UNIVERSITE AIX-MARSEILLE

Centre Européen de Recherche et d’Enseignement en Géosciences de l’Environnement N° attribué par la bibliothèque :

Active tectonics of the Doruneh Fault: seismogenic behavior and geodynamic role Tectonique active de la faille de Doruneh : comportement sismogénique et rôle géodynamique

THE SE Pour obtenir le grade de DOCTEUR DE L’UNIVERSITE AIX-MARSEILLE Faculté des Sciences et Techniques Discipline : Géosciences de l’environnement

Présentée et soutenue publiquement par

Yassaman FARBOD Le 12 juin 2012 au CEREGE

Directeurs de thèse: Olivier BELLIER, Esmaeil SHABANIAN et Mohammad R. ABBASSI Ecole Doctorale : Sciences de l’Environnement

JURY

Dr. Jean-François RITZ, Université Montpellier II Rapporteur Pr. Claudio FACCENNA, Université Roma III Rapporteur Dr. Michel SEBRIER, Université Pierre et Marie Curie Examinateur Pr. Olivier BELLIER, CEREGE/Université Aix-Marseille Directeur Dr. Esmaeil SHABANIAN, CEREGE/Université Aix-Marseille Co-Directeur Dr. Mohammad Reza ABBASSI, IIEES, Co-Directeur

ANNEE : 2012

Résumé

Tectonique active de la faille de Doruneh : comportement sismogénique et rôle géodynamique

Résumé

Ce travail de thèse porte sur la tectonique active du système de faille de Doruneh (DFS) situé au NE de l'Iran. Une approche combinée de géologie structurale, morpho-tectonique, géomorphologie quantitative et datation par des nucléides cosmogéniques 36Cl et 10Be nous a permis de décrire la cinématique ainsi que le rôle géodynamique du DFS dans le contexte de la collision Arabie-Eurasie. Le DFS comprend trois zones avec des caractéristiques structurales, géomorphologiques et cinématiques distinctes, (a) inverse-senestre pour la WFZ (Ouest) (b) purement senestre pour la CFZ (Centre) et (c) inverse pour l'EFZ (Est). Les âges d'abandon de trois générations de cônes alluviaux quaternaires ont été déterminés à ~12, ~36 et ~100 ka. Ces âges, combinés avec les décalages mesurés, indiquent une vitesse maximum de déplacement horizontal senestre de ~5,3 mm/an pour l'Holocène. Notre analyse d’aléa sismique indique que la longueur maximale d’un segment sismogène varie de 70 à 100 km, ce qui pourrait produire des séismes caractéristiques de magnitude 7.2 à 7.4 avec des intervalles de récurrence de ~750 ans. Nous proposons un nouveau modèle géodynamique dans lequel l’EFZ est impliquée dans une zone de cisaillement dextre d’orientation NNW entre Iran central et Eurasie. Le mouvement vers le nord de l’Iran central est accommodé dans cette zone de relais en transpression par du raccourcissement sur des failles inverses d’orientation NW, et transféré vers le Nord par des faille dextres d’orientation NNW.

Mots clés: NE Iran; faille de Doruneh; morpho-tectonique; segmentation; décrochement; déformation crustale; collision Arabie-Eurasie; datation par nucléides cosmogéniques

Abstract

Active tectonics of the Doruneh Fault: seismogenic behavior and geodynamic role

Abstract This study focuses on the active tectonics of the Doruneh Fault System (DFS) in the north- eastern part of central Iran. A combined approach of geological and morphotectonic mapping, fault kinematic analysis, as well as in situ-produced cosmogenic dating (36Cl and 10Be) allowed us to characterize the active kinematics and geodynamic role of the DFS in the context of the Arabia-Eurasia collision. The DFS comprises Western, Central and Eastern fault zones (WFZ, CFZ, EFZ) with distinct structural, geomorphic and kinematic characteristics. The WFZ is oblique reverse- left-lateral, the CFZ is pure left-lateral and the EFZ is reverse. Exposure ages of ~12, ~36 and ~100 ka have been determined for three generations of alluvial fan abandonment surfaces along the DFS. Combining geomorphic offsets and their related ages yields a maximum left- lateral slip rate of ~5.3 mm/yr for the CFZ during Holocene. The maximum length of independent seismogenic fault segments varies from ~70 to ~100 km that could produce a characteristic earthquake with a magnitude of Mw ≈7.2 to 7.4 and recurrence interval of ~750 years. We propose a geodynamic model in which the EFZ is involved in a NNW-trending dextral shear zone between Central Iran and Eurasia. This implies that the EFZ forms a complex right-lateral transpressional relay zone between the eastern and northeastern Arabia-Eurasia convergence boundaries. The northward motion of Central Iran relative to Eurasia is accommodated by shortening on NW-trending reverse faults, and is transferred northward via NNW-trending dextral faults.

Keywords: NE Iran; Doruneh fault; morpho-tectonics; fault segmentation; strike-slip faulting; continental deformation; Arabia-Eurasia collision; cosmogenic nuclides dating

Discipline : Géosciences de l’Environnement Laboratoire : CEREGE UMR 7330 Europôle de l’Arbois, BP 80 13545, Aix-en-Provence Cedex 4

Remerciements

Le travail de cette thèse rentre dans le cadre du programme franco-iranien de coopération scientifique sur le « Risque sismique » en Iran, dirigé par Denis Hatzfeld (LGIT- CNRS UJF) et Mohsen Ashtiani (IIEES), que nous remercions pour la qualité de cette « coordination ». Cette étude a eu le soutien du Ministère des Affaires Etrangères, de L’INSU–CNRS (programmes PNRN, Diety, IT…), du Ministère de la Recherche (ACI-FNS), du côté français, et de l’IIEES (Téhéran), du côté iranien. Les images SPOT ont été acquises grâce au soutien du programme ISIS (CNES). J’ai bénéficié de dix-sept mois de bourse (BGF, MAE), par l’intermédiaire du SCAC (Ambassade de France en Iran). Je remercie pour leurs soutiens Messieurs Lambert et Lhôte ainsi que Mesdames Mirbaha (SCAC) et Rambi ( CROUS, Aix- en Provence). Ce travail n’aurait pas vu le jour si je n’avais pas eu la chance de rencontrer un très grand nombre de gens formidables. Tout d’abord, je tiens à remercier mon directeur de thèse, Pr. Olivier Bellier, pour m’avoir proposé ce passionnant sujet de thèse. Je le remercie pour la qualité de son encadrement, sa patience, son soutien et sa disponibilité malgré son emploi de temps aussi chargé qu’un président! Je suis très reconnaissante à mes deux Co-directeurs de thèse, Dr. Mohhamad Reza Abbassi et Dr. Esmaeil Shabanian. Douze ans déjà que nous nous connaissons ! Ils m’ont toujours soutenu et encouragé aussi bien dans ma vie professionnelle que personnelle. Une pensée très particulière à Esmaeil qui a toujours cru en moi et qui m’a accompagné pas à pas, dés le début de ma carrière scientifique (Master) jusque à la fin de cette thèse. Merci de m’avoir supporté durant tous ces années, j’espère qu’un jour, je pourrais te rendre seulement une petite partie de tous ce que tu as fait pour moi. Je tiens également à remercier les membres du jury, Dr. Jean-François Ritz (DR CNRS, Université de Montpellier II, France), Pr. Claudio Faccenna (Université Rome III, Italie) et Dr. Michel Sébrier (DR CNRS, Université Pierre et Marie Curie, France) d’avoir accepté de juger ce travail de thèse. J’adresse également mes sincères remerciements à Dr Khaled Hessami Azar, à qui je dois beaucoup. S’il ne s’était pas investi avec autant de sérieux à la constitution et au suivi de mon dossier, probablement que je n’aurai jamais pu faire cette thèse en France. Nous lui devons également beaucoup pour la deuxième et dernière mission primordiale pour mon travail ! Celle-ci n’aurait probablement pas pu s’accomplir sans lui. Quotidiennement au CEREGE, j’ai eu un grand plaisir à bénéficier de l’aide d’autres personnes: les informaticiens: Cyrille, Julien et Antoine; à la bibliothèque, Brigitte Crubezy; à l’école Doctorale, Mme Nelil d’abord puis Isabelle Hammad, pour qui j’ai beaucoup d’estime et de respect. Sans oublier bien sur Michel Decobert, le ‘Papa’ du CEREGE avec toutes ses plaisanteries et histoires qu’il nous a raconté pendant les « pauses café ». J’ai eu le plaisir de partager le bureau 109 avec Stéphane Molliex, Clément Flaux, Fabrice Cuoq et le dernier arrivant Mohammad Gharbi, « le prophète ! ». J’étais la fille unique de ce bureau et je les ai probablement beaucoup embêté afin qu’ils rangent tout les bazars… Je les remercie tous pour leur sympathie et leur solidarité dans les moments difficiles. Je tiens à remercier plus particulièrement Clément qui m’a beaucoup aidé dans mon apprentissage du

français. Tous les jours en face de moi et de bonne humeur, toujours en pleine forme et disponible pour répondre à toutes mes questions avec beaucoup de patience. J’ai une pensée très forte envers Florence Derrieux avec qui j’ai pu supporter le froid et les difficultés dans le laboratoire « 10Be » durant ma première année de thèse. Merci à Laetitia Leanni, Frédéric Chauvet et Valery Guillou qui m’ont beaucoup appris et beaucoup aidé durant les « manips » dans les laboratoires 10Be et 36Cl. Merci également à Didier Bourlèse, Régis Braucher le Lucilla Benedetti pour leur aide et leur conseil concernant les analyses 10Be et 36Cl, ainsi que pour leur encouragement dans mon travail jusqu’à la fin de cette thèse, Je leur suis très reconnaissante. Finalement je remercie Lionel Siame, avec qui j’ai eu quelques courtes mais riches discussions, dommage qu’on n’ait pas pu plus travailler ensemble. Chaudes salutations à Leila et Morteza qui sont toujours là quoi qu’il arrive ainsi qu’à Madjid! Eh oui, je sais que j’étais très fatigante quand je n’arrêtais pas de râler pendant les pauses « déjeuner » et « café » ! mais ça y est, tout est fini maintenant… Je pense également à Sylvain, Guillaume, Cédric, Pierre et Mahabat, Luci, Guenael, Margot et Natalia... D’autres amis de l’IIEES (Téhéran, Iran), Parisa Mobayen et Maryam Farjamfar sur qui je pouvais compter sans inquiétude; mesdames Azimi et Bayani pour toutes leurs aides pendant mes missions et bien sure Mohsen, le meilleur chauffeur de l’IIEES, avec lui sur le terrain tout est possible. Un grande merci à Jules et sa famille, pour leur soutien permanant et leurs encouragements. Il est rassurant de savoir à quel point je peux compter sur vous. Et enfin et surtout ma famille, ma mère et ses forts encouragements durant toute ma vie. Une pensée tout spéciale à mon père qui n’a malheureusement pas pu rester en vie et voir ce jour. Je voudrais donc lui dédier ce manuscrit, lui qui espérait tant me voir docteur !

Contents

INTRODUCTION ...... 11

CHAPTER I ...... 13 GENERAL PRESENTATION

1. Geodynamic framework ...... 15

2. The Doruneh Fault ...... 20

3. Problems ...... 23

4. Methodology ...... 24

4.1. Structural and geomorphic analysis ...... 24

4.2. Fault kinematics and stress states ...... 24

4.3. Left-lateral slip rate determination ...... 24

5. Dissertation plan ...... 25

References ...... 26

CHAPTER II ...... 31 GEOMORPHIC AND STRUCTURAL VARIATIONS ALONG THE DORUNEH FAULT SYSTEM (NE IRAN)

Abstract ...... 33

1. Introduction ...... 34

2. Tectonic and Geological Setting ...... 36

3. The Doruneh Fault System (DFS) ...... 39

4. Fault segmentation ...... 43

4.1. Variations in the structural and geometric characteristics of the DFS ...... 43

4.1.1. The Western Fault Zone (WFZ) ...... 45

4.1.2. The Central Fault Zone (CFZ) ...... 48

4.1.3. The Eastern Fault Zone (EFZ) ...... 50

4.2. Variations in the geomorphology and fault behavior of the DFS ...... 51

5. The present-day kinematics and seismic potential of the Central Fault Zone ...... 58

5.1. The CFZ: an oblique-slip or strike-slip fault zone? ...... 59

5.2. Modern stress state along the CFZ deduced from fault kinematic analyses ...... 61

5.3. Seismic potential of the Central fault zone ...... 64

6. Discussion ...... 66

6.1. Kinematic models of the Doruneh Fault System ...... 66

6.2. Initiation of strike-slip faulting along the Doruneh Fault System ...... 71

7. Conclusion ...... 71

Acknowledgements ...... 73

References ...... 74

CHAPTER III ...... 79 TEMPORAL AND SPATIAL VARIATIONS IN LATE QUATERNARY SLIP RATES ALONG THE DORUNEH FAULT SYSTEM (NE IRAN)

Abstract ...... 81

1. Introduction ...... 82

2. Tectonic setting and general geology ...... 83

3. Cumulative left-lateral offsets along the Doruneh Fault System...... 84

4. Surface exposure dating ...... 90

4.1. Sampling strategy and analytical procedure for cosmogenic dating ...... 90

4.2. Cosmogenic dating results ...... 93

4.2.1. 10Be exposure ages from Q1 surfaces ...... 93

4.2.2. 10Be and 36Cl exposure ages from Q2 surfaces ...... 95

4.2.3. 10Be and 36Cl exposure ages from Q3 surfaces ...... 97

4.3. Timing of the abandonment of Q1, Q2, and Q3 alluvial surfaces ...... 99

5. Discussion ...... 103

5.1. Slip rates along the Doruneh Fault System ...... 103

5.2. Seismotectonic implications of the fault slip rates ...... 107

6. Conclusion ...... 109

Acknowledgement ...... 110

References ...... 111

CHAPTER IV ...... 117 THE ARABIA-EURASIA COLLISIONAL BOUNDARY IN NE IRAN, THE EASTERN TERMINATION OF THE DORUNEH FAULT SYSTEM

1. Introduction ...... 119

2. Active faulting along the Eastern Fault Zone (EFZ) ...... 122

2.1. Geological and structural evidences of right-lateral component of faulting ...... 123

2.2. Geomorphic evidence for right-lateral component of faulting ...... 126

3. Summary and Discussion ...... 131

4. Conclusion ...... 133

References ...... 134

DISCUSSION AND CONCLUSION ...... 137

SYNTHESE ...... 147

1. Introduction ...... 149

2. Contexte géodynamique ...... 149

3. Etude structurale et géométrie du système de faille ...... 151

4. Quantification de la déformation : déplacement fini et vitesse de déplacement ...... 153

5. Initiation du décrochement le long du système de faille de Doruneh...... 155

6. Comportement sismogénique et évaluation de l’aléa sismique ...... 156

7. Rôle géodynamique ...... 156

8. Conclusions et perspectives ...... 158

Bibliographie ...... 159

APPENDIX ...... 161

Figure captions ...... 165

List of tables ...... 173

Introduction

INTRODUCTION

L'Iran est une région de déformation continentale active située dans le domaine de la convergence Arabie-Eurasie. Les données GPS indiquent un mouvement vers le nord de l'Arabie par rapport à l'Eurasie a une vitesse d’environ 22 mm/an au niveau de Bahrein. Une partie significative de cette convergence est accommodée d’une part dans la subduction du Makran et d’autre part dans les collisions du Zagros, de l’Alborz et du Kopeh Dagh, sur des failles inverses et décrochantes subparallèles aux chaînes. A l’Est de l’Iran, ce mouvement de l’Arabie vers le nord se traduit par une zone de cisaillement dextre entre l’Iran Central et le bloc de Helmand (Eurasie). Les arguments tectoniques et géodésiques indiquent que ce mouvement est distribué le long de systèmes de failles décrochantes dextres d’orientation N-S (ex. les failles de Neh, Nayband et Dehshir). Au nord de ces failles dextres, à la latitude ~35°N, la faille de Doruneh s’étend sur ~600 km de manière perpendiculaire à la direction de la convergence. C’est une faille décrochante senestre qui sépare deux domaines aux styles de déformation drastiquement différents : au sud le domaine des failles décrochantes dextres subméridiennes, comprenant le bloc de Lut ; au nord la zone de déformation comprenant les chaînes d’orientation NW-SE du Binalud et du Kopeh Dagh. Notre étude vise à comprendre le comportement la faille de Doruneh et son rôle dans le transfert ou l'accommodation de la déformation. Afin de répondre à ces questionnements, nous avons cherché à caractériser la géométrie, la cinématique et le taux de déplacement de ce décrochement majeur affectant le NE de l’Iran ainsi que ses relations spatiales et temporelles avec les structures environnantes. Pour ce faire, les objectifs principaux suivis dans ce travail sont : (1) Etablir une cartographie détaillée des zones de failles actives afin de pouvoir en déduire une carte de segmentation ; (2) Mesurer les déplacements finis et les vitesses de glissement intégrées sur des échelles de temps de l’ordre de quelques dizaines de milliers

11 Introduction d’années (échelle de quelques cycles sismiques) à plusieurs millions d’années (échelle de la tectonique des plaques) ; (3) Déterminer la cinématique des déformations plio-quaternaires et en déduire les états de contraintes responsables. L’approche pluridisciplinaire développée pour atteindre ces objectifs combine des méthodes de géologie structurale, morpho- tectonique, géomorphologie quantitative et datations par des nucléides cosmogéniques. Les données utilisées au cours des différentes étapes de cette étude ont été intégrées dans un Système d’Information Géographique (SIG) et comprennent des images satellites de haute à très haute résolution (e.g., GeoEye), des modèles numériques de terrain (eg. SRTM), des cartes géologiques et des observations et échantillonnages de terrain

Notre étude est développée sur quatre chapitres : Le chapitre I présente le cadre géodynamique régional et précise la problématique soulevée par la faille de Doruneh ; Le chapitre II concerne l’étude structurale et morpho-tectonique de la partie est de la faille de Doruneh désignée Doruneh Fault System (DFS) dans la suite de ce mémoire, et séparée de la faille de Great Kavir. Les principaux résultats visés sont une carte de segmentation détaillée et un modèle cinématique du système de faille ; Le chapitre III est une étude de géomorphologie quantitative sur le DFS comprenant des mesures de décalages sur des marqueurs quaternaires et leur datation par nucléides cosmogéniques produits in situ. L’échantillonnage au long du DFS doit nous permettre de caractériser les variations des taux de glissement horizontaux et d’en déduire des indications sur le comportement sismogéniques des différents segments de faille ; Le chapitre IV est consacré à la cartographie structurale de la terminaison orientale du DFS afin de comprendre son rôle dans le transfert vers le nord de l’Iran Central par rapport à l’Eurasie. Cette partie permettra de compléter le modèle géodynamique de la convergence Arabie-Eurasie.

12

CHAPTER I

General Presentation:

Geodynamic framework, Problems, Objectives and Methodology

Chapter I

This dissertation is focused on active tectonics along the Doruneh Fault System (DFS) in NE Iran, aiming to better understand its contribution in the accommodation of ongoing deformation due to the Arabia-Eurasia convergence.

1. Geodynamic framework The Iranian plateau absorbs almost all the northward convergence between the Arabian and Eurasian plates. This convergence is mainly taken up by shortening and strike-slip faulting in the interior of the plateau, and the oceanic subduction along the Makran coast. The plateau comprises several block-like regions (e.g., Central Iran, Lut and South Caspian Basin) surrounded by complex deformation domains such as the Zagros, Alborz and Kopeh Dagh mountain ranges (Fig. 1). Both the historical (e.g., Ambraseys and Melville, 1982, Berberian and Yeats, 1999, and 2001) and instrumental (e.g., Jackson and McKenzie, 1984; Berberian, 1979; Engdahl et al., 1998) seismicity confirm that much of the deformation is localized in these mountain belts surrounding the relatively aseismic blocks of Central Iran, Lut and South Caspian Basin) (Fig. 2). The present-day Global Positioning System (GPS) measurements indicate a northward motion of Arabia toward Eurasia (Fig. 3) at a rate of ~22 mm/yr at Bahrain longitude (Sella et al., 2002; McClusky et al., 2003; Vernant et al., 2004; Reilinger et al., 2006). The Arabia-Eurasia convergence is partly taken place in the Zagros fold and thrust belt. This belt results from the closure of the Neo-Tethys ocean, following the northeast-dipping subduction under the Iranian micro-continents. The subsequent collision beginning in the Neogene between the Arabian plate and the Iranian blocks (e.g., Stöcklin, 1968; Falcon, 1974; Berberian and King, 1981; Berberian et al., 1982, and 1995; Mouthereau and Lacombe, 2011, and references therein). The convergence, at a present-day rate of 5 to 9 mm/yr (e.g., Hessami et al., 2006; Vernant et al., 2004), is accommodated by active folding and thrust faulting parallel to the belt, accompanied by strike-slip faulting parallel or oblique to the belt. East of 58°E, most of the shortening is accommodated by the active subduction of the Oman oceanic lithosphere beneath Makran (Byrne et al., 1992), with a rate of 19.5±2 mm/yr (Vernant et al., 2004). The Makran subduction is connected to the Zagros collision by the NNW-trending right-lateral Minab-Zendan fault system that slips at a rate of ~6 mm/yr deduced from both geological and geodetic measurements (Regard et al., 2005, and 2010; Peyret et al., 2009).

15 Chapter I

Figure 1. Simplified structural map of Iran located in the Arabia-Eurasia collision, with main tectono-structural division of the Iranian plateau

On the easternmost border of the Iranian plateau, the very low GPS velocity of the YAZT and ZABO stations relative to Eurasia indicates that the Helmand block (western Afghanistan) belong to stable Eurasia (Jackson and McKenzie., 1984; Vernant et al., 2004; Masson et al., 2007). This implies a right-lateral shear between the Central Iran and the Helmand block which, is accommodated along N-trending right-lateral fault systems on both sides of the Lut block (e.g., Conrad et al., 1981; Tirrul et al., 1983; Meyer and Le Dortz, 2007). The GPS velocities indicate a total shear of 16±2 mm/yr between Central Iran and the Helmand block. About 7±2 mm/yr of this shear is accommodated along the eastern border of the Lut block, i.e., the Neh Fault (Vernant et al., 2004; Masson et al., 2007). Consequently, the northern deformation domains of Alborz, Kopeh Dagh and Allah Dagh-Binalud take up

16 Chapter I the remnant deformation that is not absorbed by the Zagros Mountains and the Makran subduction.

Figure 2. (a) Distribution of earthquake epicenters from historical records (3000 B.C. to 1962 A.D.) (b) Distribution of instrumentally recorded earthquake epicenters (1964 to 1998) from the USGS catalogue. The distribution is similar in a and b and show that the seismicity is broadly confined to the Zagros, Alborz, and Kopeh Dagh mountain belts and to narrow N–S zones surrounding the Dasht-e-Lut.

The Alborz is a curved mountain belt in northern Iran (Fig. 1) that wraps around the rigid South Caspian Basin. The tectonic activity in the Alborz mountain range is due to the northward convergence of Central Iran, and the north-westwards motion of the South Caspian Basin, both with respect to Eurasia. Geological (e.g., Allen et al., 2003b; Ritz et al., 2006; Nazari, 2006) and seismological studies (e.g., Jackson et al., 2002; Tatar et al., 2007; Tatar and Hatzfeld, 2009) have suggested that the present-day deformation in the Alborz Mountains appears partitioned along range-parallel thrusts and left-lateral faults. However, the interaction between the asymmetric “V” shape structure of Alborz and the rotational motion of the South Caspian Basin relative to Eurasia produces a more complex velocity field (see Djamour et al., 2010). At the southern side of the Alborz, left-lateral faulting is well documented by focal mechanisms of earthquakes along (1) WNW-striking faults, such as the Mosha Fault, within the western part of the range, and (2) along NE-trending Firuzkuh and Astaneh faults within the eastern part of the Alborz belt (see Shabanian et al., 2012, and references therein). This left-lateral kinematics is also supported by tectonic geomorphology (e.g., Ritz et al., 2003; Solaymani et al., 2003; Bachmanov et al., 2004; Nazari et al., 2009;

17 Chapter I

Hollingsworth et al., 2010; Javidfakhr et al., 2011b; Solaymani Azad et al., 2011b) and InSAR (Shirzaei et al., 2011) investigations carried out along the major faults of the range such as the Mosha, Taleghan and Firuzkuh faults, as well as the Shahrud fault system. The southwest- to southeast-dipping Khazar Fault forms the boundary between Central Alborz and the South Caspian Basin (e.g., Berberian, 1983; Tatar et al., 2007). The fault follows the northern flank of Central Alborz, with a complex kinematics (e.g., Allen et al., 2003b; Guest et al., 2006; Nazari, 2006), and is suggested to be, in the western part, the source of the 2004 Baladeh thrust faulting earthquake, occurred at the depth of ~35 km (Tatar et al., 2007). However, recent GPS study by Djamour et al. (2010) has revealed that the faults south of the Central Alborz range (Mosha, Firuzkuh, and Astaneh faults) are mainly left-lateral, with the small vertical component of faulting. Conversely, the Khazar Fault is divided into two distinct parts: (1) the western segment is mainly a thrust fault slipping at ~6 mm/yr, with a left-lateral component of ~2-3 mm/yr; (2) the eastern segment is predominately left-lateral (~5 mm/yr) accompanied with a lesser component of thrust faulting (~2-3 mm/yr).

Figure 3. GPS velocity field show that approximately N-S shortening between Arabia and Eurasia is accommodated throughout Iran (after Vernant et al., 2007).

The northeastern part of the Arabia-Eurasia collision zone corresponds to NE Iran, including the Kopeh Dagh, Allah Dagh–Binalud, and Kuh Sorkh mountains (Fig. 4). The NW-trending Kopeh Dagh Mountains form an active deformation domain separating Central

18 Chapter I

Iran from Eurasia. Mesozoic and Tertiary sediments of the Kopeh Dagh were folded into parallel, asymmetric NW-trending folds during Oligo-Miocene compression (Stöcklin, 1968; Afshar Harb, 1979; Lyberis and Manby, 1999). These folds are obliquely dissected by active NNW-trending right-lateral and ENE-trending left-lateral strike-slip faults (e.g., Tchalenko, 1975; Afshar Harb, 1979; Shabanian et al., 2009a). The Allah Dagh–Binalud Mountains form NW-trending Mesozoic paleoreliefs south of the Kopeh Dagh (e.g., Afshar Harb, 1979) that thrusts over the northern margin of Central Iran (Alavi, 1992). The Binalud is a collection of the paleo-Tethys remnants, as well as middle and upper Mesozoic, and Cenozoic rocks deformed since pre-late Triassic (Alavi, 1992). To the NW, the Kopeh Dagh Mountains are bounded by the Main Kopeh Dagh Fault (MKDF) which is considered as the NE limit of the Arabia-Eurasia convergence (e.g., Trifonov 1978; Lyberis and Manby, 1999; Shabanian et al., 2009a). Near 57°E meridian, the Main Kopeh Dagh Fault (MKDF) is intersected by the NNW-trending right-lateral Bakharden- Fault System (BQFS) which dissects the Kopeh Dagh folds (e.g., Afshar Harb, 1979; Shabanian et al., 2009a). The BQFS extends between the MKDF to the north and the Binalud range to the southeast. Recent quantitative geological and geomorphic investigations (Shabanian et al., 2009a, 2009b) in the Kopeh Dagh and Binalud ranges show that the deformation is accommodated by strike-slip faulting along NE-trending left-lateral and NNW-trending right-lateral strike-slip faults, as well as NW-trending reverse to oblique faults. The regional deformation is achieved without any partitioning of stress or strain, neither fault block rotation as proposed by previous authors (e.g., Jackson and McKenzie, 1984; Jackson et al., 1995; Hollingsworth et al., 2006). According to available geodetic (Vernant et al., 2004; Reillinger et al., 2006; Masson et al., 2007; Tavakoli, 2007) and geological (Shabanian et al., 2009a) data, the Arabia northward motion is accommodated in northeast Iran with a rate ranging from 4 to 11mm/yr. A significant part of this deformation (8±2 mm/yr) is localized along the BQFS, suggesting that this fault system forms the boundary between Iran and Eurasia (Shabanian et al., 2009a). The right-lateral motion along the BQFS transfers to the Binalud mountain range through the Meshkan Transfer Zone and terminates to the dextral reverse Neyshabur fault system and right-lateral fault zone on the southwest and northeast side of the Binalud Mountains, respectively as a crustal-scale restraining relay zone (Shabanian, 2009; Shabanian et al., submitted to GJI). South of the Koh Sorkh range, at ~35°N latitude, the E-trending Doruneh Fault (e.g., Wellman, 1966; Tchalenko et al., 1973; Mohajer-Ashjai, 1975) separates two regions with

19 Chapter I different deformation styles. In the southern domain, i.e., the Lut block-Central Iran, the N- trending dextral strike-slip faults are dominant. The northern domain comprises the NW- trending Kopeh Dagh and Binalud mountain ranges.

2. The Doruneh Fault The Doruneh Fault has been presented for the first time by Wellman (1966) as the longest left-lateral strike-slip fault in the Iranian plateau extending over ~600-km-long from the eastern Iranian border to the Great Kavir desert (Fig. 4). The general trace of the Doruneh Fault was mapped at regional scale and indicates a northward convex geometry (Stocklin and Nabavi, 1973; Eftekhar-Nezhad et al., 1976; Huber, 1977). Despite clear Quaternary activity of the Doruneh Fault expressed in the geomorphology, no historical and instrumental earthquakes of magnitudes M ≥ 6 (Ambraseys and Melville, 1982; Jackson and McKenzie, 1984) have been recorded along this major fault. Regarding to scaling relationships (e.g., Scholz, 1982; Wells and Coppersmith, 1994), large faults such as the Doruneh Fault with length of >100 km have the potential to produce earthquakes with magnitudes higher than 7.5. Such a possible high seismic potential has been inferred by Berberian and Yeats (1999) for the Doruneh Fault. However, before any seismic hazard assessment, it should be taken into account that most faults, particularly long ones, do not rupture along their entire length during a single earthquake (King and Nabelek, 1985; Barka and Kadinsky-Cade, 1988; Knuepfer, 1989; Schwartz and Sibson, 1989; dePolo et al., 1991; McCalpin, 1996; Stewart and Taylor, 1996; Barka et al., 2002; Kondo et al., 2005; Ikeda et al., 2009). Actually, many of these apparently continuous long fault systems can be divided into discrete fault zones or segments, at a variety of scales, that have different geometries, faulting histories and seismicity (e.g., Schwartz and Coppersmith, 1984; Bruhn et al., 1987; Knuepfer, 1989; dePolo et al., 1991; Stewart and Taylor, 1996). Fault segmentation has emerged as an interdisciplinary field of earthquake research, based on segment boundaries identification. Five concepts of fault segmentation such as “earthquake”, “behavioral”, “structural”, “geological” and “geometric” have been proposed by dePolo et al (1991) and McCalpin (1996) to identify the segment boundaries from the surface observation of the fault.

20 Chapter I

Figure 4. Principal deformation domains and structural units in Central and northeastern Iran overlain on GTOPO30 shaded relief. The abbreviations are : MKDF, Main Kopeh Dagh Fault; BQFS, Bakharden-Quchan Fault System; CF, Chakaneh Fault; NF, Neyshabur Fault; KF, Kashafrud Fault; DFS, Doruneh Fault System; GKFS; Great Kavir Fault System.

Before this study, the detailed kinematics, geometry and segmentation of the Doruneh Fault were poorly known. The preliminary division of the Doruneh Fault into the eastern and western sectors (Tchalenko et al., 1973; Mohajer-Ashjai, 1975) was based on the changes in the orientation of the fault trace from NE to E near Doruneh village. Given the curved geometry of the Doruneh Fault and its mean E-W orientation, perpendicular to the northward motion of Central Iran, a clockwise rotation model has been proposed by Walker and Jackson (2004) (see also Jackson and McKenzie, 1984). This model was relied on the hypotheses that (1) the Doruneh Fault was initially an E-W linear structure, and (2) both the northern and southern sides of the fault are rigid rotating blocks. But, these hypotheses are contradicted by key observations; for instance, there is no geological data supporting the linear initial geometry of the fault. Furthermore, the geometry of individual

21 Chapter I rotating rigid blocks has never been determined, and at least, the northern side of the Doruneh Fault cannot be considered as a rigid block. The recent tectonic geomorphology study by Fattahi et al. (2007) proposes a left-lateral slip rate of ~2.5 mm/yr, and a recurrence interval of ~2000 ka for the large-magnitude earthquakes (Mw 7-7.5) along the central part of the Doruneh Fault. This slip rate has been estimated in one location, near the town. However, it is well known that horizontal offsets and consequently the slip rate may significantly vary along the strike of long strike- slip faults (e.g., Deng et al., 1986). In such a case and in a lack of knowledge of the slip distribution along the fault, a single slip rate could not be representative of the overall fault slip. In summary, despite the valuable knowledge provided by pioneer and previous workers, there are still a lot of unclear, and sometimes, inconsistent points that raise the necessity of new investigations on the active tectonics of the Doruneh Fault. The detailed knowledge of the kinematics and structure of the Doruneh Fault would help to identify its geodynamic role in the context of the Arabia-Eurasia collisional convergence. In order to achieve these objectives, at both the regional and geodynamic scales, we carried out detailed structural, kinematic and geomorphic analyses along the entire length of the Doruneh Fault System. First, after the distinction between the Great Kavir and the Doruneh Fault Systems (both constituting the general “Doruneh Fault”), we have performed a segmentation, at the scale of fault zones, along the Doruneh Fault System. This segmentation is based on detailed geomorphic and geological analyses of high- to very high-resolution satellite images (SPOT 5 and GeoEye©-Google Earth) and digital elevation models (mainly SRTM data), complemented by filed observations and available geological data (e.g., geological maps and their associated cross sections). These analyses reveal evidence of drastic changes in geological, geomorphic and structural characteristics of the fault system, leading us to establish a fault segmentation model comprising three fault zones. In the second stage, we conducted more detailed geomorphic investigations, at the site scale, along the fault zones. This allowed us to characterize the style and distribution of active faulting along each fault zone. The exposure dating of the alluvial fan abandonment surfaces offset along the central and western parts of the DFS led us determining late Quaternary fault slip rates. In the final stage, the analysis of a collection of new and original data and deduced results helped us to propose two kinematic and geodynamic models to explain how the DFS contributes to the accommodation of the Arabia-Eurasia active convergence.

22 Chapter I

3. Problems The geological and geodetic investigations reveal that the northward motion of Arabia relative to Eurasia has similar displacement rates in its eastern and northeastern boundaries (Vernant et al., 2004; Tavakoli, 2007; Shabanian et al., 2009a). In eastern Iran, the Central Iran northward motion is localized along intracontinental right- lateral strike-slip faults such as the Sistan and Neh fault systems (Conrad et al., 1981; Tirrul et al., 1983; Jackson and McKenzie, 1984). Farther north in NE Iran, the NNW-trending dextral Bakharden-Quchan and Chakaneh fault systems are known as the northeastern boundary of the Arabia-Eurasia convergence (Shabanian et al., 2009a and 2009b). All these strike-slip faults end into reverse to oblique-slip bend-terminations (e.g., Kashafrud Fault and Neyshabur Fault System - Hollingsworth et al., 2006; Shabanian, 2009; Shabanian et al., submitted to GJI), which in turn form the restraining fault relay zones (Fig. 4). The DFS, between these two deformation domains, is a key segment in the geodynamic puzzle of Arabia-Eurasia collision.

The aim of this study is to answer two simple but fundamental questions:

(1) How does the crustal strike-slip DFS take up overall tectonic block motions perpendicular to its strike? (2) What is the geodynamic role of the DFS in the accommodation and/or the transfer of the deformation resulting from the Arabia-Eurasia plate convergence?

To deal with these questions, the following main objectives need to be achieved:

(1) To provide a detailed structural pattern of active faults in the DFS for which a precise 2-D fault geometry, structural relationships and possible interactions between major faults can be described. (2) To characterize the modern stress state responsible for Quaternary deformation in order to better understand the kinematic relationships between interacting faults. (3) To identify the kinematics, style and distribution of active faulting along the DFS, and to describe the kinematics of active deformation at the regional scale. (4) To determining the DFS fault slip rate and its possible spatiotemporal variations during the Late Pleistocene and Holocene time periods. (5) To do a preliminary assessment of the seismic behavior of the DFS, providing a pertinent dataset that would pave the way for further paleoseismological studies.

23 Chapter I

4. Methodology To attain these main objectives, a multidisciplinary approach combining structural geologic, geomorphologic, and dating methods has been performed. All the original and available, multi-scale data were integrated into a Geographic Information System, allowing the spatial analysis of data during investigations.

4.1. Structural and geomorphic analysis The fault trace were mapped by identification and characterization of geomorphic scarps and offsets using analysis of satellite images with different resolutions ranging from medium to very high (Landsat ETM+, SPOT5, Quickbird and GeoEye©-Google Earth), and digital elevation models (SRTM data). This mapping was performed at both the scale of fault zones (~100 km) and the scale of fault segments (~10 km), and then has been complemented by field surveys. Consequently, fault traces were mapped in detail and their activity has been evaluated using Quaternary geomorphic markers, such as alluvial fans and drainage networks, offset along the faults. These allowed us to provide independent datasets of well-constrained cumulative offsets measured in sixty seven sites along the DFS extending over time periods of several thousands of years.

4.2. Fault kinematics and stress states To characterize the Plio-Quaternary stress state responsible for faulting along the DFS, the fault kinematic data (fault planes and their associated striations) measured in five sites have been analyzed using the inversion schema originally proposed by Carey and Brunier (1974). The chronology of different generation of striations was determined using crosscutting relationships between fault planes and/or striations. The last generation of striations has used to determine the modern stress state. These results, representing the stress state along the DFS, were integrated in the context of NE Iran to evaluate the regional significance of the stress state acting on the DFS.

4.3. Left-lateral slip rate determination Fault slip rates, integrated over the Holocene and the Pleistocene, have been determined through the dating of the geomorphic markers offset along the DFS. The choice of the dating methods depends upon the nature of the material constituting the markers and on the time scales. The surface exposure dating by the 10Be and 36Cl cosmogenic nuclides (e.g., Goss and

24 Chapter I

Philips, 2001) has been used to date the abandonment surface of alluvial fans composed of quartz-rich or carbonate clasts, respectively.

5. Dissertation plan The present study is organized around the three following chapters:

Chapter II: Geomorphic and structural variations along the Doruneh Fault System. Chapter III: Temporal and spatial variations in late Quaternary slip rates along the Doruneh Fault System. Chapter IV: The Arabia-Eurasia collisional boundary in NE Iran, eastern termination of the Doruneh Fault System.

In chapter II, along-strike variations in structural, geomorphic and behavioral characteristics of the fault system is evaluated to propose a segmentation model for the DFS, at the scale of fault zones. Based on this segmentation and the kinematics of different fault zones within the DFS, a kinematic model is presented to explain how active deformation due to the northwards Arabia-Eurasia convergence is accommodated along the DFS, and other adjusting faults in interaction with this fault system.

In chapter III, we present a systematic measurement of cumulative geomorphic offsets recorded by three generations of inset Quaternary alluvial fans. The determination of exposure ages for the abandonment surfaces of these alluvial fans leads to examine spatiotemporal variations in cumulative offsets, and consequently slip rates along the western and eastern parts of the fault system. The distribution of slip during the Holocene is used to define the extent of seismogenic fault segments and their persistent boundaries. These data has important implications in the assessment of seismic behavior of the DFS.

Finally, in chapter IV, remote-sensing geological and geomorphic investigations along the eastern termination of the DFS allow to describe the geodynamic role of the DFS in the accommodation\transfer of the northward motion of Central Iran with respect to Eurasia. This objective is achieved by detailed structural mapping of the eastern termination and by integrating all related data and interpretations at the regional scale.

25 Chapter I

References Afshar Harb, A. (1979), The stratigraphy, tectonics and petroleum geology of the Kopet Dagh region, northeastern Iran, PhD thesis, Petroleum Geology Section, Royal School of Mines, Imperial College of Science and Technology, London. Alavi, M. (1992), Thrust tectonics of the Binalood region, NE Iran, Tectonics, 11, 360-370, doi:10.1029/91TC02217. Allen, M. B., M. R. Ghassemi, M. Shahrabi, and M. Qorashi (2003), Accommodation of late Cenozoic oblique shortening in the Alborz range, northern Iran, J. Struct. Geol., 25, 659- 672. Ambraseys, N., and Melville, C., (1982), A History of Persian Earthquakes, Cambridge University Press: Cambridge, UK. Ambraseys, N.N., (1978). The relocation of the epicentres in Iran, Geophys. J. R. astr. Soc., 53, 117–121. Bachmanov, D. M., V. G. Trifonov, K. Hessami, A. I. Kozhurin, T. P. Ivanova, E. A. Rogozhin, M. C. Hademi, and F. H. Jamali (2004), Active faults in the Zagros and central Iran, Tectonophysics, 380, 221-241. Barka, A. A., and K. Kadinsky-Cade (1988), Strike-slip fault geometry in Turkey and its influence on earthquake activity, Tectonics, 7, 663–684. Barka, A., et al. (2002), The surface rupture and slip distribution of the 17 August 1999 I˙zmit earthquake, (M 7.4), North Anatolian fault, Bull. Seismol. Soc. Am., 92(1), 43-60. Berberian, F., I. D. Muir, R. J. Pankhurst, and M., Berberian (1982), Late Cretaceous and early Miocene Andean type plutonic activity in northern Makran and central Iran, Journal of Geological Society of London, 139, 605-614. Berberian, M. (1983), The southern Caspian: a compressional depression floored by a trapped, modified oceanic crust. Canadian Journal of Earth Sciences, 20, 163-183. Berberian, M. (1995), Master "blind" thrust faults hidden under the Zagros folds:active basement tectonics and surface morphotectonics, Tectonophysics, 241, 193-224. Berberian, M., (1979). Evaluation of instrumental and relocated epicentres of Iranian earthquakes, Geophys. J. R. astr. Soc., 58, 625–630. Berberian, M., and G. C. P. King (1981), Canadian Journal of Earth Science, National Research Council of Canada, 18, 210-265. Berberian, M., and R. Yeats (1999), Patterns of Historical Earthquake rupture in the Iranian Plateau, Bull. Seismol. Soc. Am., 89(1), 120-139. Berberian, M., and R. Yeats (2001), Contribution of archaeological data to studies of earthquake history in the Iranian Plateau, J. Struct. Geol., 23, 563-584. Bruhn, R. L., Sibler, P. R. and Parry, W. T. (1987), Rupture characteristics of normal faults: an example from the Wasatch fault zone, Utah. In: Continental Extensional Tectonics (edited by Coward, M. P., Dewey, J. F. and Hancock, P. L.). Spec. Pub. geol. Soc. Land., 337-353. Byrne D.E., Sykes, L.R. and Davis, D.M., Great thrust earthquakes and aseismic slip along the plate boundary of the Makran subduction zone, J. Geophys. Res., 97, 449-478. Carey, E., and B. Brunier (1974), Analyse théorique et numérique d'un modèle mécanique élémentaire appliqué à l'étude d'une population de failles, C.R. Acad. Sci., Paris, 179, 891-894. Conrad, G., R. Montigny, R. Thuizat, and M. Westphal (1981), Tertiary and Quaternary geodynamics of southern Lut (Iran) as deduced from palaeomagnetic, isotopic and structural data, Tectonics, 75, T11-T17. Deng, Q., D. Wu, P. Zhang, and S. Chen (1986), Structure and deformational character of strike-slip fault zone, PAGEOPH, 124(1/2).

26 Chapter I dePolo, C. M., D. G. Clark, D. B. Slemmons, and A. Ramallie (1991), Historical basin and range province surface faulting and fault segmentation. J. Struct. Geol., 13, 123-136. Djamour, Y., et al. (2010), GPS and gravity constraints on continental deformation in the Alborz mountain range, Iran, Geophys. J. Int., 1287-1301, doi: 10.1111/j.1365- 246X.2010.04811.x. Eftekhar-Nezhad, J., A. Aghanabati, B. Hamzehpour, V. Baroyant (1976), Geological map of Iran, Ser. J5, Kashmar sheet, Scale 1: 250 000, Geol. Surv. of Iran,Tehran. Engdahl, E.V.d.H., R.: Buland, R, (1998), Global teleseismic earthquake relocation with improved travel times and procedures for depth determination: BSSA, v. 88. Falcon, N. L. (1974), Souther Iran: Zagros Mountains, in Mesozoic-Cenozoic orogenic belts, data for orogenic studies, Geological Society of London special publication, 199-211. Fattahi, M., R. T. Walker, M. M. Khatib, A. Dolati and A. Bahroudi; (2007), Slip-rate estimate and past earthquakes on the Doruneh fault, eastern Iran, Geophys. J. Int. 168, 691–709 doi: 10.1111/j.1365-246X.2006.03248.x Gosse, J., F. M. Philips (2001), Terrestrial in situ cosmogenic nuclides: theory and application, Quaternary science reviews, 20, 1475-1560. Guest, B., G. J. Axen, P. S. Lam, J. Hassanzadeh (2006), Late Cenozoic shortening in the west central Alborz mountains, northern Iran, by combined conjugate strike-slip and thin- skinned deformation, Geosphere, 2, 35-52. doi:10.1130/GES00019.1. Hessami, K., Nilforoushan, F., and Talbot, C. (2006), Active deformation within the Zagros Mountains deduced from GPS measurements, J. Geol. Soc. Lond., 163, 143–148. Hollingsworth, J., J. Jackson, R. Walker, M. Gheitanchi, and M. Bolourchi (2006), Strike-slip faulting, rotation, and along-strike elongation in the Kopeh Dagh mountains, NE Iran, Geophys. J. Int., 166, 1161–1177, doi: 10.1111/j.1365-246X.2006.02983.x. Hollingsworth, J., M. Fattahi, R. Walker, M. Talebian, A. Bahroudi, M. J. Bolourchi, J. Jackson, and A. Copley (2010), Oroclinal bending, distributed thrust and strike-slip faulting, and the accommodation of Arabia–Eurasia convergence in NE Iran since the Oligocene, Geophys. J. Int. 181(3), 1214-1246, doi: 10.1111/j.1365-246X.2010.04591.x. Huber, H. (1977), Geological map of Iran, Sheet, Ser. No.2 and 3, North Central Iran and North East Iran, Scale sheets, scale 1:1 000 000 National Iranian Oil company, Tehran. Ikeda, M., S. Toda, S. Kobayashi, Y. Ohno, N. Nishizaka, and I. Ohno (2009), Tectonic model and fault segmentation of the Median Tectonic Line active fault system on Shikoku, Japan, Tectonics, 28, TC5006, doi: 10.1029/2008TC002349. Jackson, J., A. Haines, A. J. , and W. E. Holt (1995), The accommodation of Arabia-Eurasia plate convergence in Iran, J. Geophys. Res., 100(15), 205-219. Jackson, J., and Mckenzie, D., (1984), Active tectonics of the Alpine-Himalayan Belt between Turkey and Pakistan.: Geophys. J. R. astr.Soc, v. 77, p. 214-245. Jackson, J., Priestley, K., Allen, M., and Berberian, M. (2002), Active tectonics of the South Caspian Basin, Geophys. J. Int., 148, 214–245. Javidfakhr, B., O. Bellier, E. Shabanian, L. Siame, L. Léanni, D. Bourlès, and S. Ahmadian (2011b), Fault kinematics and active tectonics at the southeastern boundary of the eastern Alborz (Abr and Khij fault zones): Geodynamic implications for NNE Iran, Journal of Geodynamics, 52(3-4), 290-303. King, G. C. P., and J. Nabelek (1985), Role of fault bends in the initiation and termination of earthquake rupture, Science, 228, 984-987. Knuepfer, P. L. K. (1989), Implication of the characteristics of end-points of historical surface fault ruptures for the nature of fault segmentation, U.S. Geol. Surv. Open-File Rep., 193- 228. Kondo, H., Y. Awata, O. Emre, A. Dougan, S. Ozalp, F. Tokay, C. Yildirim, T. Yoshioka, and K. Okumura (2005), Slip distribution, fault geometry, and fault segmentation of the

27 Chapter I

1944 Bolu-Gerede earthquake rupture, North Anatolian fault, Turkey, Bull. Seismol. Soc . Am., 95, 1234 – 1249, doi : 10.1785/0120040194. Lyberis, N., and Manby, G. (1999), Oblique to orthogonal convergence across the Turan block in the post-Miocene, Am. Assoc. Petrol. Geol. Bull., 83(7), 1135-1160. Masson, F., M. Anvari, Y. Djamour, A. Walpersdorf, F. Tavakoli, M. Daignieres, H. Nankali, and S. Van Gorp (2007), Large-scale velocity field and strain tensor in Iran inferred from GPS measurements: new insight for the present-day deformation pattern within NE Iran, Geophys. J. Int., 170, 436-440, doi: 10.1111/j.1365-246X.2007.03477.x. McCalpin, J. P. (Eds) (1996), Paleoseismology, Academic, San Diego, California. McClusky, S., R. Reilinger, S. Mahmoud, D. Ben Sari, and A. Tealeb (2003), GPS constraints on Africa (Nubia) and Arabia plate motions, Geophys. J. Int., 155(1), 126-138. Meyer, B., and K. Le Dortz (2007), Strike-slip kinematics in Central and eastern Iran: Estimating fault slip rates averaged over the Holocene, Tectonics, 26, TC5009, doi: 10.1029/2006TC002073. Mohajer-Ashjai, A., (1975), Recent and contemporary crustal deformation in eastern Iran: London, Imperial college London. Mouthereau, F., O. Lacombe, and J. Vergés (2012), Building the Zagros collisional orogen: Timing, strain distribution and the dynamics of Arabia/Eurasia plate convergence, Tectonophysics, in press, doi: 10.1016/j.tecto.2012.01.022. Nazari, H., J-F. Ritz, A. Shafei, A. Ghassemi, R. Salamati, J. L. Michelot, and M. Massault (2009), Morphological and paleoseismological analyses of the Taleghan fault, Alborz, Iran, Geophys. J. Int., 178, 1028-1041. Nazari, H. (2006), Analyse de la tectonique récente et active dans l'Alborz Central et la région de Téhéran: Approche morphotectonique et paléoseismologique. Ph.D. thesis, Faculté des Sciences et des techniques du Languedoc, l'Université Montpellier II, France, 246 pp. Peyret, M., et al. (2009), Present-day strain distribution across theMinab-Zendan-Palami fault system from dense GPS transects, Geophys. J. Int., 179, 751-762, doi: 10.1111/j.1365- 246X.2009.04321.x. Regard, V., D. Hatzfeld, M. Molinaro, C. Aubourg, R. Bayer, O. Bellier, F. Yamini-fard, M. Peyret, and M. Abbassi (2010), The transiotion between Makran subduction and the Zagros collision: recent advances in its structure and active deformation, Geological Society of London special publication, 330, 43-64, doi: 10.1144/SP330.4. Regard, V., et al. (2005), Cumulative right-lateral fault slip rate across the Zagros–Makran transfer zone: role of the Minab–Zendan fault system in accommodating Arabia–Eurasia convergence in southeast Iran, Geophys. J. Int., 162, 177-203, doi: 10.1111/j.1365- 246X.2005.02558.x. Reilinger, R., et al. (2006), GPS constraints on continental deformation in the Africa-Arabia- Eurasia continental collision zone and implications for the dynamics of plate interactions, J. Geophys. Res., 111, B05411, doi: 10.1029/2005JB004051. Ritz, J. F., H. Nazari, A. Ghassemi, R. Salamati, A. Shafei, S. Soleymani, and P. Vernant (2006), Active transtension inside central Alborz: A new insight into northern Iran– southern Caspian geodynamics, Geology, 34, 477-480, doi: 10.1130/G22319.1. Ritz, J-F., S. Balescu, S. Solaymani, et al. (2003), Determining the long-term slip rate along the Mosha Fault, Central Alborz, Iran, Implications in terms of seismic activity, paper presented at 4th International Conference of Seismology and Earthquake Engineering, Tehran, Iran. Scholz, C. H. (1982), Scaling laws for large earthquakes: Consequences for physical models, Bull. Seismol. Soc. Am., 72, 1– 14.

28 Chapter I

Schwartz, D. P., and K. J. Coppersmith (1984), Fault Behavior and Characteristic Earthquakes' Examples From the Wasatch and San Andreas Fault Zones, J. Geophy. Res., 89(B7), 5681-5698. Schwartz, D.P. and Sibson, R.H. (Eds) (1989), Fault segmentation and controls of rupture initiation and termination, USGS Open-File Report, 89-315, Menelo park, California. Sella, F. G., H. T. Dixon, and A. Mao (2002), Revel: A model for Recent plate velocities from space geodesy, J. Geophys. Res., 107(B4), ETG 11. Shabanian, E., L. Siame, O. Bellier, L. Benedetti, and M. R. Abbassi (2009a), Quaternary slip rates along the northeastern boundary of the Arabia–Eurasia collision zone (Kopeh Dagh Mountains, Northeast Iran), Geophys. J. Int., 1-23, doi: 0.1111/j.1365- 246X.2009.04183.x. Shabanian, E., O. Bellier, L. Siame, M. R. Abbassi, D. Bourlès, and Y. Farbod, The Binalud Mountains, a key piece 1 for the geodynamic puzzle of NE Iran. Submitted to Geophys. J. Int. Shabanian, E., O. Bellier, L. Siame, N. Arnaud, M. R. Abbassi, and J.-J. Cochemé (2009b), New tectonic configuration in NE Iran: Active strike-slip faulting between the Kopeh Dagh and Binalud mountains, Tectonics, 28, TC5002, doi: 10.1029/2008TC002444. Shabanian, E., O. Bellier, M. R. Abbassi, L. Siame, and Y. Farbod (2010), Plio-Quaternary stress states in NE Iran: Kopeh Dagh and Allah Dagh-Binalud mountain ranges, Tectonophysics, 480, 280-304, doi: doi:10.1016/j.tecto.2009.10.022. Shabanian, E., V. Acocella, A. Gioncada, H. Ghasemi, and O. Bellier (2012), Structural control on volcanism in intraplate post collisional settings: late Cenozoic to Quaternary examples of Iran and Eastern Turkey, Tectonics, in press, doi: 10.1029/2011TC003042. Shirzaei M., T.R. Walter, H.R. Nankali and E.P. Holohan (2011), InSAR time series Gravity- driven deformation of Damavand volcano, Iran, detected through, Geology, 39, 251-254, doi: 10.1130/G31779.1 Solaymani Azad, S., J-F. Ritz, and M. R. Abbassi (2011b), Left-lateral active deformation along the Mosha–North Tehran fault system (Iran): Morphotectonics and paleoseismological investigations, Tectonophysics, 497, 1-14, doi:10.1016/j.tecto.2010.09.013. Solaymani, S., K. Feghhi, E. Shabanian, M. R. Abbassi, and J-F. Ritz (2003), Preliminary Paleoseismological Studies on the Mosha Fault at Mosha Valley, Rep., International Institute of Earthquake Engineering and Seismology, Tehran, Iran, 89 pp (in Persian). Stewart, M. E., and W. J. Taylor (1996), Structural analysis and fault segment boundary identification along the Hurricane fault in southwestern Utah, J. Struct. Geol., 18(8), 1017-1029. Stöcklin, J., (1968). Structural history and tectonics of Iran: A review, Am. Assoc. Petr. Geol. Bull., 52(7), 1229–1258. Stocklin, J., and M.H., Nabavi (1973), Tectonic map of Iran, Scale, 1: 2 500 000, Geol. Surv. of Iran, Tehran. Tatar, M., and D. Hatzfeld (2009), Microseismic evidence of slip partitioning for the Rudbar- Tarom earthquake (Ms 7.7) of (1990) June 20 in NW Iran, Geophisical Journal International, 176, 529-541. Tatar, M., J. Jackson, D. Hatzfeld, and E. Bergman (2007), The 28 May 2004 Baladeh earthquake (Mw 6.2) in the Alborz, Iran: implications for the geology of the South Caspian Basin margin and for the seismic hazard of Tehran, Geophys. J. Int., 170, 249- 261. Tavakoli, F., (2007), Present-day kinematics of the Zagros and east of Iran faults, Ph.D. thesis, Université Joseph Fourier, Grenoble, France (available at http://tel.archives- ouvertes.fr/tel-00285919/fr/

29 Chapter I

Tchalenko, J. S. (1975), Seismotectonics framework of the North Tehran fault, Tectonophysics, 29, 411-420. Tchalenko, J.S., Berberian, M., and Behzadi, H., (1973), Geomorphic and seismic evidence for recent activity on the Doruneh fault; Iran.: Tectonophysics, v. 19, p. 333-341. Tirrul, R., Bell, I. R., Griffis, R. J., and Camp, V. E. (1983), The Sistan suture zone of eastern Iran, Geol. Soc. Am. Bull., 94, 134 – 150. Trifonov, V. (1978), Late Quaternary tectonic movements of western and central Asia, Geol. Soc. Am. Bull., 89, 1059-1072. Vernant, P., et al. (2004), Present-day crustal deformation and plate kinematics in the Middle East constrained by GPS measurements in Iran and northern Oman, Geophys. J. Int., 157(1), 381-398, doi: 10.1111/j.1365-246X.2004.02222.x. Walker, R., and J. Jackson (2004), Active tectonics and late Cenozoic strain distribution in central and eastern Iran, Tectonics, 23, TC5010, doi: 10.1029/2003TC001529. Wellman, H.W., (1966), Active wrench faults of Iran, Afghanistan and Pakistan: Geologische Undschau, V.55, P. 716-735. Wells, D. L., and K. J. Coppersmith (1994). New Empirical Relationships among Magnitude, Rupture Length, Rupture Width, Rupture Area, and Surface Displacement. Bull. Seism. Soc. Am. 84, 974–1002.

30

CHAPTER II

Geomorphic and structural variations along the Doruneh Fault System (NE Iran)

Chapter II

Geomorphic and structural variations along the Doruneh Fault System (NE Iran)1 Yassaman Farbod1, 2 , Olivier Bellier1, Esmaeil Shabanian1, Mohammad Reza Abbassi2

1.Aix-Marseille Université, CEREGE, UMR 6635, 13545 Aix en Provence cedex 4, France CNRS, CEREGE, UMR 6635, 13545 Aix en Provence cedex 4, France IRD, CEREGE, UMR 161, 13545 Aix en Provence cedex 4, France Collège de France, CEREGE, 13545 Aix en Provence cedex 4, France, [email protected], [email protected], [email protected].

2. International Institute of Earthquake Engineering and Seismology (IIEES), Tehran, Iran [email protected].

Abstract

This paper focuses on the analysis of geomorphic, structural, and behavioral characteristics along the Doruneh Fault System (DFS), east of longitude 56°45’E. Detailed geomorphic and structural analyses of different scale satellite images and digital topographic data, accompanied with field surveys allowed us to establish a fault segmentation model in which three discrete fault zones have been recognized: (1) the western fault zone (WFZ) characterized by reverse left-lateral mechanism with left-handed step-over geometry, (2) the central fault zone (CFZ) which is pure left-lateral strike-slip and comprises nearly parallel faults, and (3) The eastern fault zone (EFZ) that is a trailing contractional imbricate fan fault- termination characterized by reverse faulting and fault-related folding. Each fault zone shows discrete geometry and kinematics implying that deformation is not uniformly accommodated along the DFS. We propose a new kinematic model to explain how the DFS accommodate the Arabia-Eurasia convergence normal to the overall fault orientation. According to this model, the DFS takes up the northward motion between Central Iran – Lut block relative to Eurasia by a complex kinematics varying from pure reverse to pure left-lateral strike-slip faulting. The kinematics of the WFZ and EFZ corresponds to the direction of the NE-trending regional compression. While, the partitioning of slip into strike-slip and reverse component of faulting on parallel faults (strain partitioning) allows the CFZ to remain pure left-lateral strike-slip. Such a model propose a way to explain how large strike-slip faults such as the DFS accommodate tectonic block motions perpendicular to strike of the faults.

Keywords: Fault segmentation, strike-slip faulting, fault kinematic inversion, strain partitioning, active tectonics.

1 Farbod, Y., O. Bellier, E. Shabanian, and M. R. Abbassi (2011), Geomorphic and structural variations along the Doruneh Fault System (Central Iran), Tectonics, 30, TC6014, doi:10.1029/2011TC002889.

33 Chapter II

1. Introduction As common features in intracontinental deformation domains, long straight strike-slip fault systems such as the Doruneh, Neh, and Nayband faults affect the seismically active Iranian plateau (Fig. 1). The seismogenic behavior of a few individual faults have locally been characterized thank to paleoseismologic studies (e.g., Hessami et al., 2003; Ritz et al., 2006; Nazari et al., 2009; Solaymani, 2009), or thank to a rich, but local, historical and archeological record spanned several thousand years (e.g., Ambraseys and Melville, 1982; Berberian and Yeats, 1999, 2001), as well as occurrences of large (M ≥ 6) instrumental earthquakes. However, the historical records of earthquakes are biased toward regions close to trade routes, and are not complete everywhere around the country. Thus, the seismogenic behavior of other large faults remains unknown due to the lack of instrumental and/or historical earthquake records along and in the vicinity of those faults. In Central Iran, the geologically active Doruneh Fault System (Fig. 1) has not sustained historical earthquakes of magnitudes M ≥ 6, raising the possibility that it could produce earthquakes of large (M ≥ 8) magnitudes (Berberian and Yeats, 1999) resulting in a high seismic hazard. The explanations for such a hypothesis could be (1) a long recurrence interval of several thousand of years (Berberian and Yeats, 1999; Fattahi et al., 2007) during which the large faults are accumulating strain and no energy release occurred by seismic or aseismic displacements, and (2) the fact that the Doruneh Fault has been far from historical trade routes that may caused a non-complete records of large historical earthquakes, which probably occurred on and around the fault. Nevertheless, there are general consensuses that fault zones, particularly long ones, do not rupture along their entire length during a single earthquake (King and Nabelek, 1985; Barka and Kadinsky-Cade, 1988; Knuepfer, 1989; Schwartz and Sibson, 1989; dePolo et al., 1991; McCalpin, 1996; Stewart and Taylor, 1996; Barka et al., 2002; Kondo et al., 2005; Ikeda et al., 2009). Actually, such apparently continuous long fault systems could be divided into discrete fault zones or segments, at a variety of scales, that have different geometries, faulting histories and seismicity (e.g., Schwartz and Coppersmith, 1984; Bruhn et al., 1987; Knuepfer, 1989; dePolo et al., 1991; Stewart and Taylor, 1996). The first stage to understand the seismogenic behavior of large strike-slip fault systems is the identification of distinct fault zones and/or fault segments together with their persistent boundaries, i.e., to establish a fault segmentation model.

34 Chapter II

Figure 1. GTOPO30 topographic image of central and northeastern Iran showing the location of the Doruneh Fault together with principal deformation domains and structural units mentioned in the text. Black arrows are GPS horizontal velocities (mm/yr) in a Eurasia-fixed reference frame (YAZT station, (Masson et al., 2007)). White arrows, south of the Doruneh Fault are geodetic-derived tectonic motions after Reilinger et al. (2006), while the white arrow in the Kopeh Dagh is the geologic rate of the western Kopeh Dagh – Eurasia northward motion (Shabanian et al., 2009a). The box in the upper left inset shows the location in the Arabia–Eurasia collision zone. Grey arrows and associated numbers represent Arabia–Eurasia plate velocities (mm/yr) after Reilinger et al. (2006).

Various types of segmentation models, based on seismological data and surface geological features, have been proposed for seismic hazard assessment within active strike-slip fault systems (e.g., dePolo et al., 1991; Lees and Nicholson, 1993; McCalpin, 1996; Nishigami, 2000; Lettis et al., 2002; Kondo et al., 2005). In ideal fault segmentations, the aim is to identify the part of a fault zone that will rupture as an independent segment and, within the segment, identify the nucleation point of the earthquake. Nevertheless, segments may present

35 Chapter II a long fault with a length of tens to hundreds of kilometers, or they may represent an individual fault rupture only a few kilometers in length. As the term “segment” was not explicitly defined, we use the term segmentation for the division of a large fault system into fault zones with distinct geomorphic, structural, and behavioral characteristics. The two-fold aim of this paper is (1) to provide evidence of geomorphic, structural, and fault behavioral variations (cf., Knuepfer, 1989) using which a fault segmentation model for the DFS, at the scale of fault zones, can be established, and (2) to address the way that active deformation is accommodated along the fault system. These two objectives lead to better understand the seismic potential of different sectors of the DFS, and the contribution of the entire fault system in the accommodation of ongoing deformation in NE Iran. The results of this study may be useful to explain how large strike-slip faults such as the DFS take up overall tectonic block motions perpendicular to their strike.

2. Tectonic and Geological Setting The Iranian plateau is deformed as a collisional domain between the converging Arabian and Eurasian plates. The convergence rate increases eastwards reaching 26±2 mm/yr at a longitude of 59°E, south of the Persian Gulf (Sella et al., 2002; (McClusky et al., 2003; Reilinger et al., 2006; Sella, 2002; Vernant et al., 2004); Reilinger et al., 2006). This plate motion results in right-lateral shear between Central Iran and Eurasia at a rate of ~16 mm/yr (e.g., (Regard et al., 2005; Vernant et al., 2004); Regard et al., 2005) involving major N- trending strike-slip fault systems (Tirrul et al., 1983; Regard et al., 2004; Walker and Jackson, 2004; Meyer and Le Dortz, 2007) (Fig. 1). Further north, in northeast Iran, the available geodetic data (Vernant et al., 2004; Masson et al., 2007; Tavakoli, 2007) indicate a northward displacement rate ranging from 4 to 9 mm/yr for the western Kopeh Dagh with respect to Eurasia. The geological rate of this northward motion is estimated at 8±2 mm/yr (Shabanian et al., 2009a). At a latitude of ~35°N, the area between northern deformation domains (i.e., the Alborz, Kopeh Dagh and Binalud mountain ranges) and the southern N-S fault systems is separated by the “Doruneh Fault” (e.g., Tchalenko et al., 1973(Mohajer-Ashjai, 1975; Tchalenko et al., 1973), which preclude their direct structural connections. The term “Doruneh Fault” has been used for the longest strike-slip fault in the Iranian plateau (Wellman, 1966), which runs east-west over a ~600 km length between longitudes of 54°E and 60°30’E (Fig. 1).

36 Chapter II

Figure 2. (a) Simplified geological map of the region affected by the Doruneh (DFS) and Great Kavir (GKFS) fault systems. Geological units are from Huber (1977).(b) Examples of pull-apart basins formed between releasing left-handed stepovers along the GKFS that cuts through dome and basin structures in the Miocene deposits (Quickbird image - GoogleEarth). (c) Detailed geological map of the Doruneh Fault System (modified after Eftekhar-Nezhad et al. (1976) and Alavi-Naini et al. (1992)) superposed on shaded relief image of the area (SRTM digital topographic data). Fault traces in (a) and (c) are based on geomorphic and structural analyses of SPOT5 and LANDSAT ETM+ (this study). Abbreviations are: BF, Bijvard Fault; DQF, Dahan-Qaleh Fault; KF, Kharturan Fault; KHF, Khaf Fault; JTF, Jangal Thrust Fault.

37 Chapter II

The general trace of the Doruneh Fault, which was mapped on regional scale geological maps (Eftekhar-Nezhad, 1972; Hubber, 1977; Stocklin and Nabavi, 1973), indicates a bow- shape northward convex structure comprising two distinct parts, with different orientations. The preliminary division of the Doruneh Fault into two sectors (Mohajer-Ashjai, 1975; Tchalenko et al., 1973) was based on the change in the orientation of the fault trace from NE to E at the longitude of Doruneh village (Fig. 2a). However, new structural and geomorphic mapping complemented by preexisting data (geological maps and their associated cross sections) indicate that the structural boundary between the western and eastern parts does not closely coincide with the location of Doruneh village. But, further west, at a longitude of 56°45’E, there is evidence of drastic changes in geologic, structural and geomorphic expression of the fault (Fig. 2a), where there is a ~40-km-long structural gap along which Quaternary deposits have not been affected by the fault activity. From there, the western part (i.e., Great Kavir Fault) runs southwest for ~275 km through the Great Kavir desert (Dasht-e Kavir), and left-laterally cuts the dome and basin structures of the folded Tertiary rocks. Along its entire length, the western part passes within the upper Miocene sandstones and marls (Upper Red Formation) without significant change in the topography on either sides of the fault. The surface geometry of the fault is characterized by left-handed arrangements of multiple fault segments. Releasing offsets between stepover fault segments occupied by pull- apart basins of different scales which are filled by Quaternary deposits (Fig. 2b). The eastern part, which we term the Doruneh Fault System (DFS), is a ~360-km-long E- trending fault system separating the pre-Oligocene paleoreliefs to the north from the folded Neogene piedmont armored by Quaternary deposits to the south (Fig. 2c). In other words, the DFS forms a geological and topographic boundary between the northern mountain ranges, with an average elevation of ~2500 m, and the southern Quaternary plain with a mean elevation of ~1200 m. Along the DFS, the moderately folded Miocene rocks, crop out in the southern side, contrast intensely folded and faulted Paleozoic to Eocene sedimentary rocks of the northern side, which have been intruded by Upper Cretaceous to upper Eocene volcanic rocks. The lack of Neogene deposits on the northern part of the DFS suggests that vertical faulting during Oligocene and Miocene along the DFS controlled the northern margin of the sedimentary basin. Such a geological characteristic sets apart the DFS from the Great Kavir Fault, which clearly cuts and laterally offset the post-Miocene geological structures through the Dasht-e Kavir lowlands (e.g., Walker and Jackson, 2004), without any evidence of vertical faulting. That means the Oligo-Miocene Dasht-e Kavir sedimentary basin spreads either side of the fault, and was not controlled by vertical faulting along the Great Kavir Fault. Hereafter,

38 Chapter II we present geomorphic and structural investigations leading us to characterize the pattern and kinematics of active faulting along the DFS, and subsequently, to establish a fault segmentation model, at the scale of fault zones.

3. The Doruneh Fault System (DFS) The DFS is an arcuate structural assemblage, with a complex history of tectonic evolution along which there is clear evidence of at least two distinct periods of faulting during the Cenozoic: (1) an initial vertical faulting that controlled the northern margin of the subsiding Neogene sedimentary basin located on the southern side of the fault, and (2) Quaternary strike-slip movements affecting post-Miocene geological structures and geomorphic landforms. There is a striking contrast between Cenozoic rocks on opposite sides of the DFS. The northern domain is characterized by Paleozoic to Eocene rocks while in the southern part, Neogene and Quaternary deposits are crop out (Fig. 2c). The lack of Miocene sedimentation on the northern side indicates direct structural control of the DFS due to its vertical movements on the geological evolution of the bounding sedimentary basin. The same sedimentary and volcano-clastic rock series of Jurassic-Cretaceous and Eocene on either sides of the fault (e.g., (Hubber, 1977) suggest the post-Eocene onset of vertical movements along the DFS. The total vertical slip of the earlier movements along the fault system is unclear; no rocks of post Eocene series have been matched across the fault system. On the southern side, the Eocene rocks are overlain by more than 2000 m of Neogene-Quaternary deposits (Hubber, 1977); Alavi-Naini et al., 1992) that have no counterpart. On the other hand, the present-day average topographic level of the Eocene paleoreliefs north side of the fault is about 1200 m above the average topographic level of the southern part. If the earlier movement was vertical, a minimum displacement of ~3200 m is necessary to achieve the present day configuration, though the general history of movement on the fault system remains unknown. Nevertheless, conspicuous geomorphic offsets recorded by alluvial fans and streams indicate a prominent left-lateral strike-slip kinematics for the late Quaternary activity of the DFS (section 4.2). Along the DFS and in nearby areas, the historical (Ambraseys and Melville, 1982) and instrumental (e.g., Tchalenko, 1973; Tchalenko et al., 1973; Jackson and Mackenzie, 1984; Fattahi et al., 2007) records of ~700 years seismicity indicate moderate seismicity despite clear late Pleistocene and Holocene activity of the fault system.

39 Chapter II

catalogue catalogue

.

)

1982

(

n (Internatl. Seis. Cent.,

EHB EHB Bulleti

-

Ambraseys and Melville

. Epicenters are from the ISC

)

1984

(

). regionsThe of maximumdestruction are basedon

Jackson and McKenzie

http://www.isc.ac.uk

-

. Historical and instrumental seismicity of the DFS and surrounding areas. Focal mechanisms are mainly taken from the Harvard

3

Figure (http://www.globalcmt.org/CMTsearch.html) and Thatcham, United Kingdom, 2009

40 Chapter II

The earliest recorded historical event is the earthquake of 1336 AD, which destroyed a wide area of 110×40 km², extended between Torbat-e Heidariyeh and Khaf (Fig. 3) causing about thirty thousand deaths (Ambraseys and Melville, 1982). The region of the maximum destruction is located along the NW-trending Jangal thrust fault that runs parallel, and to the west of the eastern termination of the DFS. Another historical earthquake occurred in 1619, and damaged several villages south of Torbat-e Heidariyeh and also totally destroyed the village of Dugh-Abad (Fig. 3). It was suggested that the earthquake was a large magnitude event causing a lot of damage and almost 800 deaths during the day (Ambraseys and Melville, 1982). The two ~45 km-apart damaged zones of the 1903 (Turshiz M=5.9, Ambraseys and Melville, 1982) and 1923 (Kajderakht M=5.5-6, Tchalenko, 1973) earthquakes are located just south of and along the main trace of the DFS, respectively (Fig. 3).

Table 1. Epicenters and source parameters of earthquakes affecting the nearby region of the Doruneh Fault System. Date Longitude Latitude Plane 1 Plane 2 Depth ID Mw (dd.mm.yyyy) (°E) (°N) Azimuth Dip Rake Azimuth Dip Rake (km) 1 26.05.1971 58.135 35.525 5.4 88 26 31 329 76 112 26

2 01.12.1972 57.922 35.428 5.4 156 65 -176 64 87 -25 33

3 09.12.1979 56.827 35.107 5.6 350 44 121 129 53 63 15

4 14.12.1994 58.604 35.094 5.2 319 32 144 80 72 63 33

5 25.02.1996 56.948 35.72 5.4 82 77 10 350 80 166 33

5 02.02.2000 58.207 35.227 5.3 83 43 79 278 48 100 26

6 31.05.2005 57.626 34.306 4.9 309 84 -1 39 89 -174 25 Focal mechanisms of the two first events are from Jackson and McKenzie (1984) and the events 3 to 8 are taken from the Harvard catalogue (http://www.globalcmt.org/CMTsearch.html). Epicenters are from ISC, EHB Bulletin, (http://www.isc.ac.uk).

During the field work, ~30 km west of the damaged area of 1923 earthquake (near Forsheh village) we observed evidence of a coseismic left-lateral surface rupture zone that is composed of overlapping fault segments (Fig. 4). The releasing structural offsets between adjusting segments are occupied by pull-apart basins, while restraining stepovers are the place of en echelon pressure ridges. The rupture zone and its associated geomorphic landforms have been developed in the lower (youngest) alluvial fan surface above the current stream bed (Fig. 4). Clear deflections of streams around the ~2-m-high pressure ridges together with a lack of incision at the top of ridges confirm the coseismic nature of the landforms. Assuming that a single event produced such a rupture zone, a total rupture length of ~11.5 km (Fig. 4) suggests a magnitude of ~6.3 for that seismic event. However, given the relatively large distance (~30 km) between the rupture zone and the location of damage zones of both 1903

41 Chapter II and 1923 earthquakes (Fig. 3), the event or events that produced these features remain unknown. Focal mechanism solutions of instrumental earthquakes (M ≥ 4.5) are shown in Figure 3, and are listed in Table 1. No instrumental, and consequently, earthquake focal mechanism can directly be assigned to the Doruneh Fault. Nearly 60 per cent of the solutions in the vicinity of the fault (a distance of ~30 km from the fault line) show prominently thrust faulting on fault planes in various directions. The focal mechanism of 1971 earthquake indicates left-lateral faulting on a southward gently-dipping (26°S) nodal plane or dextral faulting on a nearly vertical NW-SE nodal plane, both inconsistent with geometry of the Doruneh Fault. Two of other three focal mechanism solutions of 1972, 1994 and 2000 earthquakes (i.e., 1994 and 2000 earthquakes) located near the Doruneh Fault show nearly pure thrusting on E-W fault planes. The explanation for reverse faulting on E-W planes indicated by these focal mechanism solutions remains unclear; Fattahi et al. (2007) attributed the 2000 event to undefined north-dipping east-west thrusts. In section 6.1, we discuss possible relationships between these earthquakes and the Doruneh Fault in the light of geomorphic and structural data. The focal mechanism solution of 1972 earthquake indicates strike-slip faulting along a NE-SW plane. Considering both the location and fault geometry of this solution, it is more likely that the 1972 earthquake to be related to the Bijvard Fault (Fig. 3).

Figure 4. Simplified morphotectonic map of the rupture zone along the CFZ, in the Forsheh area. See text for more information.

The epicenter of 9 December 1979 earthquake is located close to the DFS, ~50 km west of Doruneh village (Fig. 3). Harvard CMT solution of this event represents nearly pure reverse faulting along N-trending nodal planes (Fig. 3). An alternative strike-slip solution has been

42 Chapter II proposed by Jackson and McKenzie (1984) showing left-lateral faulting on an ENE-trending nodal plane. This solution has been preferred by Jackson and McKenzie (1984) because it was more consistent with the general orientation of the western part of the DFS. However, a detailed mapping of the western termination of the DFS (section 0) led to evaluate the geological reliability of these two solutions indicating that the reverse mechanism (Harvard CMT solution) is more consistent with both the structural pattern and geometry of the fault system. In summary, both the instrumental and historical records of earthquakes indicate that the region affected by the DFS experienced a moderate seismicity when compared to nearby seismic regions such as the Kopeh Dagh (Tchalenko, 1975; Berberian and Yeats, 1999, 2001; Hollingsworth et al., 2006; Shabanian et al., 2009b) and northern margin of the Lut block (Berberian and Yeats, 1999 and references therein; Walker et al., 2003, 2004).

4. Fault segmentation At the regional scale, the DFS seems a continuous curved structure that runs northeast from a longitude of 56°45’E for about 125 km to ~12 km to the northeast of Bardeskan town (Fig. 5), where it turns ~15° eastward and continues for about 75 km. The fault trace takes another bend (~20°clockwise) at a longitude of 59°15’E and continues for ~160 km toward the southeast. Given that such a large scale change in the surface geometry of a fault may result in variations in both the slip rate and kinematics of the fault, each portion of the ~360- km-long DFS can represent a distinct fault zone with the characteristic geometry, structure and fault behavior. In this section we present structural and geomorphic analyses conducted along the DFS in order to establish, at the scale of fault zones, a fault segmentation model. In this way, we take heed of (1) the geometry and kinematics of discrete fault zones, (2) spatial variations in the amount and nature of Quaternary geomorphic offsets (3) the size and geometry of structural discontinuities along the fault system, and (4) changes in complexity of fault trace, in the sense of fault slip, and in the die-out of fault trace beyond which no active faulting is mapped (cf., Knuepfer, 1989).

4.1. Variations in the structural and geometric characteristics of the DFS According to our detailed structural mapping of the entire fault system, the DFS is constituted by multiple individual fault segments. A statistical analysis of the orientation of fault traces longer than 5 km clearly illustrates three structural fault portions with different

43 Chapter II

Figure 5. (a) General fault map of the DFS comprising three discrete fault zones. Rose diagrams represent the predominant orientation of the fault zones deduced from the statistic analysis of individual fault segments, which have been mapped on SPOT5 satellite images. (b) along-strike variations in the wideness of the DFS. The initiation point of the diagram is the western termination of the WFZ. (c) Spatial variations in the geometric arrangement of fault segments within the DFS are illustrated by plotting the straight length of individual fault segments versus the total overlap length of each segment that is covered by other nearby faults. Thick lines indicate the average value of the segment overlap in each fault zone. (d) along-strike variations in cumulative left-lateral offsets of the DFS. Uncertainty of the offset measurement is shown by error bars.

44 Chapter II orientations (Fig. 5a): (1) the western part between 56°45’ and 57°37’ that is principally oriented in the N75±5°E direction, (2) the central part that continues eastward (N95±5°) up to a longitude of 59°15’, and (3) the N130±10°E-trending eastern part. These variations in the fault strike are followed by changes in the width of the fault system (Fig. 5b) such that the DFS becomes wider at both the eastern and western terminations. The western part reaches ~2.5 km in width joining the relatively narrow ≤0.4-km-wide linear central part (Fig. 5b). Interestingly, the maximum width of the western part is observed at its structural terminations where the fault movements is transmitted to transpressional relay zones forming complex reverse pop-up structures. The linear pattern of the central part contrasts with the ~14-km- wide eastern part, which is composed of numerous imbricate reverse fault segments (Fig. 5b). The presence of three distinct fault zones is also confirmed by examining changes in the geometric arrangement of fault segments constituting each fault zone. For instance, along the western part, fault segments show mostly a left-stepping arrangement, whereas to the east a parallel fault arrangement is dominated. The central part of the DFS is more complex than the two other parts such that both parallel and en-echelon arrangements are found. Such a variation was illustrated by plotting the straight length of individual fault segments versus the total overlap length of each segment that is covered by other nearby faults (Fig. 5c). Within the western part, almost ~70 per cent of the total length of each fault segment is overlapped by one or more adjacent segments. In the eastern part, fault segments are nearly parallel, and each segment is totally overlapped by one or more nearby fault segments (Fig. 5c). In other words, the overlapped length of nearly all segments reaches 100 per cent of the total segment length. Along the central part, overlapping percentage decreases to a mean value of ~40 per cent (Fig. 5c), which means that the involved fault segments are structurally more independent with respect to the fault segments of two other parts. In summary, the analysis of along-strike variations in the orientation, wideness and geometric arrangement of the DFS leads us to divide the entire fault system into three distinct Western (WFZ), Central (CFZ), and Eastern (EFZ) fault zones (Fig. 5a).

4.1.1. The Western Fault Zone (WFZ) The 85-km-long WFZ comprises nine main strands forming a fault zone of several hundred meters to 2.5 km in width. Left-lateral component of faulting along its left-stepping arranged fault segments forms pull-apart troughs within releasing segment offsets (Fig. 6). The western and eastern terminations of the WFZ are marked by two fault zones intersecting the DFS. To the west, the WFZ is intersected at a right angle by a ~N160°E-trending reverse

45 Chapter II

Figure 6. SPOT5 image centered on two pull-apart basins (hatched areas) formed in releasing offsets between overlapping segments of the WFZ; see Figure 7 for the location.

fault zone (Kharturan) beyond which, except for diffused minor faulting, the surface expression of the DFS is died out (Fig. 7). The Kharturan fault zone forms the western boundary of uplifted areas on the northern side of the DFS, where Neogene and Quaternary deposits have been tilted and uplifted against the western playa (i.e., Dasht-e Kavir). On the western side of the Kharturan fault zone, the Dasht-e Kavir (Great Kavir) depression reaches its lowest (750 m) elevation (Fig. 7). These structural and geomorphic relationships suggest that the relative westward (left-lateral) motion of the northern block of the DFS should be principally taken up by reverse faulting along the Kharturan fault zone. Such a mechanism can explain the topographic contrast between the western and eastern sides of the reverse Kharturan fault zone, the die-out of the DFS surface expression, as well as the location of the deepest part of the Dasht-e Kavir playa. If is accepted, the focal mechanism of the 09 December 1979 earthquake, which indicate reverse faulting with little right-lateral component on a NNW-trending nodal plane (section Erreur ! Source du renvoi introuvable.), can be representative of a seismic event on the reverse Kharturan fault zone at the western termination of the DFS (Fig. 3). To the east, the western fault zone is intersected by a ~N50°E-trending fault zone (Dahan-Qaleh) which runs NE over a length of ~50 km (Figs 7 and 8). Cumulative offsets recorded by both well-

46 Chapter II stratified geologic units and geomorphic features (Fig. 8) indicate the prominent left-lateral strike-slip character of the fault zone. Eastward from the intersection area, the WFZ joins the central fault zone through a drastic change in the geomorphic expression of the DFS (section 4.2).

Figure 7. (a) Shaded relief map (SRTM data) of the WFZ. The trace of the WFZ joins at right angle the Kharturan reverse fault zone (KF), which marks a sharp topographic edge at the NE boundary of the Dasht-e Kavir (Great Kavir desert) absorbing the relative westward motion (left-lateral faulting) of the WFZ’s northern block. “P” in Figure 7a indicates a pop-up structure formed in a restraining bend along the WFZ. (b) A profile of the fault shown in Figure 7d indicating the reverse component of faulting along the WFZ. (c) Field photograph of a seasonal stream incised in Q1 alluvial fan surface obliquely displaced by the fault indicating coeval vertical and left-lateral offsets of ~6.5 m and ~5.5 m. (d) Field photograph of the fault scarp along the WFZ. The accurate location of field photographs is: Figure 7b, 35.2205°N - 57.5577°E; Figure 7c, 35.22106°N - 57.5603°E; Figure 7d, 35.2210°N - 57.5606°E.

47 Chapter II

Figure 8. Geomorphic and geological evidence of left-lateral faulting along the Dahan-Qaleh Fault (Quickbird image - GoogleEarth). Upper image illustrates cumulative left-lateral offset along the fault running through folded rock units and Quaternary landforms. Lower image shows left-lateral displacement of the fault in the folded Neogene deposits. Trace of the fault is shown by dotted black lines. White lines marks offset markers. See Figure 7 for the location.

4.1.2. The Central Fault Zone (CFZ) The CFZ runs about 140 km eastward between the longitudes of 57°37’ and 59°15’. The CFZ comprises nine fault strands with lengths ranging from ~5 to 45 km that form a relatively narrow zone of several meters to 400 m in width (Fig. 9). Along the CFZ, the frequent structural features are either push-up structures, which were formed within restraining relay zones and/or fault bends, or left-handed pressure ridges formed within stepovers (e.g., Fattahi et al., 2007) or between en echelon fault segments (Figs 9 and 10).

48 Chapter II

Figure 9. Simplified fault map of the CFZ and the surrounding areas. (a) and (b) are shaded relief maps based on SRTM digital topographic data indicating two examples of uplifted areas along the CFZ; the uplifted and subsided areas are marked by (+) and (-), respectively.

The eastern termination of the CFZ, near the village of Kajderakht, is characterized by a complex structural zone (Fig. 10), where, the main fault splits out into the northern and southern strands. The southern strand ends into a ~20-km-long restraining bend comprising SSE-striking reverse faults along which Neogene deposits on the western fault blocks have been thrust over the Quaternary plain on the east side of the faults (Fig. 10). In the same area, just south of Torbat-e Heidariyeh, the northernmost strands of the CFZ are intersected by two NNE-trending faults that border a graben-like morphology (Fig. 10). There is a striking contrast between geologic outcrops in the inside and outside of the graben (regardless its structural genesis). In the outer side, are Eocene andesitic rocks while in the interior, are middle Oligocene - late Miocene lagoonal to continental sedimentary rocks that have been covered by Quaternary deposits. The inner area is clearly less elevated (~1300 m) than the average elevation (~1650 m) of the uniformly elevated mountains on the other sides of the graben-bounding faults (Fig. 10). These geologic and geomorphic features indicate that the graben-bounding faults are two long-lived faults (at least since Early Oligocene), and that together with the southern restraining bend they represent an important boundary condition between the central and eastern fault zones.

49 Chapter II

Figure 10. Shaded relief map (SRTM digital topographic data) and simplified morphotectonic interpretation (below) of the eastern termination of the CFZ. The lower topographic profile (white line in upper panel) represents abrupt topographic changes across the fault-bounded graben (see section 4.1 for explanations) at the CFZ’s eastern termination.

4.1.3. The Eastern Fault Zone (EFZ) The ~140-km-long EFZ forms a horsetail structure or nearly more, a trailing contractional imbricate fan (e.g., Woodcock and Fischer, 1986) at the eastern termination of the DFS (Fig. 11a). The fault zone varies between 1.5 and 14 km in width, and comprises at least fifteen NE-dipping reverse faults along which Neogene to Quaternary deposits are thrusted over the Holocene plain (Fig. 11b). The imbricate reverse faults diverge toward the southeast,

50 Chapter II and some of them bound the southwest flanks of NW-trending anticlines formed on the hanging wall of the faults (Fig. 11c).

4.2. Variations in the geomorphology and fault behavior of the DFS Given that the rate and kinematics of slip on a Quaternary fault is expressed in the geomorphology of the fault, along-strike variations in the fault behavior could be examined thanks to systematic changes expressed by geomorphic features affected by the fault. The surface trace of the DFS is marked by geomorphic features such as Quaternary alluvial fans and their associated streams that have been laterally and/or vertically offset along the whole length of the fault. Detailed geomorphic mapping based on SPOT5 satellite images complemented by field observations allowed us to subdivide multiple generations of alluvial fans, which are abandoned and entrenched, into Q3, Q2 and Q1 geomorphic surfaces, from upper to lower. Each geomorphic surface exhibits specific geomorphology such as geomorphic terracing (stepping), incision pattern and the state of preservation. Naturally, the higher surfaces are more incised and consequently, less preserved relative to the lower fan surfaces.The type (vertical, horizontal or oblique) and amount of fault offsets recorded by alluvial fans vary along the strike of the fault system such that alluvial fans of the same generation, but in different structural setting, exhibit various type of fault offsets ranging from pure vertical (reverse) to pure horizontal (left-lateral). For instance, the cumulative geomorphic offsets observed along the western and central fault zones imply the respective oblique-slip (reverse left-lateral) and strike-slip (left-lateral) characters of faulting; whereas within the eastern fault zone offsets are principally vertical. In order to illustrate spatial variations in the amount of lateral offsets recorded by Q1, Q2, and Q3 fan surfaces, cumulative offsets measured along the DFS were plotted versus along- strike distance from the western termination of the fault system (Fig. 5d). The amount of cumulative left-lateral offsets systematically increases from the western to the central fault zones. For each fan generation, the largest left-lateral offsets have been recorded by Q3 (840±70 m), Q2 (400±10 m) and Q1 (95±20 m) fan surfaces along the CFZ. But, along the WFZ, cumulative left-lateral offsets recorded by the same fan generations decrease in 240±40 and 145±20 m for Q3 and Q2 surfaces, respectively (Fig. 12 and table 2). The only systematic left-lateral offset in Q1 fan surfaces are offset of ~5.5 m recorded by streams incising Q1 fan surfaces. Such an offset value should be considered as a minimum offset along the WFZ (Fig. 7c).

51 Chapter II

Figure 11. (a) Fault map of the EFZ. Quaternary fault traces are shown by thick black lines. Faults along which there is no evidence of Quaternary activity are marked by thin black lines. (b) 3D view (Quickbird image superposed on SRTM data - GoogleEarth) of a Q3 alluvial fan that has vertically been offset on a fault within the EFZ. A topographic profile (A-A’ in Figure 11b), based on SRTM data, that illustrates vertical faulting along the westernmost fault of the EFZ. (c) A growing anticline (3D view - GoogleEarth) on the hanging wall of a reverse fault of the CFZ. Quaternary deposits were titled and uplifted on the SW flank of the fold. The trace of the fault is indicated by triangles.

The maximum left-lateral offset along the CFZ is recorded by the Q3 Quch Palang alluvial fan (Wellman, 1966; Giessner et al., 1984). In that area, a ~800 m left-lateral offset was measured by Fattahi et al. (2007). Subsequently, the offset value has been reduced to ~150 m by Walker and Fattahi (2011) favoring a much smaller displacement which satisfies the IRSL ~50 ka ages (Fattahi et al., 2007) estimated for the uppermost part of fan deposits. This offset reduction is based on a suggestion that assign deposits from a Q2 and another Q3 fan on the south and north sides of the fault, respectively, to remnants of the Q3 Quch Palang fan surface

52 Chapter II

(see Walker and Fattahi, 2011). But, our detailed geomorphic mapping based on the high- resolution Quickbird and SPOT5 imageries complemented by field survey allowed us to define the border of the alluvial fan (Fig. 13).

Table 2. Geomorphic offset measurements along the DFS. Site Name Fault zone Latitude Longitude Quaternary measured surface

(°E) (°N) Q3 Q2 Q1

1 WFZ 35.0389 56.7664 - 20±10 -

2 WFZ 35.0762 56.8763 - 60±10 -

3 WFZ 35.2015 57.4409 - 100±10 -

4 WFZ 35.2076 57.4835 - 145±20 -

5 WFZ 35.2173 57.5456 240±40 - -

6 WFZ 35.2232 57.5713 - 130±20 -

7 WFZ 35.2255 57.5833 - 130±20 -

8 CFZ 35.2976 58.2422 - 400±10 -

9 CFZ 35.2933 58.2878 630±70 - -

10 CFZ 35.2962 58.3432 - 260±30 50±10

11 CFZ 35.2825 58.5676 840±70 - -

12 CFZ 35.277 58.6623 - 300±50 100±15

13 CFZ 35.2622 58.855 - 150±30 -

14 CFZ 35.2542 58.9131 350±30 - -

15 CFZ 35.241 59.009 - - 90±10

16 CFZ 35.235 59.049 - 190±30 -

17 CFZ 35.221 59.12 330±20 - -

18 EFZ 35.175 59.3226 - 60±10 -

Site name refers to the numbers marked in Figure 12a. The most important offsets playing a crucial role in offset variations along the DFS are presented in the Figures 12b-i

The geomorphic reconstruction of both the overall shape of the fan, and main streams incised in the fan surface reveals a left-lateral offset of 800±50 m. The same offset value (880±50 m) has been obtained through the geometric reconstruction of the axial trace of the fan surface (e.g., Keller et al., 2000; Shabanian et al., 2009a). This geometric method uses topographic contours of a fan surface to define the geometric axis of the alluvial fan, and avoids producing other possibilities due to interactions between the alluvial fan and other erosional/depositional adjacent features (e.g., Walker and Fattahi, 2011). Interestingly, a topographic profile across the distal part of the Quch Palang alluvial fan clearly reveals the presence of two individual landforms indicating that the lower left part surface (B-B’ in Fig. 13) belong to another alluvial fan (Q2 surface).

53 Chapter II

Figure 12. (a) General fault map of the DFS showing the location of sites in which left-lateral offset recorded by Q1, Q2, and Q3 alluvial fans have been measured. The most important offset measurements, which have a crucial role in the understanding of offset variations along the DFS, are shown in Figure 12b to Figure 12i. (b) A Q2 fan apex left-laterally offset along the WFZ. (c) The eastern boundary of a Q2 alluvial fan offset along the WFZ. (d) the Q2/Q3 terrace riser left-laterally offset along the CFZ representing the cumulative offset of the Q2 fan surface. (e) An eroded Q3 alluvial surface left-laterally offset along the CFZ. The initial shape of the alluvial fan cannot be reconstructed. A 630±70 m cumulative offset (sum of two offset along the northern and southern fault strands) recorded by streams incising in the fan surface have been measured as a minimum left-lateral offset at site 9. Note that each beheaded stream in the northern side has been compared with the first stream on the left in the southern side.

54 Chapter II

Figure 12. (continued): (f) Streams incised in a Q2 fan surface have recorded a minimum cumulative offset of the surface. The Q1/Q2 terrace riser reveals the cumulative left-lateral offset recorded by the Q1 fan surface. The middle part of the fault strands is a pressure ridge. (g) streams incised in Q2 alluvial surfaces recorded left-lateral offset of 150±30 m along the CFZ. (h) Q3 alluvial surfaces offset along the easternmost of the CFZ. The same left-lateral offset is shown by drainages on the left of the Q3 surface. The (1’) stream segment is abandoned due to the fault offset. (i) Terrace riser between Q1 and Q2 alluvial fan surfaces has been offset along three strands of the CFZ indicating the offset recorded by the Q1 fan surface.

In summary, both the geomorphic and geometric approaches yield a constrained offset of 880±70 m that is representative of the maximum cumulative left-lateral displacement along the DFS since the abandonment of the Q3 Quch Palang alluvial fan. Along the CFZ, the largest left-lateral offset (95±20 m) recorded by Q1 alluvial fan surfaces is observed in Zarmehr area (Site 15 in Fig. 12). Previously, a smaller offset of ~25 m recorded by the Q1 Shesh Taraz alluvial fan was reported by Fattahi et al. (2007).

55 Chapter II

The ~25 m offset was suggested to be representative of cumulative offset in Q1 alluvial surfaces along the CFZ (Fattahi et al., 2007; Walker and Fattahi, 2011), and has been used to estimate the slip rate of the Doruneh Fault.

Figure 13. (a) SPOT5 image and its morphotectonic interpretation centered on the Quch Palang area which shows the geomorphic setting of Q3 and Q1 alluvial fans along the CFZ. A – A’ topographic profile on a Q1 fan surface shows that there is no vertical displacement across the fault cutting the Q1 fan surface. B – B’ topographic profile across the Q3 Quch Palang alluvial fan indicates the topographic difference between the Q3 and Q2 geomorphic surfaces. (b) The geomorphic reconstruction of both the overall shape and main streams of the Quch Palang alluvial fan. (c) The geometric and topographic reconstruction of the Quch Palang fan. The location of apex defined using concentric arcs on the distal part of the fan is located ~880 m to the east of the initial apex. This offset value is the same than the offset recorded by main streams incising in the fan surface.

Actually, the ~25 m offset of T1-T2 riser indicated in Figure 8 of Fattahi et al. (2007) is clearly an apparent offset resulted from lateral erosion of the riser near the fault trace, and within the eastward continuation of the pull-apart basin (Fig. 14). Just ~50 m to the south of the southern fault trace, the T1-T2 riser return back to its initial position such that no left-

56 Chapter II lateral offset in this riser is observed (Fig. 14). Immediately east of the T1-T2 riser there is an older linear riser that does not show any left-lateral offset (Fig. 14). Given such relationships, the ~25 m offset deduced from the DGPS-derived topographic map by Fattahi et al. (2007) is an apparent offset that cannot be representative of either cumulative or minimum left-lateral offset of the Shesh Taraz Q1 fan surface. For the eastern fault zone, cumulative displacements are principally vertical, and no conspicuous evidence of lateral faulting is observed. The cumulative vertical displacements are expressed in uplifted and/or tilted Quaternary surfaces on the hanging wall of reverse faults or on the fault-bounded flanks of probably fault-related anticlines (Fig. 11b and c). The lack of left-lateral displacement in the eastern part of the DFS, thus, suggests that strike-slip faulting along the CFZ is principally taken up by reverse faulting and the associated folding along the EFZ within a trailing contractional imbricate fan structure (Fig. 11). Assuming that alluvial fan surfaces belong to each generation of Q1, Q2, and Q3 geomorphic surfaces were abandoned at the same time; spatial variations in the type and amount of cumulative offsets recorded by the fan surfaces represent along-strike changes in the kinematics and slip rate of the DFS. This assumption was examined at one site along the eastern part of the WFZ (Fig. 7). In that area, multiple generations of alluvial fans and streams have been offset by the fault providing evidence of both reverse and left-lateral fault displacements. A Q1 fan surface has recorded a minimum vertical offset of ~6.5 m that is coeval with a left-lateral offset of~5.5 m (Fig. 7c). Our measurement on the fault plane, which is crop out in the left bank of the axial stream of the fan (Fig. 7b), indicates geometry of N73°E/65°NW, and allows determining a geomorphic-derived pitch angle of ~50°NE for the main strand of the WFZ. This pitch angle is clearly higher than the average pitch angle (15±5 degrees) we measured along the CFZ (section5.2). Such a direct observation explain why the cumulative horizontal offsets recorded by Q3, Q2, and Q1 fan surfaces affected by the western fault zone are significantly less than offsets recorded by the equivalent fan surfaces along the central fault zone. On the other hand, we showed that (Fig. 5d) systematic variations in cumulative fault displacements correspond with changes in the geometry and structure of the fault system. Altogether, the consistency between long-lived geomorphic, structural and behavioral fault characteristics indicate the persistence of the fault zone boundaries using which the DFS is divided into three distinct western, central and eastern fault zones.

57 Chapter II

Figure 14. The ~25 m cumulative offset in Q1 alluvial surfaces along the Sheh Taraz river west of Kashmar reported by Fattahi et al. (2007). See text for more information.

5. The present-day kinematics and seismic potential of the Central Fault Zone The CFZ accommodates the E-W shear between the north and south tectonic domains of the DFS by pure left-lateral strike-slip faulting. The region affected by the CFZ is the place of populated large towns such as Kashmar, Bardeskan and Torbat-e Heidariyeh. We conducted

58 Chapter II more detailed geomorphic and structural investigations along the CFZ to describe the kinematics of active faulting and to reevaluate the seismic potential of the fault zone.

5.1. The CFZ; an oblique-slip or strike-slip fault zone? The present-day kinematics of the DFS has been under debate because of ambiguous geomorphic features such as Quaternary fault escarpments, with height of several tens of meters, especially along the central part of the fault system (e.g., Tchalenko et al., 1973; Mohajer-Ashjai, 1975; Fattahi et al., 2007). On the other hand, a lack of focal mechanism of earthquake that can be directly assigned to the DFS (see section 3) precludes previous workers to describe the active kinematics of the fault. In the pioneer works by Tchalenko et al. (1973) and Mohajer-Ashjai (1975), the uplift domains in either sides of the fault were interpreted as the geomorphic expression of vertical displacements, whatever their sense, on south-dipping steep fault planes. Nevertheless, the distribution of these areas on both side of the fault precluded the previous workers to conclude about the normal or reverse character of faulting. For instance, based on the basin-ward emplacement of younger alluvial fans with respect to older fan surfaces (between Kashmar and Nay), Tchalenko et al. (1973) proposed episodic vertical (normal) movements elevating the northern block (mountain side) relative to the southern side (alluvial basin), and consequently, to push younger alluvial fans being formed downhill from the fault zone. On the other hand, the presence of an elevated ridge south of the fault zone between Nay and Azghand led Mohajer-Ashjai (1975) to propose a reverse fault kinematics associated with a minor component of left-lateral faulting. Our detailed mapping of Quaternary alluvial fans reveals that the distribution of alluvial fans does not follow the pattern suggested by Tchalenko et al. (1973). In the Quch Palang area, for example, the lowest fan surface (Q1), above the active stream bed, was formed south of the fault zone and has not been affected by the fault activity. Further west (~2 km), another Q1 fan spread on the both sides of the fault zone and is left-laterally offset along two fault strands. In that area, there is no evidence of vertical fault displacements on the fan surface (Fig. 13). The same conclusion has been reached by Fattahi et al. (2007) reporting little (~1.5- 2 m) vertical displacement recorded by a Q1 alluvial fan at the outlet of the Kashmar River. Indeed, it is difficult to determine the sense of vertical movements only based on the arrangement of locally uplifted domains which are randomly distributed along a fault trace, especially when the dip direction of the fault is unclear.

59 Chapter II

Figure 15. The central part of the N-dipping oblique slip fault at the Siah-Kuh Mountain front in the north of the CFZ. The fault offsets (reverse left-lateral) both Quaternary deposits and south running streams. See Fig. 12a for the location.

Other confusing geomorphic features interpreted as vertical offsets along the DFS are topographic steps in Q2 and Q3 alluvial fan surfaces (Tchalenko et al., 1973; Fattahi et al., 2007) which are frequently observed on the northern side of the fault (Fig. 9a and b). These features, in most cases, are apparent fault escarpments when the fault transects rugged topography produced by lateral stepping alluvial fans that are alternated by erosional domains of large streams. In such morphology, the displaced part of the alluvial fans located on the southern side of the fault was missed due to ongoing erosion and/or sedimentation such that the counterpart of the fan appears as a linear escarpment. This suggestion is testified by the fact that along the same fault trace there are other fan surfaces which have left-laterally been displaced keeping their counterparts, whereas, there is no evidence of significant vertical displacements across the fault trace (Fig. 13a). In some cases, interaction between adjacent faults also leads to relatively large uplift domains. For instance, between Kashmar and Ali Abad Keshmar village, the central fault zone is intersected by a NNE-dipping reverse fault that cuts Quaternary deposits at the Siah-Kuh Mountain front (Fig. 15). The combination of left-lateral faulting along the DFS and oblique-slip reverse faulting along the NNE-dipping

60 Chapter II reverse fault forms a narrow contractional wedge (Fig. 9a) within which alluvial fan surfaces have been uplifted and intensely eroded. In summary, such an irregular uplift pattern indicates the strike-slip characteristics of the fault zone along which the position and orientation of local domains of extension and compression, and related geomorphic features depends on the bending and stepping geometry of the strike-slip fault zone, as well as the degree of transpression or transtension (e.g., Sylvester, 1988). Interestingly, our fault kinematic measurement (next section) along the central part of the DFS represents steep fault planes with low pitch angles ranging from 10 to 20 degrees, indicating the pure strike-slip character of the CFZ (Fig. 16c). Nevertheless, structural complexities such as releasing and restraining stepovers and/or bends cause different styles of deformation producing a large variety of shortening and stretching domain.

5.2. Modern stress state along the CFZ deduced from fault kinematic analyses Fault slip data consisting of fault planes and associated striations were measured in five localities distributed along the CFZ (Fig. 16). The inversion of fault kinematic data has been performed using the method originally proposed by Carey (1979). This fault kinematics inversion method computes a mean best fitting deviatoric stress tensor from a set of striated faults by minimizing the angular deviation (misfit angle) between a predicted slip vector and the observed striation (Carey and Brunier, 1974; Carey, 1979). In all fault kinematic inversion schemes, the main assumption is that slip direction indicated by the striation represents the direction of the shear stress resolved on the fault plane. The inversion results include the orientation (trend and plunge) of the principal stress axes (σ1>σ2>σ3, corresponding to maximum, intermediate and minimum stress axes, respectively) of a mean deviatoric stress tensor as well as a stress shape parameter R= (σ2−σ1)/ (σ3−σ1) that describe relative stress magnitudes ranging from 0 to 1 (e.g., Carey and Brunier, 1974; Bellier and Zoback, 1995; Shabanian et al., 2010 and references therein). At two of five sites, fault slip data were measured in Quaternary alluvial fan deposits, while at the three other sites measurements were made in Eocene and Neogene rocks (Fig. 16a and b, Table 3). Fault slip data comprising fault slip families belonging to two different tectonic regimes was manually separated into appropriate data sets. The data separation was done on the basis of geological field data using relative chronology of the striations (crosscutting relationships).

61 Chapter II

Table 3. Result of stress tensor inversion for slip data representing late Cenozoic faulting stress regimes.

Longitude Latitude Paleostress Modern stress Site (°E) (°N) Stress axis (trend/plunge) Stress axis (trend/plunge) Formation age R Rm R Rm

σ1 σ2 σ3 σ1 σ2 σ3 *1 58.466 35.287 142/00 052/00 322/90 0.57 --- 038/06 136/56 304/33 0.95 SS Eocene

2 58.591 35.280 129/13 037/08 278/75 0.78 C 038/02 308/19 135/71 0.86 C Neogene

3 58.773 35.270 351/09 206/79 082/06 0.61 SS 067/22 198/58 328/21 0.43 SS Quaternary

4 58.849 35.264 ------060/03 157/65 328/25 0.81 SS Quaternary

5 58.335 35.294 ------031/02 131/78 300/12 0.23 SS Quaternary

* For the fault data populations comprised of less than four well-distributed fault directions, a “fixed” solution (Bellier and Zoback, 1995) was applied, in which the principal stress axes are fixed to lie in horizontal and vertical planes. Such deviatoric stress tensors are only used for deducing the direction of principal stress axes, not used for R value interpretation (stress regime). See Figure 9 for the location of sites. Ages are reported from geological maps cited in the text; R, stress shape parameter; Rm, stress regime: C, compressional; SS, strike-slip.

The modern stress state was determined using the youngest striae sets measured in both pre-Quaternary rock units and Pleistocene conglomerates (Fig. 16a and Table 3). For all data sets, the calculated modern stress tensors are coherent regardless ages of the rocks in which fault slip data were measured. The modern stress state is characterized by a N45±15°E- trending σ1 and shows a predominated strike-slip tectonic regime, except for one site with compressional tectonic regime (site 2, Fig. 16a). The inversion analysis is individually performed for the oldest striae sets measured in pre- Quaternary rock units (Fig. 16b and Table 3). The calculated paleostress tensors represent a N150±20°E-trending σ1, with a compressional tectonic regime. Applying the average N45°E- trending horizontal σ1 (modern stress) to average geometry (N95±5°E) of the CFZ implies dominant left-lateral strike-slip faulting, which is consistence with late Quaternary geomorphic expression of the fault zone. Both the modern and paleostress states we computed are compatible with corresponding stresses axes in NE Iran, which were deduced from the inversion of both geologically and seismically (in the case of modern stress state) determined fault slip vectors (Shabanian et al., 2010; Javidfakhr et al., 2011).

62 Chapter II

Figure 16. Results of the inversion analysis of fault kinematic data measured along the CFZ. Azimuths of σ1 (maximum stress) axis for modern stress (a) and paleostress (b) states are presented on the upper part of each figure. Numbers refer to site names detailed in Table 3. The lower hemisphere stereograms of fault planes and associated slip vectors (arrows on fault planes) together with deduced stress directions (large arrows) are shown on the lower parts. Histograms show distribution of deviation angles between the measured and calculated slip vectors (e.g., Bellier and Zoback, 1995). (c) Distribution of pitch angles measured on main fault planes within and parallel to the CFZ; about 80 per cent of pitch angles are lower than 20 degrees.

63 Chapter II

5.3. Seismic potential of the Central fault zone The CFZ is presumably a hazardous fault zone that is favorably oriented for movement in the present-day stress field, and could produce damaging large (M≥7) earthquakes (e.g., Fattahi et al., 2007). Generally, the assessment of earthquake faulting and the associated seismic hazard relies heavily on the assumption that future large earthquakes will occur in the same regions as historical events (Coppersmith, 1988; Adams et al., 1995). Thus, the temporal and spatial distribution of past earthquakes in addition to the knowledge of active faulting have been the principal sources of information used to forecast where large events will likely occur in the future. In the region affected by the CFZ, however, a lack of adequate historical and instrumental records of seismicity has been a fundamental obstacle in the way of both the deterministic and probabilistic approaches of seismic hazard assessments. In spite of this difficulty, the combination of geomorphic, kinematic, and structural characteristics of the CFZ allowed us to evaluate the seismic potential of the fault zone. On one hand, we used the Wells and Coppersmith (1994) relation between surface rupture length (SRL for strike-slip faults) and the moment magnitude (M). For this purpose, SRL corresponds to between 50 and 100 per cent of the maximum surface length of the fault (see Berberian and Yeats, 1999), which can be reactivated during an earthquake. But, the structural characteristics of the CFZ propose two possible maximum surface lengths. If we assume that the whole length of the CFZ would reactivate as a single seismogenic fault segment, the total length of 140 km of the fault zone yields a maximum magnitude of 7.2 – 7.6. While, if each main fault strand of the CFZ represents an independent seismogenic segment, a length of ~48 km for the longest fault strand consist with earthquakes of M≤7. We also analyzed possible coseismic displacements that are found along the fault zone. Around the village of Forsheh (Fig. 17), two kariz lines (qanat; traditional Iranian system of irrigation composed of underground canals drawing water from mountain sources by gravity) provide evidence of coseismic fault offsets. East of the Forsheh village, two ~1-km-apart kariz lines show left-lateral offsets of ~4 and ~8 m (Fig. 17). The largest offset (~8 m) corresponds to the kariz line that is suggested to be relatively older given the repaired form of the Karizes (Fig. 17b). Further west, between Shesh Taraz and Kashmar, Fattahi et al., (2007) found a series of stream offset along the CFZ representing nearly the same offset values averaged at ~4.7 m.

64 Chapter II

Figure 17. (a) SPOT5 image of a Q1 alluvial fan (east of Forsheh village) affected by the CFZ. Two kariz lines were left-laterally offset along a coseismic rupture zone. (b) The relatively older kariz line is offset of 8±1 m, where a pull-apart through was formed due to left-lateral faulting of the fault. (c) The younger kariz shows an offset of 4±0.5 m. In the same area, along another fault trace, a stream was offset of ~9 m that is consistent with the offset shown in (Figure 17c) observed along the northern fault segment.

65 Chapter II

The consistency of these offset values observed at different places along the CFZ may implies a possible characteristic offset (cf., Schwartz and Coppersmith, 1984) of ~4-5 m for the central part of the DFS. If is accepted, the cumulative offset of 8 m could be achieved by successive (recurring) occurrences of two earthquakes of 7.1≤M≤7.4 (Wells and Coppersmith, 1994 relations), which may indicate the reactivation of an independent ~50-km-long seismogenic fault segment up to a ~80-km-long fault zone. The ~8 m cumulative left-lateral offset along the CFZ may lead also to estimate a possible recurrence interval dividing the time over which the events occurred by the number of earthquakes during that time interval (e.g., Wallace, 1970; Budding et al., 1991). Unfortunately, it is not possible at present to establish the age of the displaced kariz lines. The oldest known kariz in Iran was constructed in the second millennium B.C. at the northern edge of the Dasht-e Kavir (Wulff, 1968; Berberian and Yeats, 1999 and references therein). Assuming that the oldest kariz line offset by the CFZ is as old as 4000 yr, the successive occurrence of two seismic faulting events yields a possible maximum recurrence interval of ~2000 years for large earthquakes of M ≤ 7.4. Our presumed result seems coherent with that is suggested by Fattahi et al. (2007), and may help to characterize the seismogenic behavior of the CFZ. However, in the absence of historical records of large earthquakes along the CFZ we cannot rule out the possibility of creep faulting. Detailed paleoseismological investigations are necessary to differentiate the contribution of seismic and possible aseismic faulting along the DFS. Nevertheless, according to our data and deduced results indicating that the three western, central, and eastern fault zones of the DFS have had distinct evolution history during the Middle to Late Pleistocene, it is unlikely that the “Doruneh Fault“ could rupture along the entire ~600 km length to produce earthquakes of M ≥ 8.

6. Discussion 6.1. Kinematic models of the Doruneh Fault System The geodynamic role of the Doruneh Fault is one of the debates in the active tectonics of the Iranian plateau. Actually, two perpendicular right-lateral (i.e., N-S strike-slip fault systems) and left-lateral (e.g., Doruneh Fault) sets of faults in eastern Iran accommodate crustal shortening in ways that the relationship between the overall convergence vector and slip on strike-slip faults is nearly unclear (e.g., Allen et al., 2006; Walker and Jackson, 2004; Jackson and McKenzie, 1984). Both geological (e.g., Walker and Jackson, 2004; Meyer and Le Dortz, 2007) and geodetic GPS (Vernant et al., 2004; Masson et al., 2005; Reilinger et al.,

66 Chapter II

2006) studies indicate the northward motion of Central Iran – Lut relative to Eurasia along N- S strike-slip fault systems (section 2). While, previous geological works (e.g., Walker and Jackson, 2004; Fattahi et al., 2007) have suggested nearly pure left-lateral strike-slip faulting along the entire DFS, which is perpendicular to the overall northward motion of the southern regions (Fig. 1). The kinematics of the DFS is controlled by various interplaying factors such as the present-day regional stress state (section 5.2), structural complexities and geometry of the DFS (Sections 4.1 and 4.2), as well as kinematic interactions between the DFS and other intersecting fault zones (this section). A vector addition of GPS-derived velocities at two SHIR and KASH stations (Masson et al., 2007) indicate an overall motion (Fig. 18a) toward NE for the relative displacement between KASH station in the southern DFS block and SHIR station in the northern block (Fig. 18b). The NE direction of the motion is perpendicular to the strike of the EFZ implying reverse / thrust faulting along the fault zone (Fig. 18c). The acute angle (~45°) between the strike of both the CFZ and WFZ and the overall GPS motion implies oblique-slip reverse left-lateral faulting along these fault zones (Fig. 18). Interestingly, the predicted GPS-derived motions along both the EFZ and WFZ are consistent with the present-day kinematics of these fault zones deduced from our geomorphic and structural data (section 4.1). But, the predicted GPS-derived oblique-slip reverse left-lateral faulting along the CFZ is not compatible with the pure left-lateral faulting demonstrated in this study (sections 4 and 5.2) and by Fattahi et al. (2007). In this geodynamic context, the question to know is how the CFZ remains pure strike-slip, while, the NE relative motion between the southern and northern fault blocks implies a component of vertical faulting along the fault zone (Fig. 18a and c)? Actually, such a discrepancy is explained by the partitioning of overall slip along the CFZ and other adjacent parallel faults (Fig. 18c). The northward component of the NE relative motion is partially transferred northward along the Dahan-Qaleh fault zone (section 4.1) and is taken up by reverse faults at the northern side of the CFZ such as the and Kuh Sorkh faults (e.g., Fattahi et al., 2006; Hollingsworth et al., 2010). The Dahan-Qaleh fault zone divides the northern side of the DFS into two individual fault blocks and merges into the DFS (Fig. 18). Left-lateral faulting along the Dahan-Qaleh fault zone implies a SW translation of its western block relative to the WFZ. Along the WFZ, this relative movement is resolved into reverse and left-lateral components of faulting, which the reverse component is predominant (Fig. 18a).

67 Chapter II

Figure 18. A simplified tectonic model to describe the present-day kinematics of the DFS based on the data and deduced results presented in this study. Yellow arrow is the GPS-derived vector (3.3 mm/yr) of active motion between the southern and northern blocks of the DFS. This velocity vector (Figure 15a) was calculated from the relative motion between SHIR and KASH stations (Figure 15b) reported by Masson et al. (2007). White arrows illustrate the theoretical direction of the contraction suggested along the WFZ and EFZ. According to our geomorphic and structural data, there is no evidence of shortening across the CFZ suggesting that the contractional component of deformation is transferred northward, and is accommodated by geological reverse faulting and folding to the north of the CFZ. It is noteworthy that the 2.5 mm/yr presented in this model is an “assumed maximum” slip rate derived from differential GPS velocities between SHIR and KASH stations, which is different from and independent to the geomorphic-derived 2.5 mm/yr slip rate proposed by Fattahi et al. (2007).

This mechanism clearly explains the drastic change in the geomorphology of the DFS, and the decreasing left-lateral offsets from the CFZ to the WFZ just west of the intersection point between the Dahan-Qaleh fault zone and WFZ (section 0). The eastward component of the NE relative motion is assumed to be principally accommodated by left-lateral faulting along the DFS, which in turn, contributes to thrust faulting along the imbricate reverse faults at the EFZ restraining fault termination (Fig. 18c). Interestingly, slip partitioning along the CFZ can explain the occurrence of the 14/12/1994 and 02/02/2000 earthquakes that indicate nearly pure thrust faulting on E-W fault planes parallel to the CFZ (section 3, Fig. 3).

68 Chapter II

Preexisting tectonic models are principally based on fault block rotation theory (Jackson and McKenzie, 1984; Walker et al., 2004; Walker and Jackson, 2004). In all these models, the focus of discussions has been the curvature, rather than kinematics of the Doruneh Fault. The model proposed by Walker and Jackson (2004) assumes that eastward increasing slip rates on N-trending sets of right-lateral faults (e.g. Dehshir, Anar, Nayband and Nehbandan faults) could cause clockwise block rotation around vertical axes, which in turn, allows the Doruneh Fault to accommodate the northward motion of Central Iran relative to Eurasia. The maximum curvature in the trace of the Doruneh Fault has been related to the highest slip rate on the Sistan suture zone. However, key observations oppose some basics of this model. Firstly, the definition of block rotation is unclear. Regardless the fact that block rotation models are based on the continuum deformation hypothesis, which requires specific conditions (e.g., Thatcher, 1995; 2003 and references therein; Shabanian et al., 2009), the first step to establish such a model is to define the geometry of individual rotating rigid blocks. In theory, the “planar” Doruneh Fault separates two distinct tectonic domains. The southern side comprises several block-like regions bounded by N-S right-lateral faults (Fig. 1). The northern side (NE Iran) is a complex deformation domain affected by ~WNW-trending reverse / thrust faults together with other differently directed strike-slip faults (Fattahi et al., 2006; Hollingsworth et al., 2010; Shabanian et al., 2009a, 2009b; Shabanian et al., 2010). Neither the northern nor the southern sides of the Doruneh Fault can be considered as individual rigid blocks. Moreover, the non-uniform rotation of the Doruneh Fault (walker and Jackson, 2004) requires that assumed rigid blocks at either side to be deformed. Indeed, if the fault bounded domains are deformable, the possible change in the orientation of the Doruneh Fault is only a simple structural deflection, not block rotation. Secondly, there is no evidence indicating that the 600-km-long Doruneh Fault, with an unknown geological history, was initially an E-trending straight fault (Walker and Jackson, 2004). In addition, the suggestion of an initial E-W Doruneh Fault is already discarded by the present-day orientation of the fault; the eastern and western parts of the Doruneh Fault imply rotations in opposite directions with respect to an assumed initial E-W line (Fig. 19a). The more fundamental issue is that individual fault slip rates, which have been used to explain the eastward increasing clockwise rotation of the Doruneh Fault (Walker and Jackson, 2004), cannot be directly integrated in the geodynamic context.

69 Chapter II

Figure 19. Simplified tectonic model to examine the hypothesis of clockwise block rotation proposed by Walker and Jackson (2004). The main faults of central and eastern Iran are shown in Figure 16a. Thick grey line indicates the hypothetical E-W geometry of the DFS suggested by Walker and Jackson (2004). To achieve the present-day geometry of the DFS the fault trace needs to rotate clockwise in the eastern part and counterclockwise in the western part. (b) The model-derived geometry of the DFS caused by eastward increasing slip rates along the N-S right-lateral faults.

But, those to be converted to relative displacements with respect to Eurasia. More precisely, despite the eastward increasing slip rates of two adjacent fault-bounded blocks, the northward motion of the fault-bounded blocks relative to Eurasia increase westwards as slip rates of other western faults are added (Fig. 19b). In such a way, the westernmost block (Central Iran) moves faster than all eastern blocks implying the highest shear rate and largest northward displacement for the westernmost block with respect to Eurasia (e.g., Reilinger et al., 2006). As a result, the initial E-W Doruneh Fault assumed by Walker and Jackson (2004) would changes into a final NW-striking fault, incompatible with the present-day geometry of the Doruneh Fault (Fig. 19b). As it is explained above, neither eastward increasing fault slip rates nor westward increasing shear rates relative to Eurasia can explain the present-day geometry of the Doruneh Fault. Aside from all these discrepancies, it is difficult to explain different kinematics of the WFZ, CFZ, and EFZ by a clockwise rotation of the Doruneh Fault.

70 Chapter II

6.2. Initiation of strike-slip faulting along the Doruneh Fault System Along the DFS, the maximum left-lateral offset (840±70 m) is recorded by a Quaternary alluvial fan surface (Q3). But, the total cumulative displacement of the fault is unknown. Nevertheless, various cross-cutting relationships between faults and other geological structures within the fault system allowed us to estimate a possible onset age for the present- day strike-slip faulting along the DFS. As discussed in section 3, early post Eocene vertical movements (normal or reverse faulting) of the DFS had continued to the Miocene time causing the subsidence of the southern Neogene sedimentary basin with respect to the northern pre-Oligocene paleo-reliefs. Subsequently, Miocene marls and sandstones together with Pliocene conglomerates have been folded and faulted at the eastern termination of the DFS. These contractional structures are likely due to the compressional tectonic regime which, along the EFZ, is partly resulted from left-lateral strike-slip faulting along the CFZ (section 6.1). Such a suggestion implies that left- lateral strike slip faulting along the DFS may have started after Miocene or during Pliocene. On the other hand, the knife-cut trace of the Great Kavir Fault runs through the Dasht-e Kavir depression, and left-laterally displaced the post Miocene dome and basin structures (Fig. 2b). This cross-cutting relationship indicates a maximum Pliocene age (≤5 Ma) for the initiation of strike slip faulting along the Great Kavir Fault. These age estimates are rather consistent with the post-Miocene (~4 Ma) tectonic reorganization that caused regional strike- slip faulting in NE Iran (Shabanian et al., 2009a, 2009b).

7. Conclusion The structural and geomorphic investigations presented in this paper revealed that the curved shape DFS is a structural assemblage of three distinct fault zones, i.e., WFZ, CFZ, and EFZ; instead of, a 400-km-long uniform structure. Each fault zone is characterized by its own geometry and kinematics leading to discrete structural and geomorphic characteristics. It seems, thus, unlikely that the “Doruneh Fault”, including the Great Kavir Fault, could rupture along the entire ~600 km length to produce earthquakes of M ≥ 8. The onset of strike-slip faulting along both the DFS and Great Kavir Fault is suggested at ≤5 Ma that is probably concurrent with the major tectonic reorganization reported in NE Iran (Shabanian et al., 2009a, 2009b). In the kinematic model presented in this study (Fig. 18c), the overall northeastward motion between Central Iran - Lut and NE Iran is taken up by reverse left-lateral oblique-slip faulting

71 Chapter II along the WFZ and reverse faulting along the EFZ. In the central part, partitioning of slip into strike-slip and reverse component of faulting on parallel faults allows the CFZ to remain pure left-lateral strike-slip, while the overall convergence vector and slip on the fault are perpendicular. Such a model helps to better understand the geodynamic role of the DFS in the accommodation of ongoing Arabia-Eurasia convergence. Our data and the deduced results indicate that even continuous large strike-slip faults may be divided into several discrete fault zones to take up more easily overall tectonic motions. Such a deformation pattern may has further implications in other tectonic domains affected by perpendicular sets of strike-slip faults that accommodate crustal shortening in ways that the relationship between the overall convergence vector and slip on faults is drastically different.

72 Chapter II

Acknowledgements This work was funded by the INSU-CNRS (France) and partly by the International Institute of Earthquake Engineering and Seismology (IIEES, Iran), supervised by D. Hatzfeld, M. Ghafouri Ashtiani, and A.A. Tasnimi. We thank M. Zare (IIEES) for support and administrative assistance. Funding was provided by the Dyeti and PNRN programs (INSU- CNRS), and ACI FNS program (French Ministry of Research), within the above mentioned co-operative agreement. SPOT images were provided thanks to the ISIS projects n° 76 (©CNES 2007 to 2008, distribution SPOT images S.A.). Y. Farbod benefits of a Foreign Affair Ministry (Ministère des Affaires Etrangères, France) grant through French Embassy in Iran. We thank V. Grimault, Ch. Duhamel and the staff of the SCAC of the French Embassy in Tehran, for their support. We are grateful to R.T. Walker, K. Hessami Azar, an anonymous reviewer, and an Associate Editor for helpful and constructive reviews. The editor Onno Oncken is acknowledged for help and handling the manuscript. The General government of Khorassan-e Jonubi (Southern Khorassan) province has efficiently helped us during two field works in 2007 and 2010.

73 Chapter II

References Adams, J., P. W. Basham, and S. Halchuk (1995), ortheastern North American earthquake potential - new challenges for seismic hazard mapping, Geological Survey of Canada Current Research, 1995-D, 91-99. Alavi-Naini, M., M. J. Vaezi-Pour, N. Alavi Tehrani, A. Behrouzi, and M. H. Kholghi (1992), Geological Map of Torbat-e Heidariyeh, 1:250 000, Geological survey of Iran. Allen, M. E Blanc, R. Walker, J. Jackson, M. Talebian, M. Ghassemi (2006), Contrasting styles of convergence in the Arabia-Eurasia collision: Why escape tectonics does not occur in Iran, Geological Society of America, Special Paper 409, 579-589. Ambraseys, N., and C. Melville (1982), A History of Persian Earthquakes, Cambridge University Press, Cambridge, UK. Barka, A. A., and K. Kadinsky-Cade (1988), Strike-slip fault geometry in Turkey and its influence on earthquake activity, Tectonics, 7, 663–684. Barka, A., et al. (2002), The surface rupture and slip distribution of the 17 August 1999 I˙zmit earthquake, (M 7.4), North Anatolian fault, Bull. Seismol. Soc. Am., 92(1), 43-60. Bellier, O., and M. L. Zoback (1995), Recent state of stress change in the Walker Lane zone, western Basin and Range Province, United States, Tectonics, 14(3), 564-593. Berberian, M., and R. Yeats (1999), Patterns of Historical Earthquake rupture in the Iranian Plateau, Bull. Seismol. Soc. Am., 89, 120-139. Berberian, M., and R. Yeats (2001), Contribution of archaeological data to studies of earthquake history in the Iranian Plateau, J. Struct. Geol., 23, 563-584. Bruhn, R. L., Sibler, P. R. and Parry, W. T. (1987), Rupture characteristics of normal faults: an example from the Wasatch fault zone, Utah. In: Continental Extensional Tectonics (edited by Coward, M. P., Dewey, J. F. and Hancock, P. L.). Spec. Pub. geol. Soc. Land., 337-353. Budding, E., D., Schwartz, and D., Oppenheim (1991), Slip rate, earthquake recurrence, and seismogenic potential of the Rodgers Creek Fault Zone, northern California: initial results, Geophysical Research Letters, 18(3), 447-450. Carey, E. (1979), Recherche des directions principales de contraintes associées au jeu d'une population de failles., Rev. Geol. Dyn. Geogr. Phys, 21, 57-66. Carey, E., and B. Brunier (1974), Analyse théorique et numérique d'un modèle mécanique élémentaire appliqué à l'étude d'une population de failles, C.R. Acad. Sci., Paris, 179, 891-894. Coppersmith K.J. (1988), Temporal and spatial clustering of earthquake activity in the central and eastern United States, Seismological Research Letters, 59, 299-30. dePolo, C. M., D. G. Clark, D. B. Slemmons, and A. Ramallie (1991), Historical basin and range province surface faulting and fault segmentation J. Struct. Geol., 13, 123-136. Eftekhar-Nezhad, J., A. Aghanabati, B. Hamzehpour, V. Baroyant (1976), Geological map of Iran, Ser. J5, Kashmar sheet, Scale 1: 250 000, Geol. Surv. of Iran,Tehran. Fattahi, M., et al. (2006), Holocene slip-rate on the Sabzevar thrust fault, NE Iran, determined using optically stimulated luminescence (OSL), Earth and Planetary Science Letters, 245, 673–684, doi:10.1016/j.epsl.2006.03.027. Fattahi, M., R. T. Walker, M. M. Khatib, A. Dolati and A. Bahroudi; (2007), Slip-rate estimate and past earthquakes on the Doruneh fault, eastern Iran, Geophys. J. Int. 168, 691–709 doi: 10.1111/j.1365-246X.2006.03248.x Giessner, K. H., Hagedorn, and M., Sarvati (1984), Geomorphological studies in the Kashmar region (NE Iran), N. Jb. Geol. Palaont. Abh., 168, 545–557. Hessami, K., F. Jamali, and H. Tabassi (2003), Major active faults of Iran, scale 1:2 500 000, International Institute of Earthquake Engineering and Seismology, Iran, -Tehran.

74 Chapter II

Hollingsworth, J., et al. (2010), Oroclinal bending, distributed thrust and strike-slip faulting, and the accommodation of Arabia–Eurasia convergence in NE Iran since the Oligocene, Geophys. J. Int., 181, 1214–1246, doi: 10.1111/j.1365-246X.2010.04591.x. Hollingsworth, J., J. Jackson, R. Walker, M. Gheitanchi, and M. Bolourchi (2006), Strike-slip faulting, rotation, and along-strike elongation in the Kopeh Dagh mountains, NE Iran, Geophys. J. Int., 166, 1161–1177, doi: 10.1111/j.1365-246X.2006.02983.x. Huber, H. (1977), Geological map of Iran, Sheet, Ser. No.2 and 3, North Central Iran and North East Iran, Scale sheets, scale 1:1 000 000 National Iranian Oil company, Tehran. Ikeda, M., S. Toda, S. Kobayashi, Y. Ohno, N. Nishizaka, and I. Ohno (2009), Tectonic model and fault segmentation of the Median Tectonic Line active fault system on Shikoku, Japan, Tectonics, 28, TC5006, doi: 10.1029/2008TC002349. International Seismological Centre, EHB Bulletin, http://www.isc.ac.uk, Internatl. Seis. Cent., Thatcham, United Kingdom, 2009. Jackson, J., and D. Mckenzie (1984), Active tectonics of the Alpine-Himalayan Belt between Turkey and Pakistan, Geophys. J. R. astr. Soc, 77(1), 214-245. Javidfakhr, B., O. Bellier, E. Shabanian, S. Ahmadian, and A. Saidi (2011), Plio-Quaternary tectonic regime changes in the transition zone between Alborz and Kopeh Dagh mountain ranges (NE Iran), Tectonophysics, 506(1-4), 86-108, doi:10.1016/j.tecto.2011.04.013. Keller, E.A., D.B., Seaver, D.L., Laduzinsky, D.L., Johnson and T.L., Ku,T.L (2000) Tectonic geomorphology of active folding over buried reverse faults: San Emigdio Mountain front, southern San Joaquin Valley, California, Geol.Soc. Am. Bull., 112(1), 86–97. King, G. C. P., and J. Nabelek (1985), Role of fault bends in the initiation and termination of earthquake rupture, Science, 228, 984-987. Knuepfer, P. L. K. (1989), Implication of the characteristics of end-points of historical surface fault ruptures for the nature of fault segmentation, U.S. Geol. Surv. Open-File Rep, 193- 228. Kondo, H., Y. Awata, O. Emre, A. Dougan, S. Ozalp, F. Tokay, C. Yildirim, T. Yoshioka, and K. Okumura (2005), Slip distribution, fault geometry, and fault segmentation of the 1944 Bolu-Gerede earthquake rupture, North Anatolian fault, Turkey, Bull. Seismol. Soc . Am., 95, 1234 – 1249, doi : 10.1785/0120040194. Lees, J.M., and C. Nicholson (1993), Three-dimensional tomography of the 1992 southern California earthquake sequence: Constraints on dynamic earthquake rupture, Geology, 21, 387 – 390, doi:10.1130/0091-7613(1993)021<0387:TDTOTS>2.3.CO;2 Lettis, W., J. Bachhuber, R. Witter, C. Brankman, C. E. Randolph, A. Barka, W. D. Page, and A. Kaya (2002), Influence of Releasing Step-Overs on Surface Fault Rupture and Fault Segmentation: Examples from the 17 August 1999 izmit earthquake of the North Anatolian Fault, Turkey, Bull. Seismol. Soc. Am., 92(1), 19-42. Masson, F., et al., (2005), Seismic versus aseismic deformation in Iran inferred from earthquake and geodetic data, Geophys. J. Int., 160, 217-226, doi : 10.1111/j.1365- 246X.2004.02465.x. Masson, F., M. Anvari, Y. Djamour, A. Walpersdorf, F. Tavakoli, M. Daignieres, H. Nankali, and S. Van Gorp (2007), Large-scale velocity field and strain tensor in Iran inferred from GPS measurements: new insight for the present-day deformation pattern within NE Iran, Geophys. J. Int., 170, 436-440, doi: 10.1111/j.1365-246X.2007.03477.x. McCalpin, J. P. (Eds) (1996), Paleoseismology, Academic, San Diego, California. McClusky, S., R. Reilinger, S. Mahmoud, D. Ben Sari, and A. Tealeb (2003), GPS constraints on Africa (Nubia) and Arabia plate motions, Geophys. J. Int., 155(1), 126-138.

75 Chapter II

Meyer, B., and K. Le Dortz (2007), Strike-slip kinematics in Central and eastern Iran: Estimating fault slip rates averaged over the Holocene, Tectonics, 26, TC5009, doi: 10.1029/2006TC002073. Mohajer-Ashjai, A. (1975), Recent and contemporary crustal deformation in eastern Iran, Ph.D. thesis, Imperial college, London. Nazari, H., et al., (2009), First evidence for large earthquakes on the Deshir Fault, Central Iran Plateau, Terra Nova, 00, 1-10, doi: 10.1111/j.1365-3121.2009.00892.x. Nishigami, K. (2000), Deep crustal heterogeneity along and around the San Andreas fault system in central California and its relation to the segmentation, J. Geophys. Res., 105, 7983–7998, doi:10.1029/1999JB900381. Regard, V., et al. (2005), Cumulative right-lateral fault slip rate across the Zagros–Makran transfer zone: role of the Minab–Zendan fault system in accommodating Arabia–Eurasia convergence in southeast Iran, Geophys. J. Int., 162, 177-203, doi: 10.1111/j.1365- 246X.2005.02558.x. Regard, V., O. Bellier, J.-C. Thomas, M. R. Abbassi, J. Mercier, E. Shabanian, K. Feghhi, and S. Soleymani (2004), Accommodation of Arabia-Eurasia convergence in the Zagros- Makran transfer zone, SE Iran: A transition between collision and subduction through a young deforming system, Tectonics, 23, TC4007, doi: 10.1029/2003TC001599,. Reilinger, R., et al. (2006), GPS constraints on continental deformation in the Africa-Arabia- Eurasia continental collision zone and implications for the dynamics of plate interactions, J. Geophys. Res., 111, B05411, doi: 10.1029/2005JB004051. Ritz, J. F., H. Nazari, A. Ghassemi, R. Salamati, A. Shafei, S. Soleymani, and P. Vernant (2006), Active transtension inside central Alborz: A new insight into northern Iran– southern Caspian geodynamics, Geology, 34, 477-480, doi: 10.1130/G22319.1. Schwartz, D. P., and K. J. Coppersmith (1984), Fault Behavior and Characteristic Earthquakes' Examples From the Wasatch and San Andreas Fault Zones, J. Geophys. Res., 89(B7), 5681-5698. Schwartz, D.P. and Sibson, R.H. (Eds) (1989), Fault segmentation and controls of rupture initiation and termination, USGS Open-File Report 89-315, Menelo park, California. Sella, F. G., H. T. Dixon, and A. Mao (2002), Revel: A model for Recent plate velocities from space geodesy, J. Geophys. Res., 107(B4), ETG 11. Shabanian, E., L. Siame, O. Bellier, L. Benedetti, and M. R. Abbassi (2009a), Quaternary slip rates along the northeastern boundary of the Arabia–Eurasia collision zone (Kopeh Dagh Mountains, Northeast Iran), Geophys. J. Int., 1-23, doi: 0.1111/j.1365- 246X.2009.04183.x. Shabanian, E., O. Bellier, L. Siame, N. Arnaud, M. R. Abbassi, and J.-J. Cochemé (2009b), New tectonic configuration in NE Iran: Active strike-slip faulting between the Kopeh Dagh and Binalud mountains, Tectonics, 28, TC5002, doi: 10.1029/2008TC002444. Shabanian, E., O. Bellier, M. R. Abbassi, L. Siame, and Y. Farbod (2010), Plio-Quaternary stress states in NE Iran: Kopeh Dagh and Allah Dagh-Binalud mountain ranges, Tectonophysics, 480, 280-304, doi: doi:10.1016/j.tecto.2009.10.022. Solaymani Azad. S., (2009), Seismic hazard assessment for Tehran, Tabriz and Zanjan cities (NW Iran) based on morphotectonics and paleoseismology, Ph.D. thesis, University of Montpellier, France, 150 p. Stewart, M. E., and W. J. Taylor (1996), Structural analysis and fault segment boundary identification along the Hurricane fault in southwestern Utah, J. Struct. Geol., 18(8), 1017-1029. Stocklin, J., and M.H., Nabavi (1973), Tectonic map of Iran, Scale, 1: 2 500 000, Geol. Surv. of Iran, Tehran.

76 Chapter II

Sylvester, A. (1988), Strike-slip faults, Geological Society of America Bulletin, 100, 1666- 1703. Tavakoli, F., (2007), Present-day kinematics of the Zagros and east of Iran faults, Ph.D. thesis, Université Joseph Fourier, Grenoble, France (available at http://tel.archives- ouvertes.fr/tel-00285919/fr/ Tchalenko, J. S. (1973), The Kashmar (Turshiz) 1903 and Torbat-e-Heidariyeh (south) earthquakes in Central Khorassan (Iran), Annali de geofisca, 26(1), 29-40. Tchalenko, J. S., and M. Berberian (1975), Dasht-e-Bayez fault, Iran: Earthquake and earlier related structures in bedrock, Geol. sco.Amr 86, 703-709. Tchalenko, J. S., M. Berberian, and H. Behzadi (1973), Geomorphic and seismic evidence for recent activity on the Doruneh Fault, Iran, Tectonophysics, 19, 333-341. Thatcher, W. (1995), Microplate versus continuum descriptions of active tectonic deformation, J. Geophys. Res., 100, 3885 – 3894, doi:10.1029/94JB03064. Thatcher, W. (2003), GPS constraints on the kinematics of continental deformation, Int. Geol. Rev., 45, 191 – 212, doi:10.2747/0020-6814.45.3.191. Tirrul, R., I. R. Bell, R. J. Griffis, and V. E. Camp (1983), The Sistan suture zone of eastern Iran, Geological Society of America Bulletin, 94, 134-150. Vernant, P., et al. (2004), Present-day crustal deformation and plate kinematics in the Middle East constrained by GPS measurements in Iran and northern Oman, Geophys. J. Int., 157(1), 381-398, doi: 10.1111/j.1365-246X.2004.02222.x. Walker, R., and J. Jackson (2004), Active tectonics and late Cenozoic strain distribution in central and eastern Iran, Tectonics, 23, TC5010, doi: 10.1029/2003TC001529. Walker, R., J. Jackson, and C. Baker (2003), Surface expression of thrust faulting in eastern Iran: source parameters and surface deformation of the 1978 Tabas and 1968 Ferdows earthquake sequences Geophys. J. Int. 152, 749-765. Walker, R., J. Jackson, and C. Baker (2004), Active faulting and seismicity of the Dasht-e- Bayaz region, eastern Iran, Geophys. J. Int. 157, 265-282, 10.1111/j.1365- 2966.2004.02179.x. Walker, R.T. and M. Fattahi, (2011), A framework of Holocene and Late Pleistocene environmental change in eastern Iran inferred from the dating of periods of alluvial fan abandonment, river terracing, and lake deposition, Quaternary Science Reviews, 30, 1257-1272, doi:10.1016/j.quascirev.2011.03.004. Wallace, R. E. (1970), Earthquake recurrence intervals on the San Andreas Fault, Geological Society of American Bulletin, 81, 2875-2890. Wellman, H. W. (1966), Active wrench faults of Iran Afghanistan and Pakistan, Geologische undschau, 55, 716-735. Wells, D., and K.J. Coppersmith (1994), New empirical relationships among Magnitude, Rupture length, Rupture Width, Rupture Area, and surface Displacement, Bull. Seismol. Soc. Am., 84(4), 974-1002. Woodcock, N. H., and M. Fischer (1986), Strike-slip duplexes, J. Struct. Geol. 8, 725-735. Wulff, H. E. (1968), The Qanats of Iran, Sci. Am., 218, 94-105, doi: 10.1038/scientificamerican0468-94.

77

CHAPTER III

Temporal and spatial variations in late Quaternary slip rates along the Doruneh Fault System (NE Iran)

Chapter III

Temporal and spatial variations in late Quaternary slip rates along the Doruneh Fault System (NE Iran)2

Yassaman Farbod1, *, Esmaeil Shabanian1, Olivier Bellier1, Mohammad Reza Abbassi2, Régis Braucher1, Lucilla Benedetti1, Didier Bourlès1, Khaled Hessami Azar2

1 Aix-Marseille Univ., CEREGE, UMR 6635, 13545 Aix en Provence cedex 4, France CNRS, CEREGE, UMR 6635, 13545 Aix en Provence cedex 4, France IRD, CEREGE, UMR 161, 13545 Aix en Provence cedex 4, France Collège de France, CEREGE, 13545 Aix en Provence cedex 4, France

2 International Institute of Earthquake Engineering and Seismology (IIEES), Tehran, Iran

Abstract The Doruneh Fault System (DFS) is one of the major active strike-slip faults in the Arabia- Eurasia collision zone. Despite its geological activity, no historical or instrumental large earthquake (M≥6.5) was recorded along the fault system. The DFS comprises three fault zones, i.e., WFZ, CFZ, and EFZ, with distinct kinematics and surface geometry. However, the rate, distribution, and behavior of slip on these fault zones are unknown, so, their seismic behavior remains mysterious. We reconstructed sixty seven geomorphic offsets recorded by Q1, Q2, and Q3 alluvial fans and their associated geomorphic markers displaced along the WFZ and CFZ. Respective abandonment ages of ~12, ~36, and ~100 ka were determined for the Q1, Q2, and Q3 surfaces using in situ-produced 10Be and 36Cl nuclides. These allowed us to determine three sets of individual left-lateral slip rates, and consequently, to describe the distribution of slip along the fault zones. These slip rates that are averaged over three time periods of ~12, ~36, and ~100 ka, reveal that (1) the long-term slip behavior along both the WFZ and CFZ is characteristic, and (2) the overall slip rates of the two fault zones remain nearly constant during late Pleistocene with a maximum rate of ~8.3 mm/yr accommodated along the mid-length of the CFZ. During the Holocene, however, the slip behavior of the fault seems more complex; the CFZ is divided into two independent segments, with characteristic but symmetrical slip distributions. These segments are separated by a persistent boundary which has not been ruptured during the last ~12 ka. A maximum Holocene slip rate of ~5.3 mm/yr is estimated for left-lateral faulting along the CFZ. The maximum length of independent seismogenic fault segments varies from 70 to 100 km, which could produce characteristic earthquakes with magnitude of Mw≈7.2-7.4. Our results reconfirm that long strike-slip fault zones, do not rupture along their entire length during a single earthquake.

2 Paper in preparation

81 Chapter III

1. Introduction Quantitative data, particularly offset amounts and slip rates on major faults, hold the key to evaluate slip behavior of the faults (e.g., Ritz et al., 1995; Van der Woerd et al., 2000; Tapponnier et al., 2001). Geological slip rates derived from the relevant age and amount of geomorphic/geologic fault offsets are averaged over thousands to several million years (e.g., Ritz et al., 2003; Siame et al., 1997; Philip et al., 2001; Hessami et al., 2006; Rizza et al., 2011; Shabanian et al., 2009a), and cover a large number of seismic cycles. These slip rates lead to the better understanding of (1) the role of active faults in the accommodation of regional deformation, and (2) the seismic behavior of the faults that helps in the seismic hazard assessment and the development of physical earthquake models. Such data are essential in the areas that suffer from a lack of well-documented paleoseismicity and historical earthquakes records. The ~400-km-long Doruneh Fault System (DFS) is one of the largest strike-slip faults affecting the Iranian plateau (Wellman, 1966; Tchalenko et al., 1973; Farbod et al., 2011). Despite recent active tectonic studies(Berberian and Yeats, 1999; Fattahi et al., 2007; Farbod et al., 2011), the slip behavior and seismogenic characteristics of the DFS remain Poorly known. The structural and geomorphic investigations by Farbod et al. (2011) led to establish a segmentation model according which (1) the DFS is divided into three fault zones, with distinct structural and geomorphic characteristics, and (2) independent seismic segments could not exceed the length of ~140 km. However, the distribution and rate of slip on the fault zones, as well as the true extent of seismic segments have remained unknown. In this study, we used GeoEye© images (Google Earth) to reconstruct left-lateral geomorphic offsets recorded by three generations of alluvial fans (Q1, Q2, and Q3), which have been cut by the DFS. This new data set completes and, in some cases, constrains the geomorphic offsets reported by Farbod et al. (2011). Exposure ages of abandonment surfaces were determined using in situ-produced 10Be and 36Cl cosmogenic nuclides. These new and original data were used to evaluate the spatial distribution of, and temporal variations in, Quaternary left-lateral slip rates on the DFS. The resulting Late Pleistocene to Holocene slip rates helped us to better understand the contribution of the DFS in the accommodation of the Arabia-Eurasia convergence. This study paves the way for further paleoseismological investigations characterizing the long-term slip behavior, and the extent of independent seismogenic segments of the DFS. Our data and derived results may improve understanding of fault behavior and seismic hazard along large intracontinental strike-slip faults.

82 Chapter III

Figure 1. Structure of central and northeastern Iran showing the location of the Doruneh Fault System (DFS) together with principal deformation domains and structural units overlain on GTOPO30 shaded digital elevation model. Black arrows are GPS horizontal velocities (mm/yr) from Masson et al. (2007) in a Eurasia-fixed reference frame. The box in the upper left inset shows the location in the Arabia–Eurasia collision zone. Grey arrows and associated numbers represent Arabia–Eurasia plate velocities (mm/yr) after Reilinger et al. (2006).

2. Tectonic setting and general geology The Iranian plateau is deformed between the converging Arabian and Eurasian plates at a maximum rate of 26±2 mm/yr at ~59°E (e.g., Vernant et al., 2004; Reilinger et al., 2006). The ongoing deformation is principally taken up by strike-slip faulting (e.g., Talebian and Jackson, 2002; Walker and Jackson, 2004; Authemayou et al., 2006; Solaymani Azad et al., 2011; Ritz et al., 2006; Regard et al., 2005; Le Dortz et al., 2009; Shabanian et al., 2009b; Farbod et al., 2011). The Iranian plateau is not undergoing major crustal thickening at the lithospheric scale (e.g., Jackson et al., 1995; McClusky et al., 2003), the plateau seems

83 Chapter III deformed over a broad area, with vertically-coherent deformation (Kaviani et al., 2009). However, shortening is currently focused on the Zagros (5-9 mm/yr; Hessami et al., 2006) and Alborz Mountains (~7 mm/yr; Djamour et al., 2010). In Central Iran, the DFS separates the lithospheric N-S dextral strike-slip faults like the Nayband and Neh faults from the NW-trending Binalud and Kopeh Dagh deformation domains (Fig. 1). The DFS comprises three western (WFZ), central (CFZ) and eastern (EFZ) fault zones (Fig. 2) with distinct structural, geomorphic and kinematic characteristics (Farbod et al., 2011). The CFZ is pure left-lateral strike-slip, while the WFZ is reverse left-lateral oblique-slip. The EFZ is predominantly reverse. That is, these three fault zones accommodate the northward motion between Central Iran and Eurasia in different ways ranging from pure compression to the partitioning of deformation on parallel strike-slip and reverse faults (i.e., strain partitioning) (see Farbod et al., 2011). The DFS separates the pre-Oligocene paleoreliefs to the north (~2500 m elevation) from the Neogene folded rocks (~1200 m elevation) to the south, armored by Quaternary deposits (Fig. 2). On the northern side, the Kuh-e-Sorkh Mountains, Precambrian to Cenozoic sedimentary rocks were intruded by Upper Cretaceous to Upper Eocene volcanic rocks (Fig. 2). The south-sloping Quaternary piedmont is covered by series of alluvial fans, mostly cut by the DFS (e.g., Tchalenko et al., 1973; Farbod et al., 2011).

3. Cumulative left-lateral offsets along the Doruneh Fault System Three successive generations of alluvial fans, which form inset Q1, Q2 and Q3 (from younger to older) regional geomorphic surfaces, have left-laterally been offset along the DFS. The measurement and analysis of eighteen geomorphic offsets recorded by these surfaces allowed Farbod et al. (2011) to characterize along-strike variations in the kinematic of the fault system. We analyzed high resolution (pixel size of 50 cm) GeoEye© images (from Google Earth) to create a new data set of cumulative offsets measured in sixty seven sites distributed along the DFS (Table 1). These new images also allowed us to revise and constrain geomorphic offsets presented by Farbod et al. (2011) (see Table 1). We reconstruct left-lateral offsets recorded by alluvial fan morphologies (fan shape and associated main drainages) and/or terrace risers between two successive Q1/Q2 or Q2/Q3 alluvial fan surfaces.

84

Chapter III

es es i.e. Western

Naini et al. (1992). Fault traces are based on

-

Nezhad et al. (1976) and Alavi

-

n affected by DFS simplified after Eftekhar

and structural analysis of SPOT5 and LANDSAT ETM+ by Farbod et al., (2011). The blue lines indicate the three DFS fault zon

. Geological map of regio

2

Figure geomorphic FaultZone (WFZ), Central FaultZone (CFZ) and EasternZone Fault (EFZ)

85 Chapter III

Left-lateral offsets are only occurred along the WFZ and CFZ such that beyond the two eastern and western end-points there is no evidence of left-lateral faulting along the DFS. Obviously, the number of offset measurements corresponds to the degree of preservation of the Quaternary landforms; so, we were able to measure 7, 24, and 36 offsets recorded by Q3, Q2, and Q1 surfaces, respectively (Fig. 3a; Table 1). The analysis of Q3 surface in seven sites led us to characterize the overall pattern of left- lateral displacements along the CFZ, with a maximum offset of ~840±70 m at the mid-length of the fault zone that decreases to ~160 m at both end (Fig. 3b). The measurement of twenty four offsets in Q2 fans and related landforms provided a more detailed displacement pattern, but similar to the Q3 offsets, along both the WFZ and CFZ (Fig. 3b).

Table 1. Detailed characteristics of the measurements of the cumulative left-lateral offset along the DFS. Sites locations are on Fig. 3a. The Quality field indicates the confidence level (A: very high; B: high; C: medium; D: low) of the reconstructed offset markers Site Longitude Latitude Geomorphic Offset Piercing point Quality number (°E) (°N) Surface (m) 1 56.749 35.032 Q1 0 Western fault termination A 3 57.082 35.129 Q1 16 ± 4 Fan apex A 4 57.317 35.177 Q1 15 ± 3 Fan shape and incising drainages A 5 57.44 35.201 Q1 12 ± 2 Q1/Q modern riser offset; modern fan erosion A 8 57.473 35.206 Q1 25 ± 5 Drainages incising on the Q1 fan; fan borders D 9 57.481 35.207 Q1 15 ± 5 Drainages incising on the Q1 fan; fan borders C 12 57.5 35.209 Q1 29 ± 5 Entire fan Shape ; drainages A 13 57.507 35.212 Q1 20 ± 10 Entire fan shape; individual drainage E B 14 57.519 35.213 Q1 27 ± 6 Drainages incising the fan A 18 57.558 35.220 Q1 18 ± 4 Streams with particular morphology; Eastern Q1/Q2 riser B 20 57.58 35.225 Q1 13 ± 3 Drainages and small Q1 fans C 21 57.582 35.225 Q1 14 ± 2 Q1/Q2 riser B 23 57.615 35.236 Q1 25 ± 3 Incising drainages (minimum possible offset) D 24 57.864 35.276 Q1 50 ± 20 Q1/Q3 possible riser C 25 58.145 35.294 Q1 55 ± 5 Western fan border and main drainage A 26 58.165 35.297 Q1 65 ± 15 Eastern and western fan border; two symmetric risers A 28 58.233 35.297 Q1 40 ± 10 Q1/Q2 riser ; drainages - fan shape A 30 58.269 35.297 Q1 25 ± 5 Q1/Q2 riser; fan geomorphology C 32 58.282 35.297 Q1 30 ± 10 Fan border B 35 58.304 35.295 Q1 50 ± 10 Eastern and western border of Q1 fan A 36 58.318 35.295 Q1 60 ± 20 Eastern fan border; one drainage C 37 58.338 35.295 Q1 60 ± 10 Q1/Q2 riser ; drainages A 40 58.38 35.289 Q1 40 ± 10 General reconstruction of borders of two Q1 fans C 41 58.395 35.288 Q1 10 ± 2 Border of several Q1 fans A 42 58.4 35.287 Q1 0 Gap of faulting in Q1 surfaces A 43 58.442 35.286 Q1 0 Gap of faulting in Q1 surfaces A 44 58.443 35.286 Q1 13 ± 3 General morphology of a Q1 fan C 45 58.46 35.287 Q1 17 ± 2 Q1 ridge; riser ; overall geomorphology A 46 58.5 35.286 Q1 27 ± 10 Risers - Eastern border; Eastern river; a kariz well B 47 58.549 35.285 Q1 25 ± 5 Fan morphology; series of drainages A 50 58.66 35.278 Q1 60 ± 10 Q1/Q2 riser; minimum offset in Q2 streams A 52 58.76 35.271 Q1 55 ± 5 Incising drainages ; fan border A 53 58.76 35.271 Q1 55 ± 5 Main streams incising the fan B 55 58.873 35.261 Q1 60 ± 10 pull-apart - Western fan border; drainage ; Eastern cultivated depression B 57 58.926 35.252 Q1 40 ± 10 Suspected reconstruction; Overall shape D 60 58.954 35.248 Q1 40 ± 5 Q1/Q2 riser A 62 59.009 35.242 Q1 50 ± 10 Q1/Q2 riser C 69 59.497 35.127 Q1 0 Last possible left-lateral offset -

86 Chapter III

Table 1 (continued). Site Longitude Latitude Geomorphic Offset Piercing point Quality

number (°E) (°N) Surface (m) 1 56.749 35.032 Q2 0 Western fault termination A 2 56.766 35.038 Q2 25 ± 5 Eastern fan border B 6 57.44 35.201 Q2 60 ± 10 Q2/Q1 riser offset; modern fan erosion A 7 57.454 35.204 Q2 50 ± 10 Fans - feeding drainages B 11 57.469 35.205 Q2 70 ± 10 Entier fan shape ;- drainages C 15 57.528 35.214 Q2 90 ± 30 Drainages ; fan bounding drainages B 16 57.537 35.215 Q2 100 ± 20 Drainages incising Q2 ; fan morphology A 17 57.549 35.217 Q2 110 ± 30 Drainages incising Q2 D 19 57.571 35.223 Q2 130 ± 15 Western fan border; Incising drainages A 22 57.584 35.226 Q2 140 ± 40 Fan border; Apex; Feeding stream A 27 58.194 35.298 Q2 230 ± 20 2 drainage basin in Q3 or older; Main river; General Q2 shape C 29 58.245 35.297 Q2 190 ± 20 Q2/Q3 riser ; Incising drainage A 31 58.271 35.297 Q2 200 ± 70 Eastern border of Q2 fan C 33 58.284 35.297 Q2 220 ± 70 Isolated Q2 fan A 38 58.343 35.295 Q2 200 ± 50 Incising drainage A 39 58.373 35.291 Q2 180 ± 20 Incising drainages with irregular intervals B 48 58.553 35.284 Q2 300 ± 50 Fan apex; border and feeding drainage A 51 58.669 35.278 Q2 300 ± 50 Eastern fan border; Drainages;Eastern fan apex A 54 58.855 35.263 Q2 100 ± 20 Fan borders; Incising drainages; Eastern Q2/Q3 riser A 58 58.932 35.251 Q2 210 ± 20 Suspected reconstruction ; overall shape D 61 58.965 35.247 Q2 190 ± 30 Abandoned and beheaded drainages in Q2 B 63 59.009 35.242 Q2 190 ± 30 Abandoned and beheaded drainages in Q3 B 64 59.049 35.235 Q2 210 ± 50 Overall form and incising drainages C 66 59.131 35.22 Q2 170 ± 50 Eastern border - main bordering stream of Q2 fan; Western Q2/Q3 riser D 67 59.323 35.175 Q2 70 ± 20 Western fan border A 69 59.497 35.127 Q2 0 Last possible Left-lateral offset - 1 56.749 35.032 Q3 0 Western fault termination A 8 57.492 35.208 Q3 160 ± 100 Drainages incising Q3 deposits (just a possibility) D 34 58.29 35.297 Q3 630 ± 70 Drainages incising Q3 deposits (just a possibility) D 49 58.57 35.281 Q3 840 ± 70 Fan morphology - Main drainages A 56 58.921 35.253 Q3 330 ± 80 Q3 remnants and incising drainages C 59 58.932 35.251 Q3 350 ± 70 Suspected reconstruction (overall shape) D 65 59.115 35.222 Q3 290 ± 50 Incising drainage C 68 59.331 35.174 Q3 170 ± 30 Overall form and incising drainages C 69 59.497 35.127 Q3 0 Last possible Left-lateral offset A

The CFZ is characterized by larger left-lateral displacement relative to the WFZ. A drastic decrease in offset amounts, from ~140 to ~70 m, marks the boundary between the two fault zones. Along the CFZ, a maximum offset of ~300 m was measured. This important variation in offset values may represent possible segment boundaries. Nevertheless, the low density of points precludes any interpretation on this conclusion (Fig. 3b). The thirty six cumulative offsets recorded by Q1 related features give a nearly complete image of the slip distribution along the fault zones. The ~16 m maximum offset along the WFZ contrasts the ~60 m offset measured along the CFZ (Fig. 3b). In the localities between the longitudes of 58.400°E and 58.442°E, Q1 alluvial fans have not been cut by the fault, which is thus clearly covered by Q1 alluvial deposits (Fig. 4). Given the lack of faulting in this part, the CFZ could be divided into two eastern and western segments (see section 5.1), with symmetrical patterns of lateral displacement (Fig. 3b).

87

Chapter III

lateral lateral

-

ve ve left

strike variations of the cumulative

-

offset measurementoffset shown is bar. by error

lateral offset measurements observed on Quaternary geomorphic markers along DFS. Details on the cumulati

-

(numbered) are given in Table 1. Sites with asterix have been dated by cosmogenic method. (b) Along

(a) (a) Sites location of left

. .

3

measurements

lateral offsets, recorded by the fan surfacesQ1, and Q3.Q2 Uncertaintyof the

-

offset left Figure

88 Chapter III

example of the Q1

An

E longitude where the Q1

°

E and 58.442

°

fault. (b) and (d) show enlargement of the fault trace west and east of this zone, respectively. (c)

. (a) GeoEye satellite image (from Google Earth, ©2012 Google, ©2012 GeoEye) of the DFS between 58.4 4

Figure alluvial fans have not been cut by the alluvial fan affected not fault. by the

89 Chapter III

4. Surface exposure dating 4.1. Sampling strategy and analytical procedure for cosmogenic dating We used in situ-produced 10Be and 36Cl cosmogenic nuclides resulting from spallation and muonic reactions in quartz-rich and carbonate rocks, respectively (see review in Gosse and Philips, 2001; Dunaï, 2010). The exposure ages of Q1, Q2, and Q3 alluvial fan surfaces offset along the DFS were estimated in four sites along the CFZ and two sites along the WFZ. The bedrocks cropped out in the catchment areas of alluvial fans are mostly andesite rocks, which are locally intruded by granitic bodies, and rarely carbonate rocks (Fig. 2). This is a serious problem for sampling, specially in depth-profiles, so that carbonate or quartz-rich clasts suitable for 36Cl and 10Be cosmogenic dating are rare, precluding us the collection of larger amounts of surface samples and/or the sampling in depth-profiles. Despite these difficulties, a total 66 samples, 40*30 cm in size, were carefully collected from the well-preserved and nearly pristine parts of the fan surfaces. Among these best 66 samples, in terms of the geomorphic setting and mineral composition, only less than fifty per cent of the samples (30 samples) provided materials necessary for chemical analyses. In two cases, along the WFZ, catchment areas incised in Cretaceous limestones allowed us to use in situ-produced 36Cl cosmogenic nuclide, which helped us completing the 10Be exposure age for the alluvial abandonment surfaces offset along the CFZ and WFZ. The quartz-rich samples were prepared for AMS 10Be measurements following chemical procedures adapted from Brown et al. (1991) and Merchel and Herpers (1999). Samples were first crushed and sieved to obtain fraction comprised between 1 and 0.250 mm. After a magnetic separation, the non-magnetic fraction underwent selective etchings in fluorosilicic

(H2SiF6) and hydrochloric acids (HCl) to eliminate all mineral phases but quartz. Quartz minerals were etched in hydrofluoric acid to eliminate potential surface contamination by 10Be produced in the atmosphere. The cleaned quartz minerals were then completely dissolved in hydrofluoric acid after addition of ~100 µl of an in-house 3.10-3 g/g (9Be) carrier solution prepared from a deep-mined phenakite mineral (Merchel et al., 2008). Beryllium was separated from these solutions by anion and cation exchange columns, followed by Be extraction by alkaline precipitations. The final precipitate was dried, and heated at 900°C to obtain BeO. Beryllium oxides were mixed with Niobium for target packing and samples were analyzed by Accelerator Mass Spectrometry (AMS). All 10Be concentrations are normalized to 10Be/9Be SRM 4325 NIST reference material with an assigned value of (2.79±0.03) 10-11. This standardization is equivalent to 07KNSTD within rounding error. The 10Be half-life of (1.39±0.01)106 years used is that recently recommended by Korschinek et al. (2010) and

90 Chapter III

Chmeleff et al. (2010) according to their two independent measurements. Analytical uncertainties (reported as 1σ) include a conservative 0.5% external uncertainty based on long- term measurements of standards, a one-sigma statistical error on counted 10Be events, and the uncertainty associated with the chemical blanks correction (Arnold et al., 2010). Three chemical blanks prepared with the samples yield 10Be/9Be ratio of 1.9×10-15, 2.1×10-15 and 5.3×10-15. To determine production rates, scaling factors for latitude and altitude corrections were calculated according to Stone (2000) and using a modern 10Be production rate at sea level and high latitude of 4.5±0.3 atoms/g-SiO2/yr to account for the reevaluation of absolute calibration of 10Be AMS standards proposed by Nishiizumi et al. (2007). For the carbonate samples, the treatment such as grinding, leaching and chemical extraction of chlorine by precipitation of silver chloride, was performed following the methodology describe by Stone et al. (1996). The samples were spiked with known quantity (~0.5 ml) of stable chlorine carrier (e.g., Desilets et al., 2006). Chlorine-36 has two production pathways in superficial rocks which include cosmic ray interaction (spallation reactions of Ca, K, Ti and Fe, capture of low-energy epithermal and thermal neutrons of 35Cl and direct capture of slow negative muons by 40Ca and 39K) and radiogenic production by disintegration of U and Th. To determine the proportion of radiogenic 36Cl requires measuring the concentration of U and Th in the target mineral (Bierman et al. 1995; Zreda et al. 1991; Stone et al. 1998 and 1996; Gosse & Phillips 2001; Schimmelpfennig et al. 2009). It is therefore important to know precisely the chemical composition of the target mineral to determine the rate of production and thus correctly interpret the measured 36Cl concentrations. Major elemental compositions of rock samples were determined (Table 2a) by the Inductively Coupled Plasma-Optical Emission Spectrometry technique by the Centre National de la Recherche Scientifique (SARM, CNRS-Nancy). The age of samples were calculated using the Ca concentration in the dissolved part of the samples according to Schimmelpfennig et al. (2009) excel spreadsheet. The production rate proposed by Schimmelpfennig et al. (2011) found more appropriate with respect to the other rates (e.g. Stone et al., 1998). For all sampling sites, corrections for shielding by the surrounding topography, snow cover and sample geometry, following Dunne et al. (1999), have negligible impact on the surface production rates. The concentrations of in situ-produced 36Cl and 10Be nuclides are given in Table (2 and 3). To illustrate cosmogenic exposure ages and their associated uncertainties, we used the sum of the Gaussian probability distribution (e.g., Deino and Potts, 1992) according to Taylor (1997):

91 Chapter III

2 2 (t ai ) / 2 i psum (t) e / i 2 (1) i

Where t is time, ai is the exposure age of sample i and (2σi) is the associated uncertainty. Statistically, a probability value less than 0.05 indicates that there is a significant amount of non analytical error in the data set, and that one or more samples are outliers. In such a case, cumulative frequency plots are generally bimodal in shape, with the secondary peak identifying outliers. The probability of 95%, i.e., 1.96σ, was chosen to calculate uncertainties associated with exposure ages: (2)

The Chi-square inversion test (Ward and Wilson, 1978) has also been used, in order to minimize the differences between theoretical and measured concentration.

Table 2. (a) Elemental composition of the bulk rock (per cent) Sample Water Al2O3 Cao Fe2O3 K2O MgO MnO Na2O P2O5 SiO2 TiO2 Th U ANA II-01 0.42 0.348 51.155 0.529 0.055 1.314 0.0137 N.D. N.D. 3.66 0.027 0.29 1.048 ANA II-02 0.56 0.203 32.72 1.468 0.052 18.807 0.03 N.D. N.D. N.D. 0.019 0.13 1.873 ANA II-03 0.39 0.425 33.76 0.22 0.084 18.007 0.008 N.D. N.D. 0.97 0.033 0.33 1.238 ANA 10-01 0.43 0.236 33.61 0.287 0.048 19.19 0.0121 N.D. N.D. N.D. 0.024 0.16 1.858 ANA 10-03 0.41 0.15 31.825 0.663 0.027 19.777 0.0146 N.D. N.D. N.D. 0.016 0.10 1.824 ANA 10-04 0.45 0.514 32.21 0.326 0.108 19.223 0.0081 N.D. N.D. 0.82 0.032 0.37 1.08 ANA 10-06 0.65 0.535 33.96 0.387 0.09 16.563 0.0169 N.D. N.D. 2.57 0.035 0.40 1.511 DOR-01 0.44 0.135 36.44 0.502 0.022 15.69 0.0117 N.D. N.D. N.D. 0.016 0.07 1.081 DOR-02 0.38 0.347 31.685 0.241 0.054 19.39 0.0084 N.D. N.D. 1.19 0.03 0.21 0.487 DOR-03 0.66 0.173 51.785 0.053 0.027 3.173 0.0099 N.D. N.D. N.D. 0.021 0.07 2.884

(b) Sample characteristic of 36Cl exposure ages Sample surface Latitude Longituge Elevation Cao Chlorine 36Cl production rate 36Cl Age min Age Max (°N) (°E) (m) (per cent) (ppm) (atom/g/ yr) (106Atom/g) (ka) (ka) ANAII-01 Q2 35.24 57.61 972 54.25 80 46.86 1.66 ± 0.11 38.3 ± 4.2 38.6 ± 1.9 ANAII-02 Q2 35.24 57.61 973 32.25 119 35.56 1.50 ± 0.11 41.3 ± 4.7 40.6 ± 2.0 ANAII-03 Q2 35.24 57.61 973 32.91 104 34.92 1.42 ± 0.09 40.6 ± 4.4 40.1 ± 2.0 ANA10-01 Q3 35.24 57.61 1001 33.89 111 36.43 1.59 ± 0.12 42.4 ± 4.8 41.8 ± 2.1 ANA10-03 Q3 35.24 57.61 1000 31.27 269 56.72 3.85 ± 0.34 70.2 ± 9.0 66.0 ± 3.3 ANA10-04 Q3 35.24 57.61 1004 30.79 133 36.01 2.78 ± 0.2 77.3 ± 9.0 74.7 ± 3.7 ANA10-06 Q3 35.24 57.61 999 35.88 92 35.41 3.23 ± 0.2 98.2 ± 11.1 97.2 ± 4.9 DOR-01 Q3 35.23 57.57 990 34.53 117 37.44 2.65 ± 0.19 74.3 ± 6.8 72.3 ± 3.6 DOR-02 Q3 35.23 57.57 993 32.82 129 37.67 3.41 ± 0.26 96.1 ± 11.5 91.6 ± 4.6 DOR-03 Q3 35.23 57.57 1005 53.32 35 42.52 2.7 ± 0.12 70.2 ± 7.6 72.5 ± 3.6

(a) Measurements of the elements were undertaken at the SARM facility (Nancy, France). N.D is the non detectable elements. (b) In situ-produced 36Cl concentration and modeled ages of sample from the three alluvial fans, affected by DFS. 36Cl concentrations were undertaken at the ASTER-CEREGE facility (Aix-en-Provence, France). 36Cl ages have been calculated using method of Schimmelpfennig et al. (2009). The production rates are those of Schimmelpfennig et al. (2010). Age min is the exposure ages assuming negligible erosion on the fan surface. Age max is the exposures ages assuming a maximum denudation rate (1m/Ma, from LeDortz et al., 2011). Considering this erosion rate, the 36Cl ages are rejuvenated because of the high quantity of chlorine in the samples.

92 Chapter III

Table 3. Sample characteristics of 10Be exposure ages Sample Surface Longitude Latitude Elevation 10Be productaion rate 10Be Age Min Age Max (°E) (°N) (m) (atoms/g-SiO2/yr) (Atom/g-SiO2) (ka) (ka) KHA 07-2 Q1 58.335 35.294 1099 9.6767 136221 ± 20923 13.8 ± 2.1 14 ± 2.14 KHA 07-3 Q1 58.337 35.293 1102 9.6983 113362 ± 7388 11.5 ± 0.7 11.6 ± 0.7 KHA 07-6 Q1 58.337 35.293 1101 9.6910 216499 ± 14747 22 ± 1.6 22.4 ± 1.5 KHA 07-7 Q1 58.338 35.293 1102 9.6983 255471 ± 22321 26 ± 2.3 26.5 ± 2.3 AZ II-03 Q1 58.839 35.263 1316 11.3627 246699 ± 11897 21.4 ± 1.0 21.7 ± 1.0 AZ II-04 Q1 58.839 35.264 1320 11.3964 257468 ± 13338 22.2 ± 1.1 22.6 ± 1.2 AZ II-05 Q1 58.839 35.264 1320 11.3964 192020 ± 8701 16.5 ± 0.7 16.8 ± 0.7 AZ II-06 Q1 58.839 35.264 1320 11.3964 156048 ± 12382 13.4 ± 1.1 13.6 ± 1.1 AZ II-07 Q1 58.839 35.264 1320 11.3964 287008 ± 8538 24.8 ± 0.7 25.3 ± 0.7 AZ 10-01 Q2 58.854 35.275 1385 11.9551 527182 ± 15858 43.6 ± 1.3 45.2 ± 1.4 AZ 10-03 Q2 58.854 35.275 1383 11.9378 359963 ± 12688 29.7 ± 1.0 30.4 ± 1.1 AZ 10-08 Q2 58.854 35.276 1389 11.9905 328054 ± 10746 26.9 ± 0.9 27.5 ± 0.9 AZ 10-09 Q2 58.854 35.276 1390 11.9992 609639 ± 18122 50.3 ± 1.5 52.40 ± 1.5 MAZ 07-2 Q2 58.239 35.295 1071 9.4753 331413 ± 43584 34.5 ± 4.5 35.5 ± 4.6 MAZ 07-4 Q2 58.239 35.294 1067 9.4465 502841 ± 46495 52.8 ± 4.9 55.0 ± 5.1 MAZ 07-6 Q2 58.239 35.294 1065 9.4322 329875 ± 20871 34.5 ± 2.2 35.5 ± 2.2 MAZ 07-7 Q2 58.239 35.295 1058 9.3828 337368 ± 64936 35.5 ± 6.8 36.5 ± 7.0 ESM 07-3 Q2 58.666 35.275 1345 11.6103 209169 ± 20816 17.7 ± 1.7 18.0 ± 1.78 ESM 07-6 Q2 58.664 35.274 1334 11.5167 586977 ± 31640 50.5 ± 2.7 52.6 ± 2.8 ESM 07-9 Q3 58.6731 35.281 1361 11.7488 1204036 ± 45597 102.8 ± 3.9 112.1 ± 4.2

Note: In situ-produced 10Be concentration and modeled ages of the surface sample from the alluvial fans, offset by DFS. 10Be concentrations analytical uncertainties (reported as 1σ) include a conservative 0.5% external uncertainty based on long-term measurements of standards, a one sigma statistical error on counted 10Be events, and the uncertainty associated with the chemical blanks correction (Arnold et al., 2010). 10Be concentrations were undertaken at the ASTER-CEREGE facility (Aix-en-Provence, France). Age min is the exposure ages assuming negligible erosion on the fan surface. Age max is the exposures ages assuming a maximum denudation rate (1m/Ma, from LeDortz et al., 2011). The samples collected from the Nay alluvial fan are named by ESM.

4.2. Cosmogenic dating results Our first estimates of cosmogenic exposure ages are based on two assumptions: (1) the exposure time during transport and temporary storage of samples in the drainage basin is negligible, i.e., zero-inheritance, and (2) erosion depth is low, owing to the arid climate condition and a very low uplift due to the predominant strike-slip tectonic regime of the area, i.e., zero-erosion. Given these hypotheses, the minimum exposure ages will be discussed as raw data in this section (Table 2b and 3). A second set of exposure ages were calculated using an erosion depth of 1 m/Ma (see section 4.3) allowing us to determine upper and lower age estimates for the abandonment of alluvial surfaces (Table 2b and 3).

4.2.1. 10Be exposure ages from Q1 surfaces To verify the simultaneous abandonment of Q1 surfaces, two Q1 surfaces were sampled at two ~45-km-apart localities along the CFZ (Figs 3b and 5). In the first site, Azghand fan surface, five samples were analyzed from Quartz-rich granite/granodiorite clast embedded in

93 Chapter III

Figure 5. (a) Azghand and (b) Khalilabad sampling sites on the Q1 fan surfaces. Left panel show the geomorphologic interpretation of Quaternary inset abandonment fan surfaces overlain on GeoEye images (from Google Earth, ©2012 Google, ©2012 GeoEye). The white circles show the position of the surface sampling. Labels are the in-situ produced 10Be exposure age as presented in Table3. Right panel show field photographs of the relatively well-preserved sampled part of the Q1 fan surface. (c) and (d) are the age probability distribution of 10Be exposure ages for the Azghand and Khalilabad sites, respectively. The thick curve corresponds to the Gaussian distribution of summed ages and the thin grey curves represent the age probability for each individual sample. (e) Age probability distribution of all Azghand and Khalilabad samples, showing two populations of ages. Weighted mean ages of 12.3±2.9 and 23.8±4.6 ka are calculated by Chi-square test for each population. The thin blue curve represents the age rejected in chi-square test. The conservative 2σ uncertainties are calculated for the weighted mean ages.

94 Chapter III the fan surface (Fig. 5a). The samples from the Azghand fan surface yield exposure ages range between 13.4±1 and 24.8±0.7 ka (Table 3). The Gaussian age probability distribution of these samples show two populations of the exposure ages: the first one ranging between ~13 and ~16 ka and the second one ranging between ~21 and ~24 ka (Fig. 5c). In the second site, north of Khalilabad village (Fig. 5b), among the collected quartz-rich granitic cobbles, four samples were analyzed which yield an exposure ages range between 11.5±0.7 and 26±2.3 ka (Table 3). The Gaussian age probability distribution on the Khalilabad site shows also two sets of exposure ages: the first one between ~11 and ~14 ka and the second one between ~22 and ~26 ka (Fig. 5d). At first glance, the same age distributions shown by samples collected from the Azghand and Khalilabad fans may suggest the synchronicity of these Q1 alluvial surfaces. However, the sum of the Gaussian age probability distributions (Psum) for all samples shows again two populations of the exposure age ranging between ~11 and ~16 ka for the first population and ~21 to ~26 ka for the second population (Fig. 5e). This distribution could represent a complex exposure history of the dated clasts (see section 4.3).

4.2.2. 10Be and 36Cl exposure ages from Q2 surfaces The Q2 surface was sampled in four localities along the CFZ and WFZ (Fig. 3b). Those are Azghand, Nay, Mazdeh (quartz-rich samples for 10Be dating) and Anabad (carbonate samples for 36Cl dating) alluvial fan surfaces (Fig. 6). Four samples were analyzed from granite/granodiorite clast embedded in the Azghand Q2 fan surface (Fig. 6a) and yield the 10Be exposure ages ranging between 26.9±0.9 and 50.3±1.5 ka (Table 3). The Gaussian age probability distribution (Psum) of the dated samples shows an age population between ~27 and ~30 ka and two older ages at 43.6±1.3 and 50.3±1.5 ka (Fig. 6d). Seven samples were collected from the Nay Q2 surface, ~17 km west of the Azghand Q2 alluvial fan. Only two samples have provided minimum pure quartz masses necessary for the analytical procedure of 10Be dating (Fig. 7a). Those yield different exposure ages of 17.7±1.7 and 50.5±2.7 ka, which are insufficient to deduce a mean exposure age, but, the older age is coherent with the maximum exposure ages estimates in Azghand site. Further west, in Mazdeh fan surface (Fig. 6b), the 10Be exposure ages deduced from four granitic samples indicate ages ranged between 34.5±2.2 and 52.8±4.9 ka (Table 3) that are close to the sample ages from the Azghand Q2 surface.

95 Chapter III

Figure 6. (a) Azghand, (b) Mazdeh and (c) Anabad sampling sites. Top panel show the Q2 sampled fan surface on GeoEye images (from Google Earth, ©2012 Google, ©2012 GeoEye). White circles are the sample positions and labels indicate the 10Be and 36Cl exposure ages and the offset measurements. Middle panel show field photographs of the sampled surface. Bottom panel is the age probability distribution of the in situ-produced 10Be (d and e) and 36Cl (f) exposure ages. Thick black curves are Gaussian probability sum of the ages, and grey thin curves represent the age probability for each individual sample in each site. (g) Age probability distribution of all 10Be and 36Cl samples showing two populations of exposure ages. The weighted mean age of 36.5±6.3 and 50.5±3 ka are calculated by Chi-square test for each population. The thin blue curves represent the ages rejected in chi-square test and the thin red curve represents the outlier sample which has not been included in the total statistics. The conservative 2σ uncertainties are calculated for the weighted mean ages.

96 Chapter III

The Gaussian age probability distribution of these four dated samples indicates a sharp peak at 34.6±1.9 ka and an older age of 52.8±4.9 ka (Fig. 6e). Finally in the westernmost site at Anabad Q2 fan surface, three selected carbonate clasts were dated which yield a 36Cl exposure age, ranging between 38.3±4.2 and 41.3±4.7 ka. The Gaussian age probability distribution shows a single sharp peak at 39.9±2.5 ka (Figs 6c and f). The sum of the Gaussian probability distributions (Psum) of thirteen 10Be and 36Cl exposure ages show again two age populations, considering the ESM07-3 (17.7±1.7 ka) as an outliers (Fig. 6g). The first age population is ranging between ~27 and ~43 ka and the second population is concentrated at ~50 ka. Such an age pattern is similar to the age distribution (regardless the age values) shown by Q1 surfaces (section 4.2.1), confirming the complex exposure history of the clasts embedded in Q1 and Q2 alluvial fan surfaces offset along the DFS (see section 4.3 for the age discussion).

4.2.3. 10Be and 36Cl exposure ages from Q3 surfaces A total 16 quartz-rich and carbonate samples were collected from three Q3 alluvial fan surfaces offset along the CFZ and WFZ (Figs 3b and 7). Among five quartz-rich samples collected from the Nay Q3 alluvial fans (Fig. 7a), only one sample from the Nay fan surface provided minimum pure quartz masses necessary for the analytical procedure of 10Be exposure dating and yield an exposure age at 102.8±3.9 ka (Table 3).

The analysis of four carbonate samples from rare, nearly well-preserved parts of the Anabad Q3 surfaces (Fig. 7b) gave 36Cl exposure ages that vary from 42.4±4.8 to 98.2±11 ka with three ages clustering around 79.8±5.5 ka (Fig. 7d). Three other samples from the Doruneh Q3 fan surface (Fig. 7c) indicate 36Cl exposure ages of 70.2±7.5, 74.2±8.6 and 96±11.5 ka, with weighted mean age at 76.7±5.1 ka (Fig. 7e). The analysis of the Gaussian probability distributions (Psum) of together seven 36Cl and one 10Be exposure ages indicates two sharp peak at 72.7±6.7 and 101.8±9.5 ka by considering the ANA10-01 (42.4±4.8 ka) as an outlier age (Fig. 7f).

97 Chapter III

Figure 7. (a) Nay, (b) Anabad and (c) Doruneh sampling sites. Left panel show the geomorphologic interpretation of Quaternary inset abandonment fan surfaces overlain on GeoEye images (from Google Earth, ©2012 Google, ©2012 GeoEye). White circles are surface sample locations and labels indicate the 10Be and 36Cl exposure ages and the offset measurements. Right panel show field photographs of the sampled Q3 surface. Bottom panel is the age probability distribution of Anabad (d) and Doruneh (e) sites. (f) Age probability distribution of together 10Be and 36Cl samples which show two populations of exposure ages. The weighted mean age of 72.7±6.7 and 101.8.5±9.5 ka are calculated by Chi-square test for each population. The thin red curve represents the outlier sample which has not been included in the total statistics. The conservative 2σ uncertainties are calculated for the weighted mean ages.

98 Chapter III

4.3. Timing of the abandonment of Q1, Q2, and Q3 alluvial surfaces The exposure age distribution of the samples collected from Q1 and Q2 fan surfaces show a bimodal pattern consistent of two populations of exposure age (Figs 5e and 6g). Interestingly, independent Q1 or Q2 surfaces display similar populations in the Gaussian age probability distribution (Psum). Age dispersions could sometimes be due to the sampling strategy and/or analytical errors during laboratory analyses. However, the best preserved parts of the alluvial fans were sampled with a special care, and the independent exposure ages from all Q1 and Q2 alluvial surfaces show a similar and systematic dispersion (section 4.2). These exclude the sampling cause for having two populations of ages. The exposure ages of each fan surface are also examined by the Chi-square test which confirms the presence of two populations in the exposure ages. This test present that this bimodal age population is not due to analytical errors, and should indicates regional causes such as a complex exposure history of clasts embedded in the Q1 and Q2 alluvial surfaces. The sample exposure ages rely on two assumptions of no inheritance and no erosion. These exposure ages would be different from true ages of the alluvial abandonment surface. This is mainly due to decreases in the concentration of cosmogenic nuclides in samples by erosion and/or increases the concentration by pre-deposition exposure of samples, i.e., inheritance (Anderson et al., 1996). In the case of erosion, the younger population of exposure age in a bimodal age distribution, like exposure age patterns of Q1 and Q2 surfaces, represents the exposure age of the alluvial surfaces after a certain depth of erosion that underestimates the true age of the abandonment of the surface. In such a case, the older peak is closer to the true abandonment age of the surface (e.g., Zreda and Phillips, 1994; Phillips et al., 1997; Brown et al., 2005). When erosion is negligible, at least on the most preserved parts of the surface, the older population of exposure age could be attributed to nuclide concentrations inherited from a complex exposure history prior to the last deposition. So, older exposure ages overestimate the true age of the abandonment, while the younger age cluster is close to the true abandonment age of the surface (e.g., Le Dortz et al., 2009; 2011). The knowledge of local erosion rates is needed to distinguish between these two end- member hypotheses. In the southern regions of Central Iran, with the same hyper-arid climatic and strike-slip tectonic regimes, Le Dortz et al. (2009 and 2011) reported a low denudation rate of <1 m/Ma by analyzing cosmogenic 10Be and 36Cl in surface cobbles and near surface amalgams collected from alluvial fan surfaces (abandoned between 10 and 300 ka). We calculated the second set of ages considering the 1 m/Ma maximum erosion rate (Table 2b and 3) in order to discuss the two alternative erosion and inheritance hypotheses. Applying

99 Chapter III this rate of erosion, we obtain exposure ages, which are slightly older than the first set of ages relied on the zero-erosion hypothesis (section 4.2). Nevertheless, the age distributions are similar, and still show a bimodal pattern (Fig. 8). Interestingly, for each fan surface, the younger population of the second set of ages (erosion ≤ 1 m/Ma) is still younger than the older peaks of the first set of ages (zero-erosion). This indicates that the bimodal age distributions should not be resulted from the erosion of surfaces and the upward migration of clasts to the land surface (e.g., Thomas, 1989; Cooke et al., 1993; Wells et al., 1994). Moreover the OSL/IRSL dating along the Doruneh fault by Fattahi et al. (2007), lead to maximum abandonment age of ~10 ka for the younger quaternary fan surface (Q1, this study). This age confirms our approach in which the younger age population is close to the real exposure age of fan surface by considering zero erosion.

Figure 8. Age probability distribution of the Q1 and Q2 fan surfaces, calculated with a denudation rate of 1m/Ma. Distribution shows two populations of exposure ages on both surfaces. The weighted mean ages are calculated by Chi-square test with 2σ uncertainties. The thick curve corresponds to the Gaussian probability sum of the ages and the thin grey curves represent the age probability for each individual sample. The thin blue curves represent the ages rejected in chi-square test and the thin red curve represents represent the outlier samples which have not been included in the total statistics.

When erosion is not significant, the bimodal pattern of exposure ages may be explained by pre-depositional nuclide concentrations in the samples (e.g., Le Dortz et al., 2009 and 2011). This suggestion appears reasonable given following reasons: (1) drainage catchments feeding the alluvial fans are rather large (~20 km2), with gently-slopping long axial streams (7-15 km), which imply a long transport history and complex exposure of the clasts presently embedded in the alluvial surfaces, and (2) almost all younger alluvial surfaces (Q2, Q1) are

100 Chapter III inset in older fan surfaces (Q3, Q2). The inheritance due to a long transport history (the case 1) is likely to be constant or in graduated stages for clasts coming from the same catchment area, while, the formation of inset alluvial fans in successive deposition/erosion stages could explain the bimodal age distributions. That is, the terracing geomorphic configuration of inset surfaces allows clasts, exposed on upper surfaces, to be detached from older alluvial fans, being deposited on lower inset surfaces due to gravitational movements or during occasional flooding periods (e.g., Le Dortz et al., 2009 and 2011; Schmidt et al., 2011; Le Béon et al., 2010; Ritz et al., 2006a). This hypothesis could be examined by comparing exposure ages of casts embedded in two inset alluvial surfaces.

Figure 9. (a) 10Be age probability distribution of Q1 fan surface in Azghand site showing two ages population ranging from ~13 to ~16 ka and ~21 to ~24 ka. (b) 10Be age probability distribution of Q2 abandonment fan surface in Azghand site showing an age population between ~27 and ~30 ka and two older ages of 43.6±1.3 and 50.3±1.5 ka. (c) Age probability distribution of both Q1 and Q2 inset abandonment surfaces. Black curve is the Gaussian probability sum of the ages indicating three populations of exposure age.

101 Chapter III

In the Azghand area, two inset Q1 and Q2 alluvial fan surfaces (section 4.2.1and 4.2.2) were sampled. The exposure ages for the Q1 clasts show two populations of ages: the first one between ~13 and ~16 ka and the second one between ~21 and ~24 ka (Fig. 9a). The clasts from the Q2 surface shows an age set varied between ~27 and ~30 ka (Fig. 9b) and two older ages at 43.6±1.3 and 50.3±1.5 ka (section 4.2.2). The mean age of older Q1 samples (~23 ka) is near to the mean age of younger Q2 samples (~27 ka) suggesting that older samples on the Q1 surface may have come from the Q2 surface. This is statistically confirmed by the Gaussian probability sum of the age distributions for both Q1 and Q2 samples that shows only three populations of ages (Fig. 9c). Accordingly, we retain the second scenario to explain the bimodal distribution of exposure ages suggesting that exposure ages of younger samples are more close to the true age of the well-preserved abandonment surfaces (Fig. 10). In such a case the abandonment ages of 12±3 and 36.5±6 ka are proposed for the Q1 and Q2 fan surfaces, respectively. These weighted mean ages deduced from the Chi-square test of the younger population of ages for each fan surface. The conservative 95% uncertainty is also calculated for the younger age populations.

Figure 10. Upper panel: schematic model of the three abandonment inset surfaces. The surface clasts of upper (older) surface could have fallen to the lower (younger) surface during flooding periods. Lower panel: 10Be and 36Cl exposure ages of the three Q1, Q2 and Q3 fan surfaces along the DFS. Red boxes are the 2σ uncertainties for the weighted mean age of each surface.

Q3 alluvial fan surfaces underwent greater depths of erosion than Q2 and Q1 surfaces, as indicated by drainages more deeply incising the entire surface of Q3 fans, and nearly rounded

102 Chapter III ridges. On the other hand, there is no alluvial surface above the Q3 alluvial surfaces. This reduces considerably the systematic inheritance possibility in Q3 samples, contrast to erosion (see section 4.1), leading us to interpret the exposure ages as minimum age for the Q3 abandonment surfaces. Meanwhile, we incorporate all exposure ages, ranging from 73±7 to 102±10 ka, to estimate a more conservative age for the abandonment of the Q3 fan surfaces. Further new data using more adapted sampling strategies/analytical protocols (e.g., Schimmelpfennig et al., 2009; Schmidt et al., 2011), and integrating other appropriate Quaternary dating methods (e.g., Rizza et al., 2011) would constrain the true age of Q3 surface abandonment.

5. Discussion 5.1. Slip rates along the Doruneh Fault System At sixty seven sites, we measured cumulative left-lateral displacements recorded by either alluvial fan shape incised by streams or terrace risers (Table 1; see also Farbod et al., 2011). In situ-produced 10Be (in quartz-rich clasts) and 36Cl (in carbonate clasts) cosmogenic nuclides were used to determine the age of the alluvial surface abandonment. The age of the abandonment surfaces at the base of the terrace risers were taken to give the age of the risers offsets (e.g., Tapponnier et al., 2001). The age of the offsets recorded by main stream consist with the age of the host abandonment surfaces (e.g., Regard et al., 2005; Regard et al., 2006. Authemayou et al., 2009; Shabanian et al., 2009b). In total, we have determined sixty seven independent slip rates, along the CFZ and WFZ (Table 4 Fig. 11), averaged over ~12 to ~100 ka. At seven sites, Q3 alluvial fans that are cut by the DFS represent the largest cumulative offset at each locality (Table 1). These geomorphic offsets vary between ~850 and ~160 m, from the mid-length towards the ends of the CFZ (Fig. 3b). Farther east, strike-slip faulting dies out at the longitude of 59.497°E. Along the WFZ, there is no counterpart for Q3 alluvial fans to allow such an offset measurement. Nevertheless, lateral offsets in younger Q1 and Q2 geomorphic markers die out at the longitude of 56.749°E indicating the western termination (fault tip) of the WFZ. The age of these abandonment surfaces could varies from ~73 to ~102 ka (section 4.3), although the older age estimate seems more plausible considering the degree of preservation of Q3 surfaces. Applying the age of 102±9.5 Ka to the offsets recorded by Q3 surfaces yields slip rates ranging from 1.6±1 to 8.3±1 mm/yr (Fig. 11a; Table 4). That is, the left-lateral slip

103 Chapter III rate decreases from a maximum of 8.3±1 mm/yr in the middle of the CFZ to a minimum of 1.6±1 mm/yr toward the fault ends (Fig. 11a; Table 4). At twenty four sites along both the CFZ and WFZ, we measured left-lateral offsets recorded by, Q2 alluvial fans (Fig. 3b; Table 1) that are usually inset in older Q3 alluvial fans. Along the CFZ, the cumulative displacements observed in the centre and ends of the fault zone, ranged between 300±50 m and 70±20 m, respectively (Fig. 3b). For the WFZ, the Q2 offsets display a slight variation from 70±10 m, at the junction with the CFZ, to 25±5 m near the western end of the fault zone (Fig. 3b; Table 1).

Table 4. Horizontal slip rate along the DFS estimates over three time periods of ~12 (Q1), ~36 (Q2)and ~100 (Q3) Site Longitude Latitude Offset Age Slip rate Site Longitude Latitude Offset Age Slip rate (°E) (°N) (m) (ka) (mm/yr) (°E) (°N) (m) (ka) (mm/yr) 1 56.749 35.032 0 102 ± 9.5 4 57.317 35.177 15 ± 3 12.3 ± 2.9 1.2 ± 0.4

10 57.492 35.208 160 ± 100 102 ± 9.5 1.6± 1.0 5 57.44 35.201 12 ± 2 12.3 ± 2.9 1.0 ± 0.3

34 58.290 35.297 630 ± 70 102 ± 9.5 6.2 ± 0.9 8 57.473 35.206 25 ± 5 12.3 ± 2.9 2.0 ± 0.6

49 58.570 35.281 840 ± 70 102 ± 9.5 8.3 ± 1.0 9 57.481 35.207 15 ± 5 12.3 ± 2.9 1.2 ± 0.5

Q3 56 58.921 35.253 330 ± 80 102 ± 9.5 3.2 ± 0.8 12 57.5 35.209 29 ± 5 12.3 ± 2.9 2.4 ± 0.7

59 58.932 35.251 350 ± 70 102 ± 9.5 3.4 ± 0.8 13 57.507 35.212 20 ± 10 12.3 ± 2.9 1.6 ± 0.9

65 59.115 35.222 290 ± 50 102 ± 9.5 2.8 ± 0.6 14 57.519 35.213 27 ± 6 12.3 ± 2.9 2.2 ± 0.7

68 59.331 35.174 170 ± 30 102 ± 9.5 1.7 ± 0.3 18 57.558 35.22 18 ± 4 12.3 ± 2.9 1.5 ± 0.5

69 59.497 35.127 0 102 ± 9.5 0 20 57.58 35.225 13 ± 3 12.3 ± 2.9 1.1 ± 0.3

1 56.749 35.032 0 36.5 ± 6.3 0 21 57.582 35.225 14 ± 2 12.3 ± 2.9 1.1 ± 0.3

2 56.766 35.038 25 ± 5 36.5 ± 6.3 0.7 ± 0.2 23 57.615 35.236 25 ± 3 12.3 ± 2.9 2.0 ± 0.5

6 57.440 35.201 60 ± 10 36.5 ± 6.3 1.6 ± 0.4 24 57.864 35.276 50 ± 20 12.3 ± 2.9 4.1 ± 1.9

7 57.454 35.204 50 ± 10 36.5 ± 6.3 1.4 ± 0.4 25 58.145 35.294 55 ± 5 12.3 ± 2.9 4.5 ± 1.1

11 57.469 35.205 70 ± 10 36.5 ± 6.3 1.9 ± 0.4 26 58.165 35.297 65 ± 15 12.3 ± 2.9 5.3 ± 1.7

15 57.528 35.214 90 ± 30 36.5 ± 6.3 2.5 ± 0.9 28 58.233 35.297 40 ± 10 12.3 ± 2.9 3.3 ± 1.1

16 57.537 35.215 100 ± 20 36.5 ± 6.3 2.7 ± 0.7 30 58.269 35.297 25 ± 5 12.3 ± 2.9 2.0 ± 0.6

17 57.549 35.217 110 ± 30 36.5 ± 6.3 3.0 ± 1.0 32 58.282 35.297 30 ± 10 12.3 ± 2.9 2.4 ± 1.0

19 57.571 35.223 130 ± 15 36.5 ± 6.3 3.6 ± 0.7 35 58.304 35.295 50 ± 10 12.3 ± 2.9 4.1 ± 1.3 Q1 22 57.584 35.226 140 ± 40 36.5 ± 6.3 3.8 ± 1.3 36 58.318 35.295 60 ± 20 12.3 ± 2.9 4.9 ± 2.0

27 58.194 35.298 230 ± 20 36.5 ± 6.3 6.3 ± 1.2 37 58.338 35.295 60 ± 10 12.3 ± 2.9 4.9 ± 1.4

29 58.245 35.297 190 ± 20 36.5 ± 6.3 5.2 ± 1.1 40 58.38 35.289 40 ± 10 12.3 ± 2.9 3.3 ± 1.1

31 58.271 35.297 200 ± 70 36.5 ± 6.3 5.5 ± 2.1 41 58.395 35.288 10 ± 2 12.3 ± 2.9 0.8 ± 0.3 Q2 33 58.284 35.297 220 ± 70 36.5 ± 6.3 6.0 ± 2.2 42 58.4 35.287 0 12.3 ± 2.9 0

38 58.343 35.295 200 ± 50 36.5 ± 6.3 5.5 ± 1.7 43 58.442 35.286 0 12.3 ± 2.9 0

39 58.373 35.291 180 ± 20 36.5 ± 6.3 4.9 ± 1.0 44 58.443 35.286 13 ± 3 12.3 ± 2.9 1.1 ± 0.3

48 58.553 35.284 300 ± 50 36.5 ± 6.3 8.2 ± 2.0 45 58.46 35.287 17 ± 2 12.3 ± 2.9 1.4 ± 0.4

51 58.669 35.278 300 ± 50 36.5 ± 6.3 8.2 ± 2.0 46 58.5 35.286 27 ± 10 12.3 ± 2.9 2.2 ± 1.0

54 58.855 35.263 100 ± 20 36.5 ± 6.3 2.7 ± 0.7 47 58.549 35.285 25 ± 5 12.3 ± 2.9 2.0 ± 0.6

58 58.932 35.251 210 ± 20 36.5 ± 6.3 5.8 ± 1.1 50 58.66 35.278 60 ± 10 12.3 ± 2.9 4.9 ± 1.4

61 58.965 35.247 190 ± 30 36.5 ± 6.3 5.2 ± 1.2 52 58.76 35.271 55 ± 5 12.3 ± 2.9 4.5 ± 1.1

63 59.009 35.242 190 ± 30 36.5 ± 6.3 5.2 ± 1.2 53 58.76 35.271 55 ± 5 12.3 ± 2.9 4.5 ± 1.1

64 59.049 35.235 210 ± 50 36.5 ± 6.3 5.8 ± 1.7 55 58.873 35.261 60 ± 10 12.3 ± 2.9 4.9 ± 1.4

66 59.131 35.220 170 ± 50 36.5 ± 6.3 4.7 ± 1.6 57 58.926 35.252 40 ± 10 12.3 ± 2.9 3.3 ± 1.1

67 59.323 35.175 70 ± 20 36.5 ± 6.3 1.9 ± 0.6 60 58.954 35.248 40 ± 5 12.3 ± 2.9 3.3 ± 0.9

69 59.497 35.127 0 36.5 ± 6.3 0 62 59.009 35.242 50 ± 10 12.3 ± 2.9 4.1 ± 1.3

1 56.749 35.032 0 12.3 ± 2.9 0 69 59.497 35.127 0 12.3 ± 2.9 0 Q1 3 57.082 35.129 16 ± 4 12.3 ± 2.9 1.3 ± 0.4

Applying the 36.5±6.3 ka abandonment age of the Q2 surfaces to these cumulative offsets, we have estimated slip rates that vary between 1.9±0.6 and 8.2±1.9 mm/yr (Fig. 11b and Table 4), with the maximum rate in the mid-length of the CFZ (Fig. 11b). Interestingly, the spatial distribution of these rates follows the same pattern as the Q3-related slip rates. At the

104 Chapter III junction between the CFZ and WFZ, the minimum slip rate (1.9±0.4 mm/yr) on the CFZ corresponds to the maximum slip rate along the WFZ (Fig. 11b). This maximum slip rate decreases to 0.7±0.2 mm/yr close to the western termination of the fault zone (Fig. 11b). Investigating all Q1 alluvial fans along the DFS, we were able to measure thirty six left- lateral offsets ranging between 10±2 and 65±15 m (Fig. 3b; Table 1). Suggesting the Q1 surface abandonment age at 12.3±2.9 ka (Fig. 11c; Table 4), we obtained slip rates ranging between 1±0.3 and 5.3±1.7 mm/yr (Fig. 11c). The overall trend of variation in the slip rates averaged over ~12 ka is similar to the rates averaged over the longer periods of ~36 and ~100 ka. However, the Holocene slip history of the CFZ seems slightly different from its slip behavior during a longer time scale. Considering the distribution of slip recorded by Q1 surfaces, two distinct fault segments are recognized (Fig. 11c). Those are separated by a ~3 km-long gap along which there is no evidence of faulting in Q1 surfaces (Fig. 4). The slip distribution on a fault segment is nearly mirrored in the other segment showing the same minimum (~1 mm/yr) and maximum (~5 mm/yr) slip rates (Fig. 11c; Table 4). The three independent sets of slip rate, averaged over ~12, ~36 and ~100 ka (abandonment surface ages of Q1, Q2 and Q3, respectively), show nearly similar overall distributions of slip along both the WFZ and CFZ (Fig. 11d). Given the uncertainty of slip rate estimates, we suggest that the rate of left-lateral slip along the WFZ and CFZ has remained constant, at least since the Late Pleistocene. However, at a shorter time scale, the Holocene slip distribution significantly varies along the CFZ. That is, the CFZ is divided into the 80- and 100-km-long west and east segments, respectively, along which the slip distribution is identical during the Late Pleistocene and Holocene time spans, while the boundary between these segments has remained persistent during at least the Holocene (Fig. 11c). This gap of faulting in the mid-length of the CFZ contrasts the sharp peak of slip that has been recorded during the Late Pleistocene. The inconsistency of the Late Pleistocene and Holocene slip distribution may be explained by a complex slip behavior of the fault zone during the shorter time period, or a change in the fault slip behavior since the Holocene. Considering the constant overall slip rate and the long-term characteristic behavior of the fault zone since the late Pleistocene (Fig. 11d), we favor the first scenario, suggesting that the persistent boundary of the fault segments could be ruptured due to the sequential shift of slip towards the centre of the fault zone in the time intervals as large as ~12 ka. However, more detailed paleoseismological data are needed to opt for each hypothesis.

105 Chapter III

Figure 11. (a), (b) and (c) Overall slip rate distribution along the central and western part of the DFS, recorded by the Q3, Q2 and Q1 fan surfaces, respectively. (d) Along-strike variation of the slip rates estimated from all abandonment surfaces over three time periods of ~12, ~36 and ~100 ka.

106 Chapter III

Whatever the model, the discrepancy of the Late Pleistocene and Holocene slip distributions does not favor significant creep motion along the CFZ. The fact that the Holocene deformation is accommodated by two distinct segments separated by a persistent boundary has also important implications in the evaluation of seismic behavior of the CFZ (see section 5.2). Along the DFS, Fattahi et al. (2007) estimated a Holocene slip rate of ~2.5 mm/yr. However this estimate is, based on an apparent offset in parallel terrace risers at the left bank of the Shesh Taraz river, which are disordered due to left-lateral displacement of the fault (section 4.2; see also Farbod et al., 2011). This is, thus, far from being representative of the Holocene slip rate of the DFS. Moreover, the slip rate of the strike-slip faults with the non uniform slip distribution (e.g., Shabanian et al., 2009b; this study) cannot be determined using only a single estimate except for specific sites representing the maximum cumulative offset along the faults. Farbod et al. (2011) have estimated a short-term left-lateral slip rate of ~2.5 mm/yr using the present-day GPS velocities of three KASH, SHIR and MSHN stations in northeast Iran (Vernant et al., 2004; Masson et al., 2007). This estimate is based on the assumption that the deformation accommodated between the DFS and SHIR station is not significant. Interestingly, the KASH station located within the fault zone close to the persistent segment boundary, where the Holocene slip rate of the CFZ decreases to ~2 mm/yr (Fig. 11c). So, this is not surprising that the ~2.5 mm/yr GPS-derived rate is two times slower than the 5.3 mm/yr slip rates of the fault. In addition, in an elastic rebound model, the KASH station, in the best condition, could record only a half of the interseismic elastic deformation being accommodated across the northern fault block, because the farthest GPS station is located from the fault zone the highest velocity is recorded.

5.2. Seismotectonic implications of the fault slip rates Our data and results improve the segmentation model of the DFS proposed by Farbod et al. (2011) by the precise determination of (1) the boundaries between the WFZ, CFZ and EFZ fault zones, and (2) the persistent boundary between independent eastern and western segment of the CFZ. The systematic measurement of left-lateral offsets recorded by Late Pleistocene to Holocene landforms led us to conclude that the overall left-lateral slip rate of the WFZ and CFZ has remained nearly constant since at least the Late Pleistocene. For a given time period, the slip rate varies spatially along the strike of the fault zones (Fig. 11d), but the overall distribution of slip for the three time periods of ~12, ~36, and ~100 ka is similar. This

107 Chapter III illustrates a long-term characteristic slip model (c.f., Schwartz and Coppersmith, 1984) for the western and central zones of the DFS. Accepting to this suggestion, one may expect that (1) the displacement per event remains constant at each locality on the fault, and (2) the size of large earthquakes is nearly constant with infrequent moderate events. Both the instrumental and historical records of earthquakes indicate that the region affected by the DFS experienced a low to moderate seismicity when compared to nearby seismic regions such as the Kopeh Dagh and the northern margin of the Lut block (see Farbod et al., 2011, and references therein). This may be principally due to the fact that the fault has been far from historical trade routes implying a non-complete record of large historical earthquakes, which probably occurred on and around the fault. At the first stage to understand the seismogenic behavior of the DFS, Farbod et al. (2011) have established a fault segmentation model at the scale of fault zones. In this study, considering the Holocene slip distribution on the CFZ, two eastern and western fault segments separated by a persistent boundary are recognized (section 5.1). The 100-km-long eastern fault segment slips at 5.3±3 mm/yr, and could produce characteristic events with magnitude of Mw≈7.4. Along this fault segment, Farbod et al. (2011) have presented a coseismic offset of ~4 m recorded by a kariz line (underground irrigation system; qanat). The offset locates at the longitude of 58.77°E, where the peak of Holocene cumulative slip on the fault is observed (Fig. 11c), and may consist with the maximum coseismic displacement along the fault. Such an offset could also be produced by an event as large as Mw≈7.4, and reinforce the hypothesis of characteristic slip of the fault. The ~5.3 mm/yr slip rate of the fault segment and the ~4m characteristic coseismic displacement per event allow us to calculate the recurrence interval of ~750 years for such Mw 7.4 earthquakes. The similar (~630-1400 yr) recurrence interval, is proposed by Pezzo et al. (2012) through the InSar interferometry analyses. Our estimation is shorter than the upper bound of the ~2000 years recurrence time proposed by Farbod et al. (2011), which was based on the maximum possible age of ~4 ka (Wulff, 1968; Berberian and Yeats, 1999) for the offset kariz line. The ~80-km-long western segment of the CFZ slips at ~5.3 mm/yr, and is able to produce large characteristic earthquakes of Mw≈7.3, with the recurrence time of probably ~650 years. This suggestion seems consistent with the average coseismic displacements between 2.6 and 3.6 m occurred along the fault segment. In the same way, for the ~70-km-long WFZ, large earthquakes with a maximum magnitude of Mw≈7.2 are suggested. However, the related slip rates represent only the strike-slip part of deformation that is accommodated along this reverse-sinistral, oblique-slip fault zone (Farbod et al., 2011). The lack of knowledge of

108 Chapter III vertical displacement rate along the WFZ has precluded us to estimate the total slip rate, and consequently, a possible recurrence time interval for such large events. Nevertheless, these results are based on the relatively long-term slip distribution of the DFS and should be considered as the first approximation of the seismic behavior of the DFS. Our detailed geomorphic data pave the way for further detailed paleoseismological investigations like trenching, which would help in the better understanding of the unclear seismic behavior of the largest strike-slip fault in the Iranian plateau.

6. Conclusion We evaluated the slip distribution along the DFS through the systematic measurement of sixty seven left-lateral offsets recorded by late Quaternary alluvial fan morphologies (i.e., fan surfaces and associated geomorphic features). At six sites, abandonments of the three generations of Q1 (12.3±2.9 ka), Q2 (36.5±6.3 ka), and Q3 (101.8±9.5 ka) inset alluvial surfaces were determined using in situ-produced 36Cl and 10Be cosmogenic nuclides. For the inset Q1 and Q2 surfaces, we found considerable amount of inheritance in the cosmogenic nuclides concentration due to clasts reworked from the upper alluvial fans, while erosion seems negligible. The relevant ages and geomorphic offsets allowed us to estimate sixty seven independent left-lateral slip rates, characterizing the slip distribution along the CFZ and WFZ during the Late Pleistocene and Holocene. The similar slip distributions during three ~100, ~36, and ~12 ka time spans imply the long-term characteristic slip behavior of the fault zones. Along the CFZ, a Late Pleistocene overall slip rate of 8.3±1 mm/yr is estimated, while over the shorter Holocene time span, the fault zone is divided into two fault segments, with symmetrical slip distribution, that slip at 5.3±1.7 mm/yr. The boundary between these segments remains persistent at least during the last ~12 ka, and could be ruptured during a future large earthquake (Mw ≥7) along the fault. The rate of 1.9±0.4 mm/yr is estimated for left-lateral slip on the reverse-sinistral WFZ. However, the vertical slip rate of the fault zone remains unknown. Our results reconfirm that the strike-slip fault, particularly long ones, do not rupture along their entire length during a single earthquake (e.g., King and Nabelek, 1985; Barka and Kadinsky-Cade, 1988; Knuepfer, 1989; Schwartz and Sibson, 1989; DePolo et al., 1991; McCalpin, 1996; Stewart and Taylor, 1996; Barka et al., 2002; Kondo et al., 2005; Ikeda et al., 2009). As for the DFS, the seismogenic segment lengths vary between ~70 and ~100 km, which are able to produce characteristic earthquakes with magnitudes of Mw≈7.2-7.4 with the recurrence interval of ~750 years.

109 Chapter III

Acknowledgement This work was funded by the INSU-CNRS (France) and the International Institute of Earthquake Engineering and Seismology (IIEES, Iran). Funding was provided by the Dyeti and PNRN programs (INSU-CNRS), as well as ACI FNS program (French Ministry of Research). SPOT images were provided thanks to the ISIS program (©CNES 2007 to 2008, distribution SPOT images S.A.). We thank V. Grimault, D. Lambert, P.A. Lhôte and the staff of the SCAC of the French Embassy in Tehran, for their support. The measurement of 10Be and 36Cl cosmogenic nuclide concentrations, performed at the ASTER AMS national facility (CEREGE, Aix en Provence), are supported by the INSU/CNRS, the French Ministry of Research and Higher Education, IRD and CEA. The ASTER team is acknowledged for the measurements (M. Arnold, G. Aumaître) and their help (L. Leanni, V. Guillou, F. Chauvet) to Y. F. during chemical analyses.

110 Chapter III

References Alavi-Naini, M., M. J. Vaezi-Pour, N. Alavi Tehrani, A. Behrouzi, and M. H. Kholghi (1992), Geological Map of Torbat-e Heidariyeh, 1:250 000, Geological survey of Iran. Anderson, R. S., J. L. Repka, and G. S. Dick (1996), Explicit treatment of inheritence in dating depositional surfaces using insitue 10Be and 26Al, Geology, 96, 47-51. Arnold, M., S. Merchel, D. L. Bourlès, R. Braucher, L. Benedetti, R. C. Finkel, G. Aumaître, A. Gottdang, and M. Klein (2010), The French Accelerator Mass Spectrometry Facility ASTER: Improved Performance and Developments, Nuclear Instruments and Methods in Physics Research Section B, Beam Interactions with Materials and Atoms, 268, 1954- 1959. Authemayou, C., D. Chardon, O. Bellier, Z. Malekzadeh, E. Shabanian, and M. R. Abbassi (2006), Late Cenozoic partitioning of oblique plate convergence in the Zagros fold-and- thrust belt (Iran), Tectonics, 25(TC3002), doi: doi:10.1029/2005TC001860. Authemayou, C., O. Bellier, D. Chardon, L. Benedetti, Z. Malekzade, C. Claude, B. Angeletti, E. Shabanian, and M. R. Abbassi (2009), Quaternary slip-rates of the Kazerun and the Main Recent Faults: active strike-slip partitioning in the Zagros fold-and-thrust belt, Geophys. J. Int., 178, 524-540, doi: 10.1111/j.1365-246X.2009.04191.x. Barka, A. A., and K. Kadinsky-Cade (1988), Strike-slip fault geometry in Turkey and its influence on earthquake activity, Tectonics, 7, 663–684. Barka, A., et al. (2002), The surface rupture and slip distribution of the 17 August 1999 I˙zmit earthquake, (M 7.4), North Anatolian fault, Bull. Seismol. Soc. Am., 92(1), 43-60. Berberian, M., and R. Yeats (1999), Patterns of Historical Earthquake rupture in the Iranian Plateau, Bull. Seismol. Soc. Am., 89(1), 120-139. Bierman, P. R., A. R. Gillespie, M. W. Caffee, and D. Elmore (1995), Estimating erosion rates and exposure ages with 36Cl produced by neutron activation, Geochimica et Cosmochimica Acta, 59, 3779-3798. Brown, E. T., J. M., Edmond, G. M., Raisbeck, F., Yiou, M. D. Kurz, and E. J. Brook (1991), Examination of surface exposure ages of Antarctic moraines using in-situ produced 10Be and 26Al, Geochimica et Cosmochimica Acta, 55, 2269-2283. Brown, E. T., P. Molnar, and B. D.L (2005), Slip-Rate Measurements on the Karakorum Fault May Imply Secular Variations in Fault Motion, Science, 309(5739), doi: 10.1126/science.1112508. Chmeleff, J., F. Blanckenburg, K. Kossert, and D. Jakob (2010), Determination of the 10Be half-life by multicollector ICP-MS and liquid scintillation counting, Nuclear Instruments and Methods in Physics 856 Research Section B, Beam Interactions with Materials and Atoms, 268(2), 192-199. Cooke, R. U., A. Warren, and A. Goudie (1993), Desert Geomorphology, UCL Press, London. Deino, A. and R. Potts (1992), Age-probability spectra from examination of single-crystal 40Ar/39Ar dating results: Examples from Olorgesailie, Southern Kenya Rift, Quaternary International, 13/14, 47-53. dePolo, C. M., D. G. Clark, D. B. Slemmons, and A. Ramallie (1991), Historical basin and range province surface faulting and fault segmentation, J. Struct. Geol., 13, 123-136. Desilets, D., Zreda, M., Almasi, P.F. & Elmore, D., (2006), Determination of cosmogenic 36Cl in rocks by isotope dilution: innovations, validation and error propagation, Chem. Geol., 233, 185–195. Djamour, Y., et al. (2010), GPS and gravity constraints on continental deformation in the Alborz mountain range, Iran, Geophys. J. Int., 1287-1301, doi: 10.1111/j.1365- 246X.2010.04811.x.

111 Chapter III

Dunaï, T. J. (2010), Cosmogenic Nuclides: Principles, Concepts and Applications in the Earth Surface Sciences, Cambridge University Press, New York, ISBN-13 978-0-521-87380-2. Dunne, J., D. Elmore, and P. Muzikar (1999), Scaling factors for the rates of production of cosmogenic nuclides for geometric shielding and attenuation at depth on sloped surfaces, Geomorphology, 27(1-2), 3-11. Eftekhar-Nezhad, J., A. Aghanabati, B. Hamzehpour, and V. Baroyant (1976), Geological map of Kashmar Quadrangle. 1; 250 000, Geological survey of Iran. Farbod, Y., O. Bellier, E. Shabanian, and M. R. Abbassi (2011), Geomorphic and structural variations along the Doruneh Fault System (Central Iran), Tectonics, 30, TC6014, doi:10.1029/2011TC002889. Fattahi, M., R. T. Walker, M. M. Khatib, A. Dolati and A. Bahroudi; (2007), Slip-rate estimate and past earthquakes on the Doruneh fault, eastern Iran, Geophys. J. Int. 168, 691–709 doi: 10.1111/j.1365-246X.2006.03248.x Gosse, J. C., and F. M. Phillips (2001), Terrestrial in situ cosmogenic nuclides: theory and application, Quaternary Science Reviews, 20, 1475-1560. Hessami, K., F. Jamali, and H. Tabassi (2003), Major active faults of Iran, scale 1:2 500 000, International Institute of Earthquake Engineering and Seismology, Iran, -Tehran. Hessami, K., Nilforoushan, F., and Talbot, C. (2006), Active deformation within the Zagros Mountains deduced from GPS measurements, J. Geol. Soc. Lond., 163, 143–148. Ikeda, M., S. Toda, S. Kobayashi, Y. Ohno, N. Nishizaka, and I. Ohno (2009), Tectonic model and fault segmentation of the Median Tectonic Line active fault system on Shikoku, Japan, Tectonics, 28, TC5006, doi: 10.1029/2008TC002349. Jackson, J., J. Haines, and W. Holt (1995), The accomodation of Arabia-Eurasia plate convergenec in Iran, J. Geophys. Res., 100(B8), 15205-15219. Kaviani, A., D. Hatzfeld, A. Paul, M. Tatar, and K. Priestley (2009), Shear-wave splitting, lithospheric anisotropy, and mantle deformation beneath the Arabia-Eurasia collision zone in Iran, Earth and Planetary Sciences, 286, 371-378, doi: 10.1016/j.epsl.2009.07.003. King, G. C. P., and J. Nabelek (1985), Role of fault bends in the initiation and termination of earthquake rupture, Science, 228, 984-987. Knuepfer, P. L. K. (1989), Implication of characteristics of end-points of historical surface fault ruptures for the nature of fault segmentation, U.S. Geol. Surv. Open-File Rep, 193- 228. Kondo, H. et al. (2005), Slip distribution, fault geometry, and fault segmentation of the 1944 Bolu-Gerede earthquake rupture, North Anatolian Fault, Turkey. Bull. Seismol. Soc. Am. 95, 1234-1249. Korschinek, G., A. Bergmaier, T. Faestermann, U. C. Gerstmann, K. Knie, G. Rugel, A. Wallner, I. Dillmann, G. Dollinger, Ch. Lierse von Gostomski, K. Kossert, M. Maiti, M. Poutivtsev, and A. Remmert (2010), A new value for the half-life of 10Be by Heavy-Ion 39 Elastic Recoil Detection and liquid scintillation counting: Nuclear Instruments and Methods in Physics Research Section B, Beam Interactions with Materials and Atoms, 268(2), 187-191. Le Béon, M., Y. Klinger, M. AlQaryouti, A. S. Mériaux, R. C. Finkel, A. Elias, O. Mayyas, F. J. Ryerson, and P. Tapponnier (2010), Early Holocene and Late Pleistocene slip rates of the southern Dead Sea Fault determined from 10Be cosmogenic dating of offset alluvial deposits, J. Geophys. Res., 115, B11414, doi:10.1029/2009JB007198. Le Dortz, K. et al., (2009), Holocene right-slip rate determined by cosmogenic and OSL dating on the Anar fault, Central Iran, Geophys. J. Int., 179, 700–710 doi:10.1111/j.1365- 246X.2009.04309.x.

112 Chapter III

Le Dortz, K., B. Meyer, M. Sébrier, R. Braucher, H. Nazari, L. Benedetti, M. Fattahi, D. Bourlès, M. Foroutan, L. Siame, A. Rashidi and M. D. Bateman (2011), Dating inset terraces and offset fans along the Dehshir Fault (Iran) combining cosmogenic and OSL methods, Geophys. J. Int. 1-28, doi: 10.1111/j.1365- 246X.2011.05010.x. Masson, F., M. Anvari, Y. Djamour, A. Walpersdorf, 930 F. Tavakoli, M. Daignières, H. Nankali, and S. Van Gorp (2007), Large-scale velocity field and strain tensor in Iran inferred from GPS measurements: new insight for the present-day deformation pattern within NE Iran, Geophys. J. Int. 170, 436-440, doi:10.1111/j.1365-246X.2007.03477.x. McCalpin, J. P. (Eds) (1996), Paleoseismology, Academic, San Diego, California. McClusky, S., R. Reilinger, S. Mahmoud, D. Ben Sari, and A. Tealeb (2003), GPS constraints on Africa (Nubia) and Arabia plate motions, Geophys. J. Int. 155(1), 126-138. Merchel, S., and U. Herpers (1999), An update on radiochemical separation techniques for the determination of long-lived radionuclides via accelerator mass spectrometry, Radiochimica Acta, 84, 215-219. Merchel, S., M. Arnold, G. Aumaître, L. Benedetti, D. L. Bourlès, R. Braucher, V. Alfimov, S.P.H.T. Freeman, P.Steier, and A. Wallner (2008), Towards more precise 10Be and 36Cl data from measurements at the 10-14 level: Influence of sample preparation, Nuclear Instruments and Methods in Physics Research Section B: Beam Interactions with Materials and Atoms, 266(22), 4921-4926. Nishiizumi K., M. Imamura, M. W. Caffee, J. R. Southon, R. C. Finkel, and J. McAninch (2007), Absolute calibration of 10Be AMS standards, Nuclear Instruments and Methods in Physics Research, B258, 403-413. Pezzo, G., C. Tolomei, S. Atzori, S. Salvi, E. Shabanian, O. Bellier, and Y. Farbod (2012), New kinematic constraints of the western Doruneh fault, north-eastern Iran, from interseismic deformation analysis, Geophys. J. Int. in press. Philip, H., A. Avagyan, A. Karakhanian, J.-F. Ritz, and S. Rebai (2001), Estimating slip rates and recurrence intervals for strong earthquakes along an intracontinental fault: example of the Pambak-Sevan-Sunik fault (Armenia), Tectonophysics, 343, 205-232. Phillips, F. M., M. G. Zreda, J. C. Gosse, J. Klein, E. Evenson, B. , R. D. Hall, O. A. Chadwick , and P. Sharma (1997), Cosmogenic 36Cl and 10Be ages of Quaternary glacial and fluvial deposits of the Wind River Range,Wyoming, GSA Bulletin, 109(11), 1453- 1463. Regard, V., et al. (2005), Cumulative right-lateral fault slip rate across the Zagros–Makran transfer zone: role of the Minab–Zendan fault system in accommodating Arabia–Eurasia convergence in southeast Iran, Geophys. J. Int., 162, 177-203, doi: 10.1111/j.1365- 246X.2005.02558.x. Regard, V., O. Bellier, R. Braucher, F. Gasse, D. Bourlès, J. Mercier, J.-C. Thomas, M. R. Abbassi, E. Shabanian, and S. Soleymani (2006), 10Be dating of alluvial deposits from Southeastern Iran (the Hormoz Strait area), Palaeogeography, Palaeoclimatology, Palaeoecology, 242, 36-53, doi: 10.1016/j.palaeo.2006.05.012. Reilinger, R., et al. (2006), GPS constraints on continental deformation in the Africa-Arabia- Eurasia continental collision zone and implications for the dynamics of plate interactions, J. Geophys. Res., 111, B05411, doi: 10.1029/2005JB004051. Ritz, J. F., H. Nazari, A. Ghassemi, R. Salamati, A. Shafei, S. Soleymani, and P. Vernant (2006), Active transtension inside central Alborz: A new insight into northern Iran– southern Caspian geodynamics, Geology, 34(6), 477-480, doi: 10.1130/G22319.1. Ritz, J. F., R. Vassallo, R. Braucher, E. T. Brown, S. Carretier, and D. L. Bourlès (2006a), Using in situ–produced 10Be to quantify active tectonics in the Gurvan Bogd mountain range (Gobi-Altay, Mongolia), in Siame, L. L., D. L. Bourlès, and E. T. Brown, eds., Application of cosmogenic nuclides to the study of Earth surface processes: The practice

113 Chapter III

and the potential, Geological Society of America Special Paper, 415, 87-110, doi: 10.1130/2006.2415(06). Ritz, J.F., Brown, E.T., Bourlès, D.L., Philip, H., Schlupp, A., Raisbeck, G.M., Yiou, F., Enkhtuvshin, B., (1995), Slip rates along active faults estimated with cosmic-ray- exposure dates: application to the Bogd fault, Gobi-Altaï, Mongolia, Geology, 23, 1019e1022. Ritz, J.-F., et al. (2003), Late Pleistocene to Holocene slip rates for the Gurvan Bulag thrust fault (Gobi-Altay, Mongolia) estimated with 10Be dates, J. Geophys. Res., 108(B3), doi: 10.1029/2001JB000553. Rizza, M., Mahan, S., Ritz, J.-F., Nazari, H., Hollingsworth, J., and Salamati, R., (2011), Using luminescence dating from coarse matrix material to estimate fault slip-rate in arid domain: Example of the Astaneh Fault (Iran), Quaternary Geochronology, 6(3-4), 390- 406. Schimmelpfennig, I., Benedetti, L., Finkel, R., Pik, R., Blard, P-H., Bourl`es, D., Burnard, P. & Williams, A., (2009), Sources of in-situ 36Cl in basaltic rocks. Implications for calibration of production rates, Quat.Geoch., 4(6), 441–461. Schimmelpfennig, I., L. Benedetti, V. Garreta, R. Pik, P.-H. Blard, P. Burnard, D. Bourles, R. Finkel, K. Ammond, and T. Dunai (2011), Calibration of cosmogenic 36Cl production rates from Ca and K spallation in lava flows from Mt. Etna (38 N, Italy) and Payun Matru (36 S, Argentina), Geochimica et Cosmochimica Acta, 75(10), doi: 10.1016/j.gca.2011.02.013. Schmidt, S., R. Hetzel, J. Kuhlmann, F. Mingorance, and V. A. Ramos (2011), A note of caution on the use of boulders for exposure dating of depositional surfaces, Earth and Planetary Science Letters, 302, 60-70. Schwartz, D. P., and K. J. Coppersmith (1984), Fault Behavior and Characteristic Earthquakes' Examples From the Wasatch and San Andreas Fault Zones, J. Geophy. Res., 89(B7), 5681-5698. Schwartz, D.P. and Sibson, R.H. (Eds) (1989), Fault segmentation and controls of rupture initiation and termination, USGS Open-File Report, 89-315, Menelo park, California. Shabanian, E., L. Siame, O. Bellier, L. Benedetti, and M. R. Abbassi (2009b), Quaternary slip rates along the northeastern boundary of the Arabia–Eurasia collision zone (Kopeh Dagh Mountains, Northeast Iran), Geophys. J. Int. 1-23, doi: 0.1111/j.1365- 246X.2009.04183.x. Shabanian, E., O. Bellier, L. Siame, N. Arnaud, M. R. Abbassi, and J.-J. Cocheme (2009a), New tectonic configuration in NE Iran: Active strike-slip faulting between the Kopeh Dagh and Binalud mountains, Tectonics, 28, TC5002, doi: 10.1029/2008TC002444. Siame, L. L., D. L. Bourlès, M. Sebrier, O. Bellier, J. C. Castano, M. Araujo, M. Perez, G. M. Raisbeck, and F. Yiou (1997), Cosmogenic dating ranging from 20 to 700 ka of a series of alluvial fan surfaces affected by the El Tigre fault, Argentina, Geology, 25(11), 975- 978. Solaymani Azad, S., J.-F. Ritz, and M. R. Abbassi (2011), Left-lateral active deformation along the Mosha-North Tehran fault system (Iran):Morphotectonics and paleoseismological investigations, Tectonophysics, 497, 1-14, doi: 10.1016/j.tecto.2010.09.013. Stewart, M. E., and W. J. Taylor (1996), Structural analysis and fault segment boundary identification along the Hurricane fault in southwestern Utah, J. Struct. Geol., 18(8), 1017-1029. Stone J.O., (2000), Air pressure and cosmogenic isotope production, J. Geophys. Res., 105(B10), 23 753–23 759.

114 Chapter III

Stone, J.O., Evans, J.M., Fifield, L.K., Allan, G.L. & Cresswell, R.G., (1998), Cosmogenic chlorine-36 production in calcite by muons, Geochim. Cosmochim. Acta, 62, 433–454. Stone, J.O.H., Allan, G.L., Fifield, L.K. & Cresswell, R.G., (1996), Cosmogenic chlorine-36 from calcium spallation, Geochim. Cosmochim. Acta, 60, 679–692. Talebian, M., and J. Jackson (2002), Offset on the Main Recent Fault of NW Iran and implications for the late Cenozoic tectonics of the Arabia–Eurasia collision zone, Geophys. J. Int., 150, 422-439. Tapponnier, P., F. J. Ryerson, J. Van der Woerd, A.-S. Mériaux, and C. Lasserre (2001), Long-term slip rates and characteristic slip: keys to active fault behaviour and earthquake hazard, Earth and Planetary Sciences, 333 483-494. Taylor, J. R. (1997), An Introduction to Error Analysis, The Study of Uncertainties in Physical Measurements, 2nd ed., University Science Books, Sausalito, CA. Tchalenko, J. S., M. Berberian, and H. Behzadi (1973), Geomorphic and seismic evidence for recent activity on the Doruneh Fault, Iran, Tectonophysics, 19, 333-341. Thomas, D.S.G. (1989), Arid zone geomorphology, Halsted press, New York. Van der Woerd, J., F. J. Ryerson, P. Tapponnier, A.-S. Meriaux, Y. Gaudemer, B. Meyer, R. C. Finkel, M. W. Caffee, Zhao Guoguang, and Xu Zhiqin (2000), Uniform Slip-Rate along the Kunlun Fault:Implications for seismic behaviour and large-scale tectonics, Geophysical Research Letter, 27(16), 2353-2356. Vernant, P., et al. (2004), Present-day crustal deformation and plate kinematics in the Middle East constrained by GPS measurements in Iran and northern Oman, Geophys. J. Int. 157(1), 381-398, doi: 10.1111/j.1365-246X.2004.02222.x. Walker, R., J. Jackson, and C. Baker (2004), Active faulting and seismicity of the Dasht-e- Bayaz region, eastern Iran, Geophys. J. Int. 157, 265-282, 10.1111/j.1365- 2966.2004.02179.x. Ward, G. K., and S. R. Wilson (1978), Procedures for comparing and combining radiocarbon age determinations: a critique, Archaeometry, 20(1), 19-31. Wellman, H. W. (1966), Active wrench faults of Iran Afghanistan and Pakistan, Geologische undschau, 55, 716-735. Wells, D., and k. Coppersmith (1994), New empirical relationships among Magnitude, Rupture length, Rupture Width, Rupture Area, and surface Displacement, Bull. Seismol. Soc. Am., 84(4), 974-1002. Wulff, H. E. (1968), The Qanats of Iran, Sci. Am., 218, 94-105, doi: 10.1038/scientificamerican0468-94. Zreda, M. G., F. M. Phillips, and D. Elmore (1994), Cosmogenic 36Cl accumulation in unstable landforms, 2. Simulations and measurements on eroding moraines, Water Resources Research, 30, 3127-3136. Zreda, M.G., Philips, F.M., Elmore, D., Kubik, P.W., Sharma, P. & Dorn, R.I., (1991). Cosmogenic chlorine-36 production rates in terrestrial rocks, Earth planet. Sci. Lett., 105, 94–109

115

CHAPTER IV

The Arabia-Eurasia collisional boundary in NE Iran, the eastern termination of the Doruneh Fault System

Chapter IV

The Arabia-Eurasia collisional boundary in NE Iran, the eastern termination of the Doruneh Fault System

1. Introduction In the Iranian plateau, tectonic deformations result from the Arabia–Eurasia convergence. This convergence accommodates by (1) crustal shortening and strike-slip faulting in intracontinental deformation domains such as the Zagros, Alborz and Kopeh Dagh mountain ranges, and (2) the active subduction of Makran. The overall kinematics of the Arabia-Eurasia convergence has been broadly understood through seismicity and GPS studies (e.g., Jackson et al., 1995; Vernant et al., 2004; Reilinger et al., 2006, Hessami et al., 2006; Masson et al., 2007; Djamur et al., 2010) that show a northward motion of Arabia with respect to Eurasia. This northward motion is a maximum rate of 26±2 mm/yr at the longitude of ~59°E (Sella et al., 2002; McClusky et al., 2003; Vernant et al., 2004; Reilinger et al., 2006). In eastern Iran, this plate motion results in right-lateral shear on major N-trending dextral fault systems (Fig. 1) on both sides of the Lut block (Tirrul et al., 1983; Regard et al., 2004; Walker and Jackson, 2004; Meyer and LeDortz, 2007). The GPS velocities indicate a total shear of 16±2 mm/yr between Central Iran and the Helmand block among which about 7±2 of this shear is accommodated along the eastern border of the Lut block (Vernant et al., 2004; Masson et al., 2007). Farther north, in northeast Iran, the available geodetic (Vernant et al., 2004; Resillinger et al., 2006; Masson et al., 2007; Tavakoli, 2007) and geological (Shabanian et al., 2009a) data, indicate that the Arabia northward motion is accommodated in northeast Iran with a rate ranging from 4 to 11mm/yr. A significant part of this deformation (8±2 mm/yr) is localized along the Bakharden Quchan Fault Ssystem, suggesting that this fault system forms the boundary between Iran and Eurasia (Shabanian et al., 2009a). The nearly similar rates at both the eastern and northeastern boundaries of the Arabia- Eurasia collision imply that the majority of the northward motion is transferred from the south to the north. In the Kopeh Dagh Mountains the northward motion of Central Iran relative to Eurasia (8±2 mm/yr) is principally accommodated by strike-slip faulting localized along the dextral Bakharden-Quachan Fault System, (BQFS; Shabanian et al., 2009a). A part of this right-lateral motion is transferred to the south (i.e. Binalud) through the Meshkan transfer zone (i.e. Chakaneh fault system, Shabanian et al. 2009b) and end into the

119 Chapter IV dextral-reverse Nayshabur fault system and the right-lateral Mashhad fault zone on the southwest and northeast sides of the Binalud Mountains (Shabanian et al., 2009b). This right- lateral strike-slip faulting along the crustal scale fault systems allowed Central Iran, translate northeastward with respect to Eurasia (Shabanian et al., 2009b).

Figure 1. Principal deformation domains and structural units in Central and northeastern Iran overlain on GTOPO30 shaded digital elevation model. Black arrows are GPS horizontal velocities (mm/yr) from Masson et al. (2007) in a Eurasia-fixed reference frame. The box in the upper left inset shows the location in the Arabia– Eurasia collision zone. Grey arrows and associated numbers represent Arabia–Eurasia plate velocities (mm/yr) after Reilinger et al. (2006).

Farther south at latitude of ~35°N the ~400-km-long Doruneh Fault System (DFS; Farbod et al., 2011), separates the NW-trending Kopeh Dagh and Binalud deformation domains from the crustal N-trending dextral strike-slip faults such as the Nayband and Neh faults (Fig. 1). Despite of the recent active tectonic studies (Walker and Jackson., 2004; Fattahi et al., 2007;

120 Chapter IV

Qaleh Fault; KF,

-

FS FS fault zones i.e.

sured sured in the Quaternary

Naini et al. (1992). Fault traces are based

-

. (1976) and Alavi

Nezhad et al

-

Z) and Eastern Fault Zone (EFZ). Abbreviations are: BF, Bijvard Fault; DQF, Dahan

along the CFZ.

on geomorphic and structural analysis of SPOT5 and LANDSAT ETM+ by Farbod et Western al., Fault Zone (WFZ), (2011).Central Fault Zone The (CF blue lines indicate the three D Kharturan Fault; KHF, Khaf Fault; JTF, Jangal Thrust Fault. (b) deposits Results of the inversion analysis of fault kinematic data mea Figure2. (a) Geological map of region affected by DFS simplified after Eftekhar

121 Chapter IV

Farbod et al., 2011), the role of the DFS in the accommodation and/or transfer of the Arabia- Eurasia convergence remain unknown. According to the nearly identical rates of Arabia- Eurasia convergence at its both eastern and northeastern boundaries the DFS could not absorb the important part of the convergence. In this case, the key question is to know how the Arabia-Eurasia northward motion is transferred across the DFS perpendicular to the strike of this predominantly strike-slip fault system. Our structural and geomorphic investigations (chapter II, Farbod et al., 2011) reveal that the curved shape of the DFS is a structural assemblage of three distinct western (WFZ), central (CFZ) and eastern (EFZ) fault zones, instead a ~400-km-long uniform structure (Fig. 2). Each fault zone has individual structural, geomorphic and kinematic characteristics. The CFZ is pure left-lateral strike-slip. The WFZ is reverse left-lateral oblique-slip while the EFZ is predominantly reverse (chapter II, Farbod et al., 2011). These active fault kinematics are consistent with the NE direction of the present-day maximum horizontal stress axes (σ1), deduced from the fault data inversion, along the CFZ (Farbod et al., 2011) and in the NE of Iran (Shabanian et al., 2010). However interactions between the DFS and other intersecting faults have caused kinematics complexity along the fault system (Farbod et al., 2011; this chapter). The main objectives of this chapter are (1) to describe the fault kinematics of the NW- trending EFZ, i.e. eastern termination of the DFS, (2) to reevaluate the geodynamic role of the DFS in the accommodation of the Arabia-Eurasia convergence, and (3) to identify the continuation of the Iran-Eurasia plate boundary in the gap between the eastern and northeastern boundaries of the Arabia Eurasia collision zone. To achieve these objectives, a detail structural and geomorphic mapping was carried out along the EFZ through remote sensing analyses of high resolution SPOT5 (pixel size of 5m) and GeoEye (Google Earth, pixel size of 30cm) satellite images, SRTM digital topographic data and regional scale geological maps. These new data were completed by available data in NE Iran (Shabanian et al., 2009a; 2009b; 2010) and helped proposing a model to describe how the active deformation is accommodated along the eastern termination of the DFS.

2. Active faulting along the Eastern Fault Zone (EFZ) The transition between E-trending left-lateral CFZ and NW-trending reverse EFZ occurs in ~30-km-wide zone between the longitude of 59° and 59.5°E, where the left-lateral cumulative offset dies out eastwards. The geometry and kinematic of the DFS drastically change at longitude of ~59.5°E, by taking ~30° bend towards the southeast. The ~100-km-long EFZ

122 Chapter IV forms a trailing contractional imbricate fan at the eastern termination of the DFS (Fig. 2 and 3). The fault zone varies between 1.5 and 14 km in width and comprise the NE-dipping reverse faults (Farbod et al., 2011), implying a wide domain of distributed deformation. Within the EFZ, the traces of the faults have two general strike of ~N110°E which intersected and displaced by the ~N140-150°E-trending faults. Together these two intersected orientations provide a general trend of N130±10°E for the EFZ (Fig. 3). According to Farbod et al. (2011), left-lateral faulting along the CFZ is principally accommodate by thrust faulting along the imbricate reverse faults and at the regional scale the EFZ forms the restraining termination of the DFS. This reverse active faulting along the EFZ is consistent with the regional ~N40°E trending maximum horizontal stress axes in NE Iran (Shabanian et al., 2010; Farbod et al., 2011) (Fig. 2).

Figure 3. Shaded relief map (SRTM data) of the Eastern Fault Zone (EFZ). The general E-trending of the CFZ takes a ~30° (clockwise) bend at longitude 59.5°E. The geometry and kinematic of the fault change drastically in the EFZ. KHF, Khaf Fault; JF, Jangal Fault.

2.1. Geological and structural evidences of right-lateral component of faulting Within the EFZ, series of N140-150°E-trending strike-slip faults cut and right-laterally displaced the Eocene-Oligocene rocks. These right-lateral faults end into the NW-trending

123 Chapter IV

Figure 4. (a) Geologic map (after Alavi Naini et al., 1982) showing the evidence of dextral faulting at longitude of ~60°E. The N150°E-trending right-lateral fault displaced the Eocene-Oligocene outcrops and join to the oblique to reverse NW-trending faults further south. The Oligocene fold axes show a right stepping arrangement with an oblique orientation with respect to the fault trace and confirm the right-lateral faulting along this fault. (b) SPOT image of the displaced Eocene-Oligocene outcrops (c) Shaded relief map based on SRTM digital topographic data showing two sigmoidal ridges with right-lateral stepping arrangement. (d) simplified model of a possible shear zone within this area.

124 Chapter IV reverse faults which accommodate the right-lateral motion as the restraining bend. For instance, at the longitude of ~60°E, the northernmost faults of the EFZ which separate Eocene from Oligocene deposits is right-laterally cut by three N150°E-trending dextral faults (Fig. 4a and b). Among these three faults the eastern one turns to the east and change the orientation to the ~N120°E (Fig. 4b). This bounds a crescent-like topographic ridge of Precambrian outcrops in the hanging wall (Fig. 4b). The middle NNW-trending dextral fault continues to the south and cuts and deflects the Oligocene folds before join to ~N120°E–trending fault (Fig. 4b).

Figure 5. (a) Series of right-lateral faults which displaced the Oligocene conglomerate and join to a NW-trending dextral fault. (b) SPOT image of both NNW-trending dextral and NW-trending dextral-reverse fault which displaced a series of streams incised in the Neogene to Quaternary deposits. (c) Sketch of the same dextral fault.

Interestingly these two sigmoidal ridges indicate a right-stepping arrangement and could represent a right-lateral shear zone (Fig. 4c). Moreover at both side of this dextral fault (i.e. middle one), three Oligocene fold axes show right-stepping arrangement, oblique to the fault trace. This could reconfirm the presence of a right-lateral shear zone in this area (Fig. 4b&d).

125 Chapter IV

At the longitude of 60.30°E, another series of N140°E-trending dextral faults displaced right-laterally the Oligocene-Miocene sandstone deposits (Fig. 5a). These dextral faults obliquely join to a major N110°E-trending fault (Fig. 5b) which also displaced right-laterally three incised streams (Fig. 5b and c). At the Figure 6 we present a simplified geological map (after Alavi Naini et al., 1982) of the area between the longitudes of ~60°E and 60.30°E. The fault bounded outcrops of Precambrian metamorphic rocks surrounding by Oligocene deposits, indicate a right-handed arrangement within the EFZ. Regarding to numerous major and minor faults bounding the Precambrian rocks, it is difficult to identify the segments that right-laterally offset this outcrops but their right-handed stepping arrangement could consider as an evidence of right- lateral faulting along the NNW-trending faults (Fig. 6).

2.2. Geomorphic evidence for right-lateral component of faulting The Quaternary alluvial fans are the predominant geomorphic features offset along the DFS. Our detailed geomorphic mapping based on SPOT5 satellite images complemented by filed survey allow us to subdivide three generations of inset alluvial fans at the regional scale, from older to younger Q3, Q2 and Q1. Along the strike-slip faults, regarding to horizontal motion of the blocks, the inset fans could show a lateral shift with respect to older fan. This shift corresponds to the sense of faulting. For instance along the left-lateral strike-slip CFZ, the inset fans show a left-handed arrangement (chapter II, Farbod et al., 2011). Along the EFZ, these alluvial fans are mostly tilted and uplifted on the hanging wall of reverse faults which affected the alluvial plain (Fig. 11b, Chapter II). But east of the longitude of ~60°E a right-handed arrangement of inset alluvial fans has observed (Fig. 7). This arrangement should be due to the right-lateral displacement of feeding rivers with respect to older fans downhill from the fault. The digital topographic elevation model of this region indicates basin catchments shifted right-laterally along ~N135°E-trending faults (Fig. 8a and b).

126 Chapter IV

Figure 6. Geological map of the region presented in Figure 4 (simplified after Alai Naini, et al., 1982). The right- handed arrangement of Precambrian metamorphosed outcrops indicate a dextral faulting along NW-trending faults. Fault traces are exaggerated (thick trace) to better illustrate right-stepping arrangement. The abbreviations in Legend are the same than in Figure 4.

127 Chapter IV

Figure 7. SPOT images and their geomorphic interpretations of inset Quaternary alluvial fans with a right- handed arrangement. This inset arrangement is due to the right-lateral shift of alluvial fans relative to their catchment basins. See Figure 8 for the locations.

Figure 8. (a) and (b) shaded relief maps and their geomorphic interpretation which show the catchment basins and axial drainage that are offset right laterally with respect to the deposited alluvial fans.

128 Chapter IV

Figure 9. SPOT image indicates a NW-trending fault within the EFZ which thrust the Mio-Pliocene conglomerate onto the alluvial fans. The evidences of right-lateral component of faulting are the deflected gullies and hills and the alluvial fans and their axial drainages which offset right laterally with respect to their catchments.

In the area between the longitude of 59.97°E and 60.08°E, the evidence of right-lateral component of faulting has also found along a ~N120°E-trending fault which thrusts the Mio- Pliocene conglomerate on the Quaternary alluvial fans. For instance a series of gullies and rills show a right-lateral deflection and three Q2 alluvial fans are displaced right-laterally relative to their catchments (Fig. 9).

129 Chapter IV

Figure 10. SPOT image of a Q3 fan surface within the EFZ. The axial drainage shows a right-lateral deflection along a NW-trending fault. See Figure 3 for location.

Farther north of the EFZ at 35.10°N/60.21°E, we also found the geomorphic evidence of right-lateral faulting along the same orientation of the fault (N140-150°E) in the EFZ. In this area along a ~N150°E-trending fault, two Quaternary alluvial fans have been right-laterally offset with respect to their feeding catchments (Fig. 11). This may suggest a regional dextral faulting along the NNW-trending faults in the NE Iran.

. Figure 11. SPOT image of a Q2 alluvial fan offset right-laterally from its feeding drainage and catchment along a NNW-trending fault north of the EFZ. Location (60.21°E-35.10°N) is indicated in Figure 3.

130 Chapter IV

3. Summary and Discussion The EFZ is characterized as the trailing contractional imbricate fan termination of the DFS with predominant reverse faulting kinematics (Fig. 3). This fault zone located between two eastern and northeastern boundaries of the Arabia-Eurasia collision. In such a configuration we focus our geological and geomorphic studies along this fault zone in order to understand its geodynamic role in the transfer of the Arabia-Eurasia convergence. The eastern boundary of the Arabia-Eurasia collision is located along the N-trending dextral fault system, east of the Lut block (e.g. Neh fault system). To the NE of Iran Shabanian et al. (2009b) suggest that NE boundary of the Arabia-Eurasia collision zone is localized along the series of NNW-trending dextral fault systems such as Bakharden-Quchan and Chakaneh fault systems (Shabanian et al., 2009a and 2009b). These dextral faults end into reverse bend termination such as Kashafrud fault (Hollingsworth et al., 2006) and Neyshabur fault system (Shabanian et al., 2009b; Shabanian et al., submitted), which in turn form the restraining fault relay zones. These fault systems join the Main Kopeh Dagh fault to the Binalud deformation domain. In the region between Binalud Mountains and the EFZ, the Analyses of fault kinematic data in two sites, located at 59.54°E-35.85°N and 60.12°E-35.25°N (Shabanian et al., 2010), show a right-lateral component of faulting along the NNW-trending fault planes (Fig. 12a). We also presented the geomorphic evidence of dextral faulting along NNW-trending fault at 60.21°E-35.10°N, north of the EFZ (Fig. 11). This indicates that the right-lateral faulting along the NNW-trending fault within the Kopeh Dagh and Binalud Mountains (Shabanian et al 2009a and 2009 b) is extending southward. Our detailed geomorphic and structural investigations reveal the presence of the NNW- trending dextral strike-slip faults within the EFZ. These strike-slip faults are distributed throughout the EFZ and interact with the WNW- to NW-trending reverse faults at the eastern termination of the DFS. The crosscutting fault relationships in this zone imply that the eastern contractional termination of the DFS has been involved in a NNW-wards right-lateral shear between Central Iran and Eurasia. Accordingly, the Central Iran–Eurasia relative motion is taken up by shortening on the NW-trending (~N110°E) reverse faults and is transferred northward via the NNW-trending (~N150°E) dextral faults distributed within the EFZ. A slight component of dextral faulting can also occur on reverse faults with intermediate orientation. It is noteworthy that all these fault kinematics are compatible with the present-day stress state in NE Iran characterized by a NE-trending σ1 axis. Combining all these data and deduced results, we proposed a geodynamic model (Fig. 12),

131 Chapter IV

Figure 12. (a) Simplified active fault map of NE Iran (after Shabanian et al., 2009a; 2009b and this study) together with the main tectonic domains of NE Iran. (b) Schematic kinematic model showing crustal boundary of the Arabia-Eurasia convergence. The interaction between discontinuous dextral faults separated by compressional to transpressional relay zones allows the Central Iran to be translated northward relative to Eurasia. The white arrows show the overall motion of Central Iran with respect to Eurasia. The yellow arrows indicate the NE direction of the σ1 in NE Iran. The abbreviations are: MKDF, Main Kopeh Dagh Fault; BQ, Bakharden-Quchan Fault; CF, Chakaneh Fault; NF, Neyshabur Fault; KF, Kashafrud Fault , GKFS; Great Kavir Fault System; DFS, Doruneh Fault System; EFZ, Eastern Fault Zone; KHF, Khaf Fault; NHF, Neh Fault.

132 Chapter IV according which Arabia-Eurasia convergence is transferred to the north through the eastern termination of the DFS. In such a model the EFZ forms a complex right-lateral transpressional relay zone between the crustal eastern and northeastern boundaries of the Arabia-Eurasia convergence. Such a structural interaction between discontinuous dextral fault systems separated by compressional to transpressional relay zones allows the western Iranian side (Central Iran and Lut block) to be translated northward relative to Eurasia. As an important geodynamic constrain, we suggest that the E-W-trending DFS is also translated northward with respect to Eurasia. That is, the crustal scale DFS does not make a significant boundary condition in the GPS velocity field of the Iranian plateau. This later explains also why the GPS velocities at KASH station (e.g., Vernant et al., 2004) located in the middle of the DFS show a northward motion relative to Eurasia.

4. Conclusion The data and results presented in this chapter reveal the presence of NNW-trending dextral faults distributed throughout the EFZ. This could indicate that this dominantly reverse fault zone is involved in the right-lateral shear zone between Central Iran and Eurasia. This study suggest that the northward motion of Central Iran relative to Eurasia is accommodated by both shortening on NW-trending reverse faults and transfer to the north through NNW- trending dextral strike-slip faults distributed within the EFZ.

We propose a geodynamic model in which the Arabia-Eurasia convergence is transferred to the north through the eastern termination of the DFS. In this model the EFZ forms a complex right-lateral transpressional relay zone between the crustal eastern and northeastern Arabia-Eurasia convergence boundaries. The interaction of discontinuous dextral fault systems separated by compressional to transpressional relay zones in NE Iran compose the crustal boundary of the Arabia-Eurasia convergence. Along this boundary the Central Iran as well as the DFS translates to the north with respect to Eurasia. This indicates that the E-trending DFS does not make a significant boundary condition in the GPS velocity field of the Iranian plateau. To improve our proposed geodynamic model the field surveys along the EFZ in order to better understand its active faulting is necessary. The similar detailed tectonic geomorphology investigations between the Neh and Doruneh Fault Systems would complete this model all along the Arabia-Eurasia collision boundary

133 Chapter IV

References Alavi-Naini, M and A. Behruzi . (1982), Geological map of , 1:250 000, Geological Survey of Iran. Alavi-Naini, M., M. J. Vaezi-Pour, N. Alavi Tehrani, A. Behrouzi, and M. H. Kholghi (1992), Geological Map of Torbat-e Heidariyeh, 1:250 000, Geological survey of Iran. Djamour, Y., et al. (2010), GPS and gravity constraints on continental deformation in the Alborz mountain range, Iran, Geophys. J. Int., 1287-1301, doi: 10.1111/j.1365- 246X.2010.04811.x. Eftekhar-Nezhad, J., A. Aghanabati, B. Hamzehpour, and V. Baroyant (1976), Geological map of Kashmar Quadrangle. 1; 250 000, Geological survey of Iran. Farbod, Y., O. Bellier, E. Shabanian, and M. R. Abbassi (2011), Geomorphic and structural variations along the Doruneh Fault System (Central Iran), Tectonics, doi:10.1029/2011TC002889,. Hessami, K., Nilforoushan, F., and Talbot, C. (2006), Active deformation within the Zagros Mountains deduced from GPS measurements, J. Geol. Soc. Lond., 163, 143–148. Hollingsworth, J., J. Jackson, R. Walker, M. Gheitanchi, and M. Bolourchi (2006), Strike-slip faulting, rotation, and along-strike elongation in the Kopeh Dagh mountains, NE Iran, Geophys. J. Int., 166, 1161–1177, doi: 10.1111/j.1365-246X.2006.02983.x. Jackson, J., and D. Mckenzie (1984), Active tectonics of the Alpine-Himalayan Belt between turkey and Pakistan, Geophys. J. R. astr. Soc, 77(1), 214-245. Jackson, J., J. Haines, and W. Holt (1995), The accomodation of Arabia-Eurasia plate convergence in Iran, J. Geophys. Res., 100(B8), 15205-15219. Masson, F., M. Anvari, Y. Djamour, A. Walpersdorf, F. Tavakoli, M. Daignieres, H. Nankali, and S. Van Gorp (2007), Large-scale velocity field and strain tensor in Iran inferred from GPS measurements: new insight for the present-day deformation pattern within NE Iran, Geophys. J. Int., 170, 436-440, doi: 10.1111/j.1365-246X.2007.03477.x. McClusky, S., R. Reilinger, S. Mahmoud, D. Ben Sari, and A. Tealeb (2003), GPS constraints on Africa (Nubia) and Arabia plate motions, Geophys. J. Int. 155(1), 126-138. Meyer, B., and K. Le Dortz (2007), Strike-slip kinematics in Central and eastern Iran: Estimating fault slip rates averaged over the Holocene, Tectonics, 26, TC5009, doi: 10.1029/2006TC002073. Regard, V., O. Bellier, J.-C. Thomas, M. R. Abbassi, J. Mercier, E. Shabanian, K. Feghhi, and S. Soleymani (2004), Accommodation of Arabia-Eurasia convergence in the Zagros- Makran transfer zone, SE Iran: A transition between collision and subduction through a young deforming system, Tectonics, 23, TC4007, doi: 10.1029/2003TC001599,. Reilinger, R., et al. (2006), GPS constraints on continental deformation in the Africa-Arabia- Eurasia continental collision zone and implications for the dynamics of plate interactions, J. Geophys. Res., 111, B05411, doi: 10.1029/2005JB004051. Sella, F. G., H. T. Dixon, and A. Mao (2002), Revel: A model for Recent plate velocities from space geodesy, J. Geophys. Res., 107(B4), ETG 11. Shabanian, E., L. Siame, O. Bellier, L. Benedetti, and M. R. Abbassi (2009a), Quaternary slip rates along the northeastern boundary of the Arabia–Eurasia collision zone (Kopeh Dagh Mountains, Northeast Iran), Geophys. J. Int., 1-23, doi: 0.1111/j.1365- 246X.2009.04183.x. Shabanian, E., O. Bellier, L. Siame, N. Arnaud, M. R. Abbassi, and J.-J. Cochemé (2009b), New tectonic configuration in NE Iran: Active strike-slip faulting between the Kopeh Dagh and Binalud mountains, Tectonics, 28, TC5002, doi: 10.1029/2008TC002444. Shabanian, E., O. Bellier, M. R. Abbassi, L. Siame, and Y. Farbod (2010), Plio-Quaternary stress states in NE Iran: Kopeh Dagh and Allah Dagh-Binalud mountain ranges, Tectonophysics, 480, 280-304, doi: doi:10.1016/j.tecto.2009.10.022.

134 Chapter IV

Tirrul, R., I. R. Bell, R. J. Griffis, and V. E. Camp (1983), The Sistan Sistan suture zone of eastern Iran, Geological Society of America Bulletin, 94, 134-150. Vernant, P., et al. (2004), Present-day crustal deformation and plate kinematics in the Middle East constrained by GPS measurements in Iran and northern Oman, Geophys. J. Int., 157(1), 381-398, doi: 10.1111/j.1365-246X.2004.02222.x. Walker, R., and J. Jackson (2004), Active tectonics and late Cenozoic strain distribution in Central and eastern Iran, Tectonics, 23, TC5010, doi: 10.1029/2003TC001529.

135

DISCUSSION AND CONCLUSION

Discussion and Conclusion

This dissertation focuses on the active tectonics of the Doruneh Fault System (DFS), NE Iran. For this aim, we analyzed a combination of original remote-sensing geological and geomorphic observations, field (fault kinematics and morphotectonics) and laboratory (cosmogenic exposure ages) data complemented by available historical and instrumental seismicity, as well as published GPS velocities in NE Iran. This led us reconstructing a coherent tectonic framework of the DFS, and allowed us to answer the following fundamental questions:

What are the structural and kinematic characteristics of the DFS?

What are the rate and distribution of slip along the fault system?

What is the geodynamic role of the DFS in the Arabian-Eurasian convergence? That is, how the Central Iran-Eurasia northward motion is taken up / transferred by the E-W Doruneh Fault System?

The analysis of satellite images (Landsat 7 ETM+, SPOT5, QuickBird and GeoEye) with different resolution, and digital topographic model (SRTM) combined with direct field observations allowed us to perform a detailed mapping of Quaternary and active faults at both the regional and site scales. The entire fault system is mapped with special cares to the relay zones and intersection points between different fault zone or faults, as well as the extent of fault segments affecting Holocene alluvial deposits. Using these data, we have evaluated the structural significance of faults examining the distribution of late Quaternary faulting along the DFS. Considering the variations in structural and kinematic characteristics along the strike of the DFS, we have proposed a fault segmentation model (chapter II; Farbod et al., 2011). Accordingly, the curved shape geometry of the DFS is a structural assemblage of three distinct western (WFZ), central (CFZ) and eastern (EFZ) fault zones, instead of, a ~400-km- long uniform structure. Each fault zone is characterized by its own geometry and kinematics leading to discrete structural and geomorphic characteristics: (a) the ~70-km-long WFZ, characterized by reverse left-lateral mechanism with left-handed step-over geometry, (b) the ~185-km-long CFZ which is pure left-lateral strike-slip and comprises nearly parallel faults and finally (c) the ~100-km-long EFZ, is characterized by reverse faulting and fault-related folding which formed a trailing contractional imbricate fan fault-termination. These distinct structural and kinematic characteristics of each fault zone indicate that the deformation cannot

139 Discussion and Conclusion be uniformly accommodated along the DFS. To explain such variations, and to describe the way that the DFS accommodates the Arabia-Eurasia convergence normal to the overall fault orientation we proposed a new kinematic model. In this model, DFS takes up the northward motion between Central Iran–Lut block relative to Eurasia by a complex kinematics varying from pure reverse to pure left-lateral strike-slip faulting. The kinematics of the WFZ and EFZ corresponds to the direction of the NE-trending regional compression. In the central part, partitioning of slip into strike-slip and reverse component of faulting on parallel faults (strain partitioning) allows the CFZ to remain pure left-lateral strike-slip fault, while the overall convergence vector and slip on the fault are perpendicular (Farbod et al., 2011). Subsequently, we have focused on the kinematics and tectonic geomorphology of the CFZ. Fault slip data (fault plan and associated striations) were measured in five localities along the CFZ. The inversion of these fault kinematic data indicates a N45±15°E-trending maximum horizontal stress (σ1) axis which, of the modern stress state, is coherent with the N30±10°E - direction of the σ1 stress axis in NE Iran (Shabanian et al., 2010). The distribution of horizontal slip along the DFS was evaluated through the measurement of three sets of cumulative left-lateral offsets, at sixty seven sites, recorded by late Quaternary geomorphic markers such as alluvial fans (Q1, Q2 and Q3, from younger to older) and their associated geomorphic features, as well as terrace risers between inset alluvial fan surfaces. The left-lateral offsets are found only along the CFZ and WFZ such that beyond the two eastern and western end-points there is no evidence of left-lateral faulting along the DFS. The overall distribution of slip along these two fault zones indicates the maximum cumulative offsets (whatever the age) occurred in the mid-length of the CFZ and decrease towards the both end of the fault zone. Such a distribution pattern is the same for all cumulative offsets recorded by the three generation of alluvial fans, i.e. Q1, Q2 and Q3. The abandonment surfaces of these inset alluvial fans are dated using in situ-produced 10Be and 36Cl cosmogenic nuclides. Accordingly, the exposure ages of 12.3±2.9, 36.5±6.3 and 101.8±9.5 ka have been determined for the abandonment surfaces of Q1, Q2 and Q3 alluvial fans, respectively (Chapter III). The considerable, but varying amount of inheritance in the cosmogenic nuclide concentrations has been found in some samples collected from the Q1 and Q2 inset fan surfaces. We suggest this amount of inheritance is due to the complex exposure history of clasts that were reworked from upper alluvial terraces to lower inset terraces during the occasional flooding period. Along the CFZ and WFZ, the combination of the abandonment ages and geomorphic offsets yielded three independent sets of horizontal slip rates in sixty seven sites. These

140 Discussion and Conclusion independent sets of slip rate, averaged over time periods of ~12, ~36 and ~100 ka, show nearly similar overall distributions of slip along both the WFZ and CFZ, since at least the Late Pleistocene. This may suggest that the (1) left-lateral slip rate along the WFZ and CFZ has remained constant during the Late Pleistocene, with a maximum rate of ~8.3 mm/yr, and (2) the long-term “slip behavior” of the both WFZ and CFZ is characteristic (Fig.1). During the Holocene, a maximum slip rate of ~5.3 mm/yr is estimated for left-lateral faulting along the CFZ. The distribution of the Holocene slip indicates that the CFZ is divided into two ~80- and ~100-km-long western and eastern segments, respectively. These two individual segments are separated by a ~3km-long persistent boundary, where Q1 alluvial fan surfaces have not been ruptured during at least the last ~12 ka, and may represent two independent seismogenic fault segments. Along the WFZ the maximum slip rate of ~2 mm/yr is estimated for left-lateral faulting on the reverse-sinistral oblique slip WFZ. Nevertheless, the vertical slip rate and consequently, the total slip rate of the fault zone remain unknown. The results of this study also indicate that (1) continuous large strike-slip faults could be divided into several discrete fault zones to take up more easily overall tectonic motions, and (2) those do not rupture along their entire length during a single earthquake. As for the DFS, the seismogenic segment lengths vary between ~70 and ~100 km, which are able to produce characteristic earthquakes with magnitudes of Mw ≈7.2-7.4 with the possible recurrence interval of ~750 yr. Finally (Chapter IV), we conducted our remote-sensing investigations on the EFZ, i.e., eastern termination of the DFS, in order to improve our understanding of the geodynamic role of the DFS in the Arabia-Eurasia convergence. This stage is principally based on structural and geomorphic analyses of high resolution SPOT5 (Google Earth) satellite images, SRTM digital topographic data and regional scale geological maps. The results deduced from this stage have revealed the presence of NNW-trending dextral faults within the EFZ involving this dominantly reverse fault zone in a NNW-wards right-lateral shear between Central Iran and Eurasia. These strike-slip faults are distributed throughout the Eastern Fault Zone and interact with the WNW- to NW-striking reverse faults at the eastern termination of the DFS. Accordingly, the Central Iran – Eurasia relative motion is taken up by shortening on the reverse faults and is transferred northward via a distributed dextral shear zone. It is noteworthy that all these fault kinematics are compatible with the present-day stress state in NE Iran characterized by a NE-trending σ1 axis.

141

Discussion and Conclusion

long the central and western part of the DFS (CFZ and WFZ), estimated for a time period of ~12 (Q1), ~36 (Q2) and ~100

. . Overall slip rate distribution a

1

Figure (Q3) ka.

142 Discussion and Conclusion

Combining all these data and deduced results, we proposed a geodynamic model (Fig. 2) according which Arabia-Eurasia convergence is transferred to the north through the eastern termination of the DFS. At regional scale, the eastern boundary of the Arabia-Eurasia convergence is localized along the N-striking dextral Neh fault system, east of the Lut block. Farther north, in NE Iran, the NNW-striking dextral Bakharden-Quchan and Chakaneh fault systems are known as the northeastern boundary of the Arabia-Eurasia convergence (Shabanian et al., 2009a and 2009b). All these strike-slip faults end into reverse to oblique- slip bend-terminations (e.g. Kashafrud Fault and Neyshabour Fault System - Hollingsworth et al., 2006; Shabanian, 2009; Shabanian et al., submitted to GJI), which in turn form the restraining fault relay zones (Fig. 2). We suggest that the EFZ forms a complex right-lateral transpressional relay zone between the crustal eastern and northeastern convergence boundaries.

Figure 2. (a) Simplified active fault map of the NE Iran (after Shabanian et al., 2009a; 2009b and this study) together with the main tectonic domains of NE Iran. (b) Schematic kinematic model showing crustal boundary of the Arabia-Eurasia convergence. The interaction between discontinuous dextral faults separated by compressional to transpressional relay zones allow the Central Iran to be translated northward relative to Eurasia. The white arrows show the overall motion of Central Iran with respect to Eurasia. The yellow arrows indicate the NE direction of the σ1 in NE Iran. The abbreviations are: MKDF, Main Kopeh Dagh Fault; BQ, Bakharden- Quchan Fault; CF, Chakaneh Fault; NF, Neyshabur Fault; KF, Kashafrud Fault; GKFS, Great Kavir Fault System; DFS, Doruneh Fault System; EFZ, Eastern Fault Zone; KHF, Khaf Fault; NHF, Neh Fault.

143 Discussion and Conclusion

Such a structural interaction between discontinuous dextral fault systems separated by compressional to transpressional relay zones allows the western Iranian side (Central Iran and Lut block) to be translated northward relative to Eurasia. As an important geodynamic constrain, we suggest that the E-W crustal DFS is also translated northward with respect to Eurasia. This may explain how (1) the E-W trending DFS slips left-laterally perpendicular to the overall northward motion of Central Iran relative to Eurasia, and (2) Why the crustal scale DFS does not make a significant boundary condition in the GPS velocity field of the Iranian plateau (e.g., Vernant et al., 2004). Despite the data and results presented in this study, detailed paleoseismological investigations are needed to improve our knowledge of the seismogenic behavior of the WFZ and CFZ. The WFZ should be studied in terms of vertical faulting; this would help to quantify cumulative vertical versus lateral offsets, and to estimate the total slip rate of the fault zone. The improvement of the proposed geodynamic model could be achieved through field surveys along the EFZ in order to better understand the active faulting, and similar detailed tectonic geomorphology investigations between the Neh and Doruneh Fault Systems. Well-distributed paleoseismological studies (e.g., trenching) would also improve the knowledge of earthquake geology and the assessment of seismic hazard on the DFS.

144 Discussion and Conclusion

References Farbod, Y., O. Bellier, E. Shabanian, and M. R. Abbassi (2011), Geomorphic and structural variations along the Doruneh Fault System (Central Iran), Tectonics, 30, TC6014, doi:10.1029/2011TC002889. Hollingsworth, J., J. Jackson, R. Walker, M. R. Gheitanchi, and M. J. Bolourchi (2006), Strike-slip faulting, rotation, and along-strike elongation in the Kopeh Dagh mountains, NE Iran, Geophys. J. Int., 166, 1161-1177. Masson, F., M. Anvari, Y. Djamour, A. Walpersdorf, F. Tavakoli, M. Daignieres, H. Nankali, and S. Van Gorp (2007), Large-scale velocity field and strain tensor in Iran inferred from GPS measurements: new insight for the present-day deformation pattern within NE Iran, Geophys. J. Int., 170, 436-440, doi: 10.1111/j.1365-246X.2007.03477.x. Shabanian, E., L. Siame, O. Bellier, L. Benedetti, and M. R. Abbassi (2009a), Quaternary slip rates along the northeastern boundary of the Arabia–Eurasia collision zone (Kopeh Dagh Mountains, Northeast Iran), Geophys. J. Int., 1-23, doi: 0.1111/j.1365- 246X.2009.04183.x. Shabanian, E., O. Bellier, L. Siame, N. Arnaud, M. R. Abbassi, and J.-J. Cocheme (2009b), New tectonic configuration in NE Iran: Active strike-slip faulting between the Kopeh Dagh and Binalud mountains, Tectonics, 28, TC5002, doi: 10.1029/2008TC002444. Shabanian, E., O. Bellier, M. R. Abbassi, L. Siame, and Y. Farbod (2010), Plio-Quaternary stress states in NE Iran: Kopeh Dagh and Allah Dagh-Binalud mountain ranges, Tectonophysics, 480, 280-304, doi:10.1016/j.tecto.2009.10.022. Shabanian, E., O. Bellier, L. Siame, M. R. Abbassi, D. Bourlès, and Y. Farbod, The Binalud Mountains, a key piece 1 for the geodynamic puzzle of NE Iran. Submitted to Geophys. J. Int Vernant, P., et al. (2004), Present-day crustal deformation and plate kinematics in the Middle East constrained by GPS measurements in Iran and northern Oman, Geophys. J. Int., 157(1), 381-398, doi: 10.1111/j.1365-246X.2004.02222.x.

145

SYNTHESE

Synthèse

Tectonique active de la faille de Doruneh : comportement sismogénique et rôle géodynamique

1. Introduction Cette thèse porte sur une des failles décrochantes majeures du Nord-Est de l’Iran, la faille de Doruneh. Nous avons tenté dans ce travail de définir la cinématique ainsi que le rôle géodynamique de cette structure dans le contexte de la collision Arabie-Eurasie. Notre étude tectonique et géomorphologique est basée sur l'analyse d’images satellitaires, de modèles numériques de terrain et de cartes géologiques, complétée par des observations de terrain ainsi que par des mesures de cinématique de faille et de datation par des nucléides cosmogéniques. Nous avons confronté ces analyses avec les données de sismicité instrumentale et historique disponibles, ainsi qu’avec les vitesses déduites des mesures GPS. En combinant l’ensemble de ces données indépendantes, nous proposons un cadre tectonique cohérent du système de faille de Doruneh (DFS), intégré dans un nouveau modèle géodynamique du NE de l’Iran.

2. Contexte géodynamique L'Iran est une région de déformation continentale active localisée dans le domaine de la convergence Arabie-Eurasie (chapitre I). Celle-ci est essentiellement accommodée par la subduction le long du Makran et la collision le long des chaînes du Zagros, de l’Alborz et du Kopeh Dagh. Dans ces domaines de collision, la déformation est absorbée par des systèmes de failles inverses et décrochantes. Les données GPS indiquent un mouvement vers le nord de l'Arabie par rapport à l'Eurasie à une vitesse d’environ 26±2 mm/an à la longitude de ~59°E (Sella et al, 2002; McClusky et al, 2003; Vernant et al, 2004; Reilinger et al., 2006). A l’Est de l’Iran, ce mouvement vers le nord se traduit par une zone de cisaillement dextre entre l’Iran Central et le bloc de Helmand (Afghanistan). Les arguments tectoniques (Tirrul et al, 1983; Regard et al, 2004; Walker et Jackson, 2004; Meyer et Le Dortz, 2007) et géodésiques (Vernant et al, 2004; Masson et al, 2007; Reilinger et al, 2006) indiquent que ce mouvement vers le nord de l’Iran Central et du Bloc de Lut est distribué le long de systèmes de failles décrochantes dextres d’orientation N-S. Les mesures GPS ont permis de calculer une vitesse globale de cisaillement dextre de 16±2 mm/an entre l’Iran Central et le bloc de

149 Synthèse

Helmand. Une partie de ce cisaillement, 7±2 mm/an, est concentrée à la limite est de bloc de Lut (Vernant et al, 2004; Masson et al, 2007; Tavakoli, 2007). Au Nord-Est de l’Iran, les données géodésiques (Vernant et al., 2004; Masson et al., 2007; Tavakoli, 2007) et géologiques (Shabanian et al., 2009a) dans la zone de déformation du Kopeh-Dagh indiquent un mouvement vers le nord de la partie ouest du Kopeh Dagh par rapport à l’Eurasie. Ce mouvement est essentiellement localisé le long du système de failles décrochantes de Bakharden-Quchan (Shabanian et al, 2009a). A la latitude de 35°N, la faille de Doruneh sépare deux domaines de déformation différents, le domaine du Kopeh Dagh et du Binalud au nord, et le bloc de Lut limité par les failles N-S dextres au sud. La faille de Doruneh a été décrite pour la première fois par Wellman (1966) comme la plus grande faille décrochante senestre affectant le plateau Iranien. Elle s'étend sur une longueur d’environ 600 km, du désert de Great Kavir à l’ouest jusqu’à la frontière Iran-Afghanistan à l’est. La géométrie convexe vers le nord de la faille a conduit de nombreux auteurs à proposer un modèle de rotation rigide de blocs pour expliquer la dynamique et la cinématique de l’ensemble de ce domaine de déformation (Jackson et McKenzie, 1984; Walker et Jackson, 2004; Fattahi et al, 2007). Selon ce modèle, la faille de Doruneh suit une rotation horaire afin d’accommoder le cisaillement dextre observé le long des failles majeures d’orientation N-S (ex. failles de Dehshir, Anar, Nayband et Nehbandan). Ainsi, le mouvement senestre de la faille de Doruneh résulterait de cette rotation. La courbure maximale de cette faille est présentée comme résultant du taux maximum de glissement au niveau de la zone de suture du Sistan. Au chapitre II, nous montrons que ce modèle ne peut pas être accepté car les parties au nord et au sud du DFS ne peuvent pas être considérées comme des blocs rigides et homogènes. De plus, l’augmentation vers l’ouest des taux de cisaillement de l’Iran central par rapport à l’Eurasie ne peut pas expliquer la géométrie courbe actuelle de la faille Doruneh. En effet, partant d’une faille originellement orientée E-W, les taux de vitesse plus importants à l’ouest qu’à l’est devraient se manifester par une orientation actuelle de faille NNW à NW. En fait, la cinématique de la faille de Doruneh ainsi que sa géométrie globalement perpendiculaire à la direction de la convergence Arabie-Eurasie sont beaucoup plus complexes qu’indiqué dans ce modèle car contrôlées par de nombreux facteurs. Les forces liées à la cinématique aux frontières de plaques, l’héritage structural et les interactions avec d’autres zones de failles en sont les principales raisons. Dans ce contexte, le modèle de rotation de blocs ne tient pas compte de variations de cinématique le long des systèmes de faille concernés. Par ailleurs, les taux de déplacement étant identiques à l’est et au nord-est de

150 Synthèse la frontière Arabie-Eurasie, la majeure partie de la convergence est donc transférée vers le nord par l’intermédiaire de la terminaison orientale de la faille de Doruneh.

Dans ce contexte, trois questions fondamentales se posent :

1- Comment le mouvement vers le nord de l'Iran central par rapport à l'Eurasie peut-il être accommodé par la faille de Doruneh d’orientation E-W? 2- Quel est le rôle géodynamique de la faille de Doruneh dans le transfert de la convergence Arabie-Eurasie? 3- Quel est le comportement sismogénique de la faille de Doruneh qui malgré son importance à l’échelle de l’Iran est caractérisée par une activité sismique relativement faible?

Durant ce travail de thèse, nous avons donc cherché à répondre à ces questions en appliquant une méthodologie rigoureuse et en suivant les objectifs successifs présentés ci- dessous :

1- Etablir une cartographie suffisamment détaillée des zones de failles actives afin de pouvoir en déduire une carte de segmentation ; 2- Mesurer les déplacements finis ainsi que les vitesses de glissement intégrées sur des échelles de temps de l’ordre de quelques dizaines de milliers d’années (échelle de quelques cycles sismiques) à plusieurs millions d’années (échelle de la tectonique des plaques) ; 3- Déterminer la cinématique des déformations plio-quaternaires et en déduire les états de contraintes responsables.

Ces différents objectifs scientifiques ont été poursuivis grâce à une approche pluri- disciplinaire combinant géologie structurale, morpho-tectonique, géomorphologie quantitative et datations par des nucléides cosmogéniques.

Les principaux résultats obtenus dans ce travail sont présentés ci-après.

3. Etude structurale et géométrie du système de faille Dans la première partie de cette thèse (chapitre II), nous avons combiné des approches de géologie structurale et de morpho-tectonique afin d’améliorer la cartographie de la faille de Doruneh et d’en déduire les caractéristiques de la déformation tout au long de cette structure. Ces approches ont été développées grâce à des analyses d’images satellitaires à différentes

151 Synthèse

échelles (Landsat ETM+, SPOT5, QuickBird et GeoEye3), combinées à celles du modèle numérique de terrain SRTM, puis calibrées et confirmées par des observations de terrain. Dans une première étape, notre cartographie nous a permis de subdiviser la faille de Doruneh en deux systèmes de faille indépendants : le système de faille de Great Kavir (GKFS) à l’ouest (270 km de long), et le système de faille de Doruneh (DFS) à l’est (360 km de long). Ces deux systèmes sont séparés par une discontinuité structurale à la longitude de 56.75°E où les dépôts quaternaires ne sont pas affectés. De plus, la terminaison ouest du DFS est connectée à une faille inverse perpendiculaire d’orientation N160°E, la faille de Khartouran. Cette faille absorbe le mouvement vers l’ouest du bloc situé au nord du DFS et limite le bassin de Great Kavir à l’est, on peut donc la considérer comme la terminaison ouest du DFS. L’essentiel de ce travail de thèse s’intéresse plus particulièrement au DFS. Le DFS a été cartographiée précisément afin d’identifier les discontinuités (zones de relais, courbures…) séparant les différents segments de faille. Nous avons ainsi noté des variations importantes des caractéristiques structurales, géométriques et cinématiques le long du système à l’origine d’un nouveau modèle de segmentation (chapitre II). Selon ce modèle, le DFS avec sa forme convexe vers le nord n’est pas une structure uniforme de ~400 km de longueur mais un accident constitué de trois zones de failles bien distinctes. Chacune de ces zones de faille est caractérisée par ses propres caractéristiques géométriques, géomorphologiques et cinématiques. Ces trois zones de failles sont : (1) la zone de faille Ouest (WFZ) de ~70 km de long caractérisée par une cinématique oblique senestre à composante inverse en relation avec un régime tectonique en transpression, (2) la zone de faille centrale (CFZ) de ~185 km de long caractérisée par une cinématique purement décrochante senestre et une déformation principalement localisée sur une seule trace de faille, (3) la zone de faille Est (EFZ) d’une longueur de ~100 km qui est très distribuée et constituée de failles imbriquées caractérisées par une composante inverse non négligeable voire dominante. Étant donné que les caractéristiques géométriques et cinématiques de chaque zone de faille sont différentes, la déformation ne peut pas être uniformément distribuée le long du DFS. Par conséquent, nous avons proposé un modèle cinématique qui permet d’expliquer les variations des caractéristiques structurales le long du DFS (chapitre II, Farbod et al, 2011). Selon ce modèle, le DFS accommode le mouvement vers le Nord de l'Iran Central et du Bloc de Lut par rapport à l'Eurasie par une cinématique complexe.

3 Sources des images satellites : USGS-GLCF pour Landsat, programmes ISIS n°76 pour SPOT, GoogleEarth pour GeoEye et QuickBird

152 Synthèse

Les cinématiques des WFZ et EFZ résultent du régime tectonique régional caractérisé par une direction de compression d’environ N40°E. Dans ce régime, le partitionnement de la deformation permet à la CFZ de rester purement décrochante senestre tandis que le mouvement vers le nord est accommodé par une série de failles inverses parallèles situées plus au nord. Notre modèle permet d'expliquer comment de grandes failles décrochantes tels que le DFS peuvent accommoder des mouvements tectoniques de blocs crustaux perpendiculaires à leurs orientations.

4. Quantification de la déformation : déplacement fini et vitesse de déplacement Dans la deuxième partie de cette thèse (chapitre III), nous nous sommes attachés à calculer les vitesses de déplacement du DFS par l’analyse de marqueurs géomorphologiques décalés et datés par la méthode des cosmonucléides produits in-situ. Dans un premier temps, nous avons fait une analyse géomorphologique et géométrique détaillée des marqueurs quaternaires décalés. Il s’agit essentiellement de cônes alluviaux et des réseaux de drainage associés. Cette analyse a été réalisée sur des images satellitaires à très haute résolution puis validée par des observations de terrain sur des secteurs clés. Elle nous a permis de subdiviser les cônes alluviaux quaternaires en trois générations significatives à l’échelle régionale que nous avons nommé surfaces Q1, Q2 et Q3, de la plus jeune à la plus ancienne. Afin d'illustrer les variations spatiales du décalage horizontal senestre le long du DFS, nous avons reconstitué la déformation sur 67 décalages de marqueurs géomorphologiques le long des WFZ et CFZ, comprenant 36, 24 et 7 mesures sur Q1, Q2 et Q3 respectivement. La distribution de ces décalages le long de ces zones de failles montre que le décalage horizontal cumulé diminue fortement de CFZ à WFZ. Cette observation est cohérente avec la cinématique oblique de la WFZ impliquant un décalage horizontal moindre que pour la CFZ purement décrochante. Le mouvement oblique sur la WFZ est confirmé pour la période actuelle par des travaux d’interférométrie radar (Pezzo et al, 2012). Grâce à la méthode de datation par âge d’exposition aux rayons cosmiques (10Be et 36Cl produit in situ), l’abandon des trois surfaces de cône a pu être daté régionalement à 12.3 ± 2.9 ka (Q1), 36.5 ± 6.3 ka (Q2) et 101.8 ± 9.5 ka (Q3). La combinaison des âges d'abandon et des décalages cumulés nous a permis d'estimer les vitesses de déplacement horizontal sur les 67 sites indépendants le long des WFZ et CFZ.

153

Synthèse

FZ), estimé pour des périodes de temps de ~12 (Q1), ~36 (Q2)

. Distribution du taux de glissement le long des zones Ouest et Centre du DFS (WFZ et C

1

Figure et ~100 (Q3) ka.

154 Synthèse

Pour l’ensemble de ces deux zones, la variation longitudinale de la vitesse mesurée sur chacune des périodes de temps de ~36 et ~100 ka montre un profil semblable (Fig.1). Cette observation suggère que pour les deux zones de failles, CFZ et WFZ, et sur les derniers 100 ka : (1) le comportement du glissement est caractéristique et (2) les vitesses de déplacement horizontal restent à peu près constantes. Cependant, pour la période Holocène (~12 ka), le comportement du glissement de la faille sur la CFZ semble plus complexe. En effet, les marqueurs montrent que la CFZ est subdivisée en deux segments caractérisés par une distribution symétrique des déplacements. Ces segments sont séparés par une zone de «gap» de ~3 km de long où les surfaces quaternaires Q1 ne sont pas affectées par la faille. Une vitesse maximale de déplacement horizontal de ~5.3 mm/an a pu être estimée le long de la CFZ pour cette période. Pour la CFZ, une vitesse maximale de ~8 mm/an est atteinte à mi-longueur de la zone de faille et diminue vers les deux extrémités. La vitesse horizontale maximum le long de la WFZ est de ~2 mm/an. Néanmoins, cette vitesse ne représente pas la vitesse de glissement réelle sur cette zone de faille qui contient également une composante verticale importante. Pour mesurer cette composante, nous aurions besoin de caractériser la géométrie de la faille ainsi que le vecteur glissement sur le plan de faille, paramètres que nous n’avons pas pu obtenir.

5. Initiation du décrochement le long du système de faille de Doruneh Le long du DFS, un décalage latéral maximum de 840±70 m est enregistré par le marqueur quaternaire le plus ancien (Q3). Toutefois, le déplacement total cumulé de la faille reste inconnu. Indirectement, grâce à des arguments géologiques, nous avons tenté de déterminer un âge pour la mise en place du décrochement sur le DFS, en utilisant notamment les relations et interactions entre les segments de faille et les autres structures géologiques impliquées dans le système de failles. Le Miocène est une période où les mouvements sur le DFS comportaient une composante verticale dominante qui a provoqué l'affaissement du bassin néogène au sud de la faille par rapport au domaine septentrional constitué de paléo-reliefs pré-oligocènes. En terminaison orientale du DFS, les formations Mio-Pliocènes sont plissées et faillées. Ces déformations compressives sur l’EFZ peuvent être expliquées par le transfert du mouvement décrochant senestre de la CFZ. Ceci impliquerait une mise en place du décrochement de la DFS après le Miocène, au cours du Pliocène. D'autre part, la faille de Great Kavir, située du DFS, déplace latéralement des séries et

155 Synthèse structures miocènes. Cela indiquerait un âge maximum pliocène (~5 Ma) à l'initiation du déplacement latéral sénestre le long de la faille de Great Kavir. Ces estimations d'âge correspondent bien à l’âge proposé (~ 4 Ma) pour la réorganisation régionale de la tectonique du Nord de l’Iran, qui voit la mise en place de grands décrochements tels que ceux du Kopeh Dagh (Shabanian et al, 2009a, 2009b). En conclusion, le mouvement décrochant des failles de Doruneh et de Great Kavir s’est mis en place à environ 5 Ma, probablement en même temps que la réorganisation tectonique majeure signalée en Iran septentrional.

6. Comportement sismogénique et évaluation de l’alea sismique Le DFS constitue l'un des grands décrochements actifs dans la zone de collision Arabie- Eurasie. Malgré une activité quaternaire bien caractérisée morphologiquement, aucun tremblement de terre historique ou instrumental important (M ≥ 6,5) n’a été enregistré le long de ce système de failles. Nos données et résultats confirment que les grandes failles décrochantes ne rompent pas sur toute leur longueur pendant un tremblement de terre mais sont segmentées et peuvent être divisées en plusieurs secteurs de failles discrets. Le DFS comprend trois zones de failles : les WFZ, CFZ et EFZ, avec des cinématiques et des géométries de surface bien distinctes. Toutefois, la distribution de vitesse sur les zones WFZ et CFZ montre un comportement de glissement caractéristique à long terme. La CFZ peut être subdivisée en deux segments de ~80 et ~100 km à l'ouest et l'est séparés par une zone de « gap » qui n'a pas rompue pendant les derniers ~12 ka. Les longueurs des segments sismogènes sur les zones de faille CFZ et WFZ varient de ~70 à ~100 km, ce qui pourrait produire des tremblements de terre caractéristiques de magnitude comprise entre 7.2 et 7.4. De plus, des décalages co-sismiques de ~4m ont été mesurés le long d’un segment de ~100 km. Si cela correspond à un décalage co-sismique maximum, alors cette mesure est en accord avec un événement de magnitude Mw≈7.4. Notre estimation de vitesse de glissement de ~5.3 mm/an pour l’Holocène couplée à ce décalage co-sismique nous permet de proposer un temps de récurrence d’environ 750 ans pour ces gros séismes.

7. Rôle géodynamique La terminaison orientale du DFS (EFZ) a un rôle géodynamique particulier dans l’accommodation de la convergence Arabie-Eurasie. Pour le comprendre, nous avons réalisé une étude morphostructurale essentiellement basée sur l’interprétation des images satellitaires et des modèles numériques de terrain (SRTM). Nos analyses géomorphologiques et

156 Synthèse géologiques indiquent que l’EFZ est une zone principalement caractérisée par des failles inverses formant une zone de déformation très distribuée dont l’agencement correspond à une terminaison en queue de cheval du DFS. Cette cinématique en faille inverse dominante résulte en première approximation du mouvement senestre le long de la CFZ et de la géométrie de cette terminaison d’orientation NW-SE. Par ailleurs, nous avons observé des décalages significatifs dextres de marqueurs géologiques et géomorphologiques dans l’EFZ sur des failles orientées NNW. En interprétation de ces décalages, nous proposons que la terminaison orientale du DFS est intégrée dans une zone de cisaillement dextre d’orientation NNW-SSE située entre l’Iran Central et l’Eurasie.

Figure 2. (a) Carte simplifiée des failles active (d’après Shabanian et al, 2009a;. 2009b et cette étude) et des principaux domaines tectoniques de NE de l’Iran. (b) Modèle schématique montrant la limite crustale de la convergence Arabie-Eurasie. L'interaction entre des failles dextres séparées par des zones de relais en compression et en transpression permet à l'Iran centrale d’être translatée vers le nord par rapport à l'Eurasie. La flèche blanche montre le mouvement global de l’Iran Central-Eurasie. La flèche jaune indique la direction nord- est de la contrainte maximale σ1 (compression) dans le nord de l’Iran. Les abréviations sont les suivantes: MKDF, Main Kopeh Dagh Fault ; BQ, Bakharden-Quchan Fault; CF, Chakaneh Fault; NF, Neyshabur Fault; KF, Kashafrud Fault; GKFS, Great Kavir Fault System ; DFS, Doruneh Fault System ; EFZ, Eastern Fault Zone ; KHF, Khaf Fault; NHF, Neh Fault.

La limite orientale de la convergence Arabie-Eurasie est localisée le long du système de faille dextre de Neh à l’est du bloc de Lut. La limite nord-est de cette collision est localisée

157 Synthèse dans les chaines du Kopeh Dagh et du Binalud le long d’une série de systèmes de failles dextres d’orientation NNW-SSE (i.e., systèmes de failles Bakharden-Quchan et Chakaneh). Ces systèmes de failles dextres se connectent aux failles inverses d’orientation NW-SE (i.e., Faille de Neyshabour) qui forment une terminaison inverse à composante dextre (Shabanian et al, 2009a; 2009b). L’EFZ est située entre ces deux limites crustales. Elle accommode le mouvement vers le nord de l’Iran central par rapport à l’Eurasie par du cisaillement dextre le long de failles d’orientation NNW et du raccourcissement sur des failles d’orientation NW. Ces failles sont distribuées sur l’ensemble de la zone de faille. Cette cinématique est cohérente avec la direction régionale de la compression (σ1 NE). L’ensemble de ces résultats nous permet de proposer un nouveau modèle géodynamique de déformation pour le DFS (Fig. 2). Selon ce modèle, l’EFZ est une zone de relais en transpression dextre située sur la limite de plaques Arabie-Eurasie. Elle participe au transfert vers le Nord de l’Iran central par rapport à l’Eurasie. La CFZ et la WFZ, situées sur le bloc d'Iran central, sont aussi translatées vers le Nord le long de cette limite crustale. Cela explique pourquoi les mesures GPS à la station de KASH, située au centre de la DFS, montrent un mouvement vers le Nord.

8. Conclusions et perspectives Les données et résultats de ce travail doctoral nous ont permis de préciser la géométrie, la cinématique et le taux de déformation de la faille de Doruneh et plus particulièrement du DFS. Nous avons également éclairci les relations temporelles et spatiales entre ce système de faille et les domaines de déformation environnants, ainsi que son rôle dans l’accommodation de la convergence Arabie-Eurasie. Nos résultats apportent donc de nouvelles contraintes dans la compréhension de la collision Arabie-Eurasie et plus généralement sur la place des décrochements dans l’accommodation de la convergence. Afin de préciser et compléter notre modèle géodynamique, des acquisitions de terrain seraient à réaliser dans l’EFZ et dans la région située entre celle-ci et la faille de Neh plus au sud. Nous pourrions ainsi améliorer notre connaissance de la cinématique des failles actives de l'EFZ. Une estimation de la composante inverse du mouvement sur la WFZ nous permettrait de quantifier les déplacements finis ainsi que les vitesses de mouvement réels. Des tranchées de paléosismologie seraient aussi indispensables pour améliorer notre connaissance du comportement sismogénique de la DFS et l'évaluation de son aléa sismique.

158 Synthèse

Bibliographie Farbod, Y., O. Bellier, E. Shabanian, and M. R. Abbassi (2011), Geomorphic and structural variations along the Doruneh Fault System (Central Iran), Tectonics, 30, TC6014, doi:10.1029/2011TC002889. Fattahi, M., R. T. Walker, M. M. Khatib, A. Dolati and A. Bahroudi; (2007), Slip-rate estimate and past earthquakes on the Doruneh fault, eastern Iran, Geophys. J. Int., 168, 691–709 doi: 10.1111/j.1365-246X.2006.03248.x Jackson, J., and D. Mckenzie (1984), Active tectonics of the Alpine-Himalayan Belt between turkey and Pakistan., Geophys. J. R. astr. Soc., 77(1), 214-245. Masson, F., M. Anvari, Y. Djamour, A. Walpersdorf, F. Tavakoli, M. Daignieres, H. Nankali, and S. Van Gorp (2007), Large-scale velocity field and strain tensor in Iran inferred from GPS measurements: new insight for the present-day deformation pattern within NE Iran, Geophys. J. Int., 170, 436-440, doi: 10.1111/j.1365-246X.2007.03477.x. McClusky, S., R. Reilinger, S. Mahmoud, D. Ben Sari, and A. Tealeb (2003), GPS constraints on Africa (Nubia) and Arabia plate motions, Geophys. J. Int., 155(1), 126-138. Meyer, B., and K. Le Dortz (2007), Strike-slip kinematics in Central and eastern Iran: Estimating fault slip rates averaged over the Holocene, Tectonics, 26, TC5009, doi: 10.1029/2006TC002073. Pezzo, G., C. Tolomei, S. Atzori, S. Salvi, E. Shabanian, O. Bellier, and Y. Farbod (2012), New kinematic constraints of the western Doruneh fault, north-eastern Iran, from interseismic deformation analysis, Geophys. J. Int., in press. Regard, V., O. Bellier, J.-C. Thomas, M. R. Abbassi, J. Mercier, E. Shabanian, K. Feghhi, and S. Soleymani (2004), Accommodation of Arabia-Eurasia convergence in the Zagros- Makran transfer zone, SE Iran: A transition between collision and subduction through a young deforming system, Tectonics, 23, TC4007, doi: 10.1029/2003TC001599,. Reilinger, R., et al. (2006), GPS constraints on continental deformation in the Africa-Arabia- Eurasia continental collision zone and implications for the dynamics of plate interactions, J. Geophys. Res., 111, B05411, doi: 10.1029/2005JB004051. Sella, F. G., H. T. Dixon, and A. Mao (2002), Revel: A model for Recent plate velocities from space geodesy, J. Geophys. Res., 107(B4), ETG 11. Shabanian, E., O. Bellier, L. Siame, N. Arnaud, M. R. Abbassi, and J.-J. Cocheme (2009b), New tectonic configuration in NE Iran: Active strike-slip faulting between the Kopeh Dagh and Binalud mountains, Tectonics, 28, TC5002, doi: 10.1029/2008TC002444. Shabanian, E., L. Siame, O. Bellier, L. Benedetti, and M. R. Abbassi (2009a), Quaternary slip rates along the northeastern boundary of the Arabia–Eurasia collision zone (Kopeh Dagh Mountains, Northeast Iran), Geophys. J. Int., 1-23, doi: 0.1111/j.1365- 246X.2009.04183.x. Tirrul, R., I. R. Bell, R. J. Griffis, and V. E. Camp (1983), The sistan suture zone of eastern Iran, Geological Society of America Bulletin, 94, 134-150. Vernant, P., et al. (2004), Present-day crustal deformation and plate kinematics in the Middle East constrained by GPS measurements in Iran and northern Oman, Geophys. J. Int., 157(1), 381-398, doi: 10.1111/j.1365-246X.2004.02222.x. Walker, R., and J. Jackson (2004), Active tectonics and late Cenozoic strain distribution in central and eastern Iran, Tectonics, 23, TC5010, doi: 10.1029/2003TC001529. Wellman, H. W. (1966), Active wrench faults of Iran Afghanistan and Pakistan, Geologische undschau, 55, 716-735.

159

APPENDIX

Appendix

New kinematic constraints of the western Doruneh fault, north-eastern Iran, from interseismic deformation analysis4

Giuseppe Pezzo1, Cristiano Tolomei 1, Simone Atzori 1, Stefano Salvi 1, Esmaeil Shabanian 2, Olivier Bellier 2 and Yassaman Farbod 2.

1 Istituto Nazionale di Geofisica e Vulcanologia, Via di Vigna Murata, 605, 00143 Rome, Italy [email protected]

2 Aix-Marseille Université, CEREGE, UMR 6635, 13545 Aix en Provence cedex 4, France

Summary We used the SBAS DInSAR analysis technique to estimate the interseismic deformation along the Western part of the Doruneh Fault System (DFS), north-eastern Iran. We processed 90 ENVISAT images from four different frames from ascending and descending orbits. Three of the ground velocity maps show a significant interseismic signal. Using a simple dislocation approach we model 2-D velocity profiles concerning three InSAR dataset relative to the western part of the DFS, obtaining a good fit to the observations. The resulting model indicates that a slip rate of ~5 mm yr-1 accumulates on the fault below 10 km depth, and that in its western sector the Doruneh fault is not purely strike-slip (left-lateral) as in its central part, but shows a significant thrust component. Based on published geological observations, and assuming that all interseismic deformation is recovered with a single event, we can estimate a characteristic recurrence interval between 630 and 1400 yr.

Keywords: DInSAR SBAS, Strike slip fault, Arabia-Eurasia convergence, Interseismic strain, Fault modelling, seismic cycle

4 Paper in press in Geophysical Journal international -GJI-S-11-0526L.R2

163

Figure captions

CHAPTER I: GENERAL PRESENTATION

Figure 1. Simplified structural map of Iran located in the Arabia-Eurasia collision, with main tectono- structural division of the Iranian plateau ...... 16

Figure 2. (a) Distribution of earthquake epicenters from historical records (3000 B.C. to 1962 A.D.) (b) Distribution of instrumentally recorded earthquake epicenters (1964 to 1998) from the USGS catalogue. The distribution is similar in a and b and show that the seismicity is broadly confined to the Zagros, Alborz, and Kopeh Dagh mountain belts and to narrow N–S zones surrounding the Dasht-e- Lut...... 17

Figure 3. GPS velocity field show that approximately N-S shortening between Arabia and Eurasia is accommodated throughout Iran (after Vernant et al., 2007)...... 18

Figure 4. Principal deformation domains and structural units in Central and northeastern Iran overlain on GTOPO30 shaded relief. The abbreviations are : MKDF, Main Kopeh Dagh fault; BQFS, Bakharden-Quchan Fault System; CF, Chakaneh Fault; NF, Neyshabur Fault; DFS, Doruneh Fault System; GKFS; Great Kavir Fault System...... 21

CHAPTER II: GEOMORPHIC AND STRUCTURAL VARIATIONS ALONG THE DORUNEH FAULT SYSTEM (NE IRAN)

Figure 1. GTOPO30 topographic image of central and northeastern Iran showing the location of the Doruneh Fault together with principal deformation domains and structural units mentioned in the text. Black arrows are GPS horizontal velocities (mm/yr) in a Eurasia-fixed reference frame (YAZT station, (Masson et al., 2007)). White arrows, south of the Doruneh Fault are geodetic-derived tectonic motions after Reilinger et al. (2006), while the white arrow in the Kopeh Dagh is the geologic rate of the western Kopeh Dagh – Eurasia northward motion (Shabanian et al., 2009a). The box in the upper left inset shows the location in the Arabia–Eurasia collision zone. Grey arrows and associated numbers represent Arabia–Eurasia plate velocities (mm/yr) after Reilinger et al. (2006)...... 35

Figure 2. (a) Simplified geological map of the region affected by the Doruneh (DFS) and Great Kavir (GKFS) fault systems. Geological units are from Huber (1977).(b) Examples of pull-apart basins formed between releasing left-handed stepovers along the GKFS that cuts through dome and basin structures in the Miocene deposits (Quickbird image - GoogleEarth). (c) Detailed geological map of the Doruneh Fault System (modified after Eftekhar-Nezhad et al. (1976) and Alavi-Naini et al. (1992)) superposed on shaded relief image of the area (SRTM digital topographic data). Fault traces in (a) and (c) are based on geomorphic and structural analyses of SPOT5 and LANDSAT ETM+ (this study). Abbreviations are: BF, Bijvard Fault; DQF, Dahan-Qaleh Fault; KF, Kharturan Fault; KHF, Khaf Fault; JTF, Jangal Thrust Fault...... 37

Figure 3. Historical and instrumental seismicity of the DFS and surrounding areas. Focal mechanisms are mainly taken from the Harvard catalogue (http://www.globalcmt.org/CMTsearch.html) and Jackson and McKenzie (1984). Epicenters are from the ISC - EHB Bulletin (Internatl. Seis. Cent., Thatcham, United Kingdom, 2009 - http://www.isc.ac.uk). The regions of maximum destruction are based on Ambraseys and Melville (1982)...... 40

165

Figure 4. Simplified morphotectonic map of the rupture zone along the CFZ, in the Forsheh area. See text for more information...... 42

Figure 5. (a) General fault map of the DFS comprising three discrete fault zones. Rose diagrams represent the predominant orientation of the fault zones deduced from the statistic analysis of individual fault segments, which have been mapped on SPOT5 satellite images. (b) along-strike variations in the wideness of the DFS. The initiation point of the diagram is the western termination of the WFZ. (c) Spatial variations in the geometric arrangement of fault segments within the DFS are illustrated by plotting the straight length of individual fault segments versus the total overlap length of each segment that is covered by other nearby faults. Thick lines indicate the average value of the segment overlap in each fault zone. (d) along-strike variations in cumulative left-lateral offsets of the DFS. Uncertainty of the offset measurement is shown by error bars...... 44

Figure 6. SPOT5 image centered on two pull-apart basins (hatched areas) formed in releasing offsets between overlapping segments of the WFZ; see Figure 7 for the location...... 46

Figure 7. (a) Shaded relief map (SRTM data) of the WFZ. The trace of the WFZ joins at right angle the Kharturan reverse fault zone (KF), which marks a sharp topographic edge at the NE boundary of the Dasht-e Kavir (Great Kavir desert) absorbing the relative westward motion (left-lateral faulting) of the WFZ’s northern block. “P” in Figure 7a indicates a pop-up structure formed in a restraining bend along the WFZ. (b) A profile of the fault shown in Figure 7d indicating the reverse component of faulting along the WFZ. (c) Field photograph of a seasonal stream incised in Q1 alluvial fan surface obliquely displaced by the fault indicating coeval vertical and left-lateral offsets of ~6.5 m and ~5.5 m. (d) Field photograph of the fault scarp along the WFZ. The accurate location of field photographs is: Figure 7b, 35.2205°N - 57.5577°E; Figure 7c, 35.22106°N - 57.5603°E; Figure 7d, 35.2210°N - 57.5606°E...... 47

Figure 8. Geomorphic and geological evidence of left-lateral faulting along the Dahan-Qaleh Fault (Quickbird image - GoogleEarth). Upper image illustrates cumulative left-lateral offset along the fault running through folded rock units and Quaternary landforms. Lower image shows left-lateral displacement of the fault in the folded Neogene deposits. Trace of the fault is shown by dotted black lines. White lines marks offset markers. See Figure 7 for the location...... 48

Figure 9. Simplified fault map of the CFZ and the surrounding areas. (a) and (b) are shaded relief maps based on SRTM digital topographic data indicating two examples of uplifted areas along the CFZ; the uplifted and subsided areas are marked by (+) and (-), respectively...... 49

Figure 10. Shaded relief map (SRTM digital topographic data) and simplified morphotectonic interpretation (below) of the eastern termination of the CFZ. The lower topographic profile (white line in upper panel) represents abrupt topographic changes across the fault-bounded graben (see section 4.1 for explanations) at the CFZ’s eastern termination...... 50

Figure 11. (a) Fault map of the EFZ. Quaternary fault traces are shown by thick black lines. Faults along which there is no evidence of Quaternary activity are marked by thin black lines. (b) 3D view

(Quickbird image superposed on SRTM data - GoogleEarth) of a Q3 alluvial fan that has vertically been offset on a fault within the EFZ. A topographic profile (A-A’ in Figure 11b), based on SRTM data, that illustrates vertical faulting along the westernmost fault of the EFZ. (c) A growing anticline (3D view - GoogleEarth) on the hanging wall of a reverse fault of the CFZ. Quaternary deposits were titled and uplifted on the SW flank of the fold. The trace of the fault is indicated by triangles...... 52

166

Figure 12. (a) General fault map of the DFS showing the location of sites in which left-lateral offset recorded by Q1, Q2, and Q3 alluvial fans have been measured. The most important offset measurements, which have a crucial role in the understanding of offset variations along the DFS, are shown in Figure 12b to Figure 12i. (b) A Q2 fan apex left-laterally offset along the WFZ. (c) The eastern boundary of a Q2 alluvial fan offset along the WFZ. (d) the Q2/Q3 terrace riser left-laterally offset along the CFZ representing the cumulative offset of the Q2 fan surface. (e) An eroded Q3 alluvial surface left-laterally offset along the CFZ. The initial shape of the alluvial fan cannot be reconstructed. A 630±70 m cumulative offset (sum of two offset along the northern and southern fault strands) recorded by streams incising in the fan surface have been measured as a minimum left-lateral offset at site 9. Note that each beheaded stream in the northern side has been compared with the first stream on the left in the southern side...... 54

Figure 12. (continued): (f) Streams incised in a Q2 fan surface have recorded a minimum cumulative offset of the surface. The Q1/Q2 terrace riser reveals the cumulative left-lateral offset recorded by the Q1 fan surface. The middle part of the fault strands is a pressure ridge. (g) streams incised in Q2 alluvial surfaces recorded left-lateral offset of 150±30 m along the CFZ. (h) Q3 alluvial surfaces offset along the easternmost of the CFZ. The same left-lateral offset is shown by drainages on the left of the Q3 surface. The (1’) stream segment is abandoned due to the fault offset. (i) Terrace riser between Q1 and Q2 alluvial fan surfaces has been offset along three strands of the CFZ indicating the offset recorded by the Q1 fan surface...... 55

Figure 13. (a) SPOT5 image and its morphotectonic interpretation centered on the Quch Palang area which shows the geomorphic setting of Q3 and Q1 alluvial fans along the CFZ. A – A’ topographic profile on a Q1 fan surface shows that there is no vertical displacement across the fault cutting the Q1 fan surface. B – B’ topographic profile across the Q3 Quch Palang alluvial fan indicates the topographic difference between the Q3 and Q2 geomorphic surfaces. (b) The geomorphic reconstruction of both the overall shape and main streams of the Quch Palang alluvial fan. (c) The geometric and topographic reconstruction of the Quch Palang fan. The location of apex defined using concentric arcs on the distal part of the fan is located ~880 m to the east of the initial apex. This offset value is the same than the offset recorded by main streams incising in the fan surface...... 56

Figure 14. The ~25 m cumulative offset in Q1 alluvial surfaces along the Sheh Taraz river west of Kashmar reported by Fattahi et al. (2007). See text for more information...... 58

Figure 15. The central part of the N-dipping oblique slip fault at the Siah-Kuh Mountain front in the north of the CFZ. The fault offsets (reverse left-lateral) both Quaternary deposits and south running streams. See Fig. 12a for the location...... 60

Figure 16. Results of the inversion analysis of fault kinematic data measured along the CFZ. Azimuths of σ1 (maximum stress) axis for modern stress (a) and paleostress (b) states are presented on the upper part of each figure. Numbers refer to site names detailed in Table 3. The lower hemisphere stereograms of fault planes and associated slip vectors (arrows on fault planes) together with deduced stress directions (large arrows) are shown on the lower parts. Histograms show distribution of deviation angles between the measured and calculated slip vectors (e.g., Bellier and Zoback, 1995). (c) Distribution of pitch angles measured on main fault planes within and parallel to the CFZ; about 80 per cent of pitch angles are lower than 20 degrees...... 63

Figure 17. (a) SPOT5 image of a Q1 alluvial fan (east of Forsheh village) affected by the CFZ. Two kariz lines were left-laterally offset along a coseismic rupture zone. (b) The relatively older kariz line is offset of 8±1 m, where a pull-apart through was formed due to left-lateral faulting of the fault. (c)

167

The younger kariz shows an offset of 4±0.5 m. In the same area, along another fault trace, a stream was offset of ~9 m that is consistent with the offset shown in (Figure 17c) observed along the northern fault segment...... 65

Figure 18. A simplified tectonic model to describe the present-day kinematics of the DFS based on the data and deduced results presented in this study. Yellow arrow is the GPS-derived vector (3.3 mm/yr) of active motion between the southern and northern blocks of the DFS. This velocity vector (Figure 15a) was calculated from the relative motion between SHIR and KASH stations (Figure 15b) reported by Masson et al. (2007). White arrows illustrate the theoretical direction of the contraction suggested along the WFZ and EFZ. According to our geomorphic and structural data, there is no evidence of shortening across the CFZ suggesting that the contractional component of deformation is transferred northward, and is accommodated by geological reverse faulting and folding to the north of the CFZ. It is noteworthy that the 2.5 mm/yr presented in this model is an “assumed maximum” slip rate derived from differential GPS velocities between SHIR and KASH stations, which is different from and independent to the geomorphic-derived 2.5 mm/yr slip rate proposed by Fattahi et al. (2007)...... 68

Figure 19. Simplified tectonic model to examine the hypothesis of clockwise block rotation proposed by Walker and Jackson (2004). The main faults of central and eastern Iran are shown in Figure 16a. Thick grey line indicates the hypothetical E-W geometry of the DFS suggested by Walker and Jackson (2004). To achieve the present-day geometry of the DFS the fault trace needs to rotate clockwise in the eastern part and counterclockwise in the western part. (b) The model-derived geometry of the DFS caused by eastward increasing slip rates along the N-S right-lateral faults...... 70

CHAPTER III: TEMPORAL AND SPATIAL VARIATIONS IN LATE QUATERNARY SLIP RATES ALONG THE DORUNEH FAULT SYSTEM (NE IRAN)

Figure 1. Structure of central and northeastern Iran showing the location of the Doruneh Fault System (DFS) together with principal deformation domains and structural units overlain on GTOPO30 shaded digital elevation model. Black arrows are GPS horizontal velocities (mm/yr) from Masson et al. (2007) in a Eurasia-fixed reference frame. The box in the upper left inset shows the location in the Arabia– Eurasia collision zone. Grey arrows and associated numbers represent Arabia–Eurasia plate velocities (mm/yr) after Reilinger et al. (2006)...... 83

Figure 2. Geological map of region affected by DFS simplified after Eftekhar-Nezhad et al. (1976) and Alavi-Naini et al. (1992). Fault traces are based on geomorphic and structural analysis of SPOT5 and LANDSAT ETM+ by Farbod et al., (2011). The blue lines indicate the three DFS fault zones i.e. Western Fault Zone (WFZ), Central Fault Zone (CFZ) and Eastern Fault Zone (EFZ) ...... 85

Figure 3. (a) Sites location of left-lateral offset measurements observed on Quaternary geomorphic markers along DFS. Details on the cumulative left-lateral offset measurements (numbered) are given in Table 1. Sites with asterix have been dated by cosmogenic method. (b) Along-strike variations of the cumulative left-lateral offsets, recorded by the fan surfaces Q1, Q2 and Q3. Uncertainty of the offset measurement is shown by error bar...... 88

Figure 4. (a) GeoEye satellite image (from Google Earth, ©2012 Google, ©2012 GeoEye) of the DFS between 58°.4E and 58°.442E longitude where the Q1 alluvial fans have not been cut by the fault. (b) and (d) show enlargement of the fault trace west and east of this zone, respectively. (c) example of the Q1 alluvial fan not affected by the fault...... 89

168

Figure 5. (a) Azghand and (b) Khalilabad sampling sites on the Q1 fan surfaces. Left panel show the geomorphologic interpretation of Quaternary inset abandonment fan surfaces overlain on GeoEye images (from Google Earth, ©2012 Google, ©2012 GeoEye). The white circles show the position of the surface sampling. Labels are the in-situ produced 10Be exposure age as presented in Table3. Right panel show field photographs of the relatively well-preserved sampled part of the Q1 fan surface. (c) and (d) are the age probability distribution of 10Be exposure ages for the Azghand and Khalilabad sites, respectively. The thick curve corresponds to the Gaussian distribution of summed ages and the thin grey curves represent the age probability for each individual sample. (e) Age probability distribution of all Azghand and Khalilabad samples, showing two populations of ages. Weighted mean ages of 12.3±2.9 and 23.8±4.6 ka are calculated by Chi-square test for each population. The thin blue curve represents the age rejected in chi-square test. The conservative 2σ uncertainties are calculated for the weighted mean ages...... 94

Figure 6. (a) Azghand, (b) Mazdeh and (c) Anabad sampling sites. Top panel show the Q2 sampled fan surface on GeoEye images (from Google Earth, ©2012 Google, ©2012 GeoEye). White circles are the sample positions and labels indicate the 10Be and 36Cl exposure ages and the offset measurements. Middle panel show field photographs of the sampled surface. Bottom panel is the age probability distribution of the in situ-produced 10Be (d and e) and 36Cl (f) exposure ages. Thick black curves are Gaussian probability sum of the ages, and grey thin curves represent the age probability for each individual sample in each site. (g) Age probability distribution of all 10Be and 36Cl samples showing two populations of exposure ages. The weighted mean age of 36.5±6.3 and 50.5±3 ka are calculated by Chi-square test for each population. The thin blue curves represent the ages rejected in chi-square test and the thin red curve represents the outlier sample which has not been included in the total statistics. The conservative 2σ uncertainties are calculated for the weighted mean ages...... 96

Figure 7. (a) Nay, (b) Anabad and (c) Doruneh sampling sites. Left panel show the geomorphologic interpretation of Quaternary inset abandonment fan surfaces overlain on GeoEye images (from Google Earth, ©2012 Google, ©2012 GeoEye). White circles are surface sample locations and labels indicate the 10Be and 36Cl exposure ages and the offset measurements. Right panel show field photographs of the sampled Q3 surface. Bottom panel is the age probability distribution of Anabad (d) and Doruneh (e) sites. (f) Age probability distribution of together 10Be and 36Cl samples which show two populations of exposure ages. The weighted mean age of 72.7±6.7 and 101.8.5±9.5 ka are calculated by Chi-square test for each population. The thin red curve represents the outlier sample which has not been included in the total statistics. The conservative 2σ uncertainties are calculated for the weighted mean ages...... 98

Figure 8. Age probability distribution of the Q1 and Q2 fan surfaces, calculated with a denudation rate of 1m/Ma. Distribution shows two populations of exposure ages on both surfaces. The weighted mean ages are calculated by Chi-square test with 2σ uncertainties. The thick curve corresponds to the Gaussian probability sum of the ages and the thin grey curves represent the age probability for each individual sample. The thin blue curves represent the ages rejected in chi-square test and the thin red curve represents represent the outlier samples which have not been included in the total statistics. .. 100

Figure 9. (a) 10Be age probability distribution of Q1 fan surface in Azghand site showing two ages population ranging from ~13 to ~16 ka and ~21 to ~24 ka. (b) 10Be age probability distribution of Q2 abandonment fan surface in Azghand site showing an age population between ~27 and ~30 ka and two older ages of 43.6±1.3 and 50.3±1.5 ka. (c) Age probability distribution of both Q1 and Q2 inset abandonment surfaces. Black curve is the Gaussian probability sum of the ages indicating three populations of exposure age...... 101

169

Figure 10. Upper panel: schematic model of the three abandonment inset surfaces. The surface clasts of upper (older) surface could have fallen to the lower (younger) surface during flooding periods. Lower panel: 10Be and 36Cl exposure ages of the three Q1, Q2 and Q3 fan surfaces along the DFS. Red boxes are the 2σ uncertainties for the weighted mean age of each surface...... 102

Figure 11. (a), (b) and (c) Overall slip rate distribution along the central and western part of the DFS, recorded by the Q3, Q2 and Q1 fan surfaces, respectively. (d) Along-strike variation of the slip rates estimated from all abandonment surfaces over three time periods of ~12, ~36 and ~100 ka...... 106

CHAPTER IV: THE ARABIA-EURASIA COLLISIONAL BOUNDARY IN NE IRAN, THE EASTERN TERMINATION OF THE DORUNEH FAULT SYSTEM

Figure 1. Principal deformation domains and structural units in Central and northeastern Iran overlain on GTOPO30 shaded digital elevation model. Black arrows are GPS horizontal velocities (mm/yr) from Masson et al. (2007) in a Eurasia-fixed reference frame. The box in the upper left inset shows the location in the Arabia–Eurasia collision zone. Grey arrows and associated numbers represent Arabia– Eurasia plate velocities (mm/yr) after Reilinger et al. (2006)...... 120

Figure2. (a) Geological map of region affected by DFS simplified after Eftekhar-Nezhad et al. (1976) and Alavi-Naini et al. (1992). Fault traces are based on geomorphic and structural analysis of SPOT5 and LANDSAT ETM+ by Farbod et al., (2011). The blue lines indicate the three DFS fault zones i.e. Western Fault Zone (WFZ), Central Fault Zone (CFZ) and Eastern Fault Zone (EFZ). Abbreviations are: BF, Bijvard Fault; DQF, Dahan-Qaleh Fault; KF, Kharturan Fault; KHF, Khaf Fault; JTF, Jangal Thrust Fault. (b) Results of the inversion analysis of fault kinematic data measured in the Quaternary deposits along the CFZ...... 121

Figure 3. Shaded relief map (SRTM data) of the Eastern Fault Zone (EFZ). The general E-trending of the CFZ takes a ~30° (clockwise) bend at longitude 59.5°E. The geometry and kinematic of the fault change drastically in the EFZ. KHF, Khaf Fault; JF, Jangal Fault...... 123

Figure 4. (a) Geologic map (after Alavi Naini et al., 1982) showing the evidence of right-lateral faulting at longitude of ~60°E. The N145°E-trending right-lateral fault displaced the Eocene- Oligocene outcrops and join to the oblique to reverse NW-trending faults further south. The Oligocene fold axes show a right stepping arrangement with an oblique orientation with respect to the fault trace and confirm the right-lateral faulting along this fault. (b) SPOT image of the displaced Eocene- Oligocene outcrops (c) Shaded relief map based on SRTM digital topographic data showing two sigmoidal ridges with right-lateral stepping arrangement. (d) simplified model of a possible shear zone within this area...... 124

Figure 5. (a) Series of right-lateral faults which displaced the Oligocene conglomerate and join to a NW-trending dextral fault. (b) SPOT image of both NNW-trending dextral and NW-trending dextral- reverse fault which displaced a serie of streams incised in the Neogen to Quaternary deposits. (c) Sketch of the same dextral fault...... 125

Figure 6. Geological map of the region presented in Figure 4 (simplified after Alai Naini, et al., 1982). The outcrops of Precambrian metamorphosed rock indicate a right-handed arrangement, indicating a dextral displacement. Fault traces are exaggerated (thick trace) to better illustrate right-stepping. The abbreviations in Legend are the same than in Figure 4...... 127

170

Figure 7. SPOT images and their geomorphic interpretations of inset Quaternary alluvial fans with a right-handed arrangement. This inset arrangement is due to the right-lateral shift of alluvial fans relative to their catchment basins. See Figure 8 for the locations...... 128

Figure 8. (a) and (b) shaded relief maps and their geomorphic interpretation which show the catchment basins and axial drainage that are offset right laterally with respect to the deposited alluvial fans. ... 128

Figure 9. SPOT image indicates a NW-trending fault within the EFZ which thrust the Mio-Pliocene conglomerate onto the alluvial fans. The evidences of right-lateral component of faulting are the deflected gullies and hills and the alluvial fans and their axial drainages which offset right laterally with respect to their catchments...... 129

Figure 10. SPOT image of a Q3 fan surface within the EFZ. The axial drainage shows a right-lateral deflection along a NW-trending fault. See Figure 3 for location...... 130

Figure 11. SPOT image of a Q2 alluvial fan offset right-laterally from its feeding drainage and catchment along a NNW-trending fault north of the EFZ. Location (60.21°E-35.10°N) is indicated in Figure 3...... 130

Figure 12. (a) Simplified active fault map of NE Iran (after Shabanian et al., 2009a; 2009b and this study) together with the main tectonic domains of NE Iran. (b) Schematic kinematic model showing crustal boundary of the Arabia-Eurasia convergence. The interaction between discontinuous dextral faults separated by compressional to transpressional relay zones allows the Central Iran to be translated northward relative to Eurasia. The white arrows show the overall motion of Central Iran with respect to Eurasia. The yellow arrows indicate the NE direction of the σ1 in NE Iran. The abbreviations are: MKDF, Main Kopeh Dagh Fault; BQ, Bakharden-Quchan Fault; CF, Chakaneh Fault; NF, Neyshabur Fault; KF, Kashafrud Fault , GKFS; Great Kavir Fault System; DFS, Doruneh Fault System; EFZ, Eastern Fault Zone; KHF, Khaf Fault; NHF, Neh Fault...... 132

DISCUSSION AND CONCLUSION

Figure 1. Overall slip rate distribution along the central and western part of the DFS (CFZ and WFZ), estimated for a time period of ~12 (Q1), ~36 (Q2) and ~100 (Q3) ka...... 142

Figure 2. (a) Simplified active fault map of the NE Iran (after Shabanian et al., 2009a; 2009b and this study) together with the main tectonic domains of NE Iran. (b) Schematic kinematic model showing crustal boundary of the Arabia-Eurasia convergence. The interaction between discontinuous dextral faults separated by compressional to transpressional relay zones allow the Central Iran to be translated northward relative to Eurasia. The white arrows show the overall motion of Central Iran with respect to Eurasia. The yellow arrows indicate the NE direction of the σ1 in NE Iran. The abbreviations are: MKDF, Main Kopeh Dagh Fault; BQ, Bakharden-Quchan Fault; CF, Chakaneh Fault; NF, Neyshabur Fault; KF, Kashafrud Fault; GKFS, Great Kavir Fault System; DFS, Doruneh Fault System; EFZ, Eastern Fault Zone; KHF, Khaf Fault; NHF, Neh Fault...... 143

SYNTHESE

Figure 1. Distribution du taux de glissement le long des zones Ouest et Centre du DFS (WFZ et CFZ), estimé pour des périodes de temps de ~12 (Q1), ~36 (Q2) et ~100 (Q3) ka...... 154

Figure 2. (a) Carte simplifiée des failles active (d’après Shabanian et al, 2009a;. 2009b et cette étude) et des principaux domaines tectoniques de NE de l’Iran. (b) Modèle schématique montrant la limite

171

crustale de la convergence Arabie-Eurasie. L'interaction entre des failles dextres séparées par des zones de relais en compression et en transpression permet à l'Iran centrale d’être translatée vers le nord par rapport à l'Eurasie. La flèche blanche montre le mouvement global de l’Iran Central-Eurasie. La flèche jaune indique la direction nord-est de la contrainte maximale σ1 (compression) dans le nord de l’Iran. Les abréviations sont les suivantes: MKDF, Main Kopeh Dagh Fault ; BQ, Bakharden-Quchan Fault; CF, Chakaneh Fault; NF, Neyshabur Fault, GKFS, Great Kavir Fault System ; DFS, Doruneh Fault System ; EFZ, Eastern Fault Zone ; KHF, Khaf Fault; NHF, Neh Fault...... 157

172

List of tables

CHAPTER II

Table 1. Epicenters and source parameters of earthquakes affecting the nearby region of the Doruneh Fault System...... 41

Table 2. Geomorphic offset measurements along the DFS...... 53

Table 3. Result of stress tensor inversion for slip data representing late Cenozoic faulting stress regimes...... 62

CHAPTER III

Table 1. Detailed characteristics of the measurements of the cumulative left-lateral offset along the DFS. Sites locations are on Fig. 3a. The Quality field indicates the confidence level (A: very high; B: high; C: medium; D: low) of the reconstructed offset markers ...... 86

Table 2. (a) Elemental composition of the bulk rock (per cent) ...... 92

Table 3. Sample characteristics of 10Be exposure ages ...... 93

Table 4. Horizontal Slip rate along the DFS estimates over three time periods of ~12, ~36 and ~100 ...... 104

173

174