudder that should cause the fall of its arched roofs and its very solid and heavy construction. In the same way in the convent of San Jacinto whose interior new factory, was built with bad materials, came to the floor, conserving the old external one that had resisted the earthquake of 1641; and the one that was made immediately after, heavy and rough, but of solid construction, being noticed that the walls built with raw adobes made of pure mud and straw, resisted perfectly, when the poor built masonries with mezclote [some kind of varied and mixed materials] were quartered, disjointed and destroyed they didn’t fall, already decomposed the mezclote and loose fitting bodies that entered in the formation.

Another document (Larrain, 1958) estimated that of the 5000 existing houses in Caracas before the earthquake, only 2000 were left standing. Larrain was referring to the housings, and not to the religious buildings or public administration houses. Little mention is given to civic buildings, since the government had been displaced by the revolution. However, it calls attention to the omission of this information. It may be inferred that they suffered less damage. This type of construction (two-story, roomy, stone, adobe houses with lime walls and tile roofs) resisted destruction in the earthquake. Beyond the materials used in construction, it seems that the quality of the building was decisive in the survival of the structure. Houses built with greater care and better resources, such as the spacious public administration and those of the most prominent citizens in Caracas, seem to have suffered less damage, while those that were built with fewer resources and were old or poorly maintained in their structure (bahareque with cane, heavy roofs propped with fragile and weak tied supports) were less lucky. It can be inferred that the collapse of the roofs was responsible for most of the damage and fatalities. IMPACT ON THE SOCIETY The earthquake of 26 March 1812 occurred in an area unprepared to face earthquakes, and thus produced impressive losses. Alex Scott, commissioned from the United States to observe closely the new revolutionary government, calculated the losses

The social and material impacts of the 1812 earthquake in Caracas, Venezuela in Caracas and La Guaira (near the port 20 km to the north of Caracas) at about four million dollars (Scott, 1812). War across the whole territory had bankrupted the administration by 1812. Also, public funds had collapsed. The country’s largest population by percentage, mestizos (natives and slaves), was structurally poor. Losses in the state treasury must then be envisioned with respect to a population with little wealth. The loss of the churches also meant sacrifice in funds and labor, since the majority was not able to assume the responsibility of reconstruction. Economically, the earthquake affected the religious population, who hardly had the means to go to mass. In a structurally poor society, churches were in a poor state of repair and underfunded, since most of the parishioners were mestizo and not white criollos (creoles) and wealthy people. Descriptions of the ruins repeat in all the documentation, giving account of the earthquake damages and sufferings of the reconstruction. The population was devastated by the lack of resources, besides being pressured by ongoing war. In Caracas, the earthquake damage blocked roads and access to provisions. Later, when the disaster reached its maximum expression, the Consulate of Caracas, from which the Intendant Dionisio Franco (1813) had requested a loan of 100,000 pesos, responded as follows: “The same powerful causes that have influenced from the year 1810 until the present in these Provinces, reducing them to the impossibility of covering the expenses that circumstances demand, has caused the extraordinary backwardness that experiences the trade and the agriculture, particularly in this Province, that above all not only suffering losses being consequence of the political events occurred in that epoch, but also those of the great earthquake of last March 26 that had just ruined the harvesters class” (Consulate answer to Intendant, 29 January 1813). The same text later argued that among other reasons, the consulate by itself was not able to ordinarily congregate, because there was not “...in the City enough number of individuals... whom to summon for it by effect of the same things and in the emigration that there was towards the country following the earthquake in where they still remain in terms that Government’s Meetings have sometimes stopped to be made by lack of vowels….” The merchants were also paralyzed due to the blockade suffered by the region. Even so, it was frankly recognized that, although “they are the only ones who can be counted on in similar difficulties,” the same ones were in a “general poverty.” “New contributions and loans cannot be counted on due to the general poverty of these inhabitants, caused by the revolution, earthquakes and also the recovery of the legitimate Government…” (Dionisio Franco, manuscript dated 13 February 1813). To this crisis picture in the public and private funds of the administration, the characteristic poverty of the society should be added. In this way, all suffered the catastrophe, although not equally. Church restorations represented the same tragedy in Caracas as in all the towns of the region. Undoubtedly, this was due to the lack of resources of all kinds. By 1816, efforts were made to plead the situation of Caracas churches to the city council. How-

53

ever, the political passions would continue slowing things. When the city council requested the report about “the resulting destruction caused in the two revolutions that have afflicted them and by the horrendous general earthquake of March twenty six 1812,” the secretary answered: “…almost all Churches are ruined more or less and many totally. (…) that the [Metropolitan] Church for repairs, like the ruins of their building are repairing, to restore some necessary pieces of jewellery and to make the costs of the precise one for the divine cult, is using extraordinary wills, or taking borrowed money, or receiving it to census…” (Guzmán, 26 April 1816). The earthquake’s damage repairs were delayed by the economic and societal problems. The recruitment and maintenance of a labor force were drained by migration. The repair of churches, room houses, and public administration was also delayed many years because of the deep economic crisis. Similarly, the negative consequences in the population distribution in the urban core caused the city housing to be abandoned and the development of new suburbs. This seriously affected location and image of the capital: “…to which is added that thinking that most of the inhabitants not to live more in the City again, with the object of living in the uninhabited places, unroofed their houses, pulled out their doors and windows with their wood, and these pieces to build shacks where to be…” (Méndez, 18 April 1816). With the destruction caused by the 1812 earthquake, certainly the horizons of the cities, towns, and villages were changed forever. That destruction also contributed to the transformation of the colonial society. To lift a nation from that debris was not, in spite of the nationalist historiography, a heroic gesture, but an unavoidable situation accompanied by hopeless circumstances. EARTHQUAKE FATALITIES At the time of the earthquake, the city of Caracas was the most populated in the whole Province of Venezuela. According to Cunill Graü (1987), in 1812, Caracas had 50,000 inhabitants. A more precise figure is given by Díaz (1817) of 31,813 people. If we compare the population figures with the number of houses (5000 noted by Larrain, 1958), then the average number of inhabitants per home was about six people. The earthquake occurred on Holy Thursday in the middle of the afternoon when most people of the city were in the Main Square in front of the Cathedral celebrating the services. Therefore, houses were not fully occupied at the time of the earthquake. It is quite likely that had the houses been fully occupied, the number of deaths would have been higher. The number of deaths for the 1812 Caracas earthquake has been at times confused with the death totals caused by the war, the mandatory migrations, and the famine. To understand the disastrous consequences of these last aspects, Cunill Graü (1987) conducted a detailed analysis in this respect. As noted earlier, that confusion is clearly linked to the critical historical context and to the meaning of that same context subsequently for the country’s

54

Altez

later history. At a time in which such a profound transformation of society was starting to take place with the birth of a new nation, information was managed or exaggerated for political gain (Guidoboni and Ferrari, 2000; Rodríguez and Audemard, 2003). Later, when those years of crisis were transformed by historical discourse into a revolutionary story, nobody dared to question the heroic conditions of the transformation. The tragedy of the earthquake helped manifest and heighten the revolutionary ideal of heroic triumph over adversity. The nation’s genesis mythology needed its heroes to sufficiently overcome her adversaries. This is why the war, the earthquake, and all the circumstances of that moment became the most critical moments in Venezuelan history: They are the heroic genesis of the nation. However, this does not make the history of the period an unbiased record of the truth. The death figures for the 1812 earthquake vary widely, as seen in Table 2. These figures were obtained from documents from contemporary narrators and from nineteenth- and twentiethcentury investigators. They are certainly the well-known death numbers related to the earthquake in Venezuelan historiography. Between the 20,000 of Forrest (1812) and the 1000 of Roscio (1812), the difference is very significant. Of all of the estimations, only Coll y Prat (Archbishop of Caracas in 1812) declares that his figures come from an order to the priests of his diocese. Indeed, that order was extended to all his parishes on 6 April 1812 (Coll y Prat, manuscript of 31-03-1812), but the replies from his whole jurisdiction have not been found in their entirety. Fatality figures by Coll y Prat (1816) include all the parishes and towns to which the request was made, and not exclusively from the city of Caracas. Furthermore, Díaz (1817) stated that the number of inhabitants in the province indicated that 13,000 persons “died with the earthquake.” This value accounts for the entire population of the provinces in the whole jurisdiction of Venezuela. The 1000 deaths in Caracas pointed out by Roscio (1812) are derived from a seemingly very conservative number. In the San Carlos military quarter of the city, ~500 individuals are said to have perished (Delpeche, 1813). Díaz (1829) wrote an eyewitness account of 40 people dying in the collapse of one convent of the Order of Predicadores in the Square of San Jacinto. One way to derive an independent estimation of the earthquake fatalities is to consult the funeral records. However, it was only possible to find three funeral books that correspond to March 1812 (Table 3). Only two of the funeral books are from the urban area of Caracas. The third funeral book is from the Chacao parish, a small town on the outskirts of the city. The funeral book data show a total of 137 deaths from three different parishes. Extrapolation of this fatality rate to the rest of the city is probably not appropriate because it may not represent earthquake damage and building collapse in other parishes of Caracas. In total, Caracas had as many as 19 parishes inside the urban perimeter, and thus calculations of the death numbers with data coming from only two of the parishes are highly uncertain. It is significant that the death numbers registered by the priests of these parishes take into account the total of the jurisdiction, including, obviously, the housing and not only the churches.

TABLE 2. NUMBER OF DEATHS FROM THE 1812 EARTHQUAKE IN CARACAS No. of death s Source 15,000–20,000 Forrest (1812) 16,000 Ker Porter (1825–1842) 14,000 Centeno Graü (1940) 10,000–12,000 Coll y Prat (1960, originally from 1818) 10,000 Rojas (1879) 10,000 Heredia (1895) 9000–10,000 Delpeche (1813) 8000 Irvine (1818) 7000 Parra Pérez (1939) 6000–7000 Urquinaona y Pardo (1820) More than 6000 Ascanio (1813) 3000 Semple (1812) 1000 Roscio (1812)

TABLE 3. NUMBER OF DEATHS FOR THE 1812 VENEZUELAN EARTHQUAKE DERIVED FROM FUNERAL BOOKS OF THE CARACAS REGION No. of deaths Parish Source 83 La Candelaria Book of Burials, 1806–1817 38 San Pablo Book of Burials, 1808–1812 16 Chacao Book of Burials, 1797–1821

Consequently, it is a registration that embraces several (perhaps 8–10) blocks around a church. In a radius of some 10 blocks (the probable maximum for the case of La Candelaria), there were 83 deceased. With this value, it is difficult to reconcile a total Caracas death toll of 20,000 (Forrest, 1812) or 12,000 (Coll y Prat, 1818) or 6000 (Parra Pérez, 1939). Even if the La Candelaria parish death toll is increased to 100 fatalities and applied to each parish and the 500 deaths from the San Carlos Military Headquarter are added, the total number of deaths for the area of 19 parishes is 2400 people. The conservative figure of Roscio (1812) and the approximate value of 3000 by Semple (1812) are close to this calculation. The larger 100 deaths per parish is approximately a 20% increase over the funeral records from the La Candelaria parish. The 38 deaths from the San Pablo parish funeral book are more than 50% less than that of the La Candelaria parish. The total number of deaths in the earthquake across Caracas would be an even lower number if we assumed 38 deaths from each of 19 parishes. Utilizing an average fatality rate from the funeral books of 60 deaths per parish yields a death toll estimation of 1640. Together, an average of these calculations suggests that ~2000 people died in Caracas in the earthquake (Altez, 2005a, 2006). INTENSITIES The widespread destruction of Caracas city was heterogeneous in its distribution. Damage was determined largely by the differences in the construction quality and by the maintenance status of the buildings. A spatial relationship between damage and social class therefore does not exist (Cunill Graü, 1987). The condition of the building at the time of the earthquake explains why there were churches with more or less damage. Some buildings

The social and material impacts of the 1812 earthquake in Caracas, Venezuela were virtually unharmed, and others totally collapsed. There were houses that resisted shaking and others that acted as the burial ground of the people. Estimation of the intensity of the earthquake is more difficult if the extent of damage is related to construction age and quality. The heterogeneous destruction of the buildings and a lower death count found in one historical record for Caracas can now be better appreciated. Table 4 summarizes the reported structural damage in Caracas. Values range from 33% to 90% of the buildings destroyed. These values appear as broad estimations of the areal percentage of damage. Modern assignment of Modified Mercalli scale (MMS) intensity values for Caracas are based on the historical damage records and include MMS X (FUNVISIS, 1997), MMS IX (Fiedler, 1972; Grases, 1990; Altez, 2005b), and MMS VIII–IX (Altez, 2005a). Thorough investigation of the historical sources suggests a lowering of the intensity values (Altez, 2005a). In all these observations, the earthquake intensity values for Caracas are high. However, the Modified Mercalli scale allows only a general perspective of the macroseismic effects. According to the map presented in this work (Fig. 5), it is possible to assign intensities to each damage site reported in the city using the EMS scale. This map shows that the damage distribution varies and can lead to a microzonation of intensity values. The largest variety of building responses to the earthquake is around the Main Square. In the north area, the damage level is V in the EMS scale. For the rest of the city, EMS values of IV are based on the analyses of the quantity and level of destroyed houses. Toward the north, the city felt a tremor that produced general damages with an EMS intensity of V. Most of the city of Caracas felt EMS intensity IV values (Altez, 2005a). CONCLUSIONS From the analyses of the historical data, the following conclusions are drawn: 1. The intensity values for the earthquake of 26 March 1812 in the city of Caracas have been overestimated by previous investigators. The reason for the higher intensity values is the lack of a critical evaluation of the historical context of the primary sources and the initial conditions of the buildings at the time of the earthquake.

55

2. The disaster was not the action of only one negative force (i.e., the earthquake). The earthquake occurred during a time of war and at a critical transition point in Venezuelan political, economic, and societal history. The war, the paradigmatic change represented by the pass from colonial society to forced modern institutions, and the social and political ambition of the criollos produced a vulnerable historical context. 3. The damage distribution from the earthquake of 1812 in Caracas is heterogeneous and determined by the quality and state of construction. This explains why the damage does not directly correlate with the constructions of the social classes or with the size of the construction. 4. The number of earthquake victims in Caracas in 1812 may be closer to 2000, based on analyses of three funeral books from the era. This value is lower than previous estimates about the deaths cipher. 5. The number of fatalities caused by the earthquake in the city of Caracas has been generally overestimated and historically confused with the deaths caused by other factors that were occurring concurrently, including war, migration, and famine. 6. A greater number of deaths is attributed to collapse of houses and not of religious and civic buildings. Most inhabitants were at the cathedral in celebration of Holy Thursday. Those inhabitants who remained in domestic dwellings were more likely to be killed because the roofs of houses were high, heavy, and had fragile support, making them very susceptible to collapse. 7. A seismic microzoning analysis of the city shows damages distributed among intensity values III, IV, and V on the EMS scale, with 46% of buildings and 60% of houses with damage V, while 46% of the buildings show damage IV. 8. The north area of the city of Caracas experienced higher seismic intensities (EMS V) compared to the rest of the city (EMS IV). This conclusion was also reached using similar methods by Altez (2005a), Yamazaki et al. (2005), and Schmitz et al. (2005, 2008). ACKNOWLEDGMENTS Thanks are due to Franck Audemard, André Singer, Jaime Laffaille, and Franco Urbani for their advice and lessons in seismology and geology. Special thanks go to Tina Niemi for her effort in the corrections to and observations about the manuscript. This article is a contribution to the United Nations Educational,

TABLE 4. DAMAGE ESTIMATIONS FROM THE 1812 EARTHQUAKE IN THE CITY OF CARACAS Damages Source 9/10 of the city destroyed Delpeche, mentioned in Centeno Graü (1940), 15 May 1813 3000 destroyed houses Larrain (1958) Almost all the churches and 2/3 of the houses Urquinaona y Pardo (1820) 50% of the city ruined Méndez (1957, originally 1816) 7 churches completely ruined and the rest can be repaired Linares (1816) 1/3 of the houses completely ruined 1/3 of buildings fallen Coll y Prat (1960) 50% of houses fallen Díaz (1829) 8/10 of the city destroyed Ibarra (1862 )

Figure 5. Detailed damages of Caracas. The dark circles show the most destructive effects (V in “Classification of damage to masonry buildings” from EMS scale), and the light circles show the minor damages (III–IV).

The social and material impacts of the 1812 earthquake in Caracas, Venezuela Scientific and Cultural Organization–funded International Geoscience Programme IGCP 567, “Earthquake Archaeology: Archaeoseismology along the Alpine-Himalayan Seismic Zone.” REFERENCES CITED Historical Primary and Secondary Sources Actas del Cabildo de Caracas, 1972, Volume II, 1812–1814: Caracas, Venezuela, Concejo Municipal del Distrito Federal, Tipografía Vargas, 412 p. Ascanio, 1813, Antonio Ascanio, Autobiografía, may be from 1813: National Academy of History Archive, Francisco Javier Yanes Section, volume 28. Book of Burials, Adults, San Pablo Parish, 1808–1812, Book IX, 1808–1812: Caracas, Venezuela, Archdiocese Archive of Caracas, Episcopal Section. Book of Burials, Chacao Parish, No. 2, 1797–1821, Book 2, 1797–1821: Caracas, Venezuela, Archdiocese Archive of Caracas, Episcopal Section. Book of Burials, La Candelaria Parish, No. 7, 1806–1817, Libro de Entierros de la Parroquia La Candelaria: Book 7, p. 1806–1817. Coll y Prat, manuscript from March 31, 1812, Narciso Coll y Prat, Circular a Todas las Parroquias y Pueblos del Arzobispado de Caracas: Caracas, Venezuela, Archdiocese Archive of Caracas, Parishes Section. Coll y Prat, 1957, April 18, 1816: Narciso Coll y Prat, in Los Desastres del Terremoto de 1812: Crónica de Caracas, v. 32, p. 535–536. Coll y Prat, N., 1960, Exposición al rey, 1818, in Memoriales de la Independencia de Venezuela: Caracas, Venezuela, Biblioteca de la Academia Nacional de la Historia, p. 85–386. Consulate answer to Intendant, January 29, 1813, Archivo General de Indias (General Indias Archive): Caracas, Venezuela, File 824. Delpeche, L., 1813, Relación del último terremoto de Caracas: Journal de Paris, 15 May 1813, p. 1. Díaz, J.D., 1817, A los autores y agentes del 19 de abril: Gaceta de Caracas (journal from Caracas), May 21, p. 1027–1034. Díaz, J.D., 1829, Recuerdos sobre la Rebelión de Caracas: Madrid, Imprenta de León Amarita, 408 p. Dionisio Franco, manuscript dated in February 13, 1813, Archivo General de Indias (General Indias Archive): Caracas, Venezuela, File 824. Duane, W., 1968, Viaje a la Gran Colombia en los años 1822–1823: De Caracas y La Guaira a Cartagena, por la Cordillera hasta Bogotá, y de aquí en adelante por el Río Magdalena: Caracas, Venezuela, Instituto Nacional de Hipódromos, 372 p. Forrest, 1812, Forrest (captain) to admiral Stirling, Curazao, March 30, 1812, in Parra Pérez, C., 1939, Historia de la Primera República en Venezuela, Caracas, Volume II: Caracas, Venezuela, Tipografía Americana, p. 211–212. Guerra, F.-X., 1992, Modernidad e Independencia. Ensayos Sobre las Revoluciones Hispánicas: México DF, México, Fondo de Cultura Económica, 407 p. Guzmán, April 26, 1816, Juan Joseph Guzmán, in Los Desastres del Terremoto de 1812: Crónica de Caracas, 1957, v. 32, p. 537. Heredia, J.F., 1895, Memorias Sobre las Revoluciones de Venezuela: Paris, Librería de Garnier Hermanos, 304 p. Ibarra, A., 1862, Temblores y Terremotos: El Independiente (journal from Caracas), April 1862, no. 587, p. 3–4. Irvine, 1818, John Baptiste Irvine, cited by Robertson, W.S., 1918, Francisco de Miranda y la Revolución de la América Española: Bogotá, Colombia, Biblioteca de Historia Nacional, p. 335. Ker Porter, Sir R., 1997, Diario de un Diplomático Británico en Venezuela, 1825–1842: Caracas, Venezuela, Fundación Empresas Polar, Caracas, 1039 p. Larrain, J.B., 1958, Representación ante el Muy Ilustre Ayuntamiento, 15 de febrero de 1813: Boletín de la Academia Nacional de la Historia, v. 162, p. 122–127. Linares, 1816, Pablo Linares, May 18, 1816, in Los Desastres del Terremoto de 1812: Crónica de Caracas, 1957, v. 32, p. 546–548. Lynch, J., 1985, Las revoluciones hispanoamericanas, 1808–1826: Barcelona, Spain, Editorial Ariel, 435 p. Méndez, 1957, April 18, 1816: Silvestre José Méndez, in Los Desastres del Terremoto de 1812: Crónica de Caracas, 1957, v. 32, p. 550–551. Monteverde, manuscript from August 21, 1812, Domingo de Monteverde, Ordenes a los Habitantes de la Provincia de Caracas: General National

57

Archive, Gobernación y Capitanía General section, volume CCXX, Doc. 214, folio 310. Palacio Fajardo, M., 1817, An account of the earthquake of Caracas: The Quarterly Journal of Science, London, v. 2, p. 400–402. Parra Pérez, C., 1939, Historia de la Primera República en Venezuela, Volume II: Caracas, Venezuela, Tipografía Americana, 520 p. Rojas, A., 1879, La catástrofe de 1812: La Opinión Nacional (journal from Caracas), July 12, p. 2. Roscio, 1812, Juan Germán Roscio to Luis López Méndez, Caracas, April 9, 1812: John Boulton Found Archive, Microfilmed Documents, rol c-13. Scott to Monroe, manuscript, Caracas, November 16, 1812, Alexander Scott to James Monroe, Baltimore: Colección de Documentos Diplomáticos, Biblioteca Nacional, Venezuela. Semple, 1812, John Semple a Mathew Semple, Tócome, April 3, 1812, in Tres Testigos Europeos de la Primera República, 1974: Caracas, Ediciones de la Presidencia de la República, p. 86–89. Urquinaona y Pardo, P., 1820, Relación documentada del origen y progresos del trastorno de las provincias de Venezuela hasta la exoneración del Capitán General Don Domingo de Monteverde hecha en el mes de diciembre de 1813 por la guarnición de la plaza de Puerto Cabello: Madrid, Imprenta Nueva, 322 p.

Specialized Literature Altez, R., 1998, Cronometrización extemporánea: Los sismos del 26 de marzo de 1812 en Caracas y Mérida: Revista Geográfica Venezolana, v. 39, no. 1–2, p. 297–325. Altez, R., 2005a, El terremoto de 1812 en la ciudad de Caracas: Un intento de microzonificación histórica: Revista Geográfica Venezolana, Special Issue, p. 171–198. Altez, R., 2005b, Los sismos del 26 de Marzo de 1812 en Venezuela: Nuevos aportes y evidencias sobre estos eventos: Boletín Técnico IMME (Instituto de Materiales y Modelos Estructurales de la Universidad Central de Venezuela), v. 43, no. 2, p. 11–34. Altez, R., 2006, El desastre de 1812 en Venezuela. Sismos, vulnerabilidades y una patria no tan boba: Caracas, Venezuela, Universidad Católica Andrés Bello-Fundación Empresas Polar, 522 p. Altez, R., and Laffaille, J., 2006, La microzonificación sismo-histórica como complemento fundamental de la evaluación de la amenaza sísmica: Revista de la Facultad de Ingeniería, v. 21, no. 4, p. 117–127. Altez, R., Parra, I., and Urdaneta, A., 2005, Contexto y vulnerabilidad de San Antonio de Gibraltar en el siglo XVII. Una coyuntura desastrosa: Boletín de la Academia Nacional de la Historia, v. 352, p. 181–209. Audemard, F.A., 1993, Néotectonique, Sismotectonique et Aléa Sismique du Nord-Ouest du Vénézuéla (Système de failles d’Oca-Ancón) [Ph.D. thesis]: Montpellier, Université Montpellier II, France, 369 p. + appendix. Audemard, F.A., 1998, Evolution géodynamique de la façade Nord Sud-Américaine: Nouveaux apports de l’histoire géologique du Bassin de Falcón, Vénézuéla, in Transactions of the 3rd Geological Conference of the Geological Society of Trinidad and Tobago and the XIV Caribbean Geological Conference, Trinidad, 1995, Volume 2: San Fernando, Trinidad and Tobago, Geological Society of Trinidad and Tobago, p. 327–340. Audemard, F.A., and Singer, A., 1996, Active fault recognition in northwestern Venezuela and its seismogenic characterization: Neotectonic and paleoseismic approach: Geofísica Internacional, v. 35, no. 3, p. 245–255. Audemard, F.A., Machette, M., Cox, J., Hart, R., and Haller, K., 2000, Map and Database of Quaternary Faults in Venezuela and Its Offshore Regions: U.S. Geological Survey Open-File Report 00-18, 79 p. + map. Audemard, F.A., Romero, G., Rendón, H., and Cano, V., 2005, Quaternary fault kinematics and stress tensors along the southern Caribbean from microtectonic data and focal mechanism solutions: Earth-Science Reviews, v. 69, no. 3–4, p. 181–233, doi: 10.1016/j.earscirev.2004.08.001. Audemard, F.E., and Audemard, F.A., 2002, Structure of the Mérida Andes, Venezuela: Relations with the South America–Caribbean geodynamic interaction: Tectonophysics, v. 345, p. 299–327, doi: 10.1016/S0040 -1951(01)00218-9. Beltrán, C., 1994, Trazas activas y síntesis neotectónica de Venezuela a escala 1:2.000.000, in VII Congreso Venezolano de Geofísica: Caracas, Venezuela, Sociedad Venezolana de Ingenieros Geofísicos, p. 541–547. Centeno Graü, M., 1940, Estudios Sismológicos: Caracas, Academia de Ciencias Físicas, Matemáticas y Naturales, 365 p.

58

Altez

Cunill Graü, P., 1987, Geografía del Poblamiento Venezolano en el Siglo XIX: Caracas, Venezuela, Ediciones de la Presidencia de la República, 3 volumes. European Seismological Commission, Subcommission on Engineering Seismology, Working Group Macroseismic Scales, European Macroseismic Scale, 1998, EMS-98: Luxembourg, European Seismological Commission: http://www.gfz-potsdam.de/portal/gfz/Struktur/Departments/ Department+2/sec26/projects/04_seismic_vulnerability_scales_risk/ EMS-98 (accessed 21 February 2010). Fiedler, G., 1972, La liberación de energía sísmica en Venezuela, volúmenes sísmicos y mapas de isosistas, in Memorias del IV Congreso Geológico Venezolano, Volume IV: Caracas, Venezuela, p. 2441–2462. FUNVISIS (Fundación Venezolana de Investigaciones Sismológicas), 1997, Estudio neotectónico y geología de fallas activas en el Piedemonte surandino de los Andes venezolanos, Proyecto INTEVEP 95-061: Caracas, FUNVISIS, 155 p. Grases, J., 1990, Terremotos Destructores del Caribe 1502–1990: Montevideo, UNESCO-RELACIS, 132 p. Grases, J., and Rodriguez, J.A., 2001, Estimaciones de magnitud de sismos venezolanos a partir de mapas de isosistas, in Memorias del 2nd Seminario Iberoamericano de Ingeniería Sísmica: Madrid, Spain, 12 p., digital version on CD. Grases, J., Altez, R., and Lugo, M., 1999, Catálogo de sismos sentidos y destructores, Venezuela, 1530–1998: Caracas, Venezuela, Academia de Ciencias Físicas, Matemáticas y Naturales, Facultad de Ingeniería, Universidad Central de Venezuela, Editorial Innovación Tecnológica, 654 p. Guidoboni, E., and Ferrari, G., 2000, Historical variables of seismic effects: Economics levels, demographic scales and buildings techniques: Annali di Geofisica, v. 43, no. 4, p. 687–705. Mendoza Solar, E., 1910, Plano de la Ciudad de Santiago de León de Caracas en el año 1810 (Map of Caracas near 1810): Caracas, Venezuela, Litografía del Comercio, scale 1:1000, 1 sheet. Mocquet, A., 2005, Geological and architectural context of historical earthquakes in eastern Venezuela: Journal of Earthquake Engineering, v. 9, no. 1, p. 129–146, doi: 10.1142/S136324690500175X.

Olson, R.S., and Grawonski, V.T., 2003, Disasters as critical junctures? Managua, Nicaragua, 1972, and Mexico City, 1985: International Journal of Mass Emergencies and Disasters, v. 21, no. 1, p. 5–36. Rodríguez, J.A., and Audemard, F.A., 2003, Sobrestimaciones y limitaciones en los estudios de sismicidad histórica con base en casos venezolanos: Revista Geográfica Venezolana: Universidad de Los Andes, v. 44, no. 1, p. 47–75. Schmitz, M., Hernández, J., Audemard, F., Malavé, G., and Andrade, L., 2005, Proyecto de Microzonificación Sísmica en las ciudades Caracas y Barquisimeto: Serie Técnica Fundación Venezolana de Investigaciones Sismológicas No. 1, p. 260–263. Schmitz, M., Hernández, J.J., Morales, C., Molina, D., Valleé, M., Domínguez, J., Delavaud, E., Singer, A., González, M., Leal, V., and el Grupo de Trabajo del Proyecto de Microzonificación Sísmica de Caracas, 2008, Resultados principales del Proyecto de Microzonificación Sísmica en Caracas, in Conferencia 50 Aniversario de la Sociedad Venezolana de Geotecnia (SVDG), 6 al 9 de noviembre 2008: Caracas, Venezuela, Sociedad Venezolana de Geotecnia, 11 p. Schubert, C., 1984, Basin formation along Boconó–Morón–El Pilar fault system, Venezuela: Journal of Geophysical Research, v. 89, p. 5711–5718, doi: 10.1029/JB089iB07p05711. Soulas, J.-P., 1986, Neotectónica y tectónica activa en Venezuela y regiones vecinas, in Memorias del VI Congreso Geológico Venezolano, Tomo 10: Caracas, Venezuela, Sociedad Venezolana de Geólogos, p. 6639–6656. Yamazaki, Y., Audemard, F.A., Altez, R., Hernández, J., Orihuela, N., Safina, S., Schmitz, M., Tanaka, I., Kagawa, H., and Jica Study Team—Earthquake Disaster Group, 2005, Estimation of the seismic intensity in Caracas during the 1812 earthquake using seismic microzonation methodology: Revista Geográfica Venezolana, Special Issue, p. 199–216.

MANUSCRIPT ACCEPTED BY THE SOCIETY 19 MAY 2010

Printed in the USA

The Geological Society of America Special Paper 471 2010

The impact of the 1157 and 1170 Syrian earthquakes on Crusader– Muslim politics and military affairs Kate Raphael Institute of Earth and Sciences, The Hebrew University, Jerusalem 91904, Israel

ABSTRACT This paper examines the development of a crisis over a critical military-security issue raised by the severe earthquakes that destroyed defensive structures throughout Nur al-Din’s Sultanate of Syria, the Crusader Principality of Antioch, and the County of Tripoli. The earthquakes that struck Syria in 1157 and 1170 are well documented by contemporary historians. The accounts of destruction concentrate on the collapse of many fortresses and town walls. This circumstance strongly influenced regional politics and military affairs. While the first earthquake led to an increase in tension and a rise in violence between the Crusader Kingdom of Jerusalem and the Muslim Sultanate in Syria, the destruction wrought by the 1170 earthquake forced the two sides to accept a formal peace treaty. The two case studies presented here examine the impact of earthquake destruction on decision makers in the complex international arena of medieval Syria.

role in the way each case developed and the decisions made by the regional rulers. The earthquakes in medieval Syria have been studied by several scholars, who have used the historical sources to enable them to understand the seismic dynamics and assess earthquake hazards (Ambraseys and Jackson, 1998; Amiran, 1952; Amiran et al., 1994; Ben-Menahem, 1979, 1991; Guidoboni et al., 2004a, 2004b; Sbeinati et al., 2005). The impact of earthquakes on regional affairs has seldom been examined (Tucker, 1981, 1999; Little, 1999).

INTRODUCTION Most of the political and military conflicts between the Crusader kingdom and principalities and the Muslim Sultanate of Syria during the second half of the twelfth century were related to religious tensions and territorial disputes. In several cases, however, crisis was not triggered by extreme acts or shifts in the religious or political views of regional leaders, but rather by severe natural disasters, such as long periods of drought, crop failure, and earthquakes. The aim of this paper is to examine the impact of the 1157 and 1170 earthquakes that struck Syria on the political and military affairs of the region. Did such events cause an increase or decrease in the level of animosity and violence? What forces influenced decision makers? Who was responsible for the reconstruction of private and public property? The two earthquakes struck at almost the same sites; each set off a chain of political and military reactions. However, each crisis developed in a very different manner. The force of the earthquake, the scale of damage, and its geographical distribution played an important

SHORT HISTORICAL BACKGROUND The area under discussion is what was known in contemporary Arabic sources as Bilad al-Sham (Greater Syria), corresponding to the modern states of Syria, Lebanon, southeast Turkey, northern Jordan, and northern Israel. The narrow strip along the coast was ruled by the Crusader Principality of Antioch and the County of Tripoli. The Kingdom of Jerusalem ruled the

Raphael, K., 2010, The impact of the 1157 and 1170 Syrian earthquakes on Crusader–Muslim politics and military affairs, in Sintubin, M., Stewart, I.S., Niemi, T.M., and Altunel, E., eds., Ancient Earthquakes: Geological Society of America Special Paper 471, p. 59–66, doi: 10.1130/2010.2471(06). For permission to copy, contact [email protected]. © 2010 The Geological Society of America. All rights reserved.

59

60

Raphael

territory that partly correlates with the modern state of Israel and the Palestinian Authority; it also controlled some of the lands south of the Dead Sea in the region of the fortresses of Karak, today southern Jordan. The Muslim territory included the cities of central Syria from Aleppo in the north to Damascus. During the 1140s, the individual Muslim principalities in the region were slowly being united under the house of Zengi. Lands that had been conquered by the Crusaders were gradually being recovered. In 1144, ‘Imad al-Din Zengi, ruler of Mosul (d. 1146), captured the Crusader County of Edessa (modern Ruha, southeast Turkey) and established his rule over Aleppo. A decade later his son, Nur al-Din (1118–1174) had seized Damascus and made it his capital (Prawer, 1984; Runciman 1994; Riley-Smith, 1990). During the 1160s, the interests of the political entities in the region shifted far south, and the struggle between the Crusader Kingdom and the Syrian Sultanate of Nur al-Din was waged in Egypt. Egypt’s declining Fatimid dynasty (a Shiʿite dynasty) and its wealth from international trade and agriculture made it an attractive prize. Amalric (1136–1174), ruler of the Crusader Kingdom of Jerusalem, fought against Nur al-Din’s armies for the control of Cairo and the Nile Delta for almost a decade. The Crusader armies carried out five successive attacks; on one occasion they allied themselves with the Byzantine emperor in order to receive the assistance of his fleet. The Crusader assaults failed, and Amalric eventually had no choice but to acknowledge defeat. Salah al-Din (Saladin), who was one of the leading officers in this campaign, became the administrator of the newly conquered territory in Egypt. Officially, he remained a dependent of Nur al-Din. By 1170, the geopolitical balance in the region had changed considerably. Nur al-Din’s rule over Syria was consolidated, and Egypt was administered under his suzerainty. The earthquake of 1170 thus struck after a decade of intensive fighting in which large Muslim and Christian armies had constantly been on the march back and forth from Syria or the Kingdom of Jerusalem to the Nile Delta and Cairo. Lengthy sieges were conducted, and a number of open-field battles took place (Prawer, 1984; Runciman, 1994). CONTEMPORARY WRITTEN EVIDENCE ON EARTHQUAKE DAMAGE OF 1157 AND 1170 Earthquakes are habitually described by contemporary eyewitnesses in a dramatic tone of voice. Christian and Muslim chroniclers often use the same phrase “a great and terrible earthquake, far more violent than any other within the memory of men now living” (William of Tyre, v. 2, p. 370–371). This is by no means the language of exaggeration, but some caution is necessary, and more than one contemporary source is required in order to verify each case. The destruction of urban defenses and the repairs that followed are described in detail for both 1157 and 1170. Although the number of casualties was high and the damage to private property and public buildings was considerable, the rulers were more concerned with the state of their fortifications and defenses.

In 1157, town walls, towers, citadels, and fortresses were damaged throughout Nur al-Din’s territories. There is sufficient evidence to show that in the neighboring Crusader Principality of Antioch and the County of Tripoli, the scale of damage was significantly lower. The only repairs carried out by the Crusaders due to earthquake damage were to the large fortress of Hisn al-Akrad (Crac des Chevaliers) held by the Order of the Hospitallers. The Grand Master of the Order, Raymond of Le Puy, received a generous donation from Wladislas II, King of Bohemia; this financed the reconstruction works (Elisséeff, 1986). Strangely enough, the main Crusader source for this period, William of Tyre, does not mention the 1157 earthquake at all. Ibn al-Jawzi (1126–1200), the author of al-Muntazam fi Tarikh al-Umam wal-Muluk (The Order of History, the Nations and the Kings), belonged to the intellectual elite of Baghdad and was one of the most influential people in this circle (Laoust, 1986). Although he himself lived in Baghdad, he was well informed and gives a detailed report of the cities that were hit by this earthquake (Fig. 1; Table 1). He is clearly more concerned with the suffering of the local population than with the damage caused to public and domestic building.

…thirteen cities were destroyed in this earthquake, eight in the land of Islam and five in the land of the infidels. In the land of the Muslims: Aleppo, Hama, Shayzar, Kafar-Tab, Aphamia, Hims, al-Maʿarra, and Tell-Harān; and in the land of the Franks: Hisn al-Akrad, ʿArqa, Lattakia, Tripoli and Antioch. At Aleppo one hundred people were killed; as for Hama few survived and in Shyzar only a woman and a slave survived; all the rest perished. At Kafar-Tab only one person survived and at Aphamia the citadel sank. At Homs many scholars died and at Maʿarra a number of people were killed. Tell-Harān was divided in two, exposing the interior of tombs and many houses. At Hisn al-Akrad and ʿArqa all was destroyed and at Lattakia all that remained was one man and a spring and in it there was a hole with mud in its center [in which] stood a statue. Much of Tripoli was ruined and only parts of Antioch remained. (Ibn al-Jawzi, Muntazam, 1992, v. 18, p. 119)

The following is an extract from an account written by the Muslim chronicler Ibn al-Athīr (1160–1233), author of the al-Kāmil fi’l-ta’rīkh (The Complete Work of History). Ibn al-Athīr spent most of his life in Mosul. In later years, he moved to Aleppo. In contrast to Ibn al-Jawzi, he pays greater attention to the urban defenses. In Rajab this year (9 August–7 September 1157) there were many strong earthquakes in Syria, which destroyed much of the country and which caused the death of more people than could be counted. In one moment Hama, Shayzar, Kafar-Tāb, al-Maʿarra, Homs, H ⋅ is⋅ n al-Akrād, ʿArqa, Lattakia, Tripoli and Antioch were ruined. All Syria suffered damage in most of its parts, even if the damage was not total. City walls and citadels were demolished. Nūr al-Dīn Muh⋅ mūd dealt with this in an exemplary manner. He feared for the land since the city walls had been destroyed. He assembled the troops and camped on the frontiers of his land, carrying raids on Frankish territory, while working on the walls in the rest of his lands. He kept this up until he had completed

Impact of the 1157 and 1170 Syrian earthquakes on Crusader–Muslim politics and military affairs

61

Figure 1. Towns and fortresses struck by the 1157 earthquake according to Ibn al-Jawzi’s account.

all the city walls. The great number of people who were killed is sufficiently indicated by the fact that a teacher who was in his town, namely Hama, left the Koran school for some matter of business that occurred, when the earthquake came and destroyed the town. The school collapsed on all the children. The teacher said, “Not a single person came to enquire after any child of his.” (Ibn al-Athīr, 2007, v. 2, p. 87)

TABLE 1. DISTRIBUTION OF DAMAGE FROM THE 1157 EARTHQUAKE, ACCORDING TO IBN AL-JAWZI Muslim Sultanate Crusader Principality Crusader County of Nur al-Din of Antioch of Tripoli Aleppo Antioch ૚iଙn al-Akrād Hama Lattakia Tripoli Shayzar

ȾArqa

Aphamia

According to Ibn al-Athīr, while repairing the fortifications of the central Syrian cities, Nur al-Din organized raiding contingents and ordered them to attack the neighboring Crusader territories. He was not aiming at long-term political or territorial achievements, but rather it seems he feared that Baldwin III (r. 1152–1163), ruler of the Kingdom of Jerusalem, might take advantage of the poor state of the defenses and launch an attack in order to destabilize his rule and conquer part of his newly acquired lands. These fears were not unfounded, for in fact the Crusader principalities had not suffered severe damage, and the

Kafar-Tab al-MaȾarra Hims Tall-Harān

Kingdom of Jerusalem lay outside the zone of the earthquake. Baldwin did not lose much time; a large army was assembled and joined by the forces of the Prince of Antioch and the Count of Tripoli. The Count of Flanders, who was visiting the Holy Land, joined the campaign with his own men. This army entered

62

Raphael

the Sultanate and attacked the Muslim fortress of Qalʿat Yahmur (Chastel Rouge, roughly 20 km northeast of Tripoli), but to no avail; the garrison held its ground (William of Tyre, 1943, v. 2, p. 265). Ibn Qalanisi (d. 1160), a contemporary Muslim chronicler who lived in Damascus, says Nur al-Din was quick to act and recruited a Muslim army to meet them.

Mention has already been made of the departure of al-Malik al-ʿAdil Nur al-Din from Damascus with his troops towards the cities of Syria, on receipt of news that the factions of the Franks (God forsake them) were assembling together and proceeding against them, being emboldened to attack them by reason of the continuous earthquakes and shocks which afflicted them and of the destruction wrought amongst the castles, citadels and dwellings in their districts and marches. [Nur al-Din therefore took measures] to protect and defend them and to bring solace to those of the men of Hims, Kafr Tab, Hamah and elsewhere who had escaped with their lives, whereupon there assembled to join him a great host and vast numbers of men from the fortresses and provincial cities and from the Turkmens. He encamped with them in exceeding force opposite the army of the Franks in the neighborhood of Antioch and encompassed them so that not one horseman of theirs could set out to make a raid. (Ibn Qalanisi, 1932, p. 340–341; my emphasis)

Soon after, the Muslim army dispersed due to the illness of Nur al-Din. Many thought he was on his deathbed and left the field since there was no strong commander able to replace him. Once the siege around the Crusader army was broken, it resumed its march and launched an attack on the fortress of Shayzar. The fortress was saved thanks to a dispute that broke out among the Christian commanders, and a strong Isma’ili force that managed to defeat the Franks (Ibn Qalanisi, 1932, p. 342; William of Tyre, 1943, v. 2, p. 226–268). The Crusader forces returned to their own lands empty-handed. 1170, THE EARTHQUAKE THAT FORCED PEACE BETWEEN THE CRUSADER PRINCIPALITIES AND NUR AL-DIN, THE SULTAN OF SYRIA The evidence for the 1170 earthquake presented next is based on two contemporary historians. William, Archbishop of Tyre (ca. 1130–1185), whose work “The Deeds beyond the Sea” is considered one of the most reliable sources for the study of the Crusader Kingdom of Jerusalem, was probably born in Jerusalem; he spoke the local languages and was familiar with the local cultures. He received his higher education in Europe. On his return, he became close to the royal court in Jerusalem and served as a tutor to Baldwin IV, who later became king. From 1175 until his death, he was both Archbishop of Tyre and Chancellor of the Crusader Kingdom. The second source is Ibn al-Athīr, mentioned previously. In writing about the 1170 earthquake, both Ibn al-Athīr and William of Tyre emphasize the great damage to fortifications. They both describe the deep concern of each ruler for the defense of his territories.

In contrast to the 1157 earthquake, which William ignored, suggesting that the Crusader principalities were hardly damaged, here he describes the destruction in detail.

Strongly fortified cities dating from very early times were completely demolished….The largest cities of our provinces and those of Syria and Phoenicia as well, cities famous throughout the ages for their noble antiquity were prostrated. In Coelesyria, Antioch, the metropolis of several provinces and once the head of many kingdoms, was utterly overwhelmed and its entire population destroyed. The massive walls and the immensely strong towers along their circuit fell in ruins. Churches and buildings of every kind were thrown down with such violence that even now, although much labor and expense have been devoted to their restoration, they are only partially repaired. Among other places destroyed in that same province were Gabala and Laodicea, famous cities on the coast. Of the cities further inland which were still held by the enemy there were destroyed Beroea, also known as Aleppo, Shayzar, Hama, Hims and others. The number of fortresses wrecked was beyond counting. …the great and populous city of Tripoli was suddenly shaken by a violent earthquake, and scarcely a person within the walls escaped. The entire city was reduced to a heap of stones and became a burial place and common sepulcher of the citizens who perished with it. At Tyre, the most famous city of the province, the earth movement was so violent that several massive towers were overthrown. There was, however, no loss of life here. (William of Tyre, 1943, v. 2, p. 370–371)

William ends his account with a sigh of relief, reassuring his readers that his Muslim neighbors were facing similar troubles. He mentions the damage to the main towns of Syria, Aleppo, Shayzar, Hama, and Hims, and he remarks that many other towns lay in ruins (Fig. 2; Table 2). Ibn al-Athīr’s description conveys the fear and urgency that filled the Syrian ruler as he surveyed the poor state of his fortifications. Nur al-Din quickly organized garrisons to safeguard the towns where the defenses had been destroyed. In some towns, the horrendous scale of destruction and the aftershocks drove the citizens away; certain sites were completely abandoned. The sultan then set out and personally supervised the construction of some of the main citadels in the towns along the frontier with the Crusaders.

When Nūr al-Dīn received the news, he went to Baalbek to repair the damage to its wall and citadel. When, however, the news from the rest of the towns came to him, news of the destruction of their walls and citadels and their abandonment by the inhabitants he placed men in Baalbek to repair, protect and guard it and went to Homs, where he did the same, and then to Hama and then to Baʿrin. He was extremely wary of the danger for the towns from the Franks. Then he came to Aleppo, where he saw effects of the earthquake greater than elsewhere, for it had destroyed it utterly and the survivors were totally terror stricken. They were unable to shelter in their houses for fear of aftershocks. They remained in the open. Nūr al-Dīn personally took part in the repair work and so continued until he had rebuilt its walls and mosques. (Ibn al-Athīr, 2007, v. 2, p . 185–186)

Impact of the 1157 and 1170 Syrian earthquakes on Crusader–Muslim politics and military affairs

63

Figure 2. Towns and fortresses struck by the 1170 earthquake according to William of Tyre’s account.

TABLE 2. THE DISTRIBUTION OF DAMAGE FROM THE 1170 EARTHQUAKE, ACCORDING TO WILLIAM OF TYRE AND IBN AL-ATHĪR Muslim Sultanate Crusader Principality Crusader County of Nur al-Din of Antioch of Tripoli Hama Antioch Tripoli Shayzar

Lattakia

Baalbek

Gabala

It is evident that each side was aware of the fact that the neighbor’s fortifications were in a state of ruin. Nevertheless, one is left with a strong impression that each ruler suspected that his enemy might take advantage of the situation and launch a surprise attack. This notion is clearly conveyed by Ibn al-Athīr.

Tyre

Hims Aleppo Beroea (BaȾrin)

Ibn al-Jawzi’s report on the earthquake is brief and his account only mentions Aleppo.

Half of Aleppo collapsed and it was said that eighty thousand people were killed. (Ibn al-Jawzi, 1992, v. 18, p. 188; my translation)

As for the Frankish territory, the earthquake tremors also had the same effect there. They were kept busy repairing their towns, fearful of Nūr al-Dīn for them. Each side was occupied with repair work for fear of the other. (Ibn al-Athīr, 2007, v. 2, p. 186)

Yet, in order to ensure that no side would make a move against the foe and set out to raid or launch a full-scale attack, a treaty of some definition was necessary. Decision makers in times of crisis are subject to both external and internal pressures that may force them to make peace

64

Raphael

(Randle, 1973). Nur al-Din is described as committed to the security of his lands and their inhabitants. William of Tyre presents the treaty as a decision that was made out of sheer fear: fear of Nur al-Din and fear of God.

Both in our territories and in those of the enemy were found halfruined fortresses, open on every side and freely exposed to the violence and the wiles of the foe. But since each man feared that the wrath of the Stern Judge might descend upon him individually, none dared molest his fellow man. Each was engrossed in his own troubles and weighed down by the burden of his own affairs; hence none thought of injuring his neighbor. Peace brought about by the desire of all, ensued, albeit for a short interval, and a truce was arranged through fear of divine wrath. Each, while momentarily expecting the outpouring of righteous anger from heaven in punishment for his sins, refrained from acts of hostility and curbed his own evil impulses. (William of Tyre, 1943, v. 2, p. 371; my emphasis)

The Crusaders’ fear of Nur al-Din indicates a shift in the regional balance of power. Nur al-Din’s recent conquest of Egypt had made clear his determination and his military abilities. William does not speak of any disagreements or difficulties in concluding the truce. There was no exchange of prisoners of war; neither side was required to pay a tribute, no specific conditions were made, and no lands were demanded or exchanged. William’s Latin terminology is somewhat elusive. “Pax hominum studio procurata, et foedus compositum, divinorum iudiciorum timore conscriptum.” Pax may mean: a pact to end or avert hostilities; a pact granted by God; or a settlement; or simply peace. Foedus is a formal agreement between states or peoples. Conscriptum may refer here to a charter. The frequent mentioning of divine wrath and anger may well signify that William of Tyre was referring to a pact granted by God. Thus, on this reading, it is possible that there was no proper legal written document. William’s words and phrasing seem to indicate that this was a gentleman’s agreement, or a quiet mutual understanding, as suggested by Prawer (1984). Ibn al-Athīr’s report, on the other hand, leaves no room for speculation concerning the nature of this truce, and whether there was a formal agreement rather than a loose understanding. According to Ibn al-Athīr, a formal truce was indeed drawn up between Nur al-Din’s Sultanate, the Principality of Antioch, and the County of Tripoli. This is made clear in a detailed episode that took place not long after the truce was concluded. The term used by Ibn al-Athīr is hudna, meaning peace, truce, or armistice (Ibn al-Athīr, 1966, v. 11, p. 373–374). In the autumn of 1171, a Frankish force from Tripoli and Antioch seized two Muslim merchant ships. Nur al-Din was furious and accused the Franks of violating the truce, demanding that the merchandise be returned.

…Between them [the Franks] and Nur al-Din there was a truce which they treacherously broke. Nur al-Din sent to them about the matter

and about their restoring the merchants’ property they had taken. (Ibn al-Athīr, 2007, v. 2, p. 200)

The Franks ignored the sultan’s demand. Nur al-Din did not hesitate to act. He sent a force to raid the cities of Tripoli and Antioch, and a number of smaller fortresses in the neighborhood were sacked. Frankish territory was set ablaze, plundered, and a number of people were killed. The Muslim force returned with a large amount of booty. Following this destructive raid, the Franks reviewed their situation and decided to renew the truce.

The Franks made contact with him [Nur al-Din] and offered to restore what they had taken from the two ships and to renew the truce. This was accepted…. (Ibn al-Athīr, 2007, v. 2, p. 200)

REPAIRING THE DAMAGE The twelfth century witnessed an exceptionally high number of earthquakes that struck central Syria and the coast. The magnitude of the major 1157 earthquake has been estimated at 7–7.8 (Amiran et al., 1994; Ben-Menahem, 1979; Guidoboni et al., 2004b). Four strong shocks preceded the 1157 late summer earthquake; two were felt in April and two in July. The latter caused some damage in Shayzar, Hama, Kafr Tab, Aphamia, and the area of Aleppo (Ibn Qalanisi, 1932, p. 328–329). These smaller events are well documented by Ibn Qalanisi. The 1170 earthquake has been graded 7–8 (Amiran et al., 1994; BenMenahem, 1979; Guidoboni et al., 2004a). The reconstruction of large-scale fortifications was a long and expensive process. The frequent earthquakes, the tremors that preceded them, and the aftershocks that followed undermined many of these defenses and rendered them unstable and dangerous. The Latin sources provide little evidence of Frankish rulers intervening in the reconstruction work. The only information we have concerns the post-1157 repairs to the large fortress of Crac des Chevaliers. The funds for the reconstruction came from Europe, and the Grand Master of the Order of the Hospitallers oversaw the work. By the late twelfth century, the military orders were well established, almost entirely independent and often better financed than most of the Crusader kings, princes, and counts. On the Muslim side, matters seem better organized. Concerning the financing, it appears that Nur al-Din carried out a sound economic policy carefully adjusted to the financial difficulties of the local inhabitants. An interesting observation is made by Professor Lev in an article titled “The Social and Economic Policies of Nur al-Din.” Lev suggests that disrupted economic activity after the 1157 earthquake caused Nur al-Din to reduce the taxes throughout the sultanate (Lev, 2004). These steps no doubt helped to encourage and revive trade and commercial activity in local markets, enabling the country to recover at a faster pace.

Impact of the 1157 and 1170 Syrian earthquakes on Crusader–Muslim politics and military affairs Nur al-Din was actively involved in the survey and repair of his fortifications. His political power in Syria was stronger and more centralized than that of the Crusader rulers in the County of Tripoli and the Principality of Antioch, where presumably each town and fortress saw to its own defenses. In the eulogy of the Sultan, Ibn al-Athīr dedicates a long passage to Nur al-Din’s public works. Not surprisingly, he opens this chapter with the following passage:

As for the public works, he built the walls of the cities and castles of all Syria, for example Damascus, Homs, Hama, Aleppo, Shayzar, Baalbek and others. (Ibn al-Athīr, 2007, v. 2, p. 223)

The rebuilding of the defenses was seen as the Sultan’s responsibility. It seems that thanks to his sound economic policy, his treasury was well balanced, and the funds for the reconstruction work came from the sultanate’s coffers. AN “EMERGENCY PEACE” In 1170, Nur al-Din acted with care and caution, avoiding a military conflict. Why was the military and political response to the earthquake of 1170 different from that of 1157? What drove the two sides to accept a truce? The answer lies in the historical background and perhaps in a reassessment of the strength of the 1170 earthquake, which seems to have been considerably stronger and more destructive than that of 1157. The damage spread throughout the entire region, affecting both sides equally. Each side no doubt moved to overcome its own internal conflicts, social tensions, and disputes. A few months prior to the 1157 earthquake, the Crusader and Muslim forces were engaged in an ongoing struggle over the town and region of Banyas (at the southern foot of Mount Hermon), on the Muslim-Crusader frontier. The town itself was held by Humphrey of Toron. During the winter, the region was settled by nomadic tribes who reared large herds of horses on the rich pasture. They paid rent to the King of Jerusalem for the land they were using. In the winter of 1157, a Crusader force together with Baldwin III (the king did not initiate this act, but joined the raid) massed a large-scale attack. The nomadic population was slain, and a huge amount of booty was collected, including numerous horses. William of Tyre describes the results of this raid:

…the amount of booty taken in this raid was never equaled in our land. Yet this deed brought no glorious or laudable renown to our people, for they had violated a treaty of peace. (William of Tyre, 1943, v. 2, p. 256)

Nur-al-Din was determined to conquer Banyas. He besieged the town twice during the late spring and early summer of 1157. Both sieges failed owing to the arrival of Crusader reinforce-

65

ments. After the first siege, Baldwin and a large Crusader force were surprised by a Muslim ambush near Banyas. The king barely escaped, and a large number of his knights were taken hostage and paraded through the streets of Damascus. This period of raids, ambushes, and sieges around the locality of Banyas was interrupted by the earthquake that struck the region in the late summer. The destruction on the Muslim side was considerably greater than that in the Crusader territory. This had important military and political implications. The destruction of castles and town fortifications led to a state of emergency within the Muslim sultanate. Its influence on political and military affairs was striking: It changed the balance of power in the region, which for a short time tilted in favor of the Crusader Kingdom. The Crusaders seized the opportunity to attack their chief enemy, and they made a decisive move toward Syria. The level of the conflict was immediately upgraded. Nur al-Din was aware that the Crusader rulers had no interest or reason to negotiate a truce. He recruited an army that prevented the Crusaders from taking two of his fortresses, although it seems that the disputes among the Crusader leadership had an equally significant part in the failure of their campaign. As noted already, the earthquake of 1170 came after a decade of fighting in Egypt. Both sides were engaged in expensive and distant campaigns that involved large armies. Both seem to have exhausted their resources. The destruction of fortifications on both sides by the earthquake brought the conflict to a standstill. It forced Nur al-Din and Amalric to keep their forces at home and refrain from violent military acts. Under certain circumstances, severe natural disasters may change or tilt the balance of power and alter the course of events. In order to do so, they must be of extreme strength and leave behind them mass destruction, or threaten the lives of large populations (in the case of droughts and famines). The 1170 truce was an emergency policy, necessary to allow each side to recover, not from the damage of war, but from an earthquake that brought down the fortifications throughout the region. However, the disaster was not strong enough to change military or political ideology or concepts. William of Tyre clearly states this was to be a short-term agreement. He knew that once the fortifications were in order, their storerooms stocked, and garrisons reestablished and armed, the truce would not hold. CONCLUSIONS The aim of this paper was to examine the development of political and military affairs between the Crusader states and the Muslim Sultanate after the severe damage caused by the earthquakes of 1157 and 1170. While the first earthquake led to an increase in tension and a rise in violence, the destruction wrought by the 1170 earthquake forced the two sides to accept a formal peace treaty. Although there are a number of cases that clearly show that severe environmental disasters will force rival rulers to sign short-term peace treaties, there is no pattern or rule to the behavior and the way rulers make decisions in times of crisis.

66

Raphael

Each case must be carefully studied in the light of the wider international affairs in the region. The king’s or sultan’s personality, military strength, wealth, and experience play an important role in their ability to cope with large-scale disasters. While controlling earthquakes and eliminating the death and destruction that followed was impossible, maintaining peace or ending aggressions in times of a severe crisis were, under certain circumstances, relatively feasible policies. Some leaders ignored the damage, failed to read the political map, and missed the opportunity to reduce the violence; others were wiser, more experienced, and negotiated for a short-term peace treaty. The main benefactors were the people, for it was no doubt easier to attend to the damage and reach full recovery within a shorter period when the region was peaceful. ACKNOWLEDGMENTS The research for this paper was funded by the Galilee Project, affiliated with the Hebrew University of Jerusalem, Israel, and the University of York, England. I would like to thank Michal Kidron from the Cartographic Laboratory, Department of Geography, The Hebrew University of Jerusalem, for preparing the figures. This article is a contribution to the United Nations Educational, Scientific and Cultural Organization–funded International Geoscience Programme IGCP 567, “Earthquake Archaeology: Archaeoseismology along the Alpine-Himalayan Seismic Zone.” REFERENCES CITED Ambraseys, N.N., and Jackson, J.A., 1998, Faulting associated with historical and recent earthquakes in the Eastern Mediterranean region: Geophysical Journal International, v. 133, p. 390–406, doi: 10.1046/j.1365 -246X.1998.00508.x. Ambraseys, N.N., Melville, C.P., and Adams, R.D., 1994, The Seismicity of Egypt, Arabia, and the Red Sea: A Historical Review: Cambridge, UK, Cambridge University Press. Amiran, D.H.K., 1952, A revised earthquake catalogue of Palestine: Israel Exploration Journal, v. 2, p. 48–62. Amiran, D.H.K., Arieh, E., and Turcotte, T., 1994, Earthquakes in Israel and adjacent areas: Macroseismic observations since 100 B.C.E.: Israel Exploration Journal, v. 44, p. 260–305.

Ben-Menahem, A., 1979, Earthquake Catalogue for the Middle East (92 B.C.– 1980 A.D.): Bollettino di Geofisica Teorica ed Applicata, v. 21, no. 84, p. 245–310. Ben-Menahem, A., 1991, Four thousand years of seismicity along the Dead Sea rift: Journal of Geophysical Research, v. 96, p. 20,195–20,216, doi: 10.1029/91JB01936. Elisséeff, N., 1986, Hisn al-Akrad, in Encyclopedia of Islam, Volume 3 (2nd ed.): Leiden, the Netherlands, Brill, p. 503–506. Guidoboni, E., Bernardini, F., and Comastri, A., 2004a, The 1138–1139 and 1156–1159 destructive seismic crises in Syria, south-eastern Turkey and northern Lebanon: Journal of Seismology, v. 8, p. 105–127, doi: 10.1023/B:JOSE.0000009502.58351.06. Guidoboni, E., Bernardini, F., Comastri, A., and Boschi, E., 2004b, The large earthquake on 29 June 1170 (Syria, Lebanon, and central southern Turkey): Journal of Geophysical Research, v. 109, p. B07304, doi: 10.1029/2003JB002523. Ibn al-Athīr, I., 2007, The Chronicle of Ibn al-Athīr for the Crusading Period from Al-Kāmil fi’l-Ta’rīkh, Volume 2 (translated by D.S. Richards): Aldershot, Ashgate, 401 p. Ibn al-Athīr, I., and ʿIzz al-Dīn ʿAlī, 1966, Al-Kāmil fī’l-Ta’rīkh, Volume 11 (C.J. Tornberg ed.): Beirut, Dar Beirut, 585 p. Ibn al-Jawzi, 1992, Muntazam, Volume 18: Beirut, Dar Beirut. Laoust, H., “Ibn al-Djawzi,” 1986, Encyclopaedia of Islam, Volume 3 (2nd ed.): Leiden, the Netherlands, Brill, p. 751–752. Lev, Y., 2004, The social and economic policies of Nur al-Din (1146–1174): The Sultan of Syria: Der Islam, v. 81, p. 218–242, doi: 10.1515/islm .2004.81.2.218. Little, D.P., 1999, Data on earthquakes recorded by Mamluk historians, in Zachariadou, E., ed., Natural Disasters in the Ottoman Empire: Rethomnon, Crete University Press, p. 137–151. Prawer, J.A., 1984, History of the Latin Kingdom of Jerusalem, the Crusades and the First Kingdom: Jerusalem, Bialik Institute. Qalanisi, I., 1932, The Damascus Chronicle of the Crusades (translated by H.A.R. Gibb): London, Luzac and Co., 368 p. Randle, R.F., 1973, The Origins of Peace; A Study of Peacemaking and Structure of Peace Settlement: New York, The Free Press, 307 p. Riley-Smith, J., 1990, The Crusades: A Short History: London, Athlone Press, 302 p. Runciman, S.A., 1994, History of the Crusades, Volume 2: London, The Folio Society, 428 p. Sbeinati, M.R., Darawcheh, R., and Mouty, M., 2005, Catalog of historical earthquakes in and around Syria: Annali di Geofisica, v. 48, p. 347–435. Tucker, W., 1981, Natural disasters and the peasantry in Mamluk Egypt: Journal of Economic and Social History of the Orient, v. 24, pt. 2, p. 215–224, doi: 10.2307/3631995. Tucker, W., 1999, Environmental hazards, natural disasters, economic loss, and mortality in Mamluk Syria: Mamluk Study Review, v. 3, p. 109–128. William Archbishop of Tyre, 1943, A History of Deeds Done beyond the Sea (translated by E.A. Babcock and A.C. Krey), 2 vols.: New York, Columbia University Press. MANUSCRIPT ACCEPTED BY THE SOCIETY 19 MAY 2010

Printed in the USA

The Geological Society of America Special Paper 471 2010

Western Crete: From Captain Spratt to modern archaeoseismology Manolis I. Stefanakis* Classical Archaeology and Numismatics, Department of Mediterranean Studies, University of the Aegean, 1 Democratias Ave, GR-851 00 Rhodes

ABSTRACT The earliest use of seismological observation to identify and date archaeological sites in western Crete was attempted by Captain T.A.B. Spratt in the late nineteenth century. Since then, the development of the subdiscipline of archaeoseismology has offered a great deal to our understanding of western Crete, especially regarding major sites such as Phalasarna and Kissamos. This paper is a review and summary of archaeoseismology in western Crete, presenting the archaeoseismological and excavation evidence from Phalasarna and Kissamos. It also presents evidence from other archaeological sites in western Crete and expresses the potential the region has for future archaeoseismological research.

ley describes them without having been sensible of their purpose. I was instantly impressed, for several reasons, that here was the ancient or artificial port, although full 200 yards from the sea and nearly 20 feet above it. My first idea was, that the ancients had the means of hauling their vessels into it as a dry dock; but at last the coast elevation was remembered, and on measuring the sea marks at its upper level here, I found that the bed of this ancient port is now 3 or 4 feet below that level; so that I had only to imagine the coast again let down 22 feet 6 inches, the amount it has been elevated here and at Grabusa, when the sea would immediately flow into the ancient port, and float any small craft within it. Geologically the recognition of this ancient port has another interest; it establishes the recent origin of this remarkable up heaving of the western end of Crete, which, however, is not surprising, as elsewhere ancient harbours have been lifted into the air, rocks have become islets, and maritime cities or buildings placed many yards from the shore. These facts will enable me to reconcile in some instances the ancient geography with the modern, and thus to verify points otherwise very difficult. For example, Suia is noticed in the Stadiasmus as a town with a good port (πόλις εστί και λιμένα καλόν έχει), and as following next to Poekilassos, its position is easily recognized. There are so few of the ports of Crete so described in the Stadiasmus, that I naturally looked for a well-sheltered harbour. Pashley says nothing about it, and to look at the locality, few would hope to find a port. A straight and steep shingle beach, off which there is no anchorage, stretches across the mouth of the valley of Suia, and beyond the points of the hills on either side. These points, however, were sea-cliffs, formerly rising out of the beach, to about the height

PHALASARNA 1851: ARCHAEOSEISMOLOGY AT BIRTH?

I made an interesting discovery in the western part of the island, viz., that it has been subject to a series of elevations, amounting to the maximum of 24 feet 6 inches, which occurs near Poekilassos and Suia. In the middle of the island, at Messara, the Fair Havens, and Megalo Kastro, there is none. The eastern end of the island has dipped a little. The up heaving is towards the western end. I had observed it to be about 7 feet in Suda Bay many years ago; but supposed it to be of a time prior to history, although there was a freshness in the markings which might have induced me to suspect they were of a more recent date. When at Kissamo, I observed that the ancient mole was remarkably high out of the water, and the port almost choked by sand. But the latter is so common an occurrence that it did not open my eyes, although the height of the naked unhewn rocks which formed the mole ought to have done so. On going to Phalasarna I looked for its ancient port, mentioned by Scylax, and in the Stadiasmus as the Emporium; but I could find no artificial work in the sea. There is, however, a long ledge of rocks, or rather an islet which lies off it, helping to form a natural but not an artificial harbour. This satisfied me in part, till, on examining the ruins, I saw in the plain a square place, enclosed by walls and towers, more massive and solid than those of the city. Pash*[email protected]

Stefanakis, M.I., 2010, Western Crete: From Captain Spratt to modern archaeoseismology, in Sintubin, M., Stewart, I.S., Niemi, T.M., and Altunel, E., eds., Ancient Earthquakes: Geological Society of America Special Paper 471, p. 67–79, doi: 10.1130/2010.2471(07). For permission to copy, contact [email protected]. © 2010 The Geological Society of America. All rights reserved.

67

68

Stefanakis

of 23 feet; and on them the old sea level is shown distinctly by the appearance of the rock, as well as by a line of cylindrical holes, the cells of boring sea-shells, in some of which the shells still remain. Pashley speaks of the town and ruins of Suia as lying on the E. side of the torrent or valley, but takes no notice of the western side, where a little plain within a long ridge of ruined buildings, and nearly 300 yards long and 60 or 70 broad, runs parallel to the shore. This was undoubtedly the tongue of land which sheltered the port lying behind it. The position of the port itself is indicated by a hollow or flat depression of the plain, which depression would even now be overflowed by the sea, if the island was again let down to its old level. Hence it seems evident that this great elevation of the coast must be looked upon as subsequent to the existence of these ancient cities, and subsequent, therefore, to the decline of the Roman Empire. —Spratt and Leake (1854 )

Captain T.A.B. Spratt, who surveyed Crete in the early 1850s (Spratt, 1865; Psillakis, 2007), sent the above report to the Royal Geographic Society of London in 1854. He can therefore be credited with the earliest correct “reading” of the geomorphology of western Crete, having observed the coastal uplift west from Souda Bay, westward and all the way south, to modern Loutro (Figs. 1 and 2). Captain Spratt was also the first scholar to locate the ancient harbor of Phalasarna (Fig. 3), where the geomorphology was radically changed in late antiquity. He attributed this change to the uplift/ natural catastrophe he had observed affecting the west coast of Crete, assuming that this event occurred after Phalasarna’s blossoming, in Hellenistic times and connecting it—not correctly but still rationally—with the decline of the Roman empire (Spratt, 1865, chapter XIX). Thus, Captain Spratt of the Royal British Navy, archaeologist, historian, geologist, paleontologist, and naturalist (Richards, 1888), was also the first archaeoseismologist in Crete. If Spratt had been familiar with the ancient chronicler Ammianus Marcelinus (ca. A.D. 330–ca. 392), he probably would have gotten the date of the uplift correct too. Ammianus was witness to the one of the stronger earthquakes in the history of Mediterranean, which was accompanied by the most violent tsunami (Ammianus Marcelinus Res Gestae 26.10, lines 15–19), a natural catastrophe recorded by many ancient sources (Stiros, 2001, 2009; Kelly, 2004; Stiros and Drakos, 2006). The emergence of archaeoseismology, the scientific discipline that studies past earthquakes in the archaeological record, in the twentieth century, came to bridge the gap between instrumental and historical seismology, on the one hand, and paleoseismology and earthquake geology, on the other hand and marked the beginning of a fertile cooperation between different sciences, with archaeology, tectonics, sedimentology, paleontology being among the most significant. Through the production of quantitative parameters, necessary to fully describe a past earthquake, a multidisciplinary approach, and in situ analysis of the evidence provided under different contexts, archaeoseismology offers solutions to a great number of problems, such as the positioning of the destruction level of past earthquake activity, the analysis of the deformations applied to

static structures, and the analysis of the depositional characteristics of any collapsed constructions. In addition, archaeoseismology is able to construct maps of past seismic activity concerning regions under surveillance and stretches the importance of the history of the relationship between humans and the environment. In the case of historical times (after the seventh–sixth century B.C.), seismic activity may be often traced more easily through the study of historical sources (Stiros and Jones, 1996; Stiros, 1996a, 2001, p. 547–549; Caputo, 2004; Bottari, 2005; Galadini et al., 2006; Marco, 2008). Western Crete offers an excellent field for archaeoseismological study, and it is the aim of this paper to review the results of this interdisciplinary approach in western Crete, with particular reference to Phalasarna and Kissamos. RECENT ARCHAEOSEISMOLOGICAL RESEARCH IN WESTERN CRETE Phalasarna Being one of the most important sites of western Crete, Phalasarna (Fig. 4) has offered—to date—very satisfactory results in terms of excavation stratigraphy and geomorphology, thanks to which a fruitful interdisciplinary scientific collaboration has produced, not only secure dating, but also a good reconstruction of the uplifted and silted harbor. Ammianus’s report in A.D. 365 and Spratt’s observations in 1851 were verified in the years soon after 1986, when Elpida Hadjidaki began excavations at Phalasarna (Fig. 5). Trenches in Spratt’s “square place, enclosed by walls and towers,” which now lies almost 100 m away from the sea front, provided very useful information about the depth of the harbor entrance and lagoon and its political and geologic history. Evidence for blockage of the entrance channel was found during the excavation of the channel trench (Figs. 5 and 6) and has been related to the Roman invasion of Crete, in 68 B.C. under the leadership of Quintus Cecilius Metellus Creticus. The harbor was vital to Phalasarna’s well-being, and its destruction eventually led to the city’s abandonment (Hadjidaki, 1988, 1990, 1992, 2001; Frost, 1989, 1997; Frost and Hadjidaki, 1990; Pirazzoli et al., 1992). A second trench sunk in the middle of the harbor basin (Figs. 5 and 7A) illuminated the gradual silting-up of the harbor after it was blocked. Successive layers of mud, marine shells (Fig. 7B), sand, and building ruins provide an extremely secure relative chronology for the site over the centuries and its geologic history (Hadjidaki, 1988; Pirazzoli et al., 1992). At this point, geology and environmental archaeology combined with traditional archaeology for a fuller and more secure interpretation of the finds and corrected Spratt’s inspiring original reconstruction of the harbor basin (Fig. 8). Seismic and geological research conducted by Pirazzoli and this team in the early 1990s helped to reconstruct the shape of the lagoon and the process of harbor elevation and silting in relation to human intervention and

Adriatic Sea Epidamnus

Aegean Sea Sicily Ionian Islands Peloponnese Cyprus Crete Sabartha Oea

Figure 1. Map of the Eastern Mediterranean clarifying all geographically relevant places used in the paper and situating Crete in context.

EASTERN MEDITERRANEAN SEA Lepcis Magna Alexandria

North Africa

Figure 2. Detailed map of western Crete with all geographically relevant places cited in the paper.

70

Stefanakis

Figure 3. Map of Phalasarna from Spratt (1865).

natural (seismic and tsunamigenic) events, over a span of eight centuries (fourth century B.C. to fourth century A.D.). According to Pirazzoli (Fig. 9), the silting of the harbor started gradually, shortly after the blocking of its entry, as the formation of a stratum lying 20 cm above today’s sea level consisting of deeper-water species Cerastoderma glaucum has been dated by radiocarbon method between 41 B.C. and A.D. 145 (Fig. 9B). Above this stratum, at +5.70 to +5.90 m above sea level, there lies a distinctive layer of terrestrial sediments, debris, and rounded blocks of stone, washed from the surrounding area of the city, showing the arrival of a tsunami (Fig. 9C). Some 20 cm higher, on top of the tsunami sediments, a layer containing Hydrobia acuta, dated between 54 B.C. and A.D. 137, renders identification of the origin of the stratum beneath from the A.D. 66 tsunami possible (Pirazzoli et al., 1992; DomineyHowes et al., 1998; Stiros and Papageorgiou, 2001; DomineyHowes, 2002). From +6.0 m up and for ~15 cm, marine sediments continue, denoting that seawater was still coming into the harbor (Fig. 9D). Prior to A.D. 169, however, the harbor entrance became completely blocked from the sea and was only occasionally breached by storm surges (Pirazzoli et al., 1992). A second tsunami has been blamed for the uppermost layer of angular limestone blocks and rubble within a silty clay layer in the stratigraphic section seen between +6.4 and +6.7 m (Fig. 9E) (Pirazzoli et al., 1992), although is effects were relatively limited, since no marine stuff is observed (Dawson, 1996; DomineyHowes et al., 1998; Pirazzoli, 1999; Stiros and Papageorgiou,

2001; Price et al., 2002), and it does not seem to have entered very far into the town. Seismic research and radiocarbon analysis date the incident to 1530 (±40) radiocarbon years, which calibrates to ca. A.D. 365 (Pirazzoli et al., 1992). This tsunami was therefore related to the seismic event responsible for raising the coast. Although the A.D. 365 tsunami was stronger than the A.D. 66 one, it does not seem to have affected Phalasarnas’ harbor as much. If the same seismotectonic movement that generated the tsunami had also uplifted western Crete, then perhaps the coast at Phalasarna had already been uplifted by 6.6 m (Fig. 10) when the tsunami hit (Pirazzoli et al., 1992; Pirazzoli, 1999). This may also account for the absence of evidence of A.D. 365 tsunami deposits along the western and southwest coast of Crete (Scheffers and Scheffers, 2007). To sum up, archaeological, seismic, and geological studies at Phalasarna, assisted by late antique chronicles have not only identified the seismic event of A.D. 365 on the ground, but they have shown that by the time of the great earthquake and the land uplift, the city’s infamous “closed” military port (Hadjidaki, 1990, 2001; Stefanakis, 2006b) had already been blocked, abandoned, and partly silted up. The A.D. 66 tsunami wave completed the destruction begun by the Romans more than a century earlier (68 B.C.). This wave swept anything in its path on land into the harbor basin during its retreat, contributing greatly to the siltingup of the harbor. Next, in A.D. 365, the harbor of Phalasarna was transformed into a piece of dry land due to an earthquake that raised the coast almost 9 m above sea level (Pirazzoli et al., 1992; Kelletat, 1998; Stiros and Papageorgiou, 2001; Price et al., 2002; see also Hadjidaki, 1988, 1990, 2001; Frost, 1997; Zouros et al, 2002; Stefanakis, 2006a, 2006b). Its associated tsunami deposited the last—but not much—settlement remains, along with other terrestrial sediments and rubble in the harbor basin. The A.D. 365 uplift of western Crete coast offers a unique opportunity for archaeological research in Phalasarna, since the partly artificial harbor channels, basin, and installations have become part of the land, revealing much useful information on harbor architecture and use and harbor defensive constructions (Sanders, 1982; Gondicas, 1988; Hadjidaki and Stefanakis, 2003; Stefanakis, 2006a). As for the architecture, however, exposure on the land definitely made it more accessible and easier to excavate by landlubbers. In fact, the uplift exposed the site to all kinds of land-based destruction, including being robbed out for building stone and burned up in limekilns. Kissamos At a distance of 5 km east of Phalasarna, in the bay of modern Kissamos, lies Kissamos, the port of ancient Polyrrhenia, which developed after Phalasarna’s destruction and flourished during the first centuries of the Roman conquest of Crete. Archaeological and seismic research in the area indicated that it had been heavily damaged by the A.D. 365 earthquake. Captain Spratt had originally observed an uplift of 5.5 m at the

Western Crete: From Captain Spratt to modern archaeoseismology

71

Figure 4. The site of Phalasarna. Reconstruction is from Hadjidaki and Stefanakis (2003) (courtesy of Kretiko Panorama). 1—military port; 2—fort; 3–4—temples; 5—fish tank; 6—industrial area; 7—south tower; 8—northwest tower; 9—northeast tower; 10—quarries.

harbor (Spratt, 1865, chapter XVIII), a short distance west of the city, at the site “Mavros Molos” (Pologiorgi, 1985; Gondicas, 1988; Stiros and Papageorgiou, 2001; Markoulaki et al., 2004). Indeed, the Kissamos coast was also uplifted (coastal uplift estimated to 6.5 m above sea level), and the harbor is now located quite a few meters inland (Fig. 11) (Davaras, 1967a; Flemming and Pirazzoli, 1981; Pirazzoli, 1999; Stiros and Papageorgiou, 2001; Papadimitriou and Karakostas, 2008). At the city of Kissamos itself, archaeological data from more than 50 rescue excavations over the last decades by the 25th Ephoreate of Prehistoric and Classical Antiquities have contributed to our knowledge about a destructive earthquake that abruptly ended a prosperous period of the city. Archaeological data include, among others, destruction layers, demolished houses, human corpses trapped under the ruins, as well as signs of destructive fire at some point shortly after A.D. 355–361 (Pologiorgi, 1985; Stiros, 2001; Stiros and Papageorgiou, 2001; Vlazaki-Andreadaki, 2002, 2004; Markoulaki, 2002, 2006). Geologic research indicates an earthquake of a minimum

seismic intensity of XI (MM scale) from an epicenter less than 100 km away. All this, considered together with the numerous copper coins found in the destruction layer and dating up to A.D. 355–361 (reign of the roman Emperor Constant II), makes the A.D. 365 earthquake the most probable cause of the catastrophe, which led to “nearly total physical destruction of the community” (Stiros and Papageorgiou, 2001, p. 387) since the victims failed to receive a proper burial (Stiros and Papageorgiou, 2001; Papadimitriou and Karakostas, 2008). Kissamos followed the fate of Phalasarna, with a coastal elevation of 5.5 m, which resulted in the uplift and exposure of its harbor installations. Unlike Phalasarna, Kissamos did not receive any tsunami impacts, for the town was safely built within the homonymous secure closed bay. Although it has not been verified for Phalasarna, Kissamos was seized to ground by the intense earthquake and never managed to recover. For her, as for the whole of western Crete, the A.D. 365 event signified the end of pagan era and the beginning of Christian era (Stiros and Papageorgiou, 2001; Markoulaki, 2006).

72

Stefanakis

Figure 5. Plan of the harbor area after Frost and Hadjidaki (1990). Arrows indicate the spots of two important test trenches (Figs. 6 and 7).

SEISMIC EVENT OF A.D. 365 AND WESTERN CRETE The powerful tectonic earthquake that took place during the night of 21 July A.D. 365 off the southwest coast of Crete, a result of the subduction of the African under the Aegean tectonic plate (Fig. 12), raised the west coast of the island 6–9 m above sea level (Thommeret et al., 1981; Jacques and Bousquet, 1984a, 1984b; Papazachos and Papazachos, 1989; Pirazzoli et al., 1992; Stiros, 1996b, 2009; Spyropoulos, 1997; Stiros and Papageorgiou, 2001; Stiros and Drakos, 2006) (Fig. 13). According to seismic stud-

ies on Crete and Antikythera, the earthquake took place as the result of 20 m slip on a fault ~100 km long, and its epicenter was between the SW edge of Crete and the Hellenic trench (Pirazzoli et al., 1992; Kelletat, 1998; Stiros and Papageorgiou, 2001; Shaw et al., 2008). Evaluations of the event estimate the intensity of the earthquake at M >8.5. Its focal depth was between 40 and 70 km and caused a tsunami of unknown magnitude, but of an intensity reaching 5 (in Ambrasey’s 1962 scale) and a surfacewave magnitude reaching 8 (Papazachos and Papazachos, 1989; Papazachos and Dimitriou, 1991; Papadopoulos, 2001; Stiros

Western Crete: From Captain Spratt to modern archaeoseismology

Figure 6. Drawing of the channel trench, where eight huge stone blocks (A–Θ) were revealed during the excavation in 1987 (from Hadjidaki, 1988).

A

73

the southern Peloponnese, west and south Crete, and Alexandria in the Nile Delta. Destruction probably from the same event is recorded in North Africa at Leptis Magna, Oea, and Sabartha (Papazachos and Dimitriou, 1991; Pirazzoli et al., 1992; Guidoboni et al., 1994; Kelletat, 1998; Stiros, 2001, 2009; Stiros and Papageorgiou, 2001; Price et al., 2002; Dominey-Howes, 2002; Kelly, 2004; Stiros and Drakos, 2006; Papadimitriou and Karakostas, 2008). The absolute date of the uplift, however, is still controversial (Stiros and Drakos, 2006). A reassessment of radiocarbon dates on material from the solution notches on the Sphakia shoreline, based on the Bayesian approach and new calibration data, produced a date between A.D. 480 and 550 for the land uplift, suggesting that the major uplift has not been detected in archaeological record (Price et al., 2002). A more recent study however, by Shaw et al. (2008), reassessing radiocarbon dates based on material from present sea level and the uplifted paleoshoreline, came to reinforce the earlier date of A.D. 365 for western Crete’s uplift and tsunami. According to the study and the new radiocarbon dates, the A.D. 365 event had a magnitude of Mw 8.3–8.5, and the paleoshoreline “was lifted close to its present position within a few decades of the AD 365 earthquake” (Shaw et al., 2008, p. 273). ARCHAEOSEISMOLOGY AND WESTERN CRETE: POTENTIAL RESEARCH

B

Figure 7. The harbor trench. (A) Stratigraphy of the two tsunami deposits (courtesy of Kretiko Panorama). (B) A.D. 66 tsunami layer, detail (courtesy of Kretiko Panorama).

and Papageorgiou, 2001; Stiros and Drakos, 2006; Shaw et al., 2008). It was the biggest tsunami reported in and near ancient Greece, claiming thousands of lives and causing widespread devastation in various parts of eastern Mediterranean (Fig. 1), including Sicily, the Ionian Islands, Epidamnus in the Adriatic,

Even if controversy over the dating of the seismic event persists, it still remains that western Crete offers a unique opportunity for archaeoseismological study. The coastal uplift in late antiquity is a fact, while tsunami impact and seismic destruction have been identified at a number of archaeological sites (Fig. 14) (Stiros et al., 2004). Costal uplifts attributed to the A.D. 365 earthquake (Shaw et al., 2008) have been observed at the south coast: Tarrha, modern Hagia Roumeli (Weinberg, 1960; Price et al., 2002; The Sphakia Survey, 2010), has been uplifted by 6 m (Price et al., 2002), Poikilasion (Price et al., 2002; The Sphakia Survey, 2010) by 7 m (Pirazzoli, 1999; Price et al., 2002), and Phoenix, modern Loutro (Price et al., 2002; The Sphakia Survey, 2010), by 3.5–4 m (Price et al., 2002). All three sites have already been thoroughly surveyed (Moody et al., 1998; The Sphakia Survey, 2010) and studied from the seismic and geological point of view, although the original studies did not attribute the coastal uplift to the A.D. 365 earthquake but to a later event between A.D. 405 and A.D. 615 (Price et al., 2002). Uplift, known since Spratt’s time and verified by later scientific observation, can be also seen at the coastal cities: Kydonia, uplift of 2 m (Stiros and Papageorgiou, 2001); Inachorion, uplift of 8 m at the nearby sites Ormos Stomiou and Mavros Bay (Scheffers and Scheffers, 2007); Kalamyde, modern Palaiochora, port of ancient Kandanos, uplift of 7–8 m (Scheffers and Scheffers, 2007), Lissos, uplift of 7 m (Pirazzoli, 1999); and Syia, modern Sougia (Pirazzoli, 1999), uplift of 6.6 m.

Figure 8. Suggested section plan of the harbor basin in relation to ancient and modern sea level from Spratt (1865).

A

B

C

D

E

F

Figure 9. Model to explain the stratigraphy in the Phalasarna harbor basin in relation to tectonic movement and gradual harbor silting from late classical period to modern era, based on simplified stratigraphy of the harbor trench (after Pirazzoli et al., 1992). The A–F labels represent stages of MSL (mean sea level) in relation to the harbor basin, from the late fourth century B.C. to present. F—freshwater deposits; T—tsunami deposits; C—confined marine deposits; M—marine deposits.

Western Crete: From Captain Spratt to modern archaeoseismology

Figure 10. Traces of ancient sea line southeast of the harbor entrance at Phalasarna coast (courtesy E. Hadjidaki).

Remains of uplifted ancient jetty

Figure 11. Mavros Molos, ancient harbor of Kissamos. Uplifted jetty is made of unhewn blocks (from Davaras, 1967a).

75

Tsunami impact evidence (deposits) has been observed on the whole southwestern shore of Crete, from Palaiochora (ancient Kalamyde) to Gramvousa Island, north of Phalasarna. Among other sites, stratigraphic evidence of tsunami impacts is observed in the area of ancient Kalamyde, Inachorion, and Phalasarna (Scheffers and Scheffers, 2007), while the area of ancient Viennos is also worth testing. It is very interesting that although geomorphological studies in western Crete have produced enough proof of tsunami impacts from the late Holocene to the late antiquity, they have produced no absolute data or any evidence, so far, of fourth century and the A.D. 365 tsunami deposits (DomineyHowes, 2002; Stiros and Drakos, 2006; Scheffers and Scheffers, 2007; Stiros, 2009). More evidence of the destructive earthquakes of A.D. 66 and A.D. 365 probably awaits discovery at other inland archaeological sites in western Crete. A prime candidate is the big city of Polyrrhenia, in the mountains east of Phalasarna and south of Kissamos (Theophaneides, 1948; Davaras, 1967b; Sanders, 1982; Gondicas, 1988; Markoulaki, 1992), which was possibly damaged by the A.D. 66 earthquake and eventually relocated to Kissamos (Pologiorgi, 1985; Stiros and Papageorgiou, 2001). The prosperous city of Aptera (Drerup, 1951; Sanders, 1982; Niniou-Kindeli, 2006), overlooking Souda Bay, where Spratt (1865, chapter XI) observed an elevation of ~2 m, may also bear traces of this earthquake. A destroyed Roman villa of the first century B.C.–A.D. first century, in sector VI, for example, is described by the excavator as giving “…the impression as having collapsed after a strong earthquake” (Niniou-Kindeli, 1999a, p. 169; see also Niniou-Kindeli, 1994–1996, 1999b, 2003).

Figure 12. Crete situated in the Mediterranean tectonic context (after http://commons.wikimedia.org/wiki/File:Tectonic _map_Mediterranean_EN.svg).

76

Stefanakis

Figure 13. Western Crete elevation (m) shown by uplift curves (Papadimitriou and Karakostas, 2008).

Figure 14. Sites of archaeological interest of western and central Crete affected by the land elevation or by the A.D. 365 earthquake in general. 1—Aptera; 2—Kydonia; 3—Kissamos; 4—Polyrrhenia; 5—Phalasarna; 6—Inachorion; 7—Viennos; 8—Kandanos; 9—Kalamyde; 10—Hyrtakina; 11—Lissos; 12—Elyros; 13—Syia; 14—Poekilassion; 15—Tarrha; 16—Phoenix; 17—Diktynnaion; 18—Eleutherna; 19—Knossos; 20—Gortyna. (Map with uplift curves is after Pirazzoli et al., 1992.)

Many inland cities in southwestern Crete such as Kantanos (Sanders, 1982; Gondicas, 1988; Stefanakis, 2000), Hyrtakina (Sanders, 1982), and Elyros (Sanders, 1982) must have also been damaged by the destructive earthquakes. Solution notches and seismic damage should also be looked for near the ancient sites of the Diktynnaion Temple at Spatha promontory (Welter and Jantzen, 1951; Gondicas, 1988; Markoulaki, 2000; Markoulaki and Martinez, 2000–2001), Inachorion

(Gondicas, 1988) and Viennos (Gondicas, 1988) on the west coast, and Kalamyde (Hood, 1967; Gondicas, 1988) on the south. The list of the aforementioned sites of western Crete is, however, selective, and traces of the seismic event of A.D. 365 may exist in many more archaeological sites of western Crete (Hood, 1967; Sanders, 1982; Gondicas, 1988; Faure, 1989; Andreadaki-Vlazaki, 1997; Faraklas et al., 1998; The Sphakia Survey, 2010).

Western Crete: From Captain Spratt to modern archaeoseismology At the same time, research should also extend to central Crete, since Eleutherna, in the western foothills of Mount Ida, Knossos on the north coast, and Gortyna in south-central Crete (Fig. 14) also have evidence for a destructive earthquake in the second half of the fourth century A.D. (Stiros, 2001, 2009; see also Themelis, 1988; Sidiropoulos, 2004; Guidoboni et al., 1994), which could be associated with the A.D. 365 event. So many earthquake-affected sites on Crete in the fourth century A.D. give credence to Athanasius of Alexandria’s claim that in A.D. 365, more than 100 Cretan cities were destroyed by an earthquake of unprecedented magnitude, followed by a tsunami (Migne, 1857, v. 25, p. ccx). Evidence exists, but recovering it is not simply a matter of observation. Systematic archaeological survey and excavation are needed to reveal stratigraphic sections for chronological, geologic, and seismic research. For many sites, however, such research may be too late. For example, in the early 1850s, Captain Spratt produced a map of ancient Syia showing an elevated harbor (coast uplifted by 6.70 m) to the west of the ancient settlement (Fig. 15A). Today, however, the modern village of Sougia has gradually built up over the harbor basin (Fig. 15B), obscuring the archaeological record. It is important to explore and record these sites before more are buried by modern development. ACKNOWLEDGMENTS The author wishes to thank Elpida Hadjidaki, director of the excavations at Phalasarna, for her kind permission to present data from the excavations at the site, plans, and photographs, as well as Giorgos Patroudakis, publisher of Kretiko Panorama, for his permission to reproduce photographs. Thanks are also addressed to Manuel Sintubin for the invitation and encouragement to participate in the International Geo-

A

77

science Programme (IGCP) 567 on Earthquake Archaeology and to Stathis Stiros and Jennifer Moody for saving the manuscript from many mistakes. This article is a contribution to the United Nations Educational, Scientific and Cultural Organization– funded IGCP 567, “Earthquake Archaeology: Archaeoseismology along the Alpine-Himalayan Seismic Zone.” REFERENCES CITED Andreadaki-Vlazaki, M., 1997, The County of Chania through Its Monuments: Athens: Archaeological Receipts Fund, 75 p. Bottari, C., 2005, Ancient constructions as markers of tectonics deformation and strong seismic motions: Pure and Applied Geophysics, v. 162, p. 761– 765, doi: 10.1007/s00024-004-2639-6. Caputo, R., 2004, Historical seismology, archaeoseismology and palaeoseismology: Three distinct approaches to a natural phenomenon, in Archaeoseismology at the Beginning of the 21st Century (Atlas Conferences): Rome. Davaras, K., 1967a, Kastelli Kissamou: Archaeologiko Deltion, v. 22, part B2, Chronicles, p. 498–499 [in Greek]. Davaras, K., 1967b, Polyrrhenia: Archaeologiko Deltion, v. 22, part B2, Chronicles, p. 499 [in Greek]. Dawson, A.G., 1996, The geological significance of tsunamis: Zeitschrift für Geomorphologie N.F., supplement, v. 102, p. 199–210. Dominey-Howes, D.T.M., 2002, Documentary and geological records of tsunamis in the Aegean Sea region of Greece and their potential value to risk assessment and disaster management: Natural Hazards, v. 25, p. 195–224, doi: 10.1023/A:1014808804611. Dominey-Howes, D.T.M., Dawson, A.G., and Smith, D.E., 1998, Late Holocene coastal tectonics at Phalasarna, western Crete; a sedimentary study, in Stewart, I.A., and Vita-Finzi, C., eds., Coastal Tectonics: Geological Society of London Special Publication 146, p. 343–352. Drerup, H., 1951, Paläokastro-Aptara. Bericht über eine Untersuchung und Vermessung des Stadtgebietes, in Matz, F., ed., Forschungen auf Kreta, 1942: Berlin, W. de Gruyter, p. 89–98. Faraklas, N., Kataki, E., Kossyva, A., Xifaras, N., Panagiotopoulos, E., Tassoulas, G., Tsatsaki, N., and Chatzipanagioti, M., 1998, The Territories of the Ancient Cities of Crete: Rithymna 6: Rethymno, University of Crete, 242 p. [in Greek]. Faure, P., 1989, Cités antiques de Crète de l’oust: Cretan Studies, v. 1, p. 81–96. Flemming, N., and Pirazzoli, P., 1981, Archéologie des côtes de la Crète, in Ports et Villes Engloutis, Dossiers d’Archéologie, v. 50, p. 66–81.

B

Figure 15. (A) Map of Syia from Spratt (1865). (B) Aerial view of modern Sougia.

78

Stefanakis

Frost, F.J., 1989, The last days of Phalasarna: Ancient History Bulletin, v. 3, p. 15–17. Frost, F.J., 1997, Tectonics and history at Phalasarna, in Swiny, S., Hohlefelder, R.L., and Swiny, H.W., eds., Res Maritimae: Cyprus and the Eastern Mediterranean from Prehistory to Late Antiquity: Atlanta, American School of Oriental Research, p. 107–115. Frost, F.J., and Hadjidaki, E., 1990, Excavations at the Harbour of Phalasarna in Crete: Hesperia, v. 59, p. 513–527, doi: 10.2307/148300. Galadini, F., Hinzen, K.-G., and Stiros, S.C., eds., 2006, Archaeoseismology: Methodological issues and procedure: Archaeoseismology at the Beginning of the 21st Century (Atlas Conferences): Journal of Seismology, v. 10, no. 4, p. 395–414, doi: 10.1007/s10950-006-9027-x. Gondicas, D.G., 1988, Recherches sur la Crète Occidentale: Amsterdam, Adolf M. Hakkert, 365 p. Guidoboni, E., Comastri, A., and Traina, G., 1994, Catalogue of Ancient Earthquakes in the Mediterranean up to the 10th Century: Rome, Istituto Nazionale di Geofisica, 504 p. Hadjidaki, E., 1988, Preliminary report of excavation at the harbour of Phalasarna in west Crete: American Journal of Archaeology, v. 92, p. 463–479, doi: 10.2307/505244. Hadjidaki, E., 1990, Excavations at the classical and Hellenistic harbor at Phalasarna, west Crete, Greece: Acts of the 6th International Cretological Congress, v. A1, Archaeological Section: Chania, Philologikos Syllogos “O Chryssostomos,” p. 355–361. Hadjidaki, E., 1992, Phalasarna: Archeologikon Deltion, Xρονικά, v. 42 (1987), part B2, Chronicles, p. 566–567 [in Greek]. Hadjidaki, E., 2001, The Roman destruction of Phalasarna, in Higham, N., ed., Archaeology of the Roman Empire; a Tribute to the Life and Works of Professor Barri Jones: British Archaeological Reports International Series, v. 940, p. 155–166. Hadjidaki, E., and Stefanakis, M.I., 2003, The secrets of Phalasarna: Kretiko Panorama, v. 2, p. 10–135. Hood, M.S.F., 1967, Some ancient sites in southwest Crete: Annual of the British School of Athens, v. 62, p. 47–56. Jacques, F., and Bousquet, B., 1984a, Le raz de marée du 21 juillet 365—Du cataclysme local à la catastrophe cosmique: Mélanges de l’École Française de Rome, v. 96, p. 423–461. Jacques, F., and Bousquet, B., 1984b, Le cataclysme du 21 juillet 365: Phénomène régional ou catastrophe cosmique?, in Helly, B., and Pollino, A., eds., Tremblements de Terre. Histoire et Archéologie: IVèmes Rencontres Internationales d’Archéologie et d’Histoire d’Antibes: Valbonne, Editiones ADPCA, p. 183–193. Kelletat, D., 1998, Geologische Belege katastrophaler Erdkrustenbewegungen 365 AD im Raum von Kreta, in Olshausen, E., and Sonnabend, H., eds., Stuttgarter Kolloquium zur historischen Geographie des Altertums 6, 1996 Naturkatastrophen in der antiken Welt: Geographica Historica 10: Stuttgart, Frank Steiver Verlag, p. 156–161. Kelly, G., 2004, Ammianus and the Great Tsunami: Journal of Roman Studies, v. 94, p. 141–165, doi: 10.2307/4135013. Marco, S., 2008, Recognition of earthquake-related damage in archaeological sites: Examples from the Dead Sea fault zone: Tectonophysics, v. 453, p. 148–156, doi: 10.1016/j.tecto.2007.04.011. Markoulaki, St., 1992, Polyrrhenia: Archaeologiko Deltion, Chronicles, v. 42 (1987), part B2, Chronicles, 563 p. [in Greek]. Markoulaki, St., 2000, Stele Telephou: Actes of the 8th International Cretological Congress, v. A2, p. 239–257 [in Greek]. Markoulaki, St., 2002, Kentro Hygias: Kretike Estia, v. 9, p. 270–271. Markoulaki, St., 2006, Kissamos, in Andreadaki-Vlazaki, M., and NiniouKindeli, V., eds., Ancient Sites and Monuments. The Chania Prefecture: Chania, KE Ephoreate of Prehistoric and Classical Antiquities, p. 22–23 [in Greek]. Markoulaki, St., and Martinez, A.F., 2000–2001, Psefisma proxenias apo tin Kissamo: Kretike Estia, v. 8, p. 147–158. Markoulaki, St., Christodoulakos, G., and Fragkonikolaki, C., 2004, I archaia Kissamos kai I poleodomike tes organose: Creta Romana e Protobyzantina II: Padova, Bottega d’Erasmo A. Ausilio, p. 355–374 [in Greek]. Migne, J.P., 1857, Patrologiae cursus completus: Omnium SS. patrum, doctorum scriptorumque ecclesiasticorum; sive latinorum, sive graecorum, in Diotis, J., ed.: vols. 25–28 (Athanasius of Alexandria), Helleniki Patrologia (Patrologia Graeca): Athens, Centre for Patrological Editions.

Moody, J., Nixon, L., Price, S., and Rackham, O., 1998, Surveying poleis and larger sites in Sphakia, in Cavanagh, W.G., and Curtis, M., eds., PostMinoan Crete: Proceedings of the Colloquium organised by the British School at Athens and the Institute of Archaeology, University of London, November 1995: British School at Athens Studies Series 2, p. 87–95. Niniou-Kindeli, V., 1994–1996, Aptera: Kretike Estia, v. 5, p. 210–212 [in Greek]. Niniou-Kindeli, V., 1999a, Aptera (Aptara): Kretike Estia, v. 7, p. 167–175. Niniou-Kindeli, V., 1999b, Aptera: Archaeologiko Deltion, v. 49 (1994), part B2, Chronicles, p. 721. Niniou-Kindeli, V., 2003, Aptera: Archaeologiko Deltion, v. 52 (1997), part B3, Chronicles, p. 1017–1019. Niniou-Kindeli, V., 2006, Aptera, in Andreadaki-Vlazaki, M., and NiniouKindeli, V., eds., Ancient Sites and Monuments. The Chania Perfecture: Chania, KE Ephoreate of Prehistoric and Classical Antiquities, p. 8–9 [in Greek]. Papadimitriou, E., and Karakostas, V.G., 2008, Rupture model of the great A.D. 365 Crete earthquake in the southwestern part of the Hellenic Arc: Acta Geophysica, v. 56, no. 2, p. 293–312. Papadopoulos, G.A., 2001, Tsunamis in the East Mediterranean: A catalogue for the area of Greece and adjacent seas, in Proceedings of the Intergovernmental Oceanographic Commission/International Union of Geodesy and Geophysics International Workshop Tsunami Risk Assessment Beyond 2000 Theory, Practice and Plans: In Memory of Professor S.L. Soloviev, Moscow, 14–16 June 2000: Moscow, p. 34–43. Papazachos, B., and Dimitriou, P.P., 1991, Tsunamis in and near Greece and their relation to the earthquake focal mechanisms: Natural Hazards, v. 4, p. 161–170, doi: 10.1007/BF00162785. Papazachos, B., and Papazachos, K., 1989, The Earthquakes in Greece: Thessaloniki [in Greek], 356 p. Pirazzoli, P.A., 1999, Les ports antiques soulevés de la Méditeranée orientale, in Rosselló, V.M., ed., Geoarqueologia I Quartenari Litoral, Memorial Maria Pilar Fumanal: Valencia, Valencia University, p. 391–401. Pirazzoli, P.A., Ausseil-Badie, J., Giresse, P., Hadjidaki, E., and Arnold, M., 1992, Historical environmental changes at Phalasarna Harbour, west Crete: Geoarchaeology, v. 7, no. 4, p. 371–392, doi: 10.1002/gea.3340070406. Pologiorgi, M., 1985, Kissamos; the topography of an ancient polis: Archaeologika Analekta ex Athenon, v. XVIII, p. 65–79 [in Greek]. Price, S., Higham, T., Nixon, L., and Moody, J., 2002, Relative sea-level changes in Crete: Reassessment of radiocarbon dates from Sphakia and west Crete: Annual of the British School at Athens, v. 97, p. 171–200. Psillakis, M., and Psillakis, N., 2007, T.B.A. Spratt. Taxidia kai Erevnes stin Kriti tou 1850, v. 2: Herklion, Karmanor Publications, 408 p. [in Greek]. Richards, G.H., 1888, Obituary: Vice-Admiral Thomas A.B. Spratt, C.B., F.R.S.: Proceedings of the Royal Geographical Society and Monthly Record of Geography, New Monthly Series, v. 10, no. 4 (Apr.), p. 242–244. Sanders, I.F., 1982, Roman Crete. An Archaeological Survey and Gazetteer of Late Hellenistic, Roman and Early Byzantine Crete: Warminster, Aris and Phillips, 185 p. Scheffers, A., and Scheffers, S., 2007, Tsunami deposits on the coastline of west Crete (Greece): Earth and Planetary Science Letters, v. 259, p. 613– 624, doi: 10.1016/j.epsl.2007.05.041. Shaw, B., Ambraseys, N.N., England, P.C., Floyd, M.A., Gorman, G.J., Higham, T.F.G., Jackson, J.A., Nocquet, J.-M., Pain, C.C., and Piggott, M.D., 2008, Eastern Mediterranean tectonics and tsunami hazard inferred from the AD 365 earthquake: Nature Geoscience, v. 1, p. 268–276 (published online: 9 March 2008; doi: 10.1038/ngeo151). Sidiropoulos, K., 2004, Numismatic history of Roman and Protobyzantine Crete (67 BC–AD 827). Testimonia et Desiderata, in Livadiotti, M., and Simiakaki, I., eds., Creta Romana e Protobyzantina: Padova, Bottega d’Erasmo A. Ausilio. The Sphakia Survey, 2010, The Sphakia Survey, Internet edition: http:/sphakia .classics.ox.ac.uk (accessed June 2010). Spratt, C.T.A.B., 1865, Travels and Researches in Crete, Volume II: London, John van Voorst MDCCCLXV(=1865), 327 p. Spratt, C.T.A.B., and Leake, C., 1854, Extract of a Letter from Captain Spratt, R.N., on Crete: Journal of the Royal Geographical Society of London, v. 24, p. 238–239, doi: 10.2307/3698110. Spyropoulos, P.I., 1997, The Chronicle of Earthquakes in Greece: Athens, Dodone, 453 p. [in Greek].

Western Crete: From Captain Spratt to modern archaeoseismology Stefanakis, M.I., 2000, Polyrrhenia, Oreioi and Kandanos. A relationship of the second half of the third century BC: Actes of the 8th International Cretological Congress, Herakleion, v. A3, p. 249–261 [in Greek]. Stefanakis, M.I., 2006a, Natural catastrophes in the Greek and Roman world: Curse or blessing? Four cases of earthquake-generated tsunamis: Mediterranean Archaeology and Archaeometry Journal, v. 6.1, p. 61–88. Stefanakis, M.I., 2006b, Phalasarna: Un port antique, un espace d’échanges en Méditerranée, in Clément, F., Tolan, J., and Wilgaux, J., eds., Espaces d’Échanges en Méditerranée. Antiquité et Moyen-Age: Lyon, PU Rennes, p. 41–75. Stiros, S., 1996a, Identification of earthquakes from archaeological data: Methodology, criteria and limitations, in Stiros, S., and Jones, R., eds., Archaeoseismology: British School at Athens, Fitch Laboratory Occasional Paper 7, p. 129–152. Stiros, S., 1996b, Late Holocene relative sea level changes in SW Crete: Evidence of an unusual earthquake cycle: Annali di Geofisica, v. XXXIX, no. 3, p. 677–687. Stiros, S., 2001, The AD 365 Crete earthquake and possible seismic clustering during the 4–6th centuries AD in the Eastern Mediterranean: A review of historical and archaeological data: Journal of Structural Geology, v. 23, p. 545–562, doi: 10.1016/S0191-8141(00)00118-8. Stiros, S.C., 2009, The 8.5+ magnitude, AD 365 earthquake in Crete: Coastal uplift, topography changes, archaeological and historical signature: Quaternary International, doi: 10.1016/j.quaint.2009.05.005. Stiros, S., and Drakos, A., 2006, A fault-model for the tsunami-associated, magnitude ≥8.5 Eastern Mediterranean, AD 365 earthquake: Zeitschrift für Geomorphologie, supplement, v. 146, p. 125–137. Stiros, S., and Jones, R.E., eds., 1996, Archaeoseismology: Oxford, UK, Institute of Geology and Mineral Exploration/British School at Athens, 268 p.

79

Stiros, S.C., and Papageorgiou, S., 2001, Seismicity of western Crete and the destruction of the town of Kissamos at AD 365: Archaeological evidence: Journal of Seismology, v. 5, p. 381–397, doi: 10.1023/A:1011475610236. Stiros, S., Papageorgiou, S., and Markoulaki, S., 2004, The destruction of Cretan towns in AD 365, in Livadioti, M., and Simiakaki, I., eds., Creta Romana e Protobyzantina: Atti del Congresso Internazionale (Iraklion, 23–30 Settembre 2000): Padova, Bottega d’Erasmo A. Ausilio, p. 193– 223 [in Greek]. Themelis, P., 1988, Eleutherna: Kritiki Estia, v. 2, p. 298–302. Theophaneides, V., 1948, Excavational research and finds from western Crete. The province of Kissamos. (B) Excavations: Archaeologiki Ephimeris 1942–44, part B, Chronicles, p. 17–31 [in Greek]. Thommeret, Y., Laborel, J., Montaggioni, L., and Pirazzoli, P., 1981, Late Holocene shoreline changes and seismotectonic displacements in western Crete (Greece): Zeitschrift für Geomorphologie, supplement, v. 40, p. 127–149. Vlazaki-Andreadaki, M., 2002, Kissamos: Kretike Estia, v. 9, p. 266–271. Vlazaki-Andreadaki, M., 2004, Kissamos: Archaeologiko Deltion, v. 53 (1988), part B3, Chronicles, p. 864–868. Weinberg, G.D., 1960, Excavations at Tarrha 1959: Hesperia, v. 29, p. 90–108, doi: 10.2307/147333. Welter, G., and Jantzen, U., 1951, Das Diktynnaion, in Matz, F., ed., Forschungen auf Kreta, 1942: Berlin, W. de Gruyter, p. 106–117. Zouros, N., Velitzelos, E., Mountrakis, D.M., and Soulakelis, N., eds., 2002, Atlas of the Geological Monuments of the Aegean: Athens, Ministry of the Aegean, 350 p. [in Greek]. MANUSCRIPT ACCEPTED BY THE SOCIETY 19 MAY 2010

Printed in the USA

The Geological Society of America Special Paper 471 2010

Earthquake archaeology in Japan: An overview Gina L. Barnes School of Oriental and African Studies (SOAS), University of London, Thornhaugh Street, Russell Square, London WC1H 0XG, UK

ABSTRACT Earthquake archaeology developed in Japan simultaneously with that in the Mediterranean in the mid-1980s. By 1996, evidence of earthquake occurrence had been documented at 378 sites throughout the archipelago. The main features identified include various results of liquefaction, faults, landslips, and surface cracking. This evidence differs greatly from the standard Mediterranean focus on building damage, and the reasons for the very different natures of archaeoseismology in these world regions are explained herein. This article recounts the development of this new subfield, inspired by the interest of geomorphologist Sangawa Akira and taken to its most recent advances in identifying soft-sediment deformation structures by geoarchaeologist Matsuda Jun-ichirō. The evidence of earthquake activity at archaeological sites can be matched with earthquakes caused by either active fault movement or subduction. The historical record of earthquake occurrence, already documented back to 599 C.E., is extended into the prehistorical record through relative dating of artifacts and features on archaeological sites. Both the identification and the dating of the archaeological evidence of earthquakes can be challenged, though the “territorial approach” gives the data a significance that is not achieved through analysis of single sites.

THE NEW SUBDISCIPLINE

Sangawa to develop the idea that even human-made constructions in Japan can give evidence of past earthquake activity. This may sound odd to the Mediterranean archaeologists, who deal with earthquake damage to buildings and site destruction layers as a matter of course (cf. Galadini et al., 2006), but the novelty of this idea in Japan is illustrative of the difference between earthquake archaeology in Japan and elsewhere—a difference discussed in detail herein. Sangawa’s ideas were introduced to the archaeological community in 1987 through the Kodaigaku Kenkyūkai (Paleology Research Group) at Dōshisha University in Kyoto (Sangawa, 1988). Primarily through this publication, archaeologists became aware that past earthquake activity can be seen in excavations and that many unexplained features in some sites were, in fact, the unrecognized effects of earthquakes. Such discoveries are

The 1995 Kobe earthquake (officially called the HanshinAwaji Dai-shinsai in Japanese) brought to the fore a developing subdiscipline of Japanese archaeology focused on identifying the effects of earthquakes in archaeological sites. Named jishin kōkogaku (“earthquake archaeology”) in the mid-1980s, the field has been developed in Japan primarily through the initiative of one geomorphologist with the Japan Geological Survey, Sangawa Akira, who participated in the compilation of the national survey of active faults and the rendering of these onto 1:50,000 scale maps (RGAFJ, 1992). A chance encounter in his student days with the fifth-century Kondayama mounded tomb (see periodization in Table 1), which shows evidence of dike slippage along a fault line (Fig. 1), led

Barnes, G.L., 2010, Earthquake archaeology in Japan: An overview, in Sintubin, M., Stewart, I.S., Niemi, T.M., and Altunel, E., eds., Ancient Earthquakes: Geological Society of America Special Paper 471, p. 81–96, doi: 10.1130/2010.2471(08). For permission to copy, contact [email protected]. © 2010 The Geological Society of America. All rights reserved.

81

82

Barnes TABLE 1. CHRONOLOGY OF JAPANESE HISTORY Dates Period Character Ca. 15,000–800 B.C.E. Jōmon Hunting, gathering, fishing, horticulture 800 B.C.E.–C.E. 250 Yayoi Agriculture, bronze, iron 250–710 Kofun Mounded tomb culture, state formation 710–794 Nara Early court culture, archaic state 794–1185 Heian Late court culture, state decentralization 1185–1603 Medieval Warrior-dominated culture, rise of samurai 1603–1868 Edo Tokugawa family rule, feudal administration 1868–present Modern Westernization, industrialization, militarism

Figure 1. The Kondayama Tomb (K) situated in a late fourth- to fifthcentury cluster of keyhole-shaped mounded tombs of early Japanese ruling elites, located on a Pleistocene terrace in southeastern Osaka. Some tombs are moated with additional greenbelts or dikes, which exhibited displacement along the Konda fault. (Courtesy of A. Sangawa, redrawn by Durham Archaeological Services.)

continuously being published in the journal Kodaigaku Kenkyū, and Sangawa has expanded upon his original views in 1995, 2001, and 2007, in addition to many individual articles—most recently a personalized excursion through the progress of his earthquake archaeology research (Sangawa, 2009). As Sangawa was formulating his approach to archaeological data in Japan, the field of earthquake archaeology was developing overseas, particularly in the Mediterranean (Rapp, 1986, p. 365). Archaeological sites with building damage and destruction layers were well known, but it was not until the late 1980s that a systematic approach to the data began to be constructed. In 1991, an international conference in Athens “brought together around a hundred specialists from several countries ... to exchange ideas and discuss the problems of identification and study of ancient earthquakes from the complementary standpoints of their social, cultural, historical and physical effect” (Stiros and Jones, 1996, p. 1). Many of these conference papers were published in the volume entitled Archaeoseismology (Stiros and Jones, 1996). In Japan, it was the 1995 Kobe earthquake that galvanized archaeological action on a multifaceted front. Young archaeologists in the Kinai region around Osaka, Nara, and Kyoto joined together to form the “Disaster Concerned Archaeologists’ Network” (Maibun Kankei Kyūen Renraku Kaigi; see, e.g., DCAN, 1996) to assist with communications related to cultural properties damaged in the earthquake. A newsletter edited by Okamura Katsuyuki conveyed important findings, and one of the network’s outstanding contributions to the scholarship of earthquake archaeology was the joint sponsorship of a symposium together with the “Buried Cultural Properties Research Group” (Maizō Bunkazai Kenkyūkai) on evidence of earthquakes in the archaeological record. Over 150 scholars from around Japan gathered near Kobe just 19 mo. after the earthquake to present evidence on 378 sites throughout Japan assessed as having sustained earthquake damage. Their published report runs to 826 pages and contains a master list of the sites covered (DCAN, 1996). Their purpose in this

Earthquake archaeology in Japan compilation, however, was more than academic: they saw that if the cycle of earthquakes could be established through time in different regions, they could help contribute to the future of society. The Athens conference discussed and defined the terms seismic archaeology, archaeological seismicity, earthquake geology, and paleoseismology. It seems that archaeoseismology has been generally adopted as the name of the field since it interfaces with archaeology. Sangawa preferred the term seismic archaeology, but for the purposes of this article I have chosen to translate jishin kōkogaku directly as earthquake archaeology (as Sangawa also does now) and will use this term only with reference to Japan to avoid confusion. In Japan, earthquake archaeology is considered a branch of archaeology, which itself is considered a branch of history. Its sister discipline, historical seismology (Ishibashi, 2004), has concentrated on the documentary evidence for earthquakes in Japan, beginning with the earliest chronicles of 712 C.E. However, the editors of the most recent compilation of papers in the special issue of the Journal of Seismology (Galadini et al., 2006) treat archaeoseismology as a developing branch of seismology, emphasizing the refinement of quantitative seismological analyses. The Stiros and Jones volume (1996) stressed that this new subdiscipline, whatever it is called, needs full interdisciplinary status to accomplish the goals of both neotectonic geology and archaeology in terms of identification and documentation of past seismic events, reconstruction of seismic cycles, assessment of earthquake damage on cultural properties, and methods of prevention and conservation in earthquake-prone areas. Caputo and Helly (2008) have contrasted several of these new subdisciplines, but not always to the best informative level, as we shall see in the following discussion. Earthquake archaeology in Japan accomplishes only a few of Stiros and Jones’ objectives. It is a rather narrowly defined field, confined to what is discovered at archaeological sites rather than including damage to cultural properties in general. However, given the reduced rate of excavation in today’s economic retrenchment, publicly employed archaeologists (more than 7000 at the peak time in the late 1980s) are being asked to address the built environment as well, assessing extant cultural properties. Still, there is a split between what is seen as earthquake damage to standing buildings and earthquake activity in the archaeological record. For reasons detailed later herein, building damage is not a major concern in Japanese earthquake archaeology, unlike archaeoseismology in other parts of the world (Table 2). The types of earthquake evidence at archaeological sites are totally different in the Mediterranean than in Japan: in the former area, buildings and layers of destroyed cultural remains (destruction layers) are of concern, while in the latter, sediment deformation is the focus. Around 75% of the sites in the DCAN volume (1996) bore evidence of liquefaction, while fissures and faults were reported at far lower levels of incidence. This chapter will explain why buildings per se are not a prominent component of Japanese earthquake archaeology and how sediment deformation, in its more geological sense, is used instead to glean information about past earthquakes.

83

BUILDINGS Damage to Traditional and Monumental Architecture Buildings rarely survive in Japan’s archaeological record, so Japanese archaeology is, perforce, mainly post-hole archaeology. Prehistoric buildings were mostly pit houses constructed of organic materials such as wood, thatch, reed mats, and wattleand-daub; they have not survived except for their house pits and post-hole impressions, or sometimes as architectural fragments in water-logged strata (Fig. 2). From the sixth to eighth centuries, Chinese-style palatial architecture was introduced, first through Buddhist temple architecture, and then through state administrative architecture (Fig. 3). The buildings themselves were similar to the indigenous architecture, using load-bearing wooden posts and wattle-and-daub walls; however, the roofs were laid with ceramic tiles, and the posts were set on foundation stones, while the whole structure was underlain by a pounded-earth foundation platform. Bricks, whether sun-baked or fired, were unknown until the nineteenth century, though some earlier Chinese-style buildings had ceramic tile floors. Western-style brick buildings were very

TABLE 2. PREDOMINANT TYPES OF EARTHQUAKE DAMAGE AT ARCHAEOLOGICAL SITES IN THE MEDITERRANEAN COMPARED WITH JAPAN Mediterranean (Galadini et al., 2006) 1. Displacement 2. Building damage 3. Building deformation 4. Destruction layers Japan (Sangawa, 1995) 1. Fault slippage (dansō) 2. Landslides (jisuberi) 3. Cracks (kiretsu), fissures (jiwari) 4. Sand eruptions (funsa)

Figure 2. A raised storehouse of Yayoi period agriculturalists, ca. first– second century C.E., reproduced at the Toro site, Shizuoka Prefecture. The superstructure is based on rare wood fragments preserved in water-logged strata. (Author’s photo.) Total height of storehouse: 4.3 m; floor: 1.45 m above ground; floorspace: 4 × 2.5 m.

84

Barnes

Figure 3. Chinese-style architecture with foundation platforms, foundation stones, loadbearing pillars, wattle-and-daub walls, heavy roof bracketing, and ceramic roof tiles (after Kidder, 1972, Figs. 45, 52, 56 therein; author’s photo). Top: Scale drawing of the Main Hall at Yakushiji Temple, late seventh century. Lower left: Carved foundation stones and bases of pagoda central pillars and Yakushiji pagoda, built in 698 C.E. Lower right: Bracketing system of Main Hall at Tōshōdaiji Temple, postdating 759 C.E.; note uncarved foundation stones and pillar bases.

vulnerable to earthquake damage, as discovered in the 1891 Nobi earthquake (Clancey, 2006). Stone was not a common building material until the Medieval period (1185–1603) when castles began to be constructed; before that, however, tombs of the Kofun period often had burial chambers of dry-wall stone construction and coffins carved of tuff. Chinese-style buildings also used stone as facing material for the foundation platforms, for eaves-drip drainage facilities around the structure, and for the foundation stones themselves. Tombs can be considered as monumental architecture, and Caputo and Helly (2008) include them in their “buildings” category while bundling “buildings” into their “artifacts” category. Nevertheless, these authors misrepresent the Japanese archaeological record (as well as Japan’s seismic instrumentation record), indicating only a 500 yr depth back to the sixteenth century in their Figure 3 for artifacts (buildings?) in Japan. Mounded tombs (artifacts or buildings in Caputo and Helly’s terms) have been built in Japan over several centuries, dating back to the first century B.C.E., but mostly between 250 and 710 C.E. They are

particularly subject to landslips, fissures, and fault displacement, which can disrupt both their internal stone chamber structures and their external surfaces (Fig. 4). These deformations are particularly difficult to date except as occurring after tomb construction (a terminus post quem date); since the tombs are rarely reused, there are usually no later cultural materials to sandwich the seismic event. Many building foundations and facilities such as house pits, pounded-earth platforms, and stone-lined drainage canals will also show fault displacement, fissures, and crack damage (Fig. 5), but the buildings themselves have often collapsed and/or were burned in the fires that accompanied the earthquakes; rain and rot then dispersed the remains. Survival of Traditional Architecture Rapp (1986, p. 368) noted that “well built wooden structures will flex under the stress of strong earth vibrations” but that “in most structures the locations where components are joined

Earthquake archaeology in Japan

85

Figure 4. Small landslides on the Ōyama Tomb, estimated five occurrences (A–Dʹ) (courtesy of A. Sangawa, modified by Durham Archaeological Services).

Figure 5. An Early Kofun–period house pit cut diagonally by a sand dike, assigned to the earthquake of 701 C.E. Various pits and circles of burned earth are illustrated on the house floor. (After DCAN, 1996, p. 759.)

together are the weakest points.” Despite Rapp’s assessment, there are no better built wooden buildings that those in the Chinese style, and it is exactly their joinery that makes them flexible. Two aspects of this style of architecture (Fig. 3) allowed buildings to survive earthquakes if they avoided fire: their mortice and tenon joinery construction and their foundation-stone

setting. The former allowed the building to flex and shake with the tremors, then come to rest in its original state. The latter, load-bearing pillars set on stones without any other fixtures, allowed the pillars to move separately from the foundation platform, comprising a sophisticated and ingenious ancient baseisolation system.

86

Barnes

Though both China and Japan can be considered “seismic cultures” (Stiros, 1995, p. 726), there is as yet no evidence that these building techniques, dating to the late second millennium B.C.E., were active responses to earthquake damage. They might, however, be viewed as the result of “natural selection,” in that buildings of such structure survived more often than others. However, we do know that the Chinese were sensitive to earthquakes: the first seismograph was invented in China by Zhang Heng, a Han court mathematician who operated it between 132 and 139 C.E. (Pajak, 2005). His bronze seismograph has recently been re-created from fragments buried in his tomb (Xinhua News Agency, 13 June 2005). Japan has the oldest extant wooden buildings in the world, dating back to the seventh century, and most are Chinese-style architecture. Due to the ravages of fire, many original buildings have been replaced through time, but even these replacements are often several hundred years old. In 1964, Mutō Kiyoshi, professor of structural engineering at Tokyo University, ostensibly modeled his innovative flexible steel-frame lattice on the Kan’eiji Temple pagoda, one of the few buildings in Tokyo to survive the 1923 Kantō earthquake intact (NHK, 2002). The structure of a pagoda, with a central pillar set into a foundation stone and the wooden multistoried structure arranged around it with bracketing (Fig. 3), enabled the building to shake independently around the central pillar. Mutō’s flexible steel-frame lattice was first used in the Kasumigaseki Building, the first skyscraper in Japan, completed in 1968. Thus, Japan has a history of wooden architecture, and those wooden buildings that have survived through the ages are generally shrine or Chinese-style temple buildings still in use, which were usually repaired relatively quickly if they sustained earthquake damage. Most other wooden buildings, whether collapsed by earthquakes or not, succumbed to rot or fire and have disappeared from the archaeological record, unless fragments are preserved by water-logging. Thus, earthquake archaeology will recover data from building foundation remains but not superstructure. Japanese archaeologists do not generally work on standing buildings, which are assessed within the Cultural Heritage sector, particularly in conjunction with designations as National Treasures or Historic Sites. The UNESCO World Heritage Site program was inspired by Japan’s preservation efforts, dating back to the late 1800s, in designating important cultural properties for protection and maintenance. Unlike World Heritage Sites in many other parts of the world, however, Japan’s historic sites are still living buildings, not ruins. EARTHQUAKE TYPES AND ARCHAEOLOGICAL CORRELATIONS No country is more likely to be subjected to earthquakes than Japan. The archipelago sits on a conjunction of at least four plates in a subduction zone (Fig. 6). Subduction earthquakes are caused by the descent of the Pacific plate in the northeast and the Philippine plate in the southwest; there is also an incipient sub-

duction zone developing along the edge of the Japan Sea in the northwest as the backarc basin begins to close. Most subduction earthquakes to date are recorded for the Pacific seaboard, while Japan Sea–side earthquakes have so far been assigned to landbased active fault activity. Active faults, defined as rock fractures caused by pressure with vertical and/or lateral movement of rock bodies against each other, occur only in the upper 20 km of brittle crust. In Japan, these are created by the archipelago itself being squeezed between the continental Amur plate in the west (part of the Eurasian plate) and the oceanic plates to the east. The continental plate is moving eastward at up to 1 cm/yr, through Himalayan collision escape tectonics, while the oceanic plates are moving westward at 2–5 cm/yr (Taira, 2001, p. 112, 114). The stress loading on Japan, caught in between, is much higher than for subduction earthquakes, so that active fault earthquakes are both more infrequent and much stronger than ones originating in subduction. Active faults are said not to produce earthquakes of anything less than magnitude 6.5, notated here as M 6.5. Between 1975 and 1979, the Active Fault Survey Group (RGAFJ, 1992) documented thousands of faults across the archipelago and on surrounding continental shelves as known through surface and submarine features such as horizontal displacements, fault scarps, etc. These faults are ranked into three “Certainty” groups with reference to the likelihood of their having been active in the present Quaternary period: I > 90% probability; II > ~50% probability; and III lower probability of activity. Some 80% of Japanese earthquake epicenters of ≥M 6.5 in the last century have occurred on or within 5 km of a mapped active fault (RGAFJ, 1992, p. 37), and maximum magnitudes have been predicted for active fault earthquakes likely to occur in specific fault zones within Japan (RGAFJ, 1992, Fig. 5.1 therein). Earthquakes from both subduction and active fault causes are noted in many historical sources but are, of course, undifferentiated according to type. The earliest believable historical mention of an earthquake in Japan, attributed to 599 C.E., occurs in the Nihon Shoki chronicle, which was compiled in 720 C.E. The oldest actual listing of earthquakes dates to 900 C.E., documenting 700 earthquakes before 887 C.E. (Ishibashi, 2004, p. 340, 344; see also Usami, 1988). The current collations of “historical earthquakes”—defined by Ishibashi as those occurring before seismological instrumentation was developed in Japan—run to 25 volumes and are cited in Ishibashi (2004). The first modern seismographs were developed in Japan between 1880 and 1883 through the collaboration of four men (J.A. Ewing, T. Gray, S. Sekiya, and J. Milne), and from 1884, Gray-Milne seismographs began to collect data in Japan and Britain (Utsu, 2003). Thus, instrumentation records in Japan run back ~125 yr and are not limited, as implied by Caputo and Helly (2008), to the modern seismological network covering Japan today. This modern network gives immediate information on current earthquakes, and within 4 or 5 min of occurrence, the location of the epicenter, depth, and magnitude are posted on the website of the Japan Meteorological Agency (www.jma.go.jp/en/quake/).

Earthquake archaeology in Japan

87

Figure 6. The Japanese archipelago at the junction of two continental plates, the Amur (Eurasian) and Okhotsk (North American) plates, and two oceanic plates, the Pacific and Philippine plates (redrawn by Durham Archaeological Services as Barnes, 2008, Fig. 2 therein). I-STL—Itoigawa-Shizuoka tectonic line; MTL—median tectonic line.

One of the more interesting aspects of earthquake archaeology is the ability of scientists today to collate these historical and instrumentation records of earthquakes and differentiate their causes, as described in the next two sections. Subduction Earthquake Damage Figure 7, compiled by Sangawa (2001), correlates earthquake evidence sets recovered from 33 archaeological excavations with major subduction earthquakes as recorded in the historical documents after 684 C.E. The map in Figure 7 shows the archaeological sites with earthquake damage; the offshore troughs (Nankai to the left, Sagami to the right) are indicated by dotted lines. The troughs are divided into regions (A–F) of earthquake activity. The chart below this map shows the dates of the earthquakes on horizontal lines that indicate their regions of occurrence; the short vertical lines indicate the site (numbers keyed to the map) and date range of its earthquake evidence. Site damage occurs in clusters of sites located in the Tōkai (1–18), Nankai (19–30), and Kantō regions (31–33). Earthquakes in the

Nankai and Tōkai regions are generated in Nankai Trough rupture zones A–B and C–E, respectively, by Philippine plate subduction under southwestern Japan. Kantō earthquakes are generated northeast of the Sagami Trough (F) by Pacific plate subduction. The last great Kantō earthquake was in 1923, and the next big one is apparently being delayed by the wedging of a broken piece of Pacific plate in the subduction zone between the continental plate and oceanic plates (Toda et al., 2008; Miller, 2008). Two patterns are apparent in the horizontal date lines of Figure 7, indicating the regional extent of historically documented earthquakes. First, earthquakes occur simultaneously or nearly so in rupture zones A–E, as would be expected by the subduction of a single plate; however, not all zones are always attested in the documentary or archaeological records. Second, the time depths for the different zones are quite different, with far fewer earthquakes recorded in regions D–F before 1498. This has to do with the location of the ancient capitals where literacy rates were higher: areas of old literacy documented more incidents. The capitals were located in the Kinai (around sites 11–14, 15, 17) until the late twelfth century, and then in Kantō thereafter.

88

Barnes The archaeological dates indicated by vertical date ranges were obtained from stratigraphic and artifactual associations with damage assigned to earthquakes. Several sets of relative dates predate historical records at sites 8, 9, 13, 17, 18, and 26–28. These sites were assigned dates according to their cultural contents, and at least at site 8, it appears that earthquakes struck twice. The problems of relative dating of earthquakes, as well as of assigning observed damage to earthquake causes, are discussed more fully later herein. Active Fault Earthquake Damage The 1596 Fushimi earthquake in the Kinai region (Fig. 8) was copiously documented in the historical record, and several active faults are proven (through trench excavations on the fault lines) to have caused the widespread damage (Sangawa, 2001, p. 109). Such excavations have been carried out in the last few decades as part of another new discipline, earthquake geology (cf. Caputo and Pavlides, 2008). Geological excavations took place in the Arima-Takatsuki tectonic zone and included the Nodao and Higashiura faults, both on Awaji Island. Furthermore, 31 archaeological excavations in the Kinai region have revealed damage that may be correlated with the 1596 event (Fig. 8), several of which are detailed in Table 3. Identifying and Dating Earthquake Damage

Figure 7. Subduction earthquakes of southwestern Japan from the historical and archaeological records (after Sangawa, 2001, Figs. 64 and 65 therein; modified by Durham Archaeological Services). See text for explanation.

As seen in Table 3, the relative dating by archaeological artifacts is rather loose, and many assumptions are involved in choosing to assign cause of the damage to the Fushimi 1596 active fault earthquake. The subduction earthquake nearest in time, the 1605 Keichō earthquake (M 7.9) (Fig. 7, zones A–D), is the best competitor, but Sangawa states that its intensity inland would have been too weak to cause the observed liquefaction damage (seen at sites 10, 25, 26, 31). Liquefaction is acknowledged to occur at not less than M 5 and to become common at M 5.5–6 (Obermeier, 1996, p. 331); the land-based Fushimi earthquake has been assessed at M 7.5 (Kanaori and Kawakami, 1996), and the hypocenter(s) of the various faults was located much closer to the area of damage than was the Keichō earthquake. Thus, the late Medieval deformation in archaeological sites is assigned to the nearby active fault activity causing the Fushimi earthquake rather than to distant subduction activity causing the Keichō earthquake. For reference, the Kobe earthquake of 1995, caused by the Nojima fault in line with the Arima-Takatsuki tectonic line, measured M 7.3 and caused extensive liquefaction in the reclaimed land areas near Kobe as well as widespread damage in the Kinki region of Osaka, Nara, and Kyoto. A significant difference in recurrence time characterizes the two types of earthquakes: subduction (interplate) earthquakes recur within 100 yr cycles, while active fault (intraplate, plate boundary–related) earthquakes recur in periodicities of more than 1000 yr (RGAFJ, 1992, p. 41). Paleoseismologists

Earthquake archaeology in Japan

89

Figure 8. Active faults (FS) in the Kinai region and archaeological sites exhibiting damage from the 1596 Fushimi earthquake (after Sangawa, 2001, p. 88, redrawn by Durham Archaeological Services). R—Rokkō, A—Arima-Takatsuki, U—Uemachi, I—Ikoma, N—Nara, MTL—median tectonic line.

TABLE 3. DAMAGE ASSIGNED TO THE FUSHIMI EARTHQUAKE OF 1596, DATED WITH REFERENCE TO THE MEDIEVAL PERIOD (1185–1603) AND EDO PERIOD (1603–1868) 8. Imashiro-zuka tomb: A landslide of the early sixth-century mound face into the inner moat, covered with sediment dated by radiocarbon. 10. Tamakushi site: Sand liquefaction dike cut Medieval stratum, then was overlaid by Edo stratum. 13. Osaka Castle, outer moat: A landslide archaeologically dated between 1583 and 1598. 21. Ashiya ruined temple: A late seventh-century temple through which ran a 1-m-wide fissure, which collected sand and dirt clods from a rainy period, then was filled with roof tiles made only in the late sixteenth century. Contemporaneous documents describe how a typhoon hit 5 days after the 1596 earthquake, filling the fissure with sand and roof tiles. The earthquake is written to have destroyed the temple, and it was abandoned at that time. 25. Nishi-motomezuka tomb: Sand in moat liquefied to intrude Medieval stratum above, then was overlaid by Edo stratum. 26. Hyōgo-no-tsu site: Sand boils existed just underneath a fire layer dated to 1596. 31. Shimonaizen site: A sand dike cutting through a stratum containing Medieval artifacts, then covered with an Edo stratum. Note: Site numbers are keyed to Figure 8; descriptions were extracted from Sangawa (2001, p. 110–115).

interested in the elucidation of earthquake cyclicity are the main consumers of earthquake archaeology data. Based on current information, a very large combined Tōkai-Nankai earthquake is now expected before the mid-twenty-first century; it promises to affect all A–E rupture zones with great damage (Sangawa, 2001, p. 83–84). In filling out the cyclicity chart to refine expectations of future earthquakes, the main problems in Japan (as elsewhere) relate to identifying damage or deformation at archaeological sites that can be unequivocally assigned to an earthquake of a specific type and not some other cause, and to date them sufficiently.

EARTHQUAKE EVIDENCE IN SEDIMENTS Liquefaction Features Soil liquefaction was first recognized in Japan in the 1948 Fukui earthquake; studies of it increased after the 1964 Niigata earthquake (Yasuda, 2005, p. 152), when apartment buildings fell over sideways in softened sediments with relatively little damage. Liquefaction occurs only in sediments (e.g., silt, sand, gravel) saturated with water and usually requires a relatively impermeable layer (e.g., of clay) somewhere above them, which prevents

90

Barnes

the water from dispersing as pressure builds (Obermeier, 1996, p. 334–335). Ground shaking increases the pressure of the water, which fills the “pore” spaces between the sediment grains, and in extreme cases, puts the sediment particles into suspension. Ishihara and Cubrinovski (2005, p. 10) stated that “the motion in the direction of maximum intensity is the key component of the ground motion that effectively controls the development of pore water pressures and consequent ground deformation.” Sand eruptions are the usual result of subsurface liquefaction, as sediments (not just sand but silt and gravels) are pushed out to the surface with the pressurized groundwater as sand dikes, blows, or boils. Usually, the ground surface becomes depressed and cracked as water is expelled from the lower layers (Obermeier, 1996, p. 335). Obermeier also described the phenomenon of lateral spreading as an adjunct to liquefaction, happening on level ground of less than 5% incline. When an underlying stratum liquefies, the capping stratum fissures or breaks into segments, spreading laterally, especially when the cap is faced on one side by an abrupt change of slope like a terrace scarp or river bank (Obermeier, 1996, p. 335–337). Although Obermeier states that the liquefied layer itself does not undergo a change in volume (Obermeier, 1996, p. 333–334), Ishihara et al. (1997, p. 23), based on data from the 1995 Kobe earthquake lateral spreading, postulated that dilation of the overlying ground mass must have occurred, and that “it appears highly likely that the liquefied soil was sucked into the masses instead of being ejected to the surface.” The year 1985 was the year that liquefaction features were first recognized at Japanese archaeological sites (Takahama et al., 2000, p. 157), but as far as is known, lateral spreading, with its surface fissures indicative of subsurface liquefaction, has not yet been identified in the archaeological record and has not been methodologically addressed, though its existence is well attested in modern earthquakes (Ishihara et al., 1997; Kuribayashi and

Tatsuoka, 1975; Nagase et al., 2006). Galadini et al. (2006, p. 400) include liquefaction and lateral spreading as “off-site paleoseismological” effects that may damage buildings. In Japan, liquefaction dikes characteristically cut through cultural layers, or through features such as pits and ditches; the cut-and-fill relations help date the earthquake occurrence rather than contributing greatly to an understanding of destruction to a built environment. Numerous forms that liquefaction features can take at archaeological sites are illustrated by Sangawa (Fig. 9). One of the most useful aspects of earthquake archaeology is instructing archaeologists how to recognize sand dikes from their widening at the bottom and connection to the liquefied sand layer underneath, so that these features are not mistaken for canals or ditches. Another diagnostic comes from grain-size analysis: sand dike sediments fine upward, distinguishing them from fissures propagating from above, which fill in from above and fine downward. Takahama et al. (2000) have also elucidated the “draw-in” phenomenon of liquefaction, where at the end of a sand eruption, surface materials are sucked back into the top of the eruption path (Fig. 10). Often, the surface layer contains cultural materials that can be used to relative-date the eruption. Again, careful excavation will reveal the connection with the disturbed sand layer below, so that the feature—if circular—is not mistaken for a post hole. Shaking of saturated sediments can have quite opposite effects in the sedimentary record. The most extraordinary case of liquefaction in the Japanese archaeological record is at the Izumida site in Fukui Prefecture, dated to the middle of Late Yayoi (ca. 100 C.E.) (Tomiyama, 2009). Here, a large vein of cobbles up to 20 cm in diameter breached the surface; Yayoi people were apparently so surprised at this that they embedded a 35-cm-tall standing stone at the head of the eruption (Fig. 11). Conversely, at the Nishi-Sanso/Yakumo-Higashi site in Osaka, mild liquefaction features have been recovered in section (Fig. 12). Dish and pillar structures are identifiable in the upper sand stratum; dike

A F E

D C

B

G

I H

Figure 9. Liquefaction and faulted structures revealed in archaeological sites (modified from Sangawa, 2001, p. 52). I–IV—strata; black areas—cultural features; A—sand boil later than strata II and III but earlier than stratum I; B—sand dike cutting through earlier sand boil D into preexisting cultural feature; C—sand eruption cutting through existing stratum III and cultural feature but truncated by surface razing before stratum II was laid; D—sand boil cutting through existing stratum III before stratum II was laid, cut by later sand dike B; E—deflected sand eruption; F—fizzled sand eruption; G—sand dike cutting through existing strata II and III but cut by later cultural feature; H, I—normal listric fault slips; X—pillar structures, Y—dish structures; Z—disturbed area.

Earthquake archaeology in Japan

91

Figure 10. Liquefaction draw-in of cultural materials from overlying cap (after Takahama et al., 2000, their Fig. 9). (Top) Natural stratigraphy before earthquake. (Middle) Sand and gravel erupts to surface. (Bottom) Erupted material partially drawn back into eruption vent.

intrusions strike upward, some obviously abortive, while two small efforts at sill formation are visible. In sum, sand eruptions and lateral spreading are two immediate effects of liquefaction that can be incontrovertibly assigned an earthquake cause. Obermeier (1996, p. 334–337) has implied that sand dikes, sills, and dish and pillar structures are known results of liquefaction and that liquefaction itself is a result of earthquake activity; no other earth motions have the shaking severity necessary to raise pore pressure to explosive levels. Accepting this, we may assume that liquefaction features at an archaeological site, if properly identified, are evidence of past earthquake events. The advantage of finding these features at sites rather than in the great geological outdoors is the association of cultural materials that may aid in dating the events. Soft-Sediment Deformation Structures A more difficult arena in determining earthquake effects is soft-sediment deformation. Fine-grained sedimentation is characteristic of low-energy depositional environments such as ocean floors and lake bottoms. The muddy deposits that collect there can be disturbed and distorted either during or after sedimentation

Figure 11. Liquefaction eruption of ~20-cm-diameter cobbles and cultural isolation as sacred site, Late Yayoi Izumida site, Fukui Prefecture (courtesy of A. Sangawa). Cobble layer can be seen in lower strata in trench exposure.

92

Barnes

A C

B

by several processes—not only by seismic shaking but by “loading during rapid sedimentation, localized artesian conditions, or slumping” (Obermeier, 1996, p. 332). The deformation structures formed under these pressures are patterned and have been given evocative names, including ball-and-pillow, flame, water-escape structures, neptunian dikes, load casts, slumps, and convolutions. These comprise plastic deformations of the soft sediments, which then lithify over time; they can be seen either in rock formations going back to Paleoproterozoic times or in modern unlithified sediments. The challenge is to prove which ones were caused by earthquakes rather than by some other disturbance. As Guidoboni emphasized (1995, p. 10–11), structural damage and landscape changes cannot be assumed to result from earthquakes but must be proven so. Obermeier (1996, p. 332) noted that the geological literature is “replete” with references to “seismites”—just such patterned disturbances as are assumed to have been caused by earthquake shaking. The last decade has seen a concerted effort to assess critically the formation processes of soft-sediment deformation structures (Ettensohn et al., 2002; Shiki et al., 2000), and the term “seismite” has somewhat fallen out of favor as an initial descriptor. Even after sufficient investigation, a suspected seismite might be prefaced with different degrees of inclusion in that category as certain, probable, likely, or possible (Greb et al., 2002). Several authors offer criteria by which to determine an earthquake cause for deformed sediments (e.g., Rossetti and Góes, 2000; Moretti, 2000; Greb et al., 2002; Greb and Dever, 2002; Davies et al., 2004; Bowman et al., 2004; Mazumder et al., 2006). Of these seven sources cited, six specify two important criteria: exclusion of normal processes, and correlative horizons (i.e., lateral continuity) over wide areas. Four speak of the importance of regional abundance of data, so that earthquake attribution does not rely on a single site or single observation. Five specify the obvious criteria of being located in a tectonically active zone, having other tectonic evidence nearby, and occurring in conjunction with deposition (synsedimentary evidence). Three emphasize that the deformed strata should be discrete, occurring sandwiched

C

A

Figure 12. Liquefaction structures at the Nishi-Sanso/Yakumo-Higashi site, Osaka (after Sangawa, 1999, Fig. 8 therein). Eruptions of sand from stratum II into stratum I; A—disturbed area, B—pillar structures, C—dish structures.

between other undeformed strata or strata of different origins. Two note that such deformed strata should recur, possibly rhythmically or cyclically. Two emphasize the importance of deformation immediately postdepositional while still unconsolidated. Finally, individual sources cite further interesting criteria: the necessity to define a core zone of activity; the presence of complex upward deformation; the presence of long periods of quiescence; and the similarity of deformed patterns to modern cases or experimental results. Thus, we cannot immediately discuss the presence of seismites in Japanese archaeological sites without first proving their seismic origins; instead, we must first talk about soft-sediment deformation structures. The criteria for judgment should be the same as proposed already for other areas outside Japan, and it is notable that the initial forays in this direction within Japan very much take the regional or “territorial” approach common to most of the aforementioned authors, as well as Rapp (1986), Obermeier (1996), and Galadini et al. (2006). Damage at any one site or location may have been caused by local conditions, but an earthquake usually has broader geographical repercussions. This approach simply relies on recovering similar evidence of damage from a wide area (several sites’ worth) at a certain date, so that through repetition of data, the conclusion that the damage was caused by an earthquake and not a more localized situation is reinforced. A limitation of this approach, however, is chronological uncertainty that each instance of damage was caused by the same earthquake (Galadini et al., 2006, p. 407, 408): There is always the potential of collapse of different events into one, resulting in the exaggeration of event size (Galadini et al., 2006, p. 408). Matsuda Jun-ichirō’s research in the eastern Osaka Basin, with its attention to regional occurrences of soft-sediment deformation structures in the late Holocene archaeological record, is an important new direction in Japanese earthquake archaeology. A trained geomorphologist and sedimentologist working as a geoarchaeologist, Matsuda has focused on disturbance layers in the stratigraphic succession of alluvial and cultural deposits

Earthquake archaeology in Japan from postglacial Jōmon to recent times. His 2000 publication synthesizes the sequences from six archaeological sites clustered in an area 7 km E-W by 4 km N-S in the southern Osaka Plains (cf. Fig. 8, just east of site 13). The deformation layers are demonstrated to occur in a tectonically active zone at multiple, separated site intervals, and are territorially correlative (Fig. 13). These data appear to fulfill the territorial approach criterion for assessing seismogenic status; their repetitive nature at intervals of several centuries, occurring between undisturbed layers of parallel laminated strata, also argues for occasional large disturbances separated by periods of quiescence rather than routine storm or slumping conditions. For these reasons, the deformations identified by Matsuda may be considered to be seismites. At the Kitoragawa site (Fig. 14), the evidence of softsediment deformation structures in layers relatively dated by cultural materials can be correlated with six or seven historically known earthquakes, all assessed by Usami at magnitudes between M 7 and M 8.4 (Matsuda, 2000, Fig. 11 therein). This seems to go against Obermeier’s statement that “soft-sediment deformations often form at such low levels of seismic shaking

93

that the shaking poses no hazard” (1996, p. 332). Several of the named earthquakes were Nankai subduction earthquakes, which have already been shown to have affected archaeological sites in the eastern Inland Sea region in and around Osaka (Fig. 7). The unaddressed question is, what was the level of intensity of ground shaking at these various sites to create greater (liquefaction) or lesser (soft-sediment deformation) disturbance features? The features investigated by Matsuda were all generated in muddy strata submerged under standing water at the time of deformation: in deltaic back marshes, small shallow lakes, or paddy fields. They appeared in the plans and sections of archaeological trenches; Matsuda cut out blocks and returned them to the laboratory for further analysis, including X-ray radiographs of 1-cm-thick slices (Fig. 15). He identified several kinds of features in stratigraphic succession, going upward: downward fissures and microfaults at the bottom, load structures then plumose patterns, and homogenization at the top. The extensive use of radiographs in addition to sketch drawings of the structures answers Obermeier’s lament concerning liquefaction that “the deformational effects…have rarely been illustrated and discussed

Figure 13. Correlation of soft-sediment deformation zones at different archaeological sites on the Osaka Plains (modified from Matsuda, 2000, Fig. 9 therein).

Figure 14. Soft-sediment deformation zones (DZ) and anthropogenic zones (EZ) at Kitoragawa site (modified from Matsuda, 2000, Fig. 10 therein).

Figure 15. Radiographs of soft-sediment deformation structures (photograph courtesy of J.-i. Matsuda; caption based on Matsuda, 2000). (Left) Deformation zone characterized by, from the top down, liquidized deformation unit with laminae of a weak load structure from a younger deformation, a hydroplastic deformation unit with plumose pattern of light and white intermixed and depressions at its base, and a lower, white hydroplastic deformation unit. Kitajima site. (Middle) Two deformation zones. The top light-colored layer is a young deformation zone. The older zone begins with the dark liquidized deformation unit having an undulated base in response to downward pressure. This is underlain by a hydroplastic deformation unit with twisted plumose pattern; the lower area of microfaults (white vertical fracture lines) indicates a brittle deformation unit. Kitoragawa site. Lower half of this image is a normal photograph taken in daylight. (Right) Two deformation zones with load structure. The darker liquidized homogenized band through the center marks the top of the older deformation zone; the lowermost peaty mud bed is deformed into blocky shapes. The light-colored layers of the younger deformation zone are separated horizontally into a hydroplastic deformation unit with plumose patterns above and a hydroplastic deformation unit below having downward lobes that involve the older homogenized layer. Kitoragawa site. Scale bar = 10 cm.

Earthquake archaeology in Japan in vertical section, the view most useful for paleoliquefaction studies” (1996, p. 332). Several earthquakes were identified through these deformation patterns that are not attested in the historical record: one in early Heian, one in Late Kofun, and at least one in the first half of Yayoi. Prehistoric subduction earthquakes are also attested at other archaeological sites, as shown in Figure 7, though the evidence is quite different. It remains to correlate the soft-sediment deformation structures with other dated sites to construct a potential prehistoric series of earthquakes. Soft-sediment deformation studies have broadened the earthquake data-gathering capacity of archaeological excavation, but few specialists are available to do them. Japanese geoarchaeologists have unknowingly responded to Rapp’s original goal of paying more attention to the sediments in archaeoseismic excavations (Rapp, 1986, p. 370), and this methodological aspect of Japanese earthquake archaeology is well worth exporting, since many other countries are also subject to post-hole archaeology without substantial building remains. CONCLUSIONS Earthquake archaeology in Japan arose simultaneously in the mid-1980s with archaeoseismology in the Mediterranean, but it has taken off in quite different directions. Having few building structures in which to assess earthquake damage, the actual sediments at archaeological sites and their tectonic disruption have come under examination. Geomorphology rather than paleoseismology or historical seismology has been the main disciplinary inspiration; thus, it is the identification of morphological features that becomes the foremost task. Archaeologists are being encouraged to recognize features such as fault displacement, fissures, and liquefaction eruptions, while a few geoarchaeological specialists are engaged in sediment analysis for soft-sediment deformation patterns. Much of the earthquake damage recovered in archaeological excavations in Japan is geological damage to soils, sediments, and strata. The main consumers of earthquake archaeology in Japan have been paleoseismologists interested in establishing recurrence rates. Since such rates are different for subduction and active fault earthquakes, the challenge has been to divide the archaeological evidence to support one or the other series. This relies both on accurate dating and on the definition of broad areas of coseismic damage. Although relative dating by association with cultural artifacts or layers is not as precise as documentary or absolute dating, the Japanese ceramic chronology is the most refined in the world, with generational (20 yr) spans of earthenware types and 5 yr spans of stoneware types in the prehistoric and early historic periods in the best-case examples. This is better than establishing “seismic periods” (Caputo and Pavlides, 2008, p. 3) or spans of a “few decades” (Galadini et al., 2006, p. 407). That relative dates are imprecise may irritate instrumentalists used to dealing with precise measurements, but any time human activities enter the research picture, we must deal with

95

probabilities rather than certainties. Rather than ignore or discard the datings provided by archaeoseismologists, these should be taken as supporting data for working hypotheses about past earthquake occurrence. Earthquake archaeology provides a finer framework for discovering and dating past earthquakes than do trenching activities undertaken by earthquake geologists, and so the additional data should be welcomed. Moreover, extension of research interests beyond earthquake damage to former civilizations (Caputo and Helly, 2008) by including worldwide Paleolithic and Neolithic sites should enormously expand the data available. As Noller and Lightfoot (1997) did for prehistoric sites in America, Japanese archaeologists have demonstrated that any earthquake evidence that occurs in conjunction with cultural materials can be relatively dated, regardless of period or level of social development and whether or not those “artifacts” suffered “damage.” A more thorough treatment of approaches to archaeoseismology worldwide, however, would need to take account of Japanese-language sources and even the increasing body of English-language materials on the Japanese situation, including the depth of historical information about earthquakes in Japanese texts (cf. Ishibashi, 2004) and the instrumentation records that go back to the invention and deployment of the seismograph in Japan and England in the 1880s (Clancey, 2006; Utsu, 2003). Earthquake archaeology in Japan has demonstrated that sites of all ages are repositories of earthquake damage, once archaeologists learn to read the traces. Also, damage need not be restricted to buildings but can affect the actual sediments at archaeological sites. The inclusion of the full range of Japanese historic and archaeological earthquake data cannot but enhance the emerging discipline of archaeoseismology, as would information from other tectonically active regions of the world. ACKNOWLEDGMENTS I am very grateful to Sangawa Akira, Okamura Katsuyuki, and Matsuda Jun-ichirō, who inspired my interest in earthquake archaeology through their writings, who generously discussed with me the finer points of their new discipline, and who have given permission for reuse of their published material. Iain Stewart, Manuel Sintubin, and Tina Niemi facilitated making these Japanese data available in English through inclusion in their International Geoscience Program IGCP 567 project, including presentation of the data at the Seismological Society of America meetings in Santa Fe, 9 April 2008, and the Geoarchaeology meetings in Sheffield, 17 April 2009. Thanks also go to Bruce Batten for his constructive refereeing. My research on this topic has been supported by Durham University, the Arts and Humanities Research Council (AHRC) of England, the School of Oriental and African Studies (SOAS) at the University of London, and the International Research Center for Japanese Studies (Nichibunken) in Kyoto, Japan. Several illustrations were prepared by Durham Archaeological Services (DAS), Durham University.

96

Barnes

REFERENCES CITED Barnes, G.L., 2008, The making of the Japan Sea and the Japanese mountains: Understanding Japan’s volcanism in structural context: Japan Review, v. 20, p. 3–52. Bowman, D., Korjenkov, A., and Porat, N., 2004, Late-Pleistocene seismites from Lake Issyk-Kul, the Tien Shan range, Kyrghyzstan: Sedimentary Geology, v. 163, p. 211–228, doi: 10.1016/S0037-0738(03)00194-5. Caputo, R., and Helly, B., 2008, The use of distinct disciplines to investigate past earthquakes: Tectonophysics, v. 453, p. 7–19, doi: 10.1016/j.tecto .2007.05.007. Caputo, R., and Pavlides, S.B., 2008, Earthquake geology: Methods and applications: Tectonophysics, v. 453, p. 1–6, doi: 10.1016/j.tecto.2008.01.007. Clancey, G., 2006, Earthquake Nation: The Cultural Politics of Japanese Seismicity, 1868–1930: Berkeley, University of California Press, 331 p. Davies, N.S., Turner, P., and Sansom, I.J., 2004, Soft-sediment deformation structures in the Late Silurian Stubdal Formation: The result of seismic triggering: Norwegian Journal of Geology, v. 85, p. 233–243. DCAN (Disaster Concerned Archaeologists’ Network), 1996, Hakkutsu sareta jishin konseki (Excavated evidence of earthquakes): n.p., Osaka, Maibun Kankei Kyūen Renraku Kaigi, 826 p. Ettensohn, F.R., Rast, N., and Brett, C.E., eds., 2002, Ancient Seismites: Geological Society of America Special Paper 359, 190 p. Galadini, F., Hinzen, K.-G., and Stiros, S., 2006, Archaeoseismology: Methodological issues and procedure: Journal of Seismology, v. 10, special issue, p. 395–414, doi: 10.1007/s10950-006-9027-x. Greb, S.F., and Dever, G.R., Jr., 2002, Critical evaluation of possible seismites: Examples from the Carboniferous of the Appalachian basin, in Ettensohn, F.R., Rast, N., and Brett, C.E., eds., Ancient Seismites: Geological Society of America Special Paper 359, p. 109–125. Greb, S.F., Ettisohn, F., and Obermeier, F., 2002, Developing a classification scheme for seismites: Geological Society of America Abstracts with Programs, v. 34, no. 2, p. A-102; also available at http://gsa.confex.com/ gsa/2002NC/finalprogram/abstract_32750.htm (accessed February 2008). Guidoboni, E., 1995, Archaeology and historical seismology: The need for collaboration in the Mediterranean area, in Stiros, S., and Jones, R.E., eds., Archaeoseismology: Athens, British School in Athens, p. 7–13. Ishibashi, K., 2004, Status of historical seismology in Japan: Annals of Geophysics, v. 47, no. 2/3, p. 339–368. Ishihara, K., and Cubrinovski, M., 2005, Characteristics of ground motion in liquefied deposits during earthquakes: Journal of Earthquake Engineering, v. 9, Special Issue 1, p. 1–15, doi: 10.1142/S1363246905002304. Ishihara, K., Yoshida, K., and Kato, M., 1997, Characteristics of lateral spreading in liquefied deposits during the 1995 Hanshin-Awaji earthquake: Journal of Earthquake Engineering, v. 1, no. 1, p. 23–55, doi: 10.1142/ S1363246997000039. Kanaori, Y., and Kawakami, S., 1996, The 1995 7.2 magnitude Kobe earthquake and the Arima-Takatsuki tectonic line: Implications of the seismic risk for central Japan: Engineering Geology, v. 43, no. 2–3, p. 135–150, doi: 10.1016/0013-7952(96)00056-7. Kidder, J.E., Jr., 1972, Early Buddhist Japan: London, Thames & Hudson, 212 p. Kuribayashi, E., and Tatsuoka, F., 1975, Brief review of liquefaction during earthquakes in Japan: Soil and Foundation, v. 15, no. 4, p. 81–92. Matsuda, J.-i., 2000, Seismic deformation structures of the post-2300 a BP muddy sediments in Kawachi lowland plain, Osaka, Japan: Sedimentary Geology, v. 135, p. 99–116, doi: 10.1016/S0037-0738(00)00066-X. Mazumder, R., van Loon, A.J., and Arima, M., 2006, Soft-sediment deformation structures in the Earth’s oldest seismites: Sedimentary Geology, v. 186, p. 19–26, doi: 10.1016/j.sedgeo.2005.12.002. Miller, M., 2008, Breaking the slag: Nature Geoscience, v. 1, p. 730–731, doi: 10.1038/ngeo341. Moretti, M., 2000, The interpretation of soft-sediment deformation structures as seismites, in European Geophysical Society, 25th general assembly (10297006) 01/01/2000 abstracts: www.copernicus.org/EGS/egsga/nice00/ programme/abstracts/aac6959.pdf (accessed February 2008). Nagase, H., Zen, K., Hirooka, A., Yasufuku, N., Kasama, K., Kobayashi, T., Maeda, Y., Uno, K., Hashimura, K., and Chen Guangqi, 2006, Zoning for liquefaction and damage to port and harbor facilities and others during the 2005 Fukuoka-ken Seiho-Oki earthquake: Soil and Foundation, v. 46, no. 6, p. 805–816.

NHK (Japan Broadcasting Company), 2002, Kasumigaseki-Biru—Chōkōsō e no hatenaki tatakai: Jishin rettō Nihon no kakumei gijutsu (Kasumigaseki Building—Endlessly fighting towards skyscrapers: Revolutionary technologies for the Japanese earthquake-prone archipelago): Tokyo, NHK Sofutowea (DVD) [in Japanese]. Noller, J.S., and Lightfoot, K.G., 1997, An archaeoseismic approach and method for the study of active strike–slip faults: Geoarchaeology, v. 12, p. 117–135, doi: 10.1002/(SICI)1520-6548(199703)12:2<117:: AID-GEA2>3.0.CO;2-7. Obermeier, F., 1996, Using liquefaction-induced features for paleoseismic analysis, in McCalpin, J.C., ed., Paleoseismology: San Diego, Academic Press, p. 331–396. Pajak, J., 2005, Signal processing in the “Zhang Heng Seismograph” for remote sensing of impending earthquakes, in Gupta, G.S., Mukhopadhyay, S.C., and Messom, C.H., eds., Proceedings of the 1st International Conference on Sensing Technology, 21–23 November 2005, Palmerston North, New Zealand: www-ist.massey.ac.nz/conferences/icst05/proceedings/ ICST2005-Papers/ICST_112.pdf, p. 669–673 (accessed September 2008). Rapp, G., Jr., 1986, Assessing archaeological evidence for seismic catastrophes: Geoarchaeology, v. 1, p. 365–379, doi: 10.1002/gea.3340010403. RGAFJ (Research Group for Active Faults in Japan), 1992, Maps of Active Faults in Japan with an Explanatory Text: Tokyo, University of Tokyo Press, 73 p., 3 maps [English condensed version of 1991 full publication in Japanese]. Rossetti, D.F., and Góes, A.M., 2000, Deciphering the sedimentological imprint of paleoseismic events: An example from the Aptian Codó Formation, northern Brazil: Sedimentary Geology, v. 135, p. 137–156, doi: 10.1016/ S0037-0738(00)00068-3. Sangawa, A., 1988, Kōkogaku no kenkyū taishō ni mitomerareru jishin no konseki: Kodaigaku Kenkyū, v. 116, p. 1–16 [in Japanese]. Sangawa, A., 1995, Kōkogaku no shiryō kara kojishin o saguru (Searching for ancient earthquakes in archaeological data), in Ōta, Y., and Shimazaki, K., eds., Kojishin o Saguru: Tokyo, Kokon Shoin, p. 107–124 [in Japanese]. Sangawa, A., 1999, Palaeoliquefaction features at archaeological sites in Japan: Chigaku Zasshi, v. 108, no. 4, p. 391–398. Sangawa, A., 2001, Jishin: Namazu no Katsudōshi (Earthquakes: A History of Catfish Activities): Tokyo, Daikōsha, 173 p. Sangawa, A., 2007, Jishin no Nihonshi (Earthquakes in Japanese History): Tokyo, Chūō Kōron Shinsha, 268 p. Sangawa, A., 2009, A study of paleoearthquakes at archeological sites: A new interdisciplinary area between paleoseismology and archeology: Synthesiology, English edition, v. 2, p. 84–94. Shiki, T., Cita, M.B., and Gorsline, D.S., 2000, Sedimentary features of seismites, seismo-turbidites and tsunamiites—An introduction: Sedimentary Geology, v. 135, no. 1–4, p. vii–ix, doi: 10.1016/S0037-0738(00)00058-0. Stiros, S.C., 1995, Archaeological evidence of antiseismic constructions in antiquity: Annali di Geofisica, v. 35, p. 725–736. Stiros, S., and Jones, R.E., 1996, Archaeoseismology: Athens, British School at Athens, Fitch Laboratory Occasional Paper 7, 268 p. Taira, A., 2001, Tectonic evolution of the Japanese island arc system: Annual Review of Earth and Planetary Sciences, v. 29, p. 109–134, doi: 10.1146/ annurev.earth.29.1.109. Takahama, N., Ōtsuka, T., and Brahmantyo, B., 2000, A new phenomenon in ancient liquefaction, the draw-in process, its final stage: Sedimentary Geology, v. 135, p. 157–165, doi: 10.1016/S0037-0738(00)00069-5. Toda, S., Stein, R.S., Kirby, S.H., and Bozkurt, S.B., 2008, A slab fragment wedged under Tokyo and its tectonic and seismic implications: Nature Geoscience, v. 1, p. 771–776, doi: 10.1038/ngeo318. Tomiyama, M., 2009, Hayashi-Fujishima Iseki Izumida-chiku. Fukui-ken Maizou Bunkazi Chosa Houkoko, v. 106, p. 87, 89, + photos. Usami, T., 1988, A study of historical earthquakes in Japan, in Lee, W.H.K., and Shimazaki, K., eds., Historical Seismograms and Earthquakes of the World: San Diego, Academic Press, p. 276–288. Utsu, T., 2003, Historical development of seismology in Japan: Swiss National Centennial Report to the International Association of Seismology and Physics of the Earth’s Interior, Ch. 79.33 Japan, Part 2: www.iris.edu/ seismo/info/historical/Utsu2003.pdf, 19 p (last accessed August 2008). Yasuda, S., 2005, Survey of recent remediation techniques in Japan, and future applications: Journal of Earthquake Engineering, v. 9, Special Issue 1, p. 151–186, doi: 10.1142/S1363246905002213. MANUSCRIPT ACCEPTED BY THE SOCIETY 19 MAY 2010 Printed in the USA

The Geological Society of America Special Paper 471 2010

Historical earthquake catalogues and archaeological data: Achieving synthesis without circular reasoning John D. Rucker Tina M. Niemi Department of Geosciences, University of Missouri–Kansas City, Kansas City, Missouri 64110, USA

ABSTRACT The field of archaeoseismology has been plagued by a persistent problem. The problem has been the integration of several lines of evidence to produce a holistic conclusion without entering into a situation of circular reasoning, wherein the sources are used to build on each other without foundation. The four main sources of evidence are historical texts, epigraphy, archaeology, and geology. Any seismic event may appear in any or all of them, but only the most extreme events in fortuitous locations would be expected to appear in all four. This paper uses some aspects of the interpretation of the 551 C.E. earthquake in the Levant to illustrate how this circular reasoning can develop, and how it tends to corrupt the different lines of evidence. We conclude with a suggested new approach, making the database of regional seismic events both more specific and more complete. toriography such as bias, both cultural and geographic, accident of preservation, and availability, confusion, amalgamation, and conflation in the sources. Since 1994, the three primary earthquake catalogs in English for the Levant, Amiran et al. (1994), Ambraseys et al. (1994), and Guidoboni (1994), in addition to Russell’s 1980 and 1985 papers, have been widely utilized. These earthquake catalogues have been viewed as authoritative within the community of users. However, a troubling tendency toward circular reasoning has developed. It is fair to note that two recent articles have begun the process of presenting new comprehensive catalogs and separating the sources of evidence. Particularly notable in doing so are Sbeinati et al. (2005) and Guidoboni and Comastri (2005). Ambraseys (2009) also turned a more critical eye toward separating the evidence, especially in sorting out conflating historical sources. Within the subfield of archaeoseismology, a holistic approach including all of the possible sources of information is especially important, since major seismic events are by their nature regional, with sometimes far-flung and occasionally subtle effects. In this article, we attempt to show how these historical

INTRODUCTION Because the ultimate goal of all historical and archaeological research is to achieve a better understanding of ancient events and cultural lifestyles, it is best to utilize all available sources of information. An archaeological study that ignores historical sources is doomed to at best, much additional work and at worst, failure. Historians who ignore archaeological sources will certainly miss the best independent corroboration of the ancient written sources with which they are working. Both disciplines are wise to consider newly available data from unexpected and far-flung sources as they become available (the revolution that was radiocarbon dating, for example). Thus, working in the Near East, we find ourselves being informed by sources of data as disparate as core samples from the Greenland ice sheets and historical sources from all over Eurasia, as well as our own direct excavations and historical sources from the region itself. Typically, modern earthquake catalogues are constructed by compiling all possible textual references to historic earthquakes. This makes them subject to many of the common concerns of his-

Rucker, J.D., and Niemi, T.M., 2010, Historical earthquake catalogues and archaeological data: Achieving synthesis without circular reasoning, in Sintubin, M., Stewart, I.S., Niemi, T.M., and Altunel, E., eds., Ancient Earthquakes: Geological Society of America Special Paper 471, p. 97–106, doi: 10.1130/2010.2471(09). For permission to copy, contact [email protected]. © 2010 The Geological Society of America. All rights reserved.

97

98

Rucker and Niemi

records, inscriptions, archaeological excavations, and geological data can and should mutually reinforce each other, but are susceptible to misuse, resulting in a situation of runaway circular reasoning, which taints the sources of data. The problem of circular reasoning in archaeoseismology is illustrated in Figure 1. Simply put, once this circular cycle begins, both lines of evidence become corrupted. SOURCES OF EVIDENCE The four main sources of evidence are historical texts, epigraphy, archaeology, and geology. For any given earthquake, evidence from any or all of these sources may be available. Each source, however, is only capable of providing certain kinds of information, and may be completely silent on other aspects of the seismic event. It is thus extremely important that scholars keep firmly in mind what a piece of evidence is not revealing—not just what it is revealing. In many ways, historical text is the most straightforward of the four sources of earthquake data. A respected ancient historian who says “City X was destroyed by an earthquake at 9 p.m. on Monday the 19th of September” is hard to dispute. The historical record can have the advantages of specificity, and clarity of message, but it may also have the familiar historigraphical disadvantages of secondary evidence, exaggeration, and deliberate or accidental misinformation (e.g., Guidoboni and Ebel, 2009). Some problems specific to historical seismology are: There is considerable bias toward regions of denser population and earthquakes of greater magnitude. That is, earthquakes of greater magnitude and/or occurring in areas of denser populations are much more likely to enter the historic record. Also, there is a tendency toward amalgamation of earthquakes for which the occurrences were closely spaced in time. This is due both to the limitations of ancient knowledge and the vagaries of preservation and copying amongst historic texts. When we consider the widely spaced geographic locations of ancient sources, it is not surprising that

Archaeologist finds a destruction layer at Site X and interprets it as evidence for an earthquake.

Historical seismologist adds Site X to the catalogue of cities damaged by that specific earthquake.

Archaeologist uses earthquake catalogue to assign a date to the destruction layer.

Figure 1. Circular reasoning model.

moderate earthquakes with smaller felt areas, though possibly quite severe in their local effects, might escape the notice of a distant chronicler entirely. Earthquake catalogues are collections of dates and reports of the effects of earthquakes as recorded from written records. Most catalogues are thought to be complete for major M >7 earthquakes but may be silent on less severe or less widespread earthquakes. This is, of course, a summary of an extremely complicated type of evidence that must not be taken at face value without consideration of its various problems. Guidoboni and Ebel (2009) dealt with the problems of historical seismology at a length and detail not possible in this paper. Also, Karcz (2004) contains case studies helpful in sorting out the historical sources for several specific ancient earthquakes in the Levant. Another important source for historical seismology is epigraphy. This is the study of surviving inscriptions, usually on structures, statuary, or ostraca, but also including ancient text on any artifact. These are often dedicatory in nature, but include more mundane inscriptions and even graffiti. Since inscriptions are also a written record, they are often lumped together as a subcategory of the historic data. For the purposes of historical seismology, there is an important reason to separate epigraphy from historical sources. Epigraphy is usually a written primary source, without the problems of later copying or interpolation. Thus, it is a slightly but significantly different line of evidence from the rest of the historic record. By separating it, we maintain an independent line of evidence. However, the main disadvantage of inscriptions is that they are generally short, and may or may not do more than hint at the events or issues in question, often enigmatically, as we discuss later with the Areopolis inscription. Epigraphy relies on history to flesh out the details, and archaeological and geological stratigraphy to help constrain both the date and context of the evidence. Obviously, an earthquake happens when a fault ruptures, so investigations of the stratigraphy and geomorphology of seismogenic faults in the region are critical to understanding the history of earthquakes in an area. The main advantage of this line of evidence is that it can provide an absolute record of slip events on that fault. The main disadvantage is that sediments of appropriate age to constrain relevant seismic events may not be present. In addition, the chronological resolution may be too coarse to define a specific historic earthquake. Even though geological study is the baseline of seismology, in many ways, archaeological sites offer better research opportunities, because anthropogenic features may provide better information on both dating and seismic effects than is typically present in a natural geologic setting. Paleoseismic research based predominantly on geologic data also relies on historical accounts and archaeological data to interpret the record of fault movement and secondary seismic effect. Archaeological evidence has its own set of strengths and weaknesses. The first advantage to the archaeological line of evidence is that there are many more archaeological sites than there are surviving ancient texts, and thus is a wider potential distribution of data is available. The second main advantage of

Historical earthquake catalogues and archaeological data archaeological data for historical seismology is that it is free of bias—you are not looking at the evidence through the lens of an ancient historian, nor even through that of an ancient stone carver. However, the evidence is difficult to interpret and can be obscure. Any bias ultimately introduced comes from the modern scholars who interpret (or misinterpret) it. Finally and most importantly, archaeological artifacts can provide a fine resolution of time within the stratigraphic section. The main disadvantage of archaeology as a line of evidence for archaeoseismology is that the stratigraphic and structural damage data are often open to varying interpretations. Furthermore, archaeological information is only available from those sites that have been studied or excavated. Thus, it is at best incomplete, and possibly contains a sampling bias. One area in which a sampling bias already exists in the archaeological record is a definite bias toward study and excavation of larger, more obtrusive sites. This sampling bias may impact any given question in historical seismology. An important example of the ways in which these lines of evidence have been used in the past is the study of the 551 C.E. earthquake in the Levant. This example is important both in that the earthquake in question was a major seismic event, and in that we argue that the way the evidence has been used typifies the problem with circular reasoning we seek to elucidate. THE LEVANT EARTHQUAKE OF 9 JULY 551 Historical Text The main historical sources for the 9 July 551 C.E. earthquake as described by Guidoboni (1994) are Antoninus of Piacenza, Malalas, John of Ephesus, Theophanes, Agathias Scholasticus, and the accounts of the life of St. Symeon the Stylite the Younger. Contemporary or primary sources (those alive at the time of the event) include Malalas, Agathias, John of Ephesus, and Antoninus Placentinus. Later chroniclers such as Theophanes, who wrote in the eighth–ninth century C.E., and others are considered noncontemporary or secondary sources. Secondary sources recount, copy, and often embellish what they have heard or read, thus diminishing the reliability of the information. However, it can also be said that the similarity of prose between primary sources often suggests that these text have also been copied. Several lines of historical evidence in these texts indicate that the earthquake damage was centered on the coast of Phoenicia, modern Lebanon. First, Antoninus of Piacenza describes massive destruction of the cities between Tripoli (Tripolis) and Beirut (Berytus). Second, the account in St. Symeon indicates minor damage north of Beirut and “the area to the south from Tyre to Jerusalem was also preserved” (Guidoboni, 1994, p. 335). Third, the text by Agathias indicates that the law school of Beirut was transferred after the earthquake to Sidon, corroborating evidence that the intensity of seismic shaking was weaker toward the south, and structures there remained intact and functioning. Finally, the occurrence of a seismic sea wave (tsunami) in

99

551 C.E. suggests slip on an offshore fault in addition to the described coastal landslide (Elias et al., 2007). This position is partially shared by Salamon et al. (2007), although they also allow the possibility of a submarine landslide as the result of an onshore rupture. Guidoboni (1994) and a number of previous earthquake catalogers place the epicenter of the 551 C.E. earthquake in coastal Lebanon (Fig. 2A). Two sources of confusion arise pertaining to the epicenter of the 551 C.E. earthquake. First, descriptions found in the texts of Malalas and Theophanes list such widely separated areas as Syria, Arabia, Mesopotamia, Palestine, and Phoenicia as areas affected by the earthquake. The second source of confusion is that several different earthquakes occurred in Turkey, the Levant, and Egypt during the period between 551 C.E. and 561 C.E. (Russell, 1985; Guidoboni, 1994). This has led to several different interpretations of the historical texts. Ambraseys et al. (1994), reading the same historical sources, places the epicenter of the 551 earthquake in the Jordan Valley based on damage in Alexandria, Egypt (Fig. 2B). These authors state “Modern writers place the epicentral region of this event offshore from Lebanon. This is due to the bias of information from the more populous coastal region” (p. 24). Ben-Menahem in his 1979 earthquake catalogue defines a Lebanese coastal epicentral zone, but in a later publication (1991), he lists a 7 July 551 C.E. earthquake in the Gulf of Corinth. Amiran et al. (1994) in their earthquake catalog for Israel and adjacent areas for the 551 C.E. earthquake include a Latin quote (apparently from Theophanes) that translates as “a great and terrible earthquake occurred throughout the regions of Palestine, Arabia, Mesopotamia, Syria, and Phoenica” and also lists several cities damaged in the earthquake beyond the Lebanese coast, including Jerusalem, Jerash, Mt. Nebo, Areopolis, el-Lejjun, and Petra. They state that the “el-Lejjun fortress east of Kerak destroyed. Petra destroyed and never rebuilt.” The inclusion of the earthquake damage at cities in present-day Jordan (Jerash, Mt. Nebo, Areopolis [Rabbath], el-Lejjun, and Petra) is based on archaeological data presented in Russell (1985). In the conclusion of her discussion of the 551 C.E. Beirut earthquake, Guidoboni (1994, p. 336) writes: “It is therefore very likely that the surrounding regions (Arabia, Mesopotamia, Palestine and Syria) mentioned by Malalas and Theophanes, were either subject to secondary effects or to after-shocks with different epicenters.” This leaves open the possibility that historical accounts, although focusing specifically on the major damage in 551 C.E. along the important commercial centers between Beirut and Tripoli, allude to other potential seismic damage farther afield. Were there other local source moderate earthquakes within this decade of 551–561 C.E. centered in Palestine? Guidoboni (1994) describes an account by the chronicler Agathias who personally experienced an earthquake in Alexandria on 15 October 554 C.E. Agathias relates that earthquakes were uncommon in this region, and the population did not build to withstand strong earthquake ground motion. No damage is described for

100

Rucker and Niemi

Figure 2. (A) Felt area of 551 C.E. Levant earthquake (after Guidoboni, 1994). (B) Epicenter of 551 C.E. Levant earthquake (after Ambraseys et al., 1994).

Geological Data

this seismic tremor in Egypt. Could the shaking in Alexandria in 554 C.E. have been caused by an earthquake with an epicentral region located somewhere along the Dead Sea transform in Byzantine Palestine? One must ask, why has a mid-sixth-century layer of destruction been described for sites from northern to southern Jordan? Did the 9 July 551 C.E. earthquake actually cause major damage away from the Lebanese coast, contrary to the historical accounts? Was there a different local earthquake source that was largely contemporaneous to 551 C.E. that has been conflated with it? What exactly does the archaeological data really tell us about seismic activity in Jordan and the Levant in the mid-sixth century C.E.?

Darawcheh et al. (2000) recognized the discrepancy between the two divergent epicentral locations for the 551 C.E. earthquake—one offshore of the Lebanese coast (Guidoboni, 1994), and the other in the Jordan Valley (Ambraseys et al., 1994). A reevaluation of the evidence from primary and secondary Byzantine sources and an assessment of seismic parameters of the 551 C.E. Lebanese earthquake thus seemed warranted. From the available macroseismic data, Darawcheh et al. (2000) designated the Roum strike-slip fault and its potential offshore extension as the likely seismic source and calculated a surface-wave earthquake magnitude of Ms = 7.1–7.3 for the 551 C.E. earthquake. These authors also suggested that assigning earthquake damage to archaeological sites in western Jordan “may be an interpretative error.” New offshore seafloor mapping by Elias et al. (2007) identified the source rupture of the 551 C.E. earthquake as an active low-angle thrust fault located along the Lebanese coast (Fig. 3). This east-dipping Mount Lebanon thrust fault is exposed along the coast south of Beirut, where a beach terrace has been elevated by ~80 cm. Radiocarbon analyses of shells on the uplifted beach terrace date to the sixth century C.E. (Morhange et al., 2006). Comparison of the 551 C.E. uplift and older marine terraces on Mount Lebanon suggests that the repeat time of this type of earthquake may range between 1500 and 1750 yr. Elias et al. (2007) suggested a moment magnitude (Mw) of 7.5 for the 551 C.E. seismic event. Archaeological Data It is very clear that Kenneth Russell’s 1985 summary article entitled “The Earthquake Chronology of Palestine and Northwest Arabia from the Second through the Mid-Eighth Century A.D.”

Historical earthquake catalogues and archaeological data

A B

Figure 3. (A) Epicenter of the 551 C.E. earthquake along the offshore Mount Lebanon thrust fault. Abbreviations: Ar.—Arqa; Ba.—Batroun; By—Byblos; Ch.—Chekka; Sa—Sarafand. AT—Aakkar thrust; TT—Tripoli thrust; R-AF—Rankine-Aabdeh fault; RF—Roum fault; RaF—Rachaya fault; SaF—Saida fault. Image from Elias et al. (2007). (B) Example of the elevated beach terrace near Tripoli (from Elias et al., 2007).

has had a profound effect on both the entries in the regional earthquake catalogues and on defining the stratigraphy at archaeological sites in this region. This may largely be due to the fact that the publication was in a journal widely read by archaeologists. Two commonly used earthquake catalogues at the time prior to Russell (1985) were Amiran (1950–1951, 1952) and Ben-Menahem (1979).

101

Four archaeological sites in Jordan were cited by Russell (1985) supporting the interpretation that the 551 C.E. earthquake damaged the region in present-day Jordan (Mt. Nebo, Jerash, el-Lejjun, and Petra). In Petra, Russell was part of the University of Utah’s excavation team at the Temple of the Winged Lions, where they purportedly found evidence for a mid-sixth-century collapse and possible earthquake. There are very few published data on this site available for us to understand the chronological basis of this interpretation. Russell’s (1985) evidence of earthquake damage at Jerash is based on Crowfoot (1938), Mount Nebo from Saller and Bagatti (1949), and el-Lejjun from Parker (1982). We are not in a position to evaluate the archaeological evidence at Mt. Nebo or Jerash. However, recent published excavation reports from Petra and el-Lejjun do allow us to evaluate the evidence at those two sites. Russell in 1985 (p. 45) wrote “Petra, the capital of Palestina tertia, was never rebuilt after the 551 C.E. earthquake, and by the end of the sixth century C.E., its ruins had become a quarry for liming and smelting operations.” However, recent excavations at the Petra Church archaeological site refute these conclusions (Fiema, 2001a, 2001b). Scrolls found in the Petra Church provide an unprecedented record of Late Byzantine Petra (Fiema, 2002). The church was destroyed in a fire at the end of the sixth or the beginning of the seventh century C.E. The fire carbonized papyrus scrolls that were being stored in the church. These scrolls, known as the Petra Papyri, are a collection of documents predominantly relating to taxes and property ownership, dating from 537 C.E. to at least 13 April 593 C.E., and may postdate this range by several years. Therefore, the last recorded date of the Petra papyri scrolls may extend to 597 C.E. “Neither the effects of the earthquake of 551 C.E. nor the mid-sixth century C.E. plague can be discerned from the texts” of the scrolls (Fiema, 2002, p. 4). After the fire and into the seventh century C.E., the church ceased to function as an ecclesiastical structure, building materials were salvaged for reuse, and the shell of the structure was converted to a domestic complex. Fiema (2001a, 2001b) noted evidence for two earthquakes in the later phases of the Petra Church— one in the seventh century C.E. and one in the medieval to Ottoman period—at which time, no columns remained standing. As recounted already, excavations in the 1990s at the Petra Church and textual evidence from the newly translated Petra Papyri have convincingly demonstrated that the city of Petra was not apparently appreciably affected by the 551 C.E. earthquake. Unfortunately, some excavators still designate a 551 C.E. earthquake in the stratigraphic sequence at Petra. The site of el-Lejjun, a fourth-century Roman legionary fort located east of the Dead Sea, excavated by Parker (2006) over five seasons from 1980 to 1988, is in some ways an example of the way in which the correlation of the several lines of evidence should work (Fig. 4). Parker (2006) found evidence of a collapse horizon that contained coins, the most recent of which dated to 540–541 C.E. Admittedly, this does support an interpretation that this collapse horizon may be dated to the historically recorded earthquake of 551 C.E. It is possible that this is entirely correct,

102

Rucker and Niemi

Jerusalem

although unlikely based on the majority of the other evidence. It is extremely important to note, however, that Parker’s evidence in no way precludes a slightly more recent date for the destruction layer at el-Lejjun. Based on Parker’s excavations at el-Lejjun, Amiran et al. (1994) lists el-Lejjun as destroyed in 551 C.E. Thus, this piece of only archaeological evidence, which is only partially corroborated, enters the historical record as fact. What should be a fruitful cross-pollination (and might still be in this specific case) has brought into question the integrity of both lines of evidence.

Madaba

Hebron

Diban

Dead Sea

Epigraphic Data and/or Earthquakes Not in the Catalogue?

Areopolis En Boqeq

Kerak

Betthorus (el-Lejjun)

Zoara (Safi)

Tamar

31°00’

Praesidium (Fifa) Toloha (Telah)

Mampsis

rabah

Hatzeva

Wadi 'A

Feinan

Petra Bir Madkhur

Udruh

Areopolis, modern-day Ar-Rabbath (Fig. 4), located east of the Dead Sea and west of el-Lejjun, is listed as one of the cities damaged in the earthquake of 9 July 551 C.E. in the earthquake catalogue of Amiran et al. (1994). Zayadine (1971) published the translation of a dedicatory inscription found at Areopolis that states “Restored in 492 (597–598 C.E.) after the earthquake” (Fig. 5). This block was found out of context, reused in a structure, and interpreted as referring to a previously unattested earthquake (Zayadine, 1971). Russell (1985) suggested that there was a long (46 yr) time lag between the 551 C.E. earthquake and the reconstruction of Areopolis. The long recovery time was postulated to be due to “a depressed economic environment” (Russell, 1985, p. 50). This suggestion, that the general economic depression of the region caused a 46 yr delay in reconstruction after the earthquake, is unlikely for several reasons. First, earthquake reconstruction typically takes place soon after the event or not at all. For example, the Byzantine Emperor Justinian provided funds for the immediate rebuilding of the cities of

Ma'an

Gharandal 30°00’

Humeima

Gulf of Aqaba

0

N

Aila ('Aqaba)

Via

No va

Tra

ian

a

Yotvata

30 km

35°30’

Figure 4. Map of southern Jordan showing the sites and roads during the Roman period. (Modified from Talbert [2000], p. 71 and 76).

Figure 5. A dedicatory inscription found in secondary context at the ancient site of Areopolis, east of the Dead Sea, which reads: “Restored in 492 [597–598 C.E.] after the earthquake” (from Zayadine, 1971).

Historical earthquake catalogues and archaeological data the Lebanese coast after 551 C.E. (Russell, 1985, p. 45). When Pompeii, was entombed by ash from the eruption of Mount Vesuvius in 79 C.E., damage from the devastating earthquake of 62 C.E. had mostly been repaired. San Francisco was rebuilt and able to host the 1915 Panama Pacific International Exposition after the1906 earthquake and fire. As we see in the aftermath of the 2010 earthquakes in Haiti, Chile, and Mexico, societies both modern and ancient quickly repair earthquake-damaged areas (Sintubin et al., 2008). Second, if it were to occur at such a chronological distance as 46 yr after the event, it is unlikely that a dedicatory inscription would mention the earthquake. Average life expectancy would suggest that there would be relatively few people still around who remembered the event. Is it possible that the earthquake that destroyed Areopolis was a later event that is not chronicled by historical text? Since the Petra Papyri do not date beyond 593–597 C.E., could an earthquake around 597 C.E. have also caused damage in Petra? Do archaeoseismic and paleoseismic methodologies have the ability to discern between a 551 C.E. and 597 C.E. earthquake? It is interesting to note that the earthquake catalogue of Ambraseys (2009, p. 216– 217) now includes an earthquake entry for a seismic event at “
103

it creates a larger sample size, and allows for checks on otherwise singular lines of evidence. Noller (2001) presented further case studies in archaeoseismology, mostly of prehistoric sites (thus with no need to integrate the historical line of evidence). He particularly emphasized the use of archaeological sites and the features thereof as piercing points on active faults (e.g., Marco et al., 1997; Ellenblum et al., 1998; Meghraoui et al., 2003; Rucker and Niemi, 2005; Haynes et al., 2006). These can be one of the clearest types of archaeological evidence, as well as one of the most informative. Marco (2008) published a useful list of the types of seismogenic damage often encountered in archaeological excavation, as well as noting (in passing) the potential for circular reasoning. There is a temptation to utilize the historical earthquake catalogue as a means by which uncertain stratigraphic layers and archaeological dates become more certain when the date of an earthquake is assumed. We have outlined one example of the process of circular reasoning using the Levant earthquake of 551 C.E., whereby archaeologists at Petra assigned a collapse horizon to the 551 C.E. earthquake and the earthquake cataloguer included an entry for the 551 earthquake for the complete destruction and abandonment of Petra in 551 C.E. We now know from the excavations at the Petra Church and documents in the Petra Papyri (Fiema, 2002) and other Byzantine structures within Petra that occupation was continuous throughout most of the sixth century C.E. We are not the first to point out the potential pitfalls of overreliance on the limited dates of earthquakes in a catalogue in assigning archaeological ages. Ambraseys (2005, 2006) in his discussion of a Jericho earthquake at the time of the Biblical story of Joshua entering the “Promised Land” describes the circular process where an uncertain date of an alleged earthquake becomes part of the stratigraphic and archaeological dating process. This type of corruption of the archaeological data diminishes its usefulness for scientific studies. Galadini et al. (2006) also argued for a systematic and careful combination of the various lines of evidence. They emphasized the need for continual examination of methodology in archaeoseismology to control the role of the different kinds of data involved. Guidoboni (1996) in her article (p. 7) cautioned, “…it has not been possible up to now to evaluate the methodological correctness with which hypotheses about seismic effects and their importance have been reached in the course of an excavation. If the procedures followed are not scientifically transparent and correct, the results cannot be used in a scientific context.” Pseudoobjectivity, a term that Ambraseys (2006) used for the collusion of historical earthquake and archaeological data, is a problem of bias that needs to be avoided. There are several different types of data found at archaeological sites that are interpreted as earthquake damage. They include structural damage to standing buildings, repairs and modifications to structures, and stratigraphic horizons that denote collapse and destruction. Each has strengths and limitations as to the seismic information that can be collected.

104

Rucker and Niemi

Damage to Extant Structures This category includes a wide variety of structural failures that appear in standing or partially standing buildings. Cracks in masonry and across mortar, shifted and rotated building blocks, and slipped keystone blocks are examples of the types of building failures that have been interpreted as seismically induced damage (e.g., Stiros, 1996). There are severe limitations to interpretation of extant structure. Structural failure to buildings occurs naturally under the influence of gravity and the weight of the structure. Much of the structural damage can be attributed to natural processes of differential subsidence and settling without invoking seismic excitation. Extant structures in seismically active areas often experience ground shaking from multiple earthquakes over the life of the structure. Therefore, it is not possible to constrain the age of earthquake damage closer than after the construction of the building. However, similar to precariously balanced rocks (e.g., Brune, 2002), an extant historical structure that does not have damage can be a useful indicator for estimating the minimal strong earthquake ground motion over the life of the building, if there is a clear understanding of the structure’s maintenance history. Repaired or Modified Structures Archaeologists often identify various phases in a building’s use based on repairs and modification to the building plan or structure. Much of what is excavated is the lower courses of walls, floors, or living surfaces, and foundations of the walls. Revetments are buttress walls placed on the exterior of a wall to help support it. Blocked doorways, additions, and changed use of rooms are common alterations seen in ancient building plans. Revetments and blocked doorways that seal off collapsed structures are both used to interpret repairs after an earthquake. The date of the earthquake can be interpreted from the ceramic evidence, coins, and radiocarbon dating of organic material on the surface beneath the rubble or the revetment. Collapsed Structures and Destruction Horizons Earthquake collapse horizons often fall onto living surfaces and occasionally have trapped victims, as is the case for the 363 C.E. earthquake at Petra (Stucky, 1990) and the 749 C.E. earthquake at Pella in Jordan (Smith and Day, 1989). Collapse horizons that preserve whole pots in situ or smashed and scattered whole pots provide the best evidence for an instantaneous collapse of a building, which is often interpreted as evidence for an earthquake. When this horizon is in a sealed context (i.e., a floor or layer caps the horizon), it provides the best means of dating the event. Instantaneous collapse of structures and destruction horizons can also be caused by warfare or conflagrations, which is the main competing interpretation for this type of archaeological evidence.

Because an earthquake is a brief instant in time, excavation of an earthquake collapse horizon at an archaeological site provides a unique opportunity to see the association of various features and artifacts at that moment in time. In this way, an earthquake collapse horizon is similar to a shipwreck, which also provides a snapshot of a moment in time. Therefore, an historically recorded earthquake with a known date can be an extremely powerful tool for archaeology. However, the temptation to quickly attach such a known date to an unknown collapse horizon must be resisted unless there is clear evidence linking the two. Ultimately, the investigation of past earthquakes from any source helps us to develop a more informed assessment of the seismic hazards of a region. The location of active faults, the repeat time of earthquakes, the maximum expected earthquake magnitude, and the types of ground motion often measured as peak ground acceleration are used to define seismic zonations and building codes. Historical seismology, epigraphy, archaeoseismology, and paleoseismology are all disciplines that can provide vital data to assess and mitigate seismic hazards. To be useful in these terms, data need to be defined numerically and with a stated level of uncertainty. In achieving a synthesis of these lines of evidence, we suggest that a comprehensive catalog or database be created. Currently, the earthquake catalogs are lists of textual evidence with archaeological and other data sometimes grafted onto them. They could become truly multidisciplinary tools. There would be two main differences from the current practice. First, every aspect (epicenter, felt area, magnitude, etc.) of every entry would be evaluated for degree of uncertainty. None of these lines of evidence would get a “free ride”—all would require supporting evidence. Second, all individual bits of evidence would now be entered into the catalog. Since these bits of evidence would be clearly marked as to their degree of uncertainty, something approaching objectivity would gradually develop as the database grows. It may be possible, for example, that when sixth-century C.E. destruction layers stop being shoehorned into fitting the 551 C.E. earthquake, a consistent pattern showing another date, and thus another previously unrecorded seismic event, will reveal itself. CONCLUSIONS Since any given ancient earthquake may be represented in one or more of four lines of evidence, historic texts, epigraphy, archaeological structural damage, and fault ruptures, it is critical to remember that any given ancient earthquake may be totally unrepresented in any or all of these categories. Thus, the temptation to force the data from one category into matching data from another must be avoided—correlations must be real, not forced. In this paper, we argue that it is actually simpler to postulate an earthquake on or around 597 C.E. that is unrepresented in the historic sources, than it is to force the epigraphic and archaeological evidence to fit the historically recorded 551 C.E. earthquake.

Historical earthquake catalogues and archaeological data Archaeological data have not been systematically used nor critically evaluated as an independently dated source of historical seismic data. Because of some of the past problems of overreliance on the historical earthquake catalogues, the archaeological data must be reevaluated using new chronological tools. Extensive excavations and significant advances in dating archaeological strata over the past decade have now set the stage for the compilation of a regional historical earthquake database. As another important role for this research, if the function of a list of earthquakes and their intensity (i.e., the earthquake catalogue) is to be used for seismic hazard analyses, then all probable earthquakes, their source of documentation, and estimations both of the uncertainty of identification of the event and the uncertainty in the age determination need to be clearly stated and in a form that can be used by a multidisciplinary set of users. This is exactly the new form of database suggested herein. ACKNOWLEDGMENTS This paper was originally presented at the “Archaeoseismological Methodologies: Principles and Practices,” session organized by IGCP 567 at the annual meeting of the Seismological Society of America in Santa Fe, New Mexico, in 2008. We thank S. Marco and an anonymous reviewer for helpful comments. This article is a contribution to the United Nations Educational, Scientific and Cultural Organization–funded International Geoscience Programme IGCP 567, “Earthquake Archaeology: Archaeoseismology along the Alpine-Himalayan Seismic Zone.” REFERENCES CITED Ambraseys, N., 1971, Value of historical records of earthquakes: Nature, v. 232, p. 375–379, doi: 10.1038/232375a0. Ambraseys, N., 2005, Archaeoseismology and neocatastrophism: Seismological Research Letters, v. 76, no. 5, p. 560–564, doi: 10.1785/gssrl.76.5.560. Ambraseys, N., 2006, Earthquakes and archaeology: Journal of Archaeological Science, v. 33, no. 7, p. 1008–1016, doi: 10.1016/j.jas.2005.11.006. Ambraseys, N., 2009, Earthquakes in the Mediterranean and Middle East: A Multidisciplinary Study of Seismicity up to 1900: Cambridge, UK, Cambridge University Press, 968 p. Ambraseys, N., Melville, C.P., and Adams, R.D., 1994, The Seismicity of Egypt, Arabia and the Red Sea; A Historical Review: Cambridge, UK, Cambridge University Press, 181 p. Amiran, D.H.K., 1950–1951, A revised earthquake catalogue of Palestine: Israel Exploration Journal, v. 1, p. 223–245. Amiran, D.H.K., 1952, A revised earthquake catalogue of Palestine: Israel Exploration Journal, v. 2, p. 48–62. Amiran, D.H.K., Arieh, E., and Turcotte, T., 1994, Earthquakes in Israel and adjacent areas—Macroseismic observations since 100 B.C.E.: Israel Exploration Journal, v. 44, no. 3–4, p. 260–305. Ben-Menahem, A., 1979, Earthquake catalogue for the Middle East (92 BC–1990 AD): Bollettino di Geogisica Teorica e Applicata, v. 21, p. 245– 310. Ben-Menahem, A., 1991, Four thousand years of seismicity along the Dead Sea rift: Journal of Geophysical Research, v. 96, no. B12, p. 20,195– 20,216, doi: 10.1029/91JB01936. Brune, J.N., 2002, Precarious-rock constraints on ground motion from historic and recent earthquakes in Southern California: Bulletin of the Seismological Society of America, v. 92, no. 7, p. 2602–2611, doi: 10 .1785/0120000606.

105

Crowfoot, J.W., 1938, The Christian churches, in Kraeling, C.H., ed., Gerasa: City of the Decapolis: New Haven, Connecticut, American Schools of Oriental Research, p. 171–262. Darawcheh, R., Sbeinati, M.R., Margottini, C., and Paolini, S., 2000, The 9 July 551 AD Beirut earthquake, Eastern Mediterranean region: Journal of Earthquake Engineering, v. 4, no. 4, p. 403–414, doi: 10.1142/ S1363246900000229. Elias, A., Tapponnier, P., Singh, S.C., Kin, G.C.P., Briasis, A., Daëron, M., Carton, H., Sursock, A., Jacques, E., Jomaa, R., and Klinger, Y., 2007, Active thrusting offshore Mount Lebanon: Source of the tsunamigenic A.D. 551 Beirut-Tripoli earthquake: Geology, v. 35, no. 8, p. 755–758, doi: 10.1130/G23631A.1. Ellenblum, R., Marco, S., Agnon, A., Rockwell, T., and Boas, A., 1998, Crusader castle torn apart by earthquake at dawn, 20 May 1202: Geology, v. 26, p. 303–306, doi: 10.1130/0091-7613(1998)026<0303:CCTABE >2.3.CO;2. Fiema, Z.T., 2001a, Reconstructing the history of the Petra Church: Dating and phasing, in Bikai, P.M., ed., The Petra Church: Amman, American Center of Oriental Research, p. 6–137. Fiema, Z.T., 2001b, The archaeological context of the Petra Papyri, in Bikai, P.M., ed., The Petra Church: Amman, American Center of Oriental Research, p. 138–150. Fiema, Z.T., 2002, Introduction: Historical context, in Frösén, J., Arjava, A., and Lehtinen, M., eds., The Petra Papyri I: Amman, American Center of Oriental Research, p. 1–4. Galadini, F., Hinzen, K.-G., and Stiros, S., 2006, Archaeoseismology: Methodological issues and procedure: Journal of Seismology, v. 10, p. 395–414, doi: 10.1007/s10950-006-9027-x. Guidoboni, E., 1994, Catalogue of Ancient Earthquakes in the Mediterranean Area up to the 10th Century: Rome, Istituto Nazionale di Geofisica, 504 p. Guidoboni, E., 1996, Archaeology and historical seismology: The need for collaboration in the Mediterranean area, in Stiros, S., and Jones, R.E., eds., Archaeoseismology: Athens, British School at Athens, Fitch Laboratory Occasional Paper 7, p. 7–14. Guidoboni, E., and Comastri, A., 2005, Catalogue of Earthquakes and Tsunamis in the Mediterranean area from the 11th to the 15th Century: Bologna, Italy, Instituto Nazionale di Geofisica e Vulcanologia-Secretariat for Geological Agency, 1037 p. Guidoboni, E., and Ebel, J., 2009, Earthquakes and Tsunamis in the Past: Cambridge, UK, Cambridge University Press, 590 p. Guidoboni, E., Muggia, A., and Vallensise, G., 2000, Aims and methods in territorial archaeology: Possible clues to a strong fourth-century AD earthquake in the Straits of Messina (southern Italy), in McGuire, W.J., Griffiths, D.R., Hancock, P.L., and Stewart, I., eds., The Archaeology of Geological Catastrophes: Geological Society of London Special Publication 171, p. 45–70. Haynes, J., Niemi, T.M., and Atallah, M., 2006, Evidence for ground-rupturing earthquakes on the Northern Wadi Arabah fault at the archaeological site of Qasr Tilah, Dead Sea transform fault system, Jordan: Journal of Seismology, v. 10, p. 415–430, doi: 10.1007/s10950-006-9028-9. Jones, R.E., and Stiros, S., 2000, The advent of archaeoseismology in the Mediterranean, in McGuire, W.J., Griffiths, D.R., Hancock, P.L., and Stewart, I., eds., The Archaeology of Geological Catastrophes: Geological Society of London Special Publication 171, p. 25–32. Karcz, I., 2004, Implications of some early Jewish sources for estimates of earthquake hazard in the Holy Land: Annals of Geophysics, v. 47, p. 759– 792. Karcz, I., and Kafri, U., 1978, Evaluation of supposed archaeoseismic damage in Israel: Journal of Archaeological Science, v. 5, p. 237–253, doi: 10.1016/0305-4403(78)90042-0. Karcz, I., and Kafri, U., 1981, Studies in archaeoseismicity of Israel: Hisham’s palace, Jericho: Israel Journal of Earth Sciences, v. 30, p. 12–30. Karcz, I., Kafri, U., and Meshel, Z., 1977, Archeological evidence for subrecent seismic activity along the Dead Sea–Jordan Rift: Nature, v. 269, p. 234–235. Marco, S., 2008, Recognition of earthquake-related damage in archaeological sites: Examples from the Dead Sea fault zone: Tectonophysics, v. 453, p. 148–156, doi: 10.1016/j.tecto.2007.04.011. Marco, S., Agnon, A., Ellenblum, R., Eidelman, A., Basson, U., and Boas, A., 1997, 817-year-old walls offset sinistrally 2.1 m by the Dead Sea trans-

106

Rucker and Niemi

form, Israel: Journal of Geodynamics, v. 24, p. 11–20, doi: 10.1016/ S0264-3707(96)00041-5. Meghraoui, M., Gomez, F., Sbeinati, R., Van der Woerd, J., Mouty, M., Darkal, A.N., Radwan, Y., Layyous, I., Al Najjar, H., Darawcheh, R., Hijazi, F., Al-Ghazzi, R., and Barazangi, M., 2003, Evidence for 830 years of seismic quiescence from paleoseismology, archaeoseismology, and historical seismicity along the Dead Sea fault in Syria: Earth and Planetary Science Letters, v. 210, p. 35–52, doi: 10.1016/S0012-821X(03)00144-4. Morhange, C., Pirazzoli, P.A., Marriner, N., Montagioni, L.F., and Nanmour, T., 2006, Late Holocene relative sea-level changes in Lebanon, Eastern Mediterranean: Marine Geology, v. 230, p. 99–114, doi: 10.1016/j .margeo.2006.04.003. Noller, J.S., 2001, Archaeoseismology: Shaking out the history of humans and earthquakes, in Goldberg, P., Holliday, V.T., and Ferring, C.R., eds., Earth Sciences and Archaeology: New York, Kluwer Academic/Plenum Publishers, p. 143–170. Parker, S.T., 1982, Preliminary report of the 1980 season of the Central Limes Arabicus Project: Bulletin of the American Schools of Oriental Research, v. 247, p. 1–26. Parker, S.T., 1987, The Roman Frontier in Central Jordan: Interim Report on the Limes Arabicus Project, 1980–1985 (2 volumes): Oxford, British Archaeological Reports International Series, v. 340, 617 p. Parker, S.T., 2006, The Roman Frontier in Central Jordan: Final Report on the Limes Arabicus Project, 1980–1989: Washington, D.C., Dumbarton Oaks Research Library and Collection, 1104 p. Rapp, G., 1986, Assessing archaeological evidence for seismic catastrophes: Geoarchaeology, v. 1, no. 4, p. 365–379, doi: 10.1002/gea.3340010403. Rucker, J.D., and Niemi, T.M., 2005, New excavations of the city wall at Islamic Ayla in ‘Aqaba, Jordan: Annual of the Department of Antiquities of Jordan, v. 49, p. 501–508. Russell, K.W., 1980, The earthquake of May 19, A.D. 363: Bulletin of the American Schools of Oriental Research, v. 238, p. 47–64, doi: 10.2307/ 1356515. Russell, K.W., 1985, The earthquake chronology of Palestine and northwest Arabia from the 2nd through the mid-8th century A.D.: Bulletin of the

American Schools of Oriental Research, v. 260, p. 37–59, doi: 10.2307/ 1356863. Salamon, A., Rockwell, T., Ward, S.N., Guidoboni, E., and Comastri, A., 2007, Tsunami hazard evaluation of the Eastern Mediterranean: Historical analysis and selected modeling: Bulletin of the Seismological Society of America, v. 97, no. 3, p. 705–724, doi: 10.1785/0120060147. Saller, S.J., and Bagatti, B., 1949, The town of Nebo (Khirbet El-Mekhayyat) with a brief survey of the ancient Christian monuments in Transjordan: Publications of the Studium Biblicum Franciscanum, no. 7: Jerusalem, Franciscan Press. Sbeinati, M.R., Darawcheh, R., and Mouty, M., 2005, The historical earthquakes of Syria: An analysis of large and moderate earthquakes from 1365 B.C. to 1900 A.D.: Annals of Geophysics, v. 48, p. 347–435. Sintubin, M., Stewart, I., Niemi, T.M., and Altunel, E., 2008, Earthquake archaeology, just a good story?: Seismological Research Letters, v. 79, no. 6, p. 767–768, doi: 10.1785/gssrl.79.6.767. Smith, R.H., and Day, L.P., 1989, Pella of the Decapolis, Volume 2: Final Report on the College of Wooster Excavations in Area IX, The Civic Complex, 1979–1985: Wooster, Ohio, The College of Wooster, 168 p. Stiros, S.C., 1988, Archaeology, a tool to study active tectonics—The Aegean as a case study: Eos (Transactions, American Geophysical Union), v. 69, no. 50, p. 1633, 1639. Stiros, S.C., 1996, Identification of earthquakes from archaeological data: Methodology, criteria and limitations, in Stiros, S., and Jones, R.E., eds., Archaeoseismology: Athens, British School at Athens, Fitch Laboratory Occasional Paper 7, p. 129–152. Stucky, R.A., 1990, Schweizer Ausgrabungen in Ez Zantur, Petra Vorbericht der Kampagne 1988: Annual of the Department of Antiquities of Jordan, v. 34, p. 249–283. Talbert, R.J.A., editor, 2000, Barrington Atlas of the Greek and Roman World: Princeton, New Jersey, Princeton University Press, 102 p. Zayadine, F., 1971, Un séisme à Rabbat Moab (Jordanie) d’après une inscription Grecque du VIe S: Berytus, v. 20, p. 139–141. MANUSCRIPT ACCEPTED BY THE SOCIETY 19 MAY 2010

Printed in the USA

The Geological Society of America Special Paper 471 2010

Historical earthquakes in Srinagar, Kashmir: Clues from the Shiva Temple at Pandrethan Roger Bilham Cooperative Institute for Research in Environmental Science and Department of Geological Sciences, University of Colorado, Boulder, Colorado 80309-0399, USA Bikram Singh Bali M. Ismail Bhat University of Kashmir, Srinagar 190 006, Jammu and Kashmir, India Susan Hough U.S. Geological Survey 525 South Wilson Avenue, Pasadena, California 91106, USA

ABSTRACT Srinagar, the capital city of Kashmir, has been shaken numerous times by earthquakes in the past millennium, most recently by damaging earthquakes in 1885 (M 6.2, 30 km to the west) and 2005 (M 7.6, 200 km to the west) with estimated EMS (European Macroseismic Scale) intensity VI–VII. Earthquakes in Kashmir in earlier historical times are known only from fragmentary archival sources. We present and analyze unique, repeat photographs of the Pandrethan Temple near Srinagar, which we conclude can provide clues to the severity of nineteenth-century earthquakes. Photos taken in 1868 and 1885 and recently show that the temple, a 5.5-m-square masonryblock structure constructed ca. A.D. 920, was undamaged by these two earthquakes. We conclude that displaced blocks visible in the earliest extant photograph are the result of stronger shaking in the past, the most probable causal earthquake being in 1828. Considering the fragility of the structure, we conclude that anything greater than EMS intensity IX would have caused structural collapse. We thus conclude that Pandrethan has not experienced EMS intensity greater than VIII in the past 200 yr, and possibly not in the past millennium.

INTRODUCTION: THE TEMPLE The Shiva Temple at Pandrethan, at 74.860°E, 34.056°N, ∼3 km east of Srinagar, Kashmir, is believed to have been constructed ca. A.D. 913–921. Visitors describing the temple have used various spellings in their accounts: Pándenthán (Moorcroft and Trebeck, 1841), Pandriton (Hügel, 1845), Pandrenton (Vigne, 1844), Pandrynton (Temple, 1887), and Pandrinton/Padrenton

(Hunter, 1881). It consists of a symmetrical stone structure measuring 5.5 m square aligned N20W with a portal on each side. Its stone-block pyramidal roof is interrupted by an overhanging step and four small windows. Tradition has it that for religious reasons, the temple was erected at the center of a pond fed by a natural spring. Kak (1933), however, suggests that the temple was originally constructed in a swamp that has been drained, and that the temple may have been constructed as late as the early

Bilham, R., Bali, B.S., Bhat, M.I., and Hough, S., 2010, Historical earthquakes in Srinagar, Kashmir: Clues from the Shiva Temple at Pandrethan, in Sintubin, M., Stewart, I.S., Niemi, T.M., and Altunel, E., eds., Ancient Earthquakes: Geological Society of America Special Paper 471, p. 107–117, doi: 10.1130/2010.2471(10). For permission to copy, contact [email protected]. © 2010 The Geological Society of America. All rights reserved.

107

108

Bilham et al.

twelfth century. The temple now tilts at ~5° as a result of uneven settlement in the past 1000 yr. The temple was originally part of a much larger complex. The old city of Pandrethan is believed to have been founded by Asoka (ca. 200 B.C.), but to have been abandoned ca. A.D. 500 (Kak, 1933). In the tenth century it was reconstructed by a former Kashmiri King but burned down shortly thereafter. By the time of the Mughal administration (sixteenth century), most of its building materials had been scavenged for the construction of the new capital of Srinagar (Bernier, 1891; Temple, 1887). Stone relicts of early Buddhist stupas have surfaced from time to time, but the archaeological site is now covered by the Badami Bagh military cantonment, which has obliterated surface evidence of the former extent of the old city. The temple itself lies hidden from public view within the walls of the cantonment. In Figure 1, we show the 1885 earthquake epicentral region with smoothed EMS (European Macroseismic Scale) intensity VI and VII contours to the east and west of Pandrethan. The 1885 earthquake was assessed at Mw = 6.3 by Ambraseys and Douglas (2004), and as Mw = 6.2 by Szeliga et al. (2010). We visited Pandrethan in early December 2008 and again in June 2009 to learn what could be deduced about former earthquake shaking from damage to the temple. However, extensive repairs

have been undertaken in recent years, leaving few clues about former damage. Fortunately, the structure was photographed both before and after the 1885 earthquake, and before preservation was attempted. The temple was constructed from close-fitting dressed limestone blocks, with no evidence for cement between courses. Four squat, equally spaced columns support a hollow pyramidal roof of tapered blocks (Fig. 2). It is not known whether pins hold the columns in alignment, although it is possible, because the practice is evident in the pillars of the Martand Sun temple, which was constructed at the same time using similar architectural features. The rectangular and trapezoidal roof blocks were probably not doweled, since they are smaller and ornamental in function. The recent repairs and realignment of the roof blocks since Kak’s visit in 1933 have incorporated liberal quantities of cement. The ceiling of the interior described by Cole (1869) and Kak (1933) is surfaced by three layers of stone blocks that structurally serve to hold the walls together. Large triangular blocks first cover the stout corner walls and overhang the interior corners. The resulting enclosed diagonal square space is overlain by blocks parallel to the sides of the structure, and these in turn are covered by a single ceiling block. The lower surfaces of these ceiling layers are embellished with carvings (Fig. 2).

Figure 1. Kashmir Valley shaking intensities (red—1885; blue—2005). EMS (European Macroseismic Scale) intensities are shown as assessed by Martin and Szeliga (2010). The 1885 earthquake is believed to have been Mw 6.2, roughly 50 km to the west of Pandrethan, resulting in EMS VI near the temple. Intensities from the 2005 Mw 7.6 Kashmir earthquake, which occurred 200 km west of Pandrethan, were VI to VII. Areas in green indicate epicentral regions.

Historical earthquakes in Srinagar, Kashmir

A

2m

In 1665, François Bernier (1891) almost certainly visited the temple during his visit to Kashmir with Aurangzeb. He mentions the numerous ruins near Srinagar but describes in detail only the larger temples in the valley. In 1823, Moorcroft and Trebeck (1841) visited the temple twice. On their second visit on 15 May, Trebeck, in the absence of a boat, swam inside to describe its decorated ceiling to Moorcroft. They were disappointed to find no inscriptions within the temple. In 1835, Baron Hügel (1845) contented himself with a view from the banks of its surrounding pond, speculating erroneously on Buddhist sculptures contained therein. In the same year, the temple was visited by Vigne (1844), who sketched it. According to Cunningham (1848), Elphinstone visited the temple in 1846 and discovered the interior coated with plaster. Cunningham (1848) had the plaster removed and made

109

Figure 2. (A) Plan and section through the Pandrethan temple after Kak (1933). Kak’s sketch of the decorated ceiling from below is placed on his plan view facing upward. The plinth is currently underwater and extends to an unknown depth. The summit pyramid has been lost and replaced by an artificial capstone. Cunningham (1848) suggested that the original may have consisted of a frieze supporting a third overhanging pyramid, as is evident in other Kashmir temples of similar age. (B) View of the decorated ceiling, June 2009, illustrating absence of impact damage.

a sketch of the internal ceiling, and the exterior setting of the temple. This outside view was reproduced by Fergusson (1867), who repeated Cunningham’s speculation that a third overhanging tier of triangular roof may have been lost. The fact that Trebeck noted the large lotus design at the center of the ceiling (Fig. 2) that was sketched and described by Cunningham suggests that the plastered layer was not complete, or not thick. In 1859, Richard Temple (1887) described and sketched the structure, but his sketches, like those of Vigne, were not included in his published accounts. In 1860, Knight (1863), lacking a boat, sketched Pandrethan from the banks of the pond. In 1868, the Pandrethan temple was photographed from the south by John Burke (Fig. 3) and is one of three views published by Cole (1869), who writes:

110

Bilham et al.

Figure 3. (A) Our earliest glimpse of Pandrethan (left) may be a sketch published by Cunningham (1848), here reproduced by Fergusson (1867), which shows a tree with abundant foliage sprouting from the roof. In the 18 July 1860 painting of Pandrethan from the north (Knight, 1863, p. 354), the internal tree is devoid of leaves (right). “The building which alone remained in at all a perfect state was situated in a sort of pond or tank of slimy green and was quite inaccessible without a boat.” He complains of mosquitoes plaguing his sketching and ants eating his paints. (B) Left: John Burke’s 1868 photo of the Pandrethan temple from the south, compared to approximately the same view in December 2008. The three upper roof courses were missing on the SE side in 1868, but are present, though dislodged on the NW side (Fig. 4). Note scale with 1 ft gradations in doorway on left.

The small village of Pandrethan is situated on the Jhelum, about a mile and a half to the south-east of Srinagar....The Temple is close to the village, and stands in the centre of a tank of water....At the time of my visit, the water was about two feet over the floor of the Temple, and I had to obtain a small boat to enable me and my surveyors to take measurements. The stone ceiling is elaborately carved in bas-relief figures, and it is one of the most perfect pieces of ancient carving that exists in Kashmir....The pyramidal roof is divided into two portions by an ornamental band. The corner pilasters are surmounted by carved capitals, and the pediments of the porches appear to have terminated with a melon-shaped ornament. The ceiling is formed of nine blocks of stone; four resting over the angles of the cornice, reduce the opening to a square, and an upper course of four stones still further reduces the opening, which is covered by a single block decorated with a large lotus. (p. 42)

The survival unscathed of Pandrethan’s ornately carved ceiling (Fig. 2) to the present suggests that total collapse of the

monument and subsequent reconstruction have not occurred, although one cannot exclude the possibility that the structure was damaged in early historical times and repaired by expert masons. The lower thirds of Oldham’s 1887 glass half-plate negatives have been damaged by water, and it was necessary to reassemble the broken halves of his NW view digitally. The date of the photo is recorded as August 1887 in the Calcutta photographic archive of the Geological Survey of India, and Oldham’s visit is recorded by Tom LaTouche in a letter from Srinagar to his mother dated 12 August 1887 (Bilham, 2008). The early sketches, the 1868 photograph, and two others reproduced by Cole (1869) show many of the same roof blocks in the temple displaced from their original positions. The tree that grew from, or through the roof in 1887, was evidently growing in 1848 and in 1860. The similarity between the 1868 and 1887 views suggest that shaking in 1885 was insufficient to cause further damage to this masonry structure, and hence that

Historical earthquakes in Srinagar, Kashmir accelerations in Pandrethan in 1885 were less than intensity VII, consistent with shaking estimates from other sources (Martin and Szeliga, 2010). No damage was done to the temple in 2005 when intensity VII shaking was recorded in nearby Srinagar. One is left to consider the damage to the temple that is evident in the 1868 photograph. The damage to the roof as well as the displaced blocks lower in the structure are, we conclude, more consistent with earthquake damage than vandalism or natural weathering. In particular, close inspection of the roof shows consistent offsets of the six lower courses of roofing blocks, no offset of the decorated overhang and sequentially increasing offsets of the five uppermost courses (Fig. 4), suggestive of oscillatory jostling. The uppermost blocks are perched precariously. Vandals, had they removed the missing pyramid, would surely have toppled this penultimate layer (Fig. 5). DISCUSSION Buildings that survive earthquakes may do so for several reasons: their structure and assembly may be unusually resilient to any form of shaking (e.g., the pyramids of Egypt), they may be isolated by their local seismic setting (a vibratory node or some form of base isolation), or they may have been reconstructed following damage. Thus, although it is tempting to consider Pandrethan a strong motion seismometer that has recorded the past thousand years of shaking in the Kashmir Valley, conclusions about the maximum severity of shaking are likely to be purely local at best, and completely wrong, if major repairs have been made following large earthquakes. The fact that the Pandrethan temple survived the two most recently damaging earthquakes in Kashmir with no apparent damage, however, provides us with an indication of its resilience to moderate (EMS intensity VI–VII) shaking. We examine details

111

of its construction and setting that relate to its vulnerability: what intensity caused the shaking recorded in nineteenth-century photographs, and what intensity of shaking would it take to destroy the temple? Construction Its masonry construction notwithstanding, the symmetry of the Pandrethan temple (Fig. 2), and its compact design appear to be well disposed to resist moderate shaking. The thickness and structure of its foundations are unknown, but though close to the hill, they are unlikely to extend to bedrock. The tilt of the temple is easy to recognize in 1859 and 1887 photographs relative to the surface of the pond. Although the base has settled, it has not flexed, nor is there any sign of differential settlement other than tilt and roof sag. Hence, the foundations have behaved monolithically, again a recognized feature of sound seismic resistant design. Geological and Historical Setting If the local site experiences significantly lower amplification than the surrounding region, then intensities in any earthquake are likely to have been locally lower than the overall intensity experienced in the region. Although there is no historical support for the notion, it is possible that the old city of Pandrethan was damaged by an earthquake, leading to the establishment of the new capital 3 km to the west. If so, the survival of the temple may imply that the local geology at the temple site is associated with lower site amplification than occurs in the old city of Pandrethan. Though sited 100 m from the Jhelum (Figs. 6 and 7), the temple has been constructed near the slope-break to the hills to the north. The spring that fed the tank, or pond, in which the temple is located, issues from the base of this hill.

Figure 4. Close-up of Oldham’s 1887 view of the temple from the NW. Note the parted SW gable (lower right) with its missing third block, also missing in Burke’s 1868 photo and Kak’s 1933 sketch. Note the irregular offsets to the SW of the six lower blocks, and sequentially increasing offset of the five upper blocks. The summit pyramid has been lost, and is now replaced by a modern dome with a triple-orbed pinnacle.

Figure 5. Pandrethan Temple now, and before and after the 1885 earthquake (Cole, 1869; Oldham, 1887). Lefthand panels are viewed from NW and righthand panels from NNW (see Fig. 6). Roof vegetation was recorded in a photograph by Marion Doughty (1902).

Figure 6. Google map of Pandrethan temple site showing view-directions of photos and proximity to the Jhelum.

Figure 7. The temple prior to repairs in 1901 (Doughty, 1902). A boat is visible in the foreground.

114

Bilham et al.

What Shaking Intensity Would Cause the Temple to Collapse? We speculate that the temple could probably be shaken with intensity VIII without collapse, based on the dissipative effects of its blocks, especially if the columns are held in place with alignment pegs. Intensity XI or X would probably cause collapse. The weight of the triangular roof would tend to bring down the entire structure if the building drifted significantly, and it is thus difficult to envisage partial collapse. Had collapse occurred we should expect to see damage to the stone blocks and ornamentation—splintered edges of blocks, or fractured corners. Some of the wall blocks are cracked, but the only substantial corner fracture damage seems to be to the roof overhang on the south side. The original summit pyramid is missing and this may have been projected southward during shaking, splintering the overhang as it descended. The ornately carved ceiling is apparently undamaged and is frequently cited as the best-preserved tenth-century stone carving in Kashmir (Kak, 1933; Bernier, 1997). We consider it unlikely that the stonework of the ceiling could have survived collapse and reconstruction, but we cannot exclude the possibility that the temple was refurbished by expert masons after an early historical earthquake. Several stone blocks in the roof remain misaligned. These, and those that were repositioned in recent repairs, testify to the building being shaken to the extent that many blocks were moving differentially during a former earthquake. The displaced blocks are found at high levels (in the roof) and not at lower levels near the pediment. Which Earthquake(s) Damaged the Temple? In a hand-written note on the 1887 protective envelope that contained his half-plate glass negative, R.D. Oldham states “Temple at Pandrethan, Srinagar, showing stones displaced by earthquakes.” Writing this 2 yr after the 30 May 1885 Kashmir valley earthquake, his use of the plural indicates that he does not ascribe damage to any single earthquake. He was, moreover, in the Andaman Islands at the time that the 1885 earthquake occurred. From earlier photographs, it is clear that the disposition of the uppermost blocks is identical before and after the May 1885 earthquake. The 1885 earthquake resulted in estimated EMS intensity VI shaking to the east and west of Pandrethan. We conclude that that the damage to the temple was the result of one or more earthquakes prior to 1868. Since the temple was unaltered by estimated intensity VII shaking in 2005, we deduce that the causal earthquake that shifted the summit blocks was caused by shaking with intensity VIII or IX. In Table 1, we list earthquakes known to have occurred in the Kashmir Valley in the past millennium, and estimates of maximum shaking intensity deduced from the somewhat sparse descriptions of damage they contain, mostly from Srinagar (Iyengar and Sharma, 1998; Iyengar et al., 1999; Bashir et al., 2009). With the caveat that these estimates of shaking are necessarily

very approximate, earthquakes in 1555, 1736, 1779, 1784, and 1828 appear to be candidate events. We argue for later rather than earlier damage for the following reason. The tree that grew from, or through the roof in 1887, and visible in Knight’s 1860 sketch, was evidently growing in 1868 and 1887. The temple was obviously in a neglected condition prior to 1868, but the rate of growth of the tree suggests that it had not been long established. We thus conclude that the displacement of the summit blocks probably occurred in the previous several decades. The most likely event then to have caused the roof damage is the 1828 earthquake, or possibly the late eighteenth-century events. What little is known of the 1828 earthquake is mentioned by Vigne (1844), who visited Kashmir in 1839. Although he was not in Kashmir at the time of the earthquake, his account of its effects 11 yr earlier includes numerous details that he must have assembled from eyewitnesses. He writes as follows (vol. 1, p. 281–283):

On the night of the 26th of June, 1828, at half past ten, a very severe shock was felt, which shook down a great many houses, and killed a great number of people: perhaps 1000 persons were killed, and 1200 houses shaken down; although; being built with a wooden framework, the houses were less liable to fall than an edifice of brick or stone. The earth opened in several places about the city; and fetid water, and rather warm, rose rapidly from the clefts, and then subsided. These clefts, being in the soil, soon closed again, and scarcely left any traces. I saw the remains of one fifteen yards long and two wide; but it was filled up, or nearly. Huge rocks and stones came rattling down from the mountains. On that night only one shock took place; but just before sunrise there was another, accompanied by a terrific and lengthened explosion, louder than a cannon. On that day there were twenty such shocks, each with a similar explosion. The inhabitants were, of course in the open country. The river sometimes appeared to stand still, and then rushed forward. For the remaining six days of Zilheja, and the whole of the two next months of Moharrem and Safur, there were never less than 100, and sometimes 200 or more shocks in the day, all accompanied with an explosion; but it was remarked, that when the explosion was loudest, the shock was the less. On the sixth day, there was one very bad shock, and on the fifteenth, at three o’clock, was the worst, and there were three out of the whole number that were very loud. At the end of the two above-mentioned months, the number decreased to ten or fifteen in the twenty-four hours, and the noise became less, and the earthquakes gradually ceased. About this time the cholera made its appearance. A census of the dead was taken at first, but discontinued when it was found that many thousands had died in twenty one days. In Kashmir there had been no great earthquake before, within the memory of any living person, excepting one about fifty years ago, which was rather severe, that lasted, at intervals, for a week. An earthquake is mentioned in Prinsep’s tables as having taken place in A.D. 1552. Shocks are now common, and the houses are built with a wooden framework, so as to resist them. (p. 406)

Vigne (1844) knew of the Pandrethan temple, “the curious building at Pandrynton, near the present city, which stands in a

Historical earthquakes in Srinagar, Kashmir

115

TABLE 1. HISTORICAL EARTHQUAKES REPORTED IN KASHMIR (PRINCIPALLY SRINAGAR) SINCE THE TENTH CENTURY Date Maximum estimated intensities Comments Iyengar and Sharma (1998); Bashir et al. (2009) Iyengar et al. (1999) – 1 1123 No details. 2 24 September 1501 VII Three months of aftershocks. 3 1552 (not an earthquake) Cited by Vigne and copied by others but a misinterpretation of Prinsep’s p. 312 one-line entry on the ascendency of Ibrahim II and the notable terminal event of his short reign. 4 1 September 1555 XII Earthquakes continued for several (7) days. Landslides and liquefaction. Several accounts (Ambraseys and Jackson, 2003), some assign date as 1554. – 5 ca. 1560/1561 No details. – 6 1569–1577 No details. 7 23 June 1669 IV The buildings rocked like cradles. No loss of life. 8 ca. 1678/1679 VII Persistent shaking. Reconstruction needed. 9 1683 Bashir et al. (2009). 10 24 March 1736 VIII Earthquakes for 3 mos. Buildings of the city and hamlets razed to the ground (Bashir et al., 2009, list as 1735). 11 1779 VII Srinagar and hamlets flattened and aftershocks for 14 d. People took shelter in the open. Bashir et al. (2009) list event as 1778; Oldham (1883) as 1780. 12 ca. 1784/1785 VIII People thrown. Shocks persisted 6 mo. 13 1803 VII Earth ripped apart, houses collapsed, people buried under walls (Bashir et al., 2009). 14 26 June 1828 VIII Vigne (1844) 1200 houses collapsed, 15 d of aftershocks (Bashir et al., 2009). 15 1863 Bashir et al. (2009); Lawrence (1895) indicates 1864. 16 30 May 1885 VI Bashir et al. (2009); Jones (1885) M = 6.2–6.3 (Ambraseys and Douglas, 2004; Szeliga et al., 2010). 17 8 October 2005 VII Mw = 7.6 instrumental period. Note: Discrepancies in timing of a year or more can occur where historians are using secondary texts, or where chronological conversions are sometimes ambiguous. We have attempted to reconcile close dates as single events, and have indicated known discrepancies. Item 3 is a spurious event that should not be repeated. The estimated intensities are no more than educated guesses by the authors cited.

tank,” and although the Cambridge architectural historian R. Willis describes a sketch that Vigne sent him of the building, and of others, Vigne omits the sketch of Pandrethan from his book, as does Temple (1887) from the edited version of his father’s 1859 diary. Vigne groups Pandrethan among the top four wellpreserved Hindu “ruins,” which visitors should see in Kashmir (if short on time to see all 70–80 surviving monuments). Of Pandrethan, Vigne comments that “the upper part was certainly pyramidical,” implying that the summit cap-stone was absent in 1839. He also remarks that many of ruins were once much taller but have “been diminished by earthquakes, even within the memory of man” (p. 391). We conclude from these descriptions of lateral spreading and structural collapse in Srinagar in 1828 that Pandrethan was more severely shaken in 1828 than in 1885 and 2005. We provisionally ascribe the damage to the temple to this date, although clearly it may have inherited damage from previous earthquakes (Table 1). The earthquake mentioned by Vigne as occurring ~50 yr previously would be the 1778/1779 or 1784/1785 earth-

quakes, for which independent descriptions exist (Iyengar and Sharma, 1998; Bashir et al., 2009), which appear also to have been a main shock–aftershock sequence persisting for 2 wk. Prinsep’s 1552 alleged earthquake (Prinsep, 1858, p. 312), cited by Vigne (1844) and repeated by Constable (p. 395 footnote in Bernier, 1891) and by Bashir et al. (2009), refers to the start of the rule of Ibrahím II, not to the date of events during his short rule, which terminated in the year of the 1555 earthquake. In table 75, in which he summarizes Ferishta’s list of kings of Kashmir, Prinsep’s 960 A.H. (A.D. 1552) entry consists of the seven words “Ibrahím II., set up by Daulat Chakk: earthquake” and is followed, on the next line, by an entry for 1555 identifying the succeeding king as Ism’aíl. That is, the 1552 event is bogus, and although listed here, should not be repeated (Table 1). Pandrethan as a Strong-Motion Seismometer? The vivid account Vigne has left us for the 1828 earthquake appears to be largely based on effects observed in Srinagar and

116

Bilham et al.

its surroundings. The accounts of loud explosions accompanied by lesser shaking may refer to P waves from nearby aftershocks unaccompanied by surface waves. In contrast, accounts for the 1555 earthquake have survived from several parts of the valley, which seismologists have taken to imply a significant earthquake with large rupture zone (Ambraseys and Jackson, 2003). Others have concluded that this large rupture may have occurred beneath the Pir Pinjal (Hough et al., 2009). Iyengar and Sharma (1998) assigned maximum intensity XII to the 1555 earthquake, but it is not clear how this intensity was quantified or where it is considered applicable. Historical accounts tend to record the most severe damage, or the most remarkable effects of earthquakes. We suspect, however, that intensity XII (or even IX) shaking would have destroyed the Pandrethan temple. Ambraseys and Jackson (2003) estimated an approximate magnitude of 7.6 based on its felt area, but this magnitude should be used with caution. If one assumed that the Pandrethan temple has survived unscathed since its construction, this suggests that shaking in the 1828 earthquake was probably more severe at this location than shaking in 1555. The descriptions of lateral spreading, river reversals, and weeks of aftershocks indicate that a major earthquake occurred in 1828 during the Sikh administration of the valley (1819–1846), but the absence of reports from other parts of the valley, or from large cities in northern India (e.g., Lahore or Amritsar), suggests that the earthquake, unlike the 1555 earthquake, was probably close to Srinagar, and not beneath the Pir Pinjal. That shaking at any one site is more severe during a moderate earthquake than a larger event is not surprising, because shaking intensity at any site depends on myriad factors, such as source radiation pattern, distance to the event (or distance to largest asperities within an extended rupture), and threedimensional propagation effects. However, conclusions about earlier historical events are tenuous at best because one cannot know if reconstruction/repair was undertaken in early historical times by devotees of the temple. It is further possible that early earthquakes weakened the temple, rendering it more susceptible to damage in subsequent earthquakes.

structure, and cause it to be reassembled from a heap of rubble. One would thus infer an absence of intensity >IX shaking in the past millennium. This conclusion is, however, more tenuous than the conclusions one can draw about shaking during nineteenthand twentieth-century earthquakes. Estimates of shaking intensity at a single site provide little indication of earthquake magnitude; instrumental recordings of recent earthquakes reveal that very high (PGA [peak ground acceleration] >1 g) accelerations can be generated by relatively moderate (M 6.5–7) earthquakes, while surprisingly low accelerations are sometimes recorded in the near field of very large events. Our findings are thus clearly insufficient to draw conclusions about the magnitudes of earthquakes that have shaken Kashmir Valley in the past millennium, nor do our results provide upper limits to the shaking experienced in historical times in nearby Srinagar, where thick sediments in the Jhelum River valley and around lakes are likely to amplify shaking significantly. Careful analysis of other ancient monuments in the valley, in particular dating of damage, may usefully supplement the sparse historical record. We note in closing that the Pandrethan Temple serves as both an encouraging and a cautionary case study: encouraging to the extent that the structure does provide useful clues that help elucidate the earthquake history of the region; cautionary to the extent that, if not for the fortuitous existence of repeat historical photographs, one could easily be led to the same obvious but mistaken conclusion implied by R.D. Oldham’s 1887 photograph, that damage to the temple evident in 1887 was caused by the 1885 earthquake.

CONCLUSIONS

REFERENCES CITED

The survival of the small masonry tenth-century Shiva Temple at Pandrethan, near Srinagar, through more than a dozen damaging earthquakes suggests that shaking greater than intensity IX has not occurred in the past 200 yr, and possibly the past 1000 yr. The case is based on “calibration” earthquakes in 1885 and 2005 that shook the temple with intensities of VII or less. From photographs in 1868 and 1887 that show the rapid growth of a small tree that grew through cracks in its roof, we deduce that the temple had been damaged by an earlier earthquake, probably in 1828, and/or between 1778 and 1885, in which local intensities must have exceeded VIII. The absence of damage to the ornate ceiling of the Pandrethan temple and most of the temple blocks suggests that accelerations in the past millennia have been insufficient to destroy the

Ambraseys, N.N., and Douglas, J., 2004, Magnitude calibration of North Indian earthquakes: Geophysical Journal International, v. 158, p. 1–42, doi: 10 .1111/j.1365-246X.2004.02323. Ambraseys, N.N., and Jackson, D., 2003, A note on early earthquakes in India and southern Tibet: Current Science, v. 84, p. 570–582. Bernier, F., 1891, Travels in the Mogul Empire, A.D. 1656–1668 (translated by I. Brock and revised by A. Constable): London, Constable & Co., 497 p. Bernier, R.M., 1997, Himalayan Architecture: Madison, Fairleigh Dickinson University Press, 196 p. Bashir, A., Bhat, M.I., and Bali, B.S., 2009, Historical record of earthquakes in the Kashmir Valley: Journal of Himalayan Geology, v. 30, no. 1, p. 75–84. Bilham, R., 2008, Tom LaTouche and the Great Assam Earthquake of 12 June 1897; letters from the epicenter: Electronic supplement: Seismic Research Letters, v. 79, no. 3, p. 426–437. (Transcriptions of LaTouche’s letters 1882–1913 illustrated by archival photographs from Calcutta and London: http://www.seismosoc.org/publications/SRL/SRL_79/srl_79-3 _hs.html [accessed 3 August 2010].) Burke, J., 1868, “Temple of Meruvarddhanaswami at Pandrethan near Srinagar,” British Library: Shelfmark: Photo 981/1(40).

ACKNOWLEDGMENTS We thank the British Library for permission to publish Burke’s 1859 photo, and the director general of the Geological Survey of India for permission to reproduce R.D. Oldham’s photographs of Pandrethan. The investigation was funded by the U.S. National Science Foundation.

Historical earthquakes in Srinagar, Kashmir Cole, H.H., 1869, Illustrations of Ancient Buildings in Kashmir, India Museum: London, India Museum, W. H. Allen and Co., publishers to the India office, 31 p. Cunningham, A., 1848, An essay on the Arian Order of Architecture, as exhibited in the Temples of Kashmir: Journal of the Asiatic Society of Bengal, September 1848, p. 241–327. Doughty, M., 1902, Afoot through Kashmir Valleys: London, Sands and Co., 276 p. Fergusson, J., 1867, History of Indian and Eastern Architecture: London, Murray, v. 3, 756 p. Hough, S.E., Bilham, R., and Bhat, I., 2009, Kashmir Valley megaearthquakes: American Scientist, v. 97, no. 1, p. 42–49, doi: 10.1511/2009.76.1. Hügel, C.A., 1845, Travels in Kashmir and Panjab: Containing a Particular Account of the Government and Character of the Sikhs (translated by T.B. Jervis, East India Company): London, J. Petheram, v. 1845, 423 p. Hunter, W.W., 1881, Gazetteer of India, v. VIII, 284 p. Iyengar, R.N., and Sharma, D., 1996, Some earthquakes of Kashmir from historical sources: Current Science, v. 71, no. 4, p. 300–331. Iyengar, R.N., and Sharma, D., 1998, Earthquake History of India in Medieval Times: Roorkee, Central Building Research Institute, 124 p. Iyengar, R.N., Sharma, D., and Siddiqui, J.M., 1999, Earthquake history of India in medieval times: Indian Journal of History of Science, v. 34, no. 3, p. 181–237. Jones, E.J., 1885, Report on the Kashmir earthquake of 30 May 1885: Records of the Geological Survey of India, v. 18, no. 4, p. 221–227. Kak, R.C., 1933, Ancient Monuments of Kashmir: Delhi, India Society, 172 p., 1971 reprint. Knight, W.H., 1863, Diary of a Pedestrian in Cashmere and Thibet: London, Bentley, 385 p. Lawrence, W.R., 1895, The Valley of Kashmir: Gulshan, 478 p., 2007 reprint.

117

Martin, S., and Szeliga, W., 2010, A catalog of felt intensity data for 589 earthquakes in India, 1636–2008: Bulletin of the Seismological Society of America, v. 100, no. 2, p. 562–569, doi: 10.1785/0120080328. Moorcroft, W., and Trebeck, G., 1841, Travels in the Himalayan Provinces of Hindustan and the Panjab, in Ladakh and Kashmir, in Peshawar, Kabul, Kunduz, and Bokhara from 1819 to 1825, Volume 2: London, J. Murray, 565 p. Oldham, R.D., 1887, Photographic Archives #257(288): Temple at Pendrethan showing stones displaced by earthquakes, location: near Srinagar; and #258(289): Temple at Pandrethan, location: Srinagar, Kashmir: Calcutta, Geological Survey of India. Oldham, T., 1833, A catalogue of Indian earthquakes from the earliest times to the end of AD 1869 (edited by R.D. Oldham): Memoirs of the Geological Survey of India, v. 19, no. 1, p. 163–215. Prinsep, J., 1858, Essays on Indian Antiquities, Historic, Numismatic, and Palæographic, of the Late James Prinsep: To Which Are Added His Useful Tables, Illustrative of Indian History, Chronology, Modern Coinages, Weights, Measures, Etc., Volume 2: London, J. Murray, 336 p. Szeliga, W., Martin, S., Hough, S., and Bilham, R., 2010, Intensity, magnitude, location and attenuation in India for felt earthquakes since 1762: Bulletin of the Seismological Society of America, v. 100, no. 2, p. 570–584, doi: 10.1785/0120080329. Temple, R., 1887, Diary of a Journey into Jammun and Kashmir between 8th June and 8th July 1859, in Temple, R.C., ed., Journals Kept in Hyderabad, Kashmir, Sikkim and Nepal, v. 2: London, W. H. Allen, 303 p. Vigne, G.T., 1844, Travels in Kashmir, Ladak and Iskardo, the Countries Adjoining the Mountain Course of the Indus, and the Himalaya, North of Panjab, with Map, Volume 1 (2nd ed.): London, H. Colburn, 406 p. MANUSCRIPT ACCEPTED BY THE SOCIETY 19 MAY 2010

Printed in the USA

The Geological Society of America Special Paper 471 2010

Earthquakes and civilizations of the Indus Valley: A challenge for archaeoseismology Robert L. Kovach* Department of Geophysics, Stanford University, Stanford, California 94305, USA Kelly Grijalva Department of Earth and Planetary Science, University of California–Berkeley, Berkeley, California 94720, USA Amos Nur Department of Geophysics, Stanford University, Stanford, California 94305, USA

ABSTRACT Civilizations have existed in the proximity of the Indus River Valley regions of modern Pakistan and India from at least 3000 B.C. onward. Geographically, the region encompasses a swath of the Makran coast, the alluvial plain and delta of the Indus River, and the Runn of Kachchh. The regional tectonic setting is controlled by the collision of the Indian and Eurasian plates and the subduction of the Arabian plate beneath the Eurasian plate. Earthquakes have undoubtedly struck many ancient sites, but finding their footprint in a riparian environment represents a challenge for archaeoseismology. However, some insight into seismoarchaeological indicators can be gleaned from examining the earthquake effects produced by historical infrequent large-magnitude events that have occurred in the region. Studies of these earthquakes emphasize the importance of repeated reconstructions, direct faulting, river damming from seismic uplift, and coastal elevation change as indicators of past earthquakes. Examples of past earthquake effects are presented for Banbhore in the Indus Delta, Brahmanabad, and the Harappan sites of Kalibangan and Dholavira. Future hermeneutic investigations in the region need to incorporate a seismological/ tectonic perspective and not rely solely on serendipity. INTRODUCTION For many thousands of years, civilizations have occupied regions of the southwestern Indian subcontinent stretching from the Arabian Sea to the foothills of the Himalayas. The tectonic environment argues that earthquake occurrences are a continuing integral facet of the region and must have occurred in the historical past. Earthquakes often leave their mark in the myths, leg*[email protected]

ends, and written accounts of ancient peoples, the stratigraphy of their historical sites, and the structural integrity of their constructions. Such information contributes to a better understanding of the irregularities in the spatial-temporal patterns of earthquake occurrences and whether earthquakes have contributed to the abandonment of ancient occupational sites. The objective of this contribution is to present some examples of seismic damage for a few archaeological sites in the southwestern Indian subcontinent, in particular, the Indus River Valley region and the Runn of Kachchh. Much work is yet to be

Kovach, R.L., Grijalva, K., and Nur, A., 2010, Earthquakes and civilizations of the Indus Valley: A challenge for archaeoseismology, in Sintubin, M., Stewart, I.S., Niemi, T.M., and Altunel, E., eds., Ancient Earthquakes: Geological Society of America Special Paper 471, p. 119–127, doi: 10.1130/2010.2471(11). For permission to copy, contact [email protected]. © 2010 The Geological Society of America. All rights reserved.

119

120

Kovach et al.

done in this region in deciphering the effects of ancient earthquakes from the archaeological record. TECTONIC FRAMEWORK AND SEISMICITY OF THE REGION A generalized map of the region under discussion is shown in Figure 1. A more detailed tectonic map can be found in Nakata et al. (1991). The east-west Makran coast of southern Pakistan forms a section of the convergent boundary of the EurasianArabian plate. It is bounded to the east by a transform fault, the Murray Ridge, which defines the western oceanic edge of the Indian plate. On land, the Chaman fault system, including the Ornach Nal fault zone in the south, forms the eastern boundary of the Makran region, accommodating the motion between the Eurasian and Indian plate. The convergence rate between the Arabian and Eurasian plate is estimated from modern global positioning system (GPS) studies to be ~22–30 mm/yr along the Makran coast (Apel et al., 2006; Sella et al., 2002; Vernant et al., 2004). Such a high collision rate implies that a large amount of potential earthquake slip should be accumulating each century to be periodically released in large earthquakes. However, the historic seismic record of the Makran coast does not indicate many great events, arguing that perhaps much of the subducting plate motion takes place aseismically or that great earthquakes are separated by periods, longo intervallo, that exceed the seismic record. Long-term, average rates of uplift have been determined for various segments of the Makran coast by Page et al. (1979) and Vita Finzi (1980). Radiometric dating on marine shells found in various depositional horizons in marine terraces indicates a Holo-

60˚ 61˚ 62˚ 63˚ 64˚ 65˚ 66˚ 67˚ 68˚ 69˚ 70˚ 71˚ 72˚ 73˚ 74˚ 75˚E 31˚N ASIAN PLATE 30˚ 30˚ Helmand block lock INDIAN PLATE Pakistan 29˚ 29˚ 0.8 cm/yr er iv R 28˚ 28˚ s 3.5 cm/yr du In 27˚ 27˚

cene uplift rate of 0.1–0.3 cm/yr for portions of the Makran coast. The inferred recurrence time for an M ~8 event is 660–1000 yr, suggesting that 4–6 similar sized events have occurred within the past 4000 yr. The Chaman fault system, which defines the western edge of the Indian plate, has a slip rate of movement of 20–40 mm/yr based on geological offsets (Lawrence et al., 1992) and a relative plate velocity of 26–34 mm/yr (Apel et al., 2006). The rates of movement point to a recurring potential of significant amounts of seismic slip accumulation over time scales of several hundred years. The pattern of current and historical seismicity roughly mirrors the plate boundary framework of the region (Fig. 2). The Makran region of southern Pakistan and southeastern Iran is characterized as a region with a relatively low level of seismic activity and infrequent great earthquakes. Along the Makran subduction zone, most of the seismic activity takes place at focal depths less than 24 km and is confined within an ~26° northward-dipping zone beneath the coast (Engdahl et al., 2005). The spatial distribution of epicenters forms a relatively narrow east-west band. Along the coast, the seismic activity is accented by two large earthquakes that took place in the 1940s. The MW = 8.1 event of 27 November 1945 produced a tsunami, significant coastal uplift, and the eruption of several offshore mud volcanoes (Byrne et al., 1992). A second event (M = 6.9) occurred approximately in the same area as the 1945 event on 5 August 1947. Indications of earlier seismic activity in the remote Makran coastal area

E N

CFZ

31˚

26˚

Makran

2001

India Arabian Sea

23˚ 22˚ 21˚

~2.8 cm/yr

1819

26˚

25˚ 24˚

1945

y rra Mu

ge Rid

~2.9 cm/yr Kachchh

km 0 100 200

20˚

25˚ 24˚ 23˚ 22˚ 21˚ 20˚

60˚ 61˚ 62˚ 63˚ 64˚ 65˚ 66˚ 67˚ 68˚ 69˚ 70˚ 71˚ 72˚ 73˚ 74˚ 75˚

Figure 1. Regional tectonic setting of the Indus River Valley and its environs (modified from Regard et al., 2005; Vernant et al., 2004). CFZ indicates the Chaman fault zone.

Figure 2. Seismicity of the Indus Valley region from 1905 to 2005. The circles of varying size are for events of magnitude 5 or greater. The stars represent the assumed locations for pre-instrumental events producing seismic intensities of VIII or greater. Locations of the 1945 M = 8.1 Makran coast, the 1819 Runn of Kachchh, and the 2001 M = 7.6 Bhuj earthquakes are labeled. Data taken from Quittmeyer and Jacob, 1979; Endahl et al., 2005; International Seismological Center, 2006; and National Earthquake Information Center, 2006.

Earthquakes and civilizations of the Indus Valley are sparse, so that any inferences about its long-term history are poor and incomplete. Historical damaging events are reported to have occurred in 1483, 1765, and 1851 (Ambraseys and Melville, 1982; Bilham et al., 2007; Oldham and Oldham, 1883). The highest level of current seismic activity is found along the western boundary of the Indian plate, extending northeasterly from the Arabian coast up to latitude 29°N along the Chaman fault zone. The recent seismic activity of the Chaman fault zone primarily consists of moderate-sized earthquakes with maximum magnitudes less than 5. Seismic activity increases northward from the coast, and at latitude of 29°N, a zone of high seismic activity is reached, characterized by several damaging earthquakes possessing magnitudes of 7 or greater. Epicenters in this region align well with the mapped Chaman fault system but also occur to the east toward the Indus River. Several earthquakes that took place from 1931 to 1935 caused significant fatalities (West, 1934, 1936). Damaging earthquakes in the Indus River floodplain and Kachchh regions, distant from the known plate boundaries, are infrequent. These intraplate earthquakes, however, such as the 1819 event in the northern Runn (Rann) of Kachchh, have produced a significant number of fatalities and also altered the geomorphic landscape of the region. It is believed that in the historical past, the Runn of Kachchh was at a lower elevation and linked to the Arabian Sea. Gupta (1975) has demonstrated that as late as 2000 yr ago, portions of Runn had a water depth of 4 m and thus was inundated throughout the year. Today, the Runn is not perennially under water, and so its elevation has been altered in recent times. What cannot be unambiguously resolved, however, is what percentage of the present configuration was produced by deltaic sedimentation and what percentage by tectonic uplift (Sivewright, 1907)? Burnes (1835, p. 570) described an oral tradition of the local natives arguing that the Runn was formerly opened to the sea and was altered by an earthquake. Bearing in mind that many cultures do not want to separate mythology from legendary tradition, the marvelous tale given next appears to contain umbilical elements of plausibility: “… a Hindu saint, named Dhuramanat’ha, a Jogi [a holy and hospitable man] underwent penance, by standing on his head … for a period of 12 years. At that time he resumed his proper position, and God became visible to him, when a convulsion of nature took place, and the hill on which he stood split in two, the sea that lay northward of him [which is the present Runn] dried up, and the ships which than navigated it were wrecked and its harbours destroyed, with other miraculous and wonderful events.” What is important, in the context of seismoarchaeology, is that geological forces, particularly earthquakes, have greatly altered the geography of the region over the past 4000–5000 yr or so. As a consequence, this may have accelerated the demise of many ancient settlements by altering the water supply, modifying trade routes, producing the need for continual rebuilding, and ultimately forcing migration.

121

SEISMOARCHAEOLOGICAL INDICATORS Earthquakes can cause elevation changes and alter the drainage pattern of life-giving rivers. Other direct effects of earthquakes that might be present at ancient sites in the Indus Valley and its environs under archaeological investigation are often more difficult to document because individuals with seismic expertise have been rarely involved during the actual excavations themselves. The archaeological approach at many ancient sites mainly concentrates on documenting stratifications of the ground unless the investigator has a curiosity as to the seismic history of the site or is intent on finding evidence for a catastrophic event of the past. Nevertheless, earthquakes have left their imprint on many archaeological sites in the Indus Valley region. Unlike volcanic eruptions and floods that can be recognized in excavations from an examination of soil composition and stratigraphy, the traces of past earthquakes in buildings and other structures are usually more difficult to interpret. Aging of materials and burial and architectural innovations over time have to be considered. However, the effects of ancient seismic destruction can often be ascertained from a careful reading and syntheses of archaeological records and written accounts when available. Even if the evidence for earthquake damage is only suggestive, it can be bolstered by plausibility and consistency arguments, and the onus probandi is then placed upon those who wish to prove one point or another. We here point to the statement by Ambraseys (2005), who asserts that earthquakes are often looked to as an easy solution for explaining civilization and cultural gaps. With these prefatory comments being said, Table 1 enumerates a few key diagnostics that can often be gleaned from the archaeological record and that are appropriate for ancient locations in the Indus Valley region. In the following sections, we will discuss a few examples for which evidence of ancient earthquakes has been found.

TABLE 1. SOME CRITERIA FOR IDENTIFYING EARTHQUAKE OCCURRENCES 1. Direct evidence of faulting observed in excavations. 2. Skeletons of people found in nonburial positions, particularly if found buried beneath the debris of fallen structures. 3. Abrupt geomorphological changes, such as changes in river drainage associated with subsequent abandonment or reconstructions. 4. Fallen walls with no evidence of rebuilding. 5. Evidence of destruction and hasty reconstruction using debris from earlier constructions. 6. Reconstruction of walls with what might be called “antiseismic” additions, such as buttressing. 7. Well-dated building destructions correlating with historical accounts of earthquakes. Earthquakes mentioned in myths and legends. 8. Numismatic evidence such as the leaving behind of coins and valuables due to a hasty abandonment.

122

Kovach et al.

Brahmanabad An interesting set of archaeological ruins, discovered in 1854, is known as Brahmanabad-Mansura in Sind. The ruins are situated in an open sandy plain upon the bank of an ancient bed of the Indus River, ~200 km northeast of Karachi, Pakistan (Fig. 3). A broken brick tower was the only recognizable standing feature seen in 1894 (Fig. 4). The deserted site had the circuitous shape of a boot with its sole facing northwest and its leg pointing toward the southeast. It was not a large site, as the distance around the periphery was only 9.2 km. The excavations of Bellasis (1856a, 1856b) led him to believe that the site had been devastated by an ancient earthquake:

66˚

67˚

68˚

69˚

70˚E 26˚N

26˚ Brahmanabad Umarkot

25˚

25˚

Karachi

BANBHORE

712 coastline

1833 coastline

24˚

24˚ Lakhpat

Bhuj

We had not dug two feet before we came to quantities of bones … skeletons were so numerous … The human bones were found in doorways, as if the people had been attempting to escape, and others in the corners of the rooms. Many of the skeletons were in a sufficiently perfect state to show the position the body had assumed; some were upright, some recumbent, with their faces down, and some crouched in a sitting posture. One in particular, I remember, finding in a doorway: the man had evidently been rushing out of his house, when a mass of brick-work had, in its fall, crushed him to the ground, and there his bones were lying extended full length…. (p. 417)

Three arguments were given in favor of destruction by an earthquake. The observed destruction was far too complete to have been solely the work of time; an invading army would have not produced such complete destruction, yet leave behind coins and other valuables. Finally, if the city had only been deserted in an orderly manner, the inhabitants would most likely have carried off their valuables with them. Brahmanabad was not visited again by archaeologists until 1897. At this time, it was observed that much of the surficial soil cover at the site had been destroyed by local cultivators. Subsequent excavations carried out by Cousens (1906, 1912) led him to suggest that the site was also the location of Mansura, the first Arab capital in Sind, and that Mansura had been built upon the ruins of ancient Brahmanabad. Of particular relevance to the earthquake question was the observation that the earlier excavations of Bellasis were not deep enough to have reached the Brahmanabad horizon. It was in the uppermost layer, that of Mansura, or a subsequent Mansura rebuilt after the earthquake disaster, that bones, ash, broken pottery and quantities of charcoal were found that had led Bellasis to his original supposition. It now remains to place bounds on the date of the postulated earthquake disaster. Brahmanabad was conquered by Muhammad Qasim in A.D. 712, and Mansura is reputed to have been built by his son Amru. Local coins found in the upper layers belong to subsequent Arab governors of Mansura, ca. A.D. 750 (Sykes, 1857), and Arab chroniclers describe its existence until at least A.D. 1020 (Bellasis, 1856a). The terminous post quem for the earthquake that leveled the walls of Mansura, overthrew

km

23˚ 0 66˚

100 67˚

23˚ 200 68˚

69˚

70˚

Figure 3. Location of archaeological sites of Brahmanabad and Banbhore. Shoreline and the inland sea at the time of the Arab conquest of Sind in A.D. 712 are shown by thin, solid black lines (Sivewright, 1907). The routes of the Indus River and its branches to the sea in 1833 are also shown (Burnes, 1833).

Figure 4. View from the west of the ruined tower at Brahmanabad (Cousens, 1912).

Earthquakes and civilizations of the Indus Valley its housing, and crushed many of its inhabitants, therefore, is the early eleventh century. On the other hand, if it is not accepted that Mansura was reconstructed on the site of Brahmanabad, but instead was located some 8 km to the northeast, then a slightly earlier date for the terminal earthquake at Brahmanabad is suggested (Sykes, 1857; Bilham et al., 2007). Banbhore Banbhore, an archaeological site 64 km east of Karachi (Fig. 3), is located on a flat prominence on the right bank of Gharo Creek, a creek that formed the main course of the Indus prior to A.D. 1250. Gharo Creek, now filled with silt and sand, marks the northern edge of the alluvial fan of the Indus Delta. The key structure at Banbhore was its central mosque. Its earliest constructed wall was built with dressed blocks of sandy limestone. The feature of particular interest is that the mosque was subjected to three successive reconstructions with increasing deterioration in the quality of work. Based on ceramics and renovations, it is inferred that there were four principal phases of Muslim occupation under which construction and repairs took place (Ashfaque, 1969). The earliest construction, in which the foundation was laid, took place in the Umayyid (Ummeide) Period from A.D. 715 to 750. This was followed by the Abbasis (Abbasides) Period from A.D. 750 to 892, during which the mosque was damaged by an earthquake, and major repairs were carried out. The southern wall of the mosque suffered the greatest damage and was rebuilt from undressed stones picked from the rubble. In addition, buttressing was placed against the weakened eastern boundary walls. Written documents state that the ancient town of Banbhore was destroyed during the time of the military exploits of Shekh Abu Turab during the caliphate of Harun Al Rashid (A.D. 786–808) to reestablish Islam in Sind, placing the time of the earthquake sometime between A.D. 787 and 790 (Fredunbeg, 1902). The third period of Muslim occupation, called the Late Abbasid Period, covered the time interval from A.D. 892 through most of the twelfth century A.D. Earlier repairing of the boundary walls did not prove to be lasting, since there is archaeological

123

evidence for an extensive second repair using carved blocks from the original structure and rubble set in mortar. A dated inscription (294 A.H. or A.D. 906) has been found commemorating the completion of the extensive repairs to the mosque (Fig. 5). It has been speculated that this rebuilding may have been needed as a result of an earthquake in 280 A.H. (A.D. 893–894). In Tornberg’s (1865) edition of Ibn-el-Athiri, al-Kaml fi’l-Ta’rikh, volume VII, p. 465, it is stated:

And in the Shawwal of this year (280 A.H.) there was an eclipse of the moon and the people of Dabil and the whole world woke in darkness. And the darkness lasted for a long time and when it was 4 o’clock in the afternoon there blew a black wind, which continued to a third part of the night. And when the third part of the night had come, there was an earthquake, and the town was destroyed, so that only about 100 houses remained, and after this the earth shook five times more and the total number of people that were found killed below the ruins amounted to 150,000.

The description given in A’s Suyuti’s History of the Caliphs (Jarrett, 1881, p. 387–388) strongly suggests the same historical source: “In the same year (280 A.H.) came advices from Daybul (Daibul) that the moon had been eclipsed in the month of Shawwal, and that darkness had spread over the country till the afternoon when a black storm began to blow which continued for a third of the night, followed by a mighty earthquake which destroyed the whole city, and the number of those taken out from the ruins was one hundred and fifty thousand.” Mention of a disastrous earthquake on this date is given by Hoff (1840) quoting the authority of the thirteenth-century Syrian writer Bar Hebraeus (Budge, 1932), who explicitly mentions a location in outer India. However, the earthquake described in these chronicles is well-documented to have taken place in Dvin, Armenia, on 27 December 893. Dabil is the Arabic name for Dvin, leading to misconceptions that this particular earthquake may have occurred in a similarly named place in the Indus Delta (Guidoboni and Traina, 1995; Ambraseys, 2004). The location of Debal in modern-day Pakistan has not been positively identified,

Figure 5. Floral Kufic inscription found at Banbhore commemorating rebuilding after earthquake(?) in 280 A.H. (A.D. 893–894). “In the name of Allah…This is what Amir Muhammad ibn has ordered about its erection in ‘Dhu’l Qadah?’ in the year 294” (Ghafur, 1966).

124

Kovach et al.

but some scholars have argued that a location near Karachi and possibly Banbhore itself is not unlikely. Regardless of its location, however, the archaeological evidence for repeated damage and reconstruction at Banbhore lends credence to its past vulnerability to earthquakes. The final period of the settlement of Banbhore covers the time interval, described as a period of decadence and extinction, from the late twelfth to the mid-thirteenth century A.D. During this third period of reconstruction, arches were repaired, and entrances were made narrower and strengthened by the stacking of stones and bricks on either side. Khan (1960) stated that the settlement of Banbhore came to a sudden end following a violent disturbance. A large number of human skeletons were found lying in a disorderly manner, and bricks and stones were found fallen from their original structures. Kalibangan Kalibangan was an Early and Mature Harappan site that was situated on the left bank of the Ghaggar River (known in ancient times as the Sarasvati), near its intersection with the Drishadvati River (Fig. 6). These rivers were active and perennial during the time of occupation from 3000 B.C. to 1750 B.C. It is generally accepted that the final abandonment of Kalibangan was the consequence of major shifts in the courses and volume of water associated with the Sarasvati and Drishadvati Rivers. The temporal details of the changing drainage patterns of these ancient rivers, however, are a topic of continuing debate (Possehl, 1999). The Early Harappan settlement at Kalibangan, found beneath a mound labeled KLB-1, was not large. It consisted of a parallelogram-shaped walled enclosure of mud bricks with an entrance on the northwestern side (Fig. 7). Several trenches were

made during the excavation of KLB-1 that reveal definite evidence for one or more ancient earthquakes. The archaeological diggings show clear signs of fault rupture and displacement of horizons (Figs. 8 and 9). Lal et al. (2003, p. 100) stated that “… even the uppermost layers are affected by this earth-movement. All this implies that the violent shaking of the area, which in other words would mean an earthquake, brought about the end of the Early Harappan occupation at Kalibangan.” Offset stratigraphic horizons are present in trench XD2-XE2 in the southwest corner of the excavated mound. To the north, a tilted brick wall produced by some disturbance was exposed near YA17. We can speculate that surface faulting produced both disturbed structural features. The strike of the deduced rupture is N12°E, roughly parallel to the northeasterly trend of the Indian plate boundary in this region. It is possible to estimate the magnitude of this earthquake in 2700 B.C. A fault offset of 30–40 cm is present in the depositional layers in trench XD2-XE2. From the empirical relations of Wells and Coppersmith (1994), a fault offset of this size implies an earthquake magnitude of at least 6.5, and implies that the seismic intensities that struck Kalibangan were in the range of VIII to IX on the modified Mercalli intensity scale. The supposition that the earthquake caused abandonment is strengthened by the presence of an overlying layer of infertile, windblown sand covering the ruins. Kalibangan was subsequently reoccupied after a short hiatus, placing the date of the earthquake sometime between Period I (Early Harappan) and Period II (Mature Harappan), or 2700 B.C. (Lal, 1998). Final abandonment of Kalibangan occurred around 1750 B.C., presumably as a result of the change in river flow of the Sarasvati (Raikes, 1968).

E N

Figure 6. Map showing the location of Kalibangan at the junction of the paleorivers Sarasvati and Drishadvati in 2000 B.C. (generalized from Wilhelmy, 1969).

Figure 7. Site layout of fortified structure at the Mature Harappan site of Kalibangan.

Earthquakes and civilizations of the Indus Valley

125

An item of added interest to the thesis of earthquake damage is that archaeologists have found evidence for nine reconstructions at Kalibangan during the Mature Harappan occupational period. What is unusual about the archaeological site of Kalibangan is that it is not located in an area of current seismic activity. Nevertheless, the archaeological identification of an ancient earthquake(s) that is contemporary with the level of excavation is beyond conjecture. The direct evidence of this ancient earthquake is, therefore, doubly interesting. It demonstrates a cause and effect between an earthquake occurrence and abandonment and allows us to see traces of past seismic activity in an area that one would conclude from a modern-day seismicity map to be an area that is relatively “nonseismic.” Secondly it emphasizes that the seismic effects from damaging earthquakes with very large intervals of time between occurrences can be seen in archaeological excavations. Dholavira

Figure 8. Fractured and offset cultural horizons found in trench XD2-XE2 at Kalibangan. View is looking south. Evidence of ancient earthquake rupture can be seen in three places (Lal, 1998). Height of vertical wall is approximately 2 m.

Dholavira was a well-planned Harappan enclave built on a low plateau on Khadir Island, a true island ca. 4000 B.P. in the Runn of Kachchh (Fig. 10). The site is enclosed by a stone and brick wall, 5 m thick at its base, with rectangular dimensions of 700 × 750 m. Constructed on a slope between two nullahs (seasonal drainage gullies), the city contained a sophisticated watercollection system of giant reservoirs to collect and store seasonal rainwater. Archaeologists have identified seven states of occupation beginning in 2650 B.C. and extending to 1450 B.C. Stages I and II, which lasted ~150 yr, involved constructions of molded mud brick houses with a substantial plastered fortification wall surrounding a fort. Stage III saw a substantial residential area added to the north, together with reservoirs carved into the host rock (Bisht, 1991; Joshi and Bisht, 1994). Somewhere near the end of Stage III (ca. 2200 B.C.), Dholavira was struck by a major earthquake, as evidenced by slip faults

Figure 9. Stratigraphic section of trench shown in Figure 8. Faulted section is highlighted (Lal et al., 2003).

126

Kovach et al.

67˚

68˚

69˚

70˚

71˚

72˚

73˚ 73˚E

26˚

26˚N 26˚ 1030

25˚ 1668

25˚

893 1819

Runn of Kachchh

24˚

Dholavira 1956 2001 Bhuj

Ahmedabad

Kachchh Anjar

23˚

24

Surkotada

23

Lothal 0

Kathiawar Peninsula

22˚ 67˚

68˚

69˚

70˚

71˚

72˚

22 73˚

Figure 10. Map showing the location of the Harappan site of Dholavira in the Runn of Kachchh. Epicenters of the 1819 Allah Bund event and the M = 7.6 Bhuj earthquake of 2001 are shown.

Figure 11. View to the north of partially damaged entrance to fort at Dholavira from 2001 Bhuj earthquake. Structural damage from ancient earthquake in 2200 B.C. is preserved on adjacent tilted wall to the left of 2001 damaged wall. The inset shows a flexured wall in the central part of the fort damaged by an earlier earthquake (Rajendran and Rajendran, 2003). Width of flexured wall is ~1 m.

exposed in vertical sections and the displacement and tilting of walls (Fig. 11). The earthquake apparently produced extensive devastation at the site, because large-scale repairs were undertaken near the end of Stage III (Singh, 1996; Possehl, 2002). Stages IV and V, from 2200 to 1900 B.C., encompasses the Mature Harappan period of architectural flourishing, followed by gradual decline and a short period (~100 yr) of abandonment. The subsequent posturban stages had alternating periods of occupation and abandonment, characterized by an inferior level of construction compared to earlier periods. Natural events such as flooding, drought, or earthquakes are believed to have triggered the short periods of abandonment. Dholavira is located in close proximity to the epicenters of the M = 7.7 Allah Bund event of 1819 (Bilham, 1999) and the M = 7.6 Bhuj event of 2001 (Hough et al., 2002), supporting the contention that Dholavira was subjected to direct earthquake effects during Harappan times.

is that ancient earthquakes with long intervals of time between occurrences may have taken place in regions not indicated by present-day patterns of seismic activity. Traces of damage may also have disappeared during reconstruction works carried out at later times. Challenges remain for future seismoarchaeological investigations in the Indus River valley and delta region. It is hoped that future hermeneutic investigations will incorporate a seismological/tectonic perspective, so that future discoveries are not accidental.

CONCLUDING REMARKS

REFERENCES CITED

The examples presented here demonstrate that earthquakes have produced direct damage at several historical archaeological sites in the Indus Valley region. In addition, ancient earthquakes, similar to the 1819 Runn of Kachchh event, the 1945 Makran coast event, and the 2001 Bhuj event, may also have produced significant ground uplift and deformation affecting the geographic setting of early settlements. We here point to significant changes over historical time of the fluvial system of the Indus River, the consequences of river damming and subsequent flooding, and coastal elevation changes. For this reason, the lack of apparent seismic destruction at other ancient sites in the region should not be necessarily taken as evidence that a specific site was free from earthquake effects in the past. Another factor to consider

Ambraseys, N.N., 2004, Three little known early earthquakes in India: Current Science, v. 86, no. 4, p. 506–508. Ambraseys, N.N., 2005, Archaeology and neo-catastrophism: Seismological Research Letters, v. 76, p. 560–564. Ambraseys, N.N., and Melville, C.P., 1982, A History of Persian Earthquakes: Cambridge, UK, Cambridge University Press, 219 p. Apel, E., Burgmann, R., Bannerjee, P., and Nagarajan, B., 2006, Geodetically constrained Indian plate motion and implications for plate boundary deformation: Eos (Transactions, American Geophysical Union), v. 85, abs. 52T51B-1524. Ashfaque, S.M., 1969, The grand mosque of Banbhore: Pakistan Archaeology, no. 6, p. 182–209. Bellasis, A.F., 1856a, An account of the ancient and ruined city of Brahminabad, in Sind: Journal of the Bombay Branch of the Royal Asiatic Society, v. 5, p. 413–425. Bellasis, A.F., 1856b, Further Observations of the ruined city of Brahminabad, in Sind: Journal of the Bombay Branch of the Royal Asiatic Society, v. 5, p. 467–477.

ACKNOWLEDGMENTS This article is a contribution to the United Nations Educational, Scientific and Cultural Organization–funded International Geoscience Programme IGCP 567, “Earthquake Archaeology: Archaeoseismology along the Alpine-Himalayan Seismic Zone.”

Earthquakes and civilizations of the Indus Valley Bilham, R., 1999, Slip parameters for the Rann of Kachchh, India, 16 June 1819 earthquake, quantified from contemporary accounts, in Stewart, I.S., and Vita-Finzi, C., eds., Coastal Tectonics: Geological Society of London Special Publication 146, p. 295–319. Bilham, R., Lodi, S., Hough, S., Bukhary, S., Khan, A.M., and Rafeeqi, S.F., 2007, Seismic hazard in Karachi, Pakistan: Uncertain past, uncertain future: Seismological Research Letters, v. 78, no. 6, p. 601–613, doi: 10 .1785/gssrl.78.6.601. Bisht, R.S., 1991, Dholavira: A new horizon of the Indus civilization: Puratattva, v. 20, p. 71–82. Budge, E.A.W., translator, 1932, The Chronography of Gregory Abu’l Faraj, Commonly Known as Bar Hebraeus: London, Oxford University Press, 2 volumes, 151 p. Burnes, A., 1833, Substance of a geographical memoir on the Indus: Journal of the Royal Geographical Society of London, v. 3, p. 113–156, doi: 10 .2307/1797602. Burnes, A., 1835, Memoir on the eastern branch of the River Indus, giving an account of the alterations produced on it by an earthquake, also a theory of the formation of the Runn, and some conjectures on the route of Alexander the Great; drawn up in the years 1827–1828: Transactions of the Royal Asiatic Society of Great Britain and Ireland, v. 3, p. 550–588. Byrne, D.E., Sykes, L.R., and Davis, D.M., 1992, Great thrust earthquakes and aseismic slip along the plate boundary of the Makran subduction zone: Journal of Geophysical Research, v. 97, p. 449–478, doi: 10.1029/91JB02165. Cousens, H., 1906, Brahmanabad-Mansura in Sind: Annual Report 1903–04: Calcutta, Archaeological Survey of India, Office of the Superintendent of Government Printing, p. 133–144. Cousens, H., 1912, Excavations at Brahmanabad-Mansura in Sind: Annual Report 1908-09: Calcutta, Archaeological Survey of India, Office of the Superintendent of Government Printing, p. 79–87. Engdahl, E.R., Bergman, E.A., Myers, S.C., Jackson, J.A., and Priestley, K., 2005, Relocation and assessment of seismicity in Iran region: Eos (Transactions, American Geophysical Union), v. 86, Fall meeting supplement, abstract 52T51-C-1352. Fredunbeg, M.K., 1902, The History of Sind, Volume 2: Karachi, Printed at the Commissioner’s Press, 346 p. Ghafur, M.A., 1966, Fourteen Kufic inscriptions of Banbhore: Pakistan Archaeology, no. 3, p. 65–90. Guidoboni, E., and Traina, G., 1995, A new catalogue of earthquakes in the historical Armenia area from antiquity to the 12th century: Annali di Geofisica, v. 38, no. 1, p. 85–147. Gupta, S.K., 1975, Silting of the Rann of Kutch during Holocene: Indian Journal of Earth Sciences, v. 2, p. 163–175. Hoff, K. von, 1840, Geschichte der durch uberlieferung nachgewiesenen naturlichen veranderungen der erdoberflache, Volume 4: Gotha, Justus Perthes, 275 p. Hough, S., Martin, R., Bilham, R., and Atkinson, G., 2002, The 26 January 2001, Bhuj India earthquake: Observed and predicted ground motions: Bulletin of the Seismological Society of America, v. 92, p. 2061–2079, doi: 10.1785/0120010260. International Seismological Centre, 2006, International Seismological Centre Bulletin Thatcham, United Kingdom: http://www.isc.ac.uk/ (accessed 14 July 2010). Jarrett, M., H.S., 1881, History of the Caliphs by Jalalu’ddin A’s Suyuti: Calcutta, Asiatic Society, p. 387–388. Joshi, V.P., and Bisht, R.S., 1994, India and the Indus Civilization: New Delhi, National Museum Institute, p. 22–31. Khan, F.A., 1960, Preliminary Report on Banbhore Excavations: Karachi, Government of Pakistan, Department of Archaeology. Lal, B.B., 1998, India 1947–1997: New Light on the Indus Civilization: New Delhi, Aryan Books International, 135 p. Lal, B.B., Joshi, J.P., Thapar, B.K., and Bala, M., 2003, Excavations at Kalibangan: The Early Harappans (1960–1969): Archaeological Survey of India Memoir 98, 340 p.

127

Lawrence, R.D., Khan, S.H., and Nakata, T., 1992, Chaman fault, PakistanAfghanistan: Annales Tectonicae, v. 6, p. 196–223. Nakata, T., Tsutsumi, H., Khan, S.H., and Lawrence, R.D., 1991, Active Faults of Pakistan, Map Sheets and Inventories: Hiroshima University Research Center for Regional Geography Special Publication 21, 141 p. National Earthquake Information Center, 2006, National Earthquake Information Center: http://neic.usgs.gov/neis/epic/ (accessed 14 July 2010). Oldham, T., and Oldham, R.D., 1883, A catalogue of Indian earthquakes from the earliest time to the end of A.D. 1869: Memoirs of the Geological Survey of India, v. 19, pt. 2, p. 163–213. Page, W.D., Alt, J.N., Cluff, L.S., and Plafker, G., 1979, Evidence for the recurrence of large magnitude earthquakes along the Makran coast of Iran and Pakistan: Tectonophysics, v. 52, p. 533–547, doi: 10.1016/0040-1951 (79)90269-5. Possehl, G.L., 1999, Indus Age, the Beginnings: Philadelphia, University of Pennsylvania Press, 1063 p. Possehl, G.L., 2002, The Indus Civilization: A Contemporary Perspective: Walnut Creek, California, AltaMira Press, 276 p. Quittmeyer, R.C., and Jacob, K.H., 1979, Historical and modern seismicity of Pakistan, Afghanistan, northern India and southern Iran: Bulletin of the Seismological Society of America, v. 69, p. 773–823. Raikes, R.L., 1968, Kalibangan, death from natural causes: Antiquity, v. 42, no. 168, p. 286–291. Rajendran, K., and Rajendran, C.P., 2003, Seismogenesis in the stable continental regions and implications for hazard assessment: Two recent examples from India: Current Science, v. 85, no. 7, p. 896–902. Regard, V., 2005, Cumulative right-lateral fault slip rate across the ZagrosMakran transfer zone: Role of the Minab-Zedan fault system in accommodating Arabia-Eurasia convergence in southeast Iran: Geophysical Journal International, v. 162, no. 1, p. 177–203, doi: 10.1111/j.1365-246X .2005.02558.x. Sella, G.F., Dixon, T.H., and Mao, A., 2002, REVEL: A model for recent plate velocities from space geodesy: Journal of Geophysical Research, v. 107, p. 2081–2111, doi: 10.1029/2000JB000033. Singh, B.P., ed., 1996, Indian Archaeology 1991–92—A Review: New Delhi, Archaeological Survey of India, 105 p. Sivewright, R., 1907, Cutch and the Ran: The Geographical Journal, v. 29, p. 518–539, doi: 10.2307/1776171. Sykes, W.H., 1857, Relics from the buried city of Brahmunabad in Sind: Illustrated London News, v. 30, no. 846, p. 166–167. Tornberg, C.J., 1865, Ibn-el-athiri, chronicon quod perfectissimum inscribitur (12 volumes in Arabic): Leiden, E.J. Brill, v. 7, 465 p. Vernant, P., Nilforoushan, F., Hatzfeld, D., Abbassi, M.R., Vigny, C., Masson, F., Nankali, H., Martinod, J., Ashtiani, A., Bayer, R., Tavakoli, F., and Chéry, J., 2004, Present-day crustal deformation and plate kinematics in the Middle East constrained by GPS measurements in Iran and northern Oman: Geophysical Journal International, v. 157, p. 381–398, doi: 10 .1111/j.1365-246X.2004.02222.x Vita-Finzi, C., 1980, 14C dating of recent crustal movements in the Persian Gulf and Iranian Makran: Radiocarbon, v. 22, no. 3, p. 767–773. Wells, D.L., and Coppersmith, K.J., 1994, New empirical relationships among magnitude, rupture length, rupture width, rupture area, and surface displacement: Bulletin of the Seismological Society of America, v. 84, p. 974–1002. West, W.D., 1934, The Baluchistan Earthquakes of August 25th and 27th, 1931: Geological Survey of India Memoir 67, no. 1, p. 1–82. West, W.D., 1936, Preliminary geological report on the Baluchistan (Quetta) earthquake of May 31st, 1935: Records of the Geological Survey of India, v. 69, p. 203–240. Wilhelmy, H., 1969, Das urstromtal am istrand der Indusebene das Sarasvati- problem: Zeitschrift für Geomorphologie, Supplementband, v. 8, p. 76–93. MANUSCRIPT ACCEPTED BY THE SOCIETY 19 MAY 2010

Printed in the USA

The Geological Society of America Special Paper 471 2010

Comparing semiquantitative logic trees for archaeoseismology and paleoseismology: The Baelo Claudia (southern Spain) case study Christoph Grützner* Klaus Reicherter* Institute of Neotectonics and Natural Hazards, RWTH Aachen University, Lochnerstrasse 4-20, 52056 Aachen, Germany Pablo G. Silva Departamento de Geología, Universidad de Salamanca, Escuela Politécnica Superior de Ávila, Calle Hornos Caleros, 50, 05003 Ávila, Spain

ABSTRACT The Bolonia Bay, close to the Strait of Gibraltar, hosts the Roman ruins of Baelo Claudia. In the first and third century A.D., this ancient town was affected by two earthquakes. Several earthquake-related damages can be found inside the ruins, and the adjacent mountain ranges show features of Quaternary activity. Extensive paleoseismological and archaeoseismological investigations have been conducted at the archaeological site and in its environs. The first 14C dating results from damaged infrastructure are presented in this paper, together with the preliminary results of fault-trenching on one of the closest suspect seismogenic faults near the archaeological site. The observations have been quantified using the two logic trees for paleoseismology and archaeoseismology. Our results show that a mere paleoseismological classification of the geological features leads to a paleoseismic quality factor (PQF) of 0.03, which is low compared to other studies. Taking into account the additional information from archaeoseismological work (archaeoseismological quality factor [AQF] is 0.5), it becomes clear that the Baelo Claudia study site provides an opportunity for detailed earthquake investigations. Therefore, it has a high potential for reliable seismic hazard analyses. A complementary application of both logic trees is recommended in future studies if sufficient data are available. INTRODUCTION When searching for historical and prehistorical earthquakes, many uncertainties due to the natural limits of the archives and methods need to be taken into account. When employing written records of historical events, one has to take into account inaccurate or biased reports, incomplete lists, geographic deviations,

political interest, event doubling, unreliability of witnesses, etc. (Guidoboni and Traina, 1996). Paleoseismological methods are used to evaluate prehistorical events; however, these techniques are necessarily confined to the geological environment that preserves the record and to the resolution available by the methods applied (McCalpin and Nelson, 1996). Archaeoseismology is not exactly in between those two methods, but

*E-mails: [email protected]; [email protected]; [email protected].

Grützner, C., Reicherter, K., and Silva, P.G., 2010, Comparing semiquantitative logic trees for archaeoseismology and paleoseismology: The Baelo Claudia (southern Spain) case study, in Sintubin, M., Stewart, I.S., Niemi, T.M., and Altunel, E., eds., Ancient Earthquakes: Geological Society of America Special Paper 471, p. 129–143, doi: 10.1130/2010.2471(12). For permission to copy, contact [email protected]. © 2010 The Geological Society of America. All rights reserved.

129

130

Grützner et al.

it certainly combines some advantages and disadvantages of both disciplines (Ambraseys et al., 2002; Galadini et al., 2006; Reicherter et al., 2009). For seismic hazard analysis, reliable data on earthquake location, magnitudes, and recurrence intervals are required. The most precise data are provided by instrumental seismology, covering only the past 100 yr, and even less in most places. Historical archives are often far less reliable but reach back to the beginning of written records. Finally, information on earthquakes in a geological time frame can be obtained by paleoseismology, which is often lacking the reliability and accuracy necessary for precise seismic hazard analysis. Here, archaeoseismological research can extend the time span of comparatively accurate information on old events. While the seismotectonic setting of the research area has to be investigated in terms of paleoseismology, information about use, construction, and the response of the archaeological buildings provided by archaeologists, architects, and engineers, geologists, and historians needs to be acquired in order to understand the site’s potential as an archaeoseismological record. With such a vast number of variables and uncertainties, archaeoseismology is often thought to be inaccurate and therefore is questioned as to how reliably the events described might be used in seismic hazard analyses. These contradictions are important and necessary, because they force scientists to review their own work, to find weaknesses in their arguments, and to accept and develop new methods that can resolve doubts. During the last years, several proposals have been implemented in order to establish a stringent and transparent methodology and provide a quantitative analysis of archaeoseismic damages (Karcz and Kafri, 1978; Nikonov, 1988; Guidoboni and Traina, 1996; Stiros, 1996; Galadini et al., 2006; Caputo and Helly, 2008; Sintubin and Stewart, 2008; Reicherter et al., 2009). Sintubin and Stewart (2008) created a logic tree approach on archaeoseismology based on the one proposed for paleoseismology by Atakan et al. (2000). This approach is supposed to “track uncertainties in successive stages of archaeoseismological investigation” (Sintubin and Stewart, 2008, p. 2209) and might therefore constitute an important step on the way to a systematization of the methods employed in researching ancient earthquakes. In this study, we apply the logic trees of Atakan et al. (2000) and Sintubin and Stewart (2008) to a site where both paleoseismological and archaeoseismological investigations have been conducted. We test the logic trees and aim to improve the database necessary for comparative site potential estimations. Furthermore, we show that a joint analysis of both approaches can improve the assessment of a particular site in terms of earthquake investigation. The paleoseismological observations described in the following sections are based on field work conducted by the authors between 2005 and 2009 and particularly on the results of the fault-trenching studies carried out in 2008 in the vicinity of Baelo Claudia. Archaeoseismological data have been published by Silva et al. (2005, 2006, 2009) and were mainly collected by the authors during the same time span.

BAELO CLAUDIA AREA Geology and Tectonics The Roman ruins of Baelo Claudia are situated on Bolonia Bay (Cádiz, south Spain), within the Strait of Gibraltar area (Fig. 1). In the westernmost part of the Gibraltar arc, which was formed by the assemblage of the Betic Cordilleras and the Moroccan Rif, the convergent plate boundary between Eurasia and Africa is less clearly defined than in the Eastern Mediterranean (Stich et al., 2006). Instrumental seismicity records list only 14 events with magnitudes >4.5 since 1929 (IGN) in the working area (latitude 35°N–37°N, longitude 4.5°W–6.5°W). Intensities have not exceeded VII; the strongest event reached M = 5.4 in 1976 and occurred on the Moroccan side of the Strait of Gibraltar. Compared with the entire south Iberian margin, the study area marks a seismic gap in the plate-boundary zone. Historical records mention earthquakes with local intensities greater than MSK VIII in southwestern Andalusia and in the Gulf of Cádiz (compiled by Reicherter, 2001). Silva et al. (2006) described several faults in Bolonia Bay. The SW-NE–trending Cabo de Gracia fault and the La Laja fault have a clearly expressed morphology, but their active phase might date back to the late Pliocene–early Pleistocene. However, there are few hints for more recent movements, since this fault has vertically displaced late Pleistocene marine terraces belonging to the oxygen isotopic substage (OIS) 5c, isotopically dated here ca. 125 ka (Zazo et al., 1999). On the other hand, N-S–trending normal faults in the Gibraltar Strait area account for moderate instrumental seismicity in this zone (Goy et al., 1995; Silva et al., 2006). The Carrizales, La Laja, and San Bartolome range-front faults display examples of Quaternary activity around Bolonia Bay (Silva et al., 2009) in a radius of 5 km from Baelo Claudia (Fig. 2). These faults show evidence for Quaternary paleoseismological events and offset prominent mountain ranges, resulting in step-shaped profiles of the ridges. All these faults also display relevant evidence of related landslides, massive rockfalls, liquefaction, and differential ground subsidence (Silva et al., 2009), which in some cases can be catalogued as secondary or sympathetic earthquake ground effects in the classification developed for the Environmental Seismic Intensity Scale (ESI-2007) developed by Michetti et al. (2007). Neotectonics of the study area were described in detail by Silva et al. (2006) (Figs. 2 and 3). Intensely folded Tertiary flysch deposits of the Aljibe, Algeciras, and Bolonia sandstone formations (Eocene to Aquitanian) dominate the study area. During the Alpine orogeny (Burdigalian to late Tortonian), they were overthrust by turbiditic sediments of Cretaceous to Eocene age (Sanz de Galdeano, 1990; Weijermars, 1991; Silva et al., 2006). The Facinas and Almarchal formations of the flysch complex consist of very plastic clayey and sandy layers, which are involved in the large number of mass movements observed in the study area (Fig. 2). Bolonia Bay is surrounded by three steep mountain ranges of Aquitanian sandstone: the Cabo de Gracia and La Laja mountain

The Baelo Claudia (southern Spain) case study 7°00’ W

131 4°00’

5°00’ W

6°00’ W

Córdoba

37°45’ N

37°45’ N

Seville 37°15’ N

Huelva

36°45’ N

Cádiz

Málaga

SRTM outline

Fig. 2 Bolonia Bay

Tarifa

ait Str

of

r

a ralt b i G

Figure 1. Bolonia Bay, location of the Roman remains of the Baelo Claudia village close to the Gibraltar Strait in southern Spain (SRTM data). Legend: 1—Upper Miocene–Quaternary sedimentary rocks; 2—Subbetic unit; 3—Internal Subbetic; 4—Prebetic; 5—Oligocene–Lower Miocene sedimentary rocks including flysch and olistostromes; 6—Campo de Gibraltar Flysch; 7—Predorsalian, Dorsalian, and Maláguide complexes; 8—Alpujárride complex; 9—Nevado-Filábride complex; 10—Iberian Massif with cover rocks. See Figure 2 for the regional tectonic setting. Figure is modified after Reicherter et al. (2003).

ranges in the west and the San Bartolome mountain range in the east (Fig. 3). All of them are significantly affected by landslides and extraordinary rockfall events (Höbig et al., 2009; Vollmert et al., 2009). The prominent La Laja and Cabo de Gracia mountain fronts, the morphotectonic features of which are discussed later with regard to the Atakan logic tree, form kilometer-scale lineaments (Silva et al., 2009). At the latter, slickensides were observed. Several studies describe the Quaternary tectonic setting (e.g., Zazo et al., 1999; Silva et al., 2006) and paleoseismological and archaeoseismological records at the Roman ruins (Goy et al., 1994; Sillières, 1997; Alonso-Villalobos et al., 2003; Silva et al., 2005, 2006, 2009). The Roman Ruins of Baelo Claudia The Roman ruins of Baelo Claudia are the remains of a small coastal town with a population of 2000 persons in an area of ~0.5 km2, the economic welfare of whom mainly stemmed from

the fishing industry. Its strategic significance is evident through the potential to control parts of the Gibraltar Strait. Prior to the Roman period, no urban settlement was established in the area, albeit tombs in the vicinity give evidence for Neolithic settlements. The small Roman town of Baelo Claudia was founded in the late second century B.C. (described by Strabo in A.D. 17) and was an important strategic and industrial part (tuna fishing, saucemaking, and olive-oil-pressing industries as well as iron-smelting industries) of the Roman Empire. Moreover, it used to be the gateway to Tingis—the modern Tangiers in Morocco. Before the town was founded, a small well-sheltered oppidum was situated in the Sierra La Laja (Silla de Papa). In the last century, Baelo Claudia has been extensively excavated by French archaeologists from the Casa de Velázquez in Madrid (Sillières, 1997). Thanks to these efforts, Baelo Claudia is one of the most complete and best-studied Roman towns in Spain, and its main features are an excellent example of early imperial town planning (Fig. 4). The central parts and buildings have been dated to the early first century A.D., under the reign of Emperor Claudius (A.D. 41–54),

0

5

10 km

La Laja fault El Almarchal

N

Facinas

Zahara

Cabo de La Plata Cabo de Gracia

1 2

CdG fault

3

Bolonia Bay

Carrizales fault

Baelo Claudia

S.Bartolome Mt. front

4 5 Tarifa

At l a n t i c O c e a n

6

Figure 2. Neotectonics and regional geology of Bolonia Bay and its surroundings. Legend: 1—El Almarchal plastic clays; 2—Facinas plastic clays; 3—El Aljibe Flysch nappe (predominantly sandstones), traces of Betic uprighted stratification planes are marked with dashed lines; 4—flysch units activated during the neotectonic period; 5—Pliocene and Quaternary deposits (postcollisional); 6—landslides; CdG—Cabo de Gracia Fault. Figure is modified after Silva et al. (2006). See Figure 1 for location and Figure 3 for a detailed map.

0

1

2

3

5 km

N

Minerva Jupiter Juno

La Laja fault

Cabo de Gracia fault

Forum

Carrizales fault Trenching site T2

ck Ro F G Cd

ck Ro

San Bartolome mountain front

T1

T3

Isis

Roman ruins of Baelo Claudia

Figure 3. Simplified map of Bolonia Bay, showing the location of the three trenches and the outline (city wall) of the archaeological site. A most likely fault-bounded spring is situated at the Cabo de Gracia fault (CdGF) in the trenching area, draining to the north. Geophysical investigations have been carried out at the trenching site and covered almost the entire area outlined by the rectangle in the lower left (GPR and DC-Geoelectrics). Along the La Laja and the San Bartolome, additional measurements have been taken out. See Figure 2 for the legend.

Basilica

Macellum

Decu

man

us M axi

mus

N

Figure 4. Aerial photo of the central or Capitol area of Baelo Claudia, showing the main features of the imperial town. Note the Isis temple, which is off to the side from the geometrical town center with Minerva, Jupiter, and Juno temples, the forum, and the basilica.

The Baelo Claudia (southern Spain) case study during which Baelo became a municipium. Therefore, Claudia was added to the name. Our most recent studies give evidence for at least two strong earthquakes in the first and the third centuries A.D., which destroyed parts of the infrastructure and eventually forced the Romans to gradually abandon the town (Silva et al., 2005, 2009). Excavations of the extramural necropolises, however, showed that at least parts of the town were in use during the fourth to sixth centuries A.D. After the second earthquake in the third century A.D., human activity in the town declined (Silva et al., 2009), and refuse and colluvial deposits began to accumulate in the lower monumental sector of the city, which includes the Theatre, the Decumanus Maximus (main E-W road), the Capitol area, and the Roman Basilica (Collins, 1998). The Basilica and the Roman Market (Macellum) were no longer in use, and the aqueducts supplying water to the city were severely damaged, resulting in water shortage (Sillières, 1997). The findings of the most recent coins point to the reign of Constantine the Great (A.D. 306–337), which is more or less coeval with a major crisis of the Western Roman Empire (Sillières, 1997). Later, Visigothic graves carved around the ancient city wall indicate settling in the city area until the conquest of the Moors in A.D. 711, who built a military base on top of the Roman theater in Baelo Claudia. Many different indicators of archaeoseismic damage in various places within Baelo Claudia have been described and documented, and we were able to date those with archaeological findings (e.g., pottery remains and coins) and 14C dating. These allow a tentative bracketing of the occurrence of repeated strong archaeoseismic damage (intensity ≥IX MSK) at Baelo Claudia to around A.D. 40–60 and A.D. 260–290 (Silva et al., 2009). The Isis Temple Problem The Capitol, devoted to Jupiter, Minerva, and Juno, which consists of three almost identical temples, dominates the forum and central area of the Roman village (Fig. 4). These temples are thought to have been constructed between A.D. 50 and 70—during the first imperial reconstruction period in Baelo Claudia—to represent the official religion (Sillières, 1997). Aside from the central temple area, a temple dedicated to the Egyptian goddess Isis (Figs. 4 and 5) was built in the adjacent area. Especially after the annexation of Egypt in 30 B.C., the Roman Empire adopted the Isis cult and spread it throughout their entire domain. During the reign of Emperor Caligula (A.D. 37–41), the cult was eventually established in the entire Roman Empire. Archaeological studies assumed that the Isis temple of Baelo Claudia was built around the year A.D. 70, with dimensions of 29.85 m × 17.70 m (Fig. 5). The sanctuary was divided into a public cultural area with a cella and private rooms. Currently, the area of the Isis temple is only partly excavated (Fig. 6). As with the other three temples, the Isis temple was built exactly at the toe of a small topographic scarp and displays a set of particular structural deformations that can be attributed to a shallow landsliding event presumably triggered by the second episode of damage recorded in the city during the third century A.D. (Silva et al., 2009).

133

The Isis temple is one of the few buildings of the city in which evidence for two destruction events can be assessed. The temple is partially buried by a colluvial deposit of 2.5–3 m thickness and displays some of the best examples of ground-failure building damage. Beside ceramics, glass shards, shells, and bones, the colluvium also contains slag of iron smelting. A rough stratigraphy based on Roman pottery allows dating the latter collapse event to the third century A.D. The base of the colluvial deposits was dated to be 1955 ± 30 yr B.P., pointing to a first damage event in the first century A.D. (see Table 1 for the dating results). The sample from the top of the section (Fig. 6) yielded an age of 1725 ± 25 yr B.P. This top soil sample was taken from directly underneath a toppled stucco plastered wall. Radiocarbon dating indicates that (1) the construction of the Isis temple occurred earlier than previously assumed by archaeological studies (e.g., Sillières, 1997), and (2) refuse accumulated beside the Capitol area during no more than 200–250 yr. Additionally, another set of soil samples from directly underneath the fallen columns and walls within the Isis temple was collected. Fragments of toppled columns, wall, and pillar collapses are predominantly oriented in a SW direction (Silva et al., 2009), but also parallel to the aforementioned landslide. Dating of soil samples below the column fragments (Fig. 6) yields radiocarbon ages between 2020 ± 25 yr B.P. and 1195 ± 30 yr B.P. These two dates are relatively consistent and related to the destruction of the temple by seismic activity (2020 ± 25 yr B.P. and 1900 ± 30 yr B.P.). The very young age may be interpreted as having originated by later modifications (quarry use) during the Muslim Spanish Period. Taking this into account, we are able to ascribe the event of the destruction of the Isis temple to the first century A.D., relatively soon after its construction, followed by immediate reconstruction, as previously described for other structural elements such as the city wall (Sillières, 1997). However, parts of the temple area were used as a contemporary refuse tip. During the second damage event in the third century A.D., remains of the damaged walls of the temple eventually collapsed on top of the tip. After this episode, the Capitol area was covered by a 1.5–2.0-m-thick post-Roman colluvium (no longer present) and thus was protected from erosion processes, except for the area where the suspected quarry reutilization was carried out during the Muslim period. PALEOSEISMOLOGICAL OBSERVATIONS Application of Atakan’s Logic Tree and UNIPAS v. 3.0 to the Baelo Claudia Site Atakan et al. (2000) created a logic tree approach to quantify the uncertainties related to paleoseismological investigations. At each node, at least two alternatives with their respective uncertainties can be described. The tree takes into account six basic criteria (Atakan et al., 2000, p. 416): “1. tectonic setting and strain-rate; 2. site selection for detailed analysis (site selection criteria); 3. extrapolation of the conclusions drawn from the detailed site analysis to the entire fault; 4. identification of individual

134

Grützner et al.

Sample points

0

10

30

50 m

N

Min2

Min1

Isis1

Isis5

Isis2 Isis4

Isis3

Figure 5. Detailed plan of the Isis temple area, with the mapped archaeoseismic damage and sample points for 14C dating. Small arrows with lines indicate cracks and opening directions; big arrows indicate deformation directions of walls; light gray shaded areas near Isis1 and Isis2 indicate debris of fallen walls, dark gray shaded areas indicate pavement. Min1 and Min2 indicate the position of the Minerva temple sampling points; Isis1–Isis5 indicate the position of the Isis temple sampling points.

paleo-earthquakes (diagnostic criteria); 5. dating of paleo-earthquakes (type of technique); 6. paleo-earthquake size estimates (slip on individual events, correlation between trenches).” These single steps then lead to a final joint probability of the entire estimation, the probability of the preferred end solution, called Pes. For the implementation of a paleoseismological study in seismic hazard analyses, it is necessary to introduce a correction term Cri, which describes the relative importance of the study. In combination with the correction term Cri, a paleoseismic quality factor (PQF) can be calculated as follows: PQF = Pes × Cri. Atakan et al. (2000) introduced the UNIPAS v. 3.0 program, which can be downloaded from the webpage of Bergen University (http://www.geo.uib.no/seismo/software/unipas/ unipas1.html) to facilitate and automate the calculations.

Tectonic Setting and Strain Rate The study area is located at the Eurasia-Africa plate boundary (Fig. 1), where the convergence rate is ~4 mm/yr (DeMets et al., 1990). According to the suggestions of the Atakan’s logic tree, this corresponds to a quality weight factor of 0.8–1.0 (plate boundaries, high strain rates). In the environs of Baelo Claudia, moderate earthquakes (Silva et al., 2009) occur; however, they are not as common as along other plate margins (see Figs. 2 and 3 for the regional tectonic setting). Background knowledge is available and categorized as intermediate. In the field, geomorphological features such as the Cabo de Gracia fault and the La Laja fault are clearly visible from a great distance (Fig. 7). Stepped mountain ridges are interpreted as the morphological expression of the

The Baelo Claudia (southern Spain) case study

135

Isis1: 1725 ± 25 yr B.P.

Isis5: 2020 ± 25 yr B.P. Isis2: 1955 ± 30 yr B.P.

Isis3: 1900 ± 30 yr B.P.

Isis4: 1195 ± 30 yr B.P.

Figure 6. Results of 14C dating and photos of sample locations in the Isis temple area. Here, several toppled columns have a similar orientation. Those columns fallen in the oldest event fell on a clean floor, indicating that the building was not abandoned at that time. The ones giving younger 14C ages fell on colluvium accumulated after the first earthquake.

Name Isis1 Isis2 Isis3 Isis4 Isis5 Min1

Min2

TABLE 1. DATING RESULTS FROM THE TEMPLE AREA Laboratory code Sample Material KIA32686 Isis-1 Soil Sediment, leaching residue Sediment, humid acid KIA32687 Isis-2 Soil Sediment, leaching residue KIA32688 Isis-3 Soil Sediment, leaching residue Sediment, shell KIA32689 Isis-4 Soil Sediment, leaching residue KIA32690 Isis-5 Soil Sediment, leaching residue KIA32691 Min-1 Roots Sediment, leaching residue Sediment, humid acid Charcoal, leaching residue KIA32692 Min-2 Soil Sediment, leaching residue Sediment, humid acid

Carrizales fault. The visibility of paleoseismological features is considered as low to intermediate. Generally, official seismic hazard maps show low to moderate values in Bolonia Bay, while the Sagalassos study site in Turkey, used by Sintubin and Stewart (2008) to check the archaeoseismological logic tree proposal, is situated in a high hazard region. However, an evaluation based on hazard map data leads to circular reasoning as the paleoseismological investigations are not normally included in the mapped hazard estimations. Subse-

Age (radiocarbon) 1955 ± 30 yr B.P. 1885 ± 25 yr B.P. 1725 ± 25 yr B.P. 365 ± 20 yr B.P. 1900 ± 30 yr B.P. 1195 ± 30 yr B.P. 2020 ± 25 yr B.P. 390 ± 25 yr B.P. 1070 ± 40 yr B.P. 1025 ± 25 yr B.P. >1954 A.D. >1954 A.D.

quently, we assume a QWF of 0.67 as derived from the parameters implemented in the UNIPAS v. 3.0 program. Site Selection Ground penetrating radar (GPR) and geoelectric resistivity measurements were conducted along the mountain ranges in order to visualize faults and fault-related features. Most of the GPR data show the accumulation of debris at the base of steep

136

Grützner et al.

W

E

A

Sierra de la Plata

S

N

B

Silla de la Papa La Laja

Figure 7. Geomorphological features associated with neotectonic activity. (A) Georadar and geoelectrical measurements were carried out at the Cabo de Gracia fault prior to trenching; trench T1 is located perpendicular to the linear rock face in the left of the image (see Fig. 3 for location). The Sierra de la Plata cliff is visible in the background. (B) The La Laja fault as seen from San Bartolome. Hanging valleys have formed along the mountain range, shaping triangular facets (black lines). Silla de la Papa in the background is the highest peak in the Bolonia Bay area (430 m).

scarps (La Laja and Sierra de la Plata; Fig. 7) and underlying sandstones and clays subject to differential weathering. Rockfall and colluvial deposits can be identified; however, the occurrence of faulting-related colluvial wedges could not be clarified peremptorily. Only one prospection site allowed crossing the expected fault zone (Fig. 7). Due to the high conductivity of shallow weathered clays, data quality is poor for some locations. Two-dimensional (2-D) geoelectric tomography shows a clear change of apparent resistivity on a cross profile perpendicular to the suspected fault zone. This feature is a result of increased soil moisture due to a nearby spring, which is most likely faultrelated. An offshore seismic survey conducted in 2006 (Meteor Cruise M69/1) concentrated on the proposed southwest extension of the Cabo de Gracia fault (Silva et al., 2006). The data show evidence for Quaternary faulting and allow us to determine the length of the fault. Geomorphological analyses yield clear hints for fault activity, such as large linear lineaments (Fig. 7; see also Silva et al., 2009) with lichen-free ribbons at the base of the mountain fronts, triangular facets, liquefaction structures, and fresh slickensides close to the shore. The Carrizales fault offsets the Cabo de Gracia escarpment at various places and could therefore be responsible for the observed slickensides. Activity during Holocene times is likely and the two faults join at the trenching site. According to Michetti et al. (2005), despite these features, Bolonia Bay cannot be classified as a “classic” seismic landscape with clear primary surface ruptures. However, the zone displays a large number of features related to secondary and sympathetic earthquake ground effects (Silva et al., 2009). The trenching site was chosen at the location where most of the geophysical investigations were conducted and with regard to accessibility. As mentioned already, there is only one location suitable for crossing the suspected fault zone with geophysical measurements and trenching. The observations made correspond

with a QWF of 0.6–0.8 in the logic tree. The lack of shallow seismics and onshore deep reflection studies, the poor quality of some GPR profiles, and the fact that the fault-trenching sites had to be located at a very specific site and close to each other, resulted in a decrease of the QWF to 0.6, as computed by the UNIPAS v. 3.0 algorithm. Data Extrapolation The fault-trenching sites were located at the only site suitable for excavation and within a range of 100 m of each other (Fig. 3). The average depth of the open trenches was 2.2 m, with a total research length of 44 m. The fault is supposed to have a length of no more than 10 km (onshore and offshore). Instrumental seismicity in the area shows shallow and intermediate earthquakes at depths not exceeding 60 km, with a concentration of hypocenters at depths between 10 and 15 km (Silva et al., 2006). With these values, the trench to fault ratio (TFR) calculated is 0.00000015, corresponding with QWF <0.2. However, the distribution of the trenches along the fault, which is poor in our case, has to be taken into account in these calculations. The UNIPAS v. 3.0 algorithm computed a QWF of 0.37, which appears high compared to the values listed in Atakan et al. (2000). With regard to the depth of the suspected fault zone, we assign a much lower QWF of 0.2. Identification of Paleoearthquakes In the three trenches opened in February and March 2008, several indications for earthquake events were observed. Main indicators are highly deformed and offset upper clayey soils (10R 4/6 and 5YR 4/4 in Fig. 8), and the underlying weathering profiles on clays (mainly the basal 10YR 6/6 unit), which display a highly significant number of slickensides and lineaments. These

The Baelo Claudia (southern Spain) case study

of broken and rotated sandstone strata and big blocks in the main range-front faults of the area as a result of earthquake-triggered rockfalls provides further evidence. Until now, it has not been possible to correlate these features to a distinct event or multiple earthquakes. Nevertheless, several indicators strongly point to Quaternary paleoseismic activity: liquefaction structures; ruptured and displaced pebbles and slickensides found at adjacent outcrops of early Pleistocene littoral conglomerates and middle Pleistocene alluvial-fans deposits (Carrizales fault; Silva et al., 2006); rockfalls with extraordinarily long run-out distances (San Bartolome); geomorphological features (La Laja and Sierra de La Plata faulted mountain fronts); and a most likely fault-related spring (Cabo de Gracia fault). Nonseismic phenomena that could result in some of the observed features include soil creeping, groundwater-driven movements, and mere gravity rockfalls. Liquefaction and striae cannot be explained by other means. The unusually large rockfall run-out distances at the San Bartolome (up to twice as far as calculated by various simulations; Höbig et al., 2009) and the occurrence of rock avalanches in the same area might be a result of postrockfall landslides and piggyback transport of the debris. As to the steep and lineated mountain ranges and their hanging valleys, fault activity seems to be the only reasonable process (Fig. 7). A QWF of 0.4–0.6 is suggested for common abundance of nonseismic features, and thus, we decided to choose a value of 0.5.

apparently deformed basal units are buried by a gravel-rich layer of gray, sandy-silty material topped by a thin, very recent soil (Fig. 8), which is apparently undeformed. The lowermost weathered sandstones and clays at a depth of 2 m display a complex pattern of gray banding and mottling in all trenches, suggesting strong periodical variations of the water table in the development of the weathering profiles. Slickensides, sliding planes, bedding structures, fault gouges, and individual fault planes could be identified in trench T3. Additionally, the overlying clayey units showed an apparent centimetric offset probably related to fault activity as illustrated in Figure 8. The structures highlighted are representative of trenches T2 and T3. Trench T1 also displayed indicators of recent deformation (some slickensides and sliding planes adjacent to the rock face shown in Fig. 7); however, the features observed were not as clear as in the other excavations. Trench 1 was excavated in a linear NE-SW rock face in order to check activity on the major lithological contacts within the Aljibe Sandstone formations, which displayed some features of sympathetic activity (i.e., centimetric basal lichen-free ribbons). The trench showed few features typical of a bedrock fault scarp, including overturned and rotated sandstone slabs at its basal zone. The near-vertical bedrock face displayed additional slickensides, but mainly exhibited paleocurrent marks. The features observed are difficult to interpret due to the intense deformation of this sandstone unit during the Alpine orogeny. The occurrence

W 0

137

Cabo de Gracia - Trench 3 - North Wall

0

1

3

2

5

4

6

E 8

7

9

10 m

roots

1

photograph outline b181/18 b187/37 b190/46

2 5Y 7/2

5YR 4/4

5YR 5/2

5YR 6/6

b170/38

b156/35

b184/51 b172/59

7.5YR 6/4

10R 4/6

b190/46 b236/39 b159/31 b190/35 b203/36 f138/81 f194/41 b216/42 b226/42 b194/34 b196/41 fault b201/42 f223/50 b190/43 gouge b188/39 f234/41 f214/31 f236/46 j190/41 b184/35

10YR 5/4

10YR 6/1

10YR 6/6

280/80

structural data: (f) fault (j) joint (b) bedding

b190/46 b156/35

b159/31 b190/35

b203/36 b190/43

fault gouge

b194/34

b196/41

f138/81

b188/39 f236/46

j190/41

b184/35

f214/31

Figure 8. Sketch of the northern wall of trench T3 and a photographic mosaic of the outlined area with the main faulting features; color code follows the Munsell® color system. Gravel-rich, clayey units underlie the thin topsoil (gray), followed by reddish, brownish, and yellowish clays at greater depths. The lowermost unit consists of dark-yellow clay with a complex pattern of gray intercalations. Those features are characterized by internal bedding structures and striae. Strike and dip of certain features (black dots) are given with “f” for faults, “j” for joints, and “b” for bedding structures. The V-shaped structure in the center of the trench is interpreted as a fault zone.

138

Grützner et al.

Dating of Paleoearthquakes So far, dating of the features observed in the trenches has not been conducted. We dated the aforementioned structural deformation found in the area of the Isis temple. Furthermore, we obtained a radiocarbon age of a tectonically deformed and warped soil layer in the Carrizales outcrop (ancient Roman quarry), ~600 m W of the Baelo Claudia site. There, Pleistocene conglomerates are faulted, including a well-developed clayey red paleosol pervasively affected by hydroplastic deformations generating pocket collapse–like structures, which are filled by dewatered Pleistocene conglomerates and sands topped by a red paleosol (Silva et al., 2009). Radiocarbon dating of the red paleosol yielded 6572 ± 71 yr B.P. as a maximum age. This date is in agreement with the archaeological content of the couplet of eolian sandy layers burying the red soil containing Neolithic and Bronze Age remains (Goy et al., 1995). Furthermore, it is possible to determine a time frame of several thousand years using geomorphic observations and correlations between the trenching sites and other segments of the faults; however, this method is the least accurate. For the QWF, this results in a value <0.4, and we consequently assigned a QWF of 0.25.

an idea of recurrence intervals, as well as maximum magnitude estimations, the importance level is set to 3 (Cri = 6). The resulting paleoseismic quality factor (PQF) is 0.03. Currently, there is an insufficient number of comparable values from other studies. Therefore, it is difficult to draw general conclusions. Classical paleoseismological methods reach their limits in this type of study. Obviously, when compared to Atakan et al. (2000), the low Pes of 0.005 and the PQF of 0.03 imply that additional work is required in order to acquire more reliable earthquake information or to obtain new data sources. Whatever the case, most of the indicators resulting from the archaeoseismological research conducted at Baelo Claudia suggest that the causative epicenter is likely to be located S-SW of the ancient coastal city at the offshore extension of the considered faults (Silva et al., 2006, 2009). In this hypothetical scenario, the fault traces analyzed onshore will yield little to no representative information about the offshore active fault segments. As archaeoseismological results do not contribute to the logic tree discussed here, and offshore data are really scarce to be incorporated to this logic approach, an underestimation of the site’s potential is likely. In fact, the logic tree approach helps to find weak points of an investigation and to plan the further work required to narrow down uncertainties.

Size Estimates of Paleoearthquakes ARCHAEOSEISMOLOGICAL OBSERVATIONS A number of criteria can be used to estimate magnitude: (1) the length of the fault traced onshore and offshore (~10 km; criteria of Wells and Coppersmith, 1994); (2) the occurrence of liquefaction (Castilla and Audemard, 2007) and striated and ruptured pebbles (clastic pebbles in a soft clayey matrix tend to rupture and break during high-energy impulses ascribed to earthquake occurring in the schizosphere [Atwater, 1992]; furthermore, their surfaces are striated and pitted); (3) the maximum subrecent fault offset observed in trenches (0.5 m by warping [Fig. 8] of the recent soil); and/or (4) coseismic landslides and rockfalls observed in the vicinity. Silva et al. (2009) estimated the paleo-intensity of the Baelo earthquakes to be in the order of IX MSK, taking into account the amplification factors due to the poor geotechnical parameters of the ground. Considering the ground amplification impact, the considered intensity value will be equivalent to a paleomagnitude of >5.5 (Levret et al., 1994). Therefore, we assume an event with MW >5.5. Since a combination of primary and secondary evidences was employed, a QWF of 0.5 was chosen.

Sintubin and Stewart Logic Tree

Uncertainties and Application to Seismic Hazard Analysis

The logic tree presented by Sintubin and Stewart (2008) is a modification of the logic tree for paleoseismology (Atakan et al., 2000) and works in a similar way. Accordingly, the first three criteria defined deal with the probability of an earthquake affecting the study site. In the second half of the logic tree, the focus is on the damages observed. The following aspects are to be evaluated: (1) tectonic setting; (2) site environment; (3) site potential; (4) identification of damage; (5) dating of damage; and (6) regional correlation (Fig. 9). The joint probability resulting from the answer to those questions is a value between 0 and 1. Similar to the paleoseismological scheme, a site confidence level (SCL) is introduced. The SCL is divided into seven stages and corresponds to correction term C, which is between 1 and 10. A qualitative evaluation of the excavations and the quality of the excavation reports account for the determination of the SCL. The product of Pes and C then yields the archaeoseismological quality factor AQF, in analogy to the PQF of Atakan et al. (2000).

At the current state of investigations, the Atakan logic tree yields a probability of 0.005 for our solution (Fig. 9). This value is thirty times lower than the one computed on the Bree fault example (Atakan et al., 2000) in Belgium, mainly due to the limited possibility of extrapolating the trench observation to the entire fault, the uncertain earthquake patterns in the trenches, and the lack of dating. Considering that the results of this study will be used to complete the local earthquake catalogue and to give

Tectonic Setting The first criterion is identical to the one in the paleoseismologic tree of Atakan et al. (2000). According to the setting along the active plate boundary of Africa and Eurasia and to the convergence rate of 4 mm/yr, we had to choose a QWF of 0.8–1.0. On the other hand, the plate boundary is diffuse in the study area, the faults in the vicinity are visible but difficult to evaluate, and an M >6 epicenter within a radius of 10 km cannot be proven

The Baelo Claudia (southern Spain) case study

Total number of nodes: 6 Total number of branches: 12 Total number of end solutions: 64

Identification of paleoearthquakes

Extrapolation of data Site selection Tectonic setting 0.67

0.2

0.5

Paleoearthquake Dating size estimation of paleoPes1= 0.005 earthquakes 0.5 6

0.25 5

4

3

0.6

Level of importance 3 (Cr i = 6) PQF= Pes1 x Cri = 0.03

2

1 Atakan et al. (2000) Regional correlation SPF = 0.28

Dating of damage Identification of damage

Site potential Site environment Tectonic setting 0.67

0.6

0.7

0.8

139

0.7

0.4

Pes1= 0.063

6

5

4

3

2

Figure 9. The Atakan et al. (2000) logic tree for paleoseismology consists of 12 branches and six nodes at which certain probabilities must be defined (upper image). The result is the probability of the preferred end solution, Pes. In this study, Pes is 0.005, 30 times lower than the one achieved by Atakan et al. (2000) at the Bree fault example. The paleoseismic quality factor (PQF) is Pes × Cri (Cri is a correction term depending on the level of importance of the study). PQF is 0.03 in our case. UNIPAS v. 3.0 automatically computes the Pes based on the values entered and creates the graph. The logic tree of Sintubin and Stewart (2008) has been applied to the Roman ruins of Baelo Claudia (lower image). The site potential factor (SPF) is 0.28; the resulting overall probability of our preferred end solution Pes1 is 0.063; the archaeological quality factor (AQF) computes to 0.5 (Pes1 × correction term C).

SCL 6 (C = 8) AQF= Pes1 x C = 0.5

1 Sintubin & Stewart (2008)

without any doubts. Background knowledge must be categorized as intermediate, and the tectonic setting points to a value between 0.6 and 0.8. Following the estimations from the previous sections, we assume a QWF of 0.67. Site Environment Sintubin and Stewart (2008) recommend different ways to evaluate the site environment factor. Based on the landscape signature categories introduced by Michetti et al. (2005) (see also Michetti and Hancock, 1997) for active normal faults in the Mediterranean, the QWF can be determined. Another possibility is to use the INQUA (International Union for Quaternary Research) Environmental Seismic Intensity (ESI) scale for categorizing the seismic landscape (Michetti et al., 2007; Reicherter et al., 2009). Bolonia Bay and adjacent areas are situated on an diffuse plate boundary with convergence rates of ~4 mm/yr. Liquefaction structures on Quaternary deposits have been found, as well as a variety of late Pleistocene to historical landslides and rockfalls; faults can be traced over kilometers, and sparse evidence for recent tectonic activity is visible in the field (see also site selection subsection of Paleoseismological Observations). These features correspond to a QWF of 0.6–0.8. In contrast, Bolonia Bay is not a classic seismic landscape with primary surface ruptures, but it displays a large variety of secondary earthquake ground

effects. Environmental earthquake effects recorded in the area do not necessarily point to an M >6 event. Considering the similarity to the Atakan tree, we choose a QWF of 0.6. Site Potential Baelo Claudia was populated by the Romans from the late second century B.C. to the late fourth century A.D. After that, sparse paleo-Christian to Visigothic settlement is documented, and parts of the town’s ruins served as a source for building materials. With the conquest of the Iberian Peninsula, the Moors built a small garrison on the site. Over the centuries, the ruins have repeatedly been used for military purposes (Sillières, 1997; Silva et al., 2009). Anthropogenic disturbances can be observed occasionally due to this history. The upper part of the ruins is situated on a gentle slope consisting of Cretaceous clayey material, which is subject to occasional creeping. Marine terraces are present close to the shore and host the fish factories. Shallow landslides and landsliderelated deformation are documented in some of the upper parts of the ancient town. Ground settlement must be taken into account, but will be small-scaled—if at all existent—due to underground conditions. An extensive GPR survey has been conducted in the ruins to detect shallow landslide structures, to map the event-horizon

140

Grützner et al.

of the seismic events identified, and to estimate the depth of the bedrock (Silva et al., 2009) by the correlation with the different geotechnical drillings performed within the ruins (Silva et al., 2005). Additionally, archaeological remains such as walls, tombs, streets, and houses, as well as damages (fallen boulders, deformed walls), have been imaged. Many buildings of different types are recorded at Baelo Claudia, including brick and masonry houses, a city wall, a cistern, a theater, shops, a market, a forum, a temple area, roads, a necropolis, an entire fish factory district, and many more. Masonry mainly consists of durable Cretaceous thin-bedded and fine-grained sandstones. In contrast, pillars of houses are made of very porous late Tertiary calcarenites and partly Tyrrhenian beach rocks incorporated with mortar. The calcarenites and beach rocks can be easily hewn but are also fragile and vulnerable to erosion, and the ancient quarries are nearby. The third regional construction material is heavy, siliceous sandstone of the Aljibe Formation (Early Tertiary to Mid-Tertiary); however, due to its hardness, it was exclusively used for stairways or raisers. Archaeological investigations have shown that the town was reconstructed after abrupt ruin and depopulation during the Imperial Roman settlement (e.g., walls and houses were reinforced and rebuilt, an artificial terrace was created etc.; details are listed in Silva et al., 2005). After sporadic investigations in the early twentieth century, the extensive systematic excavation began in the 1970s and continues ever since, now conducted by the Autonomic Government of Junta de Andalucía. Since the beginning of the present millennium, there has been a special focus on seismically induced damages in all ongoing archaeological work. Currently, the ruins are of major touristic significance for the area. Hence, the excavations continue in order to improve the touristic appeal, to increase the number of displays and to support further investigations. There are complete yearly reports of all the excavation campaigns carried out since the year 1970, as well as thematic publications dealing with specific buildings of structural elements within the ruins (i.e., forum, basilica, Capitol) and multiple research articles and special publications on the archaeology of the city. Most of this documentation is listed in the work of Sillières (1997) and can be consulted in the Casa de Velázquez at Madrid. The facts mentioned here suggest a QWF of 0.8–1.0, if the evaluation focuses on the excavation history and on the number of the buildings. Taking into account the quality of buildings and the anthropogenic disturbances, the QWF should be on the order of 0.6–0.8. The moderate physiography would lead to values from 0.4 to 0.6, and the evidence for ground instabilities could even produce a QWF <0.4. Hence, it is necessary to weight the observations. Ground instabilities only affect parts of the city and have been investigated in detail. The building quality is generally good, and the number of monumental structures and houses is high. We assume that a value of 0.7 is most appropriate. The resulting site potential factor SPF is on the order of 0.3, i.e., 25% lower than the SPF of the Sagalassos site. Taking into account the seismo-tectonic setting and the weaker build-

ing ground of Baelo Claudia, this ratio appears to be reasonable, despite the fact that there is no satisfactory number of comparable studies available yet. Identification of Damage The following agents may have caused damages found in the ruins of Baelo Claudia: (1) seismic shaking; (2) landslides and soil creeping; (3) ground settlement; (4) post-Roman anthropogenic destruction (e.g., tomb raiders or reuse of bricks); (5) erosion; and (6) possible tsunami action, most likely due to the 1755 Lisbon earthquake. The occurrence of at least one shallow late Roman landslide has been proven for the study site. It is located in the eastern part of the ruins and was documented by Silva et al. (2009). Folded and tilted walls in the Isis temple and forum are the results of this mass movement. Several different damage patterns are attributed to rapid movements and cannot be explained by soil creeping or sliding, such as penetrative linear cracks in poor-quality walls or shock marks and oriented cracks in the Roman pavements. Despite the weak underground conditions, recent ground settlement is unlikely in most cases, since the town was inhabited by the Romans for centuries. During that time, effects of ground settlement are likely to have been repaired. It is known from archaeological reports and field observations that a reuse of building material was common after Roman abandonment and that following settlers used parts of the ruins for other purposes. However, the bricks removed cannot be misinterpreted, and the younger buildings significantly differ from Romanesque. Erosion is a major issue, especially for the buildings close to the shore. Until now, no remains of the harbor have been excavated. It is not known yet whether the harbor was only made of wood and has therefore decayed or if its remains were removed and reused. If the harbor was stone-built, coastal erosion could have led to its destruction. Today, a highly active dune system separates the ruins from the beach, so it cannot be ruled out that some remains are yet to be found beneath. Geophysical measurements on the beach sector adjacent to the ancient fish factories show hints for harbor structures (Alonso-Villalobos et al., 2003). Weathering of bricks and building blocks is common; however, it can be easily distinguished from earthquake-related damages. The aforementioned calcarenites and beach rocks exhibit significant alteration due to eolian processes and the nearby seawater, as strong winds are perennial. However, during the years after the abandonment, most of the ruins were covered by post-Roman colluvium, whereby they are or were not exposed to the surface until the second half of the twentieth century. Tsunami waves following the 1755 Lisbon event reached heights of several meters in the study area (Luque et al., 2001; Martínez Solares, 2005). A number of sedimentary indications have recently been found along the shores close to Baelo Claudia (Koster et al., 2009). However, tsunami wave action concerns only the lower part of Baelo Claudia, not exceeding 5–6 m in elevation above sea level (i.e., the Decumanus Maximus height). No evidence of tsunamites has been found in the present ruins. These patterns, together with the shallow

The Baelo Claudia (southern Spain) case study landslide, are the ones that may most easily to be confused with earthquake-induced damages. For the archaeoseismological assessment of the site, questionable features have therefore not been included in this study. Extensive mapping of corner breakouts and pop-up like arrays in the pavement, shear fracturing in individual flagstones, wall displacements, and cracked walls has been conducted (Silva et al., 2009). The data set includes hundreds of measurements and indicates a preferred spatial orientation of the damages observed as well as an abrupt shock that led to the features observed. According to Sintubin’s criteria, we assume a QWF of 0.8 because it is possible to positively identify a large number of earthquake-related damages and distinguish them from nonseismic effects. Dating of Damage Approximate dating of the damages is possible using construction periods and archaeological remains (pottery, coins, etc.) as indicators. More precise ages can be achieved by detailed stratigraphic observations and 14C dating. Radiocarbon results from the Isis temple sector yield events within a very distinct time frame as described in detail in the Isis Temple Problem section. All collected data point to two earthquake events, which occurred in the first century A.D. and third century A.D., respectively. In Baelo Claudia, a distinct event horizon is present in most parts of the ruins. The dating is coherent and supported by the archaeological stratigraphy. Sintubin and Stewart (2008) suggest a QWF between 0.6 and 0.8 in this case. We decided to choose 0.7 because there are no written historical reports on the studied events but dating seems reliable. Regional Correlation Even though the damages observed in Baelo Claudia clearly point to destructive events that led to the abandonment of an important, prospering town, regional correlation is poor. Known neighboring Roman settlements such as Mellaria or Carteia, a few kilometers east of Bolonia Bay, are not excavated or only in a manner that does not allow archaeoseismological investigation. Therefore, the QWF is assumed to be even lower than the one proposed for the Sagalassos example. As mentioned by Sintubin and Stewart (2008), the following general problem consequently occurs: if there are no comparable adjacent sites or a complete lack of studies in the environment, a QWF close to zero needs to be chosen. This would significantly lower Pes as well as all subsequent equations and lead to a significant distortion of the overall quality estimates. Any archaeological site in the surroundings of a certain study area should be investigated in terms of historic and prehistoric earthquakes in order to rate and support local evidences. Such an overlapping approach would clearly increase the confidence in the results obtained. Unfortunately, particularly when investigating very old, poorly documented events, the possibility of referring to adjacent, comparable archaeological sites often does not exist. Because there is paleoseismological evidence for seismic activity in the Bolo-

141

nia Bay area as discussed in the Paleoseismological Observations section, we follow the suggestions of Sintubin and Stewart (2008) and apply a value of 0.4. Uncertainties The resulting probability of the preferred end solution at the archaeoseismological site of Baelo Claudia is 0.063 (Fig. 9). Main differences to the Sagalassos example are the estimates on tectonic setting, site environment, and dating. Similar to the paleoseismological logic tree of Atakan et al. (2000), an additional uncertainty factor is introduced. The site confidence level SCL is a measure of rate the investigations on a specific site for the inclusion in a seismic hazard analysis. Sintubin and Stewart (2008) defined seven stages, taking into account the excavation history in terms of archaeoseismological proceedings and the excavation reports used for earthquake history research. The Baelo Claudia study corresponds to SCL 6, with the associated correction term C = 8. Decades of archaeological excavations have been carried out, and an archaeoseismological focus has been set on the work since the beginning of the current millennium. The overall resulting archaeoseismological quality factor (AQF), given by Pes × C, is 0.5. Compared to the Sagalassos case, Baelo Claudia mainly benefits from dating. The lower site potential factor (SPF) describes the higher uncertainties in the site potential. Albeit a certain site may not reach high AQFs owing to its SPF, it is also clear that studies there can lead to reliable results if the archaeological site itself has a high potential. DISCUSSION AND CONCLUSIONS The results of both logic trees applied to the Roman ruins of Baelo Claudia, albeit not easily comparable, differ significantly and indicate that the archaeoseismological observations are more reliable than the paleoseismological ones in this special case. As a stand-alone technique, paleoseismology would have to deal with many uncertainties, caused in particular by the limits of site selection and earthquake identification. Archaeoseismology provides more detailed information and yields a higher certainty on distinct events, but it shows weakness in regional correlation of data. The combination of both methods and their evaluation by means of a logic tree lead to a higher level of confidence and result in a more particularized quality estimation of the potential of the investigation site. Subjective evaluations are necessarily inherent in every work, and modern science addresses this problem with the peerreview process, among others. The logic trees are first attempts to quantify uncertainties that are often hard to express in numbers. However, even this logic tree approach reaches its limits in certain points. Some nodes of the logic trees are based on a number of different criteria, which allow calculations of a wide range of probability values. The evaluation of the tectonic setting is probably the best example for our case study. Baelo Claudia is located at the African-Eurasian plate boundary with a local convergence

142

Grützner et al.

rate of 4 mm/yr. Consequently, this would correspond to a QWF of 0.8–1.0, not taking into account that the margin is poorly defined. On the other hand, seismogenic faults within a radius of 10 km are visible, but information on their last activation is rare, and only in some cases a late Quaternary activity can be ascertained. This fact could lead to a QWF of 0.8–1.0, but also to lower values (0.6–0.8). Instrumental seismicity shows no major earthquakes in the vicinity of the ruins, therefore requiring a QWF of 0.6–0.8. When accessing the seismic hazard of a certain study site, the circular reasoning already mentioned in Tectonic Setting and Strain Rate section constitutes another issue: Seismic hazard maps may show a low value for an archaeoseismological study site, but they do not incorporate the results of paleoseismological investigations. Similar problems occur in the archaeoseismological evaluation of the site potential. A wide range of QWF can be attributed to our study site, depending on our personal weighting of the criteria. Here, we decided that the good knowledge of the underground conditions and the large number and high quality of buildings outclass the weak soils and the topography. Any other observer could easily reach a different QWF. So, even with the quantitative approach, subjectivity remains a problem that, however, might be solved or at least be mitigated by an improved scheme. In addition, the paleoseismological logic tree is primarily designed to investigate faults and related coseismic surface ruptures and hardly incorporates the large variety of secondary earthquake ground effects now parameterized in the Macroseismic ESI-2007 Scale (Michetti et al., 2007; Reicherter et al., 2009). Therefore, this fact results in a very limited application of this logic tree approach to severely damaged sites only recording secondary earthquake ground effects located a relative distance (>10 km) to the causative fault, as is the case of Baelo Claudia and the Bolonia Bay area (Silva et al., 2006, 2009). Similarly, the archaeoseismological logic tree approach designed by Sintubin and Stewart (2008) seems to have weaknesses with regard to ground conditions and seismic shaking amplification. Secondary earthquake effects are incorporated in the site environment evaluation, but ground instability facilitating secondary coseismic effects also accounts for the quality of archaeological excavations and therefore has an influence on the evaluation of the site potential. In this step, ground instability evidence diminished the site potential factor value, even with a very good excavation record, as is the case for Baelo Claudia. By circular reasoning, however, weak ground conditions may amplify seismic shaking, thereby triggering severe damage. Therefore, in a case like Baelo Claudia, in which site effects may play a vital role as to the observed damage, the occurrence of unstable ground conditions will lower the site potential factor, diminishing the final Pes estimation and resulting in a possible undervaluation of the true archaeoseismological information recorded at this specific site for future seismic hazard estimations. Due to the lack of similar studies and the absence of sufficient estimations of PQFs and AQFs values for different sites, the classification of our results in a logic tree framework is ulti-

mately not possible. Future investigations also have to prove whether the logic trees are suitable for comparing very different settings of study sites. In the probability estimations, tectonic setting and site environment are taken into account, thus providing the basis for a global approach. However, as already mentioned by Sintubin and Stewart (2008), a direct comparison of several studies in a certain region may enhance the reliability of the sites. Eventually, a growing database will also improve the estimates of a specific site, as it facilitates individual classification of observations in terms of the relevant criteria. Subsequently, the numbers we obtained can be interpreted. In any case, a more robust incorporation of secondary earthquake ground effects and their relation to ground geotechnical properties and seismic amplification in the logic tree approaches employed will be required to conduct more realistic assessments of nonfaulted sites, exclusively devastated by ground shaking, which is the case for most of the severely damaged locations during individual earthquakes. ACKNOWLEDGMENTS This study was financially supported by the German Research Foundation (DFG-project Re 1361/9). We thank the Leibniz Institute in Kiel for radiocarbon dating. The authors would like to thank the reviewers for their inspiring and helpful comments and Kay Barbara for linguistic consultation. We are grateful for everyone who was involved in the fieldwork. This chapter is a contribution to the International Geoscience Programme (IGCP) 567, “Earthquake Archaeology: Archaeoseismology along the Alpine-Himalayan Seismic Zone,” funded by the United Nations Educational, Scientific and Cultural Organization (UNESCO) and to the International Union for Quaternary Research Focus Group on Paleoseismology and Active Tectonics. REFERENCES CITED Alonso-Villalobos, F.J., Gracia-Prieto, F.J., Ménanteau, L., Ojeda, R., Benavente, J., and Martínez, J.A., 2003, Paléogeographie de l’anse de Bolonia (Tarifa, Espagne) à l’époque romaine, in Fouache, E., ed., The Mediterranean World Environment and History: Amsterdam, Elsevier S.A.S., p. 407–417. Ambraseys, N.N., Jackson, J.A., and Melville, C.P., 2002, Historical seismicity and tectonics: The case of the Eastern Mediterranean and the Middle East, in Lee, W.H.K., Kanamori, H., Jennings, P.C., and Kisslinger, C., eds., International Handbook of Earthquake and Engineering Seismology: Amsterdam, Academic Press, International Geophysics Series 81A, p. 747–763. Atakan, K., Midzi, V., Moreno Toiran, B., Vanneste, K., Camelbeeck, T., and Meghraoui, M., 2000, Seismic hazard in regions of present-day low seismic activity: Uncertainties in the paleoseismic investigations along the Bree fault scarp (Roer graben, Belgium): Soil Dynamics and Earthquake Engineering, v. 20, p. 415–427, doi: 10.1016/S0267-7261(00)00081-6. Atwater, B.F., 1992, Geologic evidence for earthquakes during the past 2000 years along the Copalis River, southern coastal Washington: Journal of Geophysical Research, v. 97, p. 1901–1919, doi: 10.1029/91JB02346. Bergen University, Institutt for Geovitenskap, 2008, UNIPAS V3.0 Program: http://www.geo.uib.no/seismo/software/unipas/unipas1.html (accessed 14 February 2010).

The Baelo Claudia (southern Spain) case study Caputo, R., and Helly, B., 2008, The use of distinct disciplines to investigate past earthquakes: Tectonophysics, v. 453, p. 7–19, doi: 10.1016/j .tecto.2007.05.007. Castilla, R.A., and Audemard, F.A., 2007, Sand blows as a potential tool for magnitude estimation of pre-instrumental earthquakes: Journal of Seismology, v. 11, p. 473–487, doi: 10.1007/s10950-007-9065-z. Collins, R., 1998, Spain: An Oxford Archaeological Guide: Oxford, UK, Oxford University Press, 328 p. DeMets, C., Gordon, R.G., Argus, D.F., and Stein, S., 1990, Current plate motions: Geophysical Journal International, v. 101, p. 425–478, doi: 10 .1111/j.1365-246X.1990.tb06579.x. Galadini, F., Hinzen, K.G., and Stiros, S., 2006, Archaeoseismology: Methodological issues and procedure: Journal of Seismology, v. 10, no. 4, p. 395–414, doi: 10.1007/s10950-006-9027-x. Goy, J.L., Zazo, C., Mörner, N.A., Hoyos, M., Somoza, L., Lario, J., Bardají, T., Silva, P.G., and Dabrio, J.C., 1994, Pop up–like deformation of a Roman floor and liquefaction structures in SW Spain as possible paleoseismic indicators: Bulletin of the International Union for Quaternary Research (INQUA) Neotectonics Comisión, v. 17, p. 42–44. Goy, J.L., Zazo, C., Silva, P.G., Lario, J., Bardají, T., and Somoza, L., 1995, Evaluación geomorfológica del comportamiento neotectónico del Estrecho de Gibraltar durante el Cuaternario, in Esteras, M., ed., El Enlace Fijo del Estrecho de Gibraltar: Madrid, Sociedad española de estudios para la comunicacion fija a través del Estrecho de Gibraltar S.A (SECEGSA), v. 2, p. 51–69. Guidoboni, E., and Traina, G., 1996, Earthquakes in medieval Sicily: A historic revision (7th–13th century): Annali di Geofisica, v. XXXIX, no. 6, p. 1205–1225. Höbig, N., Braun, A., Grützner, C., Fernández-Steeger, T., and Reicherter, K., 2009, Rock fall hazard mapping and run out simulation: A case study from Bolonia Bay, southern Spain, in Pérez-López, R., Grützner, C., Lario, J., Reicherter, K., and Silva, P.G., eds., Archaeoseismology and Palaeoseismology in the Alpine-Himalayan Collisional Zone: 1st International Union for Quaternary Research (INQUA) and International Geoscience Programme (IGCP) 567 International Workshop on Earthquake Archaeology and Palaeoseismology, 7–13 September 2009, Baelo Claudia (Cádiz, Spain): Madrid, Instituto Geográfico Nacional de España, p. 52–56. Instituto Geográfico Nacional de España: http://www.ign.es/ign/es/IGN/ SisCatalogo.jsp (accessed 14 February 2010). Karcz, I., and Kafri, U., 1978, Evaluation of supposed archaeoseismic damage in Israel: Journal of Archaeological Science, v. 5, p. 237–253, doi: 10.1016/0305-4403(78)90042-0. Koster, B., Vonberg, D., and Reicherter, K., 2009, Tsunamigenic deposits along the southern Gulf of Cádiz (southwestern Spain) caused by tsunami in 1755?, in Pérez-López, R., Grützner, C., Lario, J., Reicherter, K., and Silva, P.G., eds., Archaeoseismology and Palaeoseismology in the AlpineHimalayan Collisional Zone: 1st International Union for Quaternary Research (INQUA) and International Geoscience Programme (IGCP) 567 International Workshop on Earthquake Archaeology and Palaeoseismology, 7–13 September 2009, Baelo Claudia (Cádiz, Spain): Madrid, Instituto Geográfico Nacional de España, p. 73–75. Levret, A., Backe, C., and Cushing, M., 1994, Atlas of macroseimic maps for French earthquakes with their principal characteristics: Natural Hazards, v. 10, p. 19–46, doi: 10.1007/BF00643439. Luque, L., Lario J., Zazo, C., Goy, J.L., Dabrio, C.J., and Silva, P.G., 2001, Tsunami deposits as paleoseismic indicators: Examples from the Spanish coast: Acta Geologica Hispanica, v. 36, no. 3–4, p. 197–211. Martínez Solares, J.M., 2005, Tsunamis en el contexto de la Península Ibérica: Enseñanza de las Ciencias de la Tierra, v. 13, no. 1, p. 52–59. McCalpin, J.P., and Nelson, A.R., 1996, Introduction to paleoseismology, in McCalpin, J.P., ed., Paleoseismology: San Diego, Academic Press, p. 1–32. Michetti, A.M., and Hancock, P.L., 1997, Paleoseismology: Understanding past earthquakes using Quaternary geology: Part 1: Geodynamics, v. 24, no. I-4, p. 3–10, doi: 10.1016/S0264-3707(97)00004-5. Michetti, A.M., Audemard, F.A., and Marco, S., 2005, Future trends in paleoseismology: Integrated study of the seismic landscape as a vital tool in seismic hazard analyses: Tectonophysics, v. 408, p. 3–21, doi: 10.1016/ j.tecto.2005.05.035. Michetti, A.M., Esposito, E., Guerrieri, L., Porfido, S., Serva, L., Tatevossian, R., Vittori, E., Audemard, F., Azuma, T., Clague, J., Comerci, V., Gürpinar, A., McCalpin, J., Mohammadioun, B., Mörner, N.A., Ota, Y., and

143

Roghozin, E., 2007, Intensity Scale ESI 2007: Memorie Descrittive della Carta Geologica D’Italia, v. 74, p. 41. Nikonov, A.A., 1988, Reconstruction of the main parameters of old large earthquakes in Soviet Central Asia using the paleoseismogeological method: Tectonophysics, v. 147, p. 297–312, doi: 10.1016/0040-1951(88)90191-6. Reicherter, K., 2001, Paleoseismologic advances in the Granada basin (Betic Cordilleras, southern Spain): Acta Geologica Hispanica, v. 36, no. 3–4, p. 267–281. Reicherter, K., Jabaloy, A., Galindo-Zaldívar, J., Ruano, P., Becker-Heidmann, P., Morales, J., Reiss, S., and González-Lodeiro, F., 2003, Repeated palaeoseismic activity of the Ventas de Zafarraya fault (S Spain) and its relation with the 1884 Andalusian earthquake: International Journal of Earth Sciences, v. 92, p. 912–922, doi: 10.1007/s00531-003-0366-3. Reicherter, K., Michetti, A.M., and Silva, P.G., 2009, Introduction, in Reicherter, K., Michetti, A.M., and Silva, P.G., eds., Paleoseismology: Historical and Prehistorical Records of Earthquake Ground Effects for Seismic Hazard Assessment: Geological Society of London Special Publication 316, p. 1–10. Sanz de Galdeano, C., 1990, Geologic evolution of the Betic Cordilleras in the Western Mediterranean, Miocene to present: Tectonophysics, v. 172, p. 107–119, doi: 10.1016/0040-1951(90)90062-D. Sillières, P., 1997, Baelo Claudia: Una ciudad Romana de la Bética: Madrid, Junta de Andalucía–Casa de Velázquez, 237 p. Silva, P.G., Borja, F., Zazo, C., Goy, J.L., Bardají, T., Luque, L., Lario, J., and Dabrio, C.J., 2005, Archaeoseismic record at the ancient Roman city of Baelo Claudia (Cádiz, south Spain): Tectonophysics, v. 408, p. 129–146, doi: 10.1016/j.tecto.2005.05.031. Silva, P.G., Goy, J.L., Zazo, C., Bardají, T., Lario, J., Somoza, L., Luque, L., and González-Hernández, F.M., 2006, Neotectonic fault mapping at the Gibraltar Strait tunnel area, Bolonia Bay (south Spain): Engineering Geology, v. 84, p. 31–47, doi: 10.1016/j.enggeo.2005.10.007. Silva, P.G., Reicherter, K., Grützner, C., Bardají, T., Lario, J., Goy, J.L., Zazo, C., and Becker-Heidmann, P., 2009, Surface and subsurface paleoseismic records at the ancient Roman city of Baelo Claudia and the Bolonia Bay area, Cádiz (south Spain), in Reicherter, K., Michetti, A.M., and Silva, P.G., eds., Paleoseismology: Historical and Prehistorical Records of Earthquake Ground Effects for Seismic Hazard Assessment: Geological Society of London Special Publication 316, p. 93–121. Sintubin, M., and Stewart, I.S., 2008, A logical methodology for archaeoseismology: A proof of concept at the archaeological site of Sagalassos, southwest Turkey: Bulletin of the Seismological Society of America, v. 98, p. 2209–2230, doi: 10.1785/0120070178. Stich, D., Serpelloni, E., Mancilla, F.L., and Morales, J., 2006, Kinematics of the Iberia–Maghreb plate contact from seismic moment tensors and GPS observations: Tectonophysics, v. 426, p. 295–317, doi: 10.1016/j .tecto.2006.08.004. Stiros, S., 1996, Identification of earthquakes from archaeological data: Methodology, criteria and limitations, in Stiros, S., and Jones, R.E., eds., Archaeoseismology: British School at Athens, Fitch Laboratory Occasional Paper 7, p. 129–152. Vollmert, A., Reicherter, K., and Grützner, C., 2009,The origin of rockfalls and the formation of hanging valleys along the La Laja range front (Tarifa, S. Spain), in Pérez-López, R., Grützner, C., Lario, J., Reicherter, K., and Silva, P.G., eds., Archaeoseismology and Palaeoseismology in the AlpineHimalayan Collisional Zone: 1st International Union for Quaternary Research (INQUA) and International Geoscience Programme (IGCP) 567 International Workshop on Earthquake Archaeology and Palaeoseismology, 7–13 September 2009, Baelo Claudia (Cádiz, Spain): Madrid, Instituto Geográfico Nacional de España, p. 162–164. Weijermars, R., 1991, Geology and tectonics of the Betic zone, SE Spain: Earth-Science Reviews, v. 31, p. 153–236, doi: 10.1016/0012-8252(91 )90019-C. Wells, D.L., and Coppersmith, J.K., 1994, New empirical relationships among magnitude, rupture length, rupture width, rupture area, and surface displacement: Bulletin of the Seismological Society of America, v. 84, p. 974–1002. Zazo, C., Silva, P.G., Goy, J.L., Hillaire-Marcel, C., Lario, J., Bardají, T., and González, A., 1999, Coastal uplift in continental collision plate boundaries: Data from the Last Interglacial marine terraces of the Gibraltar Strait area (south Spain): Tectonophysics, p. 301, v. 95–119. MANUSCRIPT ACCEPTED BY THE SOCIETY 19 MAY 2010 Printed in the USA

The Geological Society of America Special Paper 471 2010

Long-term effect of seismic activities on archaeological remains: A test study from Zakynthos, Greece Melek Tendürüs* Institute for Geo- and Bioarchaeology, Vrije Universiteit Amsterdam, De Boelelaan 1085, 1081 HV Amsterdam, The Netherlands Gert Jan van Wijngaarden* Amsterdam Archaeological Centre, Universiteit van Amsterdam, Turfdraagsterpad 9, 1012 XT Amsterdam, The Netherlands Henk Kars* Institute for Geo- and Bioarchaeology, Vrije Universiteit Amsterdam, De Boelelaan 1085, 1081 HV Amsterdam, The Netherlands

ABSTRACT During the archaeological and geoarchaeological surveys on the island of Zakynthos, Greece, it has been noted that the distribution and preservation of archaeological remains of Zakynthos present spatially different characteristics. In general, archaeological pottery finds and architectural remains in the eastern part of the island appear to be more fragmented and more widely distributed than in the western part of the island. Due to the high seismicity in the region, the question has come up whether a correlation between seismic activity and distribution and preservation conditions of archaeological remains exists or not. In order to investigate the mentioned relationship, we looked at the cumulative effect of continuing earthquakes for the last hundred years on the island of Zakynthos. We used ground acceleration to quantify the earthquake-induced damage. The predicted cumulative destruction intensity is presented on a map, and it illustrates that we can cautiously attribute the distribution difference of the archaeological remains with different preservation conditions to the seismic activity on the island. It is hoped that this study will initiate new scientific research into the characteristics of the distribution of archaeological remains in seismically active areas. In addition, it is to be expected that this study will contribute to in situ preservation studies relating to the long-term effect of seismic activities on the archaeological record.

*E-mails: [email protected]; [email protected]; [email protected]. Tendürüs, M., van Wijngaarden, G.J., and Kars, H., 2010, Long-term effect of seismic activities on archaeological remains: A test study from Zakynthos, Greece, in Sintubin, M., Stewart, I.S., Niemi, T.M., and Altunel, E., eds., Ancient Earthquakes: Geological Society of America Special Paper 471, p. 145–156, doi: 10.1130/2010.2471(13). For permission to copy, contact [email protected]. © 2010 The Geological Society of America. All rights reserved.

145

146

Tendürüs et al.

INTRODUCTION The Mediterranean area is well known for its archaeological richness and its frequent earthquakes. It is only very recently, however, that archaeologists, historians, seismologists, geologists, and engineers have begun to collaborate systematically in research and heritage management. At the moment, the common concern of these studies is mainly limited to determining, based on the archaeological and geological-geomorphological evidences, whether or not an earthquake of significance was a major cause in the destruction at a particular site, e.g., Stiros et al. (1996), Ellenblum et al. (1998), Galadini and Galli (2001), Marco et al. (2003), Koukouvelas et al. (2005), Caputo et al. (2006), Similox-Tohon et al. (2006), Reinhardt et al. (2006), and Marco (2008). The cause of destruction in an archaeological stratigraphy (e.g., earthquake, land instability, war) can be determined by examining the distortion on structures and the in situ geological data. With developing methodologies for the recognition of earthquake-induced damages, archaeologists can ascertain the reason for massive destruction at their site more reliably. Sometimes, the destruction layer recorded in the destroyed ancient site might be dated, using the archaeological context, or be correlated with a devastating earthquake event in the region, using historical sources of the expected time period mentioning such an event. The inferred earthquake is used by seismologists in the improvement of historical earthquake catalogues for the assessment of seismic hazard; by geologists in understanding of the geodynamic characteristics of the region; and by engineers in the development of new constructional methods to mitigate the seismic risk. However, the effect of seismic activities on archaeological remains does not simply conclude with one devastating event, it is an ongoing process that may continue to destroy the archaeological remains (Papastamatiou and Psycharis, 1993; Psycharis et al., 2000; Cerone et al., 2001). In this paper, the effect of seismic activities on archaeological remains is not considered to be constrained to only one devastating event. The study concentrates on the probable continuing destruction of archaeological record by earthquakes before, during, and after the abandonment of sites. Seismic hazard studies will be used to begin to understand the archaeology of a seismically active region. The area of study, Zakynthos (western Greece), is located in one of the highest seismic activity regions of the world and has been inhabited since the Paleolithic. During the archaeological and geoarchaeological surveys on the island in the period of 2005 to 2008, it was noted that the distribution and preservation of archaeological remains of Zakynthos present spatially different characteristics. In general, archaeological pottery finds and architectural remains in the eastern part of the island appear to be more fragmented and more widely distributed than in the western part of the island. Standing architectural remains also appear to be in a better condition in the west than in the east. It is believed that there may be a correlation between the ongoing seismic activities in the region and the distribution and preser-

vation conditions of archaeological remains in Zakynthos. This paper particularly addresses the development of a method that may be used to investigate spatially such a probable link between the conditions of archaeological remains and the seismicity of the specific region. STUDY AREA Geological and Seismotectonic Background Zakynthos—the southernmost of the Ionian Islands of western Greece—lies in a tectonically complex and active area (Fig. 1; Underhill, 1989; Papazachos and Kiratzi, 1996; Barka and Reilinger, 1997; Hinsbergen et al., 2006; Lagios, et al., 2007). In particular, the Ionian Basin of the African plate subducts beneath the Aegean continental microplate of the Eurasian plate, the Apulian continental crustal part of the African plate collides with the Eurasian plate in the north, and the Cephalonia transform fault zone connects these subduction and collision zones. The seismicity of the area is the highest in Greece, mainly consisting of shallow seismic activities (Papazachos, 1990; Papazachos et al., 1993; Clément et al., 2000). Figure 2 shows the spatial distribution and the magnitude occurrences of the earthquakes that were recorded between 1901 and 2006 (except the last 3 months) with Mw ≥4.5 in the vicinity of Zakynthos, which also form our data set for the study. Within the defined area, there are 1975 events, including the most destructive earthquake of the last century in Greece, which occurred in the Ionian Islands on 12 August 1953 with a surface-wave magnitude of 7.2 (Stiros et al., 1994). Hatzidimitriou et al. (1985) calculated the return period of this and larger magnitudes of earthquakes for the vicinity as 29 yr, based on the data covering the last 81 and 181 yr periods. The important local seismicity of Zakynthos occurs along the Ionian thrust, which also divides the island into two different geological units: Pre-Apulian and Ionian (Fig. 1). The PreApulian (also called Paxos) unit is recognized on the island as the Vrachionas carbonate anticline. It is composed of thick Upper Cretaceous limestones and dominates the western part of the island. Marly limestones of Eocene and Oligocene age are exposed at its eastern slope. Miocene deposits are observed at the central lowlands as sandstones and bluish marls with gypsum intercalations. Toward the east, this range of hills suddenly leaves its place to a plain filled with recent alluvium deposits. The Ionian unit is exposed on the eastern part of the island, namely at the Vasilikos Peninsula. The peninsula mainly consists of Pliocene and Pleistocene marine mudstones and sandstones, and Triassic evaporates and limestones also crop out. Archaeological Investigations Zakynthos is mentioned in various historical sources, indicating its long and intensive habitation. It is mentioned on Linear B tablets from Pylos showing the overseas connection of the island with the mainland in the Mycenaean period (Palaima,

A test study from Zakynthos, Greece

147

Figure 1. Plate boundaries and main tectonic features of Greece (Barka and Reilinger, 1997; Clément et al., 2000) and the geological map of the Zakynthos Island (Perry and Temple, 1980; Underhill, 1989). CTF—Cephalonia transform fault, NAF—North Anatolian fault.

1991). Homer mentions the island as a part of the territory of Odysseus (Iliad II: 634; Odyssey IX: 24). In Classical and Hellenistic periods, the island often was an ally of Athens (Kalligas, 1993). Pliny the Elder mentions the particular fertility of the island (The Natural History 4: 19: 12). During the Venetian period, Zakynthos was known as “the flower of the Levant” due to its beauty and fertility (Ζois, 1955). In spite of its prosperous land and the suitable geographic location with regard to local and Mediterranean maritime traffic, the archaeology of Zakynthos is relatively little known. Sylvia Benton of the British School in Athens was the first archaeologist to systematically describe several archaeological sites on the island (Souyoudzoglou-Haywood, 1999). She excavated a Mycenaean house in Cape Kalogeras on the Vasilikos

Peninsula and a tholos tomb near Alykanas in the 1930s. Unfortunately, her results remained unpublished, and at the time of the 1953 earthquake, all the finds and records were lost. The Greek Archaeological Service carried out excavations in the early 1970s (Mylona, 2006). Their research focused on the hill rising just behind the modern town of Zakynthos and its surrounding area. Today, there is a Venetian castle with extensive British modifications standing on top of the hill facing Peloponnese. The uncovered archaeological artifacts and some standing architecture in the interior of the castle indicate occupation during the Bronze Age and during Archaic, Classical, Hellenistic, and Roman periods. Based on the archaeological evidence and ancient authors’ texts, it can be concluded that the hill used to be an ancient acropolis already fortified by the mid-fifth century B.C.

148

Tendürüs et al. tholoi and settlement remains from the Mycenaean period, and other concentrations of finds, especially pottery, from Archaic, Hellenistic, or Roman times, medieval, and early modern periods have been recorded. Archaeological Remains

Figure 2. The spatial distribution (A) and the magnitude frequency (B) of the earthquakes recorded in the vicinity of Zakynthos between 1901 and 2006 (except the last 3 months). The presented data also form the data set of this study and cover the magnitudes of earthquakes Mw 4.5 and greater.

Currently, systematic archaeological research on the island has been continuing since 2005 as a joint project of the Netherlands Institute in Athens (NIA) and the 35th Ephorate of Prehistoric and Classical Antiquities: the Zakynthos Archaeology Project (ZAP) (Van Wijngaarden et al., 2006, 2007). The project aims to gain more insight into the archaeology of the island, and it combines archaeological surveys with geographic information system (GIS), aerial photography, and geomorphologic studies to better identify the human-nature interaction in the past and also to understand the effect of today’s processes on the archaeological remains (e.g., Rink, 2005; Stoker 2006; Pieters et al., 2007; Horn Lopes, 2008; Storme, 2008; Tendürüs, 2009). In 3 years, the project showed that the island was inhabited in several periods. Lithic tools and flakes from the Paleolithic to Early Bronze Age,

Unlike many other areas in Greece, Zakynthos has very few standing archaeological remains that date back to more than a few centuries ago. The only excavation that was fully published is the Mycenaean cemetery at Kambi in the western mountains of the island (Agalopoulou, 1973). Other notable ancient remains, probably of Roman date, are found built into the little church St. Dhimitrios at Melinado, near Machairado (Fig. 3A; Foss, 1969; Kalligas 1993). Palaiokastro is another significant archaeological site on Zakynthos. It is located on the hills west of Machairado with an imposing view of the plain. The site was considered to be medieval, but the finds in 2007 and 2008 indicated activities also in prehistory and antiquity (Van Wijngaarden et al., 2009, 2010). Apart from the medieval structures, several walls, probably to be dated sometime in the period from Archaic to Hellenistic times, have been recorded on the ridges and the plateaus to the west of the top of the hill (Fig. 3C). Figure 4 presents the archaeological sites known on the island from the literature and from the surveys of the ZAP. Among these sites, there are settlement remains (i.e., architectural remains like foundations and walls), graves, and surface finds (i.e., small finds including mainly pottery sherds and stone tools). In 2006, the survey conducted at the southeastern peninsula (Vasilikos) revealed large quantities of pottery and lithics, mainly lacking of local concentrations. Archaeological ceramics that were recovered, included sherds of considerable quality, but they were generally heavily worn. Standing ancient settlement remains were only observed at Kalogeras where the thick brushes were removed. A view from the discovered various eroding walls is displayed in Figure 3B. The majority of the pottery found in association with these walls dated to the prehistoric period, in particular, to the Middle and Late Bronze Ages (Von Stein, 2009). Elsewhere at the site, Archaic and Classical finds were more abundant. In comparison to the Vasilikos Peninsula, the distribution of archaeological artifacts in the vicinity of Keri, where a pilot survey was carried out in 2005, showed more concentrations of material, probably representing archaeological sites. In a few cases, these sites could be identified with certainty on the basis of the quantity, the quality, and the diversity of the surface material. In general, the archaeological record in the southeast appears to show a higher degree of disturbance and dispersion. Figure 4 also indicates churches and monasteries of Zakynthos, the major architectural remains on the island. More of these monuments are located in or near the town of Zakynthos and in the western high areas. The churches at the town were reconstructed after the catastrophic earthquake of 1953, except the concrete reinforced church of St. Dionysios, built in 1948 (Facaros and Theodorou, 2003). Some of the churches and the

A test study from Zakynthos, Greece

149

A

C B Figure 3. A few example views from the “standing” archaeological remains on Zakynthos. (A) The church of St. Dhimitrios in Melinado containing marble columns and blocks, probably of Roman date. (B) Part of wall remains at Kalogeras, probably from the second millennium B.C. (C) Part of wall remains at the site of Palaiokastro. (D) The monastery of Skopiotissa on Mount Skopos being restored.

D

monasteries in the western part of the island were also badly and partially damaged. The most serious damages mentioned are those on the church of St. Andreas at the northwest and the little church at the east foot of the Vrachionas (Foss, 1969). The monastery of Skopiotissa on Mount Skopos was also destroyed with the earthquake, but it is currently being restored (Fig. 3D).

and concrete supports. Although short-term effects of earthquakes on the archaeological remains receive some attention, their longterm effects have been neglected in archaeological studies. Therefore, there is virtually no directly relevant literature available for our investigation. On the other hand, studies in the fields of seismology, geology, and civil engineering can help us to look into the spatial distribution of the combined effect of past earthquakes.

METHOD Quantification of Earthquake-Induced Damage Human response to the damaging effects of earthquakes on archaeological remains has only recently improved further than trying to provide stability of some standing buildings with steel

Earthquakes have six major effects. Ground motion and faulting are two of them and cause damage directly. Others are fire,

150

Tendürüs et al.

Figure 4. Geographic distribution of the archaeological sites, churches, and monasteries of Zakynthos. A, B, and C represent the intensive archaeological survey areas of the Zakynthos Archaeology Project. Areas A and C were covered in 2005 and 2006. The investigations at area B still carry on.

landslides, cliff collapse, or any mass-wasting movements, liquefaction, and tsunamis (initiated by earthquakes). These are secondary effects that cause damage indirectly (Yeats et al., 1997). In our study, we used the ground motion to quantify the damage due to seismic activities because it is the most common way of determining the level of vulnerability of a region in seismic hazard assessment studies. The level of ground shaking

caused by an earthquake at a site mainly depends on the magnitude of the earthquake, geographical proximity of the site to the seismic source, and local geological characteristics of the media. For instance, ground shaking decreases when the waves propagate outward from the source; soft sediments amplify seismic waves and create more ground shaking than hard rocks (Day, 2002).

A test study from Zakynthos, Greece The level of ground shaking, or damage, is usually measured as peak ground acceleration (PGA), which is often determined by attenuation models (Day, 2002; Bozorgnia and Campbell, 2004). An attenuation model is defined as a mathematical expression developed to estimate the peak ground acceleration at a certain distance from a seismic source, using a given data set of seismological parameters (deterministic method) or using all possible earthquake locations and magnitudes together with their expected probabilities of occurrence (probabilistic method). There are a few attenuation relationships developed specifically for Greece and its neighboring regions. The attenuation model by Theodulidis and Papazachos (1992) provides a reasonable and geographically specific model to apply to PGA seismic evaluations for Greece (Burton et al., 2003) and was selected for the calculations in this study. The model was developed studying 36 shallow earthquakes from Greece with magnitudes Ms 4.5–7.0 and four from Japan and Alaska with magnitudes Ms 7.2–7.5. The attenuation relation is: ln ag = 3.88 + 1.12 Ms − 1.65 ln ( R + 15) + 0.41 S + 0.71 P,

(1)

where ag is the peak horizontal ground acceleration in cm s–2, R is the epicentral distance in km, S is a parameter equal to zero at “alluvium” sites and equal to one at “rock” sites, and P is a parameter equal to zero for mean or 50 percentile values and one for 84 percentile values (taken as zero in our calculations). Figure 5 shows the change of PGA values projected by this attenuation model when the waves are propagating further from the epicenter of an earthquake of magnitude Ms 4.5, 6.0, 7.5, and 9.0 in the same type of geological medium. While the effect of a Ms 4.5 earthquake remains local, an earthquake with a strongermagnitude earthquake, e.g., Ms 7.5, is felt moderately at 60 km distance. Data Collection We put together a collection of the magnitude and location of past earthquakes for the vicinity of Zakynthos, forming

our data set from the recent catalogue published by the Aristotle University of Thessaloniki (Papazachos et al., 2007). The selected seismic events have moment magnitudes of 4.5 and greater and occurred between 1901 and 2006 (except the last 3 mo). The area coverage of the data set is 300 × 300 km2 within the frame bounded by the coordinates of 36.420°N–39.140°N and 19.072°E–22.500°E. The spatial distribution of earthquake events included in our data set and their magnitude occurrences are shown in Figure 2. An earthquake of Richter magnitude ML 4 is felt by almost everybody; it breaks some dishes and windows and can displace unstable objects (Table 1; U.S. Geological Survey, 2010). Since slight damages start appearing during earthquakes of ML 4, equivalent to Mw 4.5 (please refer to the study of Papazachos et al. [1997] for the relationships between the magnitudes in the region), it is set as the threshold value for our data set. Due to their unstable nature, we considered that archaeological remains will be more prone to ground shaking. Although the compiled earthquake catalogues of Greece extend back to 550 B.C.E. (Papazachos and Papazachou, 1997; Papazachos et al., 2000), we chose the earthquake records from 1901 (introduction of seismograph in Greece) onward as our data set because the historical records are lacking a considerable amount of seismic events, most importantly, in the lower-magnitude range (Ambraseys, 1996; Kouskouna and Makropoulos, 2004). Insertion of the earlier records into our data set would promote a few devastating events and omit a very large collection of smaller and unrecorded events, causing a deceptive geographical distribution in our results. Cumulative Destruction Intensities The cumulative damaging effect of past earthquakes, which we call cumulative destruction intensity, was calculated by accumulating the PGA values of each earthquake in our data set for the vicinity of Zakynthos using the attenuation relationship of Theodulidis and Papazachos (Eq. 1). First, the earthquake magnitudes of the data set were converted from moment magnitude to surface-wave magnitude using the following relationship (Papazachos et al., 1997): Mw = MS, 6.0 ≤ MS ≤ 8.0, Mw = 0.56MS + 2.66, 4.2 ≤ MS ≤ 6.0.

Figure 5. The change of peak ground acceleration (PGA) values projected by the attenuation model of Theodulidis and Papazachos (1992).

151

(2)

Second, the geological units of the island were classified as alluvium and rock for the S parameter, a dummy variable taking the value of zero for “alluvium” sites and one for “rock” sites in the attenuation relationship. Figure 6 shows the rock sites of Zakynthos in gray color, including limestones, sandstones, mudstones, and gypsum, and the alluvium sites in white. Finally, the area covered by our data set was divided into 40,000 grid points, and PGA values were calculated on every grid point for each earthquake in the data set (Fig. 7). The computed PGA values for each grid point were accumulated in the process to obtain the total

152

Tendürüs et al. TABLE 1. COMPARISON OF RICHTER MAGNITUDE AND MODIFIED MERCALLI INTENSITY SCALES Richter magnitude

Modified Mercalli intensity

Description

1.0–3.0

I

Not felt except by a very few under especially favorable conditions.

3.0–3.9

II

Felt only by a few persons at rest, especially on upper floors of buildings.

III

Felt quite noticeably by persons indoors, especially on upper floors of buildings. Many people do not recognize it as an earthquake. Standing motor cars may rock slightly. Vibrations similar to the passing of a truck. Duration estimated.

IV

Felt indoors by many, outdoors by few during the day. At night, some awakened. Dishes, windows, doors disturbed; walls make cracking sound. Sensation like heavy truck striking building. Standing motor cars rocked noticeably.

V

Felt by nearly everyone; many awakened. Some dishes, windows broken. Unstable objects overturned. Pendulum clocks may stop.

VI

Felt by all, many frightened. Some heavy furniture moved; a few instances of fallen plaster. Damage slight.

VII

Damage negligible in buildings of good design and construction; slight to moderate in well-built ordinary structures; considerable damage in poorly built or badly designed structures; some chimneys broken.

VIII

Damage slight in specially designed structures; considerable damage in ordinary substantial buildings with partial collapse. Damage great in poorly built structures. Fall of chimneys, factory stacks, columns, monuments, walls. Heavy furniture overturned.

IX

Damage considerable in specially designed structures; well-designed frame structures thrown out of plumb. Damage great in substantial buildings, with partial collapse. Buildings shifted off foundations.

X

Some well-built wooden structures destroyed; most masonry and frame structures destroyed with foundations. Rails bent.

XI

Few, if any (masonry) structures remain standing. Bridges destroyed. Rails bent greatly.

XII

Damage total. Lines of sight and level are distorted. Objects thrown into the air.

4.0–4.9

5.0–5.9

6.0–6.9

7.0 and higher

Earthquake 1 MS(1)

i R(i1, j1, 1) S(i1, j1) R(i2, j2, 1)

S(i2, j2)

R(i3, j3, 1)

j

R(i1, j1, 2) Earthquake 2 MS(2)

R(i2, j2, 2) S(i3, j3)

R(i3, j3, 2)

Figure 7. A simple sketch of the calculation procedure of cumulative destruction intensity values. There are 40,000 grid points in the defined area and 1975 earthquakes in the considered period.

Figure 6. Two basic geological units of Zakynthos: rock sites (black) and alluvium sites (white). A, B, and C represent the intensive archaeological survey areas of the Zakynthos Archaeology Project. Areas A and C were covered in 2005 and 2006. The investigations at area B continue.

A test study from Zakynthos, Greece damage induced by all the earthquakes during the considered time period, using the formula CPGA (i, j ) = ∑ k =1 ag (i, j, k ), N

i = 1… L, j = 1…L,

(3)

where i and j are the horizontal and vertical grid indices, L is the grid resolution, N is the total number of earthquakes in the data set, and ag(i, j, k) is the peak ground acceleration at grid point (i, j) caused by earthquake k. Using Equation 1, this is ultimately

#

N

CPGA i, j  - exp 3.88 1.12 Ms  k < 1.65 ln k 1

}

⎡⎣R (i , j, k ) + 15⎤⎦ + 0.41S (i, j ) + 0.71P ,

(4)

where Ms(k) is the surface magnitude of the kth earthquake, R(i, j, k) is the distance of grid point (i, j) to the epicenter of the kth earthquake, S(i, j) is either zero or one according to whether grid point (i, j) is considered alluvium or rock, and P is the percentile parameter, as described earlier for the attenuation relation given in Equation 1.

153

RESULTS AND DISCUSSION The accumulated PGA values or the cumulative destruction intensities for the last hundred years for the vicinity of Zakynthos are shown in Figure 8. The higher are the cumulative PGA values, the larger is the expected accumulated damage on archaeological remains. However, at this stage of this test study, a comparison of the cumulative destruction intensities with the accumulated damage on the archaeological remains will be possible only relatively. The central part of the island, where recent alluvial sediments are deposited, accommodates the greatest cumulative destruction intensities in the resulting map. Today, this extensive plain is very poor in terms of availability of churches and monasteries. However, since the archaeological surveys from this part are not complete yet, it is difficult to attribute the absence of ancient architectural remains in this area to seismic activities. On the other hand, in the book The Earthquakes of Greece, by Papazachos and Papazachou (1997), one of the most comprehensive accounts concerning strong earthquakes in Greece, the town of Zakynthos and the hill behind it accommodating the Venetian

Figure 8. The cumulative peak ground acceleration (PGA) values in m s–2 for the period of 1901–2006 and Mw ≥ 4.5.

154

Tendürüs et al.

castle are listed most frequently among the heavily damaged areas on the island. After the last catastrophic earthquake, the town was almost entirely rebuilt. It is also known that the castle has undergone many repairs since the Venetian period (Mylona, 2006). Accordingly, Figure 8 indicates that these two sites are subjected to high levels of cumulative destruction due to earthquakes. The cumulative PGA values of the southeastern part of the island (the Vasilikos Peninsula) are found to be comparable to those of the area of the hill behind the town of Zakynthos. The rare preservation of the settlement remains and the relatively less concentrated distribution of the small finds on the peninsula might be explained with these high destruction intensity values in the vicinity. The intense fragmentation of the constructions may result in significant displacement of the pieces by the intervention of humans and nature. Additionally, the noted damages on the church of St. Andreas on the north and the monastery of Skopiotissa on Mount Skopos (listed as number 1 and 16 in Fig. 4) and the archaeological survey results on the west show a good match with the destruction intensities of the resulting map. The church of St. Andreas takes our extra attention because of its contradictory condition with the base rock type underneath. It also agreeably follows the contour lines for the cumulative PGA values on our cumulative destruction intensity map and indicates the value of studying the probable cumulative effect of seismic activities on archaeological remains. A similar situation can be considered for the area south of the Vrachionas mountain range as well. While the archaeological sites have been assigned according to the amount and concentration of the small finds in the vicinity of Keri, the allocated archaeological sites in the central part of the Vrachionas consist of all types of find collections following suitably the picture on the cumulative PGA map. Our cumulated ground acceleration map, once more, points out the significance of the local geological conditions in the ground motion studies and illustrates the vast divergence observed between the destruction levels of alluvium and rock sites. Unfortunately, attenuation models generally use simple binary categories to describe the ground composition, in which local site conditions are classified simply as soil or rock. Campbell and Bozorgnia (2003) showed the importance of a refined geological classification, including soft soil, firm soil, soft or primarily sedimentary rock, and hard rock, in the prediction of ground acceleration. Zakynthos is formed of various geological units such as limestone (also in various ages and characteristics), gypsum, sandstone, and mudstone. The Vasilikos Peninsula mainly consists of Pliocene and Miocene sandstone, in which the shear waves travel slower than in the massive limestone outcrop on the western part of the island, causing more shaking of the ground. Some examples of shear wave velocity for different materials are given in Table 2. Results from a new attenuation relation specific to Zakynthos will certainly be more detailed and more representative. Another important point to mention is that the attenuation model includes only the direct damage caused by ground motion created by earthquakes. However, surface faulting and auxiliary

TABLE 2. P AND S WAVE VELOCITIES OF SOME SELECTED MATERIALS Material

P wave velocity (m/s)

S wave velocity (m/s)

Steel

6100

3500

Concrete

3600

2000

5500–5900

2800–3000

Granite

6400

3200

Sandstone

Basalt

1400–4300

700–2800

Limestone

5900–6100

2800–3000

Sand (unsaturated)

200–1000

80–400

Sand (saturated)

800–2200

320–880

effects such as liquefaction, landslides, rockfalls, tsunamis, fire, looting, etc., can also have considerable roles in the destruction of sites. These types of damages may also be introduced into a cumulative destruction map or be studied separately. Such studies would confine the damage to a local level. CONCLUSIONS This paper aimed to draw attention to the possible correlation between local seismic activity and the different distribution and preservation conditions of archaeological remains. It has shown that the spatial distribution of total destruction intensities of past earthquakes that affected archaeological sites can provide additional insight into the present distribution of remains. However, we recognize that, being a test study, this research will lead to many further questions and discussions on the topic. First, the qualification of the degree of weathering and of the dispersion of archaeological remains has not been done systematically yet. In the near future, we plan to develop methodologies to investigate these issues more systematically. Second, the distribution of archaeological remains in the landscape is subject to many different factors. The influence of seismic activity with regard to the distribution of archaeological material and its relationship to factors such as erosion and sedimentation, and agricultural and cultural factors, still need to be assessed. On the other hand, our study shows that relationships between seismic activities and characteristics of the archaeological record are likely. This result merits similar investigations at other archaeologically rich and seismically active areas in order to validate whether the cumulative destruction intensity maps also show spatial correlation between the ongoing seismic activity and the preservation conditions of the remains. It is hoped that this research may stimulate new studies concerning the interaction of long-term seismic activities and archaeological remains, especially in view of archaeological site preservation. This interaction depends on many factors that can be studied further by employing in situ or experimental methods. For instance, standing or collapsed remains, buried or exposed remains, stone or mud brick remains will respond differently to earthquakes. Results from such studies can potentially be

A test study from Zakynthos, Greece important as an additional decision factor in setting up excavations or making preservation plans for a site. ACKNOWLEDGMENTS We are grateful to the Greek Institute for Geological and Mineralogical Exploration (I.G.M.E.) and the 35th Ephorate of Prehistoric and Classical Antiquities for the possibility to do fieldwork on the island of Zakynthos. We also would like to thank Atılım Güneş Baydin for helping with the implementation of our model and performing the calculations in Mathematica, and Ioannis Koukouvelas for his constructive comments that significantly improved the manuscript. This chapter is a contribution to the International Geoscience Programme (IGCP) 567, “Earthquake Archaeology: Archaeoseismology along the Alpine-Himalayan Seismic Zone,” funded by the United Nations Educational, Scientific and Cultural Organization (UNESCO). REFERENCES CITED Agalopoulou, P.I., 1973, Mycenaean graves at Kambi, Zakynthos: Archaeologikon Deltion 28A, p. 198–214 (in Greek). Ambraseys, N.N., 1996, Material for the investigation of the seismicity of central Greece, in Stiros, S., and Jones, R.E., ed., Archaeoseismology: British School at Athens, Fitch Laboratory Occasional Paper 7, p. 23–36. Barka, A., and Reilinger, R., 1997, Active tectonics of the Eastern Mediterranean region: Deduced from GPS, neotectonic and seismicity data: Annali di Geofisica, v. XL, no. 3, p. 587–610. Bozorgnia, Y., and Campbell, K.W., 2004, Engineering characterization of ground motion, in Bozorgnia, Y., and Bertero, V.V., ed., Earthquake Engineering: From Engineering Seismology to Performance-Based Engineering: Boca Raton, Florida, CRC Press, Chap. 5, p. 1–74. Burton, P.W., Xu, Y., Tselentis, G.A., Sokos, E., and Aspinall, W., 2003, Strong ground acceleration seismic hazard in Greece and neighboring regions: Soil Dynamics and Earthquake Engineering, v. 23, p. 159–181, doi: 10 .1016/S0267-7261(02)00155-0. Campbell, K.W., and Bozorgnia, Y., 2003, Updated near-source ground motion (attenuation) relations for the horizontal and vertical components of peak ground acceleration and acceleration response spectra: Bulletin of the Seismological Society of America, v. 93, no. 1, p. 314–331, doi: 10.1785/0120020029. Caputo, R., Helly, B., Pavlides, S., and Papadopoulos, G., 2006, Archaeoand palaeoseismological investigations in northern Thessaly (Greece): Insights for the seismic potential of the region: Natural Hazards, v. 39, p. 195–212, doi: 10.1007/s11069-006-0023-9. Cerone, M., Croci, G., and Viskovic, A., 2001, The structural behaviour of the Colosseum, International UNESCO-ICOMOS Congress “More than two thousand years in the history of architecture,” Bethlehem, Palestine, 2001: retrieved in 2010 from the UNESCO 2000 archives: http://www.unesco .org/archi2000/bio/crocicoloss.htm. Clément, C., Hirn, A., Charvis, P., Sachpazi, M., and Marnelis, F., 2000, Seismic structure and the active Hellenic subduction in the Ionian islands: Tectonophysics, v. 329, p. 141–156, doi: 10.1016/S0040-1951(00)00193-1. Day, R.W., 2002, Geotechnical Earthquake Engineering Handbook: New York, McGraw-Hill, 700 p. Ellenblum, R., Marco, S., Agnon, A., Rockwell, T.K., and Boas, A., 1998, Crusader castle torn apart by earthquake at dawn, 20 May 1202: Geology, v. 26, p. 303–306, doi: 10.1130/0091-7613(1998)026<0303:CCTABE >2.3.CO;2. Facaros, D., and Theodorou, L., 2003, Cadogan Guides: Greece: London, Globe Pequot Press, 880 p. Foss, A., 1969, The Ionian Islands: Zakynthos to Corfu: London, Faber, 272 p. Galadini, F., and Galli, P., 2001, Archaeoseismology in Italy: Case studies and implications on long-term seismicity: Journal of Earthquake Engineering, v. 5, no. 1, p. 35–68, doi: 10.1142/S1363246901000236.

155

Hatzidimitriou, P.M., Papadimitriou, E.E., Mountrakis, D.M., and Papazachos, B.C., 1985, The seismic parameter b of the frequency-magnitude relation and its association with the geological zones in the area of Greece: Tectonophysics, v. 120, p. 141–151, doi: 10.1016/0040-1951(85)90092-7. Hinsbergen, D.J., Meer, D.G., Zachariasse, W.J., and Meulenkamp, J.E., 2006, Deformation of western Greece during Neogene clockwise rotation: International Journal of Earth Sciences, v. 95, no. 3, p. 463–490, doi: 10 .1007/s00531-005-0047-5. Homeros, 1998, Odysseia (translated by A. Erhat and A. Kadir): İstanbul, Can Yayınları, 394 p. Homeros, 1999, İlyada (translated by A. Erhat and A. Kadir): İstanbul, Can Yayınları, 592 p. Horn Lopes, J., 2008, Sun over Zakynthos: Lighting Conditions Investigated [B.Sc. thesis]: Institute for Geo- and Bioarchaeology, Faculty of Earth and Life Sciences, VU University Amsterdam, 36 p. Kalligas, P.G., 1993, Habitation on Ancient Zakynthos, in Settlements of Zakynthos from ancient times to the present day: Athens, Institute for Zakynthian Studies, p. 45–73 (in Greek). Koukouvelas, I.K., Katsonopoulou, D., Soter, S., and Xypolias, P., 2005, Slip rates on the Helike fault, Gulf of Corinth, Greece: New evidence from geoarchaeology: Terra Nova, v. 17, p. 158–164, doi: 10.1111/j.1365 -3121.2005.00603.x. Kouskouna, V., and Makropoulos, K., 2004, Historical earthquake investigations in Greece: Annals of Geophysics, v. 47, p. 723–731. Lagios, E., Sakkas, V., Papadimitriou, P., Parcharidis, I., Damiata, B.N., Chousianitis, K., and Vassilopoulou, S., 2007, Crustal deformation in the Central Ionian Islands (Greece): Results from DGPS and DInSAR analyses (1995–2006): Tectonophysics, v. 444, p. 119–145, doi: 10.1016/j.tecto .2007.08.018. Marco, S., 2008, Recognition of earthquake-related damage in archaeological sites: Examples from the Dead Sea fault zone: Tectonophysics, v. 453, p. 148–156, doi: 10.1016/j.tecto.2007.04.011. Marco, S., Hartal, M., Hazan, N., Lev, L., and Stein, M., 2003, Archaeology, history, and geology of the A.D. 749 earthquake, Dead Sea transform: Geology, v. 31, no. 8, p. 665–668, doi: 10.1130/G19516.1. Mylona, Z.A., 2006, The Castle of Zakynthos (translated by A. Doumas): Athens, Hellenic Ministry of Culture Archaeological Receipts Fund, 56 p. Palaima, T.G., 1991, Maritime matters in the Linear B, in Laffineur, R., and Basch, L., eds., Thalassa: L’Egée Préhistorique et la Mer: Actes de la 3e Rencontre égéenne internationale de l’Université de Liège, Station de recherches sous-marines et océanographiques, Calvi, Corse, 23–25 avril 1990: Annales d’Archéologie égéenne de l’Université de Liège (Aegaeum) 7, p. 273–309. Papastamatiou, D., and Psycharis, I., 1993, Seismic response of classical monuments—A numerical perspective developed at the Temple of Apollo in Bassae, Greece: Terra Nova, v. 5, no. 6, p. 591–601, doi: 10.1111/j.1365 -3121.1993.tb00309.x. Papazachos, B.C., 1990, Seismicity of the Aegean and surrounding area: Tectonophysics, v. 178, p. 287–308, doi: 10.1016/0040-1951(90)90155-2. Papazachos, B.C., and Papazachou, C., 1997, The Earthquakes of Greece: Thessaloniki, P. Ziti and Co., 304 p. Papazachos, B.C., Papaioannou, C.A., Margaris, B.N., and Theodulidis, N.P., 1993, Regionalization of seismic hazard in Greece based on seismic sources: Natural Hazards, v. 8, p. 1–18, doi: 10.1007/BF00596232. Papazachos, B.C., Kiratzi, A.A., and Karacostas, B.G., 1997, Toward a homogeneous moment-magnitude determination for earthquakes in Greece and the surrounding area: Bulletin of the Seismological Society of America, v. 87, no. 2, p. 474–483. Papazachos, B.C., Comninakis, P.E., Karakaisis, G.F., Karakostas, B.G., Papaioannou, C.A., Papazachos, C.B., and Scordilis, E.M., 2000, A Catalogue of Earthquakes in Greece and Surrounding Area for the Period 550 BC–1999: Thessaloniki, Geophysical Laboratory, University of Thessaloniki, 333 p. Papazachos, B.C., Comninakis, P.E., Scordilis, E.M., Karakaisis, G.F., and Papazachos, C.B., 2007, A Catalogue of Earthquakes in the Mediterranean and Surrounding Area for the Period 1901–2006: Department of Geophysics, University of Thessaloniki; retrieved in 2007 from http:// lemnos.geo.auth.gr/the_seisnet/WEBSITE_2005/station_index_en.html. Papazachos, C.B., and Kiratzi, A.A., 1996, A detailed study of the active crustal deformation in the Aegean: Tectonophysics, v. 253, p. 129–153, doi: 10.1016/0040-1951(95)00047-X.

156

Tendürüs et al.

Perry, L.J., and Temple, P.G., 1980, The Geological Map of Greece: Zakinthos Island: Athens, Section of Geology and Geological Mapping, Institute of Geological and Mineral Exploration (I.G.M.E.), scale 1:50,000, 1 sheet. Pieters, N., Stoker, A., Tendurus, M., and Wijngaarden, G.J., 2007, Archeologie in een extreem dynamisch landscap: Het Zakynthos Archeologie project: Tijdschrift voor Mediterrane Archeologie, v. 19, no. 1, p. 26–32. Pliny the Elder, 1855, The Natural History (translated by J. Bostock and H.T. Riley): London, Taylor and Francis, available online at the Perseus Digital Library (http://www.perseus.tufts.edu/hopper/text?doc=Perseus :text:1999.02.0137). Psycharis, I.N., Papastamatiou, D.Y., and Alexandris, A.P., 2000, Parametric investigation of the stability of classical columns under harmonic and earthquake excitations: Earthquake Engineering & Structural Dynamics, v. 29, p. 1093–1109, doi: 10.1002/1096-9845(200008)29:8<1093::AID -EQE953>3.0.CO;2-S. Reinhardt, E.G., Goodman, B.N., Boyce, J.I., Lopez, G., van Hengstum, P., Rink, W.J., Mart, Y., and Raban, A., 2006, The tsunami of 13 December A.D. 115 and the destruction of Herod the Great’s harbor at Caesarea Maritime, Israel: Geology, v. 34, no. 12, p. 1061–1064, doi: 10.1130/ G22780A.1. Rink, R., 2005, Geographical Development of Zakynthos during Late Pleistocene–Holocene, Greece [M.Sc. thesis]: Utrecht, Netherlands, Department of Physical Geology, Utrecht University, 34 p. Similox-Tohon, D., Sintubin, M., Muchez, P., Verhaert, G., Vanneste, K., Fernandez, M., Vandycke, S., Vanhaverbeke, H., and Waelkens, M., 2006, The identification of an active fault by a multidisciplinary study at the archaeological site of Sagalassos (SW Turkey): Tectonophysics, v. 420, p. 371–387, doi: 10.1016/j.tecto.2006.03.026. Souyoudzoglou-Haywood, C., 1999, The Ionian Islands in the Bronze Age and Early Iron Age 3000–800 BC: Liverpool, Liverpool University Press, 280 p. Stiros, S., Pirazzoli, P., Laborel, J., and Laborel-Deguen, F., 1994, The 1953 earthquake in Cephalonia (western Hellenic arc): Coastal uplift and halotectonic faulting: Geophysical Journal International, v. 117, p. 834–849, doi: 10.1111/j.1365-246X.1994.tb02474.x. Stiros, S., Pirazzoli, P., Rothaus, R., Papageorgiou, S., Laborel, J., and Arnold, M., 1996, On the date of construction of Lechaion, western harbor of ancient Corinth, Greece: Geoarchaeology, v. 11, no. 3, p. 251–263, doi: 10.1002/(SICI)1520-6548(199605)11:3<251::AID-GEA4>3.0.CO;2-2. Stoker, A., 2006, Remote Sensing in the Zakynthos Archaeology Project. Investigation of the Benefits and Restrictions of Historic Airphotos and Satellite Images in Landscape-Archaeological Research in Greece [M.A. thesis]: Groningen, the Netherlands, Faculty of Arts, University of Groningen, 85 p.

Storme, A.L., 2008, The Landscape at Zakynthos: Detailed Geomorphological Description of the South-Eastern Part of the Vasilikos Peninsula [M.Sc. thesis]: Institute for Geo- and Bioarchaeology, Faculty of Earth and Life Sciences, VU University Amsterdam, 246 p. Tendürüs, M., 2009, Landscape Evolution of the Southern Coastal Plain of Zakynthos, Greece, since the Middle Holocene [M.Sc. thesis]: Institute for Geo- and Bioarchaeology, Faculty of Earth and Life Sciences, VU University Amsterdam, 122 p. Theodulidis, N.P., and Papazachos, B.C., 1992, Dependence of strong ground motion on magnitude-distance, site geology and macroseismic intensity for shallow earthquakes in Greece: I. Peak horizontal acceleration, velocity and displacement: Soil Dynamics and Earthquake Engineering, v. 11, p. 387–402, doi: 10.1016/0267-7261(92)90003-V. Underhill, J.R., 1989, Late Cenozoic deformation of the Hellenide foreland, western Greece: Geological Society of America Bulletin, v. 101, p. 613– 634, doi: 10.1130/0016-7606(1989)101<0613:LCDOTH>2.3.CO;2. U.S. Geological Survey, 2010, Magnitude / Intensity Comparison in the Severity of an Earthquake: United States Geological Survey’s (USGS) General Interest Publication 1989-288-913, retrieved in 2010 from the USGS Earthquake Hazard Program: http://earthquake.usgs.gov/learn/topics/mag _vs_int.php. Van Wijngaarden, G.J., Arapogianni, X., Rink, R., and Tourloukis, V., 2006, The Zakynthos Survey 2005. Preliminary report of a pilot study, Pharos: Journal of the Netherlands Institute in Athens, v. XIII, p. 59–76. Van Wijngaarden, G.J., Sotiriou, A., Pieters, N., and Tourloukis, V., 2007, The Zakynthos Archaeology Project 2006. Preliminary report of the first season, Pharos: Journal of the Netherlands Institute in Athens, v. XIV, p. 29–46. Van Wijngaarden, G.J., Sotiriou, A., Horn Lopes, J., Gouma, M., Koster, K., Stoker, A., Susan, D., and Tourloukis, V., 2010, The Zakynthos Archaeology Project, 2008. Preliminary report on the 2008 season, Pharos: Journal of the Netherlands Institute in Athens, v. XVI, 61–83. Van Wijngaarden, G., Sotiriou, A., Pieters, N., Abed, K., and Tendürüs, M., 2009, The Zakynthos Archaeology Project. Preliminary report of the 2007 season, Pharos: Journal of the Netherlands Institute in Athens, v. XV, p. 43–57. Von Stein, I., 2009, Just in Time… Assessing Cape Kalogeros before It Is Gone [Master’s thesis]: Amsterdam, Amsterdam Archaeological Centre, University of Amsterdam, 61 p. Yeats, R.S., Sieh, K.E., and Allen, C.R., 1997, The Geology of Earthquakes: New York, Oxford University Press, 568 p. Zois, L., 1955, History of Zakythos: Athens, 439 p. (in Greek). MANUSCRIPT ACCEPTED BY THE SOCIETY 19 MAY 2010

Printed in the USA

The Geological Society of America Special Paper 471 2010

Assessment of seismically induced damage using LIDAR: The ancient city of Pınara (SW Turkey) as a case study Barış Yerli Johan ten Veen* Institute for Geology, Mineralogy, and Geophysics, Ruhr University, Universitätsstrasse 150, D-44801, Bochum, Germany Manuel Sintubin Geodynamics & Geofluids Research Group, Katholieke Universiteit Leuven, Celestijnenlaan 200E, 3001 Leuven, Belgium Volkan Karabacak C. Çağlar Yalçıner Erhan Altunel Engineering Faculty, Department of Geological Engineering, Osman Gazi University, Eskisehir, Turkey

ABSTRACT Seismic-related damages of archaeological structures play an important role in increasing our knowledge about the timing and magnitudes of historical earthquakes. Although quantitative data should form the basis of objective archaeoseismological methods, most studies still do not rely on such methods. Ground-based LIDAR (light detection and ranging) is a promising, rather new, scanning technology that determines spatial position of an object or surface and provides high-resolution threedimensional (3-D) digital data. Using LIDAR, we mapped the damage and overall attitude of a Roman theater in the ancient Lycian city of Pınara (500 B.C.–A.D. 900), located at a faulted margin of the Eşen Basin (SW Turkey). An average 0.81°°NW tilt of the 20 seating rows could be computed from the LIDAR data. Conventional compass readings of these seating rows did not provide the same results because errors involved with this method are generally >2°°. The tilt direction appears perpendicular to the NE-trending basin-margin fault, suggesting that fault-block rotation is the most likely mechanism to have induced the systematic tilt of the theater. The estimated 4 m offset on this normal fault should be seen as a rough estimate of the total displacement and was likely produced by several (more than one) earthquakes with magnitudes of M = 6–6.8. This is consistent with historical records that mention several large earthquakes during the Roman period.

*Current address: TNO Built Environment and Geosciences, P.O. Box 80015, Utrecht 3508 TA, Netherlands. Yerli, B., ten Veen, J., Sintubin, M., Karabacak, V., Yalçıner, C.Ç., and Altunel, E., 2010, Assessment of seismically induced damage using LIDAR: The ancient city of Pınara (SW Turkey) as a case study, in Sintubin, M., Stewart, I.S., Niemi, T.M., and Altunel, E., eds., Ancient Earthquakes: Geological Society of America Special Paper 471, p. 157–170, doi: 10.1130/2010.2471(14). For permission to copy, contact [email protected]. © 2010 The Geological Society of America. All rights reserved.

157

158

Yerli et al.

INTRODUCTION Archaeoseismological research commonly focuses on establishing a link between damages to archaeological building structures (e.g., tilted, toppled blocks, disorientation of structures, fallen columns, etc.) and historical or younger earthquakes. The main challenge lies in separating seismic-related damage from other, natural or human, causes for destruction. Archaeoseismology is faced with the difficulty that constructions are either too intensely destroyed, or that damage is too indistinct to give any detailed information on its nature. Because of poor spatial and temporal resolution, amalgamation or duplication of seismic events is a well-known pitfall (Ambraseys et al., 2004; Guidoboni, 2002). Therefore, the most reliable results of an archaeoseismological investigation are obtained by application of modern geoarchaeological practices (archaeological stratigraphy plus geological-geomorphological data), augmenting a quantitative approach and (if available) historical information (Galadini et al., 2006). Data obtained from archaeoseismological investigation can be either qualitative or quantitative. By qualifying damage to archaeological structures, an earthquake magnitude can be estimated based on the Medvedev-Sponheuer-Karnik (MSK) scale (Medvedev et al., 1964) or the modified Mercalli (MM) scale (Wood and Neumann, 1931). However, the assessment of earthquake magnitude is imprecise and often based on a presumed analogue with earthquake effects on modern man-made structures (e.g., Karcz and Kafri, 1978; Altunel, 1998). Recent engineering seismological models can be used to test the hypothesis that observed building damage is of seismogenic nature by seeking a systematic relation between building response and the seismic source or ground motion (e.g., Hinzen, 2005), and to test the archaeoseismic hypothesis (Galadini et al., 2006). Quantitative data can be grouped as either directional or spatial and can be applied for the purposes of geophysical engineering or kinematic analysis. Directional data include the sense of slip on faulted archaeological relics and can be used (in a similar fashion as in paleoseismology) as supplementary information to discover previously unknown earthquakes (Galli and Galadini, 2001). On the basis of the amount of offset and age of displaced structures, a realistic value of earthquake magnitude and a rough evaluation of the recurrence time can be obtained. In addition, information on faulted relics (fault direction and offset direction) can be treated as fault-kinematic data and as such be used for strain analysis (e.g., Hancock and Altunel, 1997). The parallel direction of fallen columns is often used as an indicator for earthquake damage (Stiros, 1996). Nur and Ron (1996) suggested a relationship between ground motion and fall direction, where the latter is an indicator of the direction of fault-rupture propagation. However, the behavior of columns during shaking is very complex and depends on different physical parameters, such as ground motion, material characteristics, and type of building foundation (Ambraseys, 2006). Numerical modeling studies show that the fall direction of single standing columns

is highly chaotic and can be influenced by small anisotropies in constructional elements or variations of the ground motion (Hinzen, 2009). Up until recently, spatial data of archaeological sites were mostly unavailable, but they are now easier to acquire as a result of the arrival of high-resolution laser detection techniques. Spatial data provide information on the position of archaeological relics such as walls and floors. Such positional data alone is informative, but data treatment can reveal information on tilt, torque, and dislocation of building elements. This type of information is elementary in establishing links between damage and faulting and may find its way in future finite-element simulation and reconstructions as well. Here, we test data obtained by LIDAR (light detection and ranging system) against conventionally obtained data in order to test the hypothesis of faulting-induced tilting of a Roman theater at the archaeological site of Pınara. TECTONIC SETTING Neotectonics of SW Turkey Turkey is characterized by complex neotectonic deformation that is predominantly related to the convergence of Africa and Eurasia. The North and East Anatolian fault systems are the most prominent structures that accommodate the collision of the African plate’s Arabian promontory with Eurasia (e.g., Barka and Kadinsky-Cade, 1988; Westaway, 1994; Armijo et al., 1999; ten Veen et al., 2009). Southwestern Turkey is transected by numerous faults that are thought to connect southward with major fault zones associated with the Hellenic subduction zone (Eyidoğan and Barka, 1996; ten Veen and Kleinspehn, 2002, 2003; ten Veen et al., 2004). The Fethiye-Burdur fault zone (Barka and Reilinger, 1997) is a domain characterized by many subparallel fault segments, roughly aligned in the area between Fethiye and Burdur. Different styles of deformation and a wide variety of fault orientations suggest that this zone has been affected by three different tectonic phases from late Miocene until present (ten Veen, 2004; Alçiçek, 2007; ten Veen et al., 2009). Recent Earthquake Activity in SW Turkey The Fethiye-Burdur fault zone has been attributed one of the highest seismic hazard designations in Turkey, based on the concentration of recent and historical seismicity in the area (see Fig. 1B), comparable to regions situated along the North Anatolian fault zone. Instrumental data show that most earthquakes in southwestern Turkey have a shallow focal depth (<15 km), suggesting that they are not related to African subduction, but to upper-crustal deformation instead. Subcrustal- and intermediatedepth earthquakes related to subduction occur mostly offshore, south of Rhodes and Cyprus islands, close to the Hellenic-Cyprus arc system (Ambraseys, 2001; Tan et al., 2008). In spite of the high seismic potential (Erdik et al., 1999), the Fethiye-Burdur fault zone is seismically defined by weak to

The ancient city of Pınara, SW Turkey

159

A North Anatol ian fau lt

lt fau ne an i l zo o t re u Ana t t s Ea s su Bitli

De ad S e

a fault

Isparta angle

Mediterranean Ridge

30ºE

35ºE

40ºE

B

40ºN

35ºN

moderate (M ≤ 6) earthquakes only. Large earthquakes occur only at the northeastern and at the offshore southwestern extremities (Rhodes 1957, offshore Turkey 1926, 1969, 1975, 1980, and 2001). The entire central segment of the Fethiye-Burdur fault zone is seemingly seismically quiescent; however, damage to some historical sites in the region, including Sagalassos (Sintubin et al., 2003; Similox-Tohon et al., 2006), Hierapolis (Hancock and Altunel, 1997), and Cibyra (Akyüz and Altunel, 2001), indicates that destructive historical earthquakes have occurred in the region.

0

km

Figure 1. (A) Geodynamic framework of the Eastern Mediterranean (modified after ten Veen, 2004). ST—Strabo trench; PT—Pliny trench; M M—Menderes Massif; EAEP—Eastern Aegean extensional province; FBFZ—FethiyeBurdur fault zone. (B) Seismicity map of Turkey (modified after Tan et al., 2008); dark circles represent historical earthquakes between 2100 B.C. and A.D. 1963 (Pınar and Lahn, 1952; Ergin et al., 1967, 1971; Soysal et al., 1981; Güçlü et al., 1986; Ambraseys and Finkel, 1995; Ambraseys and Jackson, 1998); white circles represent the earthquake activity between 1964 and 2004 (M ≥ 4.0; ISC, 2001). Data from databases were made available online by ISC—International Seismological Centre, USGS-NEIC—United States Geological Survey–National Earthquake Information Centre, KOERI—Kandilli Observatory and Earthquake Research Institute, EMSC—European Mediterranean Seismological Centre.

200

PINARA SITE Eşen Basin The NE-trending Eşen Basin (Figs. 1 and 2) is the southernmost of several extensional basins located in the FethiyeBurdur fault zone. It is ~15 km wide and 30 km long and formed behind the front of the Lycian Nappes close to the contact with the Beydağları autochthon (Fig. 2). The Lycian Nappes basement consists of ophiolitic rocks, limestones, and turbiditic sandstones

en Ör

Paşalı Ören

Talus

Üzümlü

Terrace

ERENDAĞ

Çaykenarı & Kocaçay Mb.

t faul dağ Eren

L Miocene- PleistoL Pliocene cene

lt fau

Haticeana DağıModern 1079 m alluvium

Araplar 1000 m

Çetilidüz Mb.

Mesozoic - Miocene

Ophiolite Nappe Elamlı Nappe

Kemer

Yeşilbarak Nappe

Fethiye

Beydağları autochthon Normal fault

ult Akdağ fa

Akdağ 3015 m

u

l t

Thrust fault

lıke nt f aul t

F

a

Babadağ 1968 m

ç

Fig. 3A

O

ğ

a

36 30’ 00’’N

Sak aÇ

a

c Ko

b

a

a

Pınara

K

ay

Eşen

e p

dağ fault

Dumanlı

t o n fr

Dumanlıdağ 1956 m

N

Sandık Dağı 1009 m

E

Kınık fa

L y c i an

M

ult

D I

T

ay

E

Eşen Ç

R

R

a p

T KA

N RA

DA

Ğ

Çayağzı

A

0

10 km

A

Kalkan

N

O

E

29 30’ 00’’E

N

Figure 2. Geological map of the Eşen Basin (modified after ten Veen, 2004; Alçiçek, 2007; Şenel, 1997a, 1997b, 1997c).

The ancient city of Pınara, SW Turkey (“flysch”). The basin is bounded and transected by a series of N- to NE-trending faults that have exercised important control on the character and distribution of sedimentary facies since the late Miocene. Several E-W–trending faults crosscut the basin and are suggested to be of Quaternary age (ten Veen, 2004), although a relationship with deep-seated basement structures is not unlikely (ten Veen et al., 2009). The Neogene infill of the basin is up to 150 m thick and consists of alluvial-fan, fluvial-deltaic and lacustrine deposits ranging in age from late Miocene to Pliocene. Pleistocene conglomeratic terraces and alluvium overlie the basin fill with an angular unconformity (ten Veen, 2004; Alçiçek, 2007). Geology of Pınara and Surroundings The ancient city of Pınara is situated along the faultcontrolled western border of the Eşen Basin, near the modern village of Minare. The archaeological site is spread over a 3–4 km2 area, where a variety of geological units are exposed (Fig. 3A). The basement along the basin margin is characterized by a pre–late Miocene thrust contact between two limestone nappes. This contact is demarcated by a relatively thin flysch unit with exotic limestone blocks up to 50 m in width (Fig. 3B). From the late Miocene onward, the thrust contact was down-faulted to the east along the basin-margin normal faults. The long faulting history is exemplified by tilted Upper Miocene fluvial deposits that abut against the marginal fault and that are unconformably overlain by gently tilted Pleistocene terrace deposits. The largest part of the city was built on a down-faulted limestone basement block, whereas tombs of the Acropolis were mainly carved in an ~80-m-high, NE-SW–trending fault face to the west (Fig. 4A). This NE-oriented fault segment is well exposed and shows faultplane undulations and striations that indicate a dip-slip normal sense of slip (Fig. 4B). No archaeological remains are present on the basement flysch, which probably was used for agricultural purposes. The Roman theater was built on a slightly westwardtilted Pleistocene terrace that rests on top of the flysch. ARCHAEOSEISMOLOGICAL POTENTIAL The city of Pınara was founded by colonists from Xantos between the fifth and fourth centuries B.C. It was one of the largest cities in the powerful Lycian league and located along the main Lycian road (Akurgal, 1978). Later, the city was under Roman and Byzantine control. An important event in the history of Lycia was the conquest by Alexander the Great in the winter of 334–333 B.C., after which Pınara accepted his domination (Bean, 1978). After the death of Alexander in 323 B.C., the Roman period started around 309 B.C. The city was affected by three earthquakes in A.D. 141 (Guidoboni, 1994; Lang, 2003; Akşit, 2006), ca. A.D. 240 (Lang, 2003; Akşit, 2006), and A.D. 1851 (Soysal et al., 1981). According to ancient accounts, Pınara received aid for the reconstruction of the city after these earthquakes (Guidoboni, 1994; Lang, 2003; Akşit, 2006). Eventually, the city lost its financial and geographic importance, and at the

161

end of the ninth century A.D., the city was abandoned. Based on the historical earthquake database and substantiated by passedon personal accounts, a severe earthquake struck the Pınara site in 1851 as well. A variety of damage features, including unidirectional tumbled column blocks, systematic dislocation of building stones, completely tilted objects, and even faulted remains, attest to earthquake-related phenomena. Many walls and buildings in the agora are strongly disturbed and often show typical sinuous arrangements of ejected building blocks with gaping joints in between. The northern temple wall shows a combination of polygonal Hellenistic and rectangle-shaped Roman masonry, indicating that the wall was reconstructed under Roman control (i.e., after A.D. 46; Wurster and Wörrle, 1978). The severe damage, dislocation, and rotation of both types of masonry show that this building suffered from a later earthquake as well. In addition, the vertical cracks and joints of numerous beams, door frames, and window frames are consistent with extension related to horizontally oscillating ground motion. Overall, the archaeological relicts in Pınara ancient city exhibit damages that are expected from earthquake(s) with intensities between VIII and IX according to the MSK scale. Given that historical accounts make no mention of any devastating war in the history of Pınara, the structural damage would appear likely to have been produced by known earthquakes reported by ancient literary sources. In the following sections, detailed analyses and interpretation of such structural damage are described. QUANTITATIVE ASSESSMENT OF DAMAGE Methods In order to test the value and merits of directional data versus spatial data in recording suspected archaeoseismic evidence, the Roman theater of Pınara was surveyed using both a conventional geological compass and a laser scanning tool. While, the attitude (dip directional and dip angle) of the theater’s seating rows was measured at 103 locations with a geological compass, a ground-based LIDAR system was used to map the entire theater. Ground-based LIDAR is widely applied in research and industry because of its high efficiency, ease of use, and the fact that its accuracy far exceeds that of conventional measuring techniques. Its first application in a neotectonic setting was in paleoseismological trench logging of the San Andreas fault zone (Niemi et al., 2004). More recently, it has been also used for archaeoseismological research to determine and analyze the deformation due to active faulting (Karabacak et al., 2007; Karabacak et al., 2008). Although both ground- and air-based LIDAR has been used by archaeological researchers before (e.g., Devereux et al., 2005; Lambers et al., 2007; Frischer and Dakouri-Hild, 2008), it was mostly used as a high precision photo-mapping tool. Here, we take advantage of the high-resolution quantitative data set it produces to objectively demonstrate damage and dislocation to archaeological structures.

162

Yerli et al.

A 250

0

500 m

A

Pınara

36º 29' 30''N

Pleistocene

Modern alluvium

Miocene L Pliocene

L Miocene-

Terrace

A'

Çaykenarı & Kocacay Mb.

29º 36' 30''E

Elamlı Nappe Yeşilbarak Nappe Normal fault Thrust fault

NW

1000 950 900 850 800 750 700 650 600 550 500 450 400 350 300 250 200 150

SE

Pinara

B Roman theatre

? 0

200

400

600

800

1000

1200

1400

1600

1800

2000

2200

2400

2600

2800

3000

3200

3400

3600

3800

4000

4200

Figure 3. (A) Detailed geological map of the ancient city of Pınara, showing the position of the main buildings. 1—Roman theater, 2—bath, 3—Odeon, 4—temple, 5—tomb, 6—agora, 7—necropolis, 8—rock tombs. (B) Geological cross section A–Aʹ (in meters).

The LIDAR instrument used here was an Optech ILRIS-3D with a class 1 laser at 1500 nm wavelength, integrated digital camera, and data sampling rate of 2500 points per second. It has a high data rate and large dynamic range from 3 m to 1500 m (Optech, 2006). The complete equipment consisted of a LIDAR instrument, a tripod, a laptop, and four heavy batteries as energy supply (Figs. 5A–5C). The scanning duration in the field depends on scanning resolution and range, size of the scanned image or land-

scape, and eventually the capacity of batteries. The basic principle of operation is that the laser light is directed at the object, and the bidirectional time is measured and converted into a distance. The reflection points are accurately positioned in the space using the distance and the laser-beam angle (Bonnaffe et al., 2007), providing a point cloud with 3-D information for the scanned object. The theater was scanned from three different places to avoid gaps in the point cloud. To obtain a detailed scanner image, we

The ancient city of Pınara, SW Turkey

163

1 km

Figure 4. (A) Eastern (front) and western (back) segments of the Kabaağaç fault in Pınara; view is to the west. Tomb hill is formed by the ~80-m-high, intervening fault block. (B) Detail of the fault surface with meter-scale undulations and near-vertical groves; the hill is ~40 m high.

1m

chose 2 cm scanning resolution. Each of the three scans took ~2.5 h in Pınara. However, the scanning was allocated in 2 d because of the capacity of batteries. After the scanning process, the LIDAR provided a 3-D image of the Roman theater in an x-y-z point cloud that includes ~5.5 million sample points (Fig. 6). The resulting high-resolution 3-D data allow deformations (e.g., tilting, dislocation, cracks) of the archaeological structures to be measured and/or calculated using the spatial analysis programs Surfer® and Petrel®. The Roman theater is oriented N-S, has a typical semicircle form and a radius of ~25 m, and the opening (Cavea) of theater is facing west. It originally had 27 seating rows, of which only 20 partially damaged (i.e., growing tree roots, weathering) rows are preserved (Fig. 5). Ten staircases divide the rows into nine sections (Cunei). The original capacity was estimated at ~3200 seats (Wurster and Wörrle, 1978). In assessing the attitude of the

theater’s seating rows, the top surfaces of individual rows were selected from the data set, resulting in 20 subsets. Assuming that the theater was accurately built, the top surfaces of each row would have originally lain in a horizontal plane. Any present-day deviation of this plane from horizontal is seen as an externally induced deformation. To obtain a clean data set, erroneous data points that are related to vegetation, fallen stones, or other artifacts were removed. At some places, the seating rows are crosscut by fractures, along which individual seating blocks are severely tilted. Since we are interested in the overall attitude of the theater, these blocks were removed from the data set as well. For each row, the range of the remaining z values (delta; Table 1) is significant with respect to the LIDAR measurement error of 0.02 m, permitting the reconstruction of a best-fit plane through the data points. The best-fit planar surfaces and their dip and azimuth values were calculated using Petrel® (Table 1).

A

B

C E

W

E

D

W

E

nt

ra

ern

a Pın

st we

Roman theater

ea

Agora

Ne K cro ste aba po rn ağ li s Pın aç ar fau a s lt eg me nt

ğa

aa

b Ka

s

Dip direction

lt-

au çf

e egm

F

Figure 5. (A) Light detection and ranging (LIDAR) scanning of the entire theater, producing the first data set; view is toward the east. (B) Scanning from the northern wing of theater, producing data set two of the theater’s southern wing; note the tilted seating blocks associated with small faults transecting the theater. (C) Scanning from southern wing, producing data set three of the theater’s northern wing. (D) Rotated building stones at the northern wing of theater. (E) Tilted and partly collapsed wall at the southern part of theater. (F) Location and dip direction of the Roman theater with respect to the basin-margin fault segments around Pınara, projected on a Google® image. (Image width: 1900 m.)

The ancient city of Pınara, SW Turkey

165

A

Figure 6. (A) Three-dimensional (3-D) digital model based on light detection and ranging (LIDAR) data with example of planar surfaces through the data points of individual seating rows. Constructions were made with Petrel®. Visualizations were made in Surfer®. (B) Map view of the theater with indication of average dip direction.

(m)

B (m)

Results The compass data show a 360° spread in dip directions, although northerly directions dominate, with an average direction of N33E (Fig. 7A). Dip angles vary between 1° and 5°, with an average of 2.6° (see Table 2; Fig. 7). A rearrangement of the data according to seating row number (in total, 20 rows are preserved) shows that both dip direction and dip angle variations are nonsystematic. This relatively large spread in data is due to the roughly weathered surfaces of the seating rows, inhibiting good compass measurements. In contrast, the seating-row attitudes based on LIDAR data show a much narrower range in dip directions and dip angles (Fig. 8). All 20 seating rows are tilted toward the NW (Fig. 9),

with angles ranging between 0.49° and 0.99°. Moreover, seating rows 1–16 show an overall increase in dip angle from 0.49° to 0.96° (Fig. 8A). This systematic increase in inclination is less apparent from rows 18 to 20. Rows 2–16 also demonstrate an almost systematic counterclockwise rotation (decrease in dip direction) going from N319E to N298E (Fig. 8B). The higher rows deviate from this trend in dip direction. Data Interpretation The systematic NW tilt of the seating rows in the Roman theater based on the LIDAR data (Fig. 9) requires explanation. The mean dip angle of 0.81° is small, far below that what the human eye can detect. However, across the width of the theater

TABLE 1. DIP DIRECTIONS, DIP ANGLES, AND STATISTICAL PARAMETERS OF THE BEST-FIT PLANES THROUGH THE LIDAR DATA OF INDIVIDUAL SEATING ROWS Row 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

Zmin 0.96 1.32 1.69 2.07 2.43 2.83 3.20 3.57 3.86 4.30 4.68 5.09 5.56 5.91 6.70 7.05 7.46 7.90 8.33 6.41

ΔZ 0.14 0.18 0.21 0.23 0.28 0.31 0.36 0.36 0.46 0.43 0.45 0.45 0.40 0.45 0.47 0.52 0.55 0.52 0.47 0.38 0.38

Zmax 1.11 1.49 1.90 2.30 2.71 3.14 3.56 3.93 4.33 4.73 5.14 5.55 5.96 6.36 7.18 7.57 8.01 8.42 8.80 6.78 Mean Δ

Mean 1.04 1.43 1.83 2.21 2.63 3.05 3.43 3.84 4.20 4.64 5.03 5.42 5.86 6.23 7.04 7.41 7.78 8.20 8.62 6.67

SD 0.05 0.06 0.07 0.07 0.08 0.09 0.10 0.11 0.13 0.12 0.12 0.13 0.09 0.12 0.11 0.11 0.13 0.13 0.12 0.09

Error 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02

Significant Dip (°) Dip direction yes 0.49 336.89 yes 0.64 319.71 yes 0.68 322.92 yes 0.72 318.40 yes 0.77 317.14 yes 0.79 312.58 yes 0.85 319.03 yes 0.87 318.87 yes 0.95 318.05 yes 0.99 313.57 yes 0.96 314.71 yes 0.95 311.38 yes 0.92 310.05 yes 0.92 306.00 yes 0.96 299.89 yes 0.94 298.08 yes 0.72 305.14 yes 0.63 319.14 yes 0.74 315.65 yes 0.65 308.63 Mean value 0.81 314.29 Standard deviation 0.14 8.61 Note: If Δz is greater than the measurement error of 2 cm, the difference in height is inferred to be significant. Mean value and standard deviation of seating rows. LIDAR—light detection and ranging; SD—standard deviation.

A

Dip direction 0 -2 0 14

3 4 0 -3 6 0

2 0 -4 0

12

3 2 0 -3 4 0

4 0 -6 0

10 8

3 0 0 -3 2 0

6 0 -8 0

6 4

2 8 0 -3 0 0

8 0 -1 0 0

2

N = 103

0 1 0 0 -1 2 0

2 6 0 -2 8 0

2 4 0 -2 6 0

1 2 0 -1 4 0

2 2 0 -2 4 0

1 4 0 -1 6 0 1 6 0 -1 8 0

2 0 0 -2 2 0 1 8 0 -2 0 0

B

C

6

Dip Direction

Dip Angle

5 4 3 2 1 0

360 330 300 270 240 210 180 150 120 90 60 30 0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

Seat Rows

Seat Rows

Figure 7. Compass measurements of dip direction and angles of top surfaces of seating rows from the Roman theater at Pınara: (A) all measurements, (B) dip angle per seating row, and (C) dip direction per seating row.

TABLE 2. AVERAGE DIP DIRECTIONS, DIP ANGLES, AND STATISTICAL PARAMETERS OF THE SEATING ROWS MEASURED BY CONVENTIONAL COMPASS Seat row 1 2 3 4 5 6 7 8 9 10

Dip direction (aspect) 062 142 035 265 088 041 035 175 004 003 Standard deviation Mean value

Dip angle (slope, °) 2.1 2.4 3.0 2.5 2.5 5.0 2.8 2.0 2.4 1.8

Seat row 11 12 13 14 15 16 17 18 19 20

2 1.8

Dip Angle

1.4 1.2 1

0.84

0.69

0.6 0.4 0.2 0 0

1

2

3

4

A

5

6

7

8

9

Dip Direction

1.6

0.8

Dip direction (aspect) 325 041 292 012 180 285 325 350 010 151 80 033

Dip angle (slope, °) 2.0 2.0 2.5 2.4 3.0 3.0 1.8 3.1 2.0 4.6 0.8 2.6

360 330 300 270 240 210 180 150 120 90 60 30 0

10 11 12 13 14 15 16 17 18 19 20

0

1

2

3

4

5

6

7

8

9

10 11 12 13 14 15 16 17 18 19 20

B

Seat Rows

Seat Rows

Figure 8. Dip angles and the dip directions for the 20 seating rows of the theater computed from light detection and ranging (LIDAR) data. The data show that the Roman theater is systematically tilted toward the NW with an average dip angle of 0.81°. (A) Dip angle per seating row. (B) Dip direction per seating row.

SE

NW

10 9

Row20 Row19 Row18 Row17 Row16 Row15 Row14 Row13 Row12 Row11 Row10 Row9 Row8 Row7 Row6 Row5 Row4 Row3 Row2 Row1

8

Z (m)

7 6 5 4 3 2 1 0

0

10

Distance (m)

20

30

Figure 9. Cross section showing seating row heights of the Roman theater projected in a N134E–N314E transect, i.e., in the maximum dip direction. All 20 rows show a significant and systematic tilt toward the NW. Note that the projection plane and dip direction are at an angle of 46° with respect to the N-S–oriented theater, such that multiple sections are projected at the same distance from origin (0). Small steps in the height profile are due to small faults observed in the theater (see also Fig. 5B).

168

Yerli et al.

(~20 m), this inclination approximates to ~0.28 m (mean delta = 0.38 m) elevation difference. This value is significant, and it is considered too high to be explained by nonsystematic construction errors. The Romans are renowned for their engineering precision and used instruments like the groma, the dioptra, and the chorobates to establish exact vertical and horizontal alignments. Construction errors due to malfunctioning of these leveling tools, if at all, are expected to be neither large nor systematic. An alternative cause for the tilting of the theater might be creep of a soft-sediment substrate, “soil creep.” However, the theater is founded in firmly cemented terrace conglomerates, a unit not likely to be affected by processes of slow and incremental creep. Coherent sliding of the entire terrace could be plausible, but this would require a tectonic mechanism that increased the underlying gradient, i.e., tilting, such that sliding could have been initiated. A mechanism for this gradient chance is fault-block rotation, and support for this tendency comes from the tilt direction of the seating rows, which is exactly perpendicular to the NE-trending basin-margin fault (Figs. 4 and 5F). The theater is located only 250 m from this fault trace (Fig. 3B), and displacement along its 60° inclined fault plane could have caused back tilting of the hanging-wall block. As a first approximation, it is assumed that for small offsets, the fault can be assumed planar and that a simple geometric relationship exists between fault offset and dip of the hanging wall. Using that approach and based on a 250 m distance and a mean seating-row dip of 0.8°, a total fault offset of 4.0 m is calculated. Since this simple approach does not take the effects of elastic response to fault slip into account, the offset value should be considered a maximum. Note that for listric faults, a rather complex geometrical relationship exists among the fault angle, offset, and the dip of sedimentary layers in the hanging-wall block (White et al., 1986). For planar domino-style normal faults, such a relationship also appears to be dependent on the dip of the basal detachment (Axen, 1988). Since the geometries of the basin-margin faults in the subsurface of the Eşen Basin are unknown, as is the presence of detachments faults, our simple geometrical approach is appropriate. The linear relationship among fault displacement, moment magnitude, and fault-rupture length given by Wells and Coppersmith (1994) infers that a total fault displacement of 4.0 m equates to a M ~7.4 earthquake and a surface rupture of ~50– 60 km. The ~20 km fault length of this fault and the VIII–IX (MSK) damage intensity in the ancient city of Pınara indicate that earthquake magnitudes would not have exceeded M ~6.5. This implies that more than one earthquake event is responsible for the total fault displacement and tilt of the theater. This inference is supported by the record of historical earthquakes, which describes three large earthquakes in ancient times in A.D. 141, 240, and 1850, respectively. However, it should be considered that fault slip that is integrated over several earthquake cycles may contain a contribution from surface processes (erosion and sedimentation) as well (Maniatis et al., 2009). Up to 15% of the

total displacement can be attributed to footwall erosion, and disregarding this effect may lead to overestimation of the surface rupture and thus the earthquake magnitude. The systematic increase in dip angle and decreasing dip direction of rows 1–16 (Fig. 8) is not explainable by a faultdirected tilt of the theater alone. Instead, it suggests that outward bending of the theater side walls (Figs. 5D–5E) also plays a role. This mechanism may have enhanced the faulting-induced tilt. An obvious drop in the dip angle of seating rows 17–20 (mean value of 0.69°), compared to rows 1–16 (mean value 0.84°), can, for instance, be explained by a tilting event during the construction of the theater, i.e., during the Roman period. Alternatively, the abnormal azimuth and dip values of the upper four rows may be related to more severe outward bending. The latter explanation is corroborated by a higher number of tilted and dislocated blocks due to roots of vegetation in the upper reaches of the theater (see Fig. 6A). CONCLUSIONS High-resolution data from the Roman theater at Pınara were used to reveal a systematic deviation of the theater seating rows from a normal, i.e., horizontal, position. Discarding the possibility of man-made construction errors or soil creep, a causal relationship between the tilt of the seating rows and slip at the nearby basin-margin fault can be envisaged. The past occurrence of several damaging earthquakes is substantiated by historical records, and overall damage to the site provides a rough guide for approximate magnitudes. This first attempt to use LIDAR data for quantitative computations is promising for future further application, such as geophysical modeling aimed at seeking a systematic relation among the seismic source, ground motion, and building response. ACKNOWLEDGMENTS We thank the volume editors for providing us the opportunity to present our study in this Special Paper. We appreciate the very supportive and constructive comments of volume editor I. Stewart, reviewer Z. Çakır, and an anonymous reviewer. We are grateful for the financial support provided by the German Science Foundation (DFG). We would like to thank Cihat Alçiçek for the discussions and support during the fieldwork. This article is a contribution to the United Nations Educational, Scientific and Cultural Organization–funded International Geoscience Programme IGCP 567, “Earthquake Archaeology: Archaeoseismology along the Alpine-Himalayan Seismic Zone.” REFERENCES CITED Akşit, İ., 2006, The Land of Light: Lycia: İstanbul, Akşit Yayıncılık, 179 p. Akurgal, E., 1978, Ancient Civilizations and Ruins of Turkey: Ankara, Türk Tarih Kurumu Basim Evi, 112 p. Akyüz, H.S., and Altunel, E., 2001, Geological and archaeological evidence for post-Roman earthquake surface faulting at Cibyra, SW Turkey: Geodinamica Acta, v. 14, p. 95–101, doi: 10.1016/S0985-3111(00)01057-3.

The ancient city of Pınara, SW Turkey Alçiçek, M.C., 2007, Tectonic development of an orogen-top rift recorded by its terrestrial sedimentation pattern: The Neogene Eşen Basin of southwestern Anatolia, Turkey: Sedimentary Geology, v. 200, p. 117–140, doi: 10.1016/j.sedgeo.2007.04.003. Altunel, E., 1998, Evidence for damaging historical earthquakes at Priene, western Turkey: Turkish Journal of Earth Sciences, v. 7, p. 25–35. Ambraseys, N.N., 2001, Reassessment of earthquakes, 1900–1999, in the Eastern Mediterranean and the Middle East: Geophysical Journal International, v. 145, p. 471–485, doi: 10.1046/j.0956-540x.2001.01396.x. Ambraseys, N.N., 2006, Earthquakes and archaeology: Journal of Archaeological Science, v. 33, p. 1008–1016, doi: 10.1016/j.jas.2005.11.006. Ambraseys, N.N., and Finkel, C.F., 1995, The Seismicity of Turkey and Adjacent Areas: A Historical Review, 1500–1800: Istanbul, Turkey, Muhittin Salih Eren, 240 p. Ambraseys, N.N., and Jackson, J.A., 1998, Faulting associated with historical and recent earthquakes in the Eastern Mediterranean region: Geophysical Journal International, v. 133, p. 390–406, doi: 10.1046/j.1365-246X .1998.00508.x. Ambraseys, N.N., Smit, P., Douglas, J., Margaris, B., Sigbjörnsson, R., Ólafsson, S., Suhadolc, P., and Costa, G., 2004, Internet site for strong-motion data: Bollettino di Geofisica Teorica Applicata, v. 45, no. 3, p. 113–129. Armijo, R., Meyer, B., Hubert, A., and Barka, A., 1999, Westward propagation of the North Anatolian fault into the northern Aegean: Timing and kinematics: Geology, v. 27, no. 3, p. 267–270, doi: 10.1130/0091 -7613(1999)027<0267:WPOTNA>2.3.CO;2. Axen, G.J., 1988, The geometry of planar domino-style normal faults above a dipping basal detachment: Journal of Structural Geology, v. 10, no. 4, p. 405–411, doi: 10.1016/0191-8141(88)90018-1. Barka, A., and Kadinsky-Cade, K., 1988, Strike-slip fault geometry in Turkey and its influence on earthquake activity: Tectonics, v. 7, no. 3, p. 663–684, doi: 10.1029/TC007i003p00663. Barka, A., and Reilinger, R., 1997, Active tectonics of Eastern Mediterranean region: Deduced from GPS, neotectonic and seismicity data: Annali di Geofisica, v. 40, no. 3, p. 587–610. Bean, G.E., 1978, Lycian Turkey: An Archaeological Guide: London, BennNorton, 197 p. Bonnaffe, F., Jennette, D., and Andrews, J., 2007, A method for acquiring and processing ground-based lidar data in difficult-to-access outcrops for use in three-dimensional, virtual-reality models: Geosphere, v. 3, no. 6, p. 501–510, doi: 10.1130/GES00104.1. Devereux, B.J., Amabable, G.S., Crow, P., and Cliff, A.D., 2005, The potential of airborne LIDAR for detection of archaeological features under woodland canopies: Antiquity, v. 79, p. 648–660. Erdik, M., Biro, Y.A., Onur, T., Sesetyan, K., and Birgören, G., 1999, Assessment of earthquake hazard in Turkey and neighboring regions: Annali di Geofisica, v. 42, no. 6, p. 1125–1138. Ergin, K., Güçlü, U., and Uz, Z., 1967, A Catalog of Earthquakes for Turkey and Surrounding Area (11 A.D. to 1964 A.D.): Istanbul Technical University, Faculty of Mines, Institute of Physics of the Earth, Technical Report 24, 169 p. Ergin, K., Güçlü, U., and Aksay, G., 1971, A Catalog of Earthquakes of Turkey and Surrounding Area (1965–1970): Istanbul Technical University, Faculty of Mines, Institute of Physics of the Earth, Technical Report 28, 184 p. Eyidoğan, H., and Barka, A., 1996, The 10 October 1995 Dinar earthquake SW Turkey: Terra Nova, v. 8, p. 479–485, doi: 10.1111/j.1365-3121.1996 .tb00773.x. Frischer, B., and Dakouri-Hild, A., 2008, Beyond Illustration: 2D and 3D Digital Technologies as Tools for Discovery in Archaeology (Frischer, B., and Dakouri-Hild, A., eds.): Oxford, British Archaeological Reports International Series 1805, 192 p. Galadini, F., Hinzen, K.-G., and Stiros, S.C., 2006, Archaeoseismology: Methodological issues and procedure: Journal of Seismology, v. 10, p. 395– 414, doi: 10.1007/s10950-006-9027-x. Galli, P., and Galadini, F., 2001, Surface faulting of archaeological relics: A review of case histories from the Dead Sea to the Alps: Tectonophysics, v. 335, p. 291–312, doi: 10.1016/S0040-1951(01)00109-3. Güçlü, U., Altınbaş, G., and Eyidoğan, H., 1986, A Catalog of Earthquakes of Turkey and Surrounding Area (1971–1975): Istanbul Technical University, Faculty of Mines, Institute of Physics of the Earth, Technical Report 30, 191 p.

169

Guidoboni, E., 1994, Catalogue of Ancient Earthquakes in the Mediterranean Area up to the 10th Century: Rome, Istituto Nazionale di Geofisica, 504 p. Guidoboni, E., 2002, Historical seismology: The long memory of the inhabited world, in Lee, W.H.K., Kanamori, H., Jennings, P.C., and Kisslinger, C., eds., International Handbook of Earthquake and Engineering Seismology: Amsterdam, Academic Press, International Geophysics Series 81A, p. 775–790. Hancock, P.L., and Altunel, E., 1997, Faulted archaeological relics at Hierapolis (Pamukkale), Turkey, in Hancock, P.L., and Michetti, A.M., eds., Paleoseismology; Understanding Past Earthquakes Using Quaternary Geology: Journal of Geodynamics, v. 24, p. 21–36. Hinzen, K.-G., 2005, The use of engineering seismological models to interpret archaeoseismological findings in Tolbiacum, Germany: A case study: Bulletin of the Seismological Society of America, v. 95, no. 2, p. 521– 539, doi: 10.1785/0120040068. Hinzen, K.-G., 2009, Simulation of toppling columns in archaeoseismology: Bulletin of the Seismological Society of America, v. 99, p. 2855–2875, doi: 10.1785/0120080241. ISC (International Seismological Centre), 2001, International Seismological Centre On-line Bulletin: http://www.isc.ac.uk/search-index.html. Karabacak, V., Altunel, E., Akyüz, S., and Yönlü, Ö., 2007, Büyük Menderes Fay Zonu üzerinde normal faylanmaya bağlı özelliklerin “Yersel LİDAR” kullanılarak belirlenmesi [Determination of normal fault characteristics with ground-based LIDAR on Büyük Menderes fault zone], in Tüysüz, O., and Erturaç, K., eds., VI. TURQUA-Turkey Quaternary Symposium, ITU Istanbul, 16–18 May 2007 online abstract volume: http://www.eies.itu .edu.tr/TURQUA/bildiri/karabacak_turqua6.pdf. Karabacak, V., Altunel, E., Yönlü, Ö., Akyüz, H.S., Yalçıner, C.Ç., and Altınok, S., 2008, Aktif faylanmaya bağlı fizyografik özelliklerin ölçülmesinde yeni bir yöntem: “Yersel LİDAR” teknolojisi [A new method for the measurement of active faulting related physiographic characteristics: Ground-based LIDAR technology]: Ulusal Jeomorfoloji Sempozyumu (UJS-2008), Onsekiz Mart Üniversitesi, Çanakkale, abstract volume, p. 301–307. Karcz, I., and Kafri, U., 1978, Evaluation of supposed archaeoseismic damage in Israel: Journal of Archaeological Science, v. 5, p. 237–253, doi: 10.1016/ 0305-4403(78)90042-0. Lambers, K., Eisenbeiss, H., Sauerbier, M., Kupferschmidt, D., Gaisecker, T., Sotoodeh, S., and Hanusch, T., 2007, Combining photogrammetry and laser scanning for the recording and modeling of the Late Intermediate Period site of Pinchango Alto, Palpa, Peru: Journal of Archaeological Science, v. 34, p. 1702–1712, doi: 10.1016/j.jas.2006.12.008. Lang, G., 2003, Klassische antike Stätten Anatoliens, Band II: St. Peter am Hart, Lang Verlag, 707 p. Maniatis, G., Kurfess, D., Hampel, A., and Heidbach, O., 2009, Slip acceleration on normal faults due to erosion and sedimentation—Results from a new three-dimensional numerical model coupling tectonics and landscape evolution: Earth and Planetary Science Letters, v. 284, no. 3–4, p. 570– 582, doi: 10.1016/j.epsl.2009.05.024. Medvedev, S., Sponheur, W., and Karnik, V., 1964, Neue Seismische Skala: Deutsche Akademie der Wissenschaften zu Berlin, Heft, v. 77, Akademie Verlag, p. 69–76. Niemi, T.M., Kayen, R., Zhang, H., Dunn, C.R., and Doolin, D.M., 2004, LiDAR imagery of the San Andreas fault zone at the Vedanta and Olema Ridge paleoseismic trench sites, Pt. Reyes, CA: Eos (Transactions, American Geophysical Union), v. 85, no. 47, Fall Meeting Supplement, abstract G13B-0811. Nur, A., and Ron, H., 1996, And the walls came tumbling down: Earthquake history in the Holy Land, in Stiros, S.C., and Jones, R.E., eds., Archaeoseismology: British School at Athens, Fitch Laboratory Occasional Paper 7, p. 75–85. Optech, 2006, ILRIS-3D Intelligent Laser Ranging and Imaging System: http:// www.optech.ca/i3dprodline-ilris3d.htm (accessed July 2009). Pınar, N., and Lahn, E., 1952, Turkish Earthquake Catalog with Descriptions: Turkey Ministry of Public Works and Settlement, The General Directorate of Construction Affairs, Serial 6, Technical Report 36, 153 p. Şenel, M., 1997a, Geological Maps of Turkey in 1:250,000 scale: Fethiye Sheet: Ankara, Mineral Research and Exploration Directorate of Turkey, 26 p. Şenel, M., 1997b, Geological Maps of Turkey in 1:100,000 scale: Fethiye L8 Sheet: Ankara, Mineral Research and Exploration Directorate of Turkey, 22 p.

170

Yerli et al.

Şenel, M., 1997c, Geological Maps of Turkey in 1:100,000 scale: Fethiye M8 Sheet: Ankara, Mineral Research and Exploration Directorate of Turkey, 15 p. Similox-Tohon, D., Sintubin, M., Muchez, P., Verhaert, G., Vannesta, K., Fernandez, M., Vandycke, S., Vanhaverbeke, H., and Waelkens, M., 2006, The identification of an active fault by a multidisciplinary study at the archaeological site of Sagalassos (SW Turkey): Tectonophysics, v. 420, p. 371–387, doi: 10.1016/j.tecto.2006.03.026. Sintubin, M., Muchez, P., Similox-Tohon, D., Verhaert, G., Paulissen, E., and Waelkens, M., 2003, Seismic catastrophes at the ancient city of Sagalassos (SW Turkey) and their implications for the seismotectonics in the Burdur–Isparta area: Geological Journal, v. 38, p. 359–374, doi: 10.1002/ gj.960. Soysal, H., Sipahioğlu, S., Kolçak, D., and Altınok, Y., 1981, Historical Earthquake Catalogue of Turkey and Surrounding Area (2100 B.C.–1900 A.D.): TÜBİTAK (Türkiye Bilimsel ve Teknolojik Araştırma Merkezi [Scientific and Technological Research Council of Turkey]) Technical Report TBAG-341. Stiros, S.C., 1996, Identification of earthquakes from archaeological data: Methodology, criteria and limitations, in Stiros, S.C., and Jones, R.E., eds., Archaeoseismology: British School at Athens, Fitch Laboratory Occasional Paper 7, p. 129–152. Tan, O., Tapirdamaz, M.C., and Yörük, A., 2008, The earthquake catalogues for Turkey: Turkish Journal of Earth Sciences, v. 17, p. 405–418. ten Veen, J.H., 2004, Extension of Hellenic forearc shear zones in SW Turkey: The Pliocene–Quaternary deformation of the Eşen Çay Basin: Journal of Geodynamics, v. 37, p. 181–204, doi: 10.1016/j.jog.2004.02.001. ten Veen, J.H., and Kleinspehn, K.L., 2002, Geodynamics along an increasingly curved convergent plate margin: Late Miocene–Pleistocene Rhodes (Greece): Tectonics, v. 21, doi: 10.1029/2001TC001287.

ten Veen, J.H., and Kleinspehn, K.L., 2003, Incipient continental collision and plate-boundary curvature: Late Pliocene–Holocene transtensional Hellenic forearc, Crete, Greece: Journal of the Geological Society of London, v. 160, p. 161–181, doi: 10.1144/0016-764902-067. ten Veen, J.H., Woodside, J.M., Zitter, T.A.C., Dumont, J.F., Mascle, J., and Volkonskaia, A., 2004, Neotectonic evolution of the Anaximander Mountains at the junction of the Hellenic and Cyprus arcs: Tectonophysics, v. 391, p. 35–65, doi: 10.1016/j.tecto.2004.07.007. ten Veen, J.H., Boulton, S., and Alçiçek, M.C., 2009, From palaeotectonics to neotectonics in the Neotethys realm: The importance of kinematic decoupling and inherited structural grain in SW Anatolia (Turkey): Tectonophysics, v. 473, p. 261–281, doi: 10.1016/j.tecto.2008.09.030. Wells, D.L., and Coppersmith, K.J., 1994, New empirical relationships among magnitude, rupture length, rupture width, rupture area, and surface displacement: Bulletin of the Seismological Society of America, v. 84, no. 4, p. 974–1002. Westaway, R., 1994, Present-day kinematics of the Middle and Eastern Mediterranean: Journal of Geophysical Research, v. 99, p. 20,203–20,223, doi: 10.1029/94JB01054. White, N.J., Jackson, J.A., and McKenzie, D.P., 1986, The relationship between the geometry of normal faults and that of the sedimentary layers in their hanging walls: Journal of Structural Geology, v. 8, p. 897–909, doi: 10 .1016/0191-8141(86)90035-0. Wood, H.O., and Neumann, F., 1931, Modified Mercalli intensity scale of 1931: Bulletin of the Seismological Society of America, v. 21, p. 277–283. Wurster, W.W., and Wörrle, M., 1978, Die Stadt Pınara: Archäologischer Anzeiger, v. 1978-1, p. 74–101. MANUSCRIPT ACCEPTED BY THE SOCIETY 19 MAY 2010

Printed in the USA

The Geological Society of America Special Paper 471 2010

Ancient earthquakes from archaeoseismic evidence during the Visigothic and Islamic periods in the archaeological site of “Tolmo de Minateda” (SE Spain) M.A. Rodríguez-Pascua* Área de Riesgos Geológicos, Instituto Geológico y Minero de España, Ríos Rosas, 23, 28003-Madrid, Spain P.G. Silva* Departamento de Geología, Universidad Salamanca, Escuela Politécnica Superior de Ávila, 05003-Ávila, Spain V.H. Garduño-Monroy* Departamento de Geología y Mineralogía, Universidad Michoacana, Morelia, Michoacán, 58060, México R. Pérez-López* Área de Riesgos Geológicos, Instituto Geológico y Minero de España, Ríos Rosas, 23, 28003-Madrid, Spain I. Israde-Alcántara* Departamento de Geología y Mineralogía, Universidad Michoacana, Morelia, Michoacán, 58060, México J.L. Giner-Robles* Departamento de Geología, Facultad de Ciencias, Universidad Autónoma de Madrid, Cantoblanco, Tres Cantos, Madrid, Spain J.L. Bischoff* Laboratory of Geochronology, U.S. Geological Survey, 345 Middlefield Road, MS 211, Menlo Park, California 94025, USA J.P. Calvo* Área de Riesgos Geológicos, Instituto Geológico y Minero de España, Ríos Rosas, 23, 28003-Madrid, Spain

ABSTRACT The ruins of the ancient settlement of “El Tolmo de Minateda” are one of the best representative archaeological sites within the Albacete Province (SE Spain), characterized by a well-preserved record for the last ~3800 yr. The present ruins record an almost continuous of occupation from the Late Bronze Age (Iberian Culture, from ca. third century B.C.) to the High Middle Ages, including intervening and successive Roman, Visigoth, and Muslim city remains. The eventual Muslim settlement was

*E-mails: Rodríguez-Pascua—[email protected]; Silva—[email protected]; Garduño-Monroy—[email protected]; Pérez-López—r.perez@ igme.es; Israde-Alcántara—[email protected]; Robles—[email protected]; Bischoff—[email protected]; Calvo—[email protected]. Rodríguez-Pascua, M.A., Silva, P.G., Garduño-Monroy, V.H., Pérez-López, R., Israde-Alcántara, I., Giner-Robles, J.L., Bischoff, J.L., and Calvo, J.P., 2010, Ancient earthquakes from archaeoseismic evidence during the Visigothic and Islamic periods in the archaeological site of “Tolmo de Minateda” (SE Spain), in Sintubin, M., Stewart, I.S., Niemi, T.M., and Altunel, E., eds., Ancient Earthquakes: Geological Society of America Special Paper 471, p. 171–184, doi: 10.1130/2010.2471(15). For permission to copy, contact [email protected]. © 2010 The Geological Society of America. All rights reserved.

171

172

Rodríguez-Pascua et al. abruptly abandoned and destroyed during the ninth–tenth centuries A.D., leaving a lack of any archaeological evidence of war or decay. Another previous anomalous archaeological episode of abrupt city abandonment and destruction is recorded during the Visigothic Period (seventh century). The archaeological record of this city supports evidence for earthquake damage linked to both periods of city abandonment and destruction, including oriented collapse of walls, watchtowers, and columns, oriented cracking of walls and column drums, as well as in situ broken pottery, abrupt abandonment of kilns, and anomalous sedimentary infilling of canals and water-supply facilities. Additionally, large-scale rockfalls containing Visigothic carved tombs are also apparently associated with both episodes, constituting one of the few instances of combined geoarchaeological evidence of earthquake ground effects ever reported. As a means of testing the theoretical archaeoseismic potential of this site, we obtained an archaeoseismic quality factor (AQF) value of 0.074.

INTRODUCTION Archaeoseismology represents a new branch of science combining archaeology and paleoseismology. The aim of this discipline is to find past earthquakes that affected ancient villages with no historic reports of seismicity. These earthquakes are important in terms of earthquake hazard assessment, allowing us to complete the earthquake catalogues linking paleoseismic and historical temporal records. However, this discipline is controversial due to the absence of a rigorous methodology (Galadini et al., 2006; Sintubin and Stewart, 2008), with few well-documented works and the necessity of multidisciplinary teams of workers, including geologists, archaeologists, architects, etc. Most of the well-documented studies have been carried out in the Mediterranean area (Stiros and Jones, 1996; Guidoboni et al., 2000; Meghraoui et al., 2003; Silva et al., 2005; Similox-Tohon et al., 2006; among others). In this context, the Iberian Peninsula presents multiple archaeological sites, although only scarce archaeoseismic analyses have been performed in this sense (e.g., Silva et al., 2005, 2009). Force (2008) noted the relevance of plate tectonics in the evolution and settlement of ancient civilizations by introducing a relationship between the closest distances of these civilizations to a major tectonic plate boundary and the temporal evolution of the society. From a tectonic point of view, the Iberian Peninsula consists of a tectonic microplate with small to moderate instrumental earthquake activity, mainly concentrated at the southeastern part of Spain (Herraiz et al., 2000; Stich et al., 2003). The archaeological site of “Tolmo de Minateda” (Albacete Province, hereafter referenced as El Tolmo) is located within an intraplate stable area, although several major active faults with paleoseismic evidence have been reported (Rodríguez-Pascua et al., 2008) (Fig. 1). Despite the relevance of this archaeological site, there are no historical references of earthquakes affecting this village; however, the well-documented excavations developed at El Tolmo, and the long-term occupation (3800 yr), configures this site as being of potential archaeoseismic interest. The work presented here addresses several lines of archaeoseismic evidence, recognized from the collapses of ancient buildings, rockfalls, pottery studies,

and archaeological reports of El Tolmo, relative to paleoseismic studies carried out in the nearby area. Moreover, we applied the logical tree method proposed by Sintubin and Stewart (2008) for archaeoseismology, with the aim of assigning a qualitative value of this site in comparison with other archaeoseismic studies and potential sites. GEOGRAPHIC AND HISTORIC CONTEXT OF “EL TOLMO” El Tolmo is located between the villages of Cordovilla and Agramon, in the southernmost area of the Albacete Province (SE of Spain) (Fig. 1). This archaeological site is close to the Betic Cordilleras frontal thrust, and adjacent to major strike-slip faults such as the Pozohondo, Lietor, and Socovos-Calasparra faults (Fig. 1B). These structures are active faults oriented NW-SE, affecting Quaternary deposits and with a trace longitude of ~90 km (Rodríguez-Pascua et al., 2003). The El Tolmo site is geographically located on a strategic natural route between the inner Iberian Meseta and the Mediterranean Sea throughout the Betic Cordilleras, along the ancient Roman way linking Complutum (Alcalá de Henares) and Cartago Nova (Cartagena). The landscape of this area shows flat terrain with isolated relict structural buttes carved on late Neogene marine sandstones by differential erosion (García del Cura et al., 1979). The El Tolmo site is located on the top of one of these buttes (Fig. 2), bounded by near vertical cliffs of around 20 m height. The only access to the top of the butte is a narrow bedrock creek carved in the NW slope of the butte, locally known as “El Reguerón” (Fig. 2, point 1). At this point, the different ancient cultures that settled at El Tolmo built strong defensive city walls, closing the narrow bedrock valley outlet. From a geomorphological point of view, this particular structural butte has an elongated shape striking NW-SE and shows multiple evidence of large rockfalls along the main cliffline (Fig. 2, points 2 and 5). The archaeological site of El Tolmo is one of the most important of the Albacete Province, and currently is classified as an Archaeological Protected Park. The ruins have been known since the early twentieth century, but the first systematic excavations

Figure 1. (A) Geographic location of the study area. (B) Detailed structural map of the archaeological site of the ancient city of “Tolmo de Minateda” (dashed square) (Albacete Province, Spain). Pozohondo, Lietor, and Socovos-Calasparra faults represent major strike-slip faults in the area.

Figure 2. Location of the archaeoseismic evidence in the city during the Visigothic and Islamic periods. Visigothic Period: (1) “El Reguerón” zone is the defensive access to the city, the area of the collapsed watchtower and the defensive wall. (2) Rock mass movements and large blocks with anthropomorphic tombs, displaced from the upper northern part of the fortification. (3) Location of the oriented toppled columns (N-S trending) and cracks (dipping 45°) in the Visigothic basilica; Islamic Period. (4) Destroyed and abandoned furnace kiln and in situ pottery completely crushed by collapsed walls. (5) Recent rock mass movement with anthropomorphic tombs on large blocks, displaced from the upper southern part of the fortification. Solid black arrows show sense and orientation of main rockfall events (1 and 5). White arrows show sense and orientation of collapse of defensive walls (1) and columns, and building walls at the monumental zone of the Visigothic city (3 and 4).

174

Rodríguez-Pascua et al.

at this site started in 1987 (Abad-Casal et al., 1998). Successive archaeological excavations revealed that El Tolmo records nearcontinuous occupation since the Late Bronze Age (fifth century B.C.) to the Spanish Muslim Period (tenth century A.D.) (AbadCasal et al., 1998, 2000a, 2000b). Neolithic rock paintings point to an occupation since at least 3800 yr B.P. During this period, different civilizations successively settled at this archaeological site. Iberian, Roman, Byzantine, Visigothic, and Islamic cultures left a well-preserved archaeological record, as well as numerous references in ancient chronicles. The primitive Iberian settlement was occupied by the Romans, founding the city of Ilunum (first century A.C.–fifth century A.D.), later renamed as Eio or Elotana by the Byzantines and Visigoths (fifth–eighth century A.D.). The last historic and archaeological record corresponds to the Muslim town of Madinat Iyih (Gutiérrez-Lloret, 2000), during the Islamic Period of the Iberian Peninsula from the eighth century to the late tenth century A.D. (Abad-Casal et al., 1998, 2000a, 2000b). The main archaeological remnant consists of a welldeveloped ancient town with Muslim districts and multiple remains of pottery and utensils over an area of ~520 m2. The Muslim districts were built over earlier ancient monumental zones located at the southern upper part of the site, showing a Byzantine basilica (A.D. 589–610), different public buildings and anthropomorphic tombs (graves) of the Visigothic Period carved in the solid sandstone substratum of the city. The lower part of the city records a strong, reinforced defensive city wall and watchtowers closing the only access to the city in the El Reguerón creek, in the northeastern slope of the butte. It is important to highlight that the transitions between the different historical periods of the site show no apparent relation with any known war episodes or disease epidemics (Abad-Casal et al., 1998); although two sudden and unexplained site destructions and desertions can be drawn from the archaeological record. The first one occurred during the Visigothic Period (early seventh century), and the second during the Islamic Period (late ninth to early tenth centuries) according to Abad-Casal et al. (1998) and Gutiérrez-Lloret (2000). ARCHAEOSEISMIC EVIDENCE OF “EL TOLMO” First Desertion of “El Tolmo” (Byzantine–Visigothic Periods, Seventh Century A.D.) During the Iberian Period (Late Bronze Age), the first defensive wall was built at the only access to the city (Fig. 2, point 1), dated about the third century B.C. (Abad-Casal et al., 1998). Later, during the Roman occupation, the Iberian Wall was reinforced and ornamented with stronger, pillow-shaped masonry of local calcarenite, including an imperial inscription referent to the Emperor Augustus dated in the year A.D. 9 (first century A.D.). This fact indicates the relevance of this site, during this cultural period, in becoming a true Roman city (municipium civium Romanorum; Abad-Casal et al., 1998). These authors point out the continuous occupation of the site from the Iberian Period (A.D. 550) until the early seventh century A.D. (the Visigoth town of Elo).

The first documented episodic abandonment occurred during the Roman Period, dated between the first and fourth centuries A.D. (Abad-Casal et al., 1998). Evidence of abandonment includes the partial detrital infilling of the freshwater supply and irrigation canals. Nevertheless, neither buildings nor defensive walls record coeval damage during this episode. This fact has been interpreted as either a change of the administration policy in the town, or a lack of the archaeological record (Abad-Casal et al., 1998), although a paleoseismic event cannot be discarded due to the presence of Holocene active faults (Rodríguez-Pascua et al., 2009). After this episode, a small Byzantine basilica was built in the upper part of the city (ca. A.D. 589–610), ornamented with plain marble columns and a large baptismal stone and becoming the site of one of the main Episcopal centers of SE Spain during the Visigothic Period. During the middle Visigothic Period, there is a lack of any archaeological record that is associated with an unexplained abandonment of El Tolmo, dated by Abad-Casal et al. (1998) around the years A.D. 600–610. Associated with this abandonment, the two watchtowers flanking the city gate at the El Reguerón defensive wall, as well as part of the city wall itself, collapsed (Figs. 3A and 3B). The watchtowers were directly founded on the solid Miocene sandstone, as the main way through the city gate, which displays long-lived, continuous, and deep wheel-tracks carved on the calcarenitic substratum (Fig. 3). Previous archaeological interpretations (Abad-Casal et al., 1998) suggested a collapse of the masonry defensive set due to a failure of the basement. However, the watchtowers and gateway zone show no apparent deep weathering and/or intense cracking. In addition the collapse event displays a noticeable directionality toward the ESE, diagonal to the main E-W topographic gradient of the El Reguerón creek (Fig. 3). The collapse of this area blocked the gateway, and the calcarenite blocks and detritus fossilized the wheel-tracks, providing a relative dating for this destruction event. All this collapsed area was covered by a thick (2–3 m) layer of detritus and stone rubble before the first systematic archaeological excavations in 1987 (Abad-Casal et al., 1998). The basilica was also damaged during this same period, showing toppled columns oriented with an approximate N-S strike and dipping toward the north (Fig. 4). Moreover, the masonry collapse associated with the basilica destruction shows the same orientation. Altunel et al. (2003) suggested that the column orientation is parallel to the fault rupture propagation that could trigger an earthquake in Cnidus. In this respect, Stiros and Jones (1996) and Silva et al. (2005) have reported seismic causes for oriented fallen columns at other ancient sites as well. Silva et al. (2005, 2009) interpreted the orientation of several toppled column drums at Baelo Claudia (Roman Period, south of Spain), pavement breakouts, pop-up arrays, and perpendicular pavement folding, as coincident with the seismic shock direction. Other archaeoseismic evidence connected with this episode is the occurrence of cracks dipping 45° toward the north affecting several columns and walls of the basilica (Fig. 5). These cracks affected the external masonry of walls and ornamented blocks

Ancient earthquakes from archaeoseismic evidence during the Visigothic and Islamic periods, SE Spain

175

Figure 3. Present view of “El Reguerón” (see Fig. 2 for location of point 1). (A) Roman defensive wall collapse detected during the archaeological excavations. Idealized defensive wall reconstruction is inspired after the work by Abad-Casal et al. (1998). (B) Oriented masonry collapse of the defensive wall at the entrance. Bricks are 1 m long.

of the basilica altar, in a similar way to those reported by Stiros and Jones (1996) as seismic origin, and also modeled by Mistler et al. (2006). This type of crack is related to shear movement of the masonry blocks from seismic source (Stiros, 1996; Nur and Burgess, 2008). Furthermore, several large-scale rockfalls (~3800 m3 from field measurement) can be observed, affecting and dislocating anthropomorphic Visigothic tombs at the northern cliff of the butte opposite to the El Reguerón defensive set (Figs. 2 and 6). These dislocated blocks are partially covered and embedded by colluvium in a similar way to that which occurred at the El Reguerón collapsed area. On the contrary, other block falls around the butte cliffs are free of colluvium, indicating a more recent activity, as we explain in the following. This fact indicates the relative age of these block-fall areas, and the preliminary assumption is that falls at the northern cliff of the butte may be coeval with the collapses of the El Reguerón zone and the basilica (Fig. 6). However, absolute (e.g., lichenometry) dating has to be performed to establish the chronology of these landslides. Ongoing lichenometric analysis (sidereal dating) of the free faces of landslided blocks will provide relevant data for future interpretations (age interval between 0 and 1000 yr). Second Abandonment of El Tolmo (Islamic Period, Ninth– Tenth Centuries A.D.) The Islamic Period in El Tolmo begins in the early eighth century, and the town was renamed as Madinat Iyih (GutiérrezLloret, 2000). This settlement carried out the rebuilding and

reinforcement of the city wall and gateway at the El Reguerón site by using several layers of mortar concrete (Fig. 2, point 1). This artificial cover fossilized and masked the former evidence of collapse occurred during the Visigoth Period (seventh century) (Fig. 7A). A ceramic industrial district was established on the ruins of the destroyed basilica, with well-documented buildings, works, and ceramic kilns (Abad-Casal et al., 1998; Fig. 7B). During this period, the eventual abandonment of the city is inferred from the lack of any archaeological record between the ninth and tenth centuries A.D. Moreover, multiple well-preserved pottery artifacts were destroyed by wall collapses in the interior of the buildings, some of them located within kilns ready to be fired (Fig. 8), which suggests a destructive episode of the whole ceramic district during this period. The water reservoirs were also filled up by sediments, and the freshwater supply and irrigation canals around the site were destroyed during this period. We have interpreted this fact as traces of abandonment of these facilities. Coeval to this abandonment episode, there are reports about large rockfalls located at the upper southeastern part of the city, destroying Visigothic anthropomorphic tombs excavated in the rock and dislocated downslope from their original position (Fig. 9). Rockfalls at the southern slope of the butte are more recent than those preserved in the northern slope connected with the Visigothic episode of destruction. The toppled sandstone blocks appear weakly weathered and uncovered by colluvium. Scars on the cliff are fresh scarps, with nearly vertical free faces displaying a very poor lichen evolution. Individual mobilized blocks can reach volumes of ~3816 m3 (one large block of 16 × 11.7 × 12 m3 and 18 minor blocks), and the total mobilized

176

Rodríguez-Pascua et al.

Figure 4. Visigothic basilica located at the top of the city of “El Tolmo.” (A) Oriented fallen columns from their original position (after Abad-Casal et al., 1998). (B) Present modern reconstruction of the original colonnade and portico of the basilica and anthropomorphic Visigothic tombs excavated on the Miocene sandstone. See point 3 in Figure 2 for location. Columns are 1.80 m long.

material at the northern slope comprises ~500 m3. In order to illustrate the rockfall susceptibility of the butte cliffs, it can be said that some historical reports seem to indicate that landslide scars were presumably reactivated as far-field effects from the well-known Lisbon earthquake (A.D. 1755). An historic chronicle from the close village of Agramon (Fig. 1B for location) literally reported: “de una montaña se desprendió mucha parte” (a large part of a mountain collapsed; Rodríguez de la Torre, 1995),

and there are many more historic reports for the 1755 event mentioning similar gravitational processes in this zone. The EMS (environmental macroseismic scale) intensity suggested for this area is VI (Martínez-Solares, 2001). Without rejecting the occurrence of rockfall reactivation during more recent strong events, these falls could be correlated with the Islamic Period destruction event discussed here. Absolute dating is also needed here to clarify the timing of rockfall.

Ancient earthquakes from archaeoseismic evidence during the Visigothic and Islamic periods, SE Spain

177

Figure 5. Cracks dipping 45° toward the north (A–B) and south (A) affecting the reconstructed basilica pillars. (A) Ornamented limestone of the principal altar and (B) masonry of the outer wall of the basilica. See point 3 in Figure 2 for location.

Figure 6. Rockfall affecting anthropomorphic tombs, assumed to have occurred during the Byzantine-Visigothic periods, showing rotated and sliding tombs. See point 2 on Figure 2 for location.

SUSPECTED SEISMOGENIC FAULTS AND GROUND EFFECTS Regarding the seismic source that could trigger earthquakes, in relation to the archaeoseismic damage mentioned in this study, Rodríguez-Pascua et al. (2008) recently reported the occurrence of active faults with paleoseismic evidence less than 13 km away

from El Tolmo. These authors established earthquake-related fault activity for the major NW-SE strike-slip fault of Pozohondo (for location, see Fig. 1). In detail, the active Tobarra-Cordovilla segment (15 km length) of this fault (Fig. 10) shows a complex graben basin deforming Quaternary lacustrine deposits and displaying well-preserved coseismic fault scarps and large cracks of metric scale cutting recent soils. Furthermore, the landscape

178

Rodríguez-Pascua et al.

Figure 8. Detailed photography of in situ crushed pottery and the toppled block of the wall that destroyed the piece.

Figure 7. Archaeological evidence of the building reconstructions during the Islamic Period (Madinat Iyih) on the ruins of the ancient Visigothic basilica (Elo). (A) Covering of the defensive wall and watchtower collapses (white arrows) (after Abad-Casal et al., 1998). See point 1 in Figure 2 for location. (B) Detailed photography of a ceramic oven from the Islamic Epoch. See point 4 in Figure 2 for location. Bricks in columns are 1 m long.

along this fault segment is controlled by active faulting, with the occurrence of a dammed lake caused by the obstruction of the drainage by late Pleistocene to Holocene surface ruptures and fault scarps (Figs. 10B and 10C). Ongoing radiocarbon dating of

paleoseismic trenches open in this fault segment will hopefully throw light on this preliminary relationship. Earthquake ground effects connected with this fault segment allow us to estimate a magnitude of Mw 6–7, and a maximum ESI (environmental seismic intensity) 2007 intensity (Michetti et al., 2007) of IX–X for the fault scarp zone (Rodríguez-Pascua et al., 2008; Silva et al., 2008). This earthquake has enough magnitude to trigger significant rockfall processes in the near-vertical cliffs of the studied structural butte located 13 km away. In fact, considering the mobilized volume of material (~3800 m3), an ESI 2007 intensity of at least VII–VIII should be associated with the presumed ground movement at the El Tolmo zone. ARCHAEOLOGICAL QUALITY FACTOR (AQF) FOR “EL TOLMO” SITE Sintubin and Stewart (2008) proposed a methodology based on the logic tree scheme similar to that proposed by Atakan et al. (2000) for paleoseismicity. This method defines the archaeological quality factor (AQF), with the purpose of decreasing the potential uncertainties from archaeoseismology. Actually, the problem of the uncertainty in paleoseismicity and archaeoseismicity represents the key value for establishing paleoearthquakes

Figure 9. Large rockfalls with anthropomorphic tombs on the top, located at point 5 in Figure 2. The slide free face appears relatively unaltered and poorly colonized by lichens. See text for further explanation.

Figure 10. (A) Trench excavated along the potential seismic normal fault segment assigned to the paleoearthquake interpreted in this work. This active segment belongs to Pozohondo fault in a transtensive tectonic regime, between the towns of Tobarra and Cordovilla (see Fig. 1B for location). (B–C) Accumulated surface ruptures cutting lacustrine Quaternary sediments.

180

Rodríguez-Pascua et al.

to be included in the seismic time record within a zone. However, criteria for geological data have been more rigorous for the interpretation of earthquake-related data than evidence coming from the archaeological record. Among other considerations, this is due to either the poor quality of buildings (concrete and masonry), or warfare or fire, which could better explain the parts of the archaeological record being interpreted as earthquake damage. In this sense, the method proposed by Sintubin and Stewart (2008) introduces a quantitative quality value (AQF) assigned to the archaeoseismic evidence by taking the tectonics of the area into account, the geomorphic evidence of faulted landscapes, and type and scope of the archaeological study amongst others. With the aim of establishing the quality (AQF) of the archaeoseismic evidence interpreted from El Tolmo, we have applied this methodology, following the scheme proposed by Sintubin and Stewart (2008), and by assuming a different QWF value (quality weight factor) in six different stages as a probability value ranging from 0 to 1. Figure 11 summarizes the QWF values interpreted from El Tolmo for each stage. These stages are as follow: Stage 1. Tectonic Setting El Tolmo is located within an active plate interior, adjacent to the frontal thrust of the Betic Cordillera, and close to the Pozohondo fault (Fig. 1). This is a major NW-SE strike-slip fault with associated paleoearthquakes of estimated magnitude of 6 < M < 7, and calculated slip rate of 0.09 mm/yr (Rodríguez-Pascua et al., 2007, 2008). We propose a QWF = 0.7.

preted as a feature in this area (13 km around the site) range from intensity IX to X in the ESI 2007 scale (Michetti et al., 2007). In addition, site destruction events are presumably associated with seismically triggered rockfalls of the archaeological remains. Accordingly, we assume QWF = 0.6. Stage 3. Site Potential This stage is defined by the quality, extension, and construction of the archaeological site. In this sense, we have shown previously that El Tolmo town was a well-constructed defensive city with strong masonry, well-founded defensive walls, and highquality concrete (Figs. 2, 3, and 4). Scarce indicators of ground instability, solid bedrock, long-term occupation by different cultures, and lack of strong anthropogenic disturbance suggest a high-quality QWF value of 0.7. Stage 4. Identification of Earthquake-Related Damage The signatures from the earthquake damage in El Tolmo are large rockfalls, oriented toppled columns, building cracks, and unexplained abrupt abandonment of the city, mainly recorded by detrital infilling within the stratigraphy of ancient canals and dams, as well as discontinuous horizons of pottery industry abandonment and/or in situ destruction of handcrafted ceramics, suggesting a good QWF value of 0.7 (Figs. 3, 4, 5, 7, and 8). Oriented cracks and defensive wall collapses support earthquake damage, rather than other destructive events.

Stage 2. Site Environment

Stage 5. Dating Earthquake-Related Damage and/or Paleoearthquakes

The paleoseismic record associated with the archaeoseismic evidence exhibits surface ruptures (Fig. 7), active segments, and fault scarps affecting recent soils with 1.6 m of accumulated vertical throw. Moreover, the landscape is characterized by fault activity forming endorheic basins, dammed lakes, and variations of the fluvial network. Furthermore, the paleoearthquakes inter-

Dating of the destructive events recorded at this site is purely archaeological and mainly based in classical pottery sequences and architectural elements. Geomorphic dating of close recent fault scarps coincident with the last city destruction event indirectly support archaeological evidence (Pérez-López et al., 2007; Rodríguez-Pascua et al., 2008). No written historical reports

Figure 11. Quality weight factor (QWF) values obtained for the ancient archaeological site of “Tolmo de Minateda,” and archaeoseismic quality factor (AQF) estimated by following the logical-tree method proposed by Sintubin and Stewart (2008). See text for further explanation. SCL—site confidence level.

Ancient earthquakes from archaeoseismic evidence during the Visigothic and Islamic periods, SE Spain supporting earthquake occurrence in the zone are available. Consequently, this stage throws poor, limited evidence, considering a QWF value of 0.3. Stage 6. Regional Correlation of Archaeoseismological Evidence This stage regards the spatial correlation of the paleoearthquake with other sites nearby. Nevertheless, El Tolmo is an isolated place with a lack of other archaeological sites within a radius of ~50 km. The assigned QWF value is 0.3. AQF Value for El Tolmo The site correction term (SCL) for this site is 4. Multiple seasons of excavations have been carried out at El Tolmo, although no one has been devoted to investigate earthquake-related damage. There are good archaeological reports and publications of the site, but there is no postfactum archaeoseismological reinterpretation before the present study. Therefore, the AQF value is defined as: AQF = QWFi × C,

(1)

where QWFi is for each stage (i from 1 to 6), and C relates to the SCL value (Sintubin and Stewart, 2008). Applying the Equation 1, the AQF value for El Tolmo is 0.074 (Fig. 11). This value is very dependent on the poor quality of QWF values of the last two stages of the methodology proposed by Sintubin and Stewart (2008) related to dating (stage 5) and regional correlation (stage 6). Additionally, the level of importance of the correction term (SCL) is a value designed to measure the seismic perspectives on archaeological excavations (Sintubin and Stewart, 2008). Therefore, new archaeoseismic evidence or findings, even strongly supported by new archaeological interpretations and/or paleoseismic data, will always lack of a long-term support of specific research, resulting in low AQF values. This fact, together with the preliminary stage of dating and the isolated character of the studied site during ancient times, makes the obtained AQF value here very low, but this mainly reflects the preliminary character of the archaeoseismic research in this area of the Betic Cordilleras. Otherwise, the tectonic setting (stage 1), site environment (stage 2), site potential (stage 3), and the earthquake damage evidence (stage 4) combine for a reasonably high scoring in their respective QWF values, between 0.6 and 0.7. DISCUSSION AND CONCLUSIONS The archaeological site of Tolmo de Minateda represents a well-preserved cultural witness with nearly continuous successive settlements during the Iberian, Roman, Byzantine, Visigothic, and Islamic periods. This long-lived occupation for at least the last 3800 yr is mainly related to the strategic defensive nature of this site, which is located in the natural main route

181

connecting the inner Iberian Meseta with the Mediterranean Sea passing the Betic Cordilleras. The site is located on the top of an isolated and elongated structural butte surrounded by nearvertical cliffs of ~20 m height. The only access to the site is a narrow rocky creek closed by a defensive wall at the toe of the butte (El Reguerón), featuring the butte itself as a natural fortification. According to the inaccessibility of the site, no evidence of war or fire damage has been reported during the transitions between the different cultures settled at El Tolmo (Abad-Casal et al., 1998). On the contrary, two as-yet unexplained episodes of abrupt abandonment of the city are witnessed in the archaeological record. Both episodes are related with severe damage in monumental and defensive buildings and occurred presumably coeval to large block-fall events (Fig. 12). This anomalous record is interpreted as archaeoseismic evidence of two different historic earthquakes: 1. The first paleoearthquake occurred during the seventh century A.D. (Visigothic Period) and displays evidence from the N-S–oriented collapse of the basilica (columns and walls), NNESSW–oriented collapse of the defensive wall and watchtowers at the El Reguerón site, covered by colluvium before the excavations started in 1987, and the destruction and partial detrital filling of freshwater canals and facilities (Abad-Casal et al., 1998). 2. The second paleoearthquake can be carefully dated between ninth and tenth centuries A.D. (Islamic period) and displays evidence from the entire collapse of the ceramic district, placed on the ancient ruins of the Byzantine basilica, the abandonment of pottery and freshwater supply facilities. As already mentioned, both episodes are presumably related to large ancient rockfalls toppling anthropomorphic Visigothic tombs around the cliffs of El Tolmo butte. Individual blocks related to these landsliding events are larger than 1386 m3 and are accumulated in two different episodes, as evidenced by the geomorphologic features of the associated colluvial talus: (1) occurrence of colluvial matrix embedding the fallen rock blocks, and (2) lichen growth of rock scars and free faces of blocks and scarps. Rockfalls located at the northern cliff of the El Tolmo butte, opposite to the collapsed defensive wall of El Reguerón location are presumably coeval to the first episode of destruction, which is also fossilized by a colluvial wedge. On the contrary, larger and more recent block slides wasted at the southern cliff zone of the El Tolmo butte could be related to the second episode of destruction, although the far-field effect of the 1755 great Lisbon earthquake is not discarded, since no absolute dating of that rockfall has been performed at present. These blocks are not embedded by colluvial material and display very weak lichen growth. Despite this archaeoseismic evidence, historical seismicity in this zone is really scarce, with maximum reported intensities of IV–V MSK (Martínez-Solares, 2001). Curiously, a unique historical event reported in this zone corresponds to a IV MSK event that occurred in the locality of Minateda, adjacent to “El Tolmo” in the year A.D. 1899 (Rodríguez de la Torre, 1995; Martínez-Solares and Mezcua, 2002). On the other hand, another remarkable record is an indirect report from the nearby locality of

182

Rodríguez-Pascua et al.

Figure 12. Chronological archaeoseismic scheme proposed in this work for the ancient city of “Tolmo de Minateda.”

Hellín, indicating the occurrence of large block-sliding event in the studied area as a far-field effect of the 1755 Lisbon earthquake (Martínez-Solares, 2001). This fact may support the relatively high landslide susceptibility triggered by seismic shaking at cliffs of the El Tolmo butte. Recent seismic events SE of the studied zone (Murcia region) demonstrated that V+ EMS to VI MSK intensities are sufficient to trigger multiple rockfalls of large boulders (10 to 25 m3) along rocky escarpments developed on similar isolated structural buttes around the epicentral area under horizontal ground accelerations of ~0.024g (Murphy-Corella, 2005; Benito et al., 2007; Silva et al., 2008). Individual falling blocks identified in this study are larger than 1386 m3, matching with ESI 2007 site intensities ≥VII (Michetti et al., 2007). In this sense, the studied zone is adjacent to the main NE-SW frontal thrust zone of the Betic Cordillera, and NW-SE strike-slip faults less than 13 km away from the site record recent paleoseismic activity during historic times, such as the Pozohondo fault. The central segment of this fault (Tobarra-Cordovilla) is located only 13 km NW of the studied site and is capable of producing

paleoearthquakes with an estimated maximum magnitude in the range 6 < M < 7 (Rodríguez-Pascua et al., 2008). This active segment modifies the present landscape and the fluvial network with the presence of dammed lakes, surface faulting, and large ground crack ruptures affecting late Quaternary lacustrine tufas and soils. Morphometric dating of the related fresher fault scarps indicates a relative age for the latest event about the year A.D. 1100 (Pérez-López et al., 2007; Rodríguez-Pascua et al., 2008). In order to avoid uncertainties and with the aim of understanding the theoretical archaeoseismic potential of El Tolmo site, we analyzed the archaeological quality factor (AQF; Sintubin and Stewart, 2008). Despite the low AQF value obtained for the studied site (AQF = 0.074), the archaeoseismic evidence showed in this study is mainly supported by the site environment, site potential, and tectonic context, especially taking into account the paleoseismic evidence linked to the close active fault segment of the Pozohondo fault. The low AQF value for El Tolmo site is very dependent on the lack of historical reports, the geographical isolation of the site, and the nature of the up-to-date

Ancient earthquakes from archaeoseismic evidence during the Visigothic and Islamic periods, SE Spain archaeological research impacting on the SCL index of the AQF factor, but also on the low slip rate estimated for the Pozohondo fault (~0.09 mm/yr). However, the significance of this study is in providing an AQF value for archaeological sites located within active intraplate areas with slow-moving faults and long recurrence time intervals (~100,000 yr) for strong events, which have not previously been analyzed from an archaeoseismic point of view. Since this will be the case for most of the archaeoseismic evidence that can be unraveled in the Alpine deformation zone of the Western Mediterranean, the studied case can be used as a guide for future research in this zone. Whatever the case, the AQF factor should be refined considering that archaeoseismology is an emergent research in worldwide areas except in the Eastern Mediterranean and Middle East. In other different areas, even long-term, largescale, multidisciplinary archaeological excavations with nice evidence of archaeoseismic damage, but previously unexplored from this point of view, will results in low AQF values. Furthermore, absolute dating for archaeoseismic and paleoseismic samples is necessary to establish more detailed relationships between the paleoseismic history of the Tobarra-Cordovilla segment and the archaeological evolution of El Tolmo de Minateda archaeological site. In this sense, this study first reports combined paleoseismic and archaeoseismic evidence consisting in large rockfall events (~3800 m3) dislocating anthropomorphic tombs carved into the bedrock. Combining the directionality of rock slides, columns, and masonry for the first episode of destruction (Fig. 2), a clear dominant NNE-SSW orientation for the ground movement can be preliminary proposed, following the criteria proposed by other authors (Altunel et al., 2003; Silva et al., 2005). ACKNOWLEDGMENTS We wish to thank the Department of Archaeology and Historic Heritage, Comunidad Autónoma de Castilla La Mancha, for support. This work was sponsored by the Spanish Project ACTISIS CGL2006-05001/BTE. This article is a contribution to the United Nations Educational Scientific and Cultural Organization–funded International Geoscience Programme IGCP 567, “Earthquake Archaeology: Archaeoseismology along the Alpine-Himalayan Seismic Zone.” REFERENCES CITED Abad-Casal, L., Gutiérrez-Lloret, S., and Sanz-Gamo, R., 1998, El Tolmo de Minateda, una historia de tres mil quinientos años: Toledo, Ed. Junta de Comunidades de Castilla–La Mancha, 161 p. (in Spanish). Abad-Casal, L., Gutiérrez-Lloret, S., and Gamo-Parras, B., 2000a, La ciudad visigoda del Tolmo de Minateda (Hellín, Albacete) y la sede episcopal de Eio, in Los Orígenes del Cristianismo en Valencia y su Entorno (Grandes Temas Arqueológicos II): Valencia, Editorial Ajuntament de València, p. 101–112 (in Spanish). Abad-Casal, L., Gutiérrez-Lloret, S., and Gamo-Parras, B., 2000b, La basílica y el baptisterio del Tolmo de Minateda (Hellín, Albacete): Archivo Español de Arqueología, v. 73, p. 193–221.

183

Altunel, E., Stewart, I.S., Barka, A., and Piccardi, L., 2003, Earthquake faulting at ancient Cnidus, SW Turkey: Turkish Journal of Earth Sciences, v. 12, p. 137–151. Atakan, K., Midzi, V., Moreno, Toiran, B., Vanneste, K., Camelbeeck, T., and Meghraoui, M., 2000, Seismic hazard in regions of present-day low seismic activity: Uncertainties in the paleoseismic investigations along the Bree fault scarp (Roer graben, Belgium): Soil Dynamics and Earthquake Engineering, v. 20, p. 415–427. Benito, B., Capote, R., Murphy-Corella, P., Gaspar-Escribano, J.M., MartínezDíaz, J.J., Tsige, M., Stich, D., García-Mayordomo, J., García, M.J., Jiménez, E.M., Insua-Arévalo, J.M., Álvarez-Gómez, J.A., and Canora, C., 2007, An overview of the damaging and low-magnitude La Paca earthquake (Mw 4.8) on 29 January 2005: Context, seismotectonics and seismic risk implications for southeast Spain: Bulletin of the Seismological Society of America, v. 97, p. 671–690, doi: 10.1785/0120050150. Force, E.R., 2008, Tectonic environments of ancient civilizations in the Eastern Hemisphere: Geoarchaeology, v. 23, no. 5, p. 644–653, doi: 10.1002/ gea.20235. Galadini, F., Hinzen, K.G., and Stiros, S.C., 2006, Archaeoseismology: Methodological issues and procedure: Journal of Seismology, v. 10, p. 395– 414, doi: 10.1007/s10950-006-9027-x. García del Cura, M.A., Ordóñez, S., and Calvo, J.P., 1979, Estudio sedimentológico de la cuenca cuaternaria de Cordovilla (Provincia de Albacete), in Actas de la III Reunión Nacional del Grupo Español de Trabajo del Cuaternario: Zaragoza, Consejo Superior de Investigaciones Científicas Editions, p. 247–257 (in Spanish). Guidoboni, E., Muggia, A., and Valensise, G., 2000, Aims and methods in territorial archaeology: Possible clues to a strong fourth-century A.D. earthquake in the Straits of Messina (southern Italy), in McGuire, W.J., Griffiths, D.R., Hancock, P.L., and Stewart, I.S., eds., The Archaeology of Geological Catastrophes: Geological Society of London Special Publication 171, p. 45–70. Gutiérrez-Lloret, S., 2000, La identificación de Madinat Iyih y su relación con la sede episcopal Elotana. Nuevas perspectivas sobre viejos problemas, in Olcina, M.H., and Soler-Díaz, J.A., eds., Scripta in Honorem E.A. Llobregat: Alicante, p. 481–501 (in Spanish). Herraiz, M., De Vicente, G., Lindo-Ñaupari, R., Giner-Robles, J.L., Simón, J.L., González-Casado, J.M., Vadillo, O., Rodríguez-Pascua, M.A., Cicuéndez, J.I., Casas, A., Cabañas, L., Rincón, P., Cortés, A.L., Ramírez, M., and Lucini, M., 2000, The recent (Upper Miocene to Quaternary) and present tectonic stress distributions in the Iberian Peninsula: Tectonics, v. 19, p. 762–786, doi: 10.1029/2000TC900006. Martínez-Solares, J.M., 2001, Los efectos en España del Terremoto de Lisboa (1 de noviembre de 1755): Madrid, Ed. Dirección General del Instituto Geográfico Nacional, 756 p. (in Spanish). Martínez-Solares, J.M., and Mezcua, J., 2002, Catálogo Sísmico de la Península Ibérica (880 A.C.–1900): Madrid, Instituto Geográfico Nacional Monografía 18, 756 p. Meghraoui, M., Gomez, F., Sbeinati, R., Van der Woerd, J., Mouty, M., Darkal, A.N., Radwan, Y., Layyous, I., Al-Najjar, H., Darawcheh, R., Fouad Hijazi, F., Al-Ghazzi, R., and Barazangi, M., 2003, Evidence for 830 years of seismic quiescence from paleoseismology, archaeoseismology and historical seismicity along the Dead Sea fault in Syria: Earth and Planetary Science Letters, v. 210, p. 35–52, doi: 10.1016/S0012-821X(03)00144-4. Michetti, A.M., Esposito, E., Guerrieri, L., Porfido, S., Serva, L., Tatevossian, R., Vittori, E., Audemard, F., Azuma, T., Clague, J., Comerci, V., Gürpinar, A., McCalpin, J., Mohammadioun, B., Mörner, N.A., Ota, Y., and Roghozin, E., 2007, Intensity scale ESI 2007: Memoria Descriptiva Carta Geologica Italiana 74, 41 p. Mistler, M., Butenweg, C., and Meskouris, K., 2006, Modeling methods of historic masonry buildings under seismic excitation: Journal of Seismology, v. 10, p. 497–510, doi: 10.1007/s10950-006-9033-z. Murphy Corella, P., 2005, Building performance during recent earthquakes in the Iberian Peninsula and surrounding regions: Proceedings of the 250th Anniversary of the 1755 Lisbon Earthquake: Lisbon, p. 446–450. Nur, A., and Burgess, D., 2008, Apocalypse: Earthquakes, Archaeology and the Wrath of God: Princeton, New Jersey, Princeton University Press, 309 p. Pérez-López, R., Rodríguez-Pascua, M.A., Giner-Robles, J.L., Calvo, J.P., Garduño-Monroy, V.H., Israde-Alcantara, I., and Bischoff, J., 2007, Calibration of the diffusion constant (K0) for dating coseismic fault scarps by using the diffusion equation: Application to the Alboraj earthquake,

184

Rodríguez-Pascua et al.

Albacete, SE Spain, in Lario, J., and Silva, P.G., eds., Contribuciones al Estudio del Periodo Cuaternario: Avila, Spain, Asociación Española para el Estudio del Cuaternario, p. 161–162 (in Spanish). Rodríguez de la Torre, F., 1995, Los terremotos en la Provincia de Albacete: Albacete, Instituto de Estudios Albacetenses, 48 p. Rodríguez-Pascua, M.A., De Vicente, G., Calvo, J.P., and Pérez-López, R., 2003, Similarities between recent seismic activity and paleoseismites during the late Miocene in the External Betic Chain: Relationship by “b” value and fractal dimension: Journal of Structural Geology, v. 25, p. 749– 763, doi: 10.1016/S0191-8141(02)00078-0. Rodríguez-Pascua, M.A., Pérez-López, R., Calvo, J.P., and García del Cura, M.A., 2008, Recent seismogenic fault activity in a late Quaternary closedlake graben basin (Albacete, SE Spain): Geomorphology, v. 102, p. 169– 178, doi: 10.1016/j.geomorph.2007.06.023. Rodríguez-Pascua, M.A., Bischoff, J., Garduño-Monroy, V.H., Pérez-López, R., Giner-Robles, J.L., Israde-Alcántara, I., Calvo, J.P., and Williams, R.W., 2009, Estimation of the tectonic slip rate from Quaternary lacustrine facies within the intraplate Albacete Province (SE of Spain): Sedimentary Geology, v. 222, p. 89–97, doi: 10.1016/j.sedgeo.2009.06.007. Silva, P.G., Borja, F., Zazo, C., Goy, J.L., Bardají, T., De Luque, L., Lario, J., and Dabrio, C.J., 2005, Archaeoseismic record at the ancient Roman city of Baelo Claudia (Cádiz, south Spain): Tectonophysics, v. 408, no. 1–4, p. 129–146, doi: 10.1016/j.tecto.2005.05.031. Silva, P.G., Rodríguez-Pascua, M.A., Pérez-López, R., Bardají, T., Lario, J., Alfaro, P., Martínez-Díaz, J.J., Reicherter, K., Giménez, J., Giner, J., Azañón, J.M., Goy, J.L., and Zazo, C., 2008, Catalogación de los efectos geológicos y ambientales de los terremotos en España en la Escala ESI2007 y su aplicación a los estudios paleosismológicos: Geotemas, v. 10, p. 318.

Silva, P.G., Reicherter, K., Grützner, Ch., Bardají, T., Lario, J., Goy, J.L., Zazo, C., and Becker-Heidmann, P., 2009, Surface and subsurface palaeoseismic records at the ancient Roman city of Baelo Claudia and the Bolonia Bay area, Cádiz (south Spain), in Reicherter, K., Michetti, A.M., and Silva, P.G., eds., Palaeoseismology: Historical and Prehistorical Records of Earthquake Ground Effects for Seismic Hazard Assessment: Geological Society of London Special Publication 316, p. 93–121. Similox-Tohon, D., Sintubin, M., Muchez, P., Verhaert, G., Vanneste, K., Fernandez-Alonso, M., Vandycke, S., Vanhaverbeke, H., and Waelkens, M., 2006, The identification of an active fault by a multidisciplinary study at the archaeological site of Sagalassos (SW Turkey): Tectonophysics, v. 420, p. 371–387, doi: 10.1016/j.tecto.2006.03.026. Sintubin, M., and Stewart, I.S., 2008, A logical methodology for archaeoseismology: A proof of concept at the archaeological site of Sagalassos, southwest Turkey: Bulletin of the Seismological Society of America, v. 98, no. 5, p. 2209–2230, doi: 10.1785/0120070178. Stich, D., Ammon, C.J., and Morales, J., 2003, Moment tensor solutions for small and moderate earthquakes in the Ibero–Maghreb region: Journal of Geophysical Research, v. 108, p. 2148, doi: 10.1029/2002JB002057. Stiros, S.C., 1996, Identification of earthquakes from archaeological data: Methodology, criteria and limitations, in Stiros, S., and Jones, R., eds., Archaeoseismology: British School at Athens, Fitch Laboratory Occasional Paper 7, p. 129–152. Stiros, S., and Jones, R.E., eds., 1996, Archaeoseismology: British School at Athens, Fitch Laboratory Occasional Paper 7, 268 p.

MANUSCRIPT ACCEPTED BY THE SOCIETY 19 MAY 2010

Printed in the USA

The Geological Society of America Special Paper 471 2010

Distinguishing damages from two earthquakes—Archaeoseismology of a Crusader castle (Al-Marqab citadel, Syria) Miklós Kázmér* Department of Palaeontology, Eötvös University, P.O. Box 120, H-1518 Budapest, Hungary Balázs Major* Department of Arabic and Islamic Studies, Péter Pázmány Catholic University, Egyetem utca 1, H-2087 Piliscsaba, Hungary

ABSTRACT Damages from two major earthquakes are identified in medieval Al-Marqab citadel (Latin: Margat) in coastal Syria. Built by the Order of St. John (Hospitallers) in the twelfth–thirteenth centuries, the hilltop fortification has masonry walls made with and without mortar, using the opus caementum technology (Roman concrete). V-shaped and U-shaped failures, single-corner and symmetrical corner collapses, and in-plane shifts of ashlar masonry walls are identified and dated by historical and archaeological methods. The azimuth of displacement is NE-SW for the older damages of the Crusader period (A.D. 1170–1285), possibly related to the A.D. 1202 earthquake. A later, NW-SE displacement occurred during the Muslim period (post1285). The 1202 earthquake produced at least VIII intensity on the MSK scale at AlMarqab, which is higher than previously considered.

INTRODUCTION After a few decades of hesitating progress, archaeoseismology is becoming a quantitative science using a rigorous methodology. At the moment, it is still uncertain whether seismic traces in destruction layers and buildings can be translated into earthquake parameters such as intensity, peak ground acceleration, magnitude, distance to epicenter, etc. (Sintubin et al., 2008). This paper is intended to be a small step toward this quantification by offering a method to distinguish among traces of successive earthquakes, to establish a temporal succession, and to identify vibration directions. However, one should not forget that the latter are not necessarily straightforward indicators of epicenter locations (Hinzen, 2008, 2009).

In the past few decades, spectacular features of failure of various archaeological monuments in the Eastern Mediterranean and Near East have attracted interpretations of earthquakes in general, especially among archaeologists (Kilian, 1980). Cautionary words by seismologists, (e.g., Ambraseys, 2005a, 2006) warned that failures attributed to earthquakes are often due to poor foundation practices, landslides, and changes in groundwater level. In the meantime, a stream of publications by geologists appeared, describing major fault-related deformation of walls and buildings. Since displacement along geological faults is an unequivocal sign of earthquakes, these papers provided a solid foundation for proper interpretation of earthquake-related damages (Hancock and Altunel, 1997; Ellenblum et al., 1998; Galli and Galadini, 2001; Sintubin et al., 2003; Altunel et al., 2003;

*E-mails: Kázmér—[email protected]; Major—[email protected]. Kázmér, M., and Major, B., 2010, Distinguishing damages from two earthquakes—Archaeoseismology of a Crusader castle (Al-Marqab citadel, Syria), in Sintubin, M., Stewart, I.S., Niemi, T.M., and Altunel, E., eds., Ancient Earthquakes: Geological Society of America Special Paper 471, p. 185–198, doi: 10.1130/2010.2471(16). For permission to copy, contact [email protected]. © 2010 The Geological Society of America. All rights reserved.

185

186

Kázmér and Major

Galli et al., 2008; Karakhanian et al., 2008a). Where the faults responsible for displacement are excavated and corresponding shifted beds are dated (e.g., Meghraoui et al., 2003; Reicherter et al., 2003), these features are among the best evidences for earthquake-related damages. Other deformations as seen on buildings, like broken window and door frames, dropped keystones, displaced and/or rotated ashlars of masonry walls (e.g., Nikonov, 1988; Korjenkov and Mazor, 1999, 2003; Akyüz and Altunel, 2001; Altunel et al., 2003; Sintubin et al., 2003; Caputo and Helly, 2005; Similox-Tohon et al., 2006, 2007; Al-Tarazi and Korjenkov, 2007; Karakhanian et al., 2008a, 2008b; Marco, 2008), and spectacular disarrangement among drums of columns in Greek and Roman temples (e.g., Stiros, 1996; Stiros et al., 2000; Bottari, 2005), crosscutting breaks in walls, and twisted walls (e.g., Kamh et al., 2008; Nur and Burgess, 2008) have often invited interpretation as earthquake damages. Theoretical background of these interpretations is often missing. Alternative explanations have been sought for decades: foundations problems and poor construction practices have been considered, while precise recording of observations has been emphasized for the benefit of subsequent researchers (e.g., Karcz and Kafri, 1978). Seismic-induced landslide damage to buildings is also a problem (Wechsler et al., 2009). Here, we attempt to follow the example of Korjenkov and Mazor (1999, 2003) using kinematic indicators borrowed from structural geology in the interpretation process. After a preliminary report on earthquake-induced damages at Al-Marqab (Kázmér, 2008), we proceed with the quantitative characterization and identification of past earthquakes using archaeoseismology. This

E

AL-MARQAB CITADEL Al-Marqab citadel (Arabic: Qal‘at Al-Marqab; medieval Latin: Margat), in the coastal region of the Syria, is one of the largest and most important medieval castles of the Levant (Figs. 1–2). The site is perched on top of a 350-m-high volcanic mountain, ~2 km inland from the coast, overlooking the town of Banyas and guarding the coastal route. It also has commanding views of the fertile plains toward Latakia in the northeast, and it dominates the Jabal Ansariyya ranges to the east. History The first castle of the site is reported to have been built by the local inhabitants in H. 454 (A.D. 1062–1063). After a brief period of Byzantine occupation starting around 1104, it was taken by the Franks (Crusaders) from the local tribes in 1117– 1118. The castle seems to have reverted to as-yet-unknown Muslim hands in the 1130s during the civil war in Antiochia. It was recaptured by Renaud II Mazoir in 1140, and then became the seat of the Mazoir family (Deschamps, 1973, p. 260–261). The Mazoirs were one of the highest-ranking baronial families in the Crusader principality of Antioch and were responsible for building most of the earliest surviving structures in the castle. In early February 1187, the Mazoirs transferred Al-Marqab and all their

40

38

36

38 N 34 E

article is a small step toward a rigorous and transparent methodology for archaeoseismology (Stiros, 1996; Galadini et al., 2006; Sintubin and Stewart, 2008).

42

38 N

AF

1114

TURKEY 37

37 1822 x Aleppo

Antakya x

Beirut x

34

1157 1404 1170

x Palmyra

1202

34

1759 x Damascus

IRAQ

1759 DS F

33

1212

33 100 km

0

JORDAN

32 34 E

35

S

1222

36

EF

35 1303

1796 Lattakia 1287 Al-Marqab x x Tartus Tripoli x YF

36

N

36

38

32 40

42

Figure 1. Location of Al-Marqab (Margat) citadel in coastal Syria. Major historical earthquakes are centered along left-lateral strike-slip faults ranging from the Dead Sea fault in the south to the East Anatolian fault in the north (modified after Sbeinati et al., 2005). Epicenter of the A.D. 1202 earthquake, extensively discussed in the text, is underlined. Epicenters of the events of 1212, 1222, and 1303 earthquakes are out of the map in Jordan, Cyprus, and Crete, respectively. DSF—Dead Sea fault system, YF— Yammouneh fault, EAF—East Anatolian fault system. EFS—Euphrates fault system. Epicenter locations are from Ambraseys (2009, electronic supplement). See also Table 1.

Distinguishing damages from two earthquakes—Archaeoseismology of a Crusader castle, Syria

187

Figure 2. Al-Marqab citadel, as seen by attacking enemy from the south, was mostly built by the crusading Order of St. John (Hospitallers) in the late twelfth century. The hilltop plateau, 350 m above sea level, is vesicular basalt lava of Pliocene age, exposed below the highest towers. Terraces of olive gardens carved in weathered basaltic strata cover the slopes. Banyas city and the Mediterranean Sea are seen in the background on the right.

landed properties to the Order of St. John (Hospitallers) due to unbearable maintenance costs related to warfare and damaging earthquakes (Burgtorf, 2007). The Hospitallers turned the castle a military, judicial, and administrative center of the region. Given the fact that Al-Marqab became the central castle of one of the most influential organizations of the age, it is not surprising that the Hospitaller period was characterized by largescale construction programs that resulted in the erection of most of the surviving buildings seen on the mountain top. The castle was put to the test several times by besieging armies. It was besieged by an army from Aleppo in 1204–1205, and again in 1231. Banyas and the lands around the castle were destroyed by the Aleppines. Attempts on the Muslim side to take the castle twice, once in 1269–1270, and subsequently in 1281 and 1282, ended in failure. Al-Marqab was finally taken by the Sultan Qalawun on 25 May 1285 after a relatively short siege of 5 wk. The sultan agreed to the peace offer of the garrison to save the castle from further damage, and the destructions caused by the siege were repaired immediately (Ibn-Abdazzahir, 1946). After the complete expulsion of the Crusaders, the castle started losing its importance, and its diminishing role in the Mamluk (1250–1517) and Ottoman periods (1517–1917) is reflected in the reduced scale of later building activities. For a lively description of the castle and the function of various buildings, see Kennedy (1994, p. 163–179). For additional details concerning the history of AlMarqab, the reader is referred to Major et al. (2010). Seismicity and Large Earthquakes Besides the relatively scarce military activity, earthquakes were another considerable factor in the building history of the castle (Table 1; Fig. 1). The major earthquakes of 1114, 1157, and 1170 are suspected to have caused considerable damage to the pre-Hospitaller castle. The earthquake of 20 May 1202, one of the strongest in the history of the region, did severe damage to the castle, but left it defensible for the time being. We do not have any mention of whether other earthquakes of the thirteenth and fourteenth centuries in northern Syria affected the fabric of the castle (1212, 1222, 1236, 1287, 1303, 1339), but the earthquake of 20 February 1404 certainly did bring down a considerable part of the castle (Sbeinati et al., 2005, p. 392). Amongst the later

earthquakes, the ones of 1752 and 1759 are very likely to have affected Al-Marqab, and the ones of 26 April 1796 and 13 August 1822 are explicitly described to have caused serious damage to the castle (Sbeinati et al., 2005, p. 379–398) (Table 1; Fig. 1). The medieval building complex of Al-Marqab occupies the whole mountain top (Figs. 2–3), covering 5.7 ha, and is made up of two basic units. In the southern part of the mountain stands the concentric citadel, covering an area of 0.9 ha, while the rest of the mountain plateau is occupied by a huge suburb, which is also enclosed in a double line of defensive walls. Some parts of this suburb were inhabited until 1959. In the first phase of the research program, the work of the Syro-Hungarian Archaeological Mission (SHAM) focused primarily on the citadel area. Objectives of the Syrian-Hungarian Archaeological Mission are archaeological excavation, architectural survey, and photogrammetry of Al-Marqab citadel; conservation and restoration of unique medieval frescoes and artifacts; scientific (geological, geophysical, archaeozoological, archaeobotanical) investigations; study of the medieval technologies; exploration and reconstruction to working use of the medieval water-collection and sewage disposal system with the aim of donating the water collected to the neighboring villages; folklore studies to document and revitalize the traditional Syrian village life and houses in the suburb; and landscape archaeology to reconstruct the medieval rural settlement pattern and provide training opportunities for students from the east and the west. These activities prepare AlMarqab for nomination as United Nations Educational, Scientific, and Cultural Organization World Heritage Site. METHODS A plan of architectural structures has been provided by a previous geodetic survey carried out within the framework of EuroMED project. This is continuously being upgraded by SHAM members as archaeological excavations proceed (Fig. 3, 5, and 8–10). During the autumn field season of 2008, we identified and surveyed various damages and failures visible on buildings and walls. Some of those attributable to earthquakes are described, illustrated, and explained here. Boundaries of displaced or collapsed portions of ruptured walls were measured by tape, compass, and clinometer: measurements of azimuth/strike

D

35.5

36.0

3

3

6

7

R

R

R

Lattakia

Litani, Bekaa

Safed

Lattakia

Aleppo

SY

LE

IS

SY

SY

LE

GR

Crete Rhodes Tripoli

SY

JO CY SY-TR

SY

Lattakia

W. Araba Cyprus N. Syria

Baalbek

Muslim (Ottoman) period

Muslim (Mamluk) period

Hospitaller period

Sbeinati et al. (2005, p. 396)

Sbeinati et al. (2005, p. 381, 392); Ambraseys (2009)

Sbeinati et al. (2005, p. 391)

Guidoboni and Comastri (1997); Sbeinati et al. (2005, p. 391); Ambraseys (2009)

Sbeinati et al. (2005, p. 374); Ambraseys (2009)

Sieberg (1932); Ambraseys (2009) Sieberg (1932); Ambraseys (2009) Sieberg (1932)

Ambraseys and Melville (1988); Daëron et al. (2005); Sbeinati et al. (2005, p. 381, 389– 391); Ambraseys (2009)

Damage described

Sbeinati et al. (2005, p. 382, 398); Ambraseys (2009)

Ambraseys and Barazangi (1989); Daëron et al. (2005); Sbeinati et al. (2005, p. 396–397); Probably Ambraseys (2009) damaging earthquakes Ambraseys and Barazangi (1989); Sbeinati et al. (2005, p. 397); Ambraseys (2009)

Damage described

No source on damage

Contemporary description

1822

8

13

36.7

36.8

3

7

R

S.E. Anatolia

TR

Sbeinati et al. (2005, p. 376–380, 382); Ambraseys (2009) Notes: Q—quality of epicenter localization, where 0 is worst and 3 is best. Ms—equivalent surface-wave magnitude, where 5 stands for small event 5.0 ≤ Ms ≤ 6.0, 6 stands for 6.0 ≤ Ms ≤ 7.0, and 7 stands for 7.0 ≤ Ms ≤ 8.0. Effects—maximum effects reported, where D stands for damaging to dwellings and public buildings, and R stands for destructive earthquake causing great damage and the loss of life. Region: CY—Cyprus; GR—Greece; IS—Israel; JO—Jordan; LE—Lebanon; SY—Syria; TR—Turkey.

26

35.9

6

D

4

33.7

3

6

8?

1796

25

35.6

2

3

11

33.1

36.4

28.0

D

6

1759

30

35.2

35.0

2

D D

R

7 6

7

10

21

20

8

36.2

35.3

3 2

3

1759

8

1303

21

35.0 32.5

30.0 34.5

36.1

7

3

1287

1 11

34.1

1752

5 5

1212 1222 1236

20

D

5

1202

29

2

6

1170

12

1404

8

1157

29

D

11

1114

Day

1339

Month

Year

TABLE 1. HISTORICAL EARTHQUAKES REPORTED OR SUSPECTED TO HAVE AFFECTED AL-MARQAB CITADEL (MODIFIED AFTER AMBRASEYS, 2009) Lat Long Quality Ms Effects Site Region Age Damage Reference (°N) (°E) of epicenter magnitude localization Antioch, Sbeinati et al. (2005, p. 369–370); 37.5 37.2 3 7 R TR PreMaras Ambraseys (2009) Hospitaller Guidoboni et al. (2004); Sbeinati et al. (2005, (alternating No source 35.3 36.4 3 7 R Apamea SY Byzantine on damage p. 371–373); Ambraseys (2009) and Muslim) Sbeinati et al. (2005, p. 373–374); period 34.7 36.4 2 7 R Shaizar SY Ambraseys (2009)

Distinguishing damages from two earthquakes—Archaeoseismology of a Crusader castle, Syria

N Figs. 7-8

Fig. 6

189

Figure 3. Plan of the southern portion of Al-Marqab citadel. Heavy lines denote buildings with aboveground walls. Light lines are circumferential walls and excavated foundations of buildings. Inset displays the location of the heavily fortified southern part and the much larger northern suburb, surrounded by weaker walls, totaling 5.7 ha together. The external double line is the modern asphalt road surrounding the hilltop.

Fig. 5

Fig. 9 0

20 m

and dip of bounding planes allow the description of failure orientation in space. Similar orientations of the stress field are inferred to form failure groups, where each stress field system occurred only once. Archaeological dating provides the ages of particular failures and related stress fields, which are correlated to past earthquakes using the seismicity catalogue (Fig. 1; Table 1). Direct and indirect archaeological evidences provide temporal constraints on failures. Direct evidence is where we can document a failure of a building of known age: the failure is certainly younger than the construction date (terminus post quem datum). Indirect evidence for dating a failure is provided where an adjacent building of known construction age does not bear the same damage (terminus ante quem datum). Direct evidence is hard and

reliable; indirect evidence is soft and is prone to several errors, e.g., the adjacent building stands on different soil, was erected by different technology, has different structure, and—ultimately—is characterized by different vibration characteristics. RESULTS Masonry Components and Types Most walls of Al-Marqab, both Crusader and Muslim, are one of two types: either stone masonry or opus caementitium, i.e., “Roman concrete” (Lamprecht, 2001) or “ancient concrete” (Ferretti and Bažant, 2006). Stone masonry is characterized by

190

Kázmér and Major

dressed stones, carved rectangular and of standard size, with or without mortar, always without metal anchors. Arches, door, and window ledges, box machicolations, and some wall heads have been constructed this way. “Roman concrete” or “ancient concrete” is a mixture of sand, lime, and added stone material and is very similar to modern concrete in appearance. Invented by the Romans, the technique survived well into the Middle Ages. Opus caementitium is often combined with traditional masonry, where an outer, visible layer of variously dressed blocks was erected with mortar. This external, regular masonry work served during construction as a mold for casting the core. Poured material served for the inner, invisible parts of the wall (Ferretti and Bažant, 2006; Mistler et al., 2006). Masonry both served aesthetic demands and provided a hard, protective layer to counter weather effects and enemy attacks. This layer often served as framework during concrete pouring only, having no supporting function when concrete hardened (Fig. 4). Walls and vaults of variable thickness, from a few tenths of a meter up to 5 m thickness, were constructed this way. This multilayer construction technology made walls of AlMarqab castle extremely durable, even without reinforcement. For assessment of earthquake damages, the external masonry is treated as consisting of discrete blocks, while the concrete wall behaves as a cohesive block. For modeling purposes, this kind of wall is treated as poor Portland cement concrete (Ferretti and Bažant, 2006). While this type of wall may deteriorate through the centuries due to creep and fatigue (Anzani et al., 1995, 2009), we can be sure that this was not the case just a few decades after construction.

FAILURES V-Shaped Failure There is a spectacular V-shaped extrusion on the donjon, the main tower of the citadel of 5 m height and 5 m width (Fig. 5). Similar features occur elsewhere in the castle. It seems that if failure were to progress, we would see a wedge-shaped block missing from the wall of the donjon. The V-shaped block is shifted toward the SW by ~20 cm. No trace of it can be seen within the donjon. Bounding surfaces are joined before reaching the hall inside. Single Corner Collapse Adjoining, possible perpendicular walls have collapsed at their joining. Collapse occurs where both walls are free-standing, i.e., unconfined at least to one side. This partial collapse produces an uneven oblique surface, cutting both walls at an angle (Fig. 6). Although of irregular shape, the pattern of collapse is comparable to a failure plane that can be interpreted as a normal fault. The smoothed surface of the failure is considered the fault plane, where the two directions necessary for geological characterization, strike and dip, can be measured and/or calculated. Because we do not have any evidence for the displacement direction of the hanging wall (fallen fragments have been cleared centuries ago), we assume dip slip. Similar failures are illustrated by Galadini et al. (2006, his Fig. 2a), Penazzi et al. (2001, their Fig. 1), and Tomaževič and Lutman (2007, their Fig. 7). Symmetrical Corner Collapse Room M3 sits on top of the vault of the kitchen. It is the sole remnant of a previous, larger cluster of rooms, which might have served as an independent kitchen. Walls that are 66 to 104 cm wide bear a barrel vault. Diagonally opposite corners have suffered symmetrical damages (Figs. 7–8). Fractures that are concave outward have developed. The NE corner collapsed in full, destroying a segment of the vault and portions of the adjacent wall (Fig. 7A). The concave fault developed in the SW corner as well, but only part of the vault collapsed: there is a 2 × 1 m hole in the top of the vault, connected by an arcuate fracture—a wouldbe failure scar—to the still-intact adjacent walls (Fig. 7B). U-Shaped Gap

Figure 4. Ashlars in the western, windward wall of the donjon are seemingly unsupported. However, their rear side is firmly embedded in Roman concrete, the cementing material of the several-meter-thick wall. Laid initially with mortar, westerly winds and rain have removed much of it throughout eight centuries. Arrow: measuring tape for scale, 20 cm long.

The top of the southern corner tower of the outer enceinte, the outer ring of walls of the Mamluk-built structure, bears a downward-concave failure. Both thin and thick portions of the tower have failed (Fig. 9). A wider than deep, downward-concave failure mode is illustrated by spectacular examples from Pompeii by Martini (1996, his scoop-like failure). It was produced by shaking of relatively

Distinguishing damages from two earthquakes—Archaeoseismology of a Crusader castle, Syria

191

Figure 5. Left: A wedge-shaped block of donjon masonry wall moved toward 240° azimuth by ~20 cm. Right: Dashed line on archaeological plan of the top of donjon indicates estimated shape of displaced wedge. Gray arrow denotes direction of displacement.

Dislodged Building Blocks

Figure 6. Failed corner of perpendicular walls at NW corner of the donjon. Approximated by a normal fault dipping ~50° to 284°NW direction. Failure is ~3 m wide at horizontal line.

thin walls (Lourenço et al., 2007, their Fig. 5), semiconfined at both ends by walls adjoining at an angle, often perpendicularly. Similar failures have been described by Similox-Tohon et al. (2006, their Fig. 5d; 2007, their Fig. 5d) and Sintubin and Stewart (2008, their Fig. 12b).

A large variety of shifted and rotated building blocks (ashlars) are seen at Al-Marqab. A shift within the plane of the wall is spectacularly shown in Figure 10. Heavily protected stone boxes extrude from the top of walls. Open bottoms allowed defenders to pour hot water, oil, or burning tar on attackers climbing the walls. Box machicolations and adjacent walls on top of the southern Mamluk tower suffered in-plane extension of several tenths of a meter, and open spaces up to 10 cm wide formed between adjoining blocks during ground shaking. Although an indirect observation, this type of damage is confidently assigned to earthquakes, even by the otherwise cautious Ambraseys (2006, p. 1010). Similar open joints are described by Sintubin et al. (2003, their Fig. 5a) and Marco (2008, his Fig. 2F), and have been reproduced by vibration experiments (Vasconcelos et al., 2006, their Fig. 7). There are many other kinds of damages observed in AlMarqab: dropped keystones, in-plane and out-of-plane failures of walls, twisted walls, rotated blocks, extruded blocks, etc., which will be treated separately. Subsoil The buildings and walls of Al-Marqab have been erected on the solid subsoil of a several-meter-thick layer of compact Pliocene basalt (Fig. 11). This rock is not prone to liquefaction, even under major earthquakes, and neither is it affected by compaction under changing groundwater level (Ambraseys, 2006). The latter is ~50 m below the citadel, as shown by the location of the public bath on the western hillside.

A

B

Figure 7. Room M3, as seen from inside, bearing symmetrically arranged damages to opposite corners due to a NE-SW–oriented vibration. (A) Collapsed NE corner of room M3, concave fracture (light curve) facing 50°NE. Failure is 4 m wide. (B) Partially damaged SW corner of room M3, concave fracture (light curve of 4.5 m span) facing 240°SW. Fallen portion of damaged vault is encircled by dashed line.

0

5m

N

A B

Figure 8. Plan of room M3. Outline of symmetrical failures is indicated by dotted lines. Letters A and B correspond to failures shown on Figure 7. Arrows indicate 50°–240° extension direction, similar to the azimuth of the V-shaped failure of the donjon. Figure 9. Symmetrical, scoop-like damage affecting top of Mamluk tower facing toward 130°SE. Both upper, thin (140 cm) and lower, thick (>3 m) portions of wall collapsed toward SE (arrow). Two box machicolations are visible on top left.

Distinguishing damages from two earthquakes—Archaeoseismology of a Crusader castle, Syria

193

Figure 11. Crusader donjon (round tower partly hidden in background) and a later addition, Muslim southern tower built by Sultan Qalawun after his successful siege in 1285, bearing a row of white ashlars in the foreground. Both were firmly erected on several-meter-thick, unweathered vesicular basalt lava flow of Pliocene age (encircled), as seen on both sides of the glacis (inclined wall). Muslim tower is 20 m wide from corner to corner.

Figure 10. (A) Battlement with box machicolation suffered in-plane extension in 120°SE direction due to extension of the supporting wall. The upper 12 rows of ashlars have been displaced. Gaps between ashlars of the white stone are particularly wide, while blocks of the lowermost white row are still adjacent to each other. This is considered to be hard evidence for vibrations affecting the top of the Mamluk tower. Box is ~2 m wide. (B) Same box machicolation viewed from inside. Besides the observation slot in the center, there is a 10-cmwide gap between adjoining ashlars on the left (encircled), testifying to in-plane shaking. Extension is parallel with box machicolation face, in 120–300°SE-NW azimuth. Walk is 70 m wide. (C) Location of the semicircular Mamluk tower within the southern part of the citadel. Thin arrow points to the box machicolation that underwent extension. Wide gray arrow indicates direction of extension.

The heaviest possible damages inflicted by pre-gunpowder warfare were created by trebuchets (highly evolved catapults), throwing stone balls up to several hundred kilograms in weight against walls and onto roofs. The southern side of the donjon wall, most exposed to incoming projectiles, bears only minor fractures of conchoidal shape, witnesses of minor hits. The only really efficient siege tactic, mining, yielded collapse of walls. This method helped Sultan Qalawun’s army to take Al-Marqab in 1285 by undermining the southern tower. No traces of the mine were found. This gravity-induced failure, subsidence, has different geometrical features than those yielded by lateral seismic shaking.

A common source of damage, original construction defects, can be excluded by examining the surviving portions of the southern sector of Al-Marqab citadel. Mortar is still rock-solid in the failed walls. Textbook examples of subsidence are missing. Therefore, a seismic origin of damages is highly probable. Dating Damages The first archaeological excavation in Al-Marqab started in 2007; therefore, a large proportion of the castle fabric is still undated. However, a relative chronology (architectural stratigraphy as understood by Galadini et al., 2006) can be readily established for the buildings studied in this project. Muslim-built portions of the circumferential wall, especially the southernmost tower in the outer enceinte, are decorated with a frieze-like white row of ashlars within the black basalt wall. These blocks bear an Arabic inscription, testifying to its construction by the sultan AlMansur Qalawun (Mamluk sultan from 1279 to 1290), who took the castle in 1285 (Fig. 12). The donjon is certainly a Hospitaller construction and thus dates from after the order acquired the castle in 1187. Besides architectural design and the sheer size of the building, which could hardly have been financed by a private lord, the first results of the geophysical surveys also seem to support this dating. A georadar survey carried out inside the castle chapel (unequivocally accepted to have been the first Hospitaller construction on the site) detected the contours of a rectangular structure. Its position and the thickness of its wall, exceeding 3 m, make it a likely

194

Kázmér and Major

Figure 12. Portion of an Arabic inscription on the southern tower bearing the name of Sultan Qalawun, who conquered Al-Marqab in A.D. 1285. The full text says “This well-guarded fortress has been conquered and this tower rebuilt by Sultan Qalawun in months of the year [H.] 684. This work was executed under the direction of Balaban al-Mansuri,” as read and published by Max van Berchem in his Voyage en Syrie, on p. 303 in the early twentieth century (fide Deschamps, 1973, p. 273).

candidate for being the residential tower of the Mazoir family. This assumption is further strengthened by the presence of an old cistern incorporated in the southern walls of the chapel that stands in the center of the conjectured Mazoir tower. The presence of rock-cut cisterns under the main towers of eminent Crusader castles in the twelfth century is very common. Because the defense of the southern part of the castle mountain requires the presence of a dominant building, the substitution of the Mazoir square tower by the chapel must have been closely followed by the building of the new donjon, which is likely to have taken place in the 1190s. The supposed construction date of the donjon soon after 1187 and the precise dating of the southern corner tower after the successful Muslim siege of 1285 put constraints on dating the earthquake damages (Fig. 13). DISCUSSION Mechanism of Damages Indirect earthquake damage to buildings is caused by ground shaking. If the frequency of earth vibrations is close to the frequency of resonance of the building, excitation will occur, damage will be pervasive, and the building will collapse. If frequencies widely differ, the building will survive, possibly intact (for the spectacular example of the Pont du Gard in France, see Volant et al., 2009). Likely, this is the primary cause why halls with lower proportions, e.g., the Main Hall, collapsed (Major et al., 2010), while tall, stout buildings like the donjon survived each earthquake for 800 yr.

The donjon of Al-Marqab, being of 20 m diameter, 24 m height, and having walls up to 5 m thick, is a robust structure. Height/thickness ratio is h/t = 5, indicating extremely strong and earthquake-resistant construction (Lourenço et al., 2007). We note that Eurocode 8 building codes allow a maximum of h/t = 9 for earthquake-resistant buildings (Anonymous, 2003). In-plan area ratio (Lourenço and Roque, 2006) is 57%, again an overly resistant structure against all kinds of earthquake resonance. Eurocode 8 recommends 5%–6% for regular structures. A minimum value of 10% is recommended for historical masonry buildings (Meli, 1998). For simplicity sake, high seismicity cases can be assumed to be those where design ground acceleration for rock-like soils exceeds 0.2g. Area to weight ratio (Lourenço and Roque, 2006) is 10.4 m2/ MN, i.e., more than 8× higher than recommended (Meli, 1998). Seismological modeling of a smaller tower in Roman Tolbiacum, Germany (8.3 m diameter, 8 m high, having an up to 3.1-m-thick wall), yielded 0.12 m horizontal and 0.06 m vertical displacement at the top of the tower in case of a M > 6.4 earthquake (EMS98 intensity IX) (Hinzen, 2005). Deformation of the Al-Marqab donjon (Fig. 5) was of similar dimensions. A 0.06 m vertical displacement is more than enough to reduce friction between ashlars of the Mamluk tower while extension of the box machicolation and adjoining walls is in progress during shaking. How Many Successive Earthquakes? The Syrian earthquake catalogue (Sbeinati et al., 2005) lists a large number of damaging earthquakes in the coastal region. The name of Al-Marqab (or Margat) is mentioned for tens of them. Probably most of them caused appreciable damage to part of the castle. Because the donjon and the towers belong to the most heavily constructed portions of the citadel, we assume that only earthquakes with the highest intensity caused any damage to them. Earthquake 1 produced the V-shaped extrusion on top of the donjon (60°–240°). This earthquake occurred after the donjon was completed and before the southern tower was built: there are no traces at all of this damage direction on the southern tower. Earthquake 1 occurred during the interval between 1187 and 1285, after Hospitallers took the castle and before Mamluk occupation. A candidate earthquake is that of 1202, this being the largest in the Middle East ever recorded (see Table 1). Earthquake 2 produced the U-shaped damage to the southern corner tower. Additionally an extension of the top of the tower and of the box machicolation occurred in 120–300° direction. We can give only a terminus post quem date: it happened after 1285, i.e., during the Muslim period of Al-Marqab. In addition, a relative intensity of this quake would be lesser than that of the 20 May 1202, since it did not cause any visible damage to the donjon. While caution must be exercised in assigning damage azimuth to epicenter direction, according to Ambraseys and Melville

Distinguishing damages from two earthquakes—Archaeoseismology of a Crusader castle, Syria Known Date of earthquakes damages

195

(1988), the epicenter of the 1202 earthquake was south of AlMarqab, in the Bekaa Valley, while all major successive earthquakes had their epicenters to the north, near Aleppo (see also Fig. 1).

Orientation of damages

2000

Implications for the 1202 Earthquake

1822

1800 1796

310-130°

1600

Muslim period

1752 1759

1400 14041408

1200

1236 1222 1212 1202 1170 1157

1285

60-240° 1187

1114 1063

Byzantine period

1303 1287

Hospitaller period

1339

1062 - first castle built

1000 Figure 13. Dating of major earthquake damages in the history of AlMarqab citadel. Known earthquakes are listed after Sbeinati et al. (2005); most damaging seismic events are underlined. Double arrows refer to vibration directions as calculated from orientation of failures. These display an earlier, 60°–240° direction as shown by V-shaped extrusion of the donjon and symmetrical extensional failure of room M3 (Figs. 7–8). This is probably due to the 20 May 1202 earthquake. The later, 310°–130°-directed vibration is seen on the southern tower, built during the Muslim period (Fig. 9). It occurred any time after 1285 and may be correlated to the 1404 (and/or 1408?) earthquake.

The 1202 earthquake, widely discussed in various seismological papers (e.g., Ambraseys and Melville, 1988; Ellenblum et al., 1998; Kovach and Nur, 2006), has been considered as the most damaging earthquake in the Middle East. However, there are various, as yet unreported problems concerning dating, intensity, epicenter, and magnitude. Each of them will be discussed here briefly. Date Two contemporary Latin sources—written within days of the earthquake—report a damaging earthquake to large part of the then-Christian territories of the Middle East, including AlMarqab, as occurring in the early hours of 20 May 1202, a Monday. In a letter dated June of the same year, Geoffrey of Donjon, master of the Knights Hospitallers, reported to King Sancho VII of Navarra that “Al-Marqab had been badly affected but could be able to hold their own against hostile incursions unless there were more tremors” (Mayer, 1972, p. 303). The letter of Philip du Plessis, master of the Knights Templars, written at about the same time as Geoffrey’s letter, describes historical and environmental events between 1 November 1201 and 2 June 1202. His date for the earthquake is also 20 May 1202. Understandably, he does not mention the Hospitaller castle of AlMarqab, but describes damages elsewhere (Mayer, 1972). Many more Arabic historians report about an earthquake in Sha’ban month of H. 597 causing extensive damages from Egypt to Syria and northward. All of them cite or copy the contemporary authors Ibn Al-Athir (1999, v. X, p. 181) and Ibn Al-Jawzi (1952, p. 477–478), who write that a major earthquake damaged the Middle East between Egypt, Syria, and toward the north in the month Sha’ban H. 597. This date equals 7 May to 4 June 1201 A.D. None of them mentions any earthquake for H. 598, i.e., A.D. 1202. Since Shaban H. 597 overlaps the 20 May date of the contemporary Latin sources, although offset by a year, we can safely assume that there was only one major earthquake in the years 1201 and 1202. An original one-year error of Arabic historians—possibly working years or decades later than the event— has been inherited in successive works. The seismic catalogue of Sbeinati et al. (2005, p. 389–391) mentions both 1201 and 1202 earthquakes, listing all of them under the 20 May 1202 event. Ambraseys and Melville (1988) provide an extensive discussion of the event. By this reasoning, we can confidently exclude the possibility of amalgamation (Ambraseys, 2005b) of two successive earthquakes by historical sources. For further detailed discussion of dating problems in Arabic sources, see Ambraseys and Melville (1988).

196

Kázmér and Major

Intensity Ambraseys and Melville (1988) suggested intensity VII on the MSK scale for Al-Marqab based on historical documents only (the reports of Geoffrey of Donjon and on the history of Al-Jawzi). They considered it to be a shallow, large-magnitude multiple event, with widely felt aftershocks and a tsunami. While modified Mercalli intensity VII is the damage threshold for many archaeological sites (Kovach and Nur, 2006), we assume that damages at Al-Marqab citadel related to the 1202 earthquake require a larger intensity. The donjon, having up to 5-m-thick walls, is certainly a more earthquake-resistant structure than any ordinary city house, even palace. Since the donjon bears magnificent traces of only one major earthquake, occurring between 1187 and 1285, we suggest that the 1202 earthquake was of intensity VIII– IX at Al-Marqab (based on Rapp, 1986, his Table I). An intensity VIII value is corroborated by Geoffrey of Donjon’s letter, where he states that although Al-Marqab was heavily damaged, it can resist enemy attacks. The donjon—intact for military purposes— is located at the southernmost tip of the citadel, fully protecting it from any siege attacking from the mountain to the south. Increase of shaking intensity is justified by the statement of Geoffrey of Donjon, that Al-Marqab was heavily damaged (in those buildings that we cannot see now, probably many of them in the suburbium); however, this statement is uncertain. The donjon is an extremely earthquake-resistant construction (see previous). The fact that it has suffered any major damage, like the V-shaped wedge extrusion, is a signature that certainly more than “some masonry walls” (Rapp, 1986) fell. Estimating intensity as MM = IX might be too heavy a statement—we did not observe any buildings yet shifted off their foundations. Epicenter An epicenter or a long fault source has been outlined by mapping historical records of damages and their intensity (Ambraseys and Melville, 1988; Kovach and Nur, 2006). The zone of strongest shaking extends along the Dead Sea fault zone from Nablus in the south to Arqa in the north, enclosing MM = VIII+ and MM = IX intensities (Ambraseys and Melville, 1988). Their map suggests that the regular left-lateral faulting of the Dead Sea transform caused the earthquake. The Upper Jordan sector certainly moved at least 1.6 m sinistrally (Ellenblum et al., 1998). An epicenter was calculated at 34.1°N, 36.1°E (Ambraseys, 2009). Landslides and/or rockfalls in Mount Lebanon near Baalbek (Ibn Al-Jawzi, 1952), and 14C data from trenching across visible portions indicate that the Yammouneh fault was active in 1202, but not at any point since (Daëron et al., 2005). However, the tsunami damaging the eastern part of Cyprus, as reported by the Arab historian Ibn Al-Jawzi (1952), suggests an offshore earthquake instead, either as a sole event or as an event associated with the activity of inland faults. The Yammouneh and other bounding faults are less than 50 km from the offshore thrust faults yielding the transpressional uplift of Mount Lebanon (Elias et al., 2003). We allow several alternatives. Alternative 1 includes a major displacement along the Dead Sea fault

that was associated with activity of offshore faults, causing a tsunami that reached Cyprus in 1202. A similar event of higher magnitude has been identified for causing the A.D. 551 Beirut-Tripoli earthquake and tsunami offshore Lebanon (Elias et al., 2007). We prefer this scenario. We cannot exclude an alternative 2, which needs to include the direction of donjon damage besides the intensity assessment. Azimuth 240° allows us to introduce a hypothesis of a shallow thrust below the Jabal Ansariyya, breaking the surface offshore. This feature allows a local source for the earthquake, with lesser magnitude to reach MM = VIII+ and even MM = IX intensities. There is a possibility of an alternative 3, consisting of an earthquake-related submarine landslide causing the tsunami. Magnitude Ambraseys and Melville (1988) assigned a magnitude of 7.5 to the 1202 earthquake. The area suffering shaking equal or greater than modified Mercalli intensity VII (the damage threshold for many archaeological sites) in 1202 is roughly 60,000 km2 as outlined by Kovach and Nur (2006). Their cross-plot of earthquake magnitude versus area of intensity VII allows an estimate of magnitude 7.8 for this earthquake. They encircle ~60,000 km2 on their Figure 3, while assuming only 20,000 km2 when reading for a magnitude 7.6 only on their Figure 2. This seems unrealistic, because M 7.8 would need 400 km of coseismic fault rupture (Meghraoui, 2010, personal commun.). However, if we accept that Al-Marqab suffered at an intensity at least VII+, and possibly VIII, then the area of VII shaking will be significantly larger than outlined by Kovach and Nur (2006). An ~50 km northward extension of the VII shaking increases the shaken area northward by at least 50 km, increasing shaken territory to 70,000 km2, and increasing calculated magnitude to 7.9. One has to bear in mind, however, that correlation of magnitude and shaken area is very weak! CONCLUSIONS We distinguished traces of two major, successive earthquakes based on failures observed in Al-Marqab castle. Dating was conducted by historical documents and archaeological dating. Earthquake 1 consisted of vibration in SW-NE plane, damaging the donjon and room M3. It was a major event between 1187 and 1285, possibly the 1202 earthquake. Earthquake 2 consisted of vibration in NE-SW plane. It damaged the southern tower + NW corner of the donjon. It was also a major but lesser event than number 1, and it occurred after 1285. Candidates are the 1404 and 1759 events reported in Sbeinati et al. (2005). ACKNOWLEDGMENTS Our thanks are due to Tamás Borosházy, architect, SyroHungarian Archaeological Mission (SHAM) member, for providing architectural plans of Al-Marqab castle. The authors are grateful for the extremely helpful remarks of Mustapha

Distinguishing damages from two earthquakes—Archaeoseismology of a Crusader castle, Syria Meghraoui (Strasbourg, France) and of an anonymous reviewer; both improved the manuscript significantly. This study was supported by Hungarian National Science Foundation grants K67.583 and PD77.585. This is Syro-Hungarian Archaeological Mission publication no. 5. This article is a contribution to the UNESCO-funded International Geoscience Programme IGCP 567, “Earthquake Archaeology: Archaeoseismology along the Alpine-Himalayan Seismic Zone.” REFERENCES CITED Akyüz, S., and Altunel, E., 2001, Geological and archaeological evidence for post-Roman earthquake surface faulting at Cibyra, SW Turkey: Geodinamica Acta, v. 14, p. 95–101, doi: 10.1016/S0985-3111(00)01057-3. Al-Tarazi, E.A., and Korjenkov, A., 2007, Archaeoseismological investigation of the ancient Ayla site in the city of Aqaba, Jordan: Natural Hazards, v. 42, p. 47–66, doi: 10.1007/s11069-006-9045-6. Altunel, E., Stewart, I.S., Barka, A., and Piccardi, L., 2003, Earthquake faulting at ancient Cnidus, SW Turkey: Turkish Journal of Earth Sciences, v. 12, p. 137–151. Ambraseys, N.N., 2005a, Archaeoseismology and neo-catastrophism: Seismological Research Letters, v. 76, no. 5, p. 560–564, doi: 10.1785/gssrl .76.5.560. Ambraseys, N.N., 2005b, The seismic activity in Syria and Palestine during the middle of the 8th century: An amalgamation of historical earthquakes: Journal of Seismology, v. 9, p. 115–125, doi: 10.1007/s10950-005 -7743-2. Ambraseys, N.N., 2006, Earthquakes and archaeology: Journal of Archaeological Science, v. 33, p. 1008–1016, doi: 10.1016/j.jas.2005.11.006. Ambraseys, N.N., 2009, Earthquakes in the Mediterranean and Middle East: A Multidisciplinary Study of Seismicity up to 1900: Cambridge, UK, Cambridge University Press, 968 p. Ambraseys, N.N., and Barazangi, M., 1989, The 1759 earthquake in the Bekaa Valley: Implications for earthquake hazard assessment in the East Mediterranean region: Journal of Geophysical Research, v. 94, p. 4007–4013, doi: 10.1029/JB094iB04p04007. Ambraseys, N.N., and Melville, C.P., 1988, An analysis of the Eastern Mediterranean earthquake of 20 May 1202, in Lee, W.H.K., Meyers, H., and Shimazaki, K., eds., Historical Seismograms and Earthquakes of the World: San Diego, Academic Press, p. 181–200. Anonymous, 2003, Comité Européen de Normalisation, Eurocode 8: Design of Structures for Earthquake Resistance, prEN 1998-1: Brussels, Comité Européen de Normalisation, 94 p. Anzani, A., Binda, L., and Melchiorri, G., 1995, Time dependent damage of rubble masonry walls, in West, H., ed., 4th International Masonry Conference Proceedings: London, British Masonry Society, p. 341–351. Anzani, A., Garavaglia, E., and Binda, L., 2009, Long-term damage of historic masonry: A probabilistic model: Construction & Building Materials, v. 23, no. 2, p. 713–724, doi: 10.1016/j.conbuildmat.2008.02.010. Bottari, C., 2005, Ancient constructions as markers of tectonic deformation and strong seismic motions: Pure and Applied Geophysics, v. 162, p. 761–765, doi: 10.1007/s00024-004-2639-6. Burgtorf, J., 2007, Die Herrschaft der Johanniter in Margat im Heiligen Land, in Czaja, R., and Sarnowsky, J., eds., Die Ritterorden als Träger der Herrschaft: Territorien, Grundbesitz und Kirche. Ordines militares: Colloquia Torunensia Historica, v. XIV, p. 27–57. Caputo, R., and Helly, B., 2005, Archaeological evidences of past earthquakes: A contribution to the Sha of Thessaly, central Greece: Journal of Earthquake Engineering, v. 9, no. 2, p. 199–222, doi: 10.1142/S136324690500192X. Daëron, M., Klinger, Y., Tapponnier, P., Elias, A., Jacques, E., and Sursock, A., 2005, Sources of the large A.D. 1202 and 1759 Near East earthquakes: Geology, v. 33, no. 7, p. 529–532, doi: 10.1130/G21352.1. Deschamps, P., 1973, Les Châteaux des Croisés en Terre Sainte III. La Défense du Comté de Tripoli et de la Principauté d’Antioche: Paris, Paul Geuthner, 273 p. Elias, A., Tapponnier, P., Jacques, E., Daëron, M., Klinger, Y., and Sursock, A., 2003, Quaternary deformation associated with the Tripoli-Roum thrust,

197

and the rise of the Lebanese coast: Geophysical Research Abstracts, v. 5, p. 10137. Elias, A., Tapponnier, P., Singh, S.S., King, G.C.P., Daëron, M., Carton, H., Sursock, A., Jacques, E., Jomaa, R., and Klinger, Y., 2007, Active thrusting offshore Mount Lebanon: Source of the tsunamigenic A.D. 551 Beirut-Tripoli earthquake: Geology, v. 35, no. 8, p. 755–758, doi: 10.1130/G23631A.1. Ellenblum, R., Marco, S., Agnon, A., Rockwell, T., and Boas, A., 1998, Crusader castle torn apart by earthquake at dawn, 20 May 1202: Geology, v. 26, no. 4, p. 303–306, doi: 10.1130/0091-7613(1998)026<0303 :CCTABE>2.3.CO;2. Ferretti, D., and Bažant, Z.P., 2006, Stability of ancient masonry towers: Stress redistribution due to drying, carbonation and creep: Cement and Concrete Research, v. 36, p. 1389–1398, doi: 10.1016/j.cemconres.2006.03.014. Galadini, F., Hinzen, K.-G., and Stiros, S., 2006, Archaeoseismology: Methodological issues and procedure: Journal of Seismology, v. 10, p. 395–414, doi: 10.1007/s10950-006-9027-x. Galli, P., and Galadini, F., 2001, Surface faulting of archaeological relics. A review of case histories from the Dead Sea to the Alps: Tectonophysics, v. 335, p. 291–312, doi: 10.1016/S0040-1951(01)00109-3. Galli, P.A.C., Naso, J.A., Capini, S., Giocoli, A., Hailemikael, S., Piscitelli, S., Rizzo, E., and Scaroina, L., 2008, Faulting of the Roman aqueduct of Venafrum (Italy); investigation methodology and preliminary results, in 31st European Seismological Commission (ESC) General Assembly, 2008, Hersonissos, Crete, Greece, Proceedings and Abstracts, p. 94–100. Guidoboni, E., and Comastri, A., 1997, The large earthquake of 8 August 1303 in Crete: Seismic scenario and tsunami in the Mediterranean area: Journal of Seismology, v. 1, p. 55–72, doi: 10.1023/A:1009737632542. Guidoboni, E., Bernardini, F., and Comastri, A., 2004, The 1138–1139 and 1156–1159 destructive seismic crises in Syria, south-eastern Turkey and northern Lebanon: Journal of Seismology, v. 8, p. 105–127, doi: 10 .1023/B:JOSE.0000009502.58351.06. Hancock, P.L., and Altunel, E., 1997, Faulted archaeological relics at Hierapolis (Pamukkale, Turkey): Journal of Geodynamics, v. 24, p. 21–36, doi: 10.1016/S0264-3707(97)00003-3. Hinzen, K.-G., 2005, The use of engineering seismological models to interpret archaeoseismological findings in Tolbiacum, Germany: A case study: Bulletin of the Seismological Society of America, v. 95, no. 2, p. 521–539, doi: 10.1785/0120040068. Hinzen, K.-G., 2008, Can ruins indicate a back azimuth?: Seismological Research Letters, v. 79, no. 2, p. 290. Hinzen, K.-G., 2009, Simulation of toppling columns in archaeoseismology: Bulletin of the Seismological Society of America, v. 99, no. 5, p. 2855– 2875, doi: 10.1785/0120080241. Ibn Abdazzahir, 1946, Ibn Abdazzahir, Esprit du guerre sainte (translated by Jean Sauvaget), in Historiens Arabes: Pages Choisies, Traduites et Présentées: Paris, Librairie d’Amérique et d’Orient, p. 126–129. Ibn Al-Athir, 1999, ’Izz ad-Din Abu’l-Hasan ’Ali ibn Abu’l-Karam Muhammad ibn Muhammad Abu ’Abd al-Karim ibn ’Abd al-Wahid al-Shaybani, Al-kamil fi’l-tarikh [The Complete History], 11 vols., ed. ’Umar ’Abd alSalam Tadmuri, Dar al-Kitab al-’Arabiyy: Beirut, Lebanon, Dar al-Kitab al-’Arabi Publisher, 3500 p. (in Arabic). Ibn Al-Jawzi, 1952, Shams al-Din Abu’l-Muzaffar Yusuf ibn Qazawghali, Mirat al-zaman fi ta’rikh al-al-a’yan (H. 590–654): [The Mirror of Times in the History of the Notables]: Haydarabad-al-Dakan, India, Dā’irat’l-Ma’arif al-’Uthmaniyya Publisher, 250 p. Kamh, G.M.E., Kallash, A., and Azzam, R., 2008, Factors controlling building susceptibility to earthquakes: 14-year recordings of Islamic archaeological sites in Old Cairo, Egypt: A case study: Environmental Geology, v. 56, p. 269–279, doi: 10.1007/s00254-007-1162-3. Karakhanian, A.S., Trifonov, V.G., Ivanova, T.P., Avagyan, A., Rukieh, M., Minini, H., Dodonov, A.E., and Bachmanov, D.M., 2008a, Seismic deformation in the St. Simeon monasteries (Qal’at Sim’an), northwestern Syria: Tectonophysics, v. 453, p. 122–147, doi: 10.1016/j.tecto.2007.03.008. Karakhanian, A.S., Trifonov, V.G., Ivanova, T.P., Dodonov, A.E., Bachmanov, D.M., Minini, H., and Al-Kafri, A.M., 2008b, Seismic influence on archaeological objects in Syria, in 31st General Assembly of the European Seismological Commission (ESC) 2008, Hersonissos, Crete, Greece, 7–12 September 2008, p. 182–189. Karcz, I., and Kafri, U., 1978, Evaluation of supposed archaeoseismic damage in Israel: Journal of Archaeological Science, v. 5, p. 237–253, doi: 10.1016/0305-4403(78)90042-0.

198

Kázmér and Major

Kázmér, M., 2008, Earthquake damages in Al-Marqab citadel, Syria, in Török, Á., and Vásárhelyi, B., eds., Mérnökgeológia-Kőzetmechanika: Budapest, Műegyetemi Kiadó (Technical University Press), p. 159–168. Kennedy, H., 1994, Crusader Castles: Cambridge, UK, Cambridge University Press, xv + 221 p. Kilian, K., 1980, Zum Ende der mykenischen Epoche in der Argolis: Jahrbuch der Römisch-Germanischen Zentralmuseums Mainz, v. 27, p. 166–195. Korjenkov, A.M., and Mazor, E., 1999, Seismogenic origin of the ancient Avdat Ruins, Negev Desert, Israel: Natural Hazards, v. 18, p. 193–226, doi: 10 .1023/A:1026488932377. Korjenkov, A.M., and Mazor, E., 2003, Archaeoseismology in Mamshit (southern Israel): Cracking a millennia-old code of earthquakes preserved in ancient ruins: Archaologischer Anzeiger, v. 2003, no. 2, p. 51–82. Kovach, R.L., and Nur, A., 2006, Earthquakes and archeology: Neocatastrophism or science?: Eos (Transactions, American Geophysical Union), v. 87, no. 32, p. 317, doi: 10.1029/2006EO320003. Lamprecht, H.-O., 2001, Opus caementitium—Bautechnik der Römer [Cemented work—Construction technique of the Romans]: Köln, Römisch-Germanisches Museum & Düsseldorf, Bau+Technik Verlag, 264 p. (in German). Lourenço, P.B., and Roque, J.A., 2006, Simplified indexes for the seismic vulnerability of ancient masonry buildings: Construction & Building Materials, v. 20, p. 200–208, doi: 10.1016/j.conbuildmat.2005.08.027. Lourenço, P.B., Krakowiak, K.J., Ferriandes, F.M., and Ramos, L.F., 2007, Failure analysis of Monastery of Jerónimos, Lisbon: How to learn from sophisticated numerical models: Engineering Failure Analysis, v. 14, p. 280–300, doi: 10.1016/j.engfailanal.2006.02.002. Major, B., Buzás, G., and Al-Ajji, E., 2010, Excavations of the Syro-Hungarian Archaeological Mission at Al-Marqab, in Edbury, P., ed., The Military Orders, Volume 5: Oxford, UK, Ashgate (in press). Marco, S., 2008, Recognition of earthquake-related damage in archaeological sites: Examples from the Dead Sea fault zone: Tectonophysics, v. 453, p. 148–156, doi: 10.1016/j.tecto.2007.04.011. Martini, K., 1996, Masonry Damage Modes at Pompeii: http://urban.arch.virginia .edu/pompeii/masonry/masonry-1.html (accessed 6 October 2008). Mayer, H.E., 1972, Two unpublished letters on the Syrian earthquake of 1202, in Hanna, S.A., ed., Medieval and Middle Eastern Studies in Honour of Aziz Suryal Atiya: Leiden, E.J. Brill, p. 295–310. (Reprinted in Kreuzzüge und lateinischer Osten, Variorum Reprints, Leiden, 1983.) Meghraoui, M., Gomez, F., Sbeinati, R., Van der Woerd, J., Mouty, M., Darkal, A.N., Darawcheh, R., Hijazi, F., Al-Ghazzi, R., and Barazangi, M., 2003, Evidence for 830 years of seismic quiescence from palaeoseismology, archaeoseismology, and historical seismicity along the Dead Sea fault in Syria: Earth and Planetary Science Letters, v. 210, p. 35–52, doi: 10.1016/ S0012-821X(03)00144-4. Meli, R., 1998, Structural Engineering of Historical Buildings: Mexico-City, Fundación ICA, 312 p. (in Spanish). Mistler, M., Butenweg, C., and Meskouris, K., 2006, Modelling methods of historic masonry buildings under seismic excitation: Journal of Seismology, v. 10, p. 497–510, doi: 10.1007/s10950-006-9033-z. Nikonov, A., 1988, On the methodology of archaeoseismic research into historical monuments, in Marinos, I., and Koukis, G., eds., The Engineering Geology of Ancient Works, Monuments and Historical Sites, Preservation and Protection: Rotterdam, Balkema, p. 1315–1320. Nur, A., and Burgess, D., 2008, Apocalypse: Earthquakes, Archaeology and the Wrath of God: Princeton, New Jersey, Princeton University Press, 319 p. Penazzi, D., Valluzzi, M.R., Saisi, A., Binda, L., and Modena, C., 2001, Repair and strengthening of historic masonry buildings in seismic areas, in International Millennium Congress “More than Two Thousand Years in the History of Architecture Safeguarding the Structure of Our Architectural Heritage”: Bethlehem (Palestine), v. 2, Section V, 7 p.; http:// home.anadolu.edu.tr/~yguney/Cemil_historical/Historical_Buildings/ oct_2006/strengthening/binda197.pdf (accessed 1 August 2010). Rapp, G., Jr., 1986, Assessing archaeological evidence for seismic catastrophes: Geoarchaeology, v. 1, p. 365–379, doi: 10.1002/gea.3340010403.

Reicherter, K.R., Jabaloy, A., Galindo-Zaldívar, J., Ruano, P., Becker-Heidmann, P., Morales, J., Reiss, S., and González-Lodeiro, F., 2003, Repeated paleoseismic activity of the Ventas de Zafarraya fault (S Spain) and its relation with the 1883 Andalusian earthquake: International Journal of Earth Sciences, v. 92, p. 912–922, doi: 10.1007/s00531-003-0366-3. Sbeinati, M.R., Darawcheh, R., and Mouty, M., 2005, The historical earthquakes of Syria: An analysis of large and moderate earthquakes from 1365 B.C. to 1900 A.D.: Annals of Geophysics, v. 48, no. 3, p. 347–435. Sieberg, A., 1932, Untersuchungen über Erdbeben und Bruchschollenbau im östlichen Mittelmeergebiet [Research on earthquakes and tectonics in the Eastern Mediterranean]: Denkschriften der Medizinisch-Naturwissenschaftlichen Gesellschaft, Jena, v. 18, no. 2, p. 161–273 (in German). Similox-Tohon, D., Sintubin, M., Muchez, Ph., Verhaert, G., Vanneste, K., Fernandez, M., Vandycke, S., Vanhaverbeke, H., and Waelkens, M., 2006, The identification of an active fault by a multidisciplinary study at the archaeological site of Sagalassos (SW Turkey): Tectonophysics, v. 420, p. 371–387, doi: 10.1016/j.tecto.2006.03.026. Similox-Tohon, D., Sintubin, M., Muchez, Ph., Verhaert, G., Vanneste, K., Fernandez, M., Vandycke, S., Vanhaverbeke, H., and Waelkens, M., 2007, Erratum to “The identification of an active fault by a multidisciplinary study at the archaeological site of Sagalassos (SW Turkey)”: Tectonophysics, v. 435, p. 55–62, doi: 10.1016/j.tecto.2007.02.011. Sintubin, M., and Stewart, I.S., 2008, A logical methodology for archaeoseismology: A proof of concept at the archaeological site of Sagalassos, southwest Turkey: Bulletin of the Seismological Society of America, v. 98, no. 5, p. 2209–2230, doi: 10.1785/0120070178. Sintubin, M., Muchez, P., Similox-Tohon, D., Verhaert, G., Paulissen, E., and Waelkens, M., 2003, Seismic catastrophes at the ancient city of Sagalassos (SW Turkey) and their implications for seismotectonics in the BurdurIsparta area: Geological Journal, v. 38, p. 359–374, doi: 10.1002/gj.960. Sintubin, M., Stewart, I.S., Niemi, T., and Altunel, E., 2008, Earthquake archaeology—Just a good story?: Seismological Research Letters, v. 79, no. 6, p. 767–768, doi: 10.1785/gssrl.79.6.767. Stiros, S.C., 1996, Identification of earthquakes from archaeological data: Methodology, criteria, and limitations, in Stiros, S.C., and Jones, R.E., eds., Archaeoseismology: British School at Athens, Fitch Laboratory Occasional Paper 7, p. 129–152. Stiros, S.C., Laborel, J., Laborel-Deguen, F., Papageorgiou, S., Evin, J., and Pirazzoli, P.A., 2000, Seismic coastal uplift in a region of subsidence: Holocene raised shorelines of Samos Island, Aegean Sea, Greece: Marine Geology, v. 170, p. 41–58, doi: 10.1016/S0025-3227(00)00064-5. Tomaževič, M., and Lutman, M., 2007, Heritage masonry buildings in urban settlements and the requirements of Eurocodes: The experience of Slovenia: International Journal of Architectural Heritage, v. 1, p. 108–130, doi: 10.1080/15583050601126186. Van Berchem, M., and Fatio, E., 1913–1915, Voyage en Syrie: Cairo, Imprimerie de l’Institut français d’archéologie orientale en Le Caire, 303 p. Vasconcelos, G., Lourenço, P.B., Mouzakis, H., and Karapitta, L., 2006, Experimental investigations on dry stone masonry walls, in Hamid, A.A., chairman, Proceedings of the First International Conference on Restoration of Heritage Masonry Structures, Cairo, Egypt, 24–27 April 2006, p. P31-1– P31-10. Volant, Ph., Levret, A., Carbon, D., Scotti, O., Combescure, D., Verdel, T., Piant, A., and Laurent, Ph., 2009, An archaeo-seismological study of the Nîmes Roman aqueduct, France: Indirect evidence for an M > 6 seismic event?: Natural Hazards, v. 49, no. 1, p. 53–77, doi: 10.1007/s11069-008 -9276-9. Wechsler, N., Katz, O., Dray, Y., Gonen, I., and Marco, S., 2009, Estimating location and size of historical earthquake by combining archaeology and geology in Umm-El-Qanatir, Dead Sea transform: Natural Hazards, v. 50, p. 27–43, doi: 10.1007/s11069-008-9315-6.

MANUSCRIPT ACCEPTED BY THE SOCIETY 19 MAY 2010

Printed in the USA

The Geological Society of America Special Paper 471 2010

Archaeoseismological studies at the temple of Amenhotep III, Luxor, Egypt Arkadi Karakhanyan Institute of Geological Sciences of the National Academy of Sciences of Armenia, 24A Marshal Baghramyan Avenue, 0019, Yerevan, Armenia Ara Avagyan GEORISK Scientific Research Company, 24A Marshal Baghramyan Avenue, 0019, Yerevan, Armenia Hourig Sourouzian The Colossi of Memnon and Amenhotep III Temple Conservation Project, German Institute of Archaeology 31, Abu el Feda Street, Cairo-Zamalek 11211, Egypt

ABSTRACT Our studies in the temple of Amenhotep III, conducted under the project on Excavation and Conservation at Kom el-Hettan, provide new information about the seismic history of ancient Thebes. Distinct signs of liquefaction are revealed at the temple site. Trenches exhibit sand dikes and sills that formed extension cracks through the mechanism of lateral spreading. Clear effects of liquefaction by lateral spreading were discovered in other monuments on the west bank of the Nile. Application of historical, archaeological, and geological methods enables us to constrain the time of the earthquake responsible for the damage in the west bank temples to between 1200 and 901 B.C. Furthermore, we find no signs of an earthquake in 27 B.C. The foot of the Thebes Plateau may conceal a basement fault with combined vertical and horizontal slip kinematics. The fault located to the southeast, near an ancient sanctuary, may correspond to either seismogenic fault surface rupture, or a secondary seismic effect manifested as subordinate rupture and ground failure. INTRODUCTION The funerary temple of Amenhotep III is located on the west bank of the Nile River, opposite to the modern city of Luxor (Fig. 1). As the capital city of Egypt for many hundreds of years, ancient Thebes, today’s Luxor, is one of the most famous archaeological sites in the world. On the Nile’s western bank, a small area of 10–12 km2 accommodates the famous Valley of Kings, Theban necropolis, and the Valley of Queens, as well as numerous temples built by the pharaohs of the Middle and New Kingdoms (Fig. 1).

The memorial temple of Amenhotep III was one of the largest temples ever built in Egypt (Fig. 2). When completed, it included a massive array of pylons, great halls, chambers, stelae, and statues that covered an area more than 385,000 m2 (Ricke et al., 1981; Weeks, 2005). The temple’s main axis stretches ~700 m from its first pylon westward to its rear wall. Its width is estimated to be ~500 m, and, with its dependences and processional ways, it stretched between the Ramesseum and the temple of Medinet Habu through to Malqata, the vast palace of Amenhotep III. Two colossal statues of Amenhotep III, each 18 m high, known as the Colossi of Memnon, once guarded the entrance of

Karakhanyan, A., Avagyan, A., and Sourouzian, H., 2010, Archaeoseismological studies at the temple of Amenhotep III, Luxor, Egypt, in Sintubin, M., Stewart, I.S., Niemi, T.M., and Altunel, E., eds., Ancient Earthquakes: Geological Society of America Special Paper 471, p. 199–222, doi: 10.1130/2010.2471(17). For permission to copy, contact [email protected]. © 2010 The Geological Society of America. All rights reserved.

199

200

Karakhanyan et al.

A B

N

Mediterranean Sea

N CAIRO

14 ile r. N

a

ESNA

10

11

Se

9

12

QENA

ER RIZECAT

d Re

13

LUXOR

200 km

5 4 2

7

8

6

3 Fig. 2

1 0

500 m

Figure 1. (A) General tectonic settings of Egypt according to Youssef (2003); the arrow indicates the site of works in Luxor. (B) West bank of the Nile opposite to Luxor: 1—the temple of Amenhotep III, 2—temple of Ramses III (Medinet Habu), 3—location of fragments of a colossal statue of Amenhotep III, 4—temple of Merenptah, 5—tomb of Khonsuiridis, 6—temple of Ramses II (the Ramesseum), 7—temple of Tuthmosis III near the Ramesseum, 8—XX Dynasty temple, 9—temple of Tuthmosis III in Deir al-Bahari, 10—Sheikh Abd el-Qurna hill with the Theban necropolis, 11—the village of ancient Kings Valley’s builders in Deir Al-Medina, 12—sanctuary dedicated to the goddess Meretseger and to the god Ptah, 13—the Valley of Queens, 14—the Kings Valley.

the gigantic temple complex. The colossi were named by a tradition popular in the ancient world. Legend said that the northern statue of Amenhotep III that was damaged by an earthquake gave a sound at dawn with the first sun rays as if it were singing. Ancient Greek travelers claimed that sound was the cry of Memnon, a mythical Ethiopian warrior slain by Achilles in the Trojan War, to Eos, his mother and goddess of the dawn. The “vocal” statue was presumably silenced forever after its restoration, which is assumed to have taken place under the reign of Septimius Severus in the third century A.D. With ~3500 yr of history, Egyptian papyri and epigraphic sources contain almost no clear earthquake accounts. Strabo’s account about destruction of the northern of the Memnon Colossi in the Amenhotep III temple is the earliest historical record available to date that bears direct evidence of strong destructive earthquake impacts in Egypt. Strabo writes: “Here are two colossal figures near one another, each consisting of a single stone. One is entire; the upper parts of the other, from the chair, are fallen down, the effect, it is said, of an earthquake” (Strabo, 1854–1857, book 17, v. 3, chap. 1, para. 46, p. 261–262). Strabo’s visit to ancient Thebes is

dated supposedly to 24–26 B.C.; the earthquake that could have destroyed the colossus is commonly related to some earlier date of 27 B.C. (Sieberg, 1932). However, evidence of an earthquake in 27 B.C., like in the cases of other, earlier seismic events possibly responsible for the damage of ancient temples in Thebes in the Pharaonic period, is rather vague and controversial. More than a century-long intense archaeological excavation has still been unable to provide clear information about earthquakes that could have destroyed the Memnon Colossi and other temples in the region of ancient Thebes. Meanwhile, an understanding of the long-term earthquake history is an important aspect of seismic hazard assessment for the Luxor region in terms of reconstruction of historical events during the Pharaonic period, preservation of the unique historical heritage of ancient Thebes, and seismic safety of the modern city of Luxor, with its dense tourist infrastructure. Our studies of 2007–2008 in the area of the funerary temple of Amenhotep III in the framework of the “Project on Excavation and Conservation at Kom el-Hettan” were aimed at elucidating the seismic history of the Amenhotep III temple and other ancient Theban temples.

Archaeoseismological studies at the temple of Amenhotep III, Luxor, Egypt

A

201

N Fig. 9 - Tr.2

5

B 5

4

Fig. 7; 8 - Tr.1

4

3 2

3

15

0m

1

Pit 1

Fig. 10; 11

2 Tr.3

1 0

100 m

Figure 2. (A) Satellite image of the temple of Amenhotep III. (B) Reconstruction of the temple of Amenhotep III by Dr. Nairi Hampikyan: 1—the colossi of Amenhotep III (the Memnon Colossi) in front of the first pylon of the temple, 2—the colossi of Amenhotep III in front of the second pylon of the temple, 3—the pylon, 4—stelae, 5—the Great Peristyle court; Tr. 1, 2, 3 and Pit 1—locations of paleoseismological trenches and pits.

SEISMICITY IN THE LUXOR REGION: BACKGROUND EVIDENCE During the first half of the twentieth century, many seismic catalogues included information about the earthquake in 27 B.C. based on the Strabo’s report about destruction of the northern colossus (Sieberg, 1932). However, the exact date, intensity, and size of that destructive impact had been argued for a long time. Ambraseys et al. (1994) determined the earthquake of 27 B.C. as a false event, referring to Quatremere (1845), and attributed destruction of the northern Memnon Colossus to deliberate mutilation by Persians, and the damage in Thebes in 27 B.C. to a revolt of local population against Rome. This suggestion by Ambraseys and his colleagues has led to exclusion of the earthquake of 27 B.C. from the main international catalogues of historical seismicity. Some authors (Abdel-Monem et al., 2004; Casciati and Borja, 2004; Haggag et al., 2008) still identify an earthquake near Luxor in 27 B.C. Accounts about other destructive historical earthquakes in Thebes and Middle Egypt during the Pharaonic period are contained in many seismological, geological, and tourist publications. Evidence for these events is more vague than the account

of the event in 27 B.C. In a radius of 200 km from Thebes, Maamoun et al. (1984) did not identify any earthquake with a magnitude higher than 5.5 over the period from 600 B.C. to A.D. 1972, while Kebeasy (1990) related the damages in the Luxor and Karnak temples to some historical earthquakes. Based on the archaeological evidence, Dolinska (2007) and Pawlikowski (1987) suggested that the temple of Tuthmosis III in Deir al-Bahari was destroyed by an earthquake, which caused a rockfall around 1100–1080 B.C., and Badawy et al. (2006) reported that Middle Egypt suffered in historical times from six major earthquakes and that the Ramses II temple on the west bank of the Nile in Luxor was almost destroyed by an ancient event. With the abundance of accounts of strong earthquakes causing damage to ancient Theban temples, the actual evidence they provide is unclear and debatable. Haggag et al. (2008) suggested that the earthquake of 600 B.C. devastated the region of Thebes (Luxor). Sieberg (1932) also identified an earthquake in Upper Egypt dated to 1200 B.C. that damaged the Abu-Simbel temple, as well as an A.D. 1899 earthquake that toppled many columns in the Karnak temple. Kink (1979) reported that an earthquake in A.D. 1969 generated a crack that propagated through pylons I, II, and IX in the Luxor temple, damaged the basement of the

202

Karakhanyan et al.

obelisk standing between pylon III and IV, and tilted it additionally. In contrast, Ambraseys et al. (1994) suggested there were no earthquakes in 600 B.C., 1210 B.C., and 1899 A.D. Between 2200 B.C. and A.D. 1900, Youssef et al. (1994) identified just one weak earthquake with a magnitude less than 5 at a distance of 90 km to the northeast of Luxor. No strong earthquakes have occurred in the Upper Egypt region during the twentieth century. The Egyptian National Seismic Network has recorded earthquakes with magnitudes ranging from 4 to ≤2 in the region of Luxor and to the south-southeast (Youssef, 2003). At a distance of 170–190 km to the northnorthwest of Luxor, regions between Sohag and Assiut suffered three earthquakes with magnitudes between 4 and 5, in 1998, 1999, and 2003 (Hassoup et al., 2000). These earthquakes were localized in an area corresponding to the supposed epicenter of the 1778–1779 earthquake with M = 4.8 (Ambraseys et al., 1994). El-Sayed et al. (1999) estimated the peak ground acceleration in the region of Luxor and the west bank at a value of 0.04–0.05g. EVIDENCE OF SEISMIC DAMAGE ON THE COLOSSI OF AMENHOTEP III The colossal sculptures of Amenhotep III, known as the Memnon Colossi, were just two of numerous statues that decorated the funerary temple of Amenhotep III in Thebes, the capital of Egypt in the New Kingdom. Erected in the fourteenth century B.C., the colossi depict pharaoh Amenhotep III of the Eighteenth Dynasty (1391–1353 B.C.), father of the heretic pharaoh Akhenaten. The temple complex consisted of three gigantic pylons of mud brick, the innermost of which was linked to the Great Peristyle court with a processional way that was lined along either side by columns, statues, or sphinxes (Fig. 2). Colossal statues carved of quartzite and alabaster were installed in front of each pylon. Two massive steles stood between the third pylon and the peristyle (Fig. 2). The Great Peristyle court was surrounded with porticoes that rested on massive sandstone columns. Millennia ago, most of them were demolished and reused in other west bank monuments. However, many of column bases are still in their original position, marking the location of the columns. The eastern, northern, and southern porticoes included three rows of columns, and the western had four rows. Colossal statues of the king stood between the columns. The temple area as a whole was enclosed with a mud brick wall. Structures of the temple of Amenhotep III repeatedly served as a source of building material for neighboring temples, even in the Pharaonic period, starting from the funerary temple of Merenptah, a XIX Dynasty pharaoh (ca. 1212–1202 B.C.). Statues, stelae, and religious attributes in ready form were adjusted for reuse also in other temples of the west bank. Later on, the area of the temple was many times robbed and exposed by many excavations. Early in the nineteenth century, agents acting on behalf of the French and British consuls “discovered” this site as a rich source of museum-rank antiquities.

Fragments of pharaoh’s statues and sphinxes reached the British Museum, Louvre, and St. Petersburg. Many other statues from the site are listed in Egyptian antiquity collections worldwide. By the beginning of the twentieth century, the temple of Amenhotep III, unlike many neighboring temples, was in a severely damaged condition. Local residents had been cultivating this area, and it was periodically inundated with seasonal floods of the Nile. Nothing but two huge figures of the Memnon Colossi and a few structural fragments were then visible on the surface. Several archaeological missions studied the temple of Amenhotep III in the twentieth century. In 1930, Ludwig Borcherdt undertook the sounding and mapping of some parts of the Great Peristyle court and the Hypostyle hall, inclusive of the colossi lying by the northern gate. Unfortunately, his notes remain still unpublished. In the 1950s, the Department of Antiquities of Egypt restored the large stele at the entrance of the Great Peristyle court. In 1964 and 1970, the Swiss Institute of Architecture and Archeological Studies excavated a few exploration trenches. Drastic landscape changes in Luxor after construction of the Aswan Dam in the 1960s have worsened the conditions for the temple of Amenhotep III. The direct impact of seasonal floods no longer affects the ruins, as it did in the past, but higher groundwater levels have led to active agricultural encroachment. This has led to an increase in soil salinity, and hence to intense weathering and erosion of the sandstone and limestone blocks of many temple structures. In 1990–1992, a photogrammetric survey of the Memnon Colossi was conducted (Stadelmann and Sourouzian, 2001). Since 1998, the Colossi of Memnon and Amenhotep III Temple Conservation Project, under the auspices of the Supreme Council of Antiquities of Egypt and the German Institute of Archaeology, has been working on Kom el-Hettan with the aim of conserving the temple precinct. The temple area has been cleaned of vegetation; the groundwater level has been lowered in the peristyle court and kept stable. Excavations have uncovered unique statues, wall foundations, and innumerable fragments of statuary and architectural elements (Stadelmann, 1984; Sourouzian, 2004; Sourouzian and Stadelmann, 2003). Construction and Reconstruction of the Colossi in the Antiquity The Memnon Colossi might represent the most promising object to inspect when looking for evidence of strong historical earthquakes in the temple of Amenhotep III. The two statues representing Amenhotep III seated were placed at the entrance of the first pylon and are now 17 m apart (Figs. 2 and 3). Originally, the colossi and their socles were cut from monolithic quartzite blocks. Presently, both the statues and their socles are in damaged condition and tilted one to another (Fig. 3). The upper part of the northern Memnon colossus statue and the rear part of its socle have been restored with composite blocks of quartzite (Figs. 3B, 3D, and 4) different from the rock used in the monolithic components of the colossi. As attested by neutron

SOUTHERN COLOSSUS

NORTHERN COLOSSUS

A B

S

N

2

3

2.5oS

4

1.5oS

1 oN

13.97 m

14.76 m

2.5oS

R1

1o N

1

1.5oS

C D

E

W

R1

RA

F2

F3

3 0-1oW

f1

1 oW 5m

f2 10.40 m

1

1.5-2oW

1.5-2oE F1 6.20 m 10.55 m

RP

Figure 3. The colossi of Amenhotep III (the Memnon Colossi) in front of the first pylon of the temple. (A, B) Eastern sides of the colossi; (C, D) northern sides of the colossi. 1, 2, 3, and 4—places of broken feet of the colossi; F1, F2, and F3—monolithic parts of the northern colossus carved from the quartzite extracted from the quarries at Gebel al Ahmar near Cairo; RA and RP—the restored parts of the northern colossus statue and the rear part of the socle carved of the quartzite originating from the quarry in Aswan; R1—the upper course of blocks in the restored part, where cramp hollows of two types were found; f1 and f2—monolithic parts of the southern colossus socle. Arrows indicate through-going cracks splitting the monolithic part of the northern colossus statue and its socle, and the southern colossus socle; 1.5–2°E is the angle of tilting of parts of the statues and socles.

204

Karakhanyan et al. SOUTHERN COLOSSUS 7m 7.97 m

N 5.78 m

NORTHERN COLOSSUS

6.67 m concrete

6.98 m

4m

19 cm 74 cm

5.15 m

4

5

2

3

1

2

3

4

1

Figure 4. Outline of the socles supporting the colossi of Amenhotep III (the Memnon Colossi) in front of the first pylon of the temple. 1—monolithic parts of the socles cut of the quartzite extracted at the quarries of Gebel al Ahmar, 2—rear section of the northern colossus socle restored with quartzite blocks from the Aswan quarry, 3—the rear section of the southern colossus socle broken into pieces, 4—through-going cracks splitting the socles of the northern and southern colossi.

activation and petrographic analyses conducted at Berkeley University, the monolithic portion of the northern Memnon colossus was cut from quartzite extracted from quarries in Gebel al Ahmar, near Cairo, which were situated more than 600 km north of the temple of Amenhotep III (Heizer et al., 1973; Stadelmann, 1984). The blocks utilized to rebuild the torso and the rear part of the socle of the northern colossus originated from a quarry in Aswan, which is located 200 km south of the temple. Bowman et al. (1984) confirmed the conclusions of Heizer et al. (1973). It is still unclear when the upper part and the socle of the northern colossus were restored. One very tentative hypothesis is that this could have happened during the rule of Roman emperor Septimius Severus around A.D. 199 (Sourouzian et al., 2006). This hypothesis is based on the observation that graffiti describing the “singing” of the statue was no longer made on it after A.D. 199 and on the tradition stating that the statue was silenced after restoration. Four courses of blocks are laid across the restored torso and head of the northern colossal statue (Figs. 3B and 3D). As reported by Sourouzian et al. (2006), hollows from anchor cramps utilized to fasten the blocks in the past were found in all of the restoration rows, and two different types of the cramp hollows can be distinguished in the upper and lower rows (Fig. 5). Type a hollow is a relatively narrow, ~18–20-cm-long groove carved in stone to a depth of 2–3 cm. This cramp hol-

low is 3 cm wide in its narrower part, but widens to 5–6 cm toward the ends, creating a swallow-tail shape (Fig. 5A). Vertical holes were carved on the wider ends to a depth of 4–5 cm. This type of hollow was certainly intended for metal cramps, commonly used by the Romans in Egypt (Kink, 1979). Copper alloy cramps still remaining in torso blocks are reported by Sourouzian et al. (2006). Cramp hollows of type b had different design: they were large, 27–30 cm long, and carved 5–7 cm deep. The width of such cramp hollow was 8–9 cm and 16–17 cm in the narrower and wider parts, respectively. Compared to type a cramp, a typical b cramp has sizably wider ends, shaping a dove-tail contour (Fig. 5B). Cramps of this design were largely applied by the Egyptians during the Pharaonic period (Kink, 1979), but their larger size excluded use of metal. They were prepared of stone or very hard wood and could be 40 × 20 cm large. Some blocks in the restored upper part of the northern colossus have holes from both types of the cramps (Fig. 5C), but it seems unlikely that they could have been applied concurrently. Anchors of type a fastened blocks 1, 2, 3, and 4 (Fig. 5C) firmly, and there was no reason to spend much time to carve additional hollows for type b anchors. Moreover, some of type b hollows are filled with lime mortar. Similarly, cramps of type a would not be necessary, if the blocks had been already linked with cramps of type b.

Archaeoseismological studies at the temple of Amenhotep III, Luxor, Egypt

205

B

A 5

4

20 cm

1

5

20 cm

C 5 1

4

2

Figure 5. Hollows remaining from the anchor cramps affixing blocks in the upper course R1 of the restored part of the northern colossus statue. (A) Cramp type a commonly used by the Romans in Egypt. (B) Type b used by the Egyptians during the Pharaonic period. (C) Combination of type a and type b cramps on the same blocks in the upper course of the restored part of the northern colossus statue. Numbers 1–5 correspond to block numbers in the course of stones.

3

1m

It is unlikely also that masons reutilized blocks of some older structure already bearing type b hollows to prepare blocks for the upper part of the colossus (1, 2, 3, and 4) because the axes of type b hollow counterparts in both pairs are oriented strictly opposite one to another both vertically and horizontally. We suggest that the cramp hollows of type b and type a are likely related to two different, respectively earlier and later episodes of restoration of the upper part of the northern colossus. The earlier restoration with stone cramps could have been made in the time of pharaohs. During the Roman restoration, the holes from the Pharaonic-type anchors were filled with mortar, and metal alloy cramps were used instead. A graffito in Greek runs undistorted across one of the two type b hollows on the left foot of the northern colossus (1 in Figs. 3B and 3D). Therefore, the cramp hollow should be older than this graffito. This inscription was carved when Emperor Hadrian visited the place on November 21, A.D. 130 (Bernard and Bernard, 1960). This is additional evidence in favor of a Pharaonic-period restoration episode preceding the presumed

Roman restoration in A.D. 199. The statues of queens standing by the sides of the throne have no feet: they were broken, and the broken surfaces bear clear signs of restoration, with vertical cramps and stone-working technique typical for the Pharaonic period (Fig. 3). Considering traditions of construction of memorial temples for pharaohs, it seems rather unlikely that the anchor holes on the feet of the colossus and the queen statues were part of the original design or a repair of damage caused during transportation of the statues. Such defects would not be allowed for the central statues of a living pharaoh, suggesting that most probably they were not related to the phase of design and installation. Quartzite blocks in the upper part of the northern Memnon colossus have marks from working with both bronze tools typical for the Pharaonic period, and trident iron chisels common in the Roman era. In addition, processed surfaces of some blocks appear eroded by long-term contact with soil. Indirectly, these observations also attest to two episodes of restorations of the upper part of the northern colossus—in the Pharaonic period and later, during the Roman rule.

206

Karakhanyan et al.

Analysis of Damage on the Colossi The most severely damaged is the northern Memnon colossus. Its upper part and the rear section of the socle were destroyed and replaced with blocks of quartzite extracted from a different quarry (Figs. 3 and 4). The preserved monolithic part of the throne is split by a through-going crack into two parts (F2 and F3). The crack cuts the monolithic part of the throne and extends into the socle as the boundary between its monolithic and composite sections made of a few separate blocks (Figs. 3 and 4). The rear section of the throne (F3) rests on the monolithic socle and on the restored blocks 1, 4, and 5 (Fig. 4) and is rotated 4°–6° counterclockwise. The statue of the southern colossus appears less damaged and bears no traces of restoration with individual blocks, but its socle is strongly damaged. A through-going crack splits it into two parts (f1 and f2 in Fig. 3), while the rear section has crashed into pieces, some of which are not preserved (Fig. 4). On both colossal statues, the frontal sections of the feet and fingers are broken off and have not been preserved. Higher sections of the feet and lower shins are split with through-going cracks (Fig. 3). In a similar manner, the feet have broken off from the figures of queens standing on both sides of each throne of the colossi. The lower rear parts of the throne on both colossal statues once resting on the socle are broken, and the corner sites of the thrones have suffered the greatest damage. Blocks 1–5 in the rear part of the northern colossus socle were installed during restoration, which would have been an infeasible task if the rear part of the colossal statue (F3) had stood in its place and rested on those blocks. Hence, either the rear part F3 was slightly raised to replace the blocks in the course of restoration, as suggested by Bowman et al. (1984), or it was toppled backward, and then raised and remounted over the repaired part of the socle. We suggest the sculptors restored the upper section of the torso (RA) and the rear section of the socle (blocks 1–5), and then they lifted and remounted the lower part of the colossal statue (F3) over the repaired part of the socle. However, these interpretations of the reconstruction of the statue are in contradiction with the historical accounts. Strabo’s account says that the upper part of the northern Memnon colossus could have been destroyed by an earthquake, but other sources ascribe the damage to warfare (Pausanias, 1898; Quatremere, 1845; Ambraseys et al., 1994). The account of Pausanias dated to the middle of the second century B.C., states:

This made me marvel, but the colossus in Egypt made me marvel far more than anything else. In Egyptian Thebes, on crossing the Nile to the so called Pipes, I saw a statue, still sitting, which gave out a sound. The many call it Memnon … This statue was broken in two by Cambyses, and at the present day from head to middle it is thrown down; but the rest is seated, and every day at the rising of the sun it makes a noise, and the sound one could best liken to that of a harp or lyre when a string has been broken. (Pausanias, 1898, book 1, v. XLII, p. 64–65)

Both the accounts of Strabo and Pausanias indicate clearly that the upper part of the colossus is broken, while the remaining statue (including F2 and F3) remains seated. In case the rear part of the throne of the northern colossus (F3) was in its place by the middle of the second century B.C., we infer that the restoration of socle blocks had been made earlier. Such conclusion is consistent with the presence of two types of cramp holes on the blocks of the upper part of the colossus and with other evidence of its restoration in the Pharaonic era (Fig. 5). Like in the case of cramps of a and b types, we assume two episodes of restoration in response to strong damage of the northern colossal statue and socle in the past. The first episode could be apparently related to the Pharaonic period, when damaged parts of the socle, feet of the colossus and queen statues, and the entire upper section of the statue were replaced, attesting to largescale destruction. Later, the upper part of the statue was again destroyed and repaired for the second time, utilizing blocks that had been preserved from the first restoration. The second restoration of the statue could have taken place in the Roman epoch, around A.D. 199. Reconstruction of Possible Seismic Impact It is possible that the destruction of the northern colossus statue and of both socles was caused by an earthquake impact and the restoration was made in an effort to repair the seismic damage. However, the statues could also have been damaged by humans. The northern colossus was damaged already by the time Strabo visited it between 24 and 26 B.C. The damage of this statue is often attributed to the invasion of the Persian army of King Cambyses in 525 B.C. Pausanias wrote about this in the second century B.C. (Pausanias, 1898). Strabo also mentions that Cambyses ruined many sanctuaries in Thebes (Strabo, 1854– 1857). Quatremere (1845) concluded that deliberate dissection of the northern colossus by Persians is the most credible cause of destruction. Ambraseys et al. (1994) agreed that Persians could have damaged the colossus and referred the large-scale destruction in Thebes in 27 B.C. to the revolt of local population against Rome mentioned by Eusebius (1846). Human aggression would primarily cause damage to the upper part of the statues of the northern and southern colossi, an effect most predictable when battering rams, hand tools, or rope toppling efforts are inferred. This may be consistent with the observed damage of the upper parts of the southern colossus and complete destruction of this upper section of the northern colossus. Faces and torso of these statues might have been mutilated also in later historical periods, e.g., during the Coptic and Islamic settlement in this area, and tradition even attributes it to the invasion of the Napoleon’s army. In contrast, it would be more difficult to assign the damage of the rear parts of the socles on both statues to man-made actions. Furthermore, the cracks that split the northern colossus, its socle, and the socle of the southern colossus are unlikely to have been caused by humans.

Archaeoseismological studies at the temple of Amenhotep III, Luxor, Egypt When an earthquake hits large-block monolithic monuments, their lowermost corner parts are damaged first (Sinopoli, 1995). These parts are destroyed on both statues of Amenhotep III: the feet of the northern colossus and of the statues of queens standing beside the throne are broken off (Figs. 3 and 4). The corner section F3 of the northern colossus throne was broken off entirely, while most severely damaged corners RC on either side of the statue were replaced with inset triangular patches (Fig. 3). The feet and rear section of the southern colossus throne were broken in the same manner. The most damaged sections on the lower part of the colossal statues are the rear part of the northern colossus throne and the feet of the southern colossus. Rear parts of socles on both colossi are also damaged. A possible model of destruction of the feet and rear sections of the thrones of the colossal statues and their socles in a seismic event is illustrated in Figures 6A and 6B. The throughgoing crack separating the lower monolithic parts (F2 and F3) of the northern colossus also dissects the monolithic and composite parts of the socle (F1 and RP in Fig. 3 and Fig. 4). This crack likely formed when the socle split and moved laterally. This offset led to the fracturing of the statue itself. Hence, formation of the main crack was most probably contemporary with the splitting of the socle (Fig. 6B). The observed inclination of socle part F1 to the west and part RP to the east by 2° alike is additional supporting data that the socle broke along line F1/RP. Casciati and Borja (2004) modeled the response of the southern colossus of Amenhotep III to a possible seismic impact by means of three-dimensional analysis of soil-foundation-structure interaction (SFSI). The results of the modeling show that the rear socle section and rear base of the statue experience the greatest stress and deformation. The analysis of Casciati and Borja

A

(2004) indicated that no distribution of seismic stresses would be capable of damaging the upper torso of the statue by itself. On the contrary, the model shows that the statue torso remains comparatively stable and is not destroyed until destruction of the lower part of the socle and of the statue itself. Similar results are reported by Sinopoli (1995). Therefore, we conclude that aggressive human actions are unable to explain the damage of the rear sections of socles on the northern and southern colossi, the formation of throughgoing cracks splitting almost along a single surface the northern statue, its socle, and the socle of the southern statue, as well as the backward falling of the rear section of the northern colossus. Destructive impact of a strong earthquake has been shown by modeling (Sinopoli, 1995; Casciati and Borja, 2004) to produce this damage pattern. The parallel orientation and opposite positions of the through-going cracks in the socles and statues of the northern and southern colossi suggest a single zone of deformation, possibly reflecting a parallel differential subsidence of the soil. As we describe in the following, a zone of lateral spreading and extension of liquefied soil induced by a strong earthquake likely caused the fracture pattern in the colossal statues. An analysis of restorations on the northern colossus allows us to suggest two episodes of repairs—one in Pharaonic and one in Roman time. Considering that the earlier stage of restorations was related to the replacement of socle blocks, reinstallation of the fallen rear section of the statue, and other, comparatively minor repairs, we suggest that it was most probably an effort to eliminate effects of a strong earthquake. It is still difficult to judge whether the presumed Roman-time restoration in A.D. 199 was undertaken in response to damage by an earthquake and whether it happened in 27 B.C.

B

S

C

207

Figure 6. Simplified model of possible destruction of the northern (Memnon) colossus from an earthquake. (A) Initial shaking impact stage. (B–C) Final stages of formation of the zone of lateral soil spreading, cracking of the socle and of the rear part of the statue, failure, and destruction of the rear part of the statue.

208

Karakhanyan et al.

PALEOSEISMOLOGICAL STUDIES IN THE AREA OF THE TEMPLE OF AMENHOTEP III Soil Conditions Early investigations of the soil conditions beneath the Memnon colossi suggested that the socles rested on limestone bedrock (Jollois and DeVilliers, 1821; Wilkinson, 1835). Wiedemann (1884) described the colossi as standing on an inhomogeneous soil or sand capable of softening. Later models (Verdel, 1993; El Shabrawi and Verdel, 1994; Casciati and Borja, 2004) suggest that the soils beneath the colossi include 6 m of alluvial silt and silty clay overlying siltstone and limestone. Boreholes drilled to the depth of 15–20 m in the temple area found soils of typical alluvial sediments of the floodplain of the Nile (MISR Laboratory, 2007). The sequence of layers from the surface to the borehole bottom includes silty clay, clayey silt, sand, and silty clay. Interlayers and lenses of fine, water-saturated sand are recorded within and in between of the three lower horizons. The upper 10.5 m of the top 15–20 m of the sediment include silty clay and clayey silt deposited by seasonal flooding of the Nile River. A 5- to 9-m-thick water-saturated layer of fine sand lies below 10.5 m depth. The lowermost layer in the boreholes (10.5–20 m) consists of dense, hard silty clay. The groundwater table lies at ~2–3.5 m. Using the empirical granulometry curves of Tsuchida and Hayashi (1971) for conditions of soil liquefaction, the substrata at the temple site are liquefiable. Trench 1 in Front of the Main Peristyle Court

7 cm

Trench 1 (Fig. 2) is located near the platform, 12 × 6 m in size, paved with stone slabs, on which the northern stele was mounted. In this trench, we documented the southern wall (1.1) and the northern wall (1.2), which are 3.5 m apart. In the base of the southern wall 1.1, there is a layer consisting of a mixture of gravel and lime mortar (d in Fig. 7). A few dikes of Trench 1-1, southern wall

NW

20 cm

17 cm

SE

fine sand, inclined to the northwest, are observed above this layer (c in Fig. 7) and intrude into the higher layer of dense clay (b in Fig. 7). We interpret the sand dikes as features of soil liquefaction. The stone slabs of the man-made platform located above the clay layer are deformed, so that the northwestern edge of the platform is 20 cm lower than the other parts, and the blocks are crashed (Fig. 7). A system of branching cracks dissects all three rows of blocks in the platform beneath the basement of the stele that fell in the eastern direction (Fig. 7). The fractured stone blocks appear to be displaced horizontally by 10–12 cm. Originally, the blocks were fastened with type b cramps. We recorded lateral chips in the edges of the cramp hollows, attesting that the blocks were pulled apart so that the anchor cramps were torn away from the affixed stone blocks. The vertical fracturing and horizontal extension in the stone blocks, as well as surface deformation of the platform, are all limited to the sand dike intrusion area only. Pieces of ceramics found in the layer of gravel mixed with lime mortar (d) and also in the dikes of liquefied sand were identified as remains of pottery dated to the Late XVIII Dynasty of the New Kingdom (1382–1295 B.C.). This period corresponds to the rule of pharaohs Amenophis III and Horemheb. The age of samples was estimated by the radiocarbon method at the Laboratory of Radiocarbon Dating of the IFAO (French Institute of Oriental Archaeology), Cairo, Egypt, in 2008. Sample 9 (cal. age of 1517–1188 B.C.), sample 10 (cal. age of 1500–1122 B.C.), and sample 19 (cal. age of 1524– 1188 B.C.) provide ante-quem dates of the liquefaction, because the sand dikes appear to crosscut the dated layers d and b (Fig. 7). We have not established a post-quem date of liquefaction on the southern wall. The northern wall 1.2 in trench 1 provides additional evidence for paleoliquefaction (Fig. 8). Similar to the southern wall, it exposes NW-inclined dikes of fine sand. A small mushroomshaped dike that originates in the fine sand layer d appears to terminate in the loam layer c. A sandstone block apparently limited further spreading of the dike upward (B in Fig. 8). The layer

Figure 7. Trench 1 near the stele, the southern wall (1.1). 1—modern soil, 2—cracked stone blocks of the platform on which the stele was installed (layer a), 3—interlayers of soil enriched with silt between the blocks, 4—lime mortar with debris, 5—dense clay (layer b), 6—sand dikes and sills (layer c), 7—leveled layer of lime mortar (layer d), 8—14C sampling locations, 9—ceramics sampling locations.

1

a

2 3

b

4 5 6

10

9

c 1

19

2

4

3

3/8

d 7

7

5 6

8

8 9

1m

Archaeoseismological studies at the temple of Amenhotep III, Luxor, Egypt

dikes that are 2–5 cm wide and up to 25 cm in length (Fig. 9B). The sand dikes do not penetrate layer b. The lowermost layer d includes sandy loam with a high percent of sand (50%–60%) and abundant ceramic sherds. Two large fractures (F in Fig. 9) in the center of trench 2 break through layers c and b. Debris of destruction layer b fills the crack and provides a post-quem date of the liquefaction event. To the south, the flank of layer c is displaced by 30 cm down along fractures F (Fig. 9). At a distance of 10–15 m to the southeast, layers b and c are again broken and displaced 30–40 cm down along a few cracks. The gap between them is filled with fragments of the destruction layer. Ages of the layers in trench 2 were determined on soil samples by radiocarbon analysis and on archaeological ceramics. Radiocarbon ages were established for six samples (Table A1). Samples 1, 2, 15, and 16 taken from layers d and c provide an ante-quem date of the formation of fractures F and displacement of layer c (sample 7 is considerably younger than cal. 1530– 1252 B.C.), and sample 18 must establish a post-quem date (cal. age 1266–8962 B.C.).

capping the dike (a) is represented by sediments rich in organics deposited by seasonal floods of the Nile River. Ceramic sherds were found in the sand of the dikes and in the overlying layers of loam as well. Samples 11, 12, and 13 were taken from the northern wall 1.2 of trench 1. Sample 11 (cal. age 1562–1251 B.C.) and sample 12 (cal. age 1211–830 B.C.) may constrain an ante-quem date of the liquefaction event because the dikes crosscut the dated layer c. Sample 13 (cal. age of 766–396 B.C.) could provide a post-quem date because it is bedded stratigraphically over layer b (Fig. 8). The latter served as the floor of some structure and did not suffer any deformation; it belongs to a period much later than the time of construction of the temple of Amenhotep III. Trench 2 to the Northeast of the Peristyle The exposure in trench 2, located near the northeastern corner of the peristyle (Fig. 2), shows a section with four layers (a–d, Fig. 9). The upper layer a corresponds to the soil enriched with organics brought by seasonal floods of the Nile. The underlying deposit, layer b, has been called the “destruction layer” by archaeologists. Layer b contains fragments of the destroyed temple structures. The lower layer c is subdivided into a watersaturated clayey sand and sand. The upper boundary of layer c displays convolute bedding (Fig. 9). Layer c contains vertical sand

Trench 3 and Pit 1 near the Second Pylon Trench 3 is located at the entrance of the second pylon of the temple (Fig. 2). The section of trench 3, from top to bottom, includes layer a of soil enriched with silt and organic sediments

Trench 1, northern wall

NW

A

209

SE

13

12

C

10 cm

Figure 8. Trench 1 near the stele, the northern wall (1.2): (A) photo of the trench, (B) trench log, and (C) close-up view of mushroom-shaped dike. 1— brecciated sandstone fragments of the stone block over the mushroom-shaped dike, 2—organic-rich soil (layer a, sediments of the seasonal floods of the Nile), 3—dense clay, presumably corresponding to the floor of some structure (b), 4—dense loam (layer c), 5—sand dikes and sills (d), 6—14C sampling locations, 7—ceramics sampling locations.

11

B 0.5 m

a

13

b 1 2

a

3

b

4

c

5

d

6 7

12

c

C 11 d 0.5 m

210

Karakhanyan et al.

A

NW

B

Trench-2 D

C

SE

a Figure 9. Trench 2 northeast of the peristyle. (A–C) Photos of the trench and (D) trench log. 1—organic-rich soil layer a (seasonal flood sediments of the Nile), 2—destruction layer (b) composed of temple structure debris, 3— dense loam (layer c1), 4—loam containing up to 50%–60% of sand (layer c), 5—14C sampling locations, 6—ceramics sampling locations.

18

b 7

17

F

Convoluted level

c1 Convoluted level

F c

16

15 1

1

2

6

5 2

1m

c

d

5

3

4

from seasonal floods of the Nile (Fig. 10) and destruction layer b of sandstone debris bedded beneath. Similar to trench 2, layer b may have been generated by earthquake damage. Materials from this layer may have been recycled for construction of other temples in the Pharaonic period, although the evidence of this is conjectural. Layer c is composed of dense compact clay from older deposits of the Nile River. The sand dike with inclusions of rounded fragments of silicified limestone and pieces of ceramics crosscuts layer b. A mushroom-shaped liquefaction feature is capped by layers that span >8 m. The surface of this feature is slightly depressed toward the center. The groundwater table in the trench matches the depth of the mushroom cap. Ceramics found in layers c and b and in the sand dike provide age constraints of the liquefaction event. Calibrated age estimates were obtained for five samples of paleosol (no. 4–8) collected from trench 3 (Table A1). Samples 4, 5, and 7 can constrain an ante-quem date of the liquefaction (sample 7 gives a cal. age of 1386–976 B.C.) because they belong to layer c, which is crosscut by the dike. It is more difficult to judge about the position of sample 6 (cal. 1310–924 B.C.) because it is not finally clear whether the surface of layer c was deformed. Most probably, it was deformed, and layer b was distorted after. In such case, the age of sample 8 (cal. 1306–901 B.C.) could provide a post-quem date or even constrain the date of the event. Trench 3 is located 5 m away from the two fallen statues of pharaoh Amenhotep III. They were similar to the Memnon

5

6

Colossi, but were smaller in size (2 in Fig. 2). Both statues were knocked down from their socles, rotated counterclockwise, and lie on the ground oriented 154°–166°SE with their right side down (Fig. 11). The head and some fragments of the northern statue were lying separately, but thrown away in the same direction. Pit 1 was excavated beneath the fallen northern statue (1 in Fig. 11). The ceramic pieces found in it were sampled and dated by the radiocarbon method to 1290–1226 B.C., an age presumably corresponding to the time when the statue fell (Annales du Service, 2006). The radiometric age of sample 3 taken from the layer in which the ceramics was found was estimated at 1392–995 B.C. by 2σ or 1270–1054 B.C. by 1σ. These dates are similar to the radiometric estimates of the ante-quem dates of the liquefaction event (cal. 1267–1048 B.C. by 1σ or 1310–924 by 2σ). Therefore, we suggest that the effects of liquefaction were related to a strong earthquake that toppled both statues in the same direction. Causes of Soil Liquefaction Soil liquefaction effects in unconsolidated sedimentary environments can be often generated not by seismicity, but such phenomena as floods, artesian settings, landslides, and others (Holzer and Clark, 1993; Li et al., 1996). The seismic origin of the features of liquefaction we identified in the area of the Amenhotep III temple still needs to be confirmed. Next, we provide a few indications in support of the supposed seismic origin.

Archaeoseismological studies at the temple of Amenhotep III, Luxor, Egypt

211

A Trench-3

a

B

NE

a

SW

b

A

c

a

d

74 m

b 8

2 6 73 m

1

5

7

c

c1

c1

d

2

3

4

c

5

6

7

4

1 1m

Figure 10. Trench 3 at the second pylon. (A) Photo of the trench and (B) trench log. 1—soil rich in silt and organics brought by seasonal floods of the Nile (layer a), 2—destruction layer (b), 3—dense clay above (layer c) and below (c1) the dike, 4—sand dike (layer d), 5—14C sampling locations, 6—ceramics sampling locations, 7—elevation mark of 74 m above sea level.

Trench 3 2(N154o) 1(N166o)

N

Pit 1 2 1

Figure 11. Archaeological mapping of the fallen colossi at the second pylon of the temple of Amenhotep III. 1 and 2— the fallen colossi, 3 and 4—pedestals of the colossi.

4 3

0

At a distance of 400 m to the N-NE from the Amenhotep III temple, we found a man-made hollow in an old Nile terrace thought to be a boundary of the palace area of Amenophis III (5 in Fig. 1). The terrace is built by paleo-Nile sediments of alternating horizontal layers of densely consolidated clay and

2m

rounded gravel. Some other Theban monuments such as the tomb of Khonsuiridis, and temples of Tuthmosis IV and Ramses II are located on the same terrace. A system of multiple extension cracks dissects the vertical wall of this 4–5-m-high man-made hollow (Fig. 12A). The

212

Karakhanyan et al.

B

Nile River

A k r

a

b

c

2m

C River

Figure 12. (A–B) System of vertical extension cracks filled with sand (indicated with arrows) in the wall of a man-made hollow of the paleo-Nile terrace at the supposed boundary of the palace area of Amenophis III. k—tomb of Khonsuiridis; r—Ramesseum; a, b, and c—crater-like hollows filled with sand, possibly corresponding to sand blows. (C) Lateral spread model (according to Youd, 1984).

cracks are mostly vertical, open from 5 to 30 cm wide, and expose minor vertical offsets of the clay and gravel layers by 30–35 cm. We recorded common plunge of the eastern wings of the cracks toward the Nile River. All cracks are filled with fine sand containing individual pebbles and displaying clear vertical zoning. Sand of this kind is found nowhere on the surface, and hence filling of the cracks from the surface can be ruled out. Moreover, on the exposure wall, we observed many blind sand-filled fractures not propagating to the surface. This fact, along with the density and vertical zoning of the sand inside the cracks and their strike perpendicular to the slope gradient, i.e., to water erosion direction, does not support the surface-filling option either. An outcrop found about 100 m south of the described hollow, on the west margin of the main road, contains fine sand deposited beneath the layer of clay and gravel in the paleo-Nile terrace that is completely similar to the sand found in the cracks. Therefore, the cracks intruded this sand from the underlying stratum by splitting layers of clay and gravel of the paleoterrace.

The drilling at the site of the temple of Amenhotep III confirms that the layer of fine sand occurs at a depth of 7–8 m, under the layer of loam and clay. The cracks strike 130°–140° and can be also traced on the top surface of the terrace as far as 10–25 m away from the edge of the hollow. The pattern illustrated in Figure 12A is a clear demonstration of typical lateral spreading (Fig. 12C) commonly caused by liquefaction during strong shaking (Youd, 1984). Accordingly, liquefaction and deformation of the underlying layer can generate vertical cracks in the upper soil, and split it into separate blocks, which then move horizontally or in the slope gradient direction (i.e., toward the river). The blocks deform and tilt to different sides, which repeatedly amplifies the destructive seismic effect and leads to strongest damage of the overlying structure. This effect likely facilitated the severe destruction of the temple of Ramses II (r in Fig. 12B), which is just 170 m north of the extension cracks. Meanwhile, the tomb of Khonsuiridis of the

Archaeoseismological studies at the temple of Amenhotep III, Luxor, Egypt seventh century B.C., located right over the cracks, suffered no damage from them (k in Fig. 12B). This monument was built of mud brick, and an ~5–6-m-high pylon has been preserved from it. Our inspection in the tomb of Khonsuiridis indicates that foundations of structures east of the pylon are absolutely undamaged, despite the fact that they were built right above an open crack filled with sand. When outcropping on the ground surface, the vertical cracks (a, b, and c in Fig. 12A) formed craters on the terrace, which is yet another observation supplementing and supporting the suggested pattern. The origin of these features from erosion by rainwater flowing down from the terrace must be excluded considering that some of these craters are still filled with dense packed sand or are 0.5–1 m away from the scarp edge that is prone to this erosion. In addition, with appreciable eastward and northward inclination of the terrace slope, crater-like hollows are recorded on the opposite southern slope. Altogether, these observations allow us to suggest that the crater-like hollows could correspond to original sand blows generated by earthquakes as a consequence of liquefaction and spouting of sand from the underlying layer through the opened fractures. We observed similar patterns of deep fractures filled with sand and gravel ascending from the lower strata in many other places, including areas of the temple of Tuthmosis III, XX Dynasty temple, and the village of Deir Al-Medina. Fronabarger (2002) described sand-filled vertical cracks in tombs TT 72 and TT 121 at the Sheih Abd el Qurna hill. A few fractures in trench 2 at the site of Amenhotep III temple that are filled with fragments of the destruction layer can also be explained by spreading effects. The fractures all strike in similar directions. The liquefaction and spreading effects detected over an area of 2.2 km2 were everywhere associated with extensive damage of the ancient temples. The most important features of the recorded structural damage are deformations of subsidence and tilting, as well as numerous cases of rotation of large blocks and falling of colossal statues that can be attributed to earthquake impacts only (Fig. 13). The estimated dates of liquefaction and falling of the statues at the second pylon of the Amenhotep III temple are similar. In many cases, we observed that the features of liquefaction and spreading were located in areas 50–60 m higher by elevation than the highest recorded boundary of seasonal floods of the Nile River, and therefore could not be affected even by the highest flood. The three rows of stone blocks in the platform located near trench 1 in the area of the Amenhotep III temple could have been deformed by intrusion of liquefaction dikes. Near the tomb of Khonsuiridis, temple of Tuthmosis III, XX Dynasty temple, and the village of Deir Al-Medina, firmly cemented sediments of the Nile terrace are dissected with sand-filled cracks that bear evidence of a sudden and short-term pulse of great hydraulic force exerted in the upward direction. Therefore, we suggest that the seismic origin of the observed effects of liquefaction is attested by the regionwide development of the liquefaction features and diversity of their forms (sand cra-

213

ters, dikes, sills, and spreading) in the context of seismic damage of the temples, lack of settings suitable for aseismic liquefaction, and indications of short-term pulse of tremendous vertical hydraulic force. Chronological Constraints In trenches 2 and 3, near the peristyle and the second pylon, respectively, destruction layer b is deposited above the layers deformed by soil liquefaction. This layer consists of sandstone and limestone fragments found nowhere near the temple. The destruction layer is clearly an anthropogenic layer. The petrographic study of fragments in the destruction layer shows that they consist of the same limestone and sandstone used for construction of structures in the temple of Amenhotep III. The archaeological interpretation is that the layer was deposited at the time of pharaohs, when masons were recycling material from structures in the temple of Amenhotep III. The so-called destruction layer might correspond to a colluvium formed as a consequence of both earthquake damage of temple structures, and construction debris from a building phase. The age and structural relations of the destruction layer with other strata may be important for determination of the date of past earthquakes. In trench 2, layer c, located below the destruction layer b, is cut by fractures. The lower boundary of destruction layer b fills the space between two fractures, while the upper boundary is even and has no deformations (Fig. 9). A possible explanation is that the upper boundary was leveled by people, and as a result does not show signs of cracking or lateral extension. If, however, the destruction layer had already existed by the time of the earthquake, the fractures would have necessarily broken through b and would have been noticeable inside it. We observe no continuation of the cracks inside the destruction layer. Hence, it is possible to conclude that the destruction layer formed after the earthquake. We can additionally support this inference based on the following observation. Many unconsolidated fragments of temple structures in destruction layer b are aligned so that their longer axes are horizontal with respect to layer c and the ground surface. If the earthquake and associated spreading cracks occurred later than the accumulation of the destruction layer, the latter’s fragments would have been chaotically filled in the vertical gap caused by the two fractures. The trench wall should have shown signs of such chaotic filling of the extension fracture with debris, but careful inspection of the wall does not record such a pattern. In contrast, the fragments in the lower part of the gap are parallel to the ground surface, suggesting a relatively regular, likely anthropogenic infilling of the cracks (Fig. 9). Most probably, people used the wreckage to fill in the crack generated by the spreading, and the space freed after layer c lowered by 30 cm, trying to even and rebuild the temple floor after an earthquake. This could also explain the leveled surface of the upper boundary of layer b. Besides, upper corners of layer c are eroded on both fractures shoulders, and the bottom of the separating gap is

A

B

C

D1

c

D2

b

d a

D3

D4

E1

E2

Figure 13. Evidence of damage in other temples on the west bank. (A) An architrave on the northern side of the Second Court in the Medinet Habu temple. (B) Nilometer located on the southern side of the Medinet Habu temple. (C) Clockwise rotation of the path (a–b) leading to the nilometer in the Medinet Habu temple and its consequent restorations (c and d). (D) Deformations of the first pylon in the Ramesseum: (D1 and D2) rotations of blocks on the opposite flanks of the pylon in the Ramesseum, (D3 and D4) wavelike bend of the central part of the pylon. (E1–E2) Fallen colossal statue of Ramses II in the Ramesseum.

Archaeoseismological studies at the temple of Amenhotep III, Luxor, Egypt filled with material of this erosion. The aggregate of these findings shows that the destruction layer formed after the earthquake. The fractures originated in layer c, which is composed of easily eroding loams. If the interval between the cracking and the filling of cracks with destruction material had been large, erosion on the edges of the cracks would have been much more considerable. Therefore, we may suggest that the filling of the gap with destruction layer material followed shortly after the earthquake, and the age of destruction layer can be regarded a close postquem date of the seismic event. According to this interpretation, it is critical to determine the age of the destruction layer. Trench 2 (Fig. 9) allows us to do this by considering age difference between samples 17 and 18, which constrain the formation of the destruction layer to a period after 1530–1252 B.C. (calibration by 2σ is applied everywhere) and before 1266–896 B.C. However, we cannot date the destruction layer earlier than construction of the temple of Amenhotep III (1382–1344 B.C.). In this regard, trench 3 may appear most informative because its sample 6 and sample 7 (Fig. 10), taken just beneath the lower boundary and in the center of the destruction layer, respectively, constrain its age between 1310–924 and 1386–976 B.C. The age estimate of sample 8 (1310–976 B.C.) for the destruction layer can be used as a post-quem date, since the earthquake and the layer must have similar ages, and it is confirmed by sample 18 from Trench 2 (1266–896 B.C.). Sample 12 from trench 1 and sample 6 from trench 3 provide the best ante-quem dates of the earthquake. Table 1 summarizes all possible earthquake dates inferred by radiometric method for all trenches. Sand dikes and sills exposed by the trenches in the temple of Amenhotep III, as well as the layers deposited below and above, yielded abundant ceramics, which attest to a historical age of the earthquake that generated widespread liquefaction in this area. The chronological analysis of ceramics sampled in the trenches was conducted by Dr. David Aston, who concluded that none of the samples contained any later material than the period of Amenophis III–Horemheb (1382−1295 B.C.).

215

Trench 1, excavated under the fallen statue of northern colossus at the second pylon, provided ceramics dated to 1290– 1226 B.C. (Annales du Service, 2006). Considering that failure of both colossi near the second pylon was most likely caused by an earthquake, the age of ceramics extracted from beneath the statue can be regarded as an ante-quem limit. Sample 3 taken from the same layer had a radiometric age estimate of 1392–995 by 2σ or 1270–1054 by 1σ. Therefore, we use the generalized values of radiometric estimates, ages of ceramics, and the dates of construction of the damaged temples to conclude that the earthquake happened between 1211 and 901 B.C. INFERRED INTENSITY, MAGNITUDE, AND EPICENTRAL DISTANCE According to the world’s statistics, the threshold magnitude capable of producing isolated liquefaction effects of limited scale is 4.5, and perfect soil liquefaction features can develop starting from magnitude 5.5 and higher (Ishihara, 1985; Ambraseys, 1988; McCalpin, 1996). Differential soil subsidence caused by liquefaction and spreading effects in the temple of Amenhotep III is ~30 cm, while the mushroom-shaped cap of the sand dike near the second pylon is more than 8 m long. Craters that were presumably formed by sand blows at the terrace opposite to the temple of Ramses II are 1–1.2 m in diameter. The length of multiple spreading cracks may range up to 30–50 m near the temple of Tuthmosis III and ~100 m near the XX Dynasty temple. The zone where lateral spreading cracks are recorded stretches parallel to the bank of the Nile from the tomb of Khonsuiridis to the XX Dynasty temple over a total distance of at least 1200 m. On the west bank, we recorded many features of soil liquefaction and lateral spreading over an area of at least 2.2 km2. The destruction of the temple of Tuthmosis III in Deir el-Bahari mentioned by Pawlikowski (1987) and Dolinska (2007) and fault motions damaging KV14 (Tausert-Setnakht) in the King Valley according to Cobbold et al. (2008) expand the affected area to 5 km2,

TABLE 1. EARTHQUAKE DATES INFERRED FROM THE PALEOSEISMOLOGICAL INVESTIGATIONS Calibrated age (B.C.) Earthquake date (B.C.) Earthquake date (B.C.)* Earthquake ante-quem date Earthquake post-quem date 1σ 2σ 1σ 2σ 1σ 2σ † † 1 1088–905 746–688 766–396 Between 1088 and 688 1211 –830 Between 1211 and 396 Sample 12 Sample 13 Sample 13 Sample 12 § § 2 1458–1370 1530–1252 1130–972 1266–896 Between 1458 and 972 Between 1530 and 896 Sample 17 Sample 17 Sample 18 Sample 18 † § † 3 1267–1048 1386–976 1208–977 Between 1267 and 977 1306–901 Between 1386 and 901 Sample 7 Sample 7 Sample 8 Sample 8 § Pit 1 1290–1226* 1392–995 N.D. N.D. After 1270 After 1392 1270–1054 Sample 3 Sample 3 *The age of ceramics from the fallen statue near the second pylon. † The inferred earthquake date is between 1211 and 901 B.C. by 2σ. § Because multiple damages are recorded in the temple of Amenhotep III, the earthquake cannot be dated to a period preceding construction of the temple in 1382–1344 B.C. Trench

216

Karakhanyan et al.

while the evidence of damage in the temples over the west and east banks documents a total area of as much as 36 km2 for soil liquefaction and destruction. According to the INQUA (International Union for Quaternary Research) (Medvedev, Shponkhoer, and Karnik) scale (Michetti et al., 2004), soil liquefaction of this size can originate with earthquake intensity ranging from VIII to X on the MM or MSK (Medvedev, Shponkhoer, and Karnik) 64 scales. In our case, earthquake intensity most probably can be estimated at IX. Ambraseys (1988), Papathanassiou et al. (2005), Youd (1984), and Kuribayashi and Tatsuoka (1975) have proposed various relations linking liquefaction intensity, causative earthquake magnitude, and epicentral distance. The minimum magnitude sufficient to generate multiple liquefaction effects on the scale compatible with what we recorded in the study area ranges between 5.5 and 6.5. The epicenter could have been located from 5 to 50 km away from the site where liquefaction features were concentrated (Ambraseys, 1988). If the epicenter was farther, inferred earthquake magnitude would have been obviously higher. The region of Luxor is located 150 km and 250 km away from the west coast and the central part of the Red Sea, respectively. The zone of the Red Sea opening represents the lithospheric boundary between African and Arabian plates, and it is certainly capable of generating strong earthquakes (Maamoun et al., 1984; Ambraseys et al., 1994; El-Sayed et al., 1999). However, considering how far Luxor is from the Red Sea, the earthquake responsible for the widespread liquefaction effects should have had a magnitude in the range of 7.3–7.7 (Ambraseys, 1988). As an alternative hypothesis, the earthquake could have been associated with the Kalabsha fault zone or other seismogenic structures near the Aswan Dam, located 50−170 km south of Luxor. Here, an earthquake of 1778–1779 with a magnitude of 4.8, and another three seismic events with magnitudes ranging from 4 to 5 occurred in 1998, 1999, and 2003 to the northnorthwest of Luxor, between Sohag and Assiut. However, if this was the epicentral area of the earthquake responsible for liquefaction on the west bank of the Nile opposite to Luxor, the 150–170 km distance implies a magnitude in the range of 7.3–7.5 (Ambraseys, 1988). Damages caused by a M 7.3–7.7 earthquake would have had to be extremely severe, covering a vast region from Alexandria to Abu Simbel. Destruction of such proportion should have been reflected in ancient structures and left evidence in historical sources, while there seems to be no evidence of this kind for the period around 1200–901 B.C., except for an earthquake that supposedly destroyed Abu Simbel in 1210 B.C. as suggested by Sieberg (1932), though disregarded by Ambraseys et al. (1994). In conclusion, we reject the hypothesis of distant and strong events and suggest a proximal localization of a 6.0–6.5 earthquake within a distance of 5–50 km from the Nile’s west bank area opposite to Luxor. If so, the search for active faults capable of generating earthquakes of this size should be confined to the Nile canyon faults.

FAULTS ON THE WEST BANK OF THE NILE OPPOSITE TO LUXOR The temple of Amenhotep III is located on the west Bank of the Nile, at the boundary of two morphological units—the modern and paleo-floodplains of the Nile, bordered by the Sheikh Abd el Qurna hills behind (8 in Fig. 14). West of the temple and the hills, a rocky massif forms a natural barrier between the Western Desert and the Nile Valley. The frontal side of the massif is dissected by many wadi, among which are the famous Kings Valley and the Valley of Queens. The stratigraphy and lithology of the west bank are well studied (e.g., Said, 1990; Pawlikowski, 2001; Fronabarge, 2002; among others). The oldest formation in the west bank is up to ~15 m thick and includes the late Paleocene limestone and the Tarawan chalk overlain with 60 m of Esna shales (Said, 1990; Pawlikowski, 2001). These sedimentary deposits are covered by the Lower Eocene limestone of the Theban Formation, which is up to 290 m thick (Said, 1990; Pawlikowski, 2001; Fronabarge, 2002). Unconformably overlying the bedrock on the Sheikh Abd el Qurna hill, there is a conglomerate bed (Said, 1990). A narrow strip of Pliocene deposits of the paleo-Nile, consisting of marl, clay, and sand, stretches from Esna to Qena along the west bank (Fig. 1A). The central part of the Nile valley represents a graben bounded on either side by large normal faults (RIGW, 1997). A thick sequence of Nile River sediments of Holocene and Pleistocene age fills the graben and discordantly overlies the Pliocene clay and shales of the paleo-Nile (Ismail et al., 2003; RIGW, 1997). According to Said et al. (2000), the principal Kings fault is roughly oriented N-S along the west side of the Kings Valley and has an estimate of up to 30 m of maximum displacement along the fault, the average value being much less. Unfortunately, Said et al. (2000) did not mention anything about the kinematics and chronology of the fault, but they did document the occurrence of many minor faults and shallow open fractures in the tomb of Ramses III and upslope from it, which are destroying the burial chambers. They also recorded a normal fault in tomb KV-5. Cobbold et al. (2008) identified normal and strike-slip faults in the Kings Valley as evidence of principally NE-SW–oriented extension. This stress field is responsible for earthquake generation at present. Some of these faults are observed in the tombs of Ramesses VI (KV9) and also in Tausert-Setnakht (KV14), where fault displacement likely occurred after the tomb construction in the historical time (Cobbold et al., 2008). Said et al. (2000) suggested that faults and fractures in the Kings Valley formed 20,000 yr ago as a result of seismic activity. In the area of Deir Al-Bahari (9 in Fig. 1), Pawlikowski and Wasilewski (2004) identified a system of jointing along with many generations of tectonic faults that have deformed and crashed the rocks. The faults are discontinuous. Cobbold et al. (2008) suggested that Theban lowlands correspond to a vast landsliding area, considering that the arcuate form of the Theban cliffs, concaved toward the Nile, is typical of

Archaeoseismological studies at the temple of Amenhotep III, Luxor, Egypt

217

N

a 5 6

7 QV 13

SMP

8 9 4

10

12

3

11 2

0 1

500 m 2

3

4

b

1

Figure 14. Faults on the west bank of the Nile opposite to Luxor. 1—listric faults, 2—scarps associated with the leftlateral transtensional flower structure, 3—lateral spreading zones, 4—soil liquefaction. Numbers in the figure indicate: 1—temple of Amenhotep III, 2—tomb of Khonsuiridis and the temple of Ramses II (the Ramesseum), 3—temple of Tuthmosis III near the Ramesseum, 4—the XX Dynasty temple, 5—the Kings Valley, 6—site on the path from the Kings Valley with recorded change of bedding angle in the Thebes Formation, 7—changed angle of bedding in the Thebes Formation near the guard post, 8—site west of the Sheikh Abd el-Qurna hill with a listric sliding plane shown in Figure 17, 11—fault at the Valley of Queens entrance, shown in Figure 15, 9 and 10—eastward extension of the fault, 12 and 13—westward extension of the same fault, QV—the Valley of Queens, SMP—fault near the sanctuary to the goddess Meretseger and to the god Ptah shown in Figures 16 and 17; a–b—line of the section shown in Figure 18.

an upper landslide scar. We conducted reconnaissance field surveys in search for evidence of any active faults on the west bank of the Nile opposite to Luxor. Despite commonly horizontal bedding of the Theban limestone, it is observed to tilt NW in many places. Fronabarge (2002) and Cobbold et al. (2008) recorded such tilt and related it to the listric sliding and block slumping from the El Qurna hill. Between the Valley of Queens and the hill of Sheih Abd el Qurna, we analyzed limestone of the Theban Formation and in many places recorded bedding angles dipping from 15° to 70° to the northwest or to the west. In addition to changes in the bedding dip, disharmonic folding of the limestone was observed. The tilt of limestone strata is most clear along a path leading from the Valley of Queens to the village of Deir al-Medina (11 in Fig. 1, Fig. 14). Closer to the top of the southern slope, beds are inclined 45°–70°NW (6 in Fig. 14). The bedding is again horizontal down along the slope up to the guard post, but with farther steep dips of the beds traced up to the western slope of the depression accommodating the village of Deir al-Medina (7 in Fig. 14). In both cases of dip change, the inclined bedding is consistent with two clear fault scarps that are traced over a distance of 3 km from the south-southwestern edge of the Valley of Queens to the northeastern flank of the Sheikh Abd el-Qurna hill (Fig. 14).

At the foot of the El-Qurna hill (8 in Fig. 14), we observed a gently dipping fault plane, which, along with the subvertical scarp located higher on the slope, creates a typical listric geometry. The presence of listric faults is well illustrated by relation of dip angles in the bedding of Esna Formation on the western and eastern margins of the village of Deir al-Medina. Therefore, the west bank has two or more listric slide surfaces, the southeastern wings of which were displaced by normal faulting as confirmed by preliminary evidence of Cobbold et al. (2008). Large slumped blocks that moved along the listric faults are present in the Valley of Queens, in the village of Deir Al-Medina, and on the El-Qurna hill. On the southeastern slope of a small hill opposite to the Valley of Queens’ entry, we observed a clear open ditch oriented to the NE (11 in Fig. 14; Fig. 15). The ditch developed in the Theban limestone but cuts through the upper layer of slope colluvium likewise. Two fault planes located by the sides of the ditch are inclined toward one another with the Esna formation intruded in between. This pattern may bear evidence of a flower-type structure of the fault plane. The subvertical wall of the scarp exposes distinct tectonic striation (Fig. 15). We analyzed and measured nine striations along the main scarp, concluding that just two of them a contain near-vertical displacement component, while the remaining seven

218

Karakhanyan et al. N

NW

SE

10 9 4 6

3

8

1 2

5 7

11

D

A

B

C

20 cm

Figure 15. (A) Fault scarp at the entrance of the Valley of Queens (11 in Fig. 14). (B–C) Examples of oblique-slip striations on the scarp. (D) Schmidt lower-hemisphere projection of measured striated fault planes.

correspond to left-lateral strike-slip motions. The fault scarp shown in Figure 15 can be traced over a distance of 2.5–3 km (9–13 in Fig. 14). The sanctuary dedicated to the goddess Meretseger and to the god Ptah is located 320 m to the northeast, inside the fault zone (10 in Fig. 14). Originally, this sanctuary was a structure carved in the rock. Presently, its roof and outer walls are almost completely collapsed. The fragments of the fallen ceiling and walls lie in the same place (Fig. 16). The sanctuary was mainly carved in the Theban limestone, but on the southwestern flank of the structure, we record a layer of Esna shales that are in tectonic contact with the Theban limestone. Fault plane striation (f1 in Fig. 16) indicates left-lateral strike-slip displacements. In the same outcrop, there is a secondary fault striking N-S and dipping 70°–80° to the west (f2 in Fig. 16). Our preliminary examination of the sanctuary suggests that it could have been destroyed by reactivation of these structures. Landslide or flood impact can be ruled out considering that the sanctuary to the god Ptah is located on a gentle hillslope. Moreover, the character of damage excludes an anthropogenic cause, while strong earthquake is the most likely cause of destruction. Arrows on a detailed view (Fig. 17) point to fault plane F2 dissecting the entrance of the sanctuary. Pieces of plaster and a drawing painted red (b and c, Fig. 17) have been preserved on the remaining structure of the sanctuary entrance A. It is possible that the motion along rupture f2 was generated by an earthquake and destroyed the entrance. The date of construction of the sanctuary dedicated to the goddess Meretseger and to the god Ptah is related

to a period between 1200 and 1150 B.C. Therefore, this interval can provide yet another ante-quem date of the earthquake. The offset along fault f2 that destroyed the entrance bears clear signs of a normal kinematics with the downthrown western wing. The offset was oriented transversely with respect to the main fault f1; hence, for the latter, we may suggest a vertical reverse-fault component in combination with a left-lateral strikeslip component of displacement. Development of listric slumps can result from active motion along a concealed, deep basement fault with subvertical plane and reverse-slip components (e.g., Naylor et al., 1994). Walters and Thomas (1982) demonstrated that high-angle reverse basement faults can develop under intense horizontal stresses associated with the strike-slip or oblique-slip tectonics. Applying the models suggested by Naylor et al. (1994) and Koyi and Skelton (2001), we infer the presence of an obliqueslip basement fault at the foot of the Thebes Plateau, along the paleo-Nile shore line (Fig. 18). During an earthquake, basement fault motions at depth could activate listric faults on the slopes of the Thebes Plateau. The fault located on the southeastern slope of the hill, opposite to the entrance of the Valley of Queens, and the sanctuary dedicated to the goddess Meretseger and to the god Ptah, could be a surface manifestation of a deep basement fault motion (Fig. 18) and could have generated the earthquake that destroyed the ancient Theban temples between 1200 and 900 B.C. However, continued investigation is necessary to resolve the character of the subsurface structure definitely.

Archaeoseismological studies at the temple of Amenhotep III, Luxor, Egypt

Te

f2 f1

Es

Fig.17A

1m

Figure 16. Sanctuary dedicated to the goddess Meretseger and to the god Ptah (10 in Fig. 14). Te—Thebes Formation, Es—Esna Formation, f1—the main fault, f2—secondary rupture.

219

ity at depth of the major basement fault. The geodynamics of the region, characterized by the collision between Africa and Eurasia, and the Red Sea opening, imply compression in the northwest-southeast direction. Hence, we may infer reverse-fault and left-lateral strike-slip components of motion for the supposed basement fault stretching along the line of Er Rizecat–Qena (Fig. 1A). Such a fault mechanism is supported by the evidence we collected for the fault located at the Valley of Queens entrance and by the data of Cobbold et al. (2008) for the Kings Valley. The strip of deformed beds emphasizing the occurrence of the basement fault from Er Rizecat to Qena spans 65–70 km, a length sufficient to generate a M = 6.5 earthquake. There is some vague evidence attesting to location of a basement fault with vertical displacements and a downthrown and tilted eastern block also under Wadi Qena on the east bank of the Nile. In the case where the Wadi Qena fault is a western extension of the Er Rizecat–Qena fault, Mmax estimates would be much higher. CONCLUSION

A

f2

b c

B

b c

a

a 1m

Figure 17. (A) Entrance to the southern room in the sanctuary dedicated to the goddess Meretseger and to the god Ptah. (B) Detail of the entrance destroyed by motion on the secondary rupture f2. a—remains of plaster colored with red paint, b—stele constructed on the fault plane, and c—other stone stelae with inscriptions and paintings.

A strip of Pliocene deposits steeply tilted 45°–70° to the northwest is recorded along the western bank of the Nile, from Er Rizecat to Qena. Distinct cirques of slump blocks have developed in the rear of the tilted beds. Apparently, the high-dipping setting has concealed any surface deformations related to activ-

The archaeoseismological investigation of the Amenhotep III temple on the west bank of the Nile River reveals extensive damage attributable to strong earthquake effects. These findings are most impressive for the famous Memnon Colossi that stood at the first pylon of the temple. Trenches and exposures near the temples of Amenhotep III, Ramses II (Ramesseum), Tuthmosis III, and the XX Dynasty temple reveal signs of liquefaction and horizontal soil spreading that correlate with serious damage of the temples. It is not possible to refer this damage to human aggression or any natural phenomenon other than that of an earthquake. Damages attributable to earthquake effects have been found also in other temples of the west bank. Our preliminary investigation indicates that similar effects could be established also for the temples of Karnak and Luxor on the east bank of the Nile. The colossi damaged in the temple of Amenhotep III and in the Luxor temple bear signs of restorations referable to the time of pharaohs. Also, the type of damage of these statues allows one to imply earthquake impact. Structural damage in the sanctuary of the goddess Meretseger and the god Ptah can be also ascribed to motion along the secondary rupture plane, associated with the main strike-slip fault. To determine the date of the earthquake, we applied data of radiometric estimations, ages of the ceramic, and temple damage analysis. The bulk of analytical evidence demonstrated no internal age conflicts and showed good convergence of dates derived from different methods. This use of the historical, paleoseismological, and archaeological methods allows us to conclude that the earthquake that destroyed the Theban temples on the west and possibly east banks of the Nile occurred between 1200 B.C. and 900 B.C. In contrast, we did not find any evidence of earthquake in 27 B.C. Moreover, the two episodes of restoration of the northern Memnon Colossus, one related to the Pharaonic period

220

Karakhanyan et al.

A NW

b

a

Te

Fig. 15

QV Te

Es

+ Tr Basement faults

B 1 2 Te 3 Es 4 Tr 5

and one to the Roman era, suggest that another earthquake could have happened. Effects of soil liquefaction and lateral spreading during the earthquake are recorded over an area of 2.2 km2 on the west bank, stretching from the village of Deir al-Medina to the XX Dynasty temple. The area of recorded earthquake damage in the temples is almost 5 km2 large. Temple damage evidence collectively for the west and east banks of the Nile may extend the area of destruction and soil liquefaction up to 36 km2. Soil subsidence rates range up to 30 cm, craters possibly formed by sand blows are 1–1.2 m in diameter, and the sand dikes are more than 8 m long. Individual lengths of the spreading cracks vary in the range from 30 to 100 m, while the total length of zone of cracking is at least 1200 m. According to the INQUA scale (Michetti et al., 2004), the earthquake intensity capable of producing soil liquefaction of this scale can be estimated at IX, and magnitude of the causative earthquake will then be not less than 6.0–6.5. Our field studies on the west bank of the Nile suggest that an oblique-slip basement fault could be located at the foot of the Thebes Plateau along the shoreline of the paleo-Nile. During the earthquake, motions along the basement fault concealed at depth could have activated the listric slump faults on the slopes of the plateau. The fault possibly located on the southeastern slope of the hill opposite the entrance of the Valley of Queens at the sanctuary dedicated to the goddess Meretseger and to the god Ptah may correspond to either direct surface evidence of the coseismic rupture generated by an earthquake, or its secondary effect in the form of subordinate rupture and ground failure. It is possible that these effects resulted from the same earthquake responsible for the destruction in the temples of ancient Thebes between 1200 and 900 B.C. A strip of the paleo-Nile Pliocene deposits consisting of clay, marl, and sand, steeply tilted 45°–70° to the northwest, is recorded along the western bank of the Nile, from Er Rizecat to Qena. Distinct cirques of slump blocks have developed in

Putty ridge 0

Figure 18. (A) Schematic section along line a–b shown in Figure 14, 1—Quaternary alluvium, 2—Thebes Formation (Te), 3—Esna Formation (Es), 4—Taraway Formation (Tr), 5—listric sliding, QV—Queen Valley. (B) Modeling of motion along an oblique-slip fault according to Naylor et al. (1994).

3 cm

Basement fault

the rear of the tilted beds. Apparently, the high-dipping setting has concealed any surface deformations related to the basement fault activity at depth. The strip of deformed beds demarcating the basement fault spreads over 65–70 km, a length sufficient to generate a M ≥6.5 earthquake. There is plausible evidence to suggest extension of the basement fault toward Wadi Qena, in which case the estimate of Mmax of the fault must be much higher. It is obvious that seismic hazard assessments for the Luxor region available for today and resulting in values of Mmax = 4.7–5.2 (Fat-Helbary et al., 2008) and 0.04–0.05g (El-Sayed et al., 1999) need to be revised. The studies we conducted in 2007–2009 are preliminary and should be continued with further archaeoseismological, paleoseismological, and geological investigations in Luxor. Such studies may help elucidate many details of seismic history for the temple of Amenhotep III (and its Memnon Colossi in particular), and for other ancient Theban temples, and provide valuable results for updated seismic hazard assessment for Luxor and improve conservation and protection of the historical heritage, population, and the tourist business. ACKNOWLEDGMENTS The studies were accomplished under the Project on Excavation and Conservation at Kom el-Hettan. We want to thank Rainer Stadelmann, Nairi Hampikyan, and the project team for advice and help in the implementation of this study. We are grateful to Zbigniew Szafranski, Herve Philip, and Miroslaw Barwik for valuable information and also to Yelena Abgaryan, Suren Arakelyan, and Arshaluis Mkrtchyan for their support in the preparation of this chapter. This article is a contribution to the United Nations Educational, Scientific and Cultural Organization– International Geoscience Programme IGCP 567, “Earthquake Archaeology: Archaeoseismology along the Alpine-Himalayan Seismic Zone.”

Archaeoseismological studies at the temple of Amenhotep III, Luxor, Egypt

10

Site

Trench 1, southern wall

Sample number 9

221

TABLE A1. RESULTS OF RADIOCARBON AGE ESTIMATION FOR TRENCH LAYERS Unit Material Calibrated age (B.C.) 1σ 2σ B Silty clay 1442–1266 1517–1188 B

Silty clay

1430–1260

1500–1122

D

1450–1269

1524–1188

C

Charcoal at the silty clay level with abundant ceramics Silty clay

1497–1370

1562–1251

C

Silty clay

1088–905

1211–830

C

Silty clay

746–688

766–396

1

D

2040–1890

2144–1866

2

D

2139–1650

2231–1882

15

D

Charcoal, in the clayey level with abundant ceramics Charcoal, clayey level with abundant ceramics Clayey level with abundant ceramics

2151–2011

2287–1901

16

C

Si lt y c l a y

1688–1530

1771–1453

Soil, upper part o f the silty clay le vel with abundant ceramics Cl a y

1458–1370

1530–1252

1130–972

1266–896

1

Silty clay

1408–1208

1452–1048

1

19

13

17 18

Trench 2

12

Trench 1, northern wall

11

1

C A C

05

C

Silty clay

1416–1261

1455–1128

06

C

Si lt y c l a y

1212–1013

1310–924

C

Si lt y c l a y

1267–1048

1386–976

C

Soil from the destruction la yer

1208–977

1306–901

1270–1054

1392–995

07 08

Trench 3

04

Pit

03

Under the northern colossus of the Clay under the wedge of altered second pylon quartzite

REFERENCES CITED Abdel-Monem, S.M., Sakr, K., Hassoup, A., Mahmoud, S., Tealeb, A., AlIbiary, M., and Mansour, M., 2004, Crustal deformation measurements and seismicity of the middle part of the Nile valley in Egypt: Egyptian Geophysical Society (EGS) Journal, v. 2, no. 1, p. 123–133. Ambraseys, N.N., 1988, Engineering Seismology: Earthquake Engineering and Structural Dynamics, v. 17, p. 1–105, doi: 10.1002/eqe.4290170101. Ambraseys, N., Melville, C., and Adams, R., 1994, The Seismicity of Egypt, Arabia and the Red Sea: Cambridge, UK, Cambridge University Press, 181 p. Annales du Service des antiquités de l’Egypte, 2006, Institut français d’archéologie orientale du Caire, Maslahat al-Athar, v. 80: Cairo, p. 323– 365. Badawy, A., Abdel-Monem, S.M., Sakr, K., and Ali, Sh.M., 2006, Seismicity and kinematic evolution of Middle Egypt: Journal of Geodynamics, v. 42, p. 28–37, doi: 10.1016/j.jog.2006.04.003. Bernard, A. and Bernard, E., 1960, Les Inscriptions Grecques et Latines du Colosse de Memnon: Centre National de la Recherche Scientifique, Institut français d’archéologie orientale: Cairo, 267 p. + 73 p. of plates. Bowman, H., Stross, F.H., Asaro, F., Hay, R.L., Heizer, R.F., and Michel, H.V., 1984, The northern Colossus of Memnon: New slants: Archaeometry Oxford, v. 26, no. 2, p. 218–229, doi: 10.1111/j.1475-4754.1984 .tb00336.x. Casciati, S., and Borja, R., 2004, Dynamic FE analysis of south Memnon Colossus including 3D soil-foundation-structure interaction: Computers & Structures, v. 82, p. 1719–1736, doi: 10.1016/j.compstruc.2004.02.026.

Comments Ante-quem date Ante-quem date Ante-quem date Ante-quem date Ante-quem date Post-quem date Ante-quem date Ante-quem date Ante-quem date Ante-quem date Ante-quem date Post-quem date Ante-quem date Ante-quem date Post-quem date Ante-quem date Post-quem date Ante-quem date

Cobbold, P., Watkinson, J., and Cosgrove, J., 2008, Faults of the pharaohs: Geoscientist, v. 18, no. 6 (available at http://www.geolsoc.org.uk/gsl/site/ GSL/lang/en/page3994.html). Dolinska, M., 2007, Temples at Deir el-Bahari in the New Kingdom 6, in Haring, B., and Klug, A., eds., Aegyptologische Tempeltagung. Funktion und Gebrauch Altaegyptischer Tempelraeume: Wiesbaden, Germany, Leiden, p. 4–7. El-Sayed, A., Vaccari, F., and Panza, G.F., 1999, Deterministic Seismic Hazard in Egypt: Miramare-Trieste, Italy, Abdus Salam International Centre for Theoretical Physics, p. 32. El Shabrawi, A., and Verdel, T., 1994, The seismic risk on ancient masonry structures studied by the use of the distinct element method. Application to an Egyptian monument, in Vasco Fassina, V., Ott, H., and Zezza, F., eds., La conservation dei monumenti nel bacino del Mediterraneo [The conservation of monuments in the Mediterranean basin]: Proceedings of the 3rd International Symposium, Venice, 22–25 June 1994: Venice, p. 373–381. Eusebius, 1846, Hieronimi interpretatio chronicae Eusebii Pamphili [The First Book of the Chronicles of Eusebius Pamphilus], in Migne, J.-P., ed., Patrologiae Cursus Completus, Series Latina, vol. 27. Fat-Helbary, R.E., Khashab, H.M., Dojcinovski, D., El Faragawy, K.O., and Abdel-Motaal, A.M., 2008, Seismicity and seismic hazard analysis around the proposed Tushka new city site, Egypt: Acta Geodynamica et Geomaterialia, v. 5, no. 4 (152), p. 389–398. Fronabarger, A., 2002, Conservation Report, 2001: Preliminary Report on the Geology and Structural Stability of Three Theban Tombs: TT 72, TT 121 and MMA 850: Serapis Research Institute and the University of

222

Karakhanyan et al.

Charleston, Charleston, South Carolina, Theban Tombs Publication Project, Tombs No. 72 (Ray) and 121 (Ahmose), p. 16. Haggag, H.M., Gaber, H.H., Sayed, A.D., and Ezzat, M., 2008, A review of the recent seismic activity in the southern part of Egypt (Upper Egypt): Acta Geodynamica et Geomaterialia, v. 5, no. 1 (149), p. 19–29. Hassoup, A., Othman, A., and Sakr, K., 2000, Intensity and micro-seismicity with the December 14, 1998, Sohag earthquake, Egypt: Al-Azhar Bulletin of Science, v. 10, no. 1, p. 283–298. Heizer, R.F., Stross, F.H., Hester, T.R., Albee, A., Perlman, I., Asaro, F., and Bowman, H., 1973, The Colossi of Memnon revisited: Recent research has established the source of the stone of the two 720-ton statues at Thebes: Science, v. 182, no. 4118, p. 1219–1225, doi: 10.1126/science .182.4118.1219. Holzer, T.L., and Clark, M.M., 1993, Sand boils without earthquakes: Geology, v. 21, p. 873–876, doi: 10.1130/0091-7613(1993)021<0873:SBWE>2.3 .CO;2. Ishihara, K., 1985, Stability of natural soil deposits during earthquakes, in Proceedings of the 11th International Conference on Soil Mechanics and Foundation Engineering: Rotterdam, the Netherlands, Balkema, v. 1, p. 321–376. Ismail, A., Neil, L., Anderson, J., Rogers, D., Abbas, M.A., and Atekwana, E.S., 2003, Hydrogeophysical Investigation at Luxor, Southern Egypt: Eos (Transactions, American Geophysical Union), v. 84, Fall Meeting supplement, abstract H31B-0452, p. 1–17. Jollois, J.B.P., and DeVilliers, R.E., 1821, Description de l’Egypte, Volume 2: Paris, Pankoucke, 69 p. Kebeasy, R.M., 1990, Seismicity, in Said, R., ed., The Geology of Egypt: Rotterdam, the Netherlands, A. A. Balkema, p. 51–59. Kink, Kh.A., 1979, The Ancient Egyptian Temple: Moscow, NAUKA, Chief Editorial Office of Oriental Literature, 102 p. [in Russian]. Koyi, H.A., and Skelton, A., 2001, Centrifuge modeling of the evolution of low-angle detachment faults from high-angle normal faults: Journal of Structural Geophysics, v. 23, p. 1179–1185. Kuribayashi, E., and Tatsuoka, F., 1975, Brief review of liquefaction during earthquakes in Japan: Soil and Foundation, v. 15, no. 4, p. 81–92. Li, Y., Craven, J., Schweig, E., and Obermeier, S., 1996, Sand boils induced by the 1993 Mississippi River flood: Could they one day be misinterpreted as earthquake-induced liquefaction?: Geology, v. 24, no. 2, p. 171–174, doi: 10.1130/0091-7613(1996)024<0171:SBIBTM>2.3.CO;2. Maamoun, M., Megahed, A., and Allam, A., 1984, Seismicity of Egypt: Bulletin of the Helwan Institute of Astronomy & Geophysics, v. 4, p. 109–160. McCalpin, J.P., ed., 1996, Paleoseismology: San Diego, Academic Press, 587 p. Michetti, A.M., Esposito, E., Gurpinar, A., Mohammadioun, B., Mohammadioun, J., Porfido, S., Rogozhin, E., Serva, L., Tatevossian, R., Vittori, E., Audermand, F., Comerci, V., Marco, S., McCalpin, J., and Morner, N.A., 2004, The INQUA scale: An innovative approach for assessing earthquake intensities based on seismically induced ground effects in natural environment, in Vittori, E., and Comerci, V., eds., Memorie Descriptive della Carta Geologica D’Italia, v. 67, 116 p. MISR Laboratory, 2007, Soil Testing, the Colossi of Memnon and Amenhotep III Conservation Project, Kom el-Hettan–Thebes, Geotechnical Investigation: Prepared for German Archaeological Mission in Egypt: Luxor, 34 p. Naylor, M.A., Laroque, J.M., and Gauthier, B.D.M., 1994, Understanding extensional tectonics: Insights from sandbox models, in Roure, F., Ellouz, N., Shein, V.S., and Skvortsov, I., eds., Geodynamic Evolution of Sedimentary Basins: Proceedings of the Moscow Symposium, 18–23 May: Paris, Edition Technip, p. 69–83. Papathanassiou, G., Pavlides, S., and Ganas, A., 2005, The 2003 Lefkada earthquake: Field observations and preliminary microzonation map based on liquefaction potential index for the town of Lefkada: Engineering Geology, v. 84, p. 12–31, doi: 10.1016/j.enggeo.2005.08.006. Pausanias, 1898, Description of Greece, Book I, XLII, p. 64–65 (English translation, vol. 1, with commentary by J.G. Frazer, 6 vols.): London, New York, 561 p. Pawlikowski, M., 1987, Ekspertyza w sprawie zagrozeniq swiatyn w Dolinie Deir el-Bahari: Krakow, Institut Geologii i Surowcow Mineralnych, AGH, p. 1–41. Pawlikowski, M., 2001, Geological investigation of a cliff archaeological site above the Hatshepsut Temple in Deir El-Bahari: Polish Archaeology in the Mediterranean, in Polish-Egyptian Archaeological and Conservation Mission of the Temple of Queen Hatshepsut at Deir el-Bahari Reports

2000, XII: Warsaw, Polish Centre of Mediterranean Archeology, Warsaw University, p. 234–235. Pawlikowski, M., and Wasilewski, M., 2004, Some remarks on jointing in the Theban Limestone in the region of Deir El-Bahari, Egypt: Geologia, v. 30, no. 1, p. 47–56. Quatremere, M., 1845, Histoire de Sultans Mamlouks de l’Egypte: London, Oriental Translation Fund, 4 vols. Ricke, H., Haeny, G., and Habachi, L., 1981, Untersuchungen im Totentempel Amenophis’ III, in Haeny, G., ed., Beitrege zur Egyptischen Bauforschung und Altertumskunde, no. 11, p. 31–37: Wiesbaden, 122 p. RIGW (Research Institute of Groundwater), 1997, Hydro-Geological Map of Luxor Area, Explanatory Notes: Research Institute of Groundwater: Kalioubia, Egypt, scale 1:100,000. Strabo, 1889, The Geography of Strabo: Literally Translated, with Notes by H.C. Hamilton and W. Falconer: London, G. Bell & Sons, Book 17, 1854 p. Said, R., 1990, The Geology of Egypt: Rotterdam, A.A. Balkema, p. 1–170. Said, R., Aubry, M.-P., Ouda, Kh., and Dupuis, C., 2000, Thebes International Geo-Archaeological Project: Geological Framework for the Preservation, Sustainable Environmental Management and Geo-Hazards of the Thebes Archaeological Sites on the West Bank of the Nile (Luxor, Egypt): Under the Auspices of Prof. Zahi Hawass, Secretary General, Supreme Council of Antiquities: Cairo, 26 p. Sieberg, A., 1932, Untersuchugeb uber erdbeben und bruchschollenbau im oslichen mittelmeergebeit; Denkschriften der medizinisch-naturwissenschaftlichen: Gesselschaft zu Jean, v. 18, no. 2, p. 161–273. Sinopoli, A., 1995, Earthquakes and large block monumental structures: Annali di Geofisica, v. XXXVIII, no. 5–6, p. 737–750. Sourouzian, H., 2004, Report on works at the Colossi of Memnon and the Temple of Amenophis III at Thebes: Mitteilungen des Deutschen Archäologischen Instituts, Kairo, v. 60, p. 171–236, pl. 25–60. Sourouzian, H., and Stadelmann, R., 2003, Report on work at the Colossi of Memnon and the Temple of Amenophis III at Thebes: Mitteilungen des Deutschen Archäologischen Instituts, Kairo, v. 59, p. 425–446, pl. 71–76. Sourouzian, H., Stadelmann, R., Madden, B., and Gayer-Anderson, T., 2006, Three seasons of work at the temple of Amenhotep III at Kom el Hettan. Part I: Work at the Colossi of Memnon: Annales du Service des antiquités de l’Égypte, v. 80, p. 323–366. Stadelmann, R., 1984, Die Herkunft der Memnon-Kolosse: Heliopolis oder Aswan?: Mitteilungen des Deutschen Archäologischen Instituts, Kairo, v. 40, p. 291–296. Stadelmann, R., and Sourouzian, H., 2001, Der Totentempel Amenophis’ III, in Theben. Grabungen und Restaurierung am Kom el-Hettân, Mitteilungen des Deutschen Archäologischen Instituts, Kairo, v. 57, p. 271–280. Tsuchida, H., and Hayashi, S., 1971, Estimation of liquefaction potential of sandy soils, in Proceedings of the 3rd Joint Meeting, U.S.-Japan Panel on Wind and Seismic Effects: Tokyo, U.S.-Japan Cooperation Program (UNJR), p. 91–101 Verdel, T., 1993, Geotechnique et Monuments Historiques, Methodes de Modelisation Appliquées à des cas Egyptiens [Ph.D. thesis]: Nancy, France, Institut National Polytechnique de Lorraine–Ecole des Mines, 293 p. Walters, J.V., and Thomas, J.N., 1982, Shear zone development in granular materials, in Esisenstein, Z., ed., Proceedings of the 4th International Conference on Numerical Methods in Geomechanics, Edmonton 1982: Edmonton, Canada, Balkema, p. 263–274. Weeks, K.R., 2005, The Illustrated Guide to Luxor: Tombs, Temples and Museums: Cairo, American University in Cairo Press, 563 p. Wiedemann, A., 1884, Agyptische Geschichte, Volume 2: Gotha, F.A. Perthes, 780 p. Wilkinson, I.G., 1835, Topography of Thebes and General View of Egypt: London, Murray, 568 p. Youd, T.L., 1984, Geologic Effects—Liquefaction Associated Ground Failure: U.S. Geological Survey Open-File Report 84-760, p. 210–232. Youssef, M.M., Ibrahim, H.A., Bakheit, A.A., and Senosy, M.M., 1994, Surface and sub-surface tectonic pattern of Sohag region, Middle Egypt: Bulletin of the Faculty of Science of the Assiut University, v. 23, no. 1-F, p. 317–360. Youssef, S.E.-H., 2003, Seismicity of Egypt: Newsletter of the European-Mediterranean Seismological Centre, v. 20, p. 19–23. MANUSCRIPT ACCEPTED BY THE SOCIETY 19 MAY 2010

Printed in the USA

The Geological Society of America Special Paper 471 2010

Archaeological evidence for Roman-age faulting in central-northern Sicily: Possible effects of coseismic deformation Giovanni Barreca Maria Serafina Barbano Serafina Carbone Carmelo Monaco* Department of Geological Sciences, University of Catania, Corso Italia, 55, 95129 Catania, Italy

ABSTRACT We analyzed displacement on man-made structures and destructive effects on the ancient Roman community in two archaeological sites located on the southern Madonie Mountains, central-northern Sicily. At the archaeological site of Mount Alburchia, a votive niche of the Late Roman period is offset ~15 cm by a NNW-SSE– striking normal fault belonging to a NNW-SSE–trending en-echelon system related to an E-W–oriented regional shear zone. Structural analysis of deformed conglomerates suggests a coseismic origin of displacement. At the Gangi Vecchio archaeological site, 3 km northeast of Mount Alburchia, the prevalence of pottery pieces and coins dated from the fourth century B.C. to the fourth century A.D. and a sudden decrease of evidence of human activity at the end of the fourth century A.D. have been emphasized by archaeologists. Furthermore, recent archaeological excavations have revealed a Roman-age grave where remains of some women and children are mixed. Because ancient Romans usually did not use common burial, this could be related to an unexpected natural disaster. This event was conceivably a strong earthquake that probably occurred in the Late Roman period. An analysis of historical catalogues suggests that this event could be referred to the A.D. 361 earthquake, the epicenter of which, in central Sicily, is poor defined. Our data represent a first step for the relocation of this seismic event northward in the Madonie Mountains area.

INTRODUCTION Archaeological evidence for an earthquake is not always clear or unambiguous (Ambraseys, 2006). Where they are verified by geological and archaeological studies, faulted architectural relics represent valuable data in understanding active tectonics and seismic hazard of a region. In many cases, they represent the missing tiles of a seismic puzzle data set, providing *[email protected]

useful information and filling gaps in historical seismicity. From this perspective, countries rich in ancient ruins, such as those bordering the Mediterranean Sea, can offer a natural laboratory for this discipline. In this area, ancient buildings damaged by earthquakes have been described since the beginning of the last century (e.g., Lanciani, 1918; Evans, 1928), even though modern and interdisciplinary methodologies have been applied only recently (e.g., Karcz and Kafri, 1978; Stiros, 1988, 1996, 2001; Noller and Lightfoot, 1997; Hancock and Altunel, 1997; Galli and Galadini, 2001; Galadini and Galli, 2004).

Barreca, G., Barbano, M.S., Carbone, S., and Monaco, C., 2010, Archaeological evidence for Roman-age faulting in central-northern Sicily: Possible effects of coseismic deformation, in Sintubin, M., Stewart, I.S., Niemi, T.M., and Altunel, E., eds., Ancient Earthquakes: Geological Society of America Special Paper 471, p. 223–232, doi: 10.1130/2010.2471(18). For permission to copy, contact [email protected]. © 2010 The Geological Society of America. All rights reserved.

223

224

Barreca et al.

Our research provides a case study on faulted ancient relics in central-northern Sicily (Fig. 1A), an area characterized by moderate seismicity, mostly related to convergence along an E-W– oriented regional shear zone (Southern Tyrrhenian system) and where Quaternary faults deform terrains showing poor evidence for late Pleistocene to recent activity. Recent field investigation carried out on the southern slope of the Madonie Mountains, which are a segment of the Sicilian-Maghrebian chain (Fig. 1A), allowed us to recognize earthquake-related damage in a Romanage archaeological site (Mount Alburchia). At first, the study was focused on archaeological evidence for faulting. Successively, in order to frame the faulting in the neotectonic context of the central-northern Sicily, a structural and geomorphologic survey was carried out. Moreover, to better constrain the age of the disastrous event, field studies were accompanied by analysis of archaeological data from other sites located in the neighboring area. Finally, collected data were compared to earthquake catalogues (Boschi et al., 1995, 1997, 2000; Working Group CPTI, 2004) to verify whether seismic events with local sources could have triggered the earthquake-related damage occurring in the area.

Currently, central-northern Sicily is characterized by moderate seismicity. Focal mechanisms are mostly characterized by strike-slip and reverse-oblique kinematics compatible with low-dip NNW-SSE– to NNE-SSW–trending P axes (Fig. 1A), roughly consistent with the general convergence between the European and the African plates (Frepoli and Amato, 2000; Neri et al., 2005; Lavecchia et al., 2007). Shallow seismicity in northern Sicily has been related to convergence along the E-W–oriented regional shear zone, which is still active along the southern Tyrrhenian Sea margin (Gueguen et al., 2002; Giunta et al., 2004). This fault system deforms very recent deposits along the coast of northwestern Sicily (Nigro et al., 2000; Tondi, 2007) and seems to control drainage pattern and catchment evolution along the northern slope of the Madonie Mountains (Barreca et al., 2008). Strike-slip faulting is also evident offshore northeastern Sicily (Finetti and Del Ben, 1986; Argnani et al., 2007), where it controls the distribution of recent seismicity.

TECTONIC SETTING

Evidence for Roman-Age Faulting

The Sicilian-Maghrebian chain (Fig. 1A) is a thin-skinned, south-verging, forelandward migrating fold-and-thrust system, developed during the Neogene to Quaternary Africa-Europe collision (Dewey et al., 1989; Ben Avraham et al., 1990; Roure et al., 1990). Since the late Miocene, frontal thrusting of the chain has been accompanied by development of out-of-sequence contractional structures (Ghisetti and Vezzani, 1984; Bello et al., 2000) and by the development of syntectonic sedimentary basins in the depressions at the rear (Grasso and Butler, 1991), the deposits of which are well preserved in the southern sector of the Madonie Mountains (Grasso et al., 1978; Barreca, 2007). Here (Fig. 1B), Serravallian-Tortonian clays, sandstones, and conglomerates, Messinian evaporites, and Lower Pliocene chalks unconformably lie upon a pile of thrust sheets constituted by chaotic Paleocene-Oligocene variegated clays (Sicilide unit) and by Upper Oligocene–Lower Miocene quartzarenitic sequences (Numidian Flysch). Contractional structures are dominant in the studied area, characterized by ~W-E–oriented large folds in Neogene deposits (Fig. 1B). Moreover, since middle Pliocene times, this region has been deformed by major NW-SE–striking, en-echelon, rightlateral, strike-slip faults with conjugate NE-SW–striking, leftlateral, strike-slip faults and associated N-S extensional structures (normal faults and joints) (Barreca, 2007). In this framework, dextral lateral motion caused large clockwise rotations (e.g., the Gangi syncline, Fig. 1B). These faults are part of a regional rightlateral W-E–trending shear zone (Fig. 1A; Giunta et al., 2000) related to the Tyrrhenian Basin opening northward (“Southern Tyrrhenian system”; Finetti and Del Ben, 1986; Finetti et al., 1996; Lentini et al., 2006).

Evidence for Roman-age faulting has been found in the archaeological site of Mount Alburchia, which is located in the northeastern sector of the Mount Alburchia ridge, ~3 km southwest of Gangi village (Fig. 1B). In this site, a sixth–fifth century B.C. necropolis and a Roman cult manufacture with religious shrines dated to fourth century A.D. are found (“edicole votive”; Naselli, 1956; Tusa, 1958; Manganaro, 1965). From a geological-structural point of view, the Mount Alburchia ridge is a gentle NE-SW–oriented syncline, made up of Middle–Upper Miocene red conglomerates and alluvial laminated sands (Jones and Grasso, 1995; Barreca, 2007), crosscut by NW-SE–striking, right-lateral faults (Fig. 1B). At a smaller scale (Fig. 2A), a NNW-SSE–trending system of en-echelon tension structures (Fig. 2B), kinematically related to the strike-slip system, has been recognized along the ridge. These structures are mostly characterized by normal motion (Fig. 2C) and decimetric to metric offset (Figs. 2D–2G). A set of NNW-SSE–striking normal faults crosscuts the archaeological site of Mount Alburchia, which is characterized by the occurrence of fourth century A.D. Roman votive shrines excavated on the Middle–Upper Miocene conglomerates (Fig. 3A). In particular, a fault showing a cumulative stratigraphic offset of ~80 cm (Fig. 3B) crosscuts a niche, the plastered roof of which is clearly offset ~15 cm (Figs. 3C and 3D). Detailed structural analysis on the bedrock conglomerates, constituted by cobbles in a sandy matrix, showed that single cobbles, mostly composed of quartzarenites, granitoid, and metamorphic rocks, are cut by the same normal faults (Figs. 3E–3G). This suggests that fault displacement of the Roman architectural relic was probably related to coseismic slip (see following).

GEOARCHAEOLOGICAL DATA

13°E

15°E

14°E

Tyrrhenian Sea

Tyr r 38°N

Se ian Ion

Fig.1B

Mt. Etna

(

Sicily Channel

ian

(

a Se

ck

37°N

( (

( ( ( (

I Å

Å Å Å

Oblique fault

Æ Æ Æ

Syncline axis

( ( ( (

Anticline axis

0

)

)

5

)

I

I

I

I

I

) )

Æ

Geological boundary

Æ ( Æ( ( ( ( (Æ ( (

I

I

I

)

(( ( ( (

(

(( ( ( (

Normal fault

I

))

(

Æ

37°45’N

Å

I

I

I I

I

I

( (

( (

(

(( ( ( (

I

(

Æ

I

I

( ( (

) )

I

I

I I

(

I

Æ

I

I

I

Æ

Thrust

Æ

(

Cherty limestones (Imerese Unit) Upper Triassic

( ( (

Gangi Vecchio

Å Å Mt. Alburchia

Fig. 2A

Limestones and dolostones (Panormide Unit). Upper Triassic

Å ( ( (

( (

Numidian Flysch Upper Oligocene-Lower Miocene

)

14°4’30’’E

)

Å

Å Å Æ Æ

Variegated clays (Sicilide Unit) Paleocene-Oligocene

Æ

Å

Clays, arenites, conglomerates Serravallian-Tortonian

(

Æ

Evaporites Messinian

) )

Æ Æ

( ( ( (

(

Chalky-limestones Lower Pliocene

Æ

Gangi

Å Å

Alluvial

) )

)

I

)

Æ

)

( ( ( ( (

) )

( (

) ) )

(

(( (

(

I

(

( (Æ( Æ Æ Æ ( ( ( ( ( ( ( ( (( ( (( ( ( ( ( ( ( ( ( ( ( ( ( ( Æ ( ( ( ( (( ( ( ( ( Petralia Soprana Æ ( ( ( ( ( ( ( ( ( ( Æ ( ( (( Æ Æ Æ( (Æ ( (( ( (( (( (( ( ( ( ( ( ( ( (( (( ( ( ( ( Å Å ( Æ ( Æ Æ Æ (( (

)

I

(( ( (

((

) ) ) ) ) )

(

(

)

Mt. S. Salvatore

14°15’E

) )

(Æ

( ( (

(

) ) )

) )

B

) ) )

15°E

14°E

40 km

20

0 13°E

)

(

Sicily Channel

(

blo n ia Pelag (

Ion

( (

(

Sicily

( ( ( (

-M n hai agh C rebian

)

an

37°N

A

nia n Sea

a

Madonie Mts.

(

(

38°N

)

Sicil i

he

Æ Æ Æ ( (

( (

10 km

Figure 1. (A) Neotectonic sketch map of Sicily. The right-lateral and oblique faults belonging to the “Southern Tyrrhenian system” (after Lentini et al., 2006) are shown. Solid black line with triangles—front of the Sicilian-Maghrebian Chain. Focal mechanisms for selected earthquakes (2.2 < M < 5.5) in central-northern Sicily located in the depth range of 0–30 km are after Neri et al. (2005) and reference therein. (B) Geological-structural sketch map of the southern slope of Madonie Mountains (see part A for location) with evidence of the Mount San Salvatore–Mount Alburchia fault trace.

226

Barreca et al.

A G

Geological boundary between Middle-Upper Miocene clays and conglomerates

D I I I I I

0

25

50 m

Fig. 3A

Normal fault Strike-slip fault Archaeological site Landslide

Mt. Alburchia

Fig. 2G

Fig. 2D 37°45’ 15’’N 0

5

10 m

Fig. 2F

Fig. 2E 0

2.5

5m

E 280 m

0

B

14°10’ 15’’E N

0

1m

/

14°10’ 45’’E

14°10’ 30’’E

C

F

N

Stress axi s

σ3 σ2 σ1

Figure 2. (A) Main tectonic structures recognized in the Mount Alburchia area (see Fig. 1B for location). (B) Rose diagram showing statistical azimuthal domain of the measured normal faults. (C) Diagram (Schmidt, lower hemisphere) showing normal fault plane attitude and calculated stress axes. (D–G) Examples of normal faults (see location in Fig. 2A).

Comparison with Archaeological Data from Neighboring Site

human activity can be interpreted as a consequence of an exceptional event, such as an unexpected natural disaster.

Historical and archaeological data show that the Gangi– Mount Alburchia area was already settled in prehistoric time and subsequently inhabited during Greek and Roman ages, when it was characterized by a bustling economy tied into the Rome world-system (Naselli, 1956; Tusa, 1958; Manganaro, 1965). About 2 km southeast of Gangi village (Fig. 1B), the old site, called Gangi Vecchio, at present is occupied by a fourteenthcentury Benedictine Abbey built on Roman-age buildings. The Roman colonization dates from the high empire period (first to fifth centuries A.D.), even though eighth- to seventh-century B.C. Greek colonial remains have been also found. Gangi Vecchio could be the fabled city of Engion mentioned by Diodorus Siculus, and local tradition also holds to a 1200 B.C. colonization from Mycenaean civilization (Storey, 2002). More recently, it was colonized by Byzantine, Arabs, and, finally, by Normans, who for defense purposes, moved to the present-day site of Gangi. In the Gangi Vecchio archaeological site, the prevalence of pottery pieces and coins dated from fourth B.C. to fourth century A.D. (Fig. 4) and a sudden decrease in human activity at the end of the fourth century A.D. have been emphasized by the archaeologists (Beck et al., 1975). Furthermore, recent archaeological excavations at the same site (Storey, 2002) have revealed a fourth century A.D. grave where remains of women and children are mixed. It is worth noting that common burials are unusual in Roman culture. This feature and the significant decrease in

POSSIBLE EFFECTS OF COSEISMIC DEFORMATION Data Analysis Many cases of displaced archaeological remains have been reported in the available literature throughout the world (e.g., Stiros, 1988; Marco et al., 1997, 2003; Noller and Lightfoot, 1997; Galli and Galadini, 2001, and references therein). This deformation process involves buildings, fortified walls, canals, and other kinds of artifacts. The displacement can be related to the activation of (1) a fault during an earthquake or (2) a shear plane in the case of (a) differential settlement or (b) landsliding not necessarily induced by a seismic event. Indeed, the displacement of foundations due to coseismic activity of a fault or due to aseismic differential settlement or sliding (along newly formed shear planes or inherited fractures and faults driving the gravitational displacement) may be very similar (e.g., Karcz and Kafri, 1978). The correct interpretation of factors conditioning the displacement results from geological and geomorphologic investigations at the archaeological site. Since vertical offset may result from different processes, we performed detailed geomorphologic and structural investigations at the archaeological site, which allowed us to exclude faulting due to aseismic differential settlement or sliding. Taking into account the strong mechanical strength of cobbles, the ruptures observed

0

A

2.5

5m

F E G

B

C

B

80 cm

0.5 m

0

D

15 cm

E

G

F 15 cm

12 cm

8 cm

Figure 3. (A) View from the north of Mount Alburchia archaeological site (see Fig. 2A for location); boxes correspond to the pictures below. (B) Normal fault that crosscuts a votive niche, showing a cumulative stratigraphic offset of 80 cm. (C–D) Detail of the displaced (15 cm) plastered roof of the votive niche. (E–G) Closely spaced normal faults in Middle– Upper Miocene conglomerates; cobbles are clearly fractured.

228

Barreca et al. Relations to the Neotectonic Context of Central-Northern Sicily and Fault Parameters

V B.C.

IV B.C.

III B.C.

II B.C.

I B.C.

0 I A.D.

II A.D.

III A.D.

IV A.D.

No coins Figure 4. Coin distribution in time for Mount Alburchia and Gangi Vecchio sites; coins have an age older than the fifth century A.D. (exposition in the Gangi Museum).

at the archaeological site of Mount Alburchia seem to be compatible with a very rapid (impulsive) stress propagation rather than the result of a slow and progressive increase of the stress field (see Moretti, 1990). In the latter case, according to the Griffith theory, ruptures would have formed along preexisting horizons of mechanical weakness, such as cobbles surfaces. Interestingly, some of these faults also cut the Roman cult manufacture. For this reason, we think that the displacement of the Roman-age shrine was the result of motion along a seismogenic fault related to a strong historical earthquake, probably occurred during the Late Roman period. This is confirmed by researches at the site of Gangi Vecchio, where all archaeological data agree with a natural disaster occurring at the end of the fourth century A.D.

The offset affecting the votive niche is probably the last displacement on a normal fault that shows a cumulative offset of 80 cm on Middle–Upper Miocene conglomerates (Fig. 3B), and the precise length of which is not valuable since it extends onto clayey terrains. This structure belongs to a NNW-SSE–trending en-echelon normal fault system that transversely deforms the Mount Alburchia Ridge (Fig. 2A). The geological and structural survey showed that these extensional structures are kinematically related (for an overview on strike-slip zone geometries, see Sylvester, 1988) to the 20-km-long, regional, NW-SE–striking, right-lateral, strike-slip fault extending from Mount San Salvatore to Mount Alburchia (Fig. 1B). This structure, here named Mount San Salvatore–Mount Alburchia fault, shows, in turn, an en-echelon arrangement with other NW-SE–oriented, rightlateral, strike-slip faults belonging to the aforementioned W-E– striking shear zone located in central-northern Sicily. This fault is not older than middle Pliocene (3.6 Ma) and shows an average stratigraphic right-lateral offset of 1.5 km (Barreca, 2007), suggesting a slip rate of ~0.4 mm/yr. In order to verify that the Mount San Salvatore–Mount Alburchia fault has been reactivated during the Holocene, a preliminary geomorphologic analysis was carried out on drainage basins of the southern slope of the Madonie Mountains (Fig. 5). The analysis was based on the identification of fluvial anomalies by overlapping the fault trace on a drainage network map obtained by 1:10,000 scale digital topography (CTR, 1987). Evidence of tectonic control on the drainage pattern, which developed during the Holocene, is emphasized by the presence of particular anomalies such as fluvial captures and systematically deflected streams. In fact, the main drainage lines of the rivers Imera, Gangi, and Salso (Fig. 5) show marked deflections in correspondence of the fault trace, probably related to recent rightlateral motion. However, the very low slip rates and the prevalent clayey lithology of the area are not favorable to the development of a typical seismic landscape (sensu Michetti, 2005). Geological and archaeological evidence, together with the geomorphologic analysis, suggests that the activity of the dextral shear zone related to the Mount San Salvatore–Mount Alburchia fault and associated extensional structures could be responsible for coseismic deformation of Mount Alburchia conglomerates and of the Roman-age votive niche. Taking into account that the coseismic slip can be measured only along one of the several normal faults associated with the main shear zone, seismic source parameters cannot be estimated for the major strike-slip structure. Nevertheless, an attempted evaluation of the possible magnitude associated with the seismic activity of this shear zone can be made by using the last displacement measured along the normal fault affecting the Roman votive niche. According to empirical relationship of Wells and Coppersmith (1994) for normal faults, taking into account the displacement of the archaeological remains (~15 cm), the observed deformation can

Archaeological evidence for Roman-age faulting in central-northern Sicily

229

14°5’E Strike-slip fault

Flow direction Fluvial capture

37°500’N

Drainage network

Mt. San Salvatore

Deflected stream

Riv

er

A

37°45’N

Salso

Riv

er

Imera

Figure 5. Shaded relief map showing fluvial anomalies along the Mount San Salvatore–Mount Alburchia fault (same location as Fig. 1B).

0 14°5’E

2,5

5 km 14°10’E

be related to an earthquake of magnitude ~6. Analysis of historical catalogues reveals that this is unusually high for this area (see following). Analysis of Historical Seismicity The recurrence interval for large-magnitude earthquakes in Italy is generally longer than 1000 yr (e.g., Valensise and Pantosti, 1992; Pantosti et al., 1993; Galadini and Galli, 1999), and the largest earthquakes usually have recurrence times of the same order as the length of the available historical records. Although the Italian catalogues report earthquakes that have occurred over a long time span (461 B.C.–A.D. 2002) (Boschi et al., 2000; Working Group CPTI, 2004), data on the effects of historical earthquake are generally sparse, at least until the thirteenth- to fourteenth-century events. So, archaeoseismology can provide a useful chronological marker for constraining possible earthquakes in the past. In Sicily, the largest amount of archaeological information is available for the period between fourth century B.C. and the fourth–fifth century A.D. In order to verify whether seismic events with local sources could have triggered the earthquake effects observed in the Gangi area, an analysis of earthquake catalogues (Boschi et al., 1995, 1997, 2000; Working Group CPTI, 2004) was carried out. In particular, the analysis concerned historical seismicity of central-

Gang i River

Mt. Alburchia

northern Sicily and the possible occurrence of an earthquake with magnitude comparable to that obtained from empirical relationships (M ~6). The target of this analysis consisted of identifying earthquakes that have been able to produce superficial effects with ground displacements comparable to that observed in the Mount Alburchia niche. According to historical and instrumental records, the seismicity of central-northern Sicily is characterized by only a few moderate earthquakes with epicenters spread from the Tyrrhenian coast to the chain and maximum magnitude of 5.5 (Fig. 6). The seismic history of Gangi, obtained by merging historical observed intensities and intensities computed by parametric catalogue and cubic attenuation law, shows that the strongest damage in this locality was caused by the A.D. 361 and 1693 earthquakes (intensity [I] = VIII–IX MCS (Mercalli-Cancani-Sieberg); Fig. 7A), while the closest relevant earthquakes were the 1818, 1819, and 1967 events (Fig. 7B). Taking into account that the 1693 event was located in southeastern Sicily, at a distance of ~200 km (Fig. 6), and that the maximum intensity estimated for the 1818, 1819, and 1967 events was VII MCS (Fig. 7A), the strongest earthquake of central-northern Sicily located near the Gangi area was the A.D. 361 event (I = VIII MCS; Mw = 6.3). This earthquake, even though poor defined, was roughly located in central Sicily (Fig. 6), as given in historical catalogues (Boschi et al., 2000).

13°E

12°E

15°E

14°E

16°E

Tyrrhenian Sea 1823

38°N 1819 1968

38°N

1818

2001 Gangi 361?

Piazza Armerina

Magnitude range 37°N

4.5<M≤5 5<M≤5.5 5.5<M≤6

Ionian Sea

1693

Sic ily

6<M≤7 7<M

Ch a nn

el

0

50

37°N

100 km

36°N

36°N 13°E

12°E

15°E

14°E

16°E

Figure 6. Epicenter map and magnitude values for historical earthquakes with M >4.5 in Sicily (after Working Group CPTI, 2004).

361 1818-19 1967

Observed

7 i

Cefalù i Pollina

6

1819

i

21

5

Geraci

i

km 2 km 1

Intensity (I)

Sea Tyrrhenian

1693

Calculated

8

1818

i

Petralia Sottana

4 3 0

500

1000

Years

1500

2000

®

Gangi i km 27

ñ

Mistretta 1967 i Capizzi i

i Cerami Nicosia

Mt. Alburchia 0

10

20 km

37°30’N

9

B

38°00’N

14°30’E

14°00’E

A

Figure 7. (A) Seismic history of the Gangi area; major observed intensities (square) are related to 1818–1819 and 1967 earthquakes (I = VII– VIII); those calculated (diamond) are related to A.D. 361 and 1693 earthquakes. (B) Distance of 1818, 1819, and 1967 earthquake epicenters from Mount Alburchia archaeological site.

Archaeological evidence for Roman-age faulting in central-northern Sicily Some historical sources regarding the A.D. 361 earthquake were cited by Guidoboni et al. (1994, p. 260): “Libanius brings together the earthquakes which occurred during the reign of Julian, mentioning amongst them a destructive event in Sicily: the greatest cities of Sicily lie in ruins. An edict issued by the emperor Honorius (395–423) to promote the construction of public buildings in Sicily and archaeological evidence for damage at about this time at the great Roman villa of Piazza Armerina in central Sicily (see Fig. 6) could be linked to this earthquake.” Moreover, historical earthquake catalogues, such as Bonito (1691) and Mongitore (1743), report similar accounts referring to this earthquake with different dates (A.D. 362–365). Actually, from the few available historical accounts, we only know that during the reign of Julian, an earthquake, or more that one earthquake, struck Sicily, but its/their location and dimensions are unknown. However, further analysis of the numerous archaeological sites of northeastern Sicily is necessary to find new information and for a more complete evaluation of this uncertain event reported in the catalogues. DISCUSSION AND CONCLUSIONS During Pliocene–Quaternary times, northern Sicily was dominated by dextral strike-slip faulting along a W-E–striking shear zone. Major faults show NW-SE direction and rare evidence of reactivation related to moderate seismicity. On the southern slope of the Madonie Mountains (central-northern Sicily), one of these structures (Mount San Salvatore–Mount Alburchia fault) is associated with an en-echelon system of NNW-SSE–trending normal faults, which crosses the archaeological site of Mount Alburchia; in particular, a fourth-century votive niche is offset ~15 cm. The structural analysis of conglomerates deformed by these faults suggests a coseismic origin of displacement. Moreover, archaeological features of the area show an important decrease in human activity, probably at the end of fourth century A.D. As a whole, geological and archaeological evidence in the study area suggests the occurrence of a natural calamity that produced displacement on man-made structures and destructive effects on the ancient Roman community, possibly in the fourth century A.D. This event was conceivably a strong earthquake. A theoretical magnitude M ~6 has been extrapolated for this inferred earthquake by empirical relationships, which is unusually high for this area, where historical events with M > 5.5, capable of producing superficial faulting, are not known. Taking into account that destructive effects on the ancient Roman community of Mount Alburchia and Gangi Vecchio sites probably occurred in the Late Roman period, the analysis of historical catalogues suggests that this event could be referred to the 361 A.D. earthquake, the epicentral location of which, in central Sicily, is poor defined. Our data represent a first step for the relocation of this seismic event northward in the Madonie Mountains area. This assertion could have a strong impact on seismic hazard assessment in central-northern Sicily, an area that has been always considered to be characterized by moderate seismicity.

231

However, both the historical catalogues and the very low fault slip rate would suggest long (more than ~1500 yr) recurrence times for M 6 earthquakes in this area. Probably, an exploratory trench along the main displacement zone would be necessary to verify if recent activity has been recorded and to evaluate more precisely its kinematics, age, and slip rate. Furthermore, new analyses should be performed at other archaeological sites of central-northern Sicily to better define this uncertain event, in order to obtain a more complete evaluation of the characteristic earthquake and its recurrence time. ACKNOWLEDGMENTS We are grateful to D. Pantosti for useful discussion and to the two anonymous reviewers and co-editor E. Altunel for their suggestions, which contributed to substantially improving the original manuscript. Many thanks are due to S. Salerno, Ph.D. student at Catania University, for logistic support, and to S. Ferraro for stimulating discussion. This paper represents a contribution to the activities of the United Nations Educational, Scientific, and Cultural Organization–funded International Geoscience Programme IGCP 567, “Earthquake Archaeology: Archaeoseismology along the Alpine-Himalayan Seismic Zone.” REFERENCES CITED Ambraseys, N.N., 2006, Earthquakes and archaeology: Journal of Archaeological Science, v. 33, p. 1008–1016, doi: 10.1016/j.jas.2005.11.006. Argnani, A., Serpelloni, E., and Bonazzi, C., 2007, Pattern of deformation around the central Aeolian Islands: Evidence from multichannel seismics and GPS data: Terra Nova, v. 19, p. 317–323, doi: 10.1111/j.1365 -3121.2007.00753.x. Barreca, G., 2007, Carta Geologica del Settore Pedemontano delle Madonie: Firenze, S.EL.CA. (Società Elaborazioni Cartografiche), scale 1:25,000. Barreca, G., Carbone, S., and Perricone, C., 2008, Integrated structural and morphometric analysis for coastal small scale drainage basins in Cefalù area (central-northern Sicily); evidence for active tectonics, in Di Stefano, A., Maniscalco, R., and Sturiale, G., eds., Tethys to Mediterranean—A Journey of Geological Discovery, Meeting in Memory of Angelo di Grande and Mario Grasso, Abstract Book, 3–5 June 2008: Catania, p. 13. Beck, P., Maccari, B., and Poisson, M.J., 1975, Prospezione archeologica a Gangi Vecchio: Archeologia Medievale, v. II, p. 382–386. Bello, M., Franchino, A., and Merlini, S., 2000, Structural model of eastern Sicily: Memorie della Società Geologica Italiana, v. 55, p. 61–70. Ben-Avraham, Z., Boccaletti, M., Cello, G., Grasso, M., Lentini, F., Torelli, L., and Tortorici, L., 1990, Principali domini strutturali originatisi dalla collisione nogenico-quaternaria nel Mediterraneo centrale: Memorie della Società Geologica Italiana, v. 45, p. 453–462. Bonito, M., 1691, Terra tremante, ovvero continuatione de’ terremoti dalla creatione del mondo sino al tempo presente, Napoli (rist. anast., Bologna, 1979), 822 p. (in Italian). Boschi, E., Ferrari, G., Gasperini, P., Guidoboni, E., Smriglio, G., and Valensise, G., 1995, Catalogo dei forti terremoti in Italia dal 461 a.C. al 1980, with (CD-ROM): Istituto Nazionale di Geofisica–Storia Geofisica Ambiente, Ozzano Emilia, 973 p. Boschi, E., Guidoboni, E., Ferrari, G., Valensise, G., and Gasperini, P., 1997, Catalogo dei forti terremoti in Italia dal 461 a.C. al 1990, with (CDROM): Istituto Nazionale di Geofisica–Storia Geofisica Ambiente, Ozzano Emilia, 644 p. Boschi, E., Guidoboni, E., Ferrari, G., Gasperini, P., Mariotti, D., and Valensise, G., 2000, Catalogue of strong Italian earthquakes from 461 a.C. to 1997: Annali di Geofisica, v. 43, no. 4, p. 843–868 and CD-ROM. CTR (Carta Tecnica Regionale), 1987, Carta Tecnica Regionale: Palermo, Assessorato Territorio e Ambiente Regione Siciliana, scale 1:10,000.

232

Barreca et al.

Dewey, J.F., Helman, M.L., Turco, E., Hutton, D.H.W., and Knott, S.D., 1989, Kinematics of the Western Mediterranean, in Coward, M.P., Dietrich, D., and Park, R.G., eds., Alpine Tectonics: Geological Society of London Special Publication 45, p. 265–283. Evans, A.J., 1928, The Palace of Minos at Knossos II: London, Macmillan, 547 p. Finetti, I., and Del Ben, A., 1986, Geophysical study of the Tyrrhenian opening: Bollettino Geofisica Teorica e Applicata, v. 28, p. 75–155. Finetti, I., Lentini, F., Carbone, S., Catalano, S., and Del Ben, A., 1996, Il sistema Appennino Meridionale–Arco Calabro–Sicilia nel Mediterraneo Centrale: Studio geologico-geofisico: Bollettino Società Geologica Italiana, v. 115, p. 529–559. Frepoli, A., and Amato, A., 2000, Spatial variation in stresses in peninsular Italy and Sicily from background seismicity: Tectonophysics, v. 317, p. 109– 124, doi: 10.1016/S0040-1951(99)00265-6. Galadini, F., and Galli, P., 1999, Paleoseismology related to the displaced Roman archaeological remains at Egna (Adige Valley, northern Italy): Tectonophysics, v. 308, p. 171–191, doi: 10.1016/S0040-1951(99)00080-3. Galadini, F., and Galli, P., 2004, The 346 A.D. earthquake (central-southern Italy): An archaeoseismological approach: Annali di Geofisica, v. 43, no. 2–3, p. 885–905. Galli, P., and Galadini, F., 2001, Surface faulting on archaeological relics: A review of case histories from Dead Sea to Alps: Tectonophysics, v. 335, p. 291–312, doi: 10.1016/S0040-1951(01)00109-3. Ghisetti, F., and Vezzani, L., 1984, Thin-skinned deformation in western Sicily: Bollettino Società Geologica Italiana, v. 103, p. 129–157. Giunta, G., Nigro, F., Renda, P., and Giorgianni, A., 2000, The Sicilian-Maghrebides Tyrrhenian margin: A neotectonic evolutionary model: Bollettino Società Geologica Italiana, v. 119, p. 553–565. Giunta, G., Luzio, D., Tondi, E., De Luca, L., Giorgianni, A., D’Anna, G., Renda, P., Cello, G., Nigro, F., and Vitale, M., 2004, The Palermo (Sicily) seismic cluster of September 2002, in the seismotectonic framework of the Tyrrhenian Sea–Sicily border area: Annals of Geophysics, v. 47, no. 6, p. 1755–1770. Grasso, M., and Butler, W.H., 1991, Tectonic controls on the deposition of late Tortonian sediments in the Caltanissetta Basins of central Sicily: Memorie della Società Geologica Italiana, v. 47, p. 313–324. Grasso, M., Lentini, F., and Vezzani, L., 1978, Lineamenti stratigrafico strutturali delle Madonie (Sicilia centro-settentrionale): Geologica Romana, v. 17, p. 45–69. Gueguen, E., Tavarnelli, E., Renda, P., and Tramutoli, M., 2002, The geodynamics of the southern Tyrrhenian Sea margin as revealed by integrated geological, geophysical and geodetic data: Bollettino Società Geologica Italiana, Volume Speciale, v. 1, p. 77–85. Guidoboni, E., Comastri, A., and Traina, G., 1994, Catalogue of Ancient Earthquakes in the Mediterranean Area up to the 10th Century: Bologna, ING, 504 p. Hancock, P.L., and Altunel, E., 1997, Faulted archaeological relics at Hierapolis (Pamukkale), Turkey: Journal of Geodynamics, v. 24, p. 21–36, doi: 10.1016/S0264-3707(97)00003-3. Jones, R.E., and Grasso, M., 1995, Paleotectonics and sediment dispersal pathways in north-central Sicily during the late Tortonian: Studi Geologici Camerti, Volume Speciale 1995/2, p. 279–291. Karcz, I., and Kafri, U., 1978, Evaluation of supposed archaeoseismic damage in Israel: Journal of Archaeological Science, v. 5, p. 237–253, doi: 10.1016/0305-4403(78)90042-0. Lanciani, R., 1918, Segni di terremoti negli antichi edifizi di Roma antica: Roma, Bollettino della Archeologia Comunale, p. 1–30. Lavecchia, G., Ferrarini, F., De Nardis, R., Visini, F., and Barbano, M.S., 2007, Active thrusting as a possible seismogenic source in Sicily (Southern Italy): Some insights from integrated structural-kinematic and seismological data: Tectonophysics, v. 445, p. 145–167, doi: 10.1016/j.tecto.2007 .07.007. Lentini, F., Carbone, S., and Guarnieri, P., 2006, Collisional and postcollisional tectonics of the Apenninic-Maghrebian orogen (southern Italy), in Dilek, Y., and Pavlides, S., eds., Post-Collisional Tectonics and Magmatism in the Eastern Mediterranean Region: Geological Society of America Special Paper 409, p. 57–81. Manganaro, G., 1965, Iscrizioni di M. Albuchia (Erbita ?), Cefalù, S. Marco di Alunzio, Tindari: Archeologia Classica, L’Erma di Bretschneider, v. 17, p. 199–204.

Marco, S., Agnon, A., Ellenblum, R., Eidelman, A., Basson, U., and Boas, A., 1997, 817-year-old walls offset sinistrally 2.1 m by the Dead Sea transform, Israel: Journal of Geodynamics, v. 24, p. 11–20, doi: 10.1016/S0264 -3707(96)00041-5. Marco, S., Hartal, M., Hazan, N., Lev, L., and Stein, M., 2003, Archaeology, history, and geology of the A.D. 749 earthquake, Dead Sea transform: Geology, v. 31, p. 665–668, doi: 10.1130/G19516.1. Michetti, A.M., 2005, Future trends in paleoseismology: Integrated study of the seismic landscape as a vital tool in seismic hazard analyses: Tectonophysics, v. 408, p. 3–21, doi: 10.1016/j.tecto.2005.05.035. Mongitore, A., 1743, Istoria cronologica de’ terremoti di Sicilia, in Della Sicilia ricercata nelle cose più memorabili, a cura di F. Serio, Volume 2: Palermo, p. 345–445. Moretti, A., 1990, Fratture di origine tettonica nei ciottoli dei conglomerati: Elementi strutturali probabilmente connessi a grandi terremoti: Rendiconti Società Geologica Italiana, v. 13, p. 77–84. Naselli, S., 1956, Monte Albura o Albuchia: Gangi (Palermo), Gangi Museum, p. 1–43. Neri, G., Barberi, G., Oliva, G., and Orecchio, B., 2005, Spatial variation of seismogenic stress orientations in Sicily, south Italy: Physics of the Earth and Planetary Interiors, v. 148, p. 175–191, doi: 10.1016/j.pepi.2004.08.009. Nigro, F., Renda, P., and Arisco, G., 2000, Tettonica recente nella Sicilia nordorientale e nelle Isole Egadi: Bollettino Società Geologica Italiana, v. 119, p. 307–319. Noller, J.S., and Lightfoot, K., 1997, An archaeoseismic approach and method for the study of active strike slip faults: Geoarchaeology, v. 12, p. 117– 135, doi: 10.1002/(SICI)1520-6548(199703)12:2<117::AID-GEA2>3.0 .CO;2-7. Pantosti, D., Schwartz, D.P., and Valensise, G., 1993, Paleoseismology along the 1980 Irpinia earthquake fault and implications for earthquake recurrence in the Southern Apennines: Journal of Geophysical Research, v. 98, p. 6561–6577, doi: 10.1029/92JB02277. Roure, F., Howell, D.G., Muller, C., and Moretti, I., 1990, Late Cenozoic subduction complex of Sicily: Journal of Structural Geology, v. 12, no. 2, p. 259–266, doi: 10.1016/0191-8141(90)90009-N. Stiros, S.C., 1988, Archaeology—A tool to study active tectonics. The Aegean as a case study: Eos (Transactions, American Geophysical Union), v. 69, December 13, p. 1636–1639. Stiros, S.C., 1996, Identification of earthquakes from archaeological data: Methodology, criteria and limitations, in Stiros, S., and Jones, R., eds., Archaeoseismology: British School at Athens, Fitch Laboratory Occasional Paper 7, p. 129–152. Stiros, S.C., 2001, The A.D. 365 Crete earthquake and possible clustering during the fourth to sixth centuries in the Eastern Mediterranean: A review of historical and archaeological data: Journal of Structural Geology, v. 23, p. 545–562, doi: 10.1016/S0191-8141(00)00118-8. Storey, G., 2002, Preliminary investigations at the site of Gangi Vecchio, Gangi, Province of Palermo, Sicily, Italy. July 22 to August 6, 2000: Report to Dr. Francesca Spatafora, Soprintendenza BB.CC.AA., Palermo, Italy, p. 1–8. Sylvester, A.G., 1988, Strike-slip faults: Geological Society of America Bulletin, v. 100, p. 1666–1703, doi: 10.1130/0016-7606(1988)100<1666:SSF >2.3.CO;2. Tondi, E., 2007, Nucleation, development and petrophysical properties of faults in carbonate grainstones: Evidences from the San Vito Lo Capo peninsula (Sicily, Italy): Journal of Structural Geology, v. 29, p. 614–628, doi: 10.1016/j.jsg.2006.11.006. Tusa, V., 1958, Scavi Archeologici a Monte Alburchia di Vincenzo Tusa. Estratto da: Palermo, Giglio di Roccia, nuova serie, 443, p. 13–15. Valensise, G., and Pantosti, D., 1992, A 125 kyr-long geological record of seismic source repeatability in the Messina Straits (southern Italy) and the 1908 earthquake: Terra Nova, v. 4, p. 472–483, doi: 10.1111/j.1365 -3121.1992.tb00583.x. Wells, D.L., and Coppersmith, K.J., 1994, New empirical relationship among magnitude, rupture length, rupture width, rupture area, and surface displacement: Bulletin of the Seismological Society of America, v. 84, no. 4, p. 974–1002. Working Group CPTI, 2004, Catalogo Parametrico dei Terremoti Italiani, versione 2004 (CPTI04), Istituto Nazionale di Geofisica e Vulcanologia, Bologna, http://emidius.mi.ingv.it/CPTI04/ (accessed 17 August 2010). MANUSCRIPT ACCEPTED BY THE SOCIETY 19 MAY 2010

Printed in the USA

The Geological Society of America Special Paper 471 2010

Faulting of the Roman aqueduct of Venafrum (southern Italy): Methods of investigation, results, and seismotectonic implications Paolo A.C. Galli* Dipartimento Protezione Civile Nazionale, Via Vitorchiano 4, 00189 Rome, Italy, and Consiglio Nazionale delle Ricerche, Istituto Geologia Ambientale Geoingegneria, Monterotondo Scalo, 00016 Rome, Italy Alessandro Giocoli Consiglio Nazionale delle Ricerche, Istituto di Metodologie per l’Analisi Ambientale, 85050 Tito Scalo (PZ), Italy Jose A. Naso Dipartimento Protezione Civile Nazionale, Via Vitorchiano 4, 00189 Rome, Italy Sabatino Piscitelli Enzo Rizzo Consiglio Nazionale delle Ricerche, Istituto di Metodologie per l’Analisi Ambientale, 85050 Tito Scalo (PZ), Italy Stefania Capini Direzione Regionale Beni Culturali e Paesaggistici del Molise, 86100 Campobasso, Italy Luigi Scaroina Istituto Nazionale di Archeologia e Storia dell’Arte, Piazza San Marco, 49, 00186 Rome, Italy

ABSTRACT We present evidence of surface faulting of a poorly known first-century B.C. aqueduct in central-southern Italy. Data were acquired by means of geological, geophysical, and geodetical surveys along the surficial trace of a primary active fault (Aquae Iuliae fault). The ~30-km-long Venafrum aqueduct presents a net vertical offset of almost 4 m at the intersection with this normal fault. This fact reveals the occurrence of repeated faulting of the Roman water supply after its construction, i.e., during large historical earthquakes, the last being one of the most violent events to happen in Italy during the Middle Ages (September 1349, Mw = 6.6). We tentatively associate the remaining offset of the aqueduct to other poorly characterized earthquakes in the area, which were not previously associated with any active fault. It is a well-known fact that the recognition of ancient earthquakes on archaeological relics is a matter of debate in archaeoseismology, being difficult at all times—and often impossible—to ascertain whether the damage observed should be related to seismic shaking or other

*[email protected] Galli, P.A.C., Giocoli, A., Naso, J.A., Piscitelli, S., Rizzo, E., Capini, S., and Scaroina, L., 2010, Faulting of the Roman aqueduct of Venafrum (southern Italy): Methods of investigation, results, and seismotectonic implications, in Sintubin, M., Stewart, I.S., Niemi, T.M., and Altunel, E., eds., Ancient Earthquakes: Geological Society of America Special Paper 471, p. 233–242, doi: 10.1130/2010.2471(19). For permission to copy, contact [email protected]. © 2010 The Geological Society of America. All rights reserved.

233

234

Galli et al. causes (i.e., wars, floods, fires, decadence, etc.). Conversely, the exceptional case represented by the faulting of an archaeological relic such as this provides certain and reliable data on the causative seismogenic source and the associated earthquakes.

INTRODUCTION The debate in archaeoseismology concerning the recognition of effects induced by ancient earthquakes on archaeological relics has been faced ever since the pioneering work done by Karcz and Kafri (1978; see also Stiros, 1996; Galadini and Galli, 1996, 2001). Scientists involved in this kind of problem know how difficult or even impossible it is to relate the damage observed to seismic shaking or else to other causes (i.e., wars, floods, fires, time decay, etc.), especially in cases concerning a single building inside an ancient settlement. Moreover, even if it is possible to assess whether the damage affecting an old (and obviously ruined) building was seismically induced, it would be risky to derive any earthquake epicentral parameters on the basis of this one single observation. On the other hand, in the exceptional case of the faulting of an archaeological object—even of one single relic—it is possible to go back to the causative earthquake through the information recorded by the fault. For instance, surface faulting always coincides with the earthquake mesoseismic area, and—in terms of energy release—always reveals the overcoming of a minimum magnitude threshold (e.g., Mw >6). When compared with a typical paleoseismological trench study, the analysis of faulted archaeological relics gives more reliable information concerning the dating of the earthquake(s), the kinematics, and the associated magnitude of the seismogenic source responsible (i.e., by measuring the slip direction and offset amount, respectively). In addition, this is particularly true of strike-slip or oblique faults, the displacement of which is hardly measurable inside a paleoseismic trench. This work focuses on an almost unknown first-century B.C. Roman aqueduct located in the Molise region (central-southern Italy). Thanks to detailed field and geophysical surveys, we singled out the underground traces of the aqueduct, particularly in the sector we supposed to have been affected by a primary active fault. Our results show the occurrence of a repeated faulting of the water supply during the past two millennia, i.e., during large historical earthquakes, the last being one of the most violent events known in Italy during the Middle Ages (September 1349, Mw = 6.6). REGIONAL SEISMOTECTONIC FRAMEWORK The investigated area is located in the carbonate chain of the Apennine range. The southern Apenninic Arc is a buried duplex system of Mesozoic–Tertiary carbonate thrust sheets overlaid by a thick pile of rootless nappes (Patacca and Scandone, 2007; Fig. 1), which is currently experiencing a strong NE-SW exten-

sion. This is shown by the studies on the active faults of the area (see Galadini and Galli, 2000; Galli and Galadini, 2003), by the (few) focal mechanisms of the local earthquakes (see Fig. 2 and references therein quoted), and recent global positioning system (GPS) analyses (see double arrows in Fig. 1; Giuliani et al., 2009).

Figure 1. Seismotectonic sketch of central-southern Italy. NAA and SAA—the Northern and Southern Apennine Arcs bounding externally on the buried fronts of the overthrusted chain. Earthquakes (Mw >5.9) are modified from Working Group CPTI (2004). Faults are from Galli et al. (2008). Arrows indicate global positioning system (GPS)– derived extension rates (values in mm/yr; from Serpelloni et al., 2006; Mantenuto et al., 2007; Mantenuto and D’Agostino, 2007; Giuliani et al., 2009).

Faulting of the Roman aqueduct of Venafrum (southern Italy)

235

Figure 2. Digital terrain map (DTM) of part of the central-southern Apennines (see location in the upper-right inset) showing the epicenters of the main historical earthquakes (Mw >5.5; from Working Group CPTI, 2004, except: 280 B.C.—Galli and Galadini, 2003; A.D. 346/355—based on data in Galadini and Galli, 2004; 847—based on data in Figliuolo and Maturano, 2002; 1349—Galli et al., 2008) and the known normal primary active faults (USFS—Upper Sangro fault system; RCAFS—Mount Rotella–Cinquemiglia–Aremogna Plains fault system; NMFS—northern Matese faults; AIF—Aquae Iuliae fault). The focal mechanisms are the Mw = 5.8 event of 5 May 1984 (Anderson and Jackson, 1987), the Mw = 4.2 event on 20 February 2008 (MedNet, 2008), and the Mw = 4.0 event of 9 August 2009 (www.ingv.it, 2009). All of them show NE-SW extension driven by NW-SE normal faults. Note the Venafrum aqueduct path, which is crossed by the Aquae Iuliae fault.

Moreover, large earthquakes (Mw >6.5) have devastated the entire region in the past (Figs. 1–2). Among these, we should mention the first-millennium A.D. events of 346/355, consisting of a complex seismic sequence that is known not only through few contemporary sources, but also through many epigraphs and scores of archaeoseismic indications collected and studied by Galadini and Galli (2004); the poorly characterized event of 847, studied by Figliuolo and Maturano (2002); and finally the second-millennium disruptive events of September 1349, December 1456 (both to be considered as the strongest seismic sequences to occur in Italy), and July 1805. While one of the main shocks of the 1456 sequence and the 1805 event have been paleoseismologically related to the activity of the fault system bordering on the northern Matese massif (together with an unknown event of the early third century B.C.; see 280 B.C. epicenter in Fig. 2 and N-Matese fault system [NMFS]; Galli and Galadini, 2003), the causative structures of the 346/355, 847, and 1349 earthquakes have remained so far either unknown or debated. Indeed, recent geological, geomorphological, and paleoseismological studies carried out by some of the authors of this paper

permitted the detailed identification and mapping of a hitherto unknown active fault (Aquae Iuliae fault, AIF in Fig. 2) bordering on the southwestern edge of the Matese massif, to which the September 1349 earthquake has been unequivocally linked (Galli and Naso, 2009). This fault affects the aqueduct path near the Venafro1 area, thus providing the opportunity for the acquisition of further valuable information on recent fault activity and earthquake recognition. VENAFRUM AQUEDUCT A giant epigraph exposed in the local Venafro Museum (Fig. 3A) accounts for the archaeological fame won by this aqueduct worldwide. This epigraph, together with other smaller inscriptions found along the aqueduct path (e.g., Fig. 3Aʹ), contains—in more than 600 words—the famous Edict of Emperor Caesar Augustus (Tabula Aquaria, dated from 17 to 11 B.C.: CIL 10, 4842; see Mommsen, 1883; Pantoni, 1961) and represents a 1

Venafro is the modern settlement and is the site of the Roman Venafrum.

B

D A C F

E

Figure 3. Different images of the Roman aqueduct of Venafrum. (A) Augustus’ Edict and (Aʹ) one of the epigraphs found along the path of the aqueduct; (B) view of the specus near the Volturno River spring and (Bʹ) near the fault zone; (C) discovery of the aqueduct vault in a ploughed field near the fault; (D) remnant of the specus hanging on a talweg; (E) the aqueduct unearthed close to Venafrum; and (F) view of the aqueduct walls deformed by a landslide near the fault zone.

Faulting of the Roman aqueduct of Venafrum (southern Italy) unique collection of regulations concerning the use and the maintenance of water supplies in Roman times, e.g.,

In regard to channels, conduits, sluices, and springs... or in regard to any other work which has been performed above or below the water level for the purpose of building or repairing the said aqueduct: it is ordered that whatever of the following operations have been done in the past are to continue in effect in the same manner, and workmen are to remake, to replace, to restore, or to repair in the same manner regardless of the number of times, and are to lay culverts and pipes of all sides, to make openings therein, and to do any other work necessary to construct the aqueduct.…

Or,

It is ordered that there shall be a cleared space of eight feet vacant on the right and the left sides of this watercourse and around those structures built to carry the water….

By the way, Venafrum was one of the 28 colonies established in Italy by Emperor Augustus, who endowed them with many public works, including the aqueduct for which the above regulations were made. It is worth noting that, different from Rome, where the water was free, the water from this aqueduct was sold to the public so that legal trials rising from offenses with respect to the water supply were under the jurisdiction of the Roman peregrine praetor. However, although the existence and the rough path of this aqueduct have been known since Ciarlanti (1644) and Cotugno

237

(1824), its structure and precise location have never been investigated, with the exception of a brief survey performed in the 1930s by Frediani (1938). The aqueduct starts from the spring of the Volturno River (548 m above sea level [asl]) near the famous ninth-century Benedictine San Vincenzo a Volturno Abbey (Fig. 2), and then, after a 200 m abrupt step, runs along a flat 31-km-long winding track, and finally arrives at Venafro (225 m asl). It runs mainly through a tunnel (called specus; see Fig. 4) inside the carbonate and marly hill slopes and goes beyond the deep talweg of many streams by means of arched bridges, most of which have now collapsed (Fig. 3D). According to our surveys, in the flat areas where the subsoil is made up of alluvial/colluvial or clayey deposits, the Romans built the aqueduct into a trench and then buried the excavation under the field surface (Fig. 4B). The specus is 0.6 m wide and 1.6 m high along its entire length with a round stone arch and an external structure in opus incertum (irregular stone masonry). We also found some trapdoors inserted above the vault (Fig. 3E), which might have been used to inspect and maintain the pipe. The inside walls are coated with hydraulic plaster (i.e., opus signinum = cocciopesto mortar), whereas the bottom is lined with bipedals (i.e., typical Roman 61 × 58 cm bricks; Fig. 4). The aqueduct was built in the first half of the first century B.C.—as deduced from a letter sent by M. Tullius Cicero— living between 106 and 42 B.C.—to his brother (Ad Quintum fratrem, 3, 1; see Cicero, 1989). It was then finished or restored by Emperor Augustus at the end of the same century. Conversely, we do not know when it ceased to function, although it is reasonable to believe that it fell into disuse during/after the fall of the Roman Empire (approximately fifth century) due to a lack of maintenance, a traumatic event, or both.

Figure 4. (A–B) Different typologies of the buried specus of the Venafrum aqueduct (3—arched stone masonry, 2—loose terrains, 1—clayey terrains). (C) View of the specus: note the integer bipedal at the bottom.

238

Galli et al.

We suppose that the Venafrum aqueduct shared the same lot with the impressive Roman drainage works of the neighboring Fucino Lake (a 4-km-long tunnel dug under a 1400-m-high carbonate mount, plus a 4-km-long canal excavated in lacustrine sediments; see 1915 epicenter in Fig. 1 for its rough location), which ceased to operate at the dawn of the sixth century due to the effects of surface faulting across the main canal and to the lack of maintenance (Galadini and Galli, 2001). In order to have an idea of the time when the aqueduct stopped taking water to Venafrum, we attempted to date the first mud-layer filling the bottom of the tunnel near the village of Pozzilli (i.e., in the hanging wall of the Aquae Iuliae fault; Fig. 5; panel F in Fig. 3). Unfortunately, we obtained an absolute age (780–410 B.C., 2σ calibrated age) that is not consistent with the history of the aqueduct, which may refer to the age of the parent material of the deposit penetrated somewhere inside the broken specus. SURVEY AND ARCHAEOSEISMIC ANALYSIS OF THE VENAFRUM AQUEDUCT Actually, neither the survey quoted by Frediani (1938) nor any other study contains analytical information concerning the location and the elevation of the aqueduct. Therefore, in order to discover and measure the aqueduct relics across the fault zone and in its surroundings, we performed a specific field, topographic, and geophysical survey. Thanks also to information obtained from the local people (e.g., Mr. Di Meo, panel C in Fig. 3), we found the aqueduct relics in a dozen different localities. Nine of

these were directly inspected and permitted—among the other things—the observations summarized in the previous section of this paper. In a few words, the elevation of each point (generally the level of the bipedal at the bottom, or the inner arch) was then measured by means of topographic leveling (associated error ±10 cm) and positioned on 1:5000 maps. We finally traced the path of the aqueduct by following the altimetric gradient between each observed segment, obtaining a detailed map from which we have derived a real topographic section along the 8500 m of the investigated track. Next, we concentrated the surveys along the fault zone previously mapped in detail by Galli and Naso (2009; Arcora site in Fig. 5), where we carried out a geomagnetic survey and explorative pits. It is worth remembering that the fault trace in this area is almost invisible at surface level due both to the high erodibility of the marly bedrock and more than two millennia of agricultural works. Therefore, the fault was investigated also through electrical resistivity tomography (ERT) analyses and the excavation of a paleoseismic trench (see location in Fig. 5; Galli and Naso, 2009). We also tried to obtain additional information by carrying out a georadar survey along both the aqueduct specus and the fault, but these analyses did not provide reliable results, probably because the tunnel is filled with and covered by the same clayey material within which it was excavated (i.e., screening the waves of the 400 MHz antenna used). Therefore, we focused our efforts on the geomagnetic survey that we performed extensively along the aqueduct track, which,

Figure 5. Digital terrain model of the area where the active Aquae Iuliae fault crosses the Roman aqueduct several times (from 1:5000 scale map, see location on Fig. 2). The indicated paleoseismological trench was excavated across the fault by Galli and Naso (2009).

Faulting of the Roman aqueduct of Venafrum (southern Italy) in turn, gave the impressive results shown in Figure 6, where the magnetic anomaly visible in the footwall (due to the magnetic susceptibility of archaeological materials, mainly bricks and plaster) depicts the aqueduct track at depth perfectly. GEOMAGNETIC TECHNIQUE Magnetic prospecting represents one of the geophysical techniques most widely applied to several archaeological problems thanks to noninvasive and high-velocity data acquisition characteristics. These allow the acquisition of a lot of highresolution magnetic data in a very small time range (Ciminale and Loddo, 2001; Piro et al., 2003; Chianese et al., 2004; Rizzo et al., 2005, and reference therein). In our case, the magnetic survey was realized using a cesium vapor magnetometer G-858 GEOMETRICS in gradiometric mode, with the two magnetic sensors set in a vertical direction at a distance of ~1 m in order to remove the diurnal variations of the natural magnetic field automatically. Furthermore, in order to accelerate the magnetic data acquisition, we followed a normal serpentine path across the survey area, taking the data in a bidirectional pattern. The magnetometric measurements were acquired using the mapped-survey mode, which allowed us to specify a priori and display the survey area moving within the investigated area by means of regular grids with a sampling rate of 5 Hz and an interlinear step of 1 m (snake acquisition mode). Due to the presence of some spikes and stripe phenomena in the original magnetic maps, we applied a filtering procedure (a spike remove and destripe procedure). On the whole, during this phase, we created 10 magnetic maps, highlighting a surface of around 14,000 m2, of which Figure 6 represents only the near-fault sector. ARCHAEOSEISMIC INSIGHTS All the collected data concerning the aqueduct location and its precise elevation are summarized in Figures 5 and 7, where the latter shows the aqueduct profile from the villages of Santa Maria Oliveto to Venafro (Fig. 7A). As can be seen, the first section from Santa Maria Oliveto to the quarry site has a 3.5/1000 gradient, and then it lowers to 2/1000 between the quarry site and Arcora (which is the fault zone), and to 1/1000 continuing toward Ivella site. From here,

Figure 6. Shaded relief elaboration of part of the geomagnetic survey performed along the aqueduct trace (Camporelle–Arcora tract in Figs. 5 and 7). The map highlights the presence of gradiometric anomalies ranging from –20 nT to 20 nT (arrows). In the lower panel, the specus is suboutcropping (~0.5 m, as checked in some pits excavated ad hoc), whereas farther west (upper panel), it disappears nearing the fault zone (i.e., it is dismantled and eroded by the progressive process of fault scarp retreat). The inset photo shows the acquisition phases.

239

240

Galli et al.

Figure 7. (A) Section of the Roman aqueduct across the Aquae Iuliae fault. Note the step detected between the two strands (Venafrum–Arcora vs. Camporelle–Santa Maria Oliveto), which occurs just at the fault crossing point. A continuous deformation revealed by lowering of the gradient occurs nearing the fault on both sides. For simplicity, this section does not show any other possible fault/aqueduct intersection between Arcora and Ivella (e.g., in Fig. 5). (B) Explicative sketch of the aqueduct in the fault zone (dashed circle in A). Relics of the aqueduct were observed, and their altitude was calculated both in the hanging wall and in the footwall (arrows). Here, the absence of the tunnel nearing the fault has been confirmed by geomagnetic analyses. At present, the area is occupied by olive trees, a concrete-lined canal, and a road, which hampered the excavation of explorative trenches.

it increases, reaching 3.2/1000, up to the Pozzilli cemetery, and turning into 2/1000 toward Venafro. For our specific aims, that is, finding possible surface faulting evidence, the most significant result relates to the net step between the last observed and leveled point at the Camporelle site (bipedal level at 244.8 m asl) and the one in Arcora (240.4 m asl), which are only ~200 m away from each other. By adopting the gradient measured between the quarry site and Camporelle (2/1000) and taking the aqueduct trace toward the Arcora site (actually, some pits excavated along the aqueduct track unearthed the specus at the forecasted elevation; see arrows in Fig. 6), the step between the two strands is at least 3.6 m high, and occurs exactly where the geological, geoelectrical, and paleoseismological analyses permit the location of the fault zone. Unfortunately, at this stage, the existence of large olive trees and the presence of a road and a concrete-lined canal in the fault zone hampered the direct excavation of the two faulted aqueduct tips, whereas for the other two crossing points visible in Figure 5, the landowners did not give us the authorization for digging. DISCUSSIONS AND CONCLUSIONS As far as the observed 3.5 m offset of the aqueduct is concerned, we must exclude errors of the Roman engineers because

this part of the pipeline was excavated in trench and in open air. It is worth noting that leveling errors in Roman works are known only for very long tunnels excavated under high mountains (e.g., the one under Mount Salviano for the drainage of Fucino Lake, central Italy; see Galli and Galadini, 2001; Galadini and Galli, 2001); however, also in these cases, the offset is in the order of centimeters or few decimeters. Considering that it would have been absolutely senseless for the Romans to have intentionally lost more than 3 m of precious hydraulic head before arriving at their final destination (=Venafro; as we verified along the entire 8-km-long path surveyed) and to have done this in a flat and clayey zone (i.e., without any morphological or lithological obstacle), we believed that this 3.6-m-high step is actually due to surface faulting. Unfortunately, in this sector, the aqueduct almost parallels the fault (Fig. 5), and due to the erosion of the raised block (e.g., fault scarp retreat processes), a dozen meters or so of its structure have been completely lost (Fig. 7B). This is confirmed by the geomagnetic analysis, which progressively “loses” the aqueduct traces as it nears the fault (Fig. 6). Figure 7B summarizes the situation surveyed at the intersection between the Aquae Iuliae fault and the aqueduct. However, the continuous deformation that affects more than 3 km of the pipeline both in the footwall and in the hanging wall

Faulting of the Roman aqueduct of Venafrum (southern Italy) (Fig. 7A) also witnesses the influence of the Aquae Iuliae fault on the aqueduct level. This deformation is typical of normal faults, as observed several times by postearthquake geodetic leveling (e.g., in Barrientos et al., 1987; Stein et al., 1988). At the moment, these archaeoseismic results coupled with those recently gathered through paleoseismological analyses (see Galli and Naso, 2009) provide robust indications of a recent activity along the previously poorly known NW-SE normal fault (Aquae Iuliae fault). The paleoseismic data permit the definite location of the large 1349 earthquake (i.e., absolute ages in trenches postdate the faulting to A.D. 1150–1270, and A.D. 1290–1420; others predate it to A.D. 1450–1650; Galli and Naso, 2009), providing also reliable epicentral parameters for this Middle Ages event. In fact, the fault length (~22 km) and the offset per event (~1 m) derived from paleoseismological data yield a Mw ~6.6 for this event (e.g., in Wells and Coppersmith, 1994; for Apenninic faults, see Galli et al., 2008), a value which fits with the macroseismic magnitude estimated through the intensity distribution (Working Group CPTI, 2004). Nevertheless, paleoseismic data did not provide any timenarrowed indication of what happened before the 1349 event, apart from the evidence of surface faulting after A.D. 240–560 and before A.D. 1020–1210 (14C calibrated age; in Galli and Naso, 2009). On the other hand, considering the aforementioned value of slip-per-event, the amount of the total offset of the Venafrum aqueduct across the fault zone (~3.6 m) is obviously inconsistent with a single coseismic rupture (e.g., the 1349 earthquake), accounting for at least two other 1349 earthquake–like surface ruptures occurring after the first century B.C. (i.e., the age of its construction). Taking into account that the offset observed across the aqueduct and inside the trenches occurred, obviously, during the same earthquake(s), we can try to combine and compare the paleoseismological and archaeoseismological hints with the historical data concerning the strong earthquakes in the area. In fact, if we look at the known earthquakes of the region during the first millennium A.D. (Figs. 1–2), and consider the distribution of their effects (see insights in Galadini and Galli, 2004; Figliuolo and Maturano, 2002), both the aforementioned A.D. 346/355 and 847 events could be considered as possible candidates for the offsets measured along the Aquae Iuliae fault. In particular, the damage data points of the Late Roman period earthquake fit with the Aquae Iuliae fault elongation (i.e., they are mainly spread in the hanging wall of the structure), while the 847 event has only two intensity data points, but both close to the fault. If this hypothesis—which does not claim to be conclusive— is right, then the Venafrum aqueduct recorded three earthquakes equally spaced in time, which yields a return time of 0.5 k.y., the shortest known period on active faults in the Apennines. If this were true, the elapsed time from the last earthquake would be today larger than the return time.

241

Actually, our working hypothesis is consistent with the high slip rates evaluated across the Aquae Iuliae fault (1–1.8 mm/yr, i.e., medium to short time period of observation; Galli and Naso, 2009) and with the extension rates derived from GPS measures in the area (~5 mm/yr divided between the northern and southern Matese bounding faults; Giuliani et al., 2009), which are the highest of the entire chain, evidencing that this area could be currently at high seismic risk. Ongoing researches, which involve the Molise Archaeological Superintendence and local and regional authorities, are aimed at excavating the other fault/aqueduct intersections (see Fig. 5) in order to obtain more reliable data on the number and the age of coseismic faulting of this Roman work. REFERENCES CITED Anderson, H., and Jackson, J., 1987, Active tectonics of the Adriatic region: Geophysical Journal of the Royal Astronomical Society, v. 91, p. 937–983. Barrientos, S.E., Stein, R.S., and Ward, S.N., 1987, Comparison of the 1959 Hebgen Lake, Montana, and the 1983 Borah Peak, Idaho, earthquakes from geodetic observations: Bulletin of the Seismological Society of America, v. 77, p. 784–808. Chianese, D., D’Emilio, M., Di Salvia, S., Lapenna, V., Ragosta, M., and Rizzo, E., 2004, Magnetic mapping, ground penetrating radar surveys and magnetic susceptibility measurements for the study of the archaeological site of Serra di Vaglio (southern Italy): Journal of Archaeological Science, v. 31, p. 633–643, doi: 10.1016/j.jas.2003.10.011. Ciarlanti, G.V., 1644, Memorie historiche del Sannio chiamato hoggi Principato Ultra, contado di Molisi e parte di terra di lavoro, provincie del Regno di Napoli: Isernia, C. Cavallo, 530 p. Cicero, M.T., 1989, 1st century, Epistulae ad Quintum fratem (A. Salvatore, ed.): Milano, Mondadori, 126 p. Ciminale, M., and Loddo, M., 2001, Aspects of magnetic data processing: Archaeological Prospection, v. 8, p. 239–246, doi: 10.1002/arp.172. Cotugno, G., 1824, Memorie istoriche di Venafro: Napoli, Società Filomantica, 348 p. Figliuolo, B., and Maturano, A., 2002, Terremoti in Italia meridionale dal IX all’XI secolo, in Marturano, A., ed., Contributi per la Storia dei Terremoti nel Bacino del Mediterraneo (secc. V–XVIII): Salerno, Laveglia, p. 33–67. Frediani, F., 1938, L’acquedotto augusteo di Venafro: Riassunto e rilievi a cura dell’Ente Volturno: Istituto di Studi Romani, Campania Romana, Studi e Materiali: Napoli, Rispoli Anonima Edizione, p. 163–185. Galadini, F., and Galli, P., 1996, Paleoseismology related to deformed archaeological remains in the Fucino Plain: Implications for subrecent seismicity in central Italy: Annali di Geofisica, v. 39, p. 925–940. Galadini, F., and Galli, P., 2000, Active tectonics in the central Apennines (Italy)—Input data for seismic hazard assessment: Natural Hazards, v. 22, p. 225–268, doi: 10.1023/A:1008149531980. Galadini, F., and Galli, P., 2001, Archaeoseismology in Italy: Case studies and implications on long-term seismicity: Journal of Earthquake Engineering, v. 5, p. 35–68, doi: 10.1142/S1363246901000236. Galadini, F., and Galli, P., 2004, The 346 A.D. earthquake (central-southern Italy): An archaeoseismological approach: Annals of Geophysics, v. 47, p. 885–905. Galli, P., and Galadini, F., 2001, Surface faulting of archaeological relics: A review of case histories from Dead Sea to the Alps: Tectonophysics, v. 335, p. 291–312, doi: 10.1016/S0040-1951(01)00109-3. Galli, P., and Galadini, F., 2003, Disruptive earthquakes revealed by faulted archaeological relics in Samnium (Molise, southern Italy): Geophysical Research Letters, v. 30, p. 1266, doi: 10.1029/2002GL016456. Galli, P., and Naso, J., 2009, Unmasking the 1349 earthquake source (southern Italy) Paleoseismological and archaeoseismological indications from the Aquae Iuliae fault: Journal of Structural Geology, v. 31, p. 128–149, doi: 10.1016/j.jsg.2008.09.007.

242

Galli et al.

Galli, P., Galadini, F., and Pantosti, D., 2008, Twenty years of paleoseismology in Italy: Earth-Science Reviews, v. 88, p. 89–117, doi: 10.1016/ j.earscirev.2008.01.001. Giuliani, R., D’Agostino, N., D’Anastasio, E., Mattone, M., Bonci, L., Calcaterra, S., Gambino, P., and Merli, K., 2009, Active extension and strain accumulation in the Molise region (Southern Apennines, central Italy): Bollettino di Geofisica Teorica ed Applicata, v. 50, p. 145–156. Karcz, I., and Kafri, U., 1978, Evaluation of supposed archaeoseismic damage in Israel: Journal of Archaeological Science, v. 5, p. 237–253, doi: 10.1016/0305-4403(78)90042-0. Mantenuto, S., and D’Agostino, N., 2007, Analysis of the active extension in the Northern Apennines (Umbria-Marche, Italy) using CGPS measurements, in Slejko, D., Riggio, A.M., and Santulin, M., eds., 26th Gruppo Nazionale di Geofisica della Terra Solida Abstracts: Trieste, p. 100–101. Mantenuto, S., Bonci, L., Calcaterra, S., D’Agostino, N., Giuliani, R., Mattone, M., and Merli, K., 2007, Analysis of active extension in the Central Apennines (Abruzzo, Italy) using GPS measurements: European Geosciences Union General Assembly, Wien, Vienna, 15–20 April, Abstracts, v. 9, p. 4341. MedNet, 2008, Mediterranean Very Broadband Seismographic Network: http:// mednet.rm.ingv.it/ (accessed December 2008). Mommsen, T., 1883, Inscriptiones Bruttiorum, Lucaniae, Campaniae, Siciliae, Sardiniae Latinae, Pars Prior: Berlin, G. Reimer, 1229 p. Pantoni, A., 1961, L’editto augusteo sull’acquedotto di Venafro e una sua replica alle fonti del Volturno: Rendiconti Pontificia Accademia di Archeologia, v. 33, p. 155–172. Patacca, E., and Scandone, P., 2007, Geology of the Southern Apennines: Bollettino della Società Geologica Italiana, v. 7, p. 75–119. Piro, S., Goodman, D., and Nishimura, Y., 2003, The study and characterization of Emperor Traiano’s Villa (Altopiani di Arcinazzo-Roma) using high-

resolution integrated geophysical surveys: Archaeological Prospection, v. 10, p. 1–25, doi: 10.1002/arp.203. Rizzo, E., Chianese, D., and Lapenna, V., 2005, Integration of magnetometric, GPR and geoelectric measurements applied to the archaeological site of Viggiano (southern Italy, Agri Valley-Basilicata): Near Surface Geophysics, v. 3, p. 13–19. Serpelloni, E., Casula, G., Galvan, A., Anzidei, M., and Baldi, P., 2006, Data analysis of permanent GPS networks in Italy and surrounding regions: Application of a distributed processing approach: Annals of Geophysics, v. 49, p. 1073–1104. Stein, R.S., King, G.C.P., and Rundle, J.B., 1988, The growth of geological structure by repeated earthquakes: 2. Field examples of continental dipslip faults: Journal of Geophysical Research, v. 93, p. 13,319–13,331, doi: 10.1029/JB093iB11p13319. Stiros, S.C., 1996, Identification of earthquakes from archaeological data: Methodology criteria and limitations, in Stiros, S., and Jones, R.E., eds., Archaeoseismology: British School at Athens, Fitch Laboratory Occasional Paper 7, p. 119–152. Wells, D.L., and Coppersmith, K.J., 1994, New empirical relationships among magnitude, rupture length, rupture width, rupture area, and surface displacement: Bulletin of the Seismological Society of America, v. 84, p. 974–1002. Working Group CPTI (Catalogo Parametrico dei Terremoti Italiani), 2004, CPTI Catalogo Parametrico dei Terremoti Italiani (version 2004, CPTI04): Istituto Nazionale di Geofisica e Vulcanologia Sezione di Milano, http:// emidius.mi.ingv.it/CPTI/ (last revision 27 May 2004).

MANUSCRIPT ACCEPTED BY THE SOCIETY 19 MAY 2010

Printed in the USA

The Geological Society of America Special Paper 471 2010

Timing of earthquake ruptures at the Al Harif Roman aqueduct (Dead Sea fault, Syria) from archaeoseismology and paleoseismology Mohamed Reda Sbeinati Department of Geology, Atomic Energy Commission, Qasr El Khair, Damascus, Syria, and Laboratory of Global Geodynamics, Institut de Physique du Globe, UMR 7516, 5 rue René Descartes, 67084 Strasbourg, France Mustapha Meghraoui* Laboratory of Global Geodynamics, Institut de Physique du Globe, UMR 7516, 5 rue René Descartes, 67084 Strasbourg, France Ghada Suleyman Directorate General of Antiquities and Museums, Department of Archeology and Archeoseismology, Damascus, Syria Francisco Gomez Department of Geological Sciences, University of Missouri, Columbia, Missouri 65211, USA Pieter Grootes Marie-Josée Nadeau Leibniz-Labor für Altersbestimmung und Isotopenforschung, Christian-Albrechts Universität, Max-Eyth Str. 11-13, D-24118 Kiel, Germany Haithem Al Najjar Department of Geology, Atomic Energy Commission, Qasr El Khair, Damascus, Syria Riad Al-Ghazzi† Higher Institute for Applied Sciences and Technology, PO Box 31983, Damascus, Syria

ABSTRACT We studied the faulted Al Harif Roman aqueduct, located on the north–trending, ~90-km-long Missyaf segment of the Dead Sea fault, using four archaeological excavations, three paleoseismic trenches, and the analysis of six tufa cores. Damage to the aqueduct wall exhibits successive left-lateral fault offsets that amount to 13.6 ± 0.2 m since the aqueduct construction, which is dated younger than 65 B.C. Radiocarbon dating of sedimentary units in trenches, building cement of the aqueduct wall, and tufa cores constrain the late Holocene aqueduct history. The building stone types,

*Corresponding author: [email protected]. † Current address: Syrian Virtual University, Ministry of Higher Education Building, Damascus, Syria. Sbeinati, M.R., Meghraoui, M., Suleyman, G., Gomez, F., Grootes, P., Nadeau, M.-J., Al Najjar, H., and Al-Ghazzi, R., 2010, Timing of earthquake ruptures at the Al Harif Roman aqueduct (Dead Sea fault, Syria) from archaeoseismology and paleoseismology, in Sintubin, M., Stewart, I.S., Niemi, T.M., and Altunel, E., eds., Ancient Earthquakes: Geological Society of America Special Paper 471, p. 243–267, doi: 10.1130/2010.2471(20). For permission to copy, contact editing@ geosociety.org. © 2010 The Geological Society of America. All rights reserved.

243

244

Sbeinati et al. related cement dating, and tufa deposits of the aqueduct indicate two reconstructionrepair episodes in A.D. 340 ± 20 and A.D. 720 ± 20. The combined analysis of trench results; successive building and repair of aqueduct wall; and tufa onsets, growths, and interruptions suggests the occurrence of four faulting events in the last ~3500 yr, with a cluster of three events in A.D. 160–510, A.D. 625–690, and A.D. 1010–1210, the latter being correlated with the 29 June 1170 large earthquake. Our study provides the timing of late Holocene earthquakes and infers a lower and upper bound of 4.9–6.3 mm/yr slip rate along the Missyaf segment of the Dead Sea fault in Syria. The inferred successive faulting events, fault segment length, and related amount of coseismic slip yield Mw = 7.3–7.5 for individual earthquakes. The identification of the temporal cluster of large seismic events suggests periods of seismic quiescence reaching 1700 yr along the Missyaf fault segment.

INTRODUCTION Large strike-slip faults are continental tectonic structures and sources of seismic strain release during recurrent large earthquakes. The ~850-km-long Dead Sea fault constitutes a northsouth–trending plate boundary that accommodates most of the left-lateral active deformation between the African (Sinai subplate) and Arabia plates (Garfunkel et al., 1981; Barazangi et al., 1993; Figs. 1A and 1B). The total left-lateral offset along the fault reaches 105 km, of which ~45 km come from post-Miocene slip as a result of seafloor spreading in the Red Sea (Quennell, 1984). However, the northern Dead Sea fault shows less than 25 km leftlateral post-Miocene offset (Trifonov et al., 1991), the ~20 km missing slip being possibly absorbed by shortening along the Palmyrides fold belt (Chaimov et al., 1990). Kinematic models of the northern Dead Sea fault imply a transpressional fault system that suggests an oblique relative plate motion and relative rotation at ~31.1°N, 26.7°E at 0.40° ± 0.028 m.y.–1 (Westaway, 2004; Gomez et al., 2006). The northern section of the Dead Sea fault (i.e., in Lebanon and Syria) can be considered among the main seismogenic zones in the region, since it has a long (since 1365 B.C.), rich, and well-documented history of large destructive earthquakes that severely damaged many ancient cities (Fig. 1A; Ambraseys and Melville, 1988a; Guidoboni et al., 1994; Sbeinati et al., 2005). In contrast, the instrumental seismicity during the last century along the plate boundary is of low level and does not reflect the hazardous nature of the fault (Salamon et al., 2003). The long-term faulting behavior needs to be investigated, and a better constraint on the rate of active faulting is required for seismic hazard assessment. The Dead Sea fault has been the source of numerous large earthquakes with surface faulting in historical time (Ambraseys and Jackson, 1998). Although no recent surface ruptures have been observed, combined analyses of historical seismology, paleoseismology, and archaeoseismology contributes to a better understanding of the relationship between large historical earthquakes (Mw >7) and fault segments. The most recent large earthquake reached Mw 7.2 and took place on 22 November 1995 offshore in the Gulf of Aqaba at the southern end of the Dead Sea fault. His-

torical earthquake-faulting–related studies include, from north to south (Fig. 1A), the 1408 earthquake and Jisr-Al-Shuggur fault (Ambraseys and Melville, 1988a), the 1157 and 1170 earthquakes and Apame and Missyaf fault segments, respectively (Meghraoui et al., 2003; Sbeinati et al., 2005), the 1202 earthquake and the Yammouneh fault (Ambraseys and Melville, 1988b; Ellenblum et al., 1998; Daeron et al., 2005, 2007), the 1759 earthquake sequence and the Serghaya-Rachaya fault branches (Gomez et al., 2003; Nemer et al., 2008), the 1837 earthquake and the Roum fault branch of the Lebanese restraining band (Nemer and Meghraoui, 2006), the A.D. 749 earthquake and the Jordan Valley fault (Marco et al., 2003; Ferry et al., 2007), and the 1068 earthquake and south Araba Valley fault (Zilberman et al., 2005). The study of historical seismic events of the Dead Sea fault and related area of maximum damage is associated with investigations into the possible extent of surface ruptures and related major geometrical barriers. Earthquake parameters that include individual or cumulative left-lateral offsets and rate of slip can be obtained from paleoearthquake studies along the Dead Sea fault. Numerous fault slip rates have been inferred from offset geological units and geomorphologic features along the Dead Sea fault, and the more recent investigations including stream offsets and paleoseismic studies yield 4–7 mm/yr measured at time scales of ~10–100 k.y. (Garfunkel et al., 1981; Ginat et al., 1998; Klinger et al., 2000; Niemi et al., 2001; Daeron et al., 2004; Gomez et al., 2007; Ferry et al., 2007; Karabacak et al., 2010) and younger than 10 ka (Marco et al., 2005; Meghraoui et al., 2003; Gomez et al., 2003; Akyuz et al., 2006). Although an accurate measurement of the present-day active deformation across the Dead Sea fault requires a dense geodetic network combined with consistent block models, the 3–6 mm/yr global positioning system (GPS) velocities appear to be comparable to the geologic rate of slip (McClusky et al., 2003; Wdowinski et al., 2004; Reilinger et al., 2006; Gomez et al., 2007; Le Beon et al., 2008; Alchalbi et al., 2009). The Dead Sea fault crosses regions with abundant archaeological sites that evidence records of direct (fault offsets) or indirect (damage to building) coseismic features. Previous studies of archaeological sites from field investigations or textual documents

Timing of earthquake ruptures at the Al Harif Roman aqueduct (Dead Sea fault, Syria)

245

Figure 1. (on this and following page). (A) Seismicity (historical before 1900 and instrumental until 2004) along the Dead Sea fault (data from merged ISC [International Seismological Centre], EMSC [Euro-Mediterranean Seismological Centre], and the APAME [Archeo-PAleoseismology in the Middle East] Project catalogues). Focal mechanism solutions are from Harvard centroid moment tensor (CMT) catalogue.

have revealed the occurrence of “earthquake storms” probably associated with the Dead Sea fault (Nur and Cline, 2000). Indirect earthquake features are, however, very often problematic, and, unless dedicated to the specific study of known historical earthquake damage (Stiros and Jones, 1996; Marco, 2008), most of archaeological reports can hardly provide usable earthquake parameters (Ambraseys, 2006). Recent studies that combine archaeoseismic excavations and paleoseismic trenching provide some constraints of the left-lateral strike-slip movements and related past earthquake events. The Jordan fault segment and related past earthquake ruptures offset the Vadum Jacob Crusader

Castle and Holocene deposits visible in trenches at Beyt Zayda near the Sea of Galilee, yielding 3–4 mm/yr slip rate (Ellenblum et al., 1998; Marco et al., 2005). North of our study area, archaeological sites are widespread in the Amik Basin, where the fault crosses the ~5000 B.C. Tell Sicantarla and reveals 42.4 ± 1.5 m cumulative left-lateral movement, thus yielding 6.0 ± 0.2 mm/yr slip rate (Altunel et al., 2009). Previous archaeo-paleoseismic work on the faulted Al Harif Roman aqueduct revealed 13.6 m left-lateral offset and 6.9 ± 0.1 mm/yr slip rate that result from at least three earthquakes (Meghraoui et al., 2003). The early aqueduct study and trench

246

Sbeinati et al.

Figure 1. (continued). (B) Fault zone (black line; Meghraoui et al., 2003; Gomez et al., 2003; Elias, 2006; Nemer et al., 2008) and global positioning system (GPS) velocities (Eurasia fixed; Reilinger et al., 2006; Gomez et al., 2007; Le Beon et al., 2008; Alchalbi et al., 2009) emphasizing the leftlateral movements between the Sinai block and Arabia plate. Thick line is strike-slip fault; thin line is thrust fault. DSF—Dead Sea Fault, EAF— East Anatolian Fault, KF—Karasu Fault.

A shed light on the relationships between the Roman building and repeated faulting events but left open questions on (1) the earthquake events scenarios and related reconstruction and repair of the aqueduct after each coseismic slip; (2) the estimated longterm averaged slip rate versus a temporal cluster of seismic events over the past 2000 yr and its comparison to the present-day geodetic rate; (3) the constraint of the ~800-yr-long temporal and spatial seismic gap on the Missyaf segment and the recurrence interval of large earthquakes along the northern Dead Sea fault. In addition to the paleoseismic trenching and archaeoseismic excavations, tufa accumulation since the aqueduct construction may constitute a real archive of the aqueduct history that records the successive earthquake damages. Our study infers that the

aqueduct remains, the related tufa deposits, and faulted Holocene sedimentary units contain comparable records of the most recent surface rupturing events along the Missyaf fault segment. In this paper, we present the study of the faulted Al Harif Aqueduct site using archaeological excavations and paleoseismic trenching across the fault zone coupled with total station surveys and the coring of tufa accumulation on the aqueduct walls. We first describe the geomorphologic features and clear late Quaternary active tectonics of the fault zone that belong to the Missyaf segment of the Dead Sea fault. Archaeological excavations of the aqueduct walls and bridge combined with three trenches dug across the nearby fault zone and related radiocarbon dating illustrate the timing of successive faulting episodes. The dated

Timing of earthquake ruptures at the Al Harif Roman aqueduct (Dead Sea fault, Syria) onset, major discontinuities, and interruptions of tufa cores are correlated to the faulting events. The analysis and interpretations of earthquake damage with probable rebuilding phase of aqueduct wall and coseismic ruptures constrain the timing of successive faulting events. The Holocene faulting activity and related seismic cycle of the Missyaf fault segment reveal the long-term seismic strain release and hence determine the potential for a future large earthquake along the Dead Sea fault. ACTIVE FAULTING AND SEISMOTECTONIC SETTING OF THE MISSYAF SEGMENT The Missyaf fault segment is a section of the northern Dead Sea fault located in the western coastal part of Syria (Figs. 1B,

247

2, and 3). The ~200-km-long northern Dead Sea fault is made of (1) a 90 ± 10-km-long linear fault zone, i.e., the Missyaf segment, limited by the Al Boqueaa and the Lebanese restraining bend to the south and the Ghab pull-apart basin to the north, and (2) the ~10-km-wide Ghab pull-apart basin and related fault accompanied in its northern termination by a complex system of fault branches when reaching the Amik Basin and Karasu Valley in Turkey (Figs. 2 and 3). The Missyaf fault segment has a nearly north-south linear trend that limits the coastal ranges to the west from the MesozoicCenozoic plateau to the east (Fig. 3). Near Al Boqueaa Basin, the fault affects the Neogene basaltic formation of the Sheen Mountains. Further north, the fault crosses the Neogene volcanic and sedimentary formations and shows ~10–50-m-wide gouge zone,

Figure 2. Major historical earthquakes (white dots) and areas of maximum damage (shaded) for the 29 June 1170 earthquake (Io = IX using EMS98 intensity definition of Grünthal, 1998) as recorded along the northern Dead Sea fault (local intensities are from Guidoboni et al., 2004b; Sbeinati et al., 2005). The shaded area of maximum damage (Io = IX) for the A.D. 1170 large earthquake is along the Missyaf fault segment and Ghab Basin (see also Figs. 1B and 3 for legend) and overlaps with the maximum intensity VIII (MSK, black dashed line) as drawn by Ambraseys (2009). DSF—Dead Sea Fault, EAF—East Anatolian Fault, KF— Karasu Fault.

248

Sbeinati et al.

Figure 3. The 90-km-long Missyaf fault segment and the Al Harif Roman aqueduct site. The background topography (SRTM 30 arc posting digital elevation model; Farr and Kobrick, 2000) clearly delineates the fault segment (arrowheads) in between the Ghab and Al Boqueaa pull-apart basins. The Roman aqueduct at Al Harif (see also Fig. 4) was designed to bring freshwater from western ranges to Apamea and Shaizar. LRB—Lebanese restraining bend.

breccias, and rupture planes that affect the Mesozoic limestone of the coastal ranges. From Missyaf city to the Ghab Basin area, the fault bounds the Mesozoic limestone mountain range to the west from the Quaternary basins and Mesozoic-Cenozoic Aleppo plateau to the east (Dubertret, 1955). The left-lateral fault exhibits a clear geomorphologic expression along strike and neotectonic features consistent with the structural characteristics of the Al Boqueaa and Ghab pullapart basins (Fig. 3). The left-lateral movements are indicated by the en-echelon right stepping fault strands, faulted alluvial fans, deflected large and small streams that flow from the western mountain range, and shutter ridges made of either volcanic or limestone units (Figs. 3 and 4). The left-lateral slip is also expressed in outcrops where fault breccias in limestones also display en-echelon structures. Estimated from main channel deflections, observed in aerial photographs or satellite images, or measured using total station, systematic left-lateral offsets visible at different scales range from as low as 9 ± 1 m along strike to a few hundred meters (Figs. 3 and 4). The instrumental seismicity along the Missyaf fault segment is scarce in comparison with that of the Lebanese restraining bend to the south or the Karasu Valley and junction with the East

Anatolian fault to the north (Figs. 1A and 1B). Although the fault zone corresponds to the Africa-Arabia plate boundary, the instrumental seismicity is low level, and magnitudes (Ml) are less than 4.5. Focal mechanisms (CMT Harvard) of the few events with Mw > 4.5 in the northern Dead Sea fault show strike-slip faulting with predominant north-south–trending left-lateral fault plane mixed with normal faulting solutions near the pull-apart basins. The historical seismicity along the northern Dead Sea fault reports the occurrence of large and destructive earthquakes in 1365 B.C., A.D. 115, 526, 859, 1063, 1139, 1156, 1170, and 1408 (Ambraseys and Melville, 1988a; Guidoboni et al., 2004a, 2004b; Sbeinati et al., 2005). Only a few historical contemporaneous manuscripts, however, account for accurate damage distribution and sometimes for coseismic surface breaks with enough details that allow the correlation with fault segments (Ambraseys and Melville, 1988b). Most contemporaneous manuscripts and inscriptions from Byzantine, Crusader, and Arabic sources provide accurate damage descriptions of castles, churches and mosques, villages, and cities, often accompanied with an estimate of casualties. Based on our work on the catalogue of historical earthquakes of Syria and paleo-archaeoseismic investigations (Mouty and Sbeinati, 1988; Meghraoui et al., 2003;

Timing of earthquake ruptures at the Al Harif Roman aqueduct (Dead Sea fault, Syria)

249

Figure 4. (A) Satellite view from Google Earth showing offset Al Harif aqueduct (black arrow) along the Dead Sea fault (white arrows); and (B) local geomorphologic framework of the aqueduct site as interpreted from part A, indicating a shutter ridge (Mesozoic limestone east of the fault) and ~200 m of left-lateral offset. Blue arrow is for stream flow. See Figure 5 for the detailed aqueduct map and location of excavations and trenches.

Sbeinati et al., 2005), we have shown that the damage distribution associated with the 29 June 1170 earthquake suggests a correlation with the Missyaf fault segment. Using numerous historical documents that report the 1170 earthquake damage, Guidoboni et al. (2004b) provided a comparable damage distribution and suggested an epicentral location on the Missyaf fault. However, overestimated damages at Aleppo (from misinterpretations of the Arabic chronicler Ibn Al Athir [1160–1233]) and poorly constrained seismotectonics inferences brought the authors to the erroneous conclusion that the 90 ± 10-km-long Missyaf fault segment alone could not have generated the Mw 1170 > 7 for the seismic event (Guidoboni et al., 2004b). Zones of maximum damage should be identified primarily from contemporary eyewitness accounts in manuscripts, corroborated by present-day field investigations on the active fault and damage of ancient buildings. Furthermore, the geometrical structures (i.e., the Al Boqueaa pull-apart basin and Lebanese restraining bend to the south and the Ghab pull-apart basin to the north) that limit the fault segment are major obstacles to a coseismic rupture propagation and, hence, constrain the earthquake fault dimension. ARCHAEOSEISMOLOGY AND PALEOSEISMOLOGY Site Description The Al-Harif aqueduct is located ~4 km north of the city of Missyaf, immediately west of a limestone shutter ridge and related ~200 m left-lateral stream deflection (Figs. 3 and 4). According to the remaining aqueduct walls and related mills

in the region, the aqueduct was built during the Roman time (younger than 65 B.C. in the Middle East) to drain freshwater collected from springs of the western mountain range to the eastern semiarid plains. The remaining ruins of the aqueduct suggest an ~40-km-long construction that may have included several bridges over streams and landscape gorges. The aqueduct building description and related age have not been reported so far in any archive, manuscript, or in the literature. There is, however, an interesting anecdotal story from the local tradition that it was built by a local prince to supply potable water to Apamea and/or Sheizar cities, located northeast of the aqueduct (Fig. 3). Apamea during that time was the most famous and strategic city during the Hellenistic and Roman period, whereas Sheizar is known to have been an important political and military fortress during the Middle Ages (Ibn Al Athir, 1982). In their description of the Dead Sea fault in Syria, Trifonov et al. (1991) mentioned the existence of a faulted aqueduct near the city of Missyaf, but neither the precise location nor the accurate amount of offset walls was given. However, this early tectonic observation was helpful and allowed us to discover the site and consider a detailed study (Meghraoui et al., 2003), which is extended here using combined methods in archaeoseismology, paleoseismology, and tufa investigations. In addition, a microtopographic survey of measurements accompanied all field studies (Fig. 5). Previous investigations on the aqueduct (Meghraoui et al., 2003) established: (1) an evaluation of its age based on an account of the large size blocks, the dating of sedimentary units below the aqueduct wall foundation, and dating of early tufa deposits on

250

Sbeinati et al.

Figure 5. Microtopographic survey (0.05 m contour lines) of the Al-Harif aqueduct and related flat alluvial terrace. The aqueduct (thin blue crosses) shows a total of 13.6 ± 0.20 m left-lateral slip along the fault zone (Meghraoui et al., 2003). Roman numbers indicate archaeoseismic excavations (in reddish and orange, labeled I to IV) and letters indicate paleoseismic trenches (in gray and black, labeled A, B, C, and E). The dragged wall fragment is located between excavation IV and trench E and is marked by a dense cluster of survey points.

the aqueduct wall, and (2) the identification of the seismic faulting origin of damage in nearby trench A. The building style, with typical bridge arch and large stone size disposition (Opus caementum), suggested a Roman age, which was confirmed by the radiocarbon dating of sedimentary layers below the walls and the early tufa deposits on the walls. The faulted aqueduct revealed 13.6 ± 0.20 m of total left-lateral offset and called for detailed investigations on the characteristics and history of successive fault movements. The aqueduct design, with an open canal on top of the 4-m-high wall, allowed freshwater and carbonate-saturated water to overflow and induce significant tufa accumulation from 0.30 m to 0.83 m in section (Figs. 6A, 6B, and 7). The carbonate-rich and cool water collected from the nearby western range is associated with a semiarid and karstic area of the Mesozoic limestone (Fig. 4) that favors rapid carbonate precipitation and tufa accumulation. The tufa deposits show successive growths of lamination carbonate with high porosity, banded texture, and rich organic

encrustations (Ford and Pedley, 1996). Field observations show that tufa accumulation developed on both eastern and western sections (from the fault line), but only on the north-facing wall, likely due to a slight tilt of the damaged aqueduct wall, probably after the two first earthquakes (Fig. 7). The following paragraphs present the field investigations, which consisted of: (1) four archaeoseismic excavations near the aqueduct walls and remains, (2) four paleoseismic trenches across the fault zone and the alluvial sediments, and (3) four cores (two cores were previously studied in Meghraoui et al., 2003) of tufa deposits collected from different sections of the aqueduct. More than 200 samples of organic matter, charcoal fragments, and tufa core pieces were taken for radiocarbon analysis in order to characterize the timing of successive faulting and related damage of the aqueduct construction. All radiocarbon dating were calibrated (2σ range, 95.4% probability density) using Oxcal v4.0 (Bronk Ramsey, 2001) and INTCAL04 calibration curve of Reimer et al. (2004).

Timing of earthquake ruptures at the Al Harif Roman aqueduct (Dead Sea fault, Syria)

251

Figure 6. (A) Schematic sketch of the aqueduct and locations of the selected cores BR-3, BR-5, and BR-6; BR-4 core sample consists of tufa accumulations at the location of the missing (broken) piece of the aqueduct wall near the fault. Mosaic of the archaeological excavation I is detailed in Figure 8B (see also location in Fig. 5). (B) Core section BR-4 showing the limit between the stone wall and tufa deposits.

Archaeoseismic Excavations The remaining aqueduct construction forms an ~50-m-long, ~5-m-high, and 0.60-m-thick wall that includes an ~15-m-high arch bridge in its eastern section (Figs. 5 and 6A). The outer part is coated by a thick layer of tufa deposits, probably due to a long period of freshwater flow. The construction material that may vary

with the successive building and repair ages is made of: (1) largesize limestone blocks (Opus quadratum, 1.0 m × 0.5 m × 0.5 m; see also http://www.romanaqueducst.info/aquasite/), similar to the typical Roman archaeological constructions and visible at the lower bridge (pier section) and wall sections, (2) medium-size limestone blocks (Opus incertum; 0.50 m × 0.30 m × 0.30 m), which form the foundation or the upper half wall section and show visible small portions of cement, and (3) small sizes of mixed stones of irregular shape with significant portions of mortar (cement), mostly visible in the apparently rebuilt part of the wall. Figures 5 and 6A also show a detached small piece of the aqueduct wall made of small-size stones and related cement ~3.5 m away from the eastern wall. Therefore, four areas (noted I to IV in Fig. 5) were excavated near the aqueduct using proper archaeological methods. The large excavation I was dug on the fault zone near the dragged wall fragment, in the area between the eastern and western aqueduct walls (Figs. 5, 8A, and 8B). The purpose of excavation is here to study the relationships between the fault zone and aqueduct. The excavation that has ~4.5 × 4.5 m surface and ~0.6 m depth exposed missed parts of the aqueduct. A buried and fallen wall piece rotated and dragged parallel to the fault and a remaining wall piece in an oblique position between two shear zones were discovered. The buried wall fragments are not

252

Sbeinati et al.

Figure 7. Schematic sections of the aqueduct western wall and related tufa deposits (B, C, D, and E indicate earlier core sections of tufa deposits (Meghraoui et al., 2003). Tufa samples AQ-Tr-B13 and AQ-Tr-D5 (Table 1) are from cores B and D, respectively. The right and left vertical sections show the relative tufa thickness of the originally built part (with Opus caementum and quadratum stones) and the rebuilt part, respectively. The plan view indicates the variation of tufa deposition and shows the core distribution and related thickness along the western wall of the aqueduct.

comparable to the Opus caementum (quadratum) of the original construction and suggest a rebuilding phase. The excavation floor displays oriented gravels and pebbles that mark the shear zones and related fault branches also visible in the inner trench section E (Figs. 8B and 8C). We collected four samples in the fallen wall sections labeled A, B, and C of excavation I (Fig. 8B): Two cement samples (AQ-CS-1 and AQ-CS-4) found in between building stones are made of typical medieval rubble mortars (mainly mud, gypsum, and lime); the two other samples (AQ-CS-3-2 and AQ-CS-3-3) are tufa deposits preserved on building stones. All four samples contained enough organic matter to allow radiocarbon dating (Table 1). Two dates of cement yield A.D. 532–641 (section A, AQ-CS-4) for the large fallen wall in excavation I and A.D. 650– 780 (section C, AQ-CS-1) for the wall fragment piece in between the walls (Fig. 8B). In addition, two tufa deposits on wall stones provide consistent ages A.D. 560–690 (section B, AQ-CS-3-2) and A.D. 639–883 (section C, AQ-CS-3-3) with cement ages. The two different cement dates of the fallen wall and dragged

Figure 8. (on this and following page). (A) View from the western aqueduct wall, the dragged wall piece, buried wall, and eastern wall (string grid is 1 m × 1 m). Log of trench-excavation E is in Figure 7C.

Figure 8. (continued). (B) Mosaic of excavation I exhibits the main fallen wall (A and B) and dragged wall piece (C), scattered wall pieces and the fault zone; note also location of cement sample CS-1-4 (see text for explanation). (C) Trench E (excavation I, north wall) exposes faulted sedimentary units below the archaeological remains and wall fragment C visible in bottom of Figure 8B; fz—fault zone; sedimentary units are similar to those of trenches A, B, and C (see also Fig. 10); and dating characteristics are in Table 1. a— present-day soil and alluvial terrace (plough zone), d—reddish alluvial fine gravel, e—dark-brown silty clay (with rich organic matter), f—gravels and pebbles in silty-clay matrix, g—massive gey clay with scattered gravels.

AQ-TB-3

AQ-TB-4

KIA 23863

KIA 23857

EH III-7S

EH III-6S

EH III-3S

EH III-10S

KIA 23895

KIA 23896

KIA 23897

KIA 23900

KIA 23893

AQ-Tr-B13

AQ-Tr-B13

AQ-CS-1

AQ-CS-3-2

KIA 16628

KIA 16628

KIA 22189

KIA 22191

Al Harif—Cores & cement KIA 16627 AQ-Tr-D5

EH II-5S

EH III-8S

KIA 23909

EH II-12S

KIA 23915

EH II-18S

EH II-10S

KIA 23911

KIA 23920

EH II-11S

KIA 23917

KIA 23910

EH II-2N

EH II-16S

KIA 23880

EH II-7S

AQ-TB-2

KIA 23862

KIA 23903

AQ-TB-1

KIA 23861

AQ-TC-S3

AQ-TA-4

KIA 23855

KIA 23860

AQ-TA-3

KIA 23856

AQ-TC-S1

BAL-TA-N47

KIA 14265

AQ-TC-S2

BAL-TN-61

KIA 14268

KIA 23858

BA-TA-N31

KIA 14264

KIA 23859

EH-I-S7

EH-I-TA-S33

AA 43993

BAL-TA-N25

KIA 14262

AA 43995

BAL-TA-N27

KIA 14263

BAL-TA-N23

Sample name

Tufa, alkali residue

Cement, alkali residue

Tufa, humic acids

Tufa, acid residue

Tufa, acid residue

Charcoal, alkali residue

Charcoal, acid residue

Charcoal, acid residue

Charcoal, acid residue

Charcoal, acid residue

Seed, alkali residue

Charcoal, acid residue

Charcoal, alkali residue

Charcoal, alkali residue

Charcoal, acid residue

Charcoal, alkali residue

Charcoal, alkali residue

Charcoal, acid residue

Charcoal, alkali residue

Charcoal, alkali residue

Charcoal, alkali residue

Charcoal, alkali residue

Charcoal, alkali residue

Charcoal, alkali residue

Charcoal, alkali residue

Charcoal, alkali residue

Charcoal, alkali residue

Charcoal, alkali residue

Charcoal, alkali residue

Charcoal, alkali residue

Charcoal, alkali residue

Charcoal, alkali residue

Charcoal, alkali residue

Charcoal, alkali residue

Charcoal, alkali residue

Fraction

0.13 0.28

f (I–E) f (I–E) a (C)

I – Fallen wall

I – Wall piece

B, 0–5 cm

B, 0–5 cm

D, 0–5 cm

d? (B)

Fault zone (B)

Fault zone (B)

e (B)

e (B)

f (C)

f (C)

f (C)

f (C)

f (C)

b1 (C)

0.99

0.50

3.72

2.39

1.13

0.12

1.05

0.24

1.19

2.40

0.03

0.97

4.66

1.54

2.11

5.62

4.16

0.49

0.17

e (I–E)

a (C)

0.38

4.23

1.64

1.77

0.24

1.83

1.71

1.02

1.95

4.33

–

0.42

4.09

5.87

Analyzed carbon (mg)

e (III)

e (III)

e (III)

e (III)

e (II)

e (II)

g (A)

a (A)

a (A)

d (A)

e (A)

e (A)

f (A)

b & c (A)

Trench-excav. unit & core level

2.9%

0.4%

41.3%

2.7%

80.1%

5.5%

64.0%

6.3%

45.2%

38.4%

4.5%

9.3%

62.3%

57.5%

51.7%

69.0%

50.7%

128.9%

9.5%

1.6%

2.1%

4.0%

49.8%

25.9%

22.4%

4.7%

22.6%

3.6%

1.0%

63.3%

–

–

24.9%

67.4%

71.9%

Carbon (content %)

1400 ± 35

1314 ± 37

1880 ± 25

2030 ± 25

1863 ± 29

2680 ± 170

4375 ± 40

2390 ± 80

2215 ± 35

2110 ± 35

3420 ±570

2525 ± 40

2160 ± 30

2150 ± 30

2135 ± 30

1465 ± 30

280 ± 25

290 ± 40

2280 ± 70

2450 ± 140

1930 ± 110

2460 ± 60

2235 ± 30

2200 ± 40

2250 ± 30

2050 ± 70

2295 ± 30

7410 ± 45

4555 ± 40

875 ± 35

1287 ± 36

2195 ± 40

2090 ± 50

2335 ± 30

1015 ± 35

C date (yr B.P.)

14

C age

14

(Continued )

A.D. 560–690

A.D. 650–780

A.D. 70–230

B.C. 110–60 A.D.

A.D. 80–240

B.C. 1300–350

B.C. 3100–2980

B.C. 800–200

B.C. 410–90

B.C. 350–40

B.C. 3400–300

B.C. 800–510

B.C. 360–90

B.C. 360–60

B.C. 360–50

A.D. 540–650

A.D. 1510–1670

A.D. 1480–1800

B.C. 550–100

B.C. 900–200

B.C. 200–400 A.D.

B.C. 770–400

B.C. 390–200

B.C. 390–160

B.C. 400–200

B.C. 350–130 A.D.

B.C. 410–210

B.C. 6400–6100

B.C. 3490–3090

A.D. 1030–1250

A.D. 650–810

B.C. 390–160

B.C. 350–30 A.D.

B.C. 520–350

A.D. 960–1060

Calibrated B.C./A.D. (95.4%)

TABLE 1. SAMPLE LIST AND RADIOCARBON DATING (ACCELEROMETER MASS SPECTROMETRY) AT THE AL HARIF AQUEDUCT SITE

Al Harif—Trenches KIA 14261

Sample ID

A.D. 900–1160 1020 ± 35 1.2% 0.58 6-8, 38–39 cm Tufa, alkali residue BR-6-8/SYR Al Harif KIA 26059

Notes: All samples have been calibrated using the Oxcal program v. 3.5 (Bronk-Ramsey, 2001) and calibration curve INTCAL04 (Reimer et al., 2004), and adopted age ranges are equivalent to calibrated 2σ ranges (95.4%), in A.D. and B.C. Trench and excavation units from trenches A, B, and C and excavations I, I–E, II, and III appear in parentheses. Location of cores B and D is in Figure 7.

B.C. 400–250 A.D.

A.D. 890–1020 1090 ± 25

2020 ± 110 1.1%

40.5% 3.44

0.15 6-1, 0–0.5 cm

5-7, 32–33.5 cm Tufa, humic acids

Tufa, alkali residue

BR-5-7/SYR Al Harif

BR-6-1/SYR Al Harif

KIA 26058

BR-5-2/SYR Al Harif

KIA 26059

A.D. 540–980

B.C. 110–130 A.D. 1995 ± 45 1.7% 0.51 5-2, 4.5–5.0 cm

A.D. 530–660

KIA 26058

Tufa, alkali residue

1310 ± 110 22.6% 0.14 4-3, 23.5–24.5 cm BR-4-3/SYR Al Harif KIA 26057

Tufa, alkali residue

A.D. 770–940 1180 ± 20

1465 ± 35 6.7%

2.3% 2.28

0.71 4-1, 10.5–11 cm

3-4, 32.5–33 cm

BR-4-1/SYR Al Harif KIA 26057

Tufa, alkali residue

BR-3-4/SYR Al Harif KIA 26056

Tufa, alkali residue

A.D. 532–641

A.D. 410–600 1570 ± 35 17.3% 0.53 3-1, 0–0.5 cm BR-3-1/SYR Al Harif KIA 26056

Tufa, alkali residue

A.D. 639–883 1283 ± 44

1497 ± 24 0.3%

1.9% 0.57

5.33 I – Fallen wall

I – Fallen wall

AQ-CS-4 KIA 22192

Tufa, alkali residue AQ-CS-3-3 KIA 22191

Cement, alkali residue

C age

14 14

C date (yr B.P.)

Carbon (content %) Analyzed carbon (mg) Trench-excav. unit & core level Fraction Sample name Sample ID

TABLE 1. SAMPLE LIST AND RADIOCARBON DATING (ACCELEROMETER MASS SPECTROMETRY) AT THE AL HARIF AQUEDUCT SITE

Calibrated B.C./A.D. (95.4%)

Timing of earthquake ruptures at the Al Harif Roman aqueduct (Dead Sea fault, Syria)

255

wall fragment can be correlated to the new tufa deposits that testify for two rebuilding phases. The dated buried fallen wall in section B (CS-3-2) obtained from a thin (~5 cm) tufa accumulation correlates with the similarly fallen wall in section A and related cement date of CS-4 (Fig. 8B; and Table 1). In section C, the tufa deposits and related dated sample CS-3-3 correlates with the cement age of CS-1. The type and size of stones (opus incertum) and thin tufa accumulation in sections A and B suggest an early rebuilding phase postdating the first damaging event that may have occurred between the first and sixth centuries A.D. The different building layout of section C made of small sizes of mixed stones of irregular shape, and dating of cement sample CS-1 (A.D. 650–780) and tufa accumulation CS-3-3 (A.D. 639– 883) indicate a repair and rebuilding period postdating a second damaging event at the end of the Byzantine time and beginning of the Islamic period (seventh to eighth century A.D.). The damaged and dragged most recent wall section C along the fault indicates the occurrence of a third event, after which the aqueduct was definitely abandoned. Three small excavations II, III, and IV (1.5 m to 3.0 m long, 1.0 m wide and 1.50 m deep) were dug in the base layer of the western aqueduct wall in order to expose its foundation and related sedimentary units underneath that predate the early building phase (Figs. 5 and 9). Excavations II and III were dug under the wall section with maximum (> 0.80 m) and minimum (~0.30 m) thickness of tufa deposition, respectively (Figs. 9A and 9B). Excavation IV, already described in Meghraoui et al. (2003), exposed the faulted foundation of the missing section of the western wall edge. The wall foundation reaches 1 m depth and shows regular patterns of medium-size cut limestone blocks (0.50 m × 0.30 m × 0.30 m) built over a dark brown clayey layer (unit e). Charcoal samples collected in excavations I (trench E), II, and III from unit e yield 14C dates with ages spanning from approximately the third century B.C. to third century A.D. (see samples AQ-TA, TB, and TC in Table 1; Figs. 8C, 9A, and 9B). Although in these excavations, the age range of unit e seems quite large (probably due to detrital charcoal mixing), the younger age, i.e., 350 B.C. to A.D. 130 (sample AQ-TA-4), is consistent with other radiocarbon ages of unit e and related stratigraphic succession in trenches (see Paleoseismic Trenches herein). In excavation II, the large stone shape (Opus quadratum) with small amount of cementing material and pottery fragments found on the same level near the building base can be correlated with the early Roman era (Fig. 9A). Large stones and tufa thickness led us to consider this section of the aqueduct wall to be in original condition, i.e., probably undamaged by large earthquakes. Excavation III (1.65 m long, 1.0 m wide, and 1.2 m deep; Fig. 9B) is similar to excavations II and IV, but the 1-m-deep wall foundation and upper section show irregular shapes of mixed medium- and small-size cut limestone blocks (0.10 × 0.20 × 0.15 m). Excavation III was realized at the location of the thinnest tufa deposits (< 0.30 m). The size of stones, cement texture, and irregular shape of building wall suggest that this building section was rebuilt (Fig. 9B). The 14C dating of unit e below the

256

A

Sbeinati et al.

B

Figure 9. Excavations II (A) and III (B) that expose the aqueduct wall foundation (see also Fig. 5) and related sedimentary unit e underneath. The difference in the size of stones between excavation II (A) and excavation III (B) implies a rebuilding phase of the latter wall.

wall yielded a comparable age range to that obtained in excavations II (see AQ-TA, TB, and TC in Table 1). Trench section E (4.30 m long, 0.70 m wide, and 1.30 m deep; Figs. 5, 8A, 8B, and 8C) was dug within excavation I in order to see in section the fault zone that affects the archaeological floor units. The trench wall exposes similar sedimentary units to those visible in excavations II, III, and IV that are affected by two main fault branches of the shear zone visible in the floor layer of excavation I. The 14C dating of samples AQ-TC-S1, S2, and S3 of units f and e indicates 900 B.C. to A.D. 400 maximum and minimum age range, respectively (Fig. 8C; Table 1), which is comparable to the age range obtained in excavations II and III for unit e (Figs. 8B and 9A; Table 1). However, as here again the large age range can be due to charcoal mixing, the dating of unit e is obtained by comparison to the dated stratigraphic succession of units in trenches (see section Paleoseismic Trenches). Paleoseismic Trenches Two trenches, B and C (Figs. 5 and 10, trenches B and C), were dug across the Dead Sea fault north of the aqueduct in addition to the previously studied trench, A (Fig. 10A; Meghraoui et al., 2003). The two trenches exposed an ~1.5-m-wide fault zone that affects a succession of 2–3-m-thick fine and coarse alluvial sedimentary layers similar to the alluvial deposits of trench A. Alluvial units visible in all trenches exhibit here similar textures, structures, and color, and correspond to the same layers that belong to the same alluvial terrace. Although the three trenches A, B, and C may not expose a completed stratigraphic section, the comparisons among sedimentary units, faulting events, archaeoseismic observations, and tufa accumulation limit the possibility of a missing earthquake event that affected the aqueduct.

In trench B (south wall), the fault zone shows three main fault branches that affect sedimentary units g to d and form a negative flower structure. The central and western main branches are truncated by unit a, which forms a stratified 0.3–0.4-m-thick deposit of coarse gravels in a sandy matrix. The eastern fault branch is buried below unit d, made of well-sorted reddish fine gravels. Unit e, a 0.2–0.5-m-thick dark-brown siltclay, thickens toward east. Units f and g are made of scattered clasts in a massive clay matrix of dark-brown and light-brown color, respectively. Although intense warping and faulting are marked by contrasting color and texture of unit e, faulted sedimentary layers of this trench do not allow the identification of all faulting events. However, buried fault branches indicate a faulting event postdated by unit d (event Y), while the other fault branches show at least another faulting event (event Z) overlain by unit a. While clearly visible in other trench walls, event Y is here likely concealed by the complex fault branches truncated by unit a. Trench C (Fig. 10C) exposes a stratigraphic succession affected by at least five main fault branches (labeled I to V in Fig. 10C). From trench bottom, fault branch I, which affects unit g, is overlain by unit f. A similar observation can be made for fault branch II, which also affects all units below unit d. Furthermore, the trench wall exposes an ~0.60-m-thick well-stratified, coarse and fine gravel layer above unit e and across the fault zone. Unit d thins significantly west of fault branch III and is overlapped by relatively thick coarse gravel units, which display a mix of fine and coarse gravels between fault branches III and IV, and unit d shows a succession of well-stratified alluvial units west of fault branch IV (Fig. 10C). Taking into account its alluvial origin made of well-stratified fine and coarse gravels, west of fault branch IV, unit d is subdivided into d1, d2, d3,

Figure 10. Trench logs A, B, and C north of the aqueduct site (see location in Fig. 5). All trenches display the Dead Sea fault zone as a negative flower structure affecting all alluvial units below unit a. Calibrated 14C dates are in Table 1. Fault branches in trench C are labeled I to V (see text for explanation). The sedimentary units are very comparable and show three to four faulting events denoted W to Z (see text for explanation). Trench log A is in meters.

258

Sbeinati et al.

and d4. Faulting movements at this site allows truncation of unit d1 (equivalent to d east of fault branch III) and sedimentation of units d2 to d4 (in a likely small pull-apart basin). Unit d3 consists of an ~0.20-m-thick dark-brown silt-sand overlain by unit d4, which is made of light-brown fine silt-sand. Below the plough zone a2, the well-stratified unit a1 shows flat-laying pebbles and gravels and intercalated fine gravels covering previous units and the fault zone. Fault branches I to V in trench C indicate a negative flower structure that intersects a sedimentary sequence and reveal at least four faulting events (Fig. 10C): (1) Event W, identified on fault branch I, is older than 800–510 B.C. (EH II-18S) in the lowermost layers of unit f and is younger than unit g, which was dated with sample EH II-5S (3400–300 B.C.). (2) Next to fault branch II, buried below unit d, the vertical offsets between unit e and units d and d1 across fault branch III, and the absence of unit e between fault branches III and IV, determine the faulting event X between unit e and unit d. Since unit d overlies an erosional surface of unit e, faulting event X may have formed a depression (i.e., a small pull-apart basin) that allowed the deposition of d1 to d4 next to a thick unit d east of fault branch III. The faulting event X is here predated by 360–90 BC (EH II-12S), 360–50 BC (EH II-11S), and 360–60 BC (EH II-10S) of the uppermost layers of unit f (event X is postdated by sample EH I-TA-S33 of trench A). (3) Faulting event Y can be identified at the westernmost fault branch V between unit d2 and unit d3. The dating of sample EH II-16S in d3 postdates event X to younger than A.D. 540–650, which we consider as a reliable age, taking into account its high carbon content (event Y is predated by sample EH I-TA-S33 of trench A). (4) Faulting event Z corresponds to the main fault branches III and IV, which are overlain by the stratified unit a2 below the plough zone. Fault rupture IV affects unit d4 and indicates that the faulting event Z is older than radiocarbon age A.D. 1480–1800 (EH II-7S) and A.D. 1510–1670 (EH II-2N) of unit a2 and younger than unit d4. Summary of Faulting Events from Archaeoseismology and Paleoseismology The analysis of faulting events from the aqueduct (damage and reconstruction) and from trenches A, B, and C can be presented as following: 1. Event W is older than unit f (i.e., 800–510 B.C.) and younger than unit g (i.e., 3400–300 B.C.) of trench C. The bracket of event W is here difficult to assess since the detrital charcoal sample in unit f was not taken from the base of unit f. According to 14C dates, the faulting event can be estimated as younger than 3400 B.C. and older than 510 B.C. However, taking into account the rate of sedimentation in unit f, we may estimate a minimum age of 962 B.C. for event W. 2. Event X, the first faulting event that affected the aqueduct, is bracketed between the first and sixth centuries A.D. In trenches, a large bracket of this event is between

350 B.C. and A.D. 30 and A.D. 650–810 (as obtained from dated units of trench A). 3. Event Y, characterized from paleoseismology, appears to be older than A.D. 650–810 (unit d, trench A) and younger than A.D. 540–650 (unit d3 in trench C). The results of archaeoseismic investigations indicate that ages of CS-1 (A.D. 650–780) and tufa accumulation CS-3-3 (A.D. 639–883) postdate event Y. 4. Event Z is the last faulting event that affected the aqueduct, after which it was definitely abandoned. In trenches A and C, event Z is older than A.D. 1480–1800, A.D. 1510–1670, and A.D. 1030–1260 and younger than A.D. 960–1060. TUFA OF THE AL-HARIF AQUEDUCT The tufa thickness accumulated on the northern face of aqueduct wall suggests a continuous water flow during a relatively long period of time and may include the record of large earthquakes that affected the aqueduct. Hence, the relationships between tufa accumulation and earthquake events are established through the simultaneous major tufa interruptions and restarts observed in different cores. Except during major changes in the water-flow conditions, the permanent water flow coming from the nearby spring was responsible for the tufa accumulation that, in principle, is not interrupted on the western wall section (with regard to the fault). On the eastern wall section (and bridge) and broken pieces of western wall, however, the tufa accumulation was likely episodic due to the earthquake damage and related faulting events; new tufa accumulation appears in subsequent building-repair. Previous radiocarbon dating of early tufa deposits (A.D. 70–230 and A.D. 80–240; Table 1) postdated the initial construction of the aqueduct and revealed a Roman age consistent with the dates obtained from the archaeological and paleoseismic investigations (Meghraoui et al., 2003). Six tufa cores (named Tr-B13, Tr-D5, and BR-3, BR-4, BR-5, and BR-6) reaching the stone construction were collected from the aqueduct wall in order to date major catastrophic events and infer the relationship with large earthquakes (Fig. 11). Tr-B13 and Tr-D5 were previously collected and analyzed mainly to date the early tufa deposits, which provide the maximum age of the aqueduct construction (Meghraoui et al., 2003). A subsequent selection of core locations on both eastern and western sections of the aqueduct wall was performed to study the completed tufa accumulation and successive growth. Figures 6 and 7 show the drilled wall location with the early cores Tr-B13 and Tr-D5 and three cores (BR-4, BR-5, and BR-6) on the western wall and one core (BR-3) on the eastern wall next to the bridge. Cores BR-5 and BR-6 correspond to the thickest tufa section. BR-4 is on the eastern edge of the west aqueduct wall, a section probably exposed after earthquake damage that induced the collapse of a 2.5-m-long wall section next to the fault zone. Each core is described to illustrate fabric (structure, texture, and color) and lamination changes, which provide evidence of tufa precipitation

Timing of earthquake ruptures at the Al Harif Roman aqueduct (Dead Sea fault, Syria)

259

Figure 11. Synthetic description of cores with lithologic content and sample number for radiocarbon dating (see Table 1 and Fig. 6 for core locations); I stands for major interruption. The very porous tufa indicates major interruptions in tufa growth (e.g., a major interruption of core growth in BR-3 is visible at ~22 cm (Br-3-4 sample; see text for explanation). The correlation between major interruptions of tufa growth and faulting events in trenches and archaeoseismic building constrains the timing of repeated earthquakes along the Missyaf segment of the Dead Sea fault.

and successive growths (Fig. 11). Although marked by a high porosity, the cores were carefully drilled in order to preserve their structure and length continuity. An analysis in progress of cores using computer tomography (CT) and climatic-stratigraphy correlation details the physico-chemical and biochemical processes of tufa growth (Grootes et al., 2006). The cores show a variety of porous, dense, and biogenic tufa with growth laminae and stromatolitic markers of different colors. The end of tufa growth (i.e., very porous tufa in Fig. 11) and onset of biogenic tufa (indicating only a seasonal growth) can be interpreted as episodes of decreased accumulation, or a significant decrease in the chemical precipitation due a major change in the environmental conditions (Fig. 11). Discontinuities of tufa deposits marked by the interruption of core growths and initiation of biogenic tufa are interpreted as major changes in environment with a possible correlation with large earthquakes. The early tufa deposits on the aqueduct wall provide A.D. 70–230 and A.D. 80–240 (samples Tr-B13 and Tr-D5 in Table 1) ages, which postdate the aqueduct building and early function (Meghraoui et al., 2003). The tufa accumulation in BR-3 (core in eastern wall near the bridge, Fig. 6) started sometimes before A.D. 410–600 (sample Br 3-1, Table 1) and may have resulted from a repair of the aqueduct with water overflowing the eastern wall (and bridge) after a major damaging event. Similarly, the location of a growth interruption (very porous tufa, Fig. 11) in BR-5 at ~6 cm after Br-5-2 (110 B.C.–A.D. 130) and onset of biogenic tufa in BR-6 after Br-6-1 (400 B.C.–A.D. 250) coincide with the occurrence of the first damaging event X. In parallel, the beginning of BR-4 and tufa accumulation at the damaged eastern edge of the western wall (Fig. 6) and sample Br-4-1, dated A.D. 530–660 (Fig. 11;

Table 1), postdates the occurrence of a major damaging event. Both Br-3-1 and Br-4-1 postdate here the record of a major damaging event that affected the aqueduct. However, while BR-4 may have accumulated only after a major damage, BR-3 deposits could only have accumulated after the repair of the aqueduct. It implies that the first major damaging event on the aqueduct took place between A.D. 70–230 and A.D. 410–600. The interruption of tufa growth in BR-3 a few centimeters before sample Br-3-4, dated A.D. 770–940, probably resulted from a second damaging event. This observation coincides with the restart of BR-4 after a major interruption 3–4 cm after Br-4-3, dated at A.D. 540–980 (Fig. 11;Table 1). Furthermore, the sharp change (second interruption) from dense tufa to biogenic tufa in BR-5 and BR-6 may also have been contemporaneous with the damaging event. The age of this second damaging event can be bracketed between Br-4-3 (A.D. 540–980) and Br-3-4 (A.D. 770–940). Unless simply broken, the definite interruption of BR-3 (~10 cm after sample Br-3-4) marks the end of water overflow on the eastern aqueduct wall (and bridge) after the second damaging event. The growth of dense tufa in BR-4 and biogenic tufa in BR-5 and BR-6 in the final sections of cores indicates a continuous water flow on the western aqueduct wall after the second damaging event. The almost simultaneous arrest of tufa growth ~2 cm after Br-5-7 (A.D. 890–1020), ~1 cm after Br-6-8 (A.D. 900–1160), and ~7 cm after Br-4-3 (A.D. 540–980) suggests the occurrence of a major damaging event. Indeed, the arrest of tufa accumulation (in core samples Br-3-4, Br-5-7, and Br-6-8) probably occurred after A.D. 900–1160 (Br-6-8, Table 1) and indicates the final stoppage of water flow over the aqueduct.

260

Sbeinati et al.

TIMING OF EARTHQUAKE FAULTING AND CORRELATION AMONG ARCHAEOSEISMIC EXCAVATIONS, PALEOSEISMIC TRENCHES, AND CORES The analysis of field data in archaeoseismology, paleoseismology, and tufa coring provides some constraints on the successive past earthquakes along the Dead Sea fault at the Al Harif Roman aqueduct site (Figs. 12 and 13). The damage and repair of the aqueduct are here related to the total 13.6 m of left-lateral fault offset since construction of the aqueduct (Fig. 5). In addition, the tufa successive growth and interruptions visible in cores provide a direct relation between the water flow and the aqueduct function east and west of the fault zone. The correlation and timing coincidence between the faulting events visible in trenches, aqueduct construction damage and repair (see also Summary of Faulting Events from Archaeoseismology and Paleoseismology section), combined with tufa growth and interruptions, provide a better constraint on the timing of the successive large earthquakes: Event W, observed in trench C, occurred before 800–510 B.C. (unit f) and after 3400–300 B.C. (unit g). This faulting event can be determined only in trench C and hence cannot be correlated with damaging events in the aqueduct archaeoseismic excavations and tufa cores. However, we suggest two possible ages for this event: (1) according to the textual inscriptions found in different archaeological sites in Syria, a damaging earthquake sequence around 1365 B.C. affected Ugharit near Latakia in Syria, and Tyre further south in Lebanon and east of the Dead Sea fault (Sbeinati et al., 2005) may be correlated to event W; or (2) the rate of sedimentation in unit f of trench C implies a minimum age of 962 B.C. for event W. Event X, identified in trenches A and C between 350 B.C.– A.D. 30 and A.D. 532–641, postdates the construction of the aqueduct (younger than 65 B.C., i.e., the onset of Roman time in the Middle East and older than A.D. 70–230 of early tufa deposits). Event X also predates the onset of BR-3 tufa growth (see Br-3-1 dated A.D. 410–600). Similarly, the tufa growth interruption in BR-5 (after Br-5-2 dated 110 B.C.–A.D. 130) and onset of tufa in BR-6 (after Br-6-1 dated 400 B.C.–A.D. 250) coincide with the occurrence of the first damaging event X. The first earthquake faulting that damaged the aqueduct took place between A.D. 70–230 and A.D. 410–600. Event Y is younger than A.D. 650–810 (unit d in trench A) and older than A.D. 540–650 (unit d3 in trench C). This event postdates the first rebuilding phase of the aqueduct recognized from the fallen wall in excavation I and related cement sample AQ-CS-4 (A.D. 532–641) and tufa sample AQ-CS-3-2 (A.D. 560–690). Event Y predates the dragged wall fragment and related cement sample AQ-CS-1 (A.D. 650–780) and tufa sample AQ-CS-3-3 (A.D. 639–883; Table 1). Core samples of tufa deposits provide a bracket of the second damaging earthquake faulting between Br-4-3 (A.D. 540–980) and Br-3-4 (A.D. 770– 940). The second interruption in both BR-5 and BR-6 may also have been contemporaneous with the damaging event. Taking

into account only the archaeoseismic results, we can conclude that event Y likely occurred between A.D. 560–690 and A.D. 650–780; however, the consistency between all dates of paleoseismic, archaeoseismic, and tufa analysis suggest an earthquake event close to A.D. 650. Cement samples CS-1 and tufa sample CS-3-3 also indicate a rebuilding period after event Y, at the end of the Byzantine time and beginning of the Islamic period (fifth to sixth century A.D.). Event Z, observed in trenches A, B, and C, is identified as younger than A.D. 960–1060, and older than A.D. 1030–1260. The definite interruption of tufa growth in all cores and mainly BR-5 and BR-6 indicates the final stoppage of water flow over the bridge section. The interruption postdates sample Br-6-8 (A.D. 900–1160) and can be correlated with the 29 June 1170 large earthquake that affected the Missyaf region (Mouty and Sbeinati, 1988; Sbeinati et al., 2005). The Missyaf segment of the Dead Sea fault experienced four large earthquakes: event W in 3400–510 B.C., event X in A.D. 70–600, event Y in A.D. 560–780 (probably close to A.D. 650), and event Z in A.D. 960–1260 (probably in A.D. 1170). Using the Oxcal program (Bronk Ramsey, 2001), an attempt of sequential ordering of dates and events, presented in Figure 12, provides a time probability density function for events W (2300–500 B.C.), X (A.D. 160–510), Y (A.D. 625–690), and Z (A.D. 1010–1210). The timing of events obtained from the correlation and sequential distribution clearly indicate a temporal clustering of three large seismic events X, Y, and Z (Fig. 12) after event W, which may indicate a relatively long period of quiescence. Although our data and observations cannot precisely constrain event W, it may be correlated with the 1365 B.C. large earthquake that affected several sites between Lattakia and Tyre, as reported in the historical seismicity catalogue of Syria (Sbeinati et al., 2005). The Missyaf fault behavior is comparable to the temporal cluster of large seismic events that have occurred on other comparable major strikeslip faults (e.g., San Andreas fault—Weldon et al., 2004; Jordan Valley fault segment of the Dead Sea fault—Ferry et al., 2007). DISCUSSION AND CONCLUSION We conducted four archaeoseismic excavations, three paleoseismic trenches, and obtained the radiocarbon dating of six cores at the Al Harif Aqueduct site along the Missyaf segment of the Dead Sea fault. The combined study allows us to obtain a better constraint on the timing of past earthquakes, with four large seismic events during the last ~3400 yr. The occurrence of three seismic events X, Y, and Z (A.D. 70–600, ca. A.D. 650, and A.D. 1170, respectively) since the construction of the aqueduct is attested by faulting events in trenches, the damage and repair of the aqueduct wall, and the tufa growth and interruptions since Roman time (Fig. 13). These results point out a temporal clustering of three large earthquakes between A.D. 70 and A.D. 1170 along the Missyaf fault segment (Fig. 14). The 90 ± 10-km-long and linear Missyaf segment experienced the A.D. 1170 earthquake recorded in trenches, aqueduct

Timing of earthquake ruptures at the Al Harif Roman aqueduct (Dead Sea fault, Syria)

261

Figure 12. (A) Calibrated dating of samples (with calibration curve INTCAL04 from Reimer et al. [2004] with 2σ age range and 95.4% probability) and sequential distribution from Oxcal program (see also Table 1; Bronk Ramsey, 2001). The Bayesian distribution computes the time range of large earthquakes (events W, X, Y, and Z) at the Al Harif aqueduct according to faulting events, construction and repair of walls, and starts and interruptions of the tufa deposits (see text for explanation). Number in brackets (in %) indicates how much the sample is in sequence; the number in % indicates an agreement index of overlap with prior distribution.

construction, and tufa deposits. In this tectonic framework, the large (10-km-wide) Ghab pull-apart basin to the north and the Al Bouqueaa pull-apart and onset of the restraining bend to the south (Fig. 3) may constitute endpoints for earthquake rupture propagation, as observed for other large continental strike-slip faults (Klinger et al., 2003; Wesnousky, 2006). The size of the Ghab Basin and the sharp bend of the Lebanese fault system may act as structural control of fault-rupture initiation and propagation. Furthermore, the damage distribution of the A.D. 1170 earthquake, well located on the Missyaf segment, is limited to the north by the A.D. 1156 large earthquake and to the south by the A.D. 1063 and A.D. 1202 earthquakes (Fig. 2; Sbeinati et al.,

2005). The 20-km-thick seismogenic layer (Brew et al., 2001) correlates with the ~90 km fault length estimated from field mapping (Fig. 3). Fault dimensions are consistent with the ~4.3 m maximum characteristic slip inferred from the warping of the aqueduct wall east of the fault (and west of the bridge). Here, we assume that successive faulting episodes maintained the early ~4.3 m warping of an already ruptured strong building. Taking an average 2.0 m coseismic slip along the fault, the obtained seismic moment is Mo = 1.05 × 1020 N m (Mw 7.3; Wells and Coppersmith, 1994), which is comparable, for instance, with the seismic moment of the 1999 Izmit large earthquake (Mw 7.4) of the North Anatolian fault.

262

Sbeinati et al.

Date (AD)

Paleoseismic trenching

Archaeological excavations

Tufa analysis

Historical seismicity Civilizations Synthesis of EQ events (see Fig. 12)

Figure 13. Correlation of results among paleoseismic trenching, archaeoseismic excavations, and tufa analysis. In paleoseismic trenching, the youngest age for event X is not constrained, but it is, however, limited by event Y. In archaeoseismic excavations, the period of first damage overlaps with that of the second damage due to poor age control. In tufa analysis, the onset and restart of Br-3 and Br-4 mark the damage episodes to the aqueduct; the growth of Br-5 and Br-6 shows interruptions (I) indicating the occurrence of major events. Except for the 29 June 1170 event, previous events have been unknown in the historical seismicity catalogue. The synthesis of large earthquake events results from the timing correlation among the faulting events, building repair, and tufa interruptions (also summarized in Fig. 12 and text). Although visible in trenches (faulting event X), archaeoseismic excavations (first damage), and first interruption of tufa growth (in Br-5 and Br-6 cores), the A.D. 160–510 age of event X has a large bracket. In contrast, event Y is relatively well bracketed between A.D. 625 and 690, with the overlapped dating from trench results, the second damage of the aqueduct, and the interruption and restart of Br-3 and onset of Br-4. The occurrence of the A.D. 1170 earthquake correlates well with event Z from the trenches, the age of third damage to the aqueduct, and the age of interruption of Br-4, Br-5, and Br-6.

The Faulted Aqueduct: Earthquake Damage and Successive Offsets The consistency among the timing of faulted sedimentary units in trenches, the age of building and repair of the aqueduct wall, and the dating of tufa interruptions and restart episodes determines the completeness of a sequence of earthquake events. The dating of three episodes of fault slip X, Y, and Z is consistent with the two phases of aqueduct wall repair, and the two interruptions of the longest tufa deposits BR-5 and BR-6, and interruptions and restart in BR-3 and BR-4. Our observations indicate that the aqueduct was repaired after the large seismic events X and Y but abandoned after the most recent faulting event Z. Building repair after a damaging earthquake is very often necessary because it is a vital remedial measure of water supply in order to avoid a decline of the local economy (Ambraseys, 2006). The repair has the benefit of leaving critical indicators of previous damage and, in some cases, of the fault slip characteristics.

For instance, the eastern wall of the Al Harif aqueduct shows a clear warping that confirms the left-lateral movement near the fault zone. As observed for coseismic surface ruptures crossing buildings, fences, and walls during large strike-slip earthquakes (Yeats et al., 1997), warped walls that may record a coseismic slip are often observed along strike-slip faulting. Warping that amounts to 4.3 m can be interpreted as the individual coseismic slip during event X. The warping can be due to the opposite lateral movements across the fault constrained by the bridge cohesion to the east and wall solidity to the west. While the western aqueduct wall section was built straight on the flat alluvial terrace and ends abruptly against the fault, only the section between the bridge and the fault zone (which is partly built on loose sediments and bridge ballast) presents some warping and dragging (possibly separated from the alluvial substratum; Fig. 14). The warped section near the bridge displays one generation of cracks filled with tufa that attests to the early bridge damage and possible correlation with event X (Meghraoui et al., 2003). Similar warped walls

Timing of earthquake ruptures at the Al Harif Roman aqueduct (Dead Sea fault, Syria)

263

Figure 14. Schematic reconstruction (with final stage from Fig. 5) of the A.D. 160–510, A.D. 625–690, and A.D. 1170 large earthquakes and related faulting of the Al Harif aqueduct. Except for the A.D. 1170 earthquake (see historical catalogue of Sbeinati et al., 2005), the dating of earthquake events are from Figure 12. The white small section is the rebuilt wall after event X (see buried wall A and B in Fig. 8B); the subsequent gray piece corresponds to the rebuilt wall after event Y (see wall section C in Fig. 8B), which was damaged and dragged after event Z. The earlier aqueduct deformation (warping of the eastern wall near the fault rupture) may have recorded ~4.3 m of coseismic left-lateral slip that remained relatively well preserved during the subsequent fault movements.

and fences were observed after the 17 August 1999 earthquake and along the North Anatolia fault in Turkey (Barka et al., 2002). Subsequent faulting movements Y and Z would have affected an already broken aqueduct wall (even if rebuilt) with less strength at the fault zone than for the initial building conditions (Fig. 14). Furthermore, the 4.3 m can be considered as a characteristic slip at the aqueduct site; such characteristic behavior with repeated same amounts of coseismic slip has already been observed and inferred from paleoseismic trenches along major strike-slip faults (Klinger et al., 2003; Rockwell et al., 2009). If the warped aqueduct wall is random and not representative of a coseismic slip, the alternative solution is quite similar if we consider a 4.5 m average individual slip from the cumulative 13.6 m left-lateral offset and the X, Y, and Z large seismic events at the aqueduct site. Earthquake Records in Cores Another key issue is the relationship between the aqueduct damage and the start and interruption of tufa accumulation with past earthquakes (Figs. 11 and 13). Indeed, the water flow may be interrupted anytime due to, for instance, the actions of man (warfare) or the onset of a drought period and climatic fluctua-

tions that may influence the water flow. These possibilities seem here unlikely because the only two interruptions in cores BR-5 and BR-6 coincide with earthquake events X and Y, and no other additional interruptions were here recorded. This is also attested by the two interruptions in cores BR-3 and BR-4 that correlate with earthquake events X and Y. The difference between the tufa accumulation in BR-4, BR-5, and BR-6 located on the wall section west of the fault, and BR-3 located on the wall section next to the bridge, east of the fault, provides a consistent aqueduct damage history (Fig. 13). The onset of BR-3 after event X is the sign of an extensive damage that tilted the bridge and allowed overflow with tufa accumulation on the aqueduct northern side. The subsequent interruption (repair) and restart of BR-3 that coincides with event Y illustrate the successive aqueduct damage. Located on the broken western wall section (Fig. 6), the onset of BR-4 after event X and restart after event Y are consistent with BR-3 tufa growth and accumulation. As illustrated in Figure 13, the coincidence among faulting events X, Y, and Z from paleoseismic trenches, the three building damage and repair episodes from archaeoseismic investigations, and tufa growth and interruption constrains the earthquake-induced damage and faulting episodes across the aqueduct.

264

Sbeinati et al.

Missyaf Segment Seismic Gap and Fault-Rupture Behavior The Al Harif aqueduct, located at the mid-distance of the Missyaf fault segment, documents the size and rate of fault slip associated with large earthquakes. The numerous stream deflections observed along the fault segment imply cumulative leftlateral coseismic offsets consistent with the total aqueduct wall displacement. Stream deflections and wall offset result from the succession of large earthquakes and illustrate the long-term and short-term fault behavior, respectively. The previous 6.9 mm/yr slip rate obtained from the temporal cluster of large earthquakes X, Y, and Z (Meghraoui et al., 2003) clearly overestimates the long-term fault behavior because it was limited to the past 2000 yr time window. The occurrence of earthquake events W, X, Y, and Z in the past 3500 yr or so and related inferred 4.3 m characteristic left-lateral fault slip lead to a slip rate of 4.9–6.3 mm/yr (Fig. 15). Although the inferred age of event W from trench C is not well constrained, the correlation with the 1365 B.C. seismic sequence and related extension of damage from Latakia (in Syria) to Tyre (in Lebanon) reported in the histori-

cal catalogue (Sbeinati et al., 2005) suggests 1700–1300 yr of seismic quiescence in the sequence. The fault activity is here comparable to the Wallace-type behavior that describes the succession of temporal clusters of large earthquakes separated by periods of seismic quiescence (Shimazaki and Nakata, 1980). The mean recurrence time of large seismic events on the Missyaf fault segment can be estimated between 550 and 850 yr during the temporal cluster. This recurrence time increases to ~1077 if we take into account the maximum estimated age of event W obtained from the whole earthquake sequence of Figure 12. A comparable ~1100 yr mean return period is obtained from an ~6000 yr paleo-earthquake record on the juxtaposed southern Yammouneh fault segment (Daeron et al., 2007). With a limited number of radiocarbon ages and a possible overlap of the 1408 and 1872 earthquake ruptures, Akyuz et al. (2006) suggested a minimum 464–549 yr recurrence interval of surface faulting in the past 1000 yr on the northern end of the Dead Sea fault. However, the inferred large range in estimates of 500–1100 yr for recurrence intervals of earthquake faulting confirms the variability of earthquake occurrence and slip rates determined by the

Figure 15. Estimated fault-slip behavior and related slip rates (obtained from regression lines) from two scenarios of possible earthquake occurrence taking into account timing for paleo-earthquakes as in Figure 12 (with average X [A.D. 160–510] 375 ± 175, average Y [A.D. 625–690] 640 ± 32, Z [A.D. 1170], and two different time frames for W [historical event of 1365 B.C. and 962 B.C.). In both cases, the two regression lines indicate a minimum and maximum slip-rate estimate. In parallel, we assume an average 4.3 m characteristic individual slip consistent with the cumulative 13.6 m measured on the aqueduct (Fig. 5). If we assume a minimum age A.D. 962 for W (according to the dating in unit f, related rate of sedimentation, and the interface between unit f and unit g in trench C), the slip rate ranges between 6.1 mm/yr and 6.3 mm/yr (dark regression line with 80% correlation coefficient), implying that a large seismic event is overdue. If we consider the historical catalogue and the 1365 B.C. earthquake sequence along the Dead Sea fault for W (gray regression line with 78% correlation coefficient), the slip rate reduces to 4.9–5.5 mm/yr. The question mark indicates that for both scenarios, a large earthquake is overdue along the Missyaf fault segment (according to the seismic gap and the 4.0 m slip deficit). The temporal cluster of three large earthquakes in less than 1000 yr suggests a Wallace model of fault behavior with periods of seismic quiescence reaching ~1700 yr.

Timing of earthquake ruptures at the Al Harif Roman aqueduct (Dead Sea fault, Syria) relatively long periods of quiescence along the Dead Sea fault (Ferry et al., 2007). The instrumental seismicity in Figure 1A shows a seismic gap in Syria that also corresponds to more than 800 yr quiescence since the A.D. 1170 earthquake along the Missyaf segment. The seismic strain distribution is time predictable if we assume a constant characteristic slip at the aqueduct location. Taking into account a minimum 962 B.C. age of event W (Fig. 15), the 6.1– 6.3 mm/yr slip rate along the fault (Fig. 15) is comparable to other slip rates along the northern Dead Sea fault obtained from geology or GPS (McClusky et al., 2003; Altunel et al., 2009; Karabacak et al., 2010). If the 1365 B.C. large earthquake involved the Missyaf fault segment, it implies a 4.9–5.5 mm/yr slip rate, in agreement with other long-term slip rates of the southern Dead Sea fault (Klinger et al., 2000; Niemi et al., 2001; Daeron et al., 2004; Marco et al., 2005; Reilinger et al., 2006; Ferry et al., 2007; Gomez et al., 2007; Le Beon et al., 2008). The estimated 5.5 mm/yr slip rate and seismic quiescence since A.D. 1170 advocate ~4.0 m slip deficit and indicate that more seismic stress accumulation (which may correspond to one to two centuries to reach a 4.3 m characteristic slip) is needed for a rupture initiation. Our study shows that the integration of results from archaeoseismology, paleoseismology, tufa deposits, and historical seismicity is helpful to constrain the timing and characteristics of past earthquakes. However, the Dead Sea fault and related Wallace-type behavior require further paleoearthquake investigations that span several temporal clusters of seismic events. ACKNOWLEDGMENTS This research was funded by the European Commission–funded APAME Project (contract ICA3-CT-2002-10024) and by the UMR 7516 of Centre National de la Recherche Scientifique in Strasbourg. This research benefited from field support from the Syrian Atomic Energy Commission (SAEC), the Directorate General of Antiquities and Museums (DGAM), and the Higher Institute of Applied Sciences and Technology (HIAST) in Damascus. We are grateful to Ibrahim Osman (director general of the Atomic Energy Commission of Syria), Muawia Barazangi (Cornell University), Abdal Razzaq Moaz and Tammam Fakoush (DGAM), and Mikhail Mouty and Khaled Al-Maleh (Damascus University) for their constant support during the 5 yr study of the Al Harif archaeo-paleoseismology site. We are indebted to Tony Nemer, Ihsan Layous, Ryad Darawcheh, Youssef Radwan, and Abdul Nasser Darkal for field assistance and to Matthieu Ferry and Ersen Aksoy for critically reading an earlier version of this manuscript. We thank the two anonymous reviewers who significantly helped to improve our manuscript. Some figures were prepared using the public domain GMT software (Wessel and Smith, 1998). This article is a contribution to the United Nations Educational, Scientific and Cultural Organization–funded International Geoscience Programme IGCP 567, “Earthquake Archaeology: Archaeoseismology along the Alpine-Himalayan Seismic Zone.”

265

REFERENCES CITED Akyuz, S., Altunel, E., Karabacak, V., and Yalciner, C., 2006, Historical earthquake activity of the northern part of the Dead Sea fault zone, southern Turkey: Tectonophysics, v. 426, p. 281–293, doi: 10.1016/j.tecto .2006.08.005. Alchalbi, A., Daoud, M., Gomez, F., McClusky, S., Reilinger, R., Abu Romeyeh, M., Alsouod, A., Yassminh, R., Ballani, B., Darawcheh, R., Sbeinati, R., Radwan, Y., Al Masri, R., Bayerly, M., Al Ghazzi, R., and Barazangi, M., 2009, Crustal deformation in northwestern Arabia from GPS measurements in Syria: Slow slip rate along the northern Dead Sea fault: Geophysical Journal International, doi: 10.1111/j.1365-246X.2009.04431.x. Altunel, E., Meghraoui, M., Karabacak, V., Akyüz, S.H., Ferry, M., Yalçıner, C., and Munschy, M., 2009, Archeological sites (Tell and Road) offset by the Dead Sea fault in the Amik Basin, southern Turkey: Geophysical Journal International, v. 179, p. 1313–1329, doi: 10.1111/j.1365-246X .2009.04388.x. Ambraseys, N.N., 2006, Earthquakes and archeology: Journal of Archaeological Science, v. 33, p. 1008–1016, doi: 10.1016/j.jas.2005.11.006. Ambraseys, N.N., 2009, Earthquakes in the Mediterranean and Middle East: A Multidisciplinary Study of Seismicity up to 1900: Cambridge, UK, Cambridge University Press, 947 p. Ambraseys, N.N., and Jackson, J.A., 1998, Faulting associated with historical and recent earthquakes in the Eastern Mediterranean region: Geophysical Journal International, v. 133, p. 390–406, doi: 10.1046/j.1365-246X .1998.00508.x. Ambraseys, N., and Melville, C.P., 1988a, Historical evidence of faulting in Eastern Anatolia and northern Syria: Annali di Geofisica, v. 28, p. 337–343. Ambraseys, N., and Melville, C.P., 1988b, An analysis of the Eastern Mediterranean earthquake of 20 May 1202, in Lee, W.H.K., Meyers, H., and Shimazaki, K., eds., Historical Seismograms and Earthquakes of the World: San Diego, Academic Press, p. 181–200. Barazangi, M., Seber, D., Chaimov, T., Best, J., Litak, R., Al-Saad, D., and Sawaf, T., 1993, Tectonic evolution of the northern Arabian plate in western Syria, in Boschi, E., et al., eds., Recent Evolution and Seismicity of the Mediterranean Region: Dordercht, the Netherlands, Kluwer Academic, p. 117–140. Barka, A., Akyüz, H.S., Altunel, E., Sunal, G., Çakir, Z., Dikbas, A., Yerli, B., Armijo, R., Meyer, B., de Chabalier, J.B., Rockwell, T., Dolan, J.R., Hartleb, R., Dawson, T., Christofferson, S., Tucker, A., Fumal, T., Langridge, R., Stenner, H., Lettis, W., Bachhuber, J., and Page, W., 2002, The surface rupture and slip distribution of the 17 August 1999 Izmit earthquake (M 7.4), North Anatolian fault: Bulletin of the Seismological Society of America, v. 92, p. 43–60, doi: 10.1785/0120000841. Brew, G., Lupa, J., Barazangi, M., Sawaf, T., Al-Imam, A., and Zaza, T., 2001, Structure and tectonic development of the Ghab Basin and the Dead Sea fault system, Syria: Journal of the Geological Society of London, v. 158, p. 665–674, doi: 10.1144/jgs.158.4.665. Bronk Ramsey, C., 2001, Development of the radiocarbon program OxCal: Radiocarbon, v. 43, p. 355–363. Chaimov, T.A., Barazangi, M., Al-Saad, D., Sawaf, T., and Gebran, A., 1990, Crustal shortening in the Palmyride fold belt, Syria, and implications for movement along the Dead Sea fault system: Tectonics, v. 9, p. 1369–1386, doi: 10.1029/TC009i006p01369. Daeron, M., Benedetti, L., Tapponnier, P., Sursock, A., and Finkel, R., 2004, Constraints on the post–25-ka slip rate of the Yammouneh fault (Lebanon) using in situ cosmogenic 36Cl dating of offset limestone-clast fans: Earth and Planetary Science Letters, v. 227, p. 105–119, doi: 10.1016/ j.epsl.2004.07.014. Daeron, M., Klinger, Y., Tapponnier, P., Elias, A., Jacques, E., and Sursock, A., 2005, Sources of the large A.D. 1202 and 1759 Near East earthquakes: Geology, v. 33, p. 529–532, doi: 10.1130/G21352.1. Daeron, M., Klinger, Y., Tapponnier, P., Elias, A., Jacques, E., and Sursock, A., 2007, 12,000-year long record of 10 to 13 paleo-earthquakes on the Yammouneh fault (Levant fault system, Lebanon): Bulletin of the Seismological Society of America, v. 97, p. 749–771, doi: 10.1785/0120060106. Dubertret, L., 1955, Carte Géologique de Syrie avec Notice Explicative: Damascus, République Syrienne, Ministère des Travaux Publics, scale 1:200,000.

266

Sbeinati et al.

Elias, A., 2006, Le Chevauchement de Tripoli-Saïda: Croissance du Mont Liban et Risque Sismique [Ph.D. thesis]: Paris, Institut de Physique du Globe de Paris, 230 p. Ellenblum, R., Marco, S., Agnon, A., Rockwell, T., and Boas, A., 1998, Crusader castle torn apart by earthquake at dawn, 20 May 1202: Geology, v. 26, p. 303–306, doi: 10.1130/0091-7613(1998)026<0303:CCTABE >2.3.CO;2. Farr, T., and Kobrick, M., 2000, Shuttle Radar Topography Mission produces a wealth of data: Eos (Transactions, American Geophysical Union), v. 81, p. 583–585. Ferry, M., Meghraoui, M., Abou Karaki, N., Al-Taj, M., Amoush, H., Al-Dhaisat, S., and Barjous, M., 2007, A 48-kyr-long slip rate history for the Jordan Valley segment of the Dead Sea fault: Earth and Planetary Science Letters, v. 260, p. 394–406, doi: 10.1016/j.epsl.2007.05.049. Ford, T.D., and Pedley, H.M., 1996, A review of tufa and travertine deposits of the world: Earth-Science Reviews, v. 41, p. 117–175, doi: 10.1016/ S0012-8252(96)00030-X. Garfunkel, Z., Zak, I., and Freund, R., 1981, Active faulting in the Dead Sea rift: Tectonophysics, v. 80, p. 1–26, doi: 10.1016/0040-1951(81)90139-6. Ginat, H., Enzel, Y., and Avni, Y., 1998, Translocated Plio-Pleistocene drainage systems along the Arava fault of the Dead Sea transform: Tectonophysics, v. 284, p. 151–160, doi: 10.1016/S0040-1951(97)00165-0. Gomez, F., Meghraoui, M., Darkal, A., Hijazi, F., Mouty, M., Sulaiman, Y., Sbeinati, R., Darawcheh, R., Al-Ghazzi, R., and Barazangi, M., 2003, Holocene faulting and earthquake recurrence along the Serghaya branch of the Dead Sea fault system in Syria and Lebanon: Geophysical Journal International, v. 153, p. 658–674, doi: 10.1046/j.1365-246X .2003.01933.x. Gomez, F., Khawlie, M., Tabet, C., Darkal, A.N., Khair, K., and Barazangi, M., 2006, Late Cenozoic uplift along the northern Dead Sea transform in Lebanon and Syria: Earth and Planetary Science Letters, v. 241, p. 913–931, doi: 10.1016/j.epsl.2005.10.029. Gomez, F., Karam, G., Khawlie, M., McClusky, S., Vernant, P., Reilinger, R., Jaafar, R., Tabet, C., Khair, K., and Barazangi, M.,, 2007a, Global positioning system measurements of strain accumulation and slip transfer through the restraining bend along the Dead Sea fault system in Lebanon: Geophysical Journal International, v. 168, p. 1021–1028, doi: 10.1111/j.1365-246X.2006.03328.x. Gomez, F., Nemer, T., Tabet, C., Khawlie, M., Meghraoui, M., and Barazangi, M., 2007b, Strain partitioning of active transpression within the Lebanese restraining bend of the Dead Sea fault (Lebanon and SW Syria): Geological Society of London Special Publication 290, p. 285–303, doi: 10.1144/290.10. Grootes, P.M., Nadeau, M.J., and Rieck, A., 2004, 14C-AMS at the LeibnizLabor: Radiometric dating and isotope research: Nuclear Instruments and Methods in Physics Research, ser. B, v. 223–224, p. 55–61. Grootes, P.M., Nadeau, M.J., Roth, S., Andersen, N., Huels, M., Meghraoui, M., and Sbeinati, R., 2006, 1000 years of usage: The life story of a Roman Aqueduct provides tectonic information: Eos (Transactions, American Geophysical Union), v. 87, no. 52, Fall Meeting Suppl., abstract T13B0500. Grünthal, G., 1998, European Macroseismic Scale, 1998: Cahiers du Centre Européen de Géodynamique et de Seismologie: Luxembourg, Conseil de l’Europe, 230 p. Guidoboni, E., Comastri, A., and Traina, G., 1994, Catalogue of Ancient Earthquakes in the Mediterranean Area up to the 10th Century: Bologna, INGSGA (Istituto Nazionale di Geofisica–Storia Geofisica Ambiente), 504 p. Guidoboni, E., Bernardini, F., and Comastri, A., 2004a, The 1138–1139 and 1156–1159 destructive seismic crisis in Syria, south-eastern Turkey and northern Lebanon: Journal of Seismology, v. 8, p. 105–127, doi: 10.1023/B:JOSE.0000009502.58351.06. Guidoboni, E., Bernardini, F., Comastri, A., and Boschi, E., 2004b, The large earthquake on 29 June 1170 (Syria, Lebanon, and central southern Turkey): Journal of Geophysical Research, v. 109, p. B07304, doi: 10.1029/2003JB002523. Ibn Al-Athir, Ezz Ad-Din (AH 491–541, A.D. 1097–1146), 1982, Al-Kamil fi Al-Tarikh (The Complete History), v. 8, 9, 10, 11, 12: Beirut, Dar Sader. Karabacak, V., Altunel, E., Meghraoui, M., and Akyüz, S., 2010, Field evidences from northern Dead Sea fault zone (south Turkey): New findings for the initiation age and slip rate: Tectonophysics, v. 480, p. 172–182, doi: 10.1016/j.tecto.2009.10.001.

Klinger, Y., Avouac, J.P., Abou Karaki, N., Dorbath, L., Bourles, D., and Reyss, J.L., 2000, Slip rate on the Dead Sea transform in northern Araba Valley (Jordan): Geophysical Journal International, v. 142, p. 755–768, doi: 10.1046/j.1365-246x.2000.00165.x. Klinger, Y., Sieh, K., Altunel, E., Akoglu, A., Barka, A., Dawson, T., Gonzalez, T., Meltzner, A., and Rockwell, T., 2003, Paleoseismic evidence of characteristic slip on the western segment of the North Anatolian fault, Turkey: Bulletin of the Seismological Society of America, v. 93, p. 2317– 2332, doi: 10.1785/0120010270. Le Beon, M., Klinger, Y., Qader Amrat, A., Agnon, A., Dorbath, L., Baer, G., Ruegg, J.-C., Charade, O., and Mayyas, O., 2008, Slip rate and locking depth from GPS profiles across the southern Dead Sea transform: Journal of Geophysical Research, v. 113, p. B11403, doi: 10.1029/2007JB005280. Marco, S., 2008, Recognition of earthquake-related damage in archaeological sites—3 examples from the Dead Sea fault zone: Tectonophysics, v. 453, p. 148–156, doi: 10.1016/j.tecto.2007.04.011. Marco, S., Hartal, M., Hazan, N., Lev, L., and Stein, M., 2003, Archaeology, history, and geology of the A.D. 749 earthquake, Dead Sea Transform: Geology, v. 31, p. 665–668, doi: doi:10.1130/G19516.1. Marco, S., Rockwell, T.K., Heimann, A., Frieslander, U., and Agnon, A., 2005, Late Holocene activity of the Dead Sea transform revealed in 3D paleoseismic trenches on the Jordan Gorge segment: Earth and Planetary Science Letters, v. 234, p. 189–205, doi: 10.1016/j.epsl.2005.01.017. McClusky, S., Reilinger, R., Mahmoud, S., Ben, D., and Tealeb, A., 2003, GPS constraints on Africa (Nubia) and Arabia plate motions: Geophysical Journal International, v. 155, p. 126–138, doi: 10.1046/j.1365-246X .2003.02023.x. Meghraoui, M., Gomez, F., Sbeinati, R., Van der Woerd, J., Mouty, M., Darkal, A., Radwan, Y., Layyous, I., Najjar, H., Darawcheh, R., Hijazi, F., AlGhazzi, R., and Barazangi, M., 2003, Evidence for 830 years of seismic quiescence from paleoseismology, archeoseismology and historical seismicity along the Dead Sea fault in Syria: Earth and Planetary Science Letters, v. 210, p. 35–52, doi: 10.1016/S0012-821X(03)00144-4. Mouty, M., and Sbeinati, R., 1988, Historical Earthquake Catalogue for Syria and Adjacent Areas: Damascus, Atomic Energy Commission of Syria, Internal Report, 70 p. Nemer, T., and Meghraoui, M., 2006, Evidence of coseismic ruptures along the Roum fault (Lebanon): A possible source for the A.D. 1837 earthquake: Journal of Structural Geology, v. 28, p. 1483–1495, doi: 10.1016/ j.jsg.2006.03.038. Nemer, T., Meghraoui, M., and Khair, K., 2008, The Rachaya-Serghaya fault system (Lebanon): Evidence of coseismic ruptures, and the A.D. 1759 earthquake sequence: Journal of Geophysical Research, v. 113, B05312, doi: 10.1029/2007JB005090. Niemi, T., Zhang, H., Atallah, M., and Harrison, J.B.J., 2001, Late Pleistocene and Holocene slip rate of the northern Wadi Araba fault, Dead Sea transform, Jordan: Journal of Seismology, v. 5, p. 449–474, doi: 10.1023/A:1011487912054. Nur, A., and Cline, E.H., 2000, Poseidon’s Horses: Plate tectonics and earthquake storms in the late Bronze Age Aegean and Eastern Mediterranean: Journal of Archaeological Science, v. 27, p. 43–63, doi: 10.1006/ jasc.1999.0431. Quennell, A.M., 1984, The western Arabia rift system, in Dixon, J.E., and Robertson, A.H.F., eds., The Geological Evolution of the Eastern Mediterranean: Geological Society of London Special Publication 17, p. 775–788. Reilinger, R., McClusky, S., Vernant, P., Lawrence, S., Ergintav, S., Cakmak, R., Ozener, H., Kadirov, F., Guliev, I., Stepanyan, R., Nadariya, M., Hahubia, G., Mahmoud, S., Sakr, K., ArRajehi, A., Paradissis, D., Al-Aydrus, A., Prilepin, M., Guseva, T., Evren, E., Dmitrotsa, A., Filikov, S.V., Gomez, F., Al-Ghazzi, R., and Karam, G., 2006, GPS constraints on continental deformation in the Africa–Arabia–Eurasia continental collision zone and implications for dynamics of plate interactions: Journal of Geophysical Research, v. 111, doi: 10.1029/ 2005JB004051. Reimer, P.J., Baillie, M.G.L., Bard, E., Bayliss, A., Beck, J.W., Bertrand, C.J.H., Blackwell, P.G., Buck, C.E., Burr, G.E., Cutler, K.B., Damon, P.E., Edwards, R.L., Fairbanks, R.G., Friedrich, M., Guilderson, T.P., Hogg, A.G., Hughen, K.A., Kromer, B., McCormac, G., Manning, S., Bronk Ramsey, C., Reimer, R.W., Remmele, S., Southon, J.R., Stuiver, M., Talamo, S., Taylor, F.W., van der Plicht, J., and Weyhenmeyer, C.E., 2004, IntCal04 Terrestrial radiocarbon age calibration, 26–0 ka BP: Radiocarbon, v. 46, p. 1029–1058.

Timing of earthquake ruptures at the Al Harif Roman aqueduct (Dead Sea fault, Syria) Rockwell, T., Ragona, D., Seitz, G., Langridge, R., Aksoy, M.E., Ucarkus, G., Ferry, M., Meltzner, A.J., Klinger, Y., Meghraoui, M., Satir, D., Barka, A., and Akbalik, B., 2009, Paleoseismology of the North Anatolian fault near the Marmara Sea: Implications for fault segmentation and seismic hazard, in Reicherter, K., Michetti, A.M., and Silva, P.G., eds., Palaeoseismology: Historical and Prehistorical Records of Earthquake Ground Effects for Seismic Hazard Assessment: Geological Society of London Special Publication 316, p. 31–54, doi: 10.1144/SP316.3. Salamon, A., Hofstetter, A., Garfunkel, Z., and Ron, H., 2003, Seismotectonics of the Sinai subplate—The Eastern Mediterranean region: Geophysical Journal International, v. 155, p. 149–173, doi: 10.1046/j.1365 -246X.2003.02017.x. Sbeinati, M.R., Darawcheh, R., and Mouty, M., 2005, The historical earthquakes of Syria: An analysis of large and moderate earthquakes from 1365 B.C. to 1900 A.D.: Annals of Geophysics, v. 48, p. 347–435. Shimazaki, K., and Nakata, T., 1980, Time-predictable recurrence model for large earthquakes: Geophysical Research Letters, v. 7, p. 279–282, doi: 10.1029/GL007i004p00279. Stiros, S., and Jones, R.E., eds., 1996, Archaeoseismology: British School of Athens, Greece, Fitch Laboratory Occasional Paper 7, 268 p. Trifonov, V.G., Tribukhin, V.M., Adzhamyan, Z., Dshallad, S., El-Khair, Y., and Ayed, K., 1991, Levant fault zone in northwest Syria: Geotectonics, v. 25, p. 145–154. Wdowinski, S., Bock, Y., Baer, G., Prawirodirdjo, L., Bechor, N., Naaman, S., Knafo, R., Forrai, Y., and Melzer, Y., 2004, GPS measurements of current crustal movements along the Dead Sea fault: Journal of Geophysical Research, v. 109, B05403, doi: 10.1029/2003JB002640.

267

Weldon, R., Scharer, K., Fumal, T., and Biasi, G., 2004, Wrightwood and the earthquake cycle: What a long recurrence record tells us about how faults work: GSA Today, v. 14, no. 9, p. 4–10, doi: 10.1130/ 1052-5173(2004)014<4:WATECW>2.0.CO;2. Wells, D.L., and Coppersmith, K.J., 1994, New empirical relationships among magnitude, rupture length, rupture width, rupture area, and surface displacement: Bulletin of the Seismological Society of America, v. 84, p. 974–1002. Wesnousky, S.G., 2006, Predicting the endpoints of earthquake ruptures: Nature, v. 444, p. 358–360, doi: 10.1038/nature05275. Wessel, P., and Smith, H.F., 1998, New, improved version of the Generic Mapping Tools released: Eos (Transactions, American Geophysical Union), v. 79, p. 579. Westaway, R., 2004, Kinematic consistency between the Dead Sea fault zone and the Neogene and Quaternary left-lateral faulting in SE Turkey: Tectonophysics, v. 391, p. 203–237, doi: 10.1016/j.tecto.2004.07.014. Yeats, R., Sieh, K., and Allen, C., 1997, The Geology of Earthquakes: Oxford, UK, Oxford University Press, 568 p. Zilberman, E., Amit, R., Porat, N., Enzel, Y., and Avner, U., 2005, Surface ruptures induced by the devastating 1068 A.D. earthquake in the southern Arava valley, Dead Sea Rift, Israel: Tectonophysics, v. 408, p. 79–99, doi: 10.1016/j.tecto.2005.05.030.

MANUSCRIPT ACCEPTED BY THE SOCIETY 19 MAY 2010

Printed in the USA

The Geological Society of America Special Paper 471 2010

Offset archaeological relics in the western part of the Büyük Menderes graben (western Turkey) and their tectonic implications Önder Yönlü* Erhan Altunel Volkan Karabacak Department of Geology, Eskişehir Osmangazi University, 26480 Eskişehir, Turkey Serdar Akyüz Department of Geology, İstanbul Technical University, 34469 İstanbul, Turkey Çağlar Yalçıner Çan Vocational College, Çanakkale Onsekiz Mart University, 17400 Çan, Çanakkale, Turkey

ABSTRACT The Büyük Menderes graben is one of the most important active tectonic structures of western Anatolia. The graben extends for a distance of ~150 km between the Denizli Basin in the east and the Aegean Sea in the west, where its trend changes to NE-SW. The main active faults are located along the northern margin of the graben, some of which have been reactivated in surface-rupturing earthquakes during the twentieth century and the historical period. Detailed investigations along the NE-SW– trending part of the Büyük Menderes graben showed that archaeological relics have been faulted by surface ruptures during the large historical earthquakes. The ancient city of Priene and an Ottoman bridge are located along the northwestern margin of the graben to the southwest of Söke and in Sazlıköy, respectively. Field observations and light detection and ranging (LIDAR) studies at both sites show that faulting has a normal component with considerable right-lateral movement. Offset archaeological features at both Priene and the Ottoman bridge are evidence for the reactivation of the graben boundary faults in the past 2000 yr. At Priene, a N-S–trending street wall is offset by 21 cm vertically and 10 cm dextrally, the eastern wall of the gymnasium is offset by 8 cm vertically, and the floor blocks of the agora are displaced by 26 cm vertically and 13 cm dextrally. The Ottoman bridge displays 76 cm vertical and 43 cm dextral offset to the southeast, which probably occurred during the 1846 earthquake.

*[email protected] Yönlü, Ö., Altunel, E., Karabacak, V., Akyüz, S., and Yalçıner, Ç., 2010, Offset archaeological relics in the western part of the Büyük Menderes graben (western Turkey) and their tectonic implications, in Sintubin, M., Stewart, I.S., Niemi, T.M., and Altunel, E., eds., Ancient Earthquakes: Geological Society of America Special Paper 471, p. 269–279, doi: 10.1130/2010.2471(21). For permission to copy, contact [email protected]. ©2010 The Geological Society of America. All rights reserved.

269

270

Yönlü et al.

INTRODUCTION Western Turkey is one of the most tectonically active regions of the world; it is currently experiencing roughly NNE-SSW stretching at 30 mm yr–1 (Le Pichon et al., 1995; Oral et al., 1995; Barka and Reilinger, 1997; McClusky et al., 2003; Reilinger et al., 2006). This active extension results in a horst-and-graben system that characterizes most of western Turkey. The E-W–trending Büyük Menderes graben is the most significant structure of this system, extending for a length of ~150 km between the Denizli Basin in the east and the Aegean Sea in the west (Fig. 1A). The graben trends approximately E-W inland, but its strike changes to NE-SW between Germencik and the Aegean Sea (Figs. 1A and 1B). For simplicity, we here call this NE-SW–trending part the Söke graben. The main active normal faults are located along the northern margin of the Büyük Menderes graben (Şengör, 1982; Paton, 1992; Seyitoğlu and Scott, 1992; Şaroğlu et al., 1992; Cohen

et al., 1995; Altunel, 1999; Hakyemez et al., 1999; Koçyiğit et al., 1999; Bozkurt, 2000; Sözbilir, 2000; Yönlü, 2008); these boundary faults have been reactivated in historical and twentiethcentury earthquakes (Ergin et al., 1967; Soysal et al., 1981; Ambraseys, 1988; Ambraseys and Finkel, 2006; Tan et al., 2008). In the Söke graben, historical earthquakes include the A.D. 68 and 1846 (I°: VIII) earthquakes (Ergin et al., 1967; Ambraseys, 1988; Tan et al., 2008), while the most recent seismic event occurred in 1955 (Öcal, 1958; Ergin et al., 1967; Ambraseys, 1988; Kalafat et al., 2007). That earthquake of 16 July 1955 (M 6.8) caused great damage in the town of Balat (ancient Miletus) and the surrounding region (Fig. 1B). McKenzie (1972) provided a faultplane solution for this event that indicated normal movement downthrown to the southeast combined with subsidiary rightlateral motion (Fig. 1B). The Büyük Menderes graben is rich in well-exposed archaeological relics, which preserve damage of historical earthquakes. Here, we focus on two specific sites: the ancient city of Priene

Figure 1 (on this and following page). (A) Digital elevation map and main tectonic features of western Anatolia.

Offset archaeological relics in the western part of the Büyük Menderes graben, western Turkey

271

Figure 1 (continued). (B) Active faults overlaid on digital elevation model in the study area. Pictures show locations of ancient structures; circle is the focal-plane solution of the 16 July 1955 Söke-Balat earthquake (Mw 6.8), provided by McKenzie (1972).

and an Ottoman bridge near Sazlıköy, both located along the northern margin of the graben. Observations of damage at both sites suggest that they were affected by large historical earthquakes, and here we examine this likelihood in more detail using new field measurement techniques, notably LIDAR (light detection and ranging). As well as presenting geological and geomorphological field observations for active faulting along the Söke graben, this paper demonstrates offset archaeological features that are convincing evidence for the reactivation of the fault in the past 2000 yr. We argue that such offset archaeological relics constitute reliable data for determining the characteristics of ancient seismic ruptures. FIELD OBSERVATIONS Geological and Geomorphological Observations The Söke graben is framed both to the northwest and southeast by high topographic escarpments (Fig. 1B). As noted already, the main faults are located on the northwestern margin of the graben, with its southeastern flank bounded by less prominent antithetic faults. Field mapping investigated both sides of the graben, but the observations presented here concentrate along the northwestern boundary because the southeastern margin of the graben shows no evidence of recent surface faulting.

The Söke graben extends for a distance of at least 30 km between Germencik in the east and the Aegean Sea in the west (Fig. 2A). The late Quaternary tectonic activity of this part of graben is characterized by prominent and well-developed fault surfaces, faulted alluvial deposits, triangular-facetted rangefront spurs, linear scarps, and sharply truncated ridges. Near the northeastern end (around Germencik), the fault trends ENEWSW and separates Quaternary sediments from Neogene units, and the fault is characterized by fault planes and ~25-m-high morphological escarpments (Fig. 2B). At Argavlı, there is a 3-km-wide step over, after which the fault continues southwest on the same trend to Sazlıköy (Fig. 2B). Between Argavlı and Sazlıköy, there is another step over within a heavily deformed zone in limestone basement. Here, the oblique slickensides on fault planes show that the dip-slip motion has a right-lateral component. The fault is marked by uplifted Quaternary sediments between Sazlıköy and Söke. West of Söke and around Atburgazı, the fault trace steps over to the south and defines the limestone-Quaternary border between Söke and the Aegean coast (Fig. 2A). The trend of the fault switches between NNESSW and ENE-WSW, but its general trend is NE-SW. Again, this westernmost section shows a distinct “seismic landscape” of triangular facets (Fig. 2C), linear topographic escarpments, and striated fault planes that are indicative of active faulting (Fig. 2D).

Figure 2. (A) Detailed active fault map of the Söke graben. (B) Truncated ridges around Morallı village (view toward northwest). (C) Triangular facet in west of Priene (view toward north). (D) About 5-m-high fault plane in the limestone near the Aegean coast (view toward northwest). (See A for locations of B, C, and D.)

Offset archaeological relics in the western part of the Büyük Menderes graben, western Turkey For millennia, this faulted corridor of land has been an important route for people and armies traveling either from west or east between the Aegean coast and the Anatolian interior, and so the Büyük Menderes graben is rich in archaeological sites. Considering that the graben-bounding fault segments are very likely to have ruptured during the numerous major earthquakes reported here in historical times, it is possible that some surface ruptures affected the ancient features that existed at the time. Here, we focus attention on two convincing sites: the ancient city of Priene and an Ottoman bridge, northeast of Priene (Figs. 1B and 2A). Archaeoseismic Damage at Priene The ancient city of Priene is located on the northwestern side of the Söke graben (Fig. 1B) on a limestone terrace ~50 m above the valley floor (Fig. 3). The southeastern edge of the terrace is bounded by a weathered fault scarp, which represents the active fault separating graben-fill deposits from the basement limestone. Earthquake damages at Priene were first documented by Altunel (1998), who recognized displaced blocks and offset walls along a NE-SW–trending corridor in the city center. Although Altunel (1998) observed normal fault throw combined with rightlateral slip, there was limited information on the magnitude of offsets. Light detection and ranging (LIDAR) measurement techniques, however, now allow us to determine the amount of motion more precisely. The LIDAR instrument used here is an Optech ILRIS-3D, which has an integrated digital camera and sampling rate of 2500 points per second in the range of ~3–1500 m. It is based on the traveltime of a laser beam between the source (scanner) and the target. Multiple laser beams sweeping the laser range allow creation of a three-dimensional (3-D) point cloud of the target with less than 1 cm resolution. These point clouds are then commonly converted into triangulated surfaces and are used for the draping of high-resolution digital photographs to create virtual-reality models of the target. Earthquake traces at Priene affect a number of faulted relics in the city center. First, the eastern wall of a N-S–trending street near the eastern entrance of the city (1 in Fig. 4) is displaced by a north-dipping fault (Fig. 5A). Altunel (1998) interpreted the damage as tilting, but LIDAR measurements reveal that there is a vertical displacement combined with dextral component; the northern side of the wall is downthrown by 21 cm (Fig. 5B) and is dextrally displaced by 10 cm (Fig. 5C). Second, a few meters north of this location (2 in Fig. 4), there is a fault cutting the eastern wall of the gymnasium; the northern side is downthrown by up to 8 cm, and the lower part of the wall is dilated (Fig. 6). Third, toward the southwest, the floor blocks of the stoa and agora are displaced (3 in Figs. 4 and 7A), and LIDAR measurements record a vertical displacement of up to 26 cm (Fig. 7B) combined with 13 cm of dextral offset (Fig. 7C). Damage at Ottoman Bridge Southeast of Sazlıköy (Fig. 8A), the fault exhibits clear morphological and geological field evidence (Fig. 8B), separating Quaternary graben deposits from the basement limestone, but

273

toward the northeast, around Sazlıköy, it passes into the Quaternary deposits (Fig. 8A) and is difficult to trace. However, assuming that the fault extends further northeast along the same trend, an Ottoman bridge (Fig. 9) is located on the fault trace (Fig. 8). The bridge was built by Ramazan Pasha in 1595 (Gökben and Sülen, 1936; Arseven, 1955), and so the structure offers the opportunity of examining fault offsets over the intervening centuries. The structure has the following characteristics: First, it is 100 m long and in plan view is not straight, but has a smooth arcuate shape (Fig. 10A). In particular, there is an abrupt bend near the southeastern entrance of the bridge (Fig. 11) that is clearly visible in the LIDAR image (Fig. 10A). Second, the bridge consists of six arches with intervening windows (Figs. 9 and 10B), and it is noticeable that there is a central window (w0) midway along the structure. In the symmetry of the design, each arch and window has a mirror image on the other side (Fig. 10B). Drawing a line along the tops of the arches (a1, a2, a3) and windows (central window, w1, w2) for both sides of the bridge, there is a consistency between a1, a2, and w1 on both sides (Fig. 10B). However, a3 is not consistent, instead being 76 cm lower on the southeastern side than the equivalent a3 on the northwestern side (Fig. 10B). In addition, although there is a window (w2) between a2 and a3 on the northwestern side, its equivalent (w2 in the southeastern side) is absent (Fig. 10B). This disruption of the symmetry coincides with the location of a sudden bend in the bridge’s plan view (Fig. 10A). On the basis of these observations, it is clear that there has been a change in the architecture of the bridge since its construction in 1595. From LIDAR analysis, the SE end of the bridge is bent by 43 cm relative to the NW end, and it has been shifted southward, consistent with right-lateral displacement on the inferred fault trace. Moreover, the fact that the southeastern a3 arch is 76 cm lower than its original position is interpreted as indicating that the underlying fault is downthrown to the southeast. On the basis of this assumption, that the eastern part of the bridge is downthrown as a result of surface faulting, the sense of motion (normal with a considerable dextral component) is consistent with geological evidence and other offset features in Priene. An important question is: If the bridge was damaged as a result of an earthquake, why was it not reconstructed? We might expect that, after the damage, the eastern side would have been removed completely and rebuilt in its original position. In this case, we would expect that a3 on the southeastern side would be at the same level with its equivalent on the northwestern side, and there would be a window (w2). However, the absence of w2 and the lower position of a3 on the southeastern side indicate that this was not the case. Furthermore, reconstruction would have meant there would not be a bend in the plan view. Instead, it seems likely that after the damage, the southeastern part was repaired as it was. In this case, it would not have been possible to raise the arch a3 back to its original (pre-earthquake) position, nor to repair the broken window (w2 in the eastern side) and to recover the dextral offset in the structure. In this scenario, the vertical and dextral structural damage remained in the bridge.

Figure 3. Location of the ancient city of Priene. Ancient city is sited on the northwestern margin of the Söke graben, in the footwall of a degraded fault scarp (view toward northeast) (dashed line indicates the location of the active fault).

Figure 4. A general city map of Priene modified from Tulay (1993). Numbers indicate locations of observed damages defined in the text. Thin line is the antithetic fault; solid line is the main active fault bounding the graben on the northwestern margin.

Figure 5. (A) Tilted and offset blocks of a N-S–trending street wall (view toward southeast). (B–C) Light detection and ranging (LIDAR) views of the street wall. Profile view of the wall (B) shows 20 cm vertical offset; plan view (C) shows 10 cm dextral offset. Each label indicates the same block.

Figure 6. A fault cutting the N-S–trending eastern wall of the gymnasium (view toward west) (number 2 in Fig. 4). Note that northern side is downthrown by ~8 cm (dashed line). Arrows show dilatations.

Figure 7. (A) Light detection and ranging (LIDAR) view of the offset floor blocks of the agora (number 3 in Fig. 4). Dashed line indicates the location of the fault. (B) North-south cross section on the LIDAR view shows 20 cm vertical offset. (C) Plan view of the offset floor of the agora shows 13 cm dextral offset (white dashed line). Black dashed line indicates the location of fault, and each label indicates the same block.

Figure 8. (A) Location of the fault (dashed line) and Ottoman bridge near Sazlıköy. (B) A southeast-facing fault plane near Sazlıköy.

Figure 9. The sixteenth-century Ottoman bridge located on the abandoned bed of a meander river. Question mark indicates the absent window w2 in the eastern part of the bridge (view toward northwest). Dashed line is the symmetry line of the bridge. W0 is the central window. Note that there are three arches on each side of the symmetry line, while there are two windows on the northwestern side of the symmetry line, and there is only one window on the southeastern side—w2 is missing.

Figure 10. (A) The plan view of the Ottoman bridge obtained by light detection and ranging (LIDAR). Note that the bridge has a smooth arch shape. A line, which is drawn along the side of the bridge, shows that it does not follow the original position (dashed line) near the eastern part of the bridge. (B) Profile view of the Ottoman bridge obtained by LIDAR. The line drawn along the top points of the arches and windows shows that a3 is 76 cm lower than its original position. Question mark indicates the missing window of the bridge.

278

Yönlü et al.

Figure 11. A view of the Ottoman bridge from southeast to northwest shows bending on the bridge.

A key question for earthquake studies is when did the bridge sustain its damage? The last large earthquake was the 1955 (M 6.8) event, which, according to Öcal (1958) and McKenzie (1972), took place near the Aegean coast (Fig. 12). The damage on the Ottoman bridge is unlikely to be related to the 1955 earthquake because there is no report of either surface faulting or reparation on the Ottoman bridge. Although there is no detailed account of the 1846 earthquake (I°: VIII), the damage to the Ottoman bridge certainly occurred after its construction in 1595. Thus, it is possible that the 1846 earthquake occurred between Söke and Germencik (Fig. 2A) and that surface faulting of 76 cm vertical and 43 cm dextral offset on the bridge were associated with this event. CONCLUSIONS The ancient city of Priene on the northern margin of the Büyük Menderes graben was affected by destructive earthquakes in the historical period. Although it is difficult to correlate observed damages to a specific past seismic event, detailed LIDAR measurements presented here provide more reliable information about the sense of motion on the fault zone. In the city itself, an offset street wall (1 in Fig. 4), an offset wall of the gymnasium (2 in Fig. 4), and offset blocks of the agora (3 in Fig. 4) have their northern side downthrown up to 26 cm along a NE-SW–trending fault, which is subparallel to the main boundary fault. The LIDAR results also confirm that the vertical slip is

Figure 12. Isoseismal map of the 16 July 1955 Söke-Balat earthquake (from Öcal, 1958). Focal-plane solution was provided by McKenzie (1972).

combined with a considerable right-lateral component. Considering that the northern side is downthrown along the NE-SW– trending fault, the fault in the city center of Priene is dipping toward the northwest. However, the main boundary fault of the graben dips to the southeast, so the fault cutting through the heart of Priene is a secondary antithetic fault, as interpreted by Altunel (1998). The sense of motion on offset archaeological relics in the city center of Priene is consistent with the sense of slip from striations on fault planes (Fig. 2D) and the fault-plane solution of the Söke-Balat earthquake (Fig. 1B). Northeast of Priene, near Sazlıköy, an Ottoman bridge located on the fault line exhibits a similar sense of displacement (Figs. 2A and 8A). Architectural observations and LIDAR measurements show that the bridge is disrupted near its southeastern entrance. A small (43 cm) but distinct bend in the plan view of the bridge (Figs. 10A and 11), a missing window (w2) on the eastern side, and an arch (a3) on the southeastern side that is 76 cm lower than its equivalent on the northwestern side (Fig. 10B) are consistent with the structure being interrupted by normal fault slip with a strong right-lateral component. We speculate that this displacement and damage occurred during the 1846 earthquake, which probably occurred between Söke and Germencik.

Offset archaeological relics in the western part of the Büyük Menderes graben, western Turkey ACKNOWLEDGMENTS This paper was a part of Önder Yönlü’s M.Sc. thesis, completed at Eskişehir Osmangazi University under the supervision of Erhan Altunel. This research is supported by TUBITAK (105Y348) and Eskişehir Osmangazi University Research Foundation (200615026 and 200615036). The authors are grateful to Musa Kadıoğlu (Department of Archaeology of Ankara University), who participated in useful discussions about the Ottoman bridge. We would like to thank Aurelia Hubert Ferrari and Iain Stewart for the constructive review of the manuscript. This article is a contribution to the United Nations Educational, Scientific, and Cultural Organization–funded International Geoscience Programme IGCP 567, “Earthquake Archaeology: Archaeoseismology along the Alpine-Himalayan Seismic Zone.” REFERENCES CITED Altunel, E., 1998, Evidence for damaging historical earthquakes at Priene, western Turkey: Turkish Journal of Earth Science, v. 7, p. 25–35. Altunel, E., 1999, Geological and geomorphological observations in relation to the 20 September 1899 Menderes earthquake, western Turkey: Journal of the Geological Society of London, v. 156, p. 241–246, doi: 10.1144/ gsjgs.156.2.0241. Ambraseys, N.N., 1988, Engineering Seismology: Earthquake Engineering, Structure and Dynamics, Volume 17, 105 p. Ambraseys, N.N., and Finkel, C.F., 2006, Türkiye’de ve komşu bölgelerde sismik etkinlikler (Seismicity of Turkey and neighbouring regions); bir tarihsel inceleme 1500–1800, Volume 4: Ankara, TUBITAK Yayınları Akademik dizi, 251 p. Arseven, C., 1955, Türk sanat tarihi (Turkish art history): İstanbul, 400 p. Barka, A., and Reilinger, R., 1997, Active tectonics of the Eastern Mediterranean region: Deduced from GPS, neotectonic and seismicity data: Annali di Geofisica, v. XL, no. 3, p. 587–610. Bozkurt, E., 2000, Timing of extension on the Büyük Menderes graben, western Turkey, and its tectonic implications, in Bozkur, E., Winchester, J.A., and Piper, J.D.A., eds., Tectonics and Magmatism in Turkey and the Surrounding Area: Geological Society of London Special Publication 173, p. 385–403. Cohen, H.A., Dart, C.J., Akyüz, H.S., and Barka, A., 1995, Syn-rift sedimentation and structural development of the Gediz and Büyük Menderes graben, western Turkey: Journal of the Earth Society, v. 152, p. 629–638. Ergin, K., Güçlü, U., and Uz, Z., 1967, A Catalog of Earthquakes for Turkey and Surrounding Area (11 A.D. to 1964 A.D.): İstanbul, Turkey, İTÜ Faculty of Mining Engineering, 321 p. Gökben, H., and Sülen, S., 1936, Aydın ili tarihi (History of Aydın): İstanbul, 110 p. Hakyemez, Y.H., Erkal, T., and Göktas, F., 1999, Late Quaternary evolution of the Gediz and Büyük Menderes grabens, western Anatolia, Turkey: Quaternary Science Reviews, v. 18, p. 549–554, doi: 10.1016/S0277 -3791(98)00096-1. Kalafat, D., Güneş, Y., Kara, M., Deniz, P., Kekovalı, K., Kulel, H.S., Gülen, L., Yılmazer, M., and Özel, N.M., 2007, A Revised and Extended Earthquake Catalogue for Turkey since 1900 (M:4.0): Bebek-İstanbul, Boğaziçi Uni-

279

versity, Kandilli Observatory and Earthquake Research Institute, Special Publication, 143 p. Koçyiğit, A., Yusufoglu, H., and Bozkurt, E., 1999, Evidence from the Gediz graben for episodic two-stage extension in western Turkey: Journal of the Geological Society of London, v. 156, p. 605–616. Le Pichon, X., Chamot-Rooke, C., Lallemant, S., Noomen, R., and Veis, G., 1995, Geodetic determination of the kinematics of Central Greece with respect to Europe: Implications for Eastern Mediterranean tectonics: Journal of Geophysical Research, v. 100, p. 12,675–12,690. McClusky, S., Reilinger, R., Mahmoud, S., Ben Sari, D., and Tealeb, A., 2003, GPS constraints on Africa (Nubia) and Arabia plate motions: Geophysical Journal International, v. 155, p. 126–138, doi: 10.1046/j.1365 -246X.2003.02023.x. McKenzie, D., 1972, Active tectonics of the Alpine-Himalayan belt: The Aegean Sea and surrounding regions: Geophysical Journal of the Royal Astronomical Society, v. 55, no. 1, p. 217–254. Öcal, N., 1958, Söke-Balat zelzelesi (Söke-Balat earthquake): İstanbul Kandili Rasathanesi Sismoloji Yayınları, v. 2 p. 34. Oral, M.B., Reilinger, R.E., Toksöz, M.N., Kong, R.W., Barka, A.A., Kınık, I., and Lenk, O., 1995, Global positioning system offers evidence of plate motions in Eastern Mediterranean: Eos (Transactions, American Geophysical Union), v. 76, no. 9, p. 9–11. Paton, S., 1992, Active normal faulting, drainage patterns and sedimentation in southwestern Turkey: Journal of the Earth Society, v. 149, p. 1031–1044. Reilinger, R., McClusky, S., Vernant, P., Lawrence, S., Ergintav, S., Cakmak, R., Ozener, H., Kadirov, F., Guliev, İ., Stepanyan, R., Nadariya, M., Hahubia, G., Mahmoud, S., Sakr, K., ArRajehi, A., Paradissis, D., Al-Aydrus, A., Prilepin, M., Guseva, T., Evren, E., Dmitrotsa, A., Filikov, S.V., Gomez, F., Al-Ghazzi, R., and Karam, G., 2006, GPS constraints on continental deformation in the Africa-Arabia-Eurasia continental collision zone and implications for the dynamics of plate interactions: Journal of Geophysical Research, v. 111, B05411, doi: 10.1029/2005JB004051. Şaroğlu, F., Emre, Ö., and Kuşçu, İ, 1992, 1:1,000,000 Türkiye diri fay haritasi (Active Fault Map of Turkey in 1:1,000,000 scale): Ankara, Turkey, Maden Tetkik ve Arama (Mineral Research & Exploration General Directorate), scale 1:1,000,000. Şengör, A.M.C., 1982, Ege’nin neotektonik evrimini yöneten etkenler (Factors governing the neotectonic evolution of the Aegean), in Erol, O., and Oygür, V., eds., Bati Anadolu’nun Genç Tektonigi ve Volkanizmasi Paneli: Congress of the Geological Society of Turkey 34, p. 59–71. Seyitoğlu, G., and Scott, B.C., 1992, The age of Büyük Menderes graben (west Turkey) and its tectonic implications: Geological Magazine, v. 129, p. 239–242, doi: 10.1017/S001675680000830X. Soysal, H., Sipahioglu, S., Kolçak, D., and Altınok, Y., 1981, Türkiye ve Çevresinin Tarihsel Deprem Katalogu (M.Ö. 2100–M.S. 1900) (Historical Earthquake Catalog of Turkey and Its Environment, 2100 B.C. to 1900 A.D.): İstanbul, TÜBİTAK Publications, 87 p. Sözbilir, H., 2000, Batı Anadolu’dan örneklerle aktif faylar ve potansiyel aktif faylar: Batı Anadolu’nun Depremselligi Sempozyumu (Symposium on seismicity of the Western Anatolia), p. 133–143. Tan, O., Tapırdamaz, M.C., and Yörük, A., 2008, Earthquake catalogues for Turkey: Turkish Journal of Earth Sciences, v. 17, p. 405–418. Tulay, A.S., 1993, Anadolu’nun Pompeisi Priene (Pompei of Anatolia; Priene): İstanbul, Turing magazine, p. 81–360. Yönlü, Ö., 2008, Morphotectonic Characteristics and Historical Earthquake Activity of the Western Part of the Büyük Menderes Graben [M.Sc. thesis]: Eskişehir, Eskişehir Osmangazi University, 107 p. MANUSCRIPT ACCEPTED BY THE SOCIETY 19 MAY 2010

Printed in the USA