THE NORTH ANATOLIAN FAULT: Anewlook

Home , 1957 Fethiye earthquakes, Armutlu Peninsula, Delice, Kelkit River, Malazgirt, Niksar

18 Mar 2005 19:32 AR AR233-EA33-02.tex AR233-EA33-02.sgm LaTeX2e(2002/01/18) P1: IKH 10.1146/annurev.earth.32.101802.120415

Annu. Rev. Earth Planet. Sci. 2005. 33:37–112 doi: 10.1146/annurev.earth.32.101802.120415 Copyright c 2005 by Annual Reviews. All rights reserved First published online as a Review in Advance on July 16, 2004

THE NORTH ANATOLIAN FAULT: ANewLook

A.M.C. S¸engor,¨ 1,3 Okan Tuys¨ uz,¨ 1,3 Caner Imren,˙ 2 Mehmet Sakınc¸,1 Haluk Eyidogan,˘ 2 Naci Gor¨ ur,¨ 1,3 Xavier Le Pichon,4 and Claude Rangin4 1Istanbul˙ Teknik Universitesi,¨ Avrasya Yer Bilimleri Enstitus¨ u,¨ Ayazaga˘ 34469, Istanbul,˙ Turkey; email: [email protected], [email protected], [email protected], [email protected] 2Istanbul˙ Teknik Universitesi,¨ Maden Fakultesi,¨ Jeoﬁzik Bol¨ um¨ u,¨ Ayazaga˘ 34469, Istanbul,˙ Turkey; email: [email protected], [email protected] 3Istanbul˙ Teknik Universitesi,¨ Maden Fakultesi,¨ Jeoloji Bol¨ um¨ u,¨ Ayazaga˘ 34469, Istanbul,˙ Turkey 4College` de France—Chaire de Geodynamique,´ Europoleˆ de l’Arbois, Batiment Laennec,¨ hall D, etage´ 2 BP 80—13545 Aix-en-Provence, France; email: [email protected], [email protected]

KeyWords strike-slip faulting, shear zone development, earthquake faulting, the Sea of Marmara, tectonics of Turkey

Dedicated to the memory of three pioneers, Ihsan˙ Ketin, Sırrı Erinc¸ and Melih Tokay, and a recent student, Aykut Barka, who burnt himself out in pursuit of the mysteries of the North Anatolian Fault.

■ Abstract The North Anatolian Fault (NAF) is a 1200-km-long dextral strike-slip fault zone that formed by progressive strain localization in a generally westerly widening right-lateral keirogen in northern Turkey mostly along an interface juxtaposing subduction-accretion material to its south and older and stiffer continental basements to its north. The NAF formed approximately 13 to 11 Ma ago in the east and propagated westward. It reached the Sea of Marmara no earlier than 200 ka ago, although shear- by California Institute of Technology on 01/09/13. For personal use only. related deformation in a broad zone there had already commenced in the late Miocene. The fault zone has a very distinct morphological expression and is seismically active.

Annu. Rev. Earth Planet. Sci. 2005.33:37-112. Downloaded from www.annualreviews.org Since the seventeenth century, it has shown cyclical seismic behavior, with century- long cycles beginning in the east and progressing westward. For earlier times, the record is less clear but does indicate a lively seismicity. The twentieth century record has been successfully interpreted in terms of a Coulomb failure model, whereby every earthquake concentrates the shear stress at the western tips of the broken segments leading to westward migration of large earthquakes. The August 17 and November 12, 1999, events have loaded the Marmara segment of the fault, mapped since the 1999 earthquakes, and a major, M ≤ 7.6 event is expected in the next half century with an approximately 50% probability on this segment. Currently, the strain in the Sea of Marmara region is highly asymmetric, with greater strain to the south of the Northern Strand. This is conditioned by the geology, and it is believed that this is generally the case for the entire North Anatolian Fault Zone. What is now needed is a more detailed 0084-6597/05/0519-0037$20.00 37 18 Mar 2005 19:32 AR AR233-EA33-02.tex AR233-EA33-02.sgm LaTeX2e(2002/01/18) P1: IKH

38 S¸ENGOR¨ ET AL.

geological mapping base with detailed paleontology and magnetic stratigraphy in the shear-related basins and more paleomagnetic observations to establish shear-related rotations.

INTRODUCTION

The North Anatolian Fault (NAF) [Figures 1 (see color insert) and 2] is one of the largest currently active strike-slip faults in the world, forming the most prominent part of a medium-size strike-slip-dominated belt of deformation, i.e., a keirogen (Ketin 1948, S¸engör 1979a; for keirogen, see S¸engör & Natal’in 1996, p. 639, note 8), in northern Turkey. It extends from the Gulf of Saros in the northern Aegean Sea to the town of Karlıova (39◦18N, 41◦01E) in Eastern Turkey for 1200 km, paralleling roughly the southern Black Sea shores and keeping a fairly regular distance of some 100 km to the coast, connecting the Aegean taphrogen (Taymaz et al. 1991, Ozeren¨ 2002, Yilmaz et al. 2002) with the East Anatolian high plateau (S¸aro˘glu 1985, Ko¸cyi˘git et al. 2001, S¸engör et al. 2003). The dextral shear associated with the NAF continues across the northern Aegean, crosses northern and central mainland Greece as a broad shear zone (termed the Grecian Shear Zone by S¸engör 1979a), and eventually links up with the Hellenic subduction zone (Dewey & S¸engör 1979, McKenzie and Jackson 1983, Le Pichon et al. 1993). Although the NAF has been subject to numerous geological, geomorphological, and geophysical (especially seismological) investigations since its recognition as a major strike-slip fault in 1948 by Ihsan˙ Ketin (Ketin 1948; see previous reviews and syntheses by Ketin 1957, 1969, 1976; Pavoni 1961; Allen 1969, 1982; Ambraseys 1969; S¸engör 1979a; Barka 1981, 1992; S¸engör & Canıtez 1982; S¸engör et al. 1982; Kiratzi 1993; also see the following symposium: Anonymous 1973), national and international interest concerning the fault has literally exploded since the catastrophic earthquakes of August 17, 1999 [Barka 1999; Barka et al. 2000a, 2002; see especially the richly documented bilingual book by Emre et al. (2003)], and November 12, 1999 (Akyüz et al. 2000, 2002). Since then, a vast

by California Institute of Technology on 01/09/13. For personal use only. amount of geological, geophysical, and geotechnical data have been gathered in the Sea of Marmara [where the probability of rupture by a large earthquake within

Annu. Rev. Earth Planet. Sci. 2005.33:37-112. Downloaded from www.annualreviews.org the next 50 years is high (Parsons et al. 2000, King et al. 2001, Atakan et al. 2002; also see Figure 13 (see color insert) last frame] and around it (Karaca & Ural 1999, Barka et al. 2000b, Do˘gan & Kurter 2000, Yaud et al. 2000, Ansal 2001, Taymaz 2001, Aksu & Yaltırak 2002, Görür et al. 2002, Toksöz 2002, Altunel & Akyüz 2003, Anonymous 2003a; also see the Rangin et al. 2001 atlas and Görür 2002, 2003), filling a previously existing gap in our knowledge of the course and character of the NAF in its western part because of its submarine location. The amount, diver- sity, quality, and the density of data collected in a few years are unparalleled in the history of geological investigations in Turkey and do not have many counterparts in the world. This great acceleration of activity in and around the Sea of Marmara also has triggered other studies along the fault. Many old problems have been looked at 18 Mar 2005 19:32 AR AR233-EA33-02.tex AR233-EA33-02.sgm LaTeX2e(2002/01/18) P1: IKH

THE NORTH ANATOLIAN FAULT 39 gan (1991), Barka (1993), Dirik (1993), Yilmaz enel (2002), and our own observations. For the sources ¸ ol (1989), Eyido˘ glu et al. (1987, 1992), Bing¨ aro˘ ¸ by California Institute of Technology on 01/09/13. For personal use only. Annu. Rev. Earth Planet. Sci. 2005.33:37-112. Downloaded from www.annualreviews.org uz (1985), S uys¨ uz et al. (2000), Barka et al. (2000a, b), Herece & Akay (2003), S The North Anatolian Keirogen (NAK). All the faults shown in this map have formed in relation to the NAK. Not all of them Figure 2 are now active, but allheavier have been lines active represent sometime the in mostthat the active the past parts keirogen 11 is of Ma. entirely Most thecompiled conﬁned are keirogen chieﬂy to potential constituting from the earthquake the T¨ area generators. structure underlain Fault known by traces as Tethyside delineated accretionary by the complexes North (see Anatolian also Fault Figure (NAF). 4). Note The faults have been et al. (1997a), Aky¨ concerning the Tethyside accretionary complexes, see Figure 4. 18 Mar 2005 19:32 AR AR233-EA33-02.tex AR233-EA33-02.sgm LaTeX2e(2002/01/18) P1: IKH

40 S¸ENGOR¨ ET AL.

from different viewpoints using different methods and new technologies (the results of these new studies have been reported, in addition to scattered papers in international literature, some of which are cited below, in the following workshop reports: Tatar et al. 2000, Altunel et al. 2001, Anonymous 2001, 2003b, Gökten et al. 2001, Emre et al. 2002; also see the following compendium: Altunel & Akyüz 2003). Unlike previous catastrophic earthquakes in Turkey, the location of the August 17 and November 12, 1999, events in a densely populated region of the country, where much of the Turkish industry is located, has led to unprecedented public interest in the NAF (e.g., Ç orlu 1999; also see Atakan et al. 2002). The fact that the city of Istanbul˙ (population ≈15 million), an international center of trade and culture from time immemorial, is now under serious large earthquake threat in the foreseeable future (e.g., Durukal et al. 2002, Erdik et al. 2003) has added an earnest international dimension to the public interest (e.g., Deli & Pérouse 1999, Pérouse 2001). New ideas on the trigger function of big earthquakes for other large shocks in large regions widened the international scientific and public interest (e.g., Papadopoulos 2002). All of this international scientific activity and public con- cern inevitably generated a vast and multifarious literature in a very short time. The purpose of this review is to present an introduction to that literature, as well as to the recent NAF literature that immediately preceded the 1999 earthquakes (see the following compendia in addition to the literature cited below: Meri¸c 1995, Bozkurt 2001) in the form of a new tectonic synthesis of the entire NAF and associated structures. Not all aspects of the work undertaken in various earth science disciplines by numerous groups can be reviewed with equal weight in the space we have at our disposal here. Our emphasis is on the geological and seismic aspects of the fault and on its western part around the Sea of Marmara, where most of the new information has been gathered. We have also tried to cite many of the less well-circulated Turkish sources, as they contain valuable data that commonly escape international attention, leading to much waste of time and duplication of effort.

by California Institute of Technology on 01/09/13. For personal use only. DISCOVERY OF AND HISTORY OF INVESTIGATIONS ALONG THE NORTH ANATOLIAN FAULT Annu. Rev. Earth Planet. Sci. 2005.33:37-112. Downloaded from www.annualreviews.org In the middle of the nineteenth century, the presence of a roughly 100-km-wide band of seismicity paralleling the course of the NAF already had been recognized (Mallet 1862, map D). This zone of seismicity was later associated with the boundary that Kober (1914, plate 14; 1921, figure 26) drew in the first quarter of the twentieth century between his Anatolian Zwischengebirge (=median massif or betwixt mountains) and the allegedly north-vergent flank of the Alpi- des. The first piece of geological field data concerning the existence of an actual fault zone within Mallet’s northern Anatolian seismic band and corresponding to part of Kober’s boundary (which he had called a Narbe = scar) was collected by Ernst Nowack (1928), who mapped a 20-km-long mylonite zone paralleling the 18 Mar 2005 19:32 AR AR233-EA33-02.tex AR233-EA33-02.sgm LaTeX2e(2002/01/18) P1: IKH

THE NORTH ANATOLIAN FAULT 41

Ulu¸cay (now So˘ganlı Ç ayı) west of Ç erke¸s (40◦49N, 32◦54E). In 1932, Nowack pointed out that this mylonite zone extended from the tip of the Gulf of Izmit˙ to Ko¸chisar (present-day Ilgaz: 40◦55N, 33◦38E) for 250 km (Nowack 1932). He interpreted this shear zone as the boundary between Suess’ (1901) Pontic arcs and the Dinarides. In 1936, Wilhelm Salomon-Calvi interpreted Nowack’s shear zone in terms of the continental drift theory of Alfred Wegener as the eastern continuation of the Tonale Line of the Alps, which he believed formed a suture zone (his Synaphie) between the Laurasian and Gondwanian elements in the structure of Anatolia. Salomon-Calvi (1936) believed that this zone could be followed eastward into Iran. In the night of December 26/27, 1939, at 0200 hours local time, a disastrous earthquake of Ms = 7.8–7.9 (see Eyido˘gan et al. 1991, Barka 1996) along this north Anatolian zone of faulting claimed the lives of some 40,000 peo- ple in and around the city of Erzincan (39◦45N, 39◦30E) (Akyol 1940; Leuchs 1940; Pamir & Ketin 1940, 1941; Salomon-Calvi 1940a; Sieberg 1940; Tillot- son 1940). This great earthquake resulted in detailed geological investigations of the fault around Erzincan (Stchepinsky 1940, Stchepinsky et al. 1940, Paréjas et al. 1942; for later studies, see Barka 1996). The 1939 Erzincan quake was followed, in rapid succession, by a series of disastrous earthquakes in Niksar-Erbaa in 1942 (Ms = 7.1); in Lâdik in 1943 (Ms = 7.3); and in Bolu, Gerede, and Ç erke¸sin1944 (Ms = 7.3; see Figure 3), which led to further studies (for references, see Ketin 1948, Eyido˘gan et al. 1991, and Barka 1996). In 1944, Necdet Egeran and Erwin Lahn emphasized that earthquake activity between 1939 and 1944 had migrated westward in northern Turkey along the structure that Salomon- Calvi had earlier called the eastward continuation of the Tonale Line. In the 1940s, interpretations of the seismic structure were made in terms of the then prevailing tectonic theories. A common denominator of all these interpretations was that the structure was seen as an integral part of the “orogenic structure” of Anatolia, and the associated earthquakes were viewed as the last death throes of an expiring orogenic belt in which “cratogenic” faults had begun to form as part of the “cra- tonization” process (Salomon-Calvi 1940a; Paréjas et al. 1942; Blumenthal 1945; Pamir 1944a,b; Egeran 1947).

by California Institute of Technology on 01/09/13. For personal use only. A completely different interpretation was offered by Ihsan˙ Ketin in 1948 that revolutionized the understanding of the structure. He noted that during all ma-

Annu. Rev. Earth Planet. Sci. 2005.33:37-112. Downloaded from www.annualreviews.org jor earthquakes in northern Turkey since 1939, the surface break always had the character of a generally east-west-striking, right-lateral fault. The vertical component of the motion always upthrew the southern block. Ketin combined these observations with the previously known courses of various young, steep, geomor- phologically distinct and seismically active shear zones along the north Anatolian earthquake belt and declared that the seismic zone in northern Turkey was the product of a major, active, right-lateral, strike-slip fault. This was the ﬁrst docu- mentation of the existence of a large and active strike-slip fault in the world [the San Andreas was not conﬁrmed to be a strike-slip fault until 1953, by Hill & Diblee (1953), despite many earlier suggestions, including Wegener’s (1915)]. Ketin further pointed out that because interior Anatolia south of the fault was 18 Mar 2005 19:32 AR AR233-EA33-02.tex AR233-EA33-02.sgm LaTeX2e(2002/01/18) P1: IKH

42 S¸ENGOR¨ ET AL. uz et al. (2000). gan et al. (1991), Barka (1996), Barka et al. (2000a), and Aky¨ by California Institute of Technology on 01/09/13. For personal use only. Annu. Rev. Earth Planet. Sci. 2005.33:37-112. Downloaded from www.annualreviews.org glu et al. (1987, 1992), Eyido˘ aro˘ ¸ Earthquakes and related fault displacements along the North Anatolian Fault (NAF) since the December 26/27, 1939, Erzincan Figure 3 earthquake. Note the remarkable east-to-westbeen migration compiled from of S the major shocks, ﬁrst emphasized by Egeran & Lahn (1944). The ﬁgure has 18 Mar 2005 19:32 AR AR233-EA33-02.tex AR233-EA33-02.sgm LaTeX2e(2002/01/18) P1: IKH

THE NORTH ANATOLIAN FAULT 43

largely aseismic, a whole Anatolian block had to be moving westward with respect to the Black Sea along the strike-slip fault that he defined. Ketin argued that to accommodate such a movement, another, however, left-lateral, fault had to exist to the south of the Anatolian block (unless the whole of Africa was also moving). His prediction was vindicated a quarter of a century later when the East Anatolian Fault was discovered (Arpat & S¸aro˘glu 1972, Seymen & Aydin 1972). S¸engör & McKenzie (1997) noted that Ketin’s 1948 paper was one of the most important harbingers of the modern, post-plate tectonics studies of continental deformation (see also S¸engör 1996). In 1948, Ketin stopped the structure short of the Sea of Marmara, but pointed out in 1953, as a result of fieldwork with Franz Rösli on the Yenice-Gönen earthquake [epicenter 40◦01N, 27◦29E(Ketin & Rösli 1953)], that the fault continued south of the Sea of Marmara through a series of young basins, forming the depressions of Bursa, Ulubat (Apolyont; in some sources and on some road signs “Uluabat,” which is actually the correct orthography; Anonymous 1977 gives the name of the lake as Ulubat, but the name of the village as Uluabat, which creates confusion), and Manyas. In 1943, Nuriye Pinar already had suggested that the three deeps of the Sea of Marmara that are located in an east-west trending larger trough (the North Marmara Trough) in the northern part of the sea, known since the surveys of H.M.S.S. Selanikˆ (Spindler et al. 1896), had been formed by a single fault that connected the Gulf of Izmit˙ with the trace of the 1912 earthquake fault on the Gelibolu (Gallipoli) Peninsula (Mihailovi¸c 1923, 1927). Pinar’s (1943) suggestion was not followed because her evidence was equivocal and she had not indicated what kind of a fault the structure she suggested was. A year later, Pfannenstiel (1944) suggested that the northern trough of the Sea of Marmara formed by a group of adjacent rhomboidal extensional basins. Egeran (1947) pointed out that the cratogenic faults of the north Anatolian seismic zone continued into the Marmara Trough (“fossédela Marmara,” p. 58), which had, however, “on a reduced scale,” the characteristics of the “Aegean tectonic zone” (Egeran 1947, p. 65). In 1968, Ketin proposed that an east-west-striking rift probably underlay the northern trough of the Sea of

by California Institute of Technology on 01/09/13. For personal use only. Marmara, echoing Pfannenstiel (1944) and Egeran (1947). This suggestion was followed by McKenzie (1972), who depicted the Sea of Marmara as one of the east-

Annu. Rev. Earth Planet. Sci. 2005.33:37-112. Downloaded from www.annualreviews.org west-striking extensional structures characterizing western Turkey and indicated that extension across it was oblique (northeast-southwest). This was disputed by Dewey & S¸engör (1979) and S¸engör (1979a) because Ganos Da˘g (formerly Tekfur Da˘gı, now I¸sıklar Da˘gı: 40◦44N, 27◦10E; elevation 924 m above sea level) just to the north of the trace of the 1912 strike-slip earthquake fault (Ambraseys & Finkel 1987, Altınok et al. 2003) appeared to them as a shortening structure above a thrust fault that they interpreted to be thrusting a 1200 m deep to its immediate east in the Sea of Marmara, indicating dextral strike-slip within the Sea of Marmara. S¸engör et al. (1985) drew a dotted line between the Gulf of Izmit˙ and the 1912 earthquake fault, emphasizing the lack of evidence of the nature of the structure of the floor of the Sea of Marmara. The large number of subsequent attempts to delineate 18 Mar 2005 19:32 AR AR233-EA33-02.tex AR233-EA33-02.sgm LaTeX2e(2002/01/18) P1: IKH

44 S¸ENGOR¨ ET AL.

the nature of the NAF under the Sea of Marmara generally followed McKenzie’s notion that motion across it was strongly oblique and that the pattern of active faults had to reflect the outlines of the lozenge-shaped bathymetric basins underlying the deeper northern trough of the Sea of Marmara (e.g., Barka & Kadinsky-Cade 1988, Wong et al. 1995, Barka 1997, Parke et al. 1999, Aksu et al. 2000, Ambraseys 2002a). The age of the NAF has long been thought to be Neogene, since Egeran & Lahn (1944) and Ketin (1948) showed that it disrupted the orogenic structure of Turkey, the youngest members of which clearly reached into the early Miocene in the northern part of the country (S¸engör & Yilmaz 1981). In 1957, Ketin wrote, “Thus, the movement of the fault has occurred after the orogeny....Itrepresents a continuous sliding process with a waxing and waning intensity, which began during the Neogene (15–20 Ma ago) and which is still active” (Ketin 1957, p. 52). In his 1969 synthesis, Ketin emphasized that the main through-going fault was a very young structure, in many parts only of Quaternary age. In his last synthesis of the NAF, Ketin (1976) pointed out that within the “rift” trough of the fault, the oldest sedimentary rocks were medial Miocene in age and that the fault therefore had to be of that age at the oldest. Although offsets had been mapped during individual earthquakes, the cumulative offset along the NAF had long remained unknown largely because for much of its course the fault parallels the dominant strike of the older tectonic units it cuts. In 1961, Pavoni estimated its offset to be between 300 and 400 km, but this was based on an erroneous correlation between the Cretaceous-Eocene eastern Pontide volcanic cover with the dominantly Mio-Pliocene volcanics of the Galatean Massif northwest of Ankara on the basis of the then already se- riously outdated 1:800,000 geological map of Turkey (Egeran & Lahn 1942– 1946). It was Ihsan˙ Seymen, a doctoral student of Ketin, who established in 1975 that the northern Neo-Tethyan suture west of Erzincan had been offset for some 85 ± 5kmbythe NAF (Seymen 1975). This has remained until the recent publication of the Atlas of the Geology of the North Anatolian Fault by Herece & Akay (2003), the best estimate of the cumulative offset of the fault (see S¸engör

by California Institute of Technology on 01/09/13. For personal use only. 1979a; Hubert-Ferrari et al. 2002; and the discussion in Westaway & Arger 2001, appendix).

Annu. Rev. Earth Planet. Sci. 2005.33:37-112. Downloaded from www.annualreviews.org The rise of plate tectonics shed new light on Turkish geology. In one of the earliest papers considering the NAF in some detail from the viewpoint of plate tectonics, Ataman et al. (1975) argued that the Tethyan sutures in northern Turkey had been a major factor in localizing the fault, but this was disputed by S¸engör & Canıtez (1982) because the neo-Tethyan suture implied by Ataman et al. does not everywhere follow the fault. S¸engör and his coworkers have summarized and synthesized the work on the NAF undertaken in the light of plate tectonics in the 1970s and the early 1980s in a series of publications (S¸engör 1979a; S¸engör & Canıtez 1982; S¸engör et al. 1982, 1983, 1985) in which they showed that the fault was indeed medial to late Miocene in age and that its offset appeared to be some- where between 50 to 100 km. S¸engör & Canıtez (1982) favored an 80 to 100 km 18 Mar 2005 19:32 AR AR233-EA33-02.tex AR233-EA33-02.sgm LaTeX2e(2002/01/18) P1: IKH

THE NORTH ANATOLIAN FAULT 45

offset in the east that possibly decreased to some 30 km in the west. S¸engör and his coworkers also underlined the importance of the basins along the NAF and within the Anatolian Scholle (see Dewey & S¸engör 1979) for an understanding of the age and nature of the fault (e.g., S¸engör et al. 1985). The time between S¸engör’s first synthesis in 1979 and the 1999 earthquakes was marked by an increased activity of geological and, to a lesser extent, geophysical work along the NAF. Much of that work has built the foundation on which the post- 1999 work was undertaken and is discussed in the present review in conjunction with it.

THE NORTH ANATOLIAN FAULT AND THE NORTH ANATOLIAN SHEAR ZONE: ELEMENTS OF THE NORTH ANATOLIAN KEIROGEN

From Figure 2, it is seen clearly that the faults associated with the North Anatolian Keirogen (NAK) constitute an entire shear zone conﬁned to the North Anatolian Tethyside accretionary complexes of latest Paleozoic to early Tertiary age. In this review, we call this shear zone the North Anatolian Shear Zone (NASZ). The NAF is only a member, albeit the most prominent member, of this shear zone. The NASZ and the NAF are elements of the NAK. There are long offshoots that take off from the main trunk of the NAF and veer toward interior Anatolia. The easternmost of these offshoots is the Ovacik Fault (Westaway & Arger 2001; OF in Figure 2), which exploits in part the suture between the Munzur and the Malatya digitations of the Kirsehir Block (Figures 2 and 4). Another major offshoot is the Sungurlu Fault (SF in Figure 2), which leaves the main branch east of Resadiye (40◦23N, 37◦20E) and ends within the inner bend of the Delice tributary of the Kizilirmak (Figure 5, indicated by the letter D). The Sungurlu Fault is also contained entirely within the Tethyside accretionary complexes and does not enter the Kirsehir Massif. Two other splays have been proposed by Ko¸cyi˘git & Beyhan (1998) between the Ovacik and the Sungurlu

by California Institute of Technology on 01/09/13. For personal use only. Faults, although we are not certain that they are offshoots of the NAF or whether they even form coherent structures. In any event, no major seismicity is associated

Annu. Rev. Earth Planet. Sci. 2005.33:37-112. Downloaded from www.annualreviews.org with them. In general, the NASZ becomes wider from east to west. Around Erzincan and some 150 to 200 km west of it, the zone is extremely narrow, hardly wider than 10 km, although near Karliova it seems to widen again (Herece & Akay 2003, appendix 13). It becomes more than 100 km wide within the Tokat lobe of the Tethyside accretionary complexes (Figure 4), but the southern members of the shear zone here seem to have been abandoned very early in its history. Although the whole lobe area is still seismically active (Figure 5), the main seismicity is concentrated in its northern part. From the Ilgaz Lobe (IL in Figure 4) westward, the shear zone broadens continuously and reaches its maximum width in the Marmara Lobe (Figure 4). 18 Mar 2005 19:32 AR AR233-EA33-02.tex AR233-EA33-02.sgm LaTeX2e(2002/01/18) P1: IKH

46 S¸ENGOR¨ ET AL. or & eng¨ ¸ or et al. (1980, 1984), S eng¨ ¸ or et al. (2003), and our own unpublished eng¨ ¸ or et al. (1982). Istanbul is a Hercynian fragment, with eng¨ ¸ uz (1999, and references therein), S uys¨ by California Institute of Technology on 01/09/13. For personal use only. Annu. Rev. Earth Planet. Sci. 2005.33:37-112. Downloaded from www.annualreviews.org sehir has a Pan-African basement with some older fragments. It has a mid-Cretaceous arc built across it and was deformed The Tethyside accretionary complexes of northern and eastern Turkey. Compiled from S Cadomian basement and a Triassic and Cretaceous-Tertiary cover redeformed by Alpide events; Menderes has a Pan-African basement, xtension; Eastern Pontides have a Pan-African basement with a strong Cimmeride imprint on which a Cretaceous and Eocene to early Figure 4 Natal’in (1996), Yılmaz et al. (1997b), Okay & T¨ observations. AL is thethe Ankara Tethyside accretionary Lobe, complexes IL in northern issubduction-accretion Turkey the complexes have sectors Ilgaz in of Lobe, places CimmerideAlpide and (i.e., separated deformations Paleo-Tethyan) the and that by EAAC Alpide elided remnants is (i.e.,Malatya the Neo-Tethyan) of digitations the Cimmerian belong the East to Continent Anatolian Cimmerian the froma Menderes-Taurus Accretionary Continent Block between Complex. as and the ﬁrst All in Cimmeride deﬁned partssome and by others Cimmeride of S Alpide magmatism brought to accretionary its into prisms. north,e contact and Munzur a by and strong Cretaceous toOligocene Eocene arc south-vergent imbrication, was followed built byQuaternary. Oligocene and Kir¸ to redeformed present during thein Eocene the late to Cretaceous Miocene to Miocene convergence. times. Some deformation to its north persisted into the 18 Mar 2005 19:32 AR AR233-EA33-02.tex AR233-EA33-02.sgm LaTeX2e(2002/01/18) P1: IKH

THE NORTH ANATOLIAN FAULT 47 that occurred between 1900 and 2001 in northern Turkey. Compiled from 4 ≥ by California Institute of Technology on 01/09/13. For personal use only. Annu. Rev. Earth Planet. Sci. 2005.33:37-112. Downloaded from www.annualreviews.org Distribution of epicenters of earthquakes with M the comprehensive ISC catalogue (2003). Figure 5 18 Mar 2005 19:32 AR AR233-EA33-02.tex AR233-EA33-02.sgm LaTeX2e(2002/01/18) P1: IKH

48 S¸ENGOR¨ ET AL.

In general, it seems that the NASZ becomes wider from east to west in harmony with the widening of the zone of accretionary complexes, notwithstanding the local widening seen in the Tokat Lobe (Figures 2, 5, and 6). Another way to judge the width of the NASZ is to look at large-scale morphology, and especially at the courses of the major Anatolian rivers. Figure 6 displays these and shows their relationships to the NASZ. In the east, the Elmalı (E in Figure 6) and the Karasu (Ka), both tributaries of the Fırat (Euphrates), are deflected right-laterally along a very narrow corridor. Farther west, the Ye¸silırmak (Y, the classical Iris) displays a broader zone of dextral deflection, and the Kızilırmak (K, the classical Halys) displays a broader one still. West of Ankara, the development of the fluvial morphology has a more complex history, but rivers lose their smooth courses once they enter the NASZ (Figure 6). Erin¸cetal. (1961a) showed that to the northeast of Ankara, around Gerede (40◦48N, 32◦12E), a Mio-Pliocene fluvial system (upper course of the Filyos, the classical Billaios; F in Figure 6) had been disrupted by the faults associated with the NASZ. Similarly, the Sakarya (S, the classical Sangarios) and its tributaries are also deflected by the NASZ, as is Susurluk (Su, the classical Macestus). The Susurluk deflection cannot be interpreted to show more than some 20+ km of dextral motion, but this is probably because (a)itisayoung river (late Pliocene; Emre et al. 1997) and (b) its further deflection is now covered by the waters of the Marmara (see Rangin et al. 2001). From the distribution of Neogene faulting and large-scale river deflection as seen in the broad curves of the Ye¸silırmak, Kızilırmak, and the Filyos and Sakarya combination, it seems possible to define a westerly widening shear zone in northern Turkey as depicted in Figure 6. If the southern Tokat Lobe faults eventually prove definitely to be parts of the NASZ, the NASZ would acquire a “pinch-and-swell structure,” but one that would still generally widen westward. That dextral shear is not confined to the NAF but is distributed in a broader shear zone is confirmed by paleomagnetic observations indicating clockwise rotations of up to 270◦ within the past 5 Ma in the central parts of NASZ (Tatar et al. 1995, Piper et al. 1997). Where observations are made in the central part of the

by California Institute of Technology on 01/09/13. For personal use only. NASZ, the shear-related clockwise rotations are found in areas as far south as 25 km from the main strand of the NAF [in contrast to earlier reports of no rotation

Annu. Rev. Earth Planet. Sci. 2005.33:37-112. Downloaded from www.annualreviews.org (Platzman et al. 1994)]. However, the rotations within the NASZ are not found to be systematic, possibly because of the early activity of R shears (Piper et al. 1996; cf. Tchalenko 1970). Farther west, around the western half of the Sea of Marmara, rotations of Miocene and younger units are spread over a width of more than 100 km (Tapırdamaz & Yaltırak 1996). In this region, the rotations are also very complex, indicating a complicated strain history in keeping with the expected evolution of a broad shear zone. By contrast, in the eastern part of the NASZ, the rotations are conﬁned to a much narrower shear zone of some 15 km width (e.g., Tatar et al. 1995). Yet another element that deﬁnes the shape of the NASZ is the basins that formed in conjunction with it, which we discuss below. 18 Mar 2005 19:32 AR AR233-EA33-02.tex AR233-EA33-02.sgm LaTeX2e(2002/01/18) P1: IKH

THE NORTH ANATOLIAN FAULT 49 silırmak; K, Kızılırmak; D, Delice; F, Filyos ˙ Iznik (lake); K, Karliova; OF, Ovacik Fault; SF, Sungurlu ˙ I, ˙ Istanbul; ˙ I, ): E, Elmali/Peri (tributary of the Murat before the construction of the Keban Dam); Ka, by California Institute of Technology on 01/09/13. For personal use only. cay)/Gerede Suyu]; S, Sakarya; Su, Susurluk. Gray letters in outline show locations of certain cities black letters Fırat (Euphrates) without Murat (outside this map)]; Y, Ye¸ Annu. Rev. Earth Planet. Sci. 2005.33:37-112. Downloaded from www.annualreviews.org = Karasu + ganlı (formerly Ulu¸ c/So˘ Map showing the North Anatolian Shear Zone (NASZ) (delimited by discontinuous lines) and the courses of the major rivers olution of the NASZ would have to be revised in the Tokat Lobe, but without touching the principles. ASZ in the southern part of the Tokat Lobe are here left out of the NASZ owing to their as yet uncertain relationship to the NAK and ault. Note that significant abrupt deflections of river courses are confined to the area of the NASZ. The faults shown to be parts of the Figure 6 traversing it. From east to west ( F N to the geometry of theev major river courses. If they eventually prove to be a part of the NAK, the analysis we present in this paper of the Karasu [Elmali/Peri and tectonic features: A, Ankara; B, Bursa; b, Bolu; E, Erzincan; [Yenice/Ara¸ 18 Mar 2005 19:32 AR AR233-EA33-02.tex AR233-EA33-02.sgm LaTeX2e(2002/01/18) P1: IKH

50 S¸ENGOR¨ ET AL.

THE TECTONIC ENVIRONMENT OF THE NORTH ANATOLIAN KEIROGEN

As a comparison of Figures 2 and 4 shows, the NAK is located almost entirely within the Anatolian Peninsula (Asia Minor). Only its westernmost extremity is located in the Sea of Marmara and in the Gelibolu (Gallipoli) Peninsula (Figures 1 and 2). Anatolia and the Sea of Marmara, together with Gelibolu, constitute parts of the late Paleozoic-Cainozoic Tethyside Orogenic System and include parts of the Cimmerides and the Alpides, products, respectively, of Paleo- and Neo-Tethys ocean basin evolution between the early Jurassic and the medial Miocene (S¸engör 1984, 1989, 1990a; S¸engör & Natal’in 1996). The geology of Turkey has been reviewed recently in the following publications, which enlarge and update the synthesis presented in S¸engör & Yılmaz (1981): Farinacci et al. (1991), S¸engör & Tatar (1996), S¸engör et al. (1996), Okay & Tüysüz (1999), Bozkurt et al. (2000), and Mittwede & Bozkurt (2001). The Tethyside accretionary complexes housing the NASZ belong both to the Cimmerides and to the Alpides. We have drawn no boundary to distinguish the two groups because the Cimmerian Continent that separated Paleo-Tethys from Neo- Tethys (S¸engör 1979b) has been extremely dismembered and reduced in width in Turkey by both the Cimmeride and Alpide collisional deformations. In places it is now completely absent, where, consequently, Cimmeride and Alpide accretionary complexes directly abut each other (Tüysüz 1990, 1993; Ko¸cyi˘git, 1991; Yılmaz et al. 1997b; Okay & Tüysüz 1999). When the NAK formed in the Neogene in response to the westerly escape of an Anatolian block (S¸engör 1979a, S¸engör et al. 1985), it clearly followed a zone of preexisting weakness within the Tethyside accretionary complexes. The Tethyside accretionary complexes are now framed by the more resistant masses of the Istanbul˙ Zone (Yılmaz et al. 1997b) and the eastern Pontides (Okay &S¸ahintürk 1997, Yılmaz et al. 1997b) to the north and the Menderes (S¸engör et al. 1984, Bozkurt & Oberhänsli 2001) and Kır¸sehir Massifs (Görür et al. 1984, Seymen 1985, Fayon et al. 2001, Gautier et al. 2002) to the south (Figure 4). All of by California Institute of Technology on 01/09/13. For personal use only. these zones have Precambrian basements. The Istanbul˙ and Eastern Pontide zones have latest Proterozoic basements (Cadomian and Pan-African, respectively), re-

Annu. Rev. Earth Planet. Sci. 2005.33:37-112. Downloaded from www.annualreviews.org deformed during the medial to late Paleozoic. The Menderes and the Kır¸sehir Massifs also have latest Proterozoic basements, although they have older events recognized by zircon dating [2.0 to 1.6 Ga, with one locality associated paleogeo- graphically with the Menderes Massif even yielding Archaean zircons (Kröner & S¸engör 1990)]. There is evidence that parts of both Menderes and Kır¸sehir and their extensions (e.g., the Bitlis Massif, the easternmost part of the Menderes- Taurus block; see Görür et al. 1984, S¸engör & Natal’in 1996; also see S¸engör 1990b) intruded by granites and granodiorites were also orogenically deformed at the same time during the late Paleozoic (Helvacı & Griffin 1984, S¸engör et al. 1984). However, the tectonic context of such events south of the Tethyside sutures has not yet been worked out because of intense post-Paleozoic disruption by rifting 18 Mar 2005 19:32 AR AR233-EA33-02.tex AR233-EA33-02.sgm LaTeX2e(2002/01/18) P1: IKH

THE NORTH ANATOLIAN FAULT 51

and drifting and subsequent repeated orogenic deformation that has largely erased and dispersed the older record. Although the zone of Tethyside accretionary complexes in northern Turkey is highly irregular, with alternating wide lobes and narrower necks, the entire zone becomes wider from east to west. If we ignore the lobes, it is narrowest near Erzincan (Figures 2 and 4) and widest around the Sea of Marmara. The North and East Anatolian Faults come together at the Karlıova junction (K in Figure 2; S¸engör 1979a, S¸aro˘glu 1985, S¸engör et al. 1985), which is located entirely within the East Anatolian Accretionary Complex, a Cretaceous to Oligocene Alpide subduction- accretion prism (S¸engör & Yılmaz 1981, S¸engör et al. 2003). It thus seems that the eastern wedge-shaped ending of the Anatolian block was also preconditioned by the contrast between the resistant masses and the weaker accretionary complex material.

OUTLINING THE NORTH ANATOLIAN KEIROGEN: BASINS RELATED TO THE NORTH ANATOLIAN SHEAR ZONE

In this section we describe the main basins related to the activity of the NASZ from west to east (see Figures 7 and 8A,B). We ﬁrst describe the basins on the main strand of the fault, which includes the newly mapped northern strand of the NAF in the Sea of Marmara (Le Pichon et al. 2001, Rangin et al. 2001, Demirba˘getal. 2003). Then, the basins on its southern strand are described. In a third section, we describe the basins on some of the splay faults of the NAF. Finally, we describe the recently explored basins of the Sea of Marmara. The basin numbers below correspond to their numbers in Figures 7 and 8A,B. Basins Along the Main Strand

THE GELIBOLU˙ (GALLIPOLI) BASIN The Gelibolu (Gallipoli) Basin is a small foredeep extending parallel with the axis of the Gelibolu Peninsula south of the Anafartalar Thrust (Yaltırak 1995) and includes those Plio-Quaternary clastic sed-

by California Institute of Technology on 01/09/13. For personal use only. imentary rocks within the Conkbayiri (in the central and southern parts of the Gelibolu Peninsula) and the Fener Pebblestone (on the isthmus joining the Penin-

Annu. Rev. Earth Planet. Sci. 2005.33:37-112. Downloaded from www.annualreviews.org sula with the Thracian mainland) formations (Sakın¸cetal. 1999, Yaltırak et al. 2000). The basin is commonly, but inappropriately, considered a part of a much larger “Thrace Neogene Basin” (Sakın¸cetal. 1999, p. 38). The sedimentary fill of the Gelibolu Basin constitutes a southerly fining, cyclic flexural basin sequence in front of the southeast-vergent Anafartalar Thrust, which dies out approximately 10 km inland from the western end of the peninsula (Kopp 1964). Northeastward, the single thrust becomes a complicated zone of anastomosing, steep thrust faults (Kopp 1964, especially plate 22). The clastics, the total thickness of which changes from approximately 80 m in the northeast to 300 m in the southwest, sit dis- to unconformably on the richly fossiliferous Upper Miocene Al¸cıtepe limestones and marls. Toward the southeast, these clastic sequences rapidly lose thickness (they 18 Mar 2005 19:32 AR AR233-EA33-02.tex AR233-EA33-02.sgm LaTeX2e(2002/01/18) P1: IKH

52 S¸ENGOR¨ ET AL. ). For geological a ( B and 8 A ). For the stratigraphy of the basins, see Figures 8 a by California Institute of Technology on 01/09/13. For personal use only. ( B Annu. Rev. Earth Planet. Sci. 2005.33:37-112. Downloaded from www.annualreviews.org and 8 A Basins deﬁning the North Anatolian Keirogen (NAK). Numbers refer to those in the text and Figure 7 in Figures 8 histories, see the text. 18 Mar 2005 19:32 AR AR233-EA33-02.tex AR233-EA33-02.sgm LaTeX2e(2002/01/18) P1: IKH

THE NORTH ANATOLIAN FAULT 53 Simpliﬁed and schematized ) a )( B Permissible ranges of basin formation times in Ma. ) c , b by California Institute of Technology on 01/09/13. For personal use only. Simpliﬁed and schematized stratigraphies of the basins of the North Anatolian Keirogen Annu. Rev. Earth Planet. Sci. 2005.33:37-112. Downloaded from www.annualreviews.org ) A ( Figure 8 (NAK). Column numbers refer to the basin numbers in Figure 7. ( stratigraphies of the basins of the NAK. ( 18 Mar 2005 19:32 AR AR233-EA33-02.tex AR233-EA33-02.sgm LaTeX2e(2002/01/18) P1: IKH

54 S¸ENGOR¨ ET AL. ) by California Institute of Technology on 01/09/13. For personal use only. Continued Annu. Rev. Earth Planet. Sci. 2005.33:37-112. Downloaded from www.annualreviews.org ( Figure 8 18 Mar 2005 19:32 AR AR233-EA33-02.tex AR233-EA33-02.sgm LaTeX2e(2002/01/18) P1: IKH

THE NORTH ANATOLIAN FAULT 55

are 15 to 20 m thick near the Asian shores of the Dardanelles and pinch out immediately to the southeast; Görür et al. 1997). They are, from base to top, mainly alluvial fan to near shore and again fluvial deposits (Yaltırak 1996, see especially his figure 1 for a cartographic depiction of the environments of sedimentation). Near Gaziköy, where the NAF Northern Strand strikes out to the Sea of Mar- mara, structures within the Fener Pebblestone and the unconformably overlying Pleistocene Altınova terrace clastics are cut by the fault (Yaltırak 1995; Figure 2). Yaltırak et al. (2000) have shown that these terraces are a part of what Sakın¸c &Yaltırak (1997) have called the Marmara Formation. Yaltırak and colleagues’ (2000) detailed, combined isotopic-stratigraphic and tectonic study has shown that the western Sea of Marmara shores have been tectonically rising since 225 ka ago at a rate of 0.4 mm/year. It is clear that this uplift is related to the activity of the northern branch of the NAF and that here the fault is younger than 225 ka. This age is in remarkable agreement with that inferred by Le Pichon et al. (2001) on the basis of the 4 km offset of the western margin of the Central Basin of the Sea of Marmara (Le Pichon et al. 2001, 2003) and a similar offset in the Central High (Armijo et al. 2002, Le Pichon et al. 2003).

THE YALOVA BASIN The Yalova Basin is an east-west elongated basin bounded in the south by an east-west-striking, mainly strike-slip fault and in the west by a northwest-southeast-striking set of dominantly normal faults (Eisenlohr 1997, Alpar & Yaltırak 2002). Toward the east, the basin peters out along an irregular, unconformable contact mainly on low-grade metamorphic rocks. The same unconformable contact delimits the basin to the north against the Eocene (Bargu & Sakın¸c 1989/1990). The main infill of the basin (Figure 8A, column 2) consists of an 800-m-thick clastic sequence beginning with conglomerates and continuing upward into sandstones and shales. Cross-bedding indicates a littoral environment. Conglomerates recur in the section, probably indicating recurrent faulting. The age of the clastic section ranges from possible Sarmatian to Lower Pliocene (Akartuna 1968, Bargu & Sakın¸c 1989/1990). These sedimentary rocks are unconformably overlain by the ca. 100-m-thick terrace fill of sandstones belonging to the Mar- by California Institute of Technology on 01/09/13. For personal use only. mara Formation (Erin¸c 1956, S¸engör et al. 1982, Sakın¸c&Yaltırak 1997) with an age range of 260 to 40 ka, but with Tyrrhenian fossils (Erin¸c 1956; ca. 100 ka).

Annu. Rev. Earth Planet. Sci. 2005.33:37-112. Downloaded from www.annualreviews.org These Pleistocene deposits continue underwater and are controlled by northwest- southeast-striking normal faults (Alpar & Yaltırak 2002; ﬁgure 6, section AR-1). These faults seem to belong to the still-active east-northeast-striking normal fault family north of the Armutlu Peninsula (Le Pichon et al. 2001, Rangin et al. 2001).

THE GOLC¨ UK-DER¨ INCE˙ BASIN The Gölcük-Derince Basin is similar in geometry to the Yalova Basin, in that it is delimited to the south by a dominantly right-lateral strike-slip fault that turns, just west of the town of Gölcük (40◦43N, 29◦49E), into a northwest-southeast-striking normal fault (Akartuna 1968, plate I). It merges eastward with the Plio-Pleistocene basin housing Lake Sapanca (Figure 7; Akartuna 1968, Emre et al. 1998). The basin is filled (Figure 8A, column 3) 18 Mar 2005 19:32 AR AR233-EA33-02.tex AR233-EA33-02.sgm LaTeX2e(2002/01/18) P1: IKH

56 S¸ENGOR¨ ET AL.

dominantly with lacustrine coarse clastic rocks, including sandstones and conglomerates, except that in the lower parts and in the upper sections marls are present. Akartuna (1968) indicates a total thickness exceeding 500 m. To the south of the Gulf of Izmit,˙ Ardel (1959) and Akartuna (1968) reported Late Miocene (Pontian) fossils from the basin fill that passes upward into Pleistocene sedimentary rocks. North of the Gulf of Izmit,˙ near Derince (40◦45N, 29◦50E) Altınlı (1968) mapped 20- to 50-m-thick Pleistocene brakish-limnic marls and sandstones that are equivalents of the sections seen to the south of the Gulf in the Gölcük area. These rocks are overlain by 40- to 100-m-thick alluvium. Akartuna (1968) indicates gentle folding of the Plio-Quaternary rocks along east-west axes.

THE ADAPAZARI BASIN The Adapazarı Basin (Inandık˙ 1952/1953, Ardel 1965) is located between the Düzce (Emre et al. 1998, 1999; Unay¨ et al. 2001) or Haraklı (Greber 1996, 1997) Fault to the south and an irregular unconformable contact on the unmetamorphosed Paleozoic rocks to the north. Greber (1997) has shown that the basin formed mainly by normal faulting on northwest-southeast-striking trends and oblique right-lateral faults with east-northeast–west-southwest strikes. The basin is filled near its margins with some visibly 150 to 200 m, and, in drill holes farther out into the basin interior, more than 550-m-thick (Bilgin 1984), coarse clastic rocks in fluviatile facies intercalated with various tuff layers, indicating concurrent volcanicity. Greber (1997) found silicified wood, indicating a humid swampy environment within the basin and a dry environment in surrounding fault-controlled uplands. Emre et al. (1998), Unay¨ & de Bruijn (1998), and Unay¨ et al. (2001) found macro- and micromammals in the finer parts of the clastic section, indicating a late Villanian–early Biharian age (Upper Pliocene–Lower Pleistocene). Greber (1996, 1997) has shown that the incessant coarse detritic sup- ply from the south indicates repeated rejuvenation of the topography throughout the deposition of the basin fill.

THE DUZCE¨ BASIN The roughly rhomboidal Düzce Basin (Ardel 1965) is located by California Institute of Technology on 01/09/13. For personal use only. between the Ç ilimli Fault to the north and the Düzce Fault to the south (Eser Teknik Sondaj ve Ticaret 2000). The eastern and the western margins are irregu-

Annu. Rev. Earth Planet. Sci. 2005.33:37-112. Downloaded from www.annualreviews.org lar, but appear to be controlled by two fault families: a northeast-southwest-striking set with mainly right-lateral offsets and a northwest-southeast-striking set that is mainly populated by normal faults. The main east-west and northeast-southwest- striking faults have been chiefly nucleated on older thrust faults within the Pon- tide basement. By contrast, the normal faults are revolutionary structures (in the sense of S¸engör et al. 1985) in that they cut the older orogenic fabric at high angles. That is why they are shorter and less continuous than the east-west and northeast-southwest sets. These latter appear to have functioned as transfer faults between the normal faults. The Düzce Basin contains an approximately 260-m- thick coarse clastic fluviatile/lacustrine sedimentary section that spans an age range from the top Pliocene to Holocene sitting mainly on Eocene volcanogenic 18 Mar 2005 19:32 AR AR233-EA33-02.tex AR233-EA33-02.sgm LaTeX2e(2002/01/18) P1: IKH

THE NORTH ANATOLIAN FAULT 57

ﬂysch. The Pliocene to Lower Pleistocene section unconformably underlies the higher Upper Pleistocene and Holocene sediments representing Gilbert deltas shed from surrounding highlands (Emre et al. 1998, Eser Teknik Sondaj ve Ticaret 2000).

THE BOLU BASIN The Bolu Basin (Ardel 1965, Aktimur et al. 1986) is delimited to the south by a sharp topographic escarpment that represents a strand of the main branch of the NAF (Ardel 1965). Its northern boundary is subparallel with the southern boundary and is also faulted, but to a much lesser extent and less continuously. The northern boundary faults die out to the east-northeast of the basin and relay their motion to the Düzce Fault to the west. The Bolu Basin is filled with coarse- to medium-grained clastic sedimentary rocks. In the lower parts, the conglomerates are badly sorted and the clasts are angular. Streams that were most likely draining fault-generated highlands nearby deposited this part, with a visible thickness of 20 m. The total thickness of the unit is probably considerably more (Ardel 1964, figure 2 shows a thickness of approximately 100 m). Upward, the fluviatile facies gives way to a lacustrine deposit represented by white claystones. These units are unconformably capped by the deposits of the present-day drainage system. The thickness of these alluvia in some places reaches 100 m. The Bolu Basin is interpreted as a pull-apart basin between the southern escarpment fault and the Düzce Fault. Both the shape of the basin and the isobaths of the water table indicate the presence of northwest-southeast-striking buried faults delimiting the basin to the east and west (Aktimur et al. 1986).

THE Ç ERKES¸-KURS¸UNLU BASIN The Ç erke¸s-Kur¸sunlu Basin (Blumenthal 1948, pp. 44–46; Pınar 1953; Barka 1985; Over¨ 1996; Bellier et al. 1997; Over¨ et al. 1997) is an obliquely shortening basin within the NASZ, bounded to the north and south by oblique-thrust faults (Figure 9A, section 4). The basin fill is considerably thinner and younger than reported by Barka (1985, 1992), who mistook the pre- Pliocene rocks that occur in an area much larger than the basin itself as part of the basin fill. Türkecan et al. (1991) have shown that the basin fill in fact consists by California Institute of Technology on 01/09/13. For personal use only. of some 200-m-thick mainly coarse continental clastics belonging to what Barka called the Ilgaz Formation. The unit is coarser in the west and includes large clasts

Annu. Rev. Earth Planet. Sci. 2005.33:37-112. Downloaded from www.annualreviews.org of the underlying late Miocene volcanic rocks. To the east, the grain size becomes finer and passes into claystones, including sand lenses representing river channels in a flood plain. It is in these finer-grained members of the Ilgaz Formation that Türkecan et al. (1991) and Unay¨ & de Bruijn (1998) found a micromammalian fauna indicating a late Pliocene age. On the basis of the high position within the Ilgaz of such fossils and the rapid deposition of the lower members, it seems pretty clear that the entire basin fill is no older than medial Pliocene. This dating is corroborated by the work of Over¨ (1996), who constrained the age of the lacustrine Ilgaz equivalents (what he called, following Barka 1985, the Lower Pontus Formation, which we think is an inappropriate appellation because it carries a terminology created by Irrlitz 1972 for eastern basins to basins unrelated 18 Mar 2005 19:32 AR AR233-EA33-02.tex AR233-EA33-02.sgm LaTeX2e(2002/01/18) P1: IKH

58 S¸ENGOR¨ ET AL. urk ¨ Ozt¨ git (1990; figure 5); 3. cyi˘ by California Institute of Technology on 01/09/13. For personal use only. Annu. Rev. Earth Planet. Sci. 2005.33:37-112. Downloaded from www.annualreviews.org Simplified, true-scale geological cross-sections across the North Anatolian Fault (NAF) and its immediately neighboring ) A ( (1980; figure 11, section K-2),and 4. 7. Tokay Bilgin (1973; (1969; section figure 16), 20), Barka Yılmaz (1985; & figure Karacık 4); (2001; figure 5. 1). Tokay (1973; section 1); 6. Ardel (1965; figure 2); Figure 9 structures within the North Anatolianmotions Shear parallel Zone with (NASZ). the Theperpendicular circles NAF to represent deduced the arrowheads from showing NAFpresent-day GPS the (for motions. component observations. sources, The of The sections see the have arrows Figure present-day been parallel partly 15). with reinterpreted Note from the the 1. section Akkan compatibility (1964; show of section the 8); the component 2. nature Ko¸ of of motion the fault at any given section with the 18 Mar 2005 19:32 AR AR233-EA33-02.tex AR233-EA33-02.sgm LaTeX2e(2002/01/18) P1: IKH

THE NORTH ANATOLIAN FAULT 59 cınlar (1946; plate of sections, by California Institute of Technology on 01/09/13. For personal use only. Annu. Rev. Earth Planet. Sci. 2005.33:37-112. Downloaded from www.annualreviews.org The sections have been partly reinterpreted from 8. Imren et al. (2001; line M97-030); 9. Imren ) B ( Figure 9 et al. (2001; line DMS-002); 10.section Imren V). et al. (2001; line M97-006); and 11. Yal¸ 18 Mar 2005 19:32 AR AR233-EA33-02.tex AR233-EA33-02.sgm LaTeX2e(2002/01/18) P1: IKH

60 S¸ENGOR¨ ET AL.

to them) to be Pliocene. However, Over’s¨ dating is based on terrestrial Ostracod analogies between northern Turkey and the island of Rhodes in southern Aegean and is therefore not reliable. In the western part of the Ç erke¸s-Kur¸sunlu Basin, there are basaltic lava flows intercalated in the white lacustrine marly limestones (Over¨ 1996; Over¨ et al. 1997). The internal structure of the Ilgaz Basin represents a south-southeast-facing overturned syncline complicated by numerous secondary folds (Figure 9A, section 4). The folds in it are appressed into trends with acuter angles to the main strand of the NAF than those folds in the eastern parts of the fault where the shear strain is higher (Figure 10; see discussion in the penultimate section of the present review). In its structure, then, the Ç erke¸s-Kur¸sunlu Basin resembles the Gelibolu Basin and probably has a similar origin (Hubert-Ferrari 2002).

THE TOSYA BASIN The Tosya Basin is similar in character to the Ç erke¸s-Kur¸sunlu Basin and is similarly oriented. It has a very similar stratigraphy (Barka 1985, 1992; Over¨ 1996; Over¨ et al. 1997). It is also bounded by two oblique thrusts that verge basinward, although here the southern thrust is less pronounced than it is in the Ç erke¸s-Kur¸sunlu Basin. The dips near the northern margin of the basin are steeper than elsewhere, but there is no overturning. Gentle folds with flank dips of up to 35◦ have subparallel axes with the basin margins (Figure 10).

THE KARGI BASIN This is a small basin located to the north of the main strand of the NAF. Its orientation is similar to the C¸ erke¸s-Kur¸sunlu and the Tosya basins and it also houses a syncline, but in this case a northwest-facing one. The basin ﬁll consists entirely of probably Pleistocene to Quaternary conglomerates making up deformed terrace deposits (Barka 1985, 1992; Dirik 1993; Over¨ et al. 1997).

THE VEZIRK˙ OPR¨ UB¨ ASIN The Vezirköprü Basin is also located to the north of the main strand of the NAF. It is bounded to the east by northwest-southeast-striking normal faults, and it seems to be a fault wedge basin similar to those described by Crowell (1974). It has a Late Miocene to Pliocene, dominantly clastic fill that by California Institute of Technology on 01/09/13. For personal use only. was deposited in limnic environments (Irrlitz 1971, Dirik 1993). The basin fill begins with coarse conglomerates near the southern master fault family and fines

Annu. Rev. Earth Planet. Sci. 2005.33:37-112. Downloaded from www.annualreviews.org northward into sandstones, claystones, and even marls. There are also silica-rich sediments probably derived from the surrounding Middle Tertiary andesitic volcanics (Dirik 1993). Upward, the section becomes ﬁner-grained with alternating claystone, sandstone, and conglomerates. The entire Upper Miocene–Pliocene sequence is some 150-m-thick and is capped by? Lower Pleistocene terrace deposits (Dirik 1993).

THE HAVZA-LADˆ IK˙ BASIN This is essentially a double basin separated by a family of north-south-striking normal faults some 5 km west of Lˆadik (40◦55N, 35◦54E). The Havza part of the basin, which is the main part, is to the west of these normal faults and is bounded to the north-northeast by the main strand of the NAF, which 18 Mar 2005 19:32 AR AR233-EA33-02.tex AR233-EA33-02.sgm LaTeX2e(2002/01/18) P1: IKH

THE NORTH ANATOLIAN FAULT 61 okmen et al. ˙ Inan (1982), Barka (1985), G¨ urk (1980), Tutkun & ¨ Ozt¨ show these and compare them with one another. See text for discussion. The fold axial trends have by California Institute of Technology on 01/09/13. For personal use only. α Annu. Rev. Earth Planet. Sci. 2005.33:37-112. Downloaded from www.annualreviews.org Map showing the axial traces of folds along the North Anatolian Fault (NAF) related to the activity of the NAF and/or to that Figure 10 of NASZ. Most traces depictedand here their are sizes too compared small with togiven the region. be The used NAF. bundle For by of that the arrows reason, in reader we to have see employed their double orientation;(1993), arrows they Yaltırak to (1995, merely summarize 1996), show Yaltırak dominant where & the fold Alpar folds axial (2002a), are Herece trends & in Akay a (2003). been compiled from the following: Bilgin (1969), Tatar (1975, 1978), 18 Mar 2005 19:32 AR AR233-EA33-02.tex AR233-EA33-02.sgm LaTeX2e(2002/01/18) P1: IKH

62 S¸ENGOR¨ ET AL.

appears here as an oblique-separation dextral strike-slip fault with a minor thrust component. The Havza part of the basin seems to have formed by extension parallel with, and by flexural bending in front of, the thrust component of the NAF. In this respect it resembles the Ridge Basin in southern California (Crowell 1982). Havza’s main sedimentary fill consists of three packages: A lower limnic (lacustrine) section, which consists mainly of an upward-fining section of some 250 to 300 m thickness. It has a pollen assemblage giving a date of late Miocene (Irrlitz 1972, as revised by Benda & Meulenkamp 1979). This limnic package gives way to a coarser, conglomerate- and sandstone-dominated package of some 150 m visible thickness [Irrlitz (1972) estimates that it must have been at least 250 m thick originally]. The third package is mainly fluviatile, with an upward-thinning grain size. It is 150 m thick and has, in its upper, lignite-bearing part latest Villanian–early Biharian mammal fossils (Unay¨ & de Bruijn 1998). The Ladik part of the basin, to the east-southeast, is narrower and appears to be a combination of a fault wedge basin and a flexural basin. Some normal faulting has been reported by Ozt¨¨ urk (1980) along the southern margin of the basin. The normal faults are displaced by conjugate strike-slip faults of small separation, indicating shortening across the basin (Figure 9A, section 3). Here, the basin fill consists mainly of conglomerates with local silty interlayers. Both the fossil content (as revised by Benda & Meulenkamp 1979) and similarity of rock succession (Ozt¨¨ urk 1979) indicate that the basin fill here corresponds to the middle, limnic-fluviatile package of the Havza part of the basin. However, the presence of Lake Lâdik in the eastern part of the basin may indicate that here similar conditions may have lasted into later times, i.e., into the Pleistocene-Holocene.

THE TAS¸OVA-ERBAA BASIN A classic pull-apart basin, the structure of the Ta¸sova- Erbaa Basin has been studied in some detail by Barka et al. (2000c). It has a visibly 700-m-thick sedimentary ﬁll (the total thickness is probably more) consisting of three main packages (Irrlitz 1972, Tutkun & Inan˙ 1982, Barka et al. 2000c): A lower, dominantly lacustrine section is some 350 m thick and sits unconformably on Eocene tuffs and sandstones. It starts with conglomerates and sandstones with by California Institute of Technology on 01/09/13. For personal use only. some claystone and lignite intercalations containing a pollen assemblage giving a late Miocene age. This passes upward into a dominantly clayey-marly section

Annu. Rev. Earth Planet. Sci. 2005.33:37-112. Downloaded from www.annualreviews.org with local conglomerate lenses. The ﬂuviatile section begins with a sandstone- claystone intercalation with abundant fossil soil horizons. Farther up, the section becomes dominantly conglomeratic with a thickness of some 400 to 500 m. Three Quaternary terraces of the Kelkit River cap the section. Barka et al. (2000c) have mapped east-northeast-trending fold axes and perva- sive normal and thrust faults. The normal faults and extensional fractures mostly strike 150◦, whereas the thrusts have common strikes around 65◦ to 70◦.Fold axial trends are scattered as a result of adaptation to basin margins and internal rotation of blocks (Figure 10). Many of the normal faults appear to have had dips of 60◦ originally, but most have since been rotated out of their original position around both vertical and horizontal axes as a result of the activity of younger fault sets. 18 Mar 2005 19:32 AR AR233-EA33-02.tex AR233-EA33-02.sgm LaTeX2e(2002/01/18) P1: IKH

THE NORTH ANATOLIAN FAULT 63

THE NIKSAR˙ BASIN The Niksar Basin is a small pull-apart basin with a pronounced sigmoidal shape (Hempton & Dunne 1983; Tatar 1996a,b; Barka et al. 2000c) probably betraying an origin as a large tension gash along the NASZ. The sedimentary infill of the Niksar Basin is similar to that of the Ta¸sova-Erbaa Basin, with the important difference of having two separate volcanic events. One is a basaltic andesite between the dominantly lacustrine lower and the dominantly fluviatile upper section. The other is a basalt that caps the fluviatile section but underlies the Quaternary terrace deposits (Tatar 1996a). Tatar’s (1996a,b) detailed structural studies have revealed that the Pliocene extension direction in the Niksar Basin was approximately N30◦E. The axes of fairly open folds (maximum flank dip 28◦) generally trend east-northeast, but some fold axes in the extreme north of the basin seem to have been rotated into a north- northeast orientation (Figure 10), as corroborated by paleomagnetic observations (Tatar et al. 1995, Piper et al. 1996).

THE SUS¸EHRIB˙ ASIN The Su¸sehri Basin is a narrow pull-apart structure later cut by the main strand of the NAF. It actually consists of two subbasins: Su¸sehri sensu stricto and Gölova (Hempton & Dunne 1983; Toprak 1988; Ko¸cyi˘git 1989, 1990; Kazancı 1993; see also Figure 9A, section 2). It has a sedimentary fill of some 750 m that can be divided into three packages: a lower lacustrine section that is approximately 200 to 250 m thick and is of late Miocene age (Irrlitz 1972, Kazancı 1993). An approximately 150- to 180-m-thick clastic section above this interfingers with limnic marls to the southeast. A 150- to 250-m-thick, upward-coarsening fluviatile section caps the entire sedimentary succession, which probably reaches into the Middle Pleistocene. It seems that much of the Quaternary section has been removed by erosion.

THE REFAHIYE˙ BASIN COMPLEX Irrlitz (1972) distinguished four basins within what he called the “Intermontane basins of the Refahiye area” (see also Tatar 1978). Three of these contain biostratigraphically reliably datable successions. by California Institute of Technology on 01/09/13. For personal use only. These are the Karnos, Bicer, and the Refahiye sensu stricto basins. The Karnos (no. 15) includes a bipartite succession beginning with coarse

Annu. Rev. Earth Planet. Sci. 2005.33:37-112. Downloaded from www.annualreviews.org conglomerates and sandstones rapidly passing into lacustrine marls and claystones. This part has a thickness of 70 to 100 m and its pollen assemblage indicates a late Miocene age. A Hipparion find at the transition of the lacustrine to fluviatile section indicates that it occurred still in the late Miocene. The 300-m-thick Plio-Pleistocene section consists mainly of coarse fluvial conglomerates. The smaller Bicer Basin (no. 16) has a similar bipartite division, but its limnic beds give a younger late Miocene age. The upper 100-m-thick fluviatile section is barren. The Refahiye basin (no. 17) (Ko¸cyi˘git 1996) resembles the Karnos more than the Bicer basin in terms of its sedimentary fill. The time of the onset of sedimentation is also similar to the Karnos Basin. 18 Mar 2005 19:32 AR AR233-EA33-02.tex AR233-EA33-02.sgm LaTeX2e(2002/01/18) P1: IKH

64 S¸ENGOR¨ ET AL.

THE ERZINCAN˙ BASIN The Erzincan Basin is a typical pull-apart structure (S¸engör 1979a, Aydın & Nur 1982, Hempton & Dunne 1983, S¸engör et al. 1985) complicated only by the influence of the Ovacik Fault to its southeast (Barka & Gülen 1989). It is the largest fault-related basin east of the Sea of Marmara. Its full sedimentary fill is not exposed. The exposed part begins with lacustrine blue claystones followed by a 250-m-thick alternation of conglomerates, sandstones, and claystones. The section ends with a 200-m-thick conglomerate of fluviatile origin. Plio-Pleistocene mafic volcanics are intercalated in the section. Seismic studies using the aftershocks of the March 13, 1992, Erzincan earthquakes (Erdik et al. 1992, Yücemen 1992, Fuenzalida et al. 1997) have been undertaken by Gaucher (1993), who estimated a variation from 600 m to 2.1 km thickness for the basin fill. More recently, Kaypak (2002) estimated a total maximum thickness around 2 to 3kmusing the seismic velocities in the basin (Figure 9A, section 1). Hubert-Ferrari et al. (2002) also inferred that the Erzincan is a “deep-rooted” basin because it seems to impede seismic rupture. Although there are so far no biostratigraphic data, Irrlitz (1972) concluded, by comparison with surrounding basins, that the basin subsidence and sedimentation must have started in late Miocene times. Most subsequent authors agree with his estimate. Erzincan also has considerable Recent to Subrecent volcanism. Both rhyolitic and mafic lavas have been reported (Ba¸s 1975, Linneman 2002). Scott Linnemann (personal communication 2002) suggested that the rhyolites probably formed by crustal melting resulting from basaltic intrusions into the crust, a conclusion that agrees with the earlier findings of Ba¸s (1975). K-Ar dating on sanidines in rhyolites has yielded two ages: 0.273 ± 0.04 ka and 0.246 ± 0.26 ka (Linneman 2002). The claim that Erzincan was a basin similar to Karlıova (for Karlıova, see S¸engör 1979a, S¸aro˘glu 1985, S¸engör et al. 1985), i.e., that it formed at the junction of the Ovacık and the NAF (Barka & Gülen 1989, Fuenzalida et al. 1997, Westaway &Arger 2001), is not supported by the Pliocene age of the Ovacık Fault. As Hempton & Dunne (1983) observed, the shape of the Erzincan Basin also belies that interpretation; although, after the Ovacık Fault became active, it contributed to the extension in the Erzincan Basin, which probably is the reason for its great depth

by California Institute of Technology on 01/09/13. For personal use only. to basement. During the activity of the Ovacık Fault (for the controversy about whether the Ovacık is a still active fault, see the discussion in Westaway & Arger

Annu. Rev. Earth Planet. Sci. 2005.33:37-112. Downloaded from www.annualreviews.org 2001), the Erzincan basin clearly operated in a mode similar to the Karlıova Basin.

Basins of the Southern Strand

THE BAYRAMIC˙ ¸(=ETIL˙ I)˙ BASIN This small graben is located along the Etili Fault Zone in the middle of the Biga Peninsula. It has a geometry suggesting a pull- apart origin, although its northern part seems not to be faulted (Yılmaz & Karacık 2001; also see Bilgin 1969, ﬁgure 20). From the geological map by Yılmaz & Karacık (2001), it may also be interpreted as a small incompatibility basin between differently oriented faults, essentially a fault-wedge basin sensu Crowell (1974). It has an approximately 150-m-thick, dominantly clastic, upward-ﬁning sedimentary 18 Mar 2005 19:32 AR AR233-EA33-02.tex AR233-EA33-02.sgm LaTeX2e(2002/01/18) P1: IKH

THE NORTH ANATOLIAN FAULT 65

ﬁll that has basalt intercalations near its upper portions. The basalts have been dated to be around 9 to 8. 4 Ma (Ercan et al. 1995, Aldanmaz et al. 2000).

THE MANYAS AND ULUBAT BASINS The Manyas and Ulubat basins actually form one basin complex separated from the southern coastal regions of the Marmara Sea by a string of mountain ranges tilted to the south along a major, coast-parallel normal-separation oblique fault. Sedimentation in this basin complex began in the Pontian (late Miocene) with fluvial sediments and rapidly developed into a lacustrine environment. In the late Pliocene, fluvial conditions dominated. Lacustrine conditions returned in the later Pliocene and became dominant in the Pleistocene (Yaltırak & Alpar 2002b). The Manyas and the Ulubat (Apolyont) lakes are the remnants of the much larger Pleistocene lakes in the area (Görür et al. 1997). In the Gönen region, the western end of the Manyas subbasin, the total thickness of the Tertiary, which is mostly the basin fill, is approximately 400 m (Yal¸cın 1997). It is clear that the late Miocene sedimentation here was fault-controlled (see Figure 9B, section 11).

THE YENIS˙ ¸EHIR˙ BASIN Yeni¸sehir is a spindle-shaped pull-apart basin that is 35 km long and has a maximum width of approximately 15 km. The southern margin is discontinuously faulted. The northern margin is probably also faulted but the alluvium covers the faults. The late Miocene is represented by the lacustrine limestones of the Gemiciköy Formation (Altinlı 1975). They are unconformably covered by the red clastics of the Alayli Formation of probable Pliocene age. The Alayli is a fluvial unit that starts coarse-grained at the base and fines upward. The Alayli is in turn covered disconformably by Quaternary alluvium. The basin seems to have started its fault-related subsidence in the late Miocene (Gen¸c 1993).

THE PAMUKOVA BASIN The Pamukova Basin is a 30-km-long and 0.2- to 6-km- wide pull-apart basin ﬁlled with the Plio-Quaternary sedimentary rocks consisting of the ﬂuvial terrace deposits with a visible thickness of 100 m and overlying 10 to

by California Institute of Technology on 01/09/13. For personal use only. 20 m of alluvium made up of meander plain and marsh deposits (Akartuna & Atan 1981, Ko¸cyi˘git 1988). The Sakarya river enters the basin at its southwest corner just to the east of Mekece (40◦27N, 30◦03E) and exits at its northeast corner north Annu. Rev. Earth Planet. Sci. 2005.33:37-112. Downloaded from www.annualreviews.org of Geyve (40◦31N, 30◦18E). This indicates a deﬂection of the river for some 22 to 26 km.

Basins of the Splay Faults Associated with the NorthAnatolian Fault

THE MERZIFON˙ BASIN The Merzifon Basin is a pull-apart basin located on the HamamözüFault Zone that splays off the main strand of the NAF. The basin is approximately 30 km long and 10 to 15 km wide. It is filled with a Neogene- Quaternary sedimentary pile that commences with the blue lacustrine clays and marls of the Miocene passing upward into calcareous, poorly consolidated 18 Mar 2005 19:32 AR AR233-EA33-02.tex AR233-EA33-02.sgm LaTeX2e(2002/01/18) P1: IKH

66 S¸ENGOR¨ ET AL.

sandstones. These are followed unconformably by Pliocene sands, gravels, and clays that are hardly consolidated. They contain volcanogenic clasts. The total thickness of the Neogene deposits is around 1000 to 1500 m. The Quater- nary is represented by alluvial channel and ﬂood-plain ﬁlls and fans (Karaalio˘glu 1973).

THE KAZOVA BASIN The Kazova, or the Almus, Basin is a fault-wedge basin located on the Sungurlu Fault. It has an undated basin fill presumed to be of Pliocene age on the basis of its poorly consolidated aspect and horizontality of its atti- tude (Bozkurt & Ko¸cyi˘git 1996). The total basin fill is approximately 200 m, although Bozkurt & Ko¸cyi˘git (1995) have estimated a vertical displacement along the northern bounding fault (the Mercimekda˘gi-Ç amdere fault set) of 750 m. They also give a 1.3 km dextral offset, but this must be regarded as a minimum, as they have been measured on subsidiary Riedel (R) and P-shears (Tchalenko 1970).

Basins Within the Sea of Marmara These basins properly belong to the basin sequence of the Northern Strand of the NAF, but they are inaccessible to direct observation. Here, we consider only the three major basins shown in Figures 9B and 11. The existence of these three bathymetric basins had become known as a result of the H.M.S.S. Selanikˆ expedition in 1894 (Spindler et al. 1896). Later, more data were collected, especially by the Turkish Navy, but the bathymetry of the northern Marmara Trough and the three main deeps only became known in detail after the campaign of the Le Suroitˆ in the Summer of 2001 (Le Pichon et al. 2001 and Rangin et al. 2001; for the southern margin of the Marmara Trough, also see Smith et al. 1995). The multichannel seismic reﬂection data collected by the Sismik-I, belonging to the Mineral Research and Exploration Directorate General of Turkey (the M.T.A.) since 1997 under the coordination of the Scientiﬁc and Technological Research Council of Turkey (TUB¨ ITAK),˙ have been in part evaluated and published by Okay et al. (1999, 2000) by California Institute of Technology on 01/09/13. For personal use only. and Imren˙ et al. (2001). The detailed marine geophysical work following the 1999 earthquakes has largely antiquated the tectonic and stratigraphic interpretations by

Annu. Rev. Earth Planet. Sci. 2005.33:37-112. Downloaded from www.annualreviews.org Okay et al. (1999, 2000). In Figure 9B, sections 8, 9, and 10 show the structure of the Ç inarcik, Central, and the Tekirda˘g basins (for locations of these basins, see Figure 11). These are approximately 2-km-deep basins (not counting the water depth!). Their fills have not been dated because the sections have not been drilled to any great depth; however, shallow drilling down to 40 m during a recent RV Marion Dufresne cruise has at least shown their Holocene to latest Pleistocene (past 40 ka) sedimentation rates. In them, the rates are more than a meter/thousand years. It is believed that during the ice ages the Sea of Marmara turned into a lacustrine basin and then the sedimentation rates were even higher (Ça˘gatay 2003). This makes the entire section in the basins Pleistocene in age, contrary to the previous estimates of Miocene to 18 Mar 2005 19:32 AR AR233-EA33-02.tex AR233-EA33-02.sgm LaTeX2e(2002/01/18) P1: IKH

THE NORTH ANATOLIAN FAULT 67

Figure 11 Active faults of the Sea of Marmara and surrounding regions. The topography is from GTOPO 30 data of the U.S. Geological Survey. The multibeam bathymetry was taken from Le Pichon et al. (2001) and Rangin et al. (2001). All known active faults have been plotted from S¸aro˘glu et al. (1992), S¸eng¨or et al. (1999), Le Pichon et al. (2001), and our own observations.

Holocene ages (e.g., Okay et. al. 1999, Yaltırak 2002). The basins are therefore very young, probably no older than the latest Pliocene. They probably formed as the NASZ was developing here along a variety of R, R, and P shears and folds (Figure 10) and were cut by the Northern Strand of the NAF at some 200 ka ago (Le Pichon et al. 2001). There may be a slight component of current extension across at least the C¸ ınarcık basin (6.4 mm/year: Figure 9B, section 8), which is creating the ongoing, very slow extension within this basin. But active structure

by California Institute of Technology on 01/09/13. For personal use only. generation is clearly dominated by strike-slip along the Northern Strand of the NAFinplaces, leading even to oblique shortening in parts of the basin margins ˙ Annu. Rev. Earth Planet. Sci. 2005.33:37-112. Downloaded from www.annualreviews.org (e.g., Imren et al. 2001).

GEOMORPHOLOGY OF THE NORTH ANATOLIAN KEIROGEN

As a comparison of Figures 1 and 2 show, both the NAF and the NASZ have sharp morphological expressions throughout their extent, which betray the youthfulness of their morphology. In fact, if we bring Figure 5 into this comparison, we would see that most structures making up both are still seismically active. Even some of the ?/rotated R shears (cf. Tchalenko 1970) (or perhaps X-shears; 18 Mar 2005 19:32 AR AR233-EA33-02.tex AR233-EA33-02.sgm LaTeX2e(2002/01/18) P1: IKH

68 S¸ENGOR¨ ET AL.

cf. Bartlett et al. 1981) associated with the NASZ are still active and generate up to Mw = 6.0 left-lateral earthquakes (e.g., Ko¸cyi˘git et al. 2001, Taymaz & Tan 2001). Seismic activity in general is more concentrated along and very near the NAF than it is elsewhere within the NASZ. This is consistent with the sharper morphology of the NAF than that of the structures elsewhere within the NASZ.

Geomorphology of the North Anatolian Fault From the Gulf of Saros to Karlıova, the morphology of the main strand of the NAF (including its very young Northern Strand) is defined by a narrow fault valley interrupted at irregular intervals by a variety of basin types, as discussed in the previous section (Erin¸cetal. 1961a; Allen 1969, 1982; Ketin 1969, 1976; Arpat & S¸aro˘glu 1975; Seymen 1975; S¸engör 1979a; Sipahio˘glu 1984; Barka 1992; Emre et al. 1998; Le Pichon et al. 2001; Rangin et al. 2001; Hubert-Ferrari et al. 2002; Ko¸cyi˘git 2003). This narrow valley is paralleled by a morphological fabric within the valley and outside it (Figure 1). The intra-vale parallel fabric is formed mainly by a variety of (a) parallel fault scarps, commonly tilted; (b) displaced and otherwise deformed erosional and constructional stream terraces; (c) pieces of deformed erosion surfaces of the valley shoulders incorporated into the fault zone; (d) shutter ridges; (e) and a multitude of whaleback ridges resulting from the anastomosing nature of the individual fault splays that form the main strand of the fault. In places, for example, between Bolu and the Ç erke¸s basins, the whaleback ridges surround sag-ponds of different sizes. A similar situation is encountered, for example, near Kü¸cükgüzel, 12 km east of Su¸sehri (40◦10N, 38◦06E), in the eastern part of the NAF (Allen 1982). Much smaller ponds characterize the fault traces in many places (e.g., at the Ismetpa¸˙ sa Station; Aytun 1982). Where the main strand forks, east of Bolu, the Yeni¸ca˘ga fault wedge basin that houses a small lake originated (Erin¸cetal. 1961b, S¸engör 1979a). Throughout the fault zone, stream courses have been deformed (see Herece &

by California Institute of Technology on 01/09/13. For personal use only. Akay 2003), in places entirely disrupting an older stream network and replacing it with a new one (e.g., Erin¸cetal. 1961a, Erin¸c 1973, Hubert-Ferrari et al. 2002, Okay

Annu. Rev. Earth Planet. Sci. 2005.33:37-112. Downloaded from www.annualreviews.org & Okay 2002). All along the main strand of the fault, consequent streams ﬂowing into the fault valley are bent clockwise, as are the courses of older streams now disrupted by the fault that cut them (e.g., Allen 1969, 1982; Ketin 1969; Seymen 1975; Hubert-Ferrari et al. 2002). In many places, “wrong” displacements, i.e., those in an anticlockwise sense, accompany those in the “right” sense, indicating capture processes. Most of these phenomena are of late Pliocene age at the oldest. Presumed older river diversions are seen in the courses of the major rivers that traverse the course of the NAF and NASZ. The most spectacular river diversions have been plotted in Figure 6. Of these, the easternmost river, the Euphrates and its tributaries, are of Pliocene age (Erin¸c 1953). They could hardly be any older, as the future sites of the Euphrates and 18 Mar 2005 19:32 AR AR233-EA33-02.tex AR233-EA33-02.sgm LaTeX2e(2002/01/18) P1: IKH

THE NORTH ANATOLIAN FAULT 69

its main tributaries were under marine waters until the Serravallian (S¸engör et al. 1985). The offset of the Elmalı/Peri Suyu system is some 70 km. The river is not only offset by the main strand of the NAF, but also by a number of parallel faults making up the NASZ here. To its west-northwest, the Karasu is also offset for approximately 70 km across the Erzincan basin. Farther to the west, the offsets of the Ye¸silırmak are more spectacular and are distributed in a wider zone. First of all, the entire river turns around a bend of more than 180◦. One major offset of the valley is between Turhal (40◦24N, 36◦05E; t in Figure 6) and the Amasya Plain (a in Figure 6), giving an offset of approximately 30 km. The other major offset is accomplished across a number of east-west-striking strike-slip faults between the Sungurlu Fault and the main strand of the NAF, which also offsets the river. All this adds an extra 50 km to the offset across the entire NASZ, bringing the total up to 80 km. The age of the Ye¸silırmak is not well established. At least its Kelkit tributary is of late Miocene age (S¸engör et al. 1985), but the entire system is probably no older than Pliocene. The Kızılırmak is the oldest of the rivers that empty into the Black Sea, and its inner Anatolian parts are clearly of late Miocene age (Izbırak˙ 1948, Akkan 1970, S¸engör et al. 1985). Its capture by external drainage into the Black Sea is most likely younger as there are no river terraces older than the Pliocene in its lower course (Ardel 1967; O. Tüysüz, unpublished data). The entire river, plus its Delice tributary (D in Figure 6), defines a mighty, east-concave bend in Central Anatolia (Ardel 1967). There is little doubt that much of the northern part of this bend is a result of the right-handed shear across the NASZ. However, there are as yet no detailed studies to substantiate this. The Sungurlu Fault at least contributed a 30 km offset to this bending. There is another sharp bend right on the main strand of the NAF, which adds another 40 km and brings the total up to 70 km. But there are a number of other, smaller jerks in the river course corresponding to smaller east-west-striking faults that obviously augment the total offset. We would not be surprised if the total dextral displacement across the NAK recorded by the great bend of the Kızılırmak is more than 100 km.

by California Institute of Technology on 01/09/13. For personal use only. The next big river diversion is represented by the Filyos. Erin¸c (1973) was of the opinion that this river system was established during the late Miocene–early

Annu. Rev. Earth Planet. Sci. 2005.33:37-112. Downloaded from www.annualreviews.org Pliocene interval, but was badly disrupted by the activity of the NAF main strand and by faults paralleling it to its south. A long upper course (the So˘ganlı Ç ayı or Ulu¸cay; Ulusu of some authors; now Gerede Ç ayı) was established by a variety of capture and diversion processes (see also Erin¸cetal. 1961a). The Sakarya is the second largest river that flows into the Black Sea. It is, at most, of late Pliocene age (Bilgin 1984, Emre et al. 1998). Since the late Pliocene, it has accumulated an at least 26 km offset as seen by its diversion in Pamukova (Ko¸cyi˘git 1988). Farther north, the river lazily flows within the Adapazarı Plain and is hardly offset at all. The final river we consider is the Susurluk. It is also of late Pliocene age (Emre et al. 1997). It shows no sharp offsets but bends clockwise across the shear zone 18 Mar 2005 19:32 AR AR233-EA33-02.tex AR233-EA33-02.sgm LaTeX2e(2002/01/18) P1: IKH

70 S¸ENGOR¨ ET AL.

of the Southern Strand of the NAF, giving a lateral visible displacement of slightly more than 20 km.

Geomorphology of the NASZ A parallel fabric outside the NAF valley characterizes the NASZ and is of a larger scale than that within the valley. Ridges 100 km or longer separated commonly by 1 to 5 km wide valleys are typical (see Figure 1). The NASZ fabric extends from the surroundings of the Sea of Marmara to where the NAF takes an east- southeasterly bend in its central part (Suzanne et al. 1990, Dirik 1993, Hubert- Ferrari et al. 2002). From the bend to Erzincan in the east, elements of the parallel fabric turn into and gently merge with the main valley of the NAF. This is particularly characteristic of the Tokat Lobe region of the Tethyside accretionary complexes, but not conﬁned to it. The same behavior is clearly seen in the disposition of the valley of the upper course of the Kızılırmak and Karasu south of Erzincan. A comparison of a geological map (e.g., the new 1:500,000-scale geological map of Turkey; S¸enel 2002) and a topographic map (e.g., Figure 1) shows that the parallel fabric outside the valley of the NAF is controlled by the strike-lines related to paleotectonics. This was known already in the beginning of the twentieth century (e.g., Philippson 1918, pp. 150–51). Such a disposition of the strike-lines (and fold axial traces) may well be because of a paleotectonic dextral fault roughly between Karlıova and the central bend of the NAF, which formed obliquely across the accretionary complexes during or just after the collision in the medial Eocene. Alternatively, such a fault may have formed during the Paleocene right-lateral jerk of Africa with respect to Europe (cf. Dewey et al. 1989). Indeed, there is no Paleocene yet found anywhere on the Tokat Lobe of the Tethyside accretionary complexes (Figure 4). Yılmaz et al. (1993) have in fact reported the presence of steep, roughly east- west-striking faults with very considerable vertical throw bringing the northern block of the future North Anatolian Fault Zone upward with respect to the south- by California Institute of Technology on 01/09/13. For personal use only. ern block. Serpentinites were introduced into these fault zones from surrounding Mesozoic m´elange units. Yılmaz et al. (1993) were unable to establish the sense

Annu. Rev. Earth Planet. Sci. 2005.33:37-112. Downloaded from www.annualreviews.org of strike-slip, although they felt certain that they were looking at a transpressional system. However, the possibility of the existence of such a strike-slip system of possibly later late Eocene times (Yılmaz et al. 1993), or simply of Paleogene times, has major implications for the cumulative offset estimated by Seymen (1975) across the NAF. If a dextral fault subparallel with the trace of the NAF was active at any time after the formation in the Hettangian-Sinemurian interval of the south-facing Atlantic-type continental margin north of the Ankara-Erzincan suture (Görür et al. 1983), then Seymen’s offset may have happened during the late Eocene and does not necessarily show the cumulative offset of the Neogene to the present NAF. Thus, independent check of Seymen’s figure is necessary, as we discuss below. 18 Mar 2005 19:32 AR AR233-EA33-02.tex AR233-EA33-02.sgm LaTeX2e(2002/01/18) P1: IKH

THE NORTH ANATOLIAN FAULT 71

PRESENT-DAY MOTIONS ALONG THE NORTH ANATOLIAN KEIROGEN The Seismicity of the NAF and the NASZ The seismicity along the NAK has been known since antiquity. Wetherefore present it in two separate sections, summarizing ﬁrst the instrumental observations.

INSTRUMENTAL OBSERVATIONS Not only is the NAF now a seismically active structure, but the entire NAK is also seismically alive (Figure 5). Most of the earthquakes with reliable epicentral locations and fault-plane solutions that fall on the trace of the fault show dextral strike-slip (Figure 12 and Table 1). Some have extensional components, perhaps indicating Riedels acting in a transtensional mode (e.g., no. 43, in Figure 12); others show the same for possible anti-Riedels (e.g., nos. 39, 41); and still others indicate the activity of normal faults in a “tension-gash orientation” [e.g., nos. 25 (rotated tension gash?), 12]. But such solutions are much fewer than those showing pure strike-slip. Earthquakes outside the NAF proper, but within the NASZ, show strike-slip parallel or subparallel with the NAF and oblique-slip extension. That movement is compatible with the overall kinematics of the NASZ and even sinistral strike-slip on rotated R-orX-shears (e.g., the June 6, 2000, Orta earthquake; Ko¸cyi˘git et al. 2001, Taymaz & Tan 2001), showing that the formation of the NAF has not yet entirely deactivated the NASZ. A seismically relatively quiescent region is seen between latitudes 36◦E and 38◦E (Figure 5). This segment corresponds to the segment broken during the December 27, 1939, earthquake of Ms = 7.8 or 7.9 (Figure 3), and the present quiescence may be because of the great amount of strain energy released during that catastrophic shock. A similar quiescent segment seen between 26◦30 E and 27◦30E along the Northern Strand of the NAF may be an after-effect of the Ms = 7.2 1912 earthquake (Mihailovi¸c 1923, 1927; Ambraseys & Finkel 1987, Rockwell et al. 2001, Altınok et al. 2003). Regular behavior in terms of migration from east to west of major earthquakes by California Institute of Technology on 01/09/13. For personal use only. since 1939 was recognized by Egeran & Lahn (1944) and Ketin & R¨osli (1953). This behavior has been modeled by Stein et al. (1997) using the Coulomb failure

Annu. Rev. Earth Planet. Sci. 2005.33:37-112. Downloaded from www.annualreviews.org of dislocations, and it successfully predicted the location of the August 17, 1999, Kocaeli earthquake. Figure 13 (see color insert) shows how the progressive failure of the NAF during the twentieth century was brought about by heightened stress concentration at the tips of the broken segments during each great earthquake (R.S. Stein & S. Bozkurt, written communication, 2003; also visit their website at http://quake.wr.usgs.gov/research/deformation/modeling/stress trig/index.html). King et al. (2001) modeled the situation after the Kocaeli earthquake and concluded that “Whatever interpretation [is placed] on the data...one or two events as great or greater than the recent one is likely to occur within the next few decades near to the northern coast of the Marmara Sea” (King et al. 2001, p. 557). Parsons et al. (2000) and Le Pichon et al. (2003) concur. 18 Mar 2005 19:32 AR AR233-EA33-02.tex AR233-EA33-02.sgm LaTeX2e(2002/01/18) P1: IKH

72 S¸ENGOR¨ ET AL. by California Institute of Technology on 01/09/13. For personal use only. Annu. Rev. Earth Planet. Sci. 2005.33:37-112. Downloaded from www.annualreviews.org ault plane solutions of 48 earthquakes that occurred along the North Anatolian Fault (NAF) between 1939 and 2003 with F 5. Fault plane solutions are shown as lower hemisphere stereographic projections with compressional quadrants shaded. Solutions ≥ Figure 12 Ms with black compressional quadrants are thecompressional most quadrants reliable are those ones determined determined by byare visual wave-form examination those modeling of obtained techniques. seismograms. Solutions by Solutions with with plottingthe darker lighter readings gray gray solutions, from compressional see Bulletins, quadrants Table 1. but are compatible with the surface characteristics of the fault. For the sources of 18 Mar 2005 19:32 AR AR233-EA33-02.tex AR233-EA33-02.sgm LaTeX2e(2002/01/18) P1: IKH

THE NORTH ANATOLIAN FAULT 73 ) Continued ( cer (1967) cer (1967) cer (1967) cer (1967) cer (1967) ¨ ¨ ¨ ¨ ¨ U¸ U¸ U¸ U¸ U¸ eference ¨ ¨ Ocal (1966) Ocal (1966) )R ◦ 8290 Taymaz et al. (1991) Taymaz et al. (1991) 43 Jackson & McKenzie (1984) 20 McKenzie (1972) 176 132178 Kocaefe & Ataman (1976) Taymaz et al. (1991) 173 Harvard Univ. (1998) 158 Canıtez & 166 Canıtez & − − − − Rake ( − − − − − − ) ◦ Dip ( ) ◦ by California Institute of Technology on 01/09/13. For personal use only. Annu. Rev. Earth Planet. Sci. 2005.33:37-112. Downloaded from www.annualreviews.org arameters and sources for fault plane solutions depicted in Figure 12 P Date Origin time Latitude Longitude Depth 20.12.1942 14:03 40.66 36.35 7.1 0 128 71 26.12.1939 23:57 39.80 39.38 7.8 0 200 61 4 Canıtez & 20.06.194326.11.194301.02.1944 15:33 22:20 03:22 40.83 40.97 30.48 41.10 33.22 33.20 6.4 7.3 0 7.3 0 176 0 269 332 76 73 77 0 173 McKenzie (1972) 31 Hodgson & Wickens (1965) 13.08.195118.03.195307.09.1953 18:3326.05.1957 19:06 03:58 40.86 06:33 40.01 32.68 40.94 27.49 40.58 33.13 6.9 31.00 7.2 0 40 6.1 348 7.0 0 150 0 281 83 84 87 72 78 14 McKenzie (1972) 179 McKenzie (1972) ABLE 1 2 1 3 4 5 6 7 8 9 No. (day.month.year) (h:min) GMT (degree) (degree) M (km) Strike ( T 1213 07.07.195714 18.09.196315 06.10.1964 05:5816 23.08.1965 16:5817 22.07.1967 14:3118 26.07.1967 39.21 14:0819 30.07.1967 40.71 16:5620 03.03.1969 40.23 40.20 18:5321 23.02.1971 29.09 40.39 01:3222 27.03.1975 28.20 5.1 40.67 00:5923 05.10.1977 26.12 6.4 0 39.54 19:41 15 05.07.1983 30.69 6.9 40.72 05:15 14 237 40.38 5.9 40.08 05:34 304 33 30.52 7.1 39.62 23:02 100 12 27.50 6.0 51 40.45 261 30 56 27.32 5.6 41.02 275 16 40 26.12 5.7 40.33 194 70 54 33.57 5.6 4 301 10 88 27.21 6.6 Canıtez & 15 72 219 5.8 50 86 6.1 8 279 15 65 06 66 76 70 254 46 McKenzie (1972) 45 McKenzie (1972) 90 160 49 McKenzie (1972) Papadopoulos et al. (1986) 173 Harvard Univ. (1998) 1011 26.05.1957 27.05.1957 09:36 11:01 40.80 40.70 30.80 31.00 6.0 5.5 0 0 114 293 24 74 157 Canıtez & 18 Mar 2005 19:32 AR AR233-EA33-02.tex AR233-EA33-02.sgm LaTeX2e(2002/01/18) P1: IKH

74 S¸ENGOR¨ ET AL. (2001) (2001) u u ul¨ ul¨ Aktar (2001) Aktar (2001) Aktar (2001) Aktar (2001) Aktar (2001) ¨ ¨ Org¨ Org¨ u& u& u& u& u& ul¨ ul¨ ul¨ ul¨ ul¨ ¨ ¨ ¨ ¨ ¨ ¨ Reference Org¨ Ozalaybey et al. (2002) Org¨ Org¨ Org¨ Harvard Univ. (1998) Org¨ ) ◦ 4 29 Taymaz ve Tan (2001) 11 Harvard Univ. (1998) 82 − 168 USGS 110 143 161 167 Harvard Univ. (1998) 102 Harvard Univ. (1998) 180 Harvard Univ. (1998) 175 Harvard Univ. (1998) − − − − − − − − − − − )Rake( ◦ Dip ( ) ◦ by California Institute of Technology on 01/09/13. For personal use only. Annu. Rev. Earth Planet. Sci. 2005.33:37-112. Downloaded from www.annualreviews.org ) Continued ( Date Origin time Latitude Longitude Depth ABLE 1 No. (day.month.year) (h:min) GMT (degree) (degree) M (km) Strike ( 4546 06.06.200047 23.08.200048 27.01.2003 02:41 06.07.2003 13:41 07:26 22:10 40.65 40.68 32.92 39.52 30.71 40.41 39.78 6.0 26.10 5.2 8 18 6.0 5.7 9 2 253 24 151 77 46 61 77 70 177 Aktar & 180 USGS 40 13.09.1999 11:55 40.76 30.07 5.8 12 293 73 164 3839 19.08.1999 31.08.199941 15:1742 29.09.1999 08:1043 11.11.199944 12.11.1999 00:13 40.65 14.02.2000 14:41 40.74 29.09 16:59 29.99 06:56 40.71 40.78 5.0 29.30 40.768 5.0 4 30.29 40.90 8.6 32.148 5.0 92 31.75 5.5 80 8 7.2 20 18 60 5.1 70 85 307 4 268 63 260 66 54 42 179 154 Aktar & 34 13.04.1998 15:14 39.18 41.10 5.3 15 272 75 2627 24.04.198828 20.04.199029 13.03.1992 20:4930 15.03.1992 23:3031 29.01.1995 17:1832 05.12.1995 16:16 40.8833 14.08.1996 04:16 40.12 14.08.1996 28.24 18:49 39.7235 40.07 01:55 39.5336 06.04.1999 39.63 02:59 39.8237 5.3 17.08.1999 39.93 15 39.43 5.4 17.08.1999 40.64 00:08 15 40.75 6.7 40.11 00:01 356 15 40.78 5.9 35.30 03:14 299 15 5.2 35.31 213 33 39.40 5.8 71 15 40.70 5.7 61 90 38.31 211 15 40.59 5.7 85 29.99 136 15 70 30.62 116 5.4 70 197 15 7.4 70 4 5.3 9 70 14 326 Harvard 8 69 Univ. (1998) 1 160 Harvard Univ. (1998) 91 176 Harvard 192 49 Univ. (1998) Harvard Univ. (1998) Harvard Univ. (1998) 87 32 175 164 Harvard Univ. (1998) Harvard Univ. (1998) T 2425 21.10.1983 18.11.1983 20:34 01:15 40.14 39.79 39.35 39.43 5.4 15 5.4 10 217 156 90 21 180 Harvard Univ. (1998) 18 Mar 2005 19:32 AR AR233-EA33-02.tex AR233-EA33-02.sgm LaTeX2e(2002/01/18) P1: IKH

THE NORTH ANATOLIAN FAULT 75

HISTORICAL SEISMICITY Figure 14 shows the interpretation that we think is the most plausible of the historical seismicity of the NAF (and, given the uncertainties in the location determinations, of the entire NAK). Fornotime interval earlier than the twentieth century can we confidently iden- tify a migratory pattern similar to the twentieth century pattern. We cannot speak of a characteristic regular, indeed cyclical, behavior of the NAF for all times. Fol- lowing the very large 1668 event, a series of earthquakes seem to have happened in the seventeenth, eighteenth, and nineteenth centuries to the west of the surmised break of the 1668 event, so that it is tempting to see here, albeit with a longer time frame, a cycle not dissimilar to the twentieth century cycle, most likely for similar reasons. The earthquake activity between 1719 and 1912 seems to have migrated generally westward, with some exceptions; however, here there are disagreements as to whether the second of the large 1766 shocks broke the Ganos Fault (see Altınok et al. 2003). Earlier than the seventeenth century, all hope of finding any regular behavior of the NAF disappears. Every century between the eleventh and the sixteenth seems to have had one major event between Refahiye (39◦54N, 38◦46E) and Karlıova. A similar recurrence pattern (except for the twelfth century) is seen between the great central bend of the NAF and its western termination. In between, we have no record as far back as the seventh century. Is this because nothing happened, nothing was recorded, or no record survived or has not yet been unearthed in the area that we have termed, for the sake of provocation, the Paphlagonian Temporal Seismic Gap? We know that the period indicated was a particularly turbulent episode in the eventful social history of Anatolia. Beginning with the Muslim incursions into the Byzantine realm in 662 and the Seljuk advance along northern Anatolia since 1015, which was followed by the Battle of Malazgirt (Manzikert: 1071) and the Crusades, the first of which (1096–1101) reached as far north as Kastamonu (41◦22N, 33◦47E; in the Byzantine sources Kastamon) and Samsun (41◦17N, 36◦20E; the classical Amisos); the Mongol advance; and the catastrophe of the Battle of Köseda˘gin1243, Anatolia was continuously ravaged by warfare and resultant displacement and/or annihilation of populations. It was not until

by California Institute of Technology on 01/09/13. For personal use only. the Ottomans were able to impose their ﬁnal authority in 1473 (the Battle of Otlukbeli) that the lands along the NASZ were able to enjoy continuous peace [for

Annu. Rev. Earth Planet. Sci. 2005.33:37-112. Downloaded from www.annualreviews.org a quick historical orientation, see the two excellent historical atlases (Da˘gtekin 1981 and Kinder and Hilgemann 1982)]. Interestingly, the 662–1473 interval pretty much coincides with the Paphlagonian Temporal Seismic Gap. However, social turbulence cannot be the whole answer because the eastern part of the NAK was not necessarily a scene of peace and prosperity in the same time interval. If anything, it was more troubled (see, in addition, Hewsen 2001). Yet from there we have records. So, the Paphlagonian Temporal Seismic Gap remains a problem to be explained. It is particularly teasing because it was followed by what were possibly the two largest earthquakes along the NAF in recorded history (1509 and 1668). Earlier in the fourth to ﬁfth and in the ﬁrst to second centuries, a biased eye may see a pattern not dissimilar to those in the 1668 to 1912 and 1939 to 1999 18 Mar 2005 19:32 AR AR233-EA33-02.tex AR233-EA33-02.sgm LaTeX2e(2002/01/18) P1: IKH

76 S¸ENGOR¨ ET AL. by California Institute of Technology on 01/09/13. For personal use only. Annu. Rev. Earth Planet. Sci. 2005.33:37-112. Downloaded from www.annualreviews.org 18 Mar 2005 19:32 AR AR233-EA33-02.tex AR233-EA33-02.sgm LaTeX2e(2002/01/18) P1: IKH

THE NORTH ANATOLIAN FAULT 77

cycles. But given the shakiness of the record even for the former of these cycles, we hesitate to say anything about the nature of the groups of events in the ﬁrst ﬁve Christian centuries except that earthquakes undoubtedly took place along the NASZ and that they were numerous in its western part, which happened to be its most civilized sector.

Motions Along the NASZ Measured by the Global Positioning System Figure 15 shows the global positioning system (GPS) vectors of motion for Turkey and surrounding regions in a fixed Eurasian reference frame after Clarke et al. (1998), McClusky et al. (2000), and Meade et al. (2002). The Anatolian Scholle east of about 31◦E longitude is now rotating, with respect to Eurasia, around an Euler pole near the Nile Delta (Le Pichon et al. 1993, Reilinger et al. 1997, McClusky et al. 2000; see also Le Pichon et al. 2003). A very remarkable thing seen in Figure 15 is that along the eastern segment of the fault east of the central bend, the motion of the Anatolian Scholle with respect to the Black Sea mountains lying north of the NAF is pure strike-slip. This is in excellent agreement with most of the best fault plate solutions of earthquakes that occurred on the NAF (nos. 27–31 in Figure 12) and the NASZ (e.g., nos. 32, 33, and 47) but not with some ←−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−− Figure 14 Historical earthquake activity along the North Anatolian Fault (NAF) for the period 400 BC to AD 2000. Only those earthquakes have been plotted that are known to have caused widespread damage. Those that have associated surface faulting are indicated with continuous lines. Those with probable surface faulting are shown by dashed lines, and those that are suspected to have possible surface faulting are indicated with dotted lines. The earthquakes have been compiled from Ambraseys (1975), Soysal et al. (1981), Ambraseys & Finkel (1991, 1997), Guidoboni et al. (1994), Ambraseys & White (1997), Ambraseys & Jackson (1998), U.S. Geological Survey Circular 1193 (Kropschot 2000), and Ambraseys (2002). The earthquake distribution given here is strongly dependent on the availability and interpretation of historical records. Given the by California Institute of Technology on 01/09/13. For personal use only. character and the state of scholarship on these records (see, for example, the discussion in Ambraseys & Finkel 1995, pp. 13–34, and compare Ambraseys & Finkel 1995 with Annu. Rev. Earth Planet. Sci. 2005.33:37-112. Downloaded from www.annualreviews.org Ambraseys 2002b on the September 10, 1509, Sea of Marmara earthquake), they most likely do not constitute an accurate picture of the seismic activity of the NAF. They clearly indicate, however, that the fault was seismically active in the period here considered. Since the sixteenth century, they also show a general distribution in time in cycles that always commenced in the east and developed westward. A biased eye can also detect a similar pattern for the first and the second centuries and for the fifth and the sixth centuries. However, these latter cannot be rigorously supported by the data and the much emphasized characteristic behavior of the NAF may not be a permanent property. Also, the gap preceding the 1668 earthquake that broke a 600-km- long segment is remarkable, although it is difficult to be sure of its existence owing to the paucity and the character of the records. 18 Mar 2005 19:32 AR AR233-EA33-02.tex AR233-EA33-02.sgm LaTeX2e(2002/01/18) P1: IKH

78 S¸ENGOR¨ ET AL. meridian. E ◦ meridians, and the transtensional relationship west of 31 E ◦ by California Institute of Technology on 01/09/13. For personal use only. and 36 E ◦ Annu. Rev. Earth Planet. Sci. 2005.33:37-112. Downloaded from www.annualreviews.org The present horizontal motions of Earth’s land surface in and around Turkey according to GPS obser- ations. Note the parallelism of the current motion with the strike of NAF in its eastern half, the transpressional Figure 15 v relationship between 31 18 Mar 2005 19:32 AR AR233-EA33-02.tex AR233-EA33-02.sgm LaTeX2e(2002/01/18) P1: IKH

THE NORTH ANATOLIAN FAULT 79

others (e.g., no. 25). Because some of the less reliable solutions also generally agree with the GPS directions (e.g., nos 1 and 2), whereas others do not (e.g., no. 12), the suspicion is awakened that not all seismic deformation follows the large- scale motions of the landmasses bordering the NASZ. Perhaps some are responses to local incompatibilities. According to the GPS observations, immediately west of the central bend of the NAF, the present-day nature of the NAF is transpressional and remains so, although in decreasing degrees, to near Bolu. Although the geology agrees (see Figure 9A, sections 4 and 5), the fault plane solutions here show either pure strike-slip or a slight normal faulting component (but compatible with northwest-southeast shortening), except earthquake no. 44, which has a considerable northwest-southeast shortening expressed by a thrust component in the fault plane solution. East of Bolu (b in Figure 2) at Yeni¸ca˘ga, the Northern Strand of the NAF turns into what would have been a transtensional orientation. Fault-plane solutions (nos. 13, 18, 38, 39, 41) east of the Sea of Marmara could be seen as corroborating this, but nos. 40 and 42 and all the well-located and well-resolved smaller earthquakes (not shown in our Figure 12, but see, e.g., Gürbüz et al. 2000, Ozalaybey¨ 2002, 2003) along the Northern Strand are essentially pure strike-slip and compatible with the orientation of the fault at their epicentral areas. Within the Sea of Marmara, there is clear strain partitioning, as emphasized by Le Pichon et al. (2001). Whereas the Main Marmara Fault (i.e., the Northern Strand within the Sea of Marmara; Le Pichon et al. 2001) gives pure dextral strike-slip solutions everywhere, fault families to its south, especially in the Ç ınarcık Basin, indicate normal faulting (see Figure 16). In contrast to the Northern Strand, the Southern Strand shows evidence for transtension as expressed by a string of basins (Yeni¸sehir, Pamukova: nos. 21 and 22 respectively in Figure 7). Those basins are isolated from one another east of Lake Iznik˙ and tied to one another by segments of the NAF, where the Southern strand is close to a pure strike-slip orientation between the Marmara Block (see Meade et al. 2002, Le Pichon et al. 2003) and the regions south of the Southern Strand of the NAF. From there westwards, most of its major splays show transtension ◦ by California Institute of Technology on 01/09/13. For personal use only. expressed in the linked basins of Bursa-Ulubat-Manyas until the 28 E longitude is reached. From there westward, all branches of the Southern Strand bend into a

Annu. Rev. Earth Planet. Sci. 2005.33:37-112. Downloaded from www.annualreviews.org transpressive orientation, which is mostly borne out by the fault-plane solutions (see nos. 19 and 20). Meade et al. (2002) and Le Pichon et al. (2001, and especially 2003) have shown that the two strands west of Adapazarı delimit an independent “Marmara block,” as shown by Dewey & S¸engör (1979) and S¸engör (1979a). As a result, the motion along the Northern Strand of the NAF is not concentric with respect to the pole near the Nile Delta, but instead about a pole located at 36◦10N, 28◦38E. This makes the Northern Strand in the Sea of Marmara (the Main Marmara Fault of Le Pichon et al. 2001) essentially a pure strike-slip fault. In fact, as Le Pichon et al. (2003) pointed out, only the Ç ınarcık Basin has a significant extensional component across it that expresses itself in a major way in present structure generation. 18 Mar 2005 19:32 AR AR233-EA33-02.tex AR233-EA33-02.sgm LaTeX2e(2002/01/18) P1: IKH

80 S¸ENGOR¨ ET AL.

Figure 16 Present-day motions of the Sea of Marmara and surrounding regions. The GPS data are from Meade et al. (2002) and were plotted with respect to a Eurasian reference frame. The slip vectors have been computed for selected earthquakes from Table 1, and Ozalaybey¨ et al. (2002, 2003) represent the motion of the hanging wall with respect to the footwall.

Another signiﬁcant result from the latest GPS measurements is Le Pichon and colleagues’ (2003) inference that strain across the Main Marmara Fault is asymmetric and the southern wall of the fault, i.e., the northern part of the Mar- mara Block delimited by the Northern and the Southern strands of the NAF, is straining more than its northern wall. The asymmetry probably results from the local geology: The fault delimits the Intra-Pontide suture zone to the north at least by California Institute of Technology on 01/09/13. For personal use only. as far west as the 28◦30E meridian and possibly farther west (S¸eng¨or & Yılmaz 1981, Yılmaz 1990, Yılmaz et al. 1997b), juxtaposing weaker late Cretaceous to

Annu. Rev. Earth Planet. Sci. 2005.33:37-112. Downloaded from www.annualreviews.org Paleocene subduction-accretion material to the south against the stronger Istanbul˙ Zone rocks to the north (see Le Pichon et al. 2003). This has important implications for the seismicity of the Marmara, and hence of the Istanbul,˙ region: The expected earthquake is likely to rupture the entire length of the Main Marmara Fault, producing an earthquake with a magnitude of 7.6. This might trigger another event of magnitude approximately 7 along the normal faults south of the C¸ ınarcık Basin (Le Pichon et al. 2003). Further implications for the seismicity of the asymmetry have not yet been investigated in any detail. A glance at Figure 2 suggests that the same asymmetric behavior may be (and may have been) characteristic of the entire length of the NAF. For most of its length, the fault is everywhere localized right on the boundary between the Tethyside 18 Mar 2005 19:32 AR AR233-EA33-02.tex AR233-EA33-02.sgm LaTeX2e(2002/01/18) P1: IKH

THE NORTH ANATOLIAN FAULT 81

accretionary complexes and the older, stiffer basement fragments to the north (also see Kaya 1988). Indeed, much of the NASZ is within the Tethyside accretionary complexes as indicated above, but the NAF nucleated nearly everywhere along a bimaterial surface (cf. Weertman 1980, Andrews & Ben-Zion 1997, Ben-Zion & Andrews 1998, Ben Zion 2001).

THE CUMULATIVE OFFSET OF THE FAULT

The cumulative offset of the NAF has long been a contentious issue. Pavoni’s first attempt was a failure because his database, the 1:800,000 geological map of Turkey (Egeran & Lahn 1942–1946), was erroneous. For many years Seymen’s (1975) estimate of the dextral offset of the Ankara-Erzincan suture zone for some 85 ± 5km(Figure 17, markers s-s) was taken as “the” offset, although S¸engör (1979a) pointed out as early as 1979 that the total displacement along the fault was less in the west than it is in the east. Bergougnan’s (1976) mapping revealed a similar offset in the same region as Seymen’s offset. Seymen’s number was disputed by others working on different parts of the entire NAF (e.g., Barka 1981, 1992; S¸aro˘glu 1988; Ko¸cyi˘git 1988, 1989), but none of these authors addressed the problem of multiple parallel faults and other structures that also take up displacement in the NASZ. Because of that, all their estimates fall short of Seymen’s (and ours in this review; see below). However, there is a genuine problem with Seymen’s (and Bergougnan’s) estimate that only became known after it was discovered that considerable strike-slip faulting before the origin of the NASZ may have displaced the continental margin on which they had measured the offset. First, Yılmaz et al. (1993) showed that there probably was significant strike-slip faulting, most likely dextral, along the future site of the NAF between Havza (40◦58N, 35◦40E) and Niksar (40◦35N, 36◦57E). Secondly, Cretaceous mélanges have been found within the mainly pre- Liassic body of Tokat Massif along narrow zones (Bozkurt et al. 1997) that are probably of strike-slip nature. If so, these would have displaced the southern margin of the Tokat Massif, Seymen’s marker “s” of the NAF, before the NAF formed by California Institute of Technology on 01/09/13. For personal use only. (Figure 17). Surprisingly, Seymen’s figure of 85 ± 5kmstill seems in good agreement with the more recent estimates measured on more reliable markers as dis-

Annu. Rev. Earth Planet. Sci. 2005.33:37-112. Downloaded from www.annualreviews.org cussed below. Somehow, the earlier faults must have created a geometry that remained suitable for estimating the total displacement. Herece & Akay (2003, table 1) tabulated all the former estimates of total offset along the NAF (except those in Hubert-Ferrari et al. 2002), which range from 7.5 km to 300 to 400 km. Their own estimates on the basis of data displayed in Herece & Akay (2003, table 2) range from 7 to 155 km. We do not think that all of the markers they used are equally reliable. The same applies to Hubert-Ferrari and colleagues’ (2002) estimates of maximum offset. For instance, their reported offset of the “two sheared folds” between the Tosya and the Vezirköprü basins (Figure 7, nos. 8 and 9, respectively) cannot be substantiated in the field because the “matchable” features give offsets of only 18 Mar 2005 19:32 AR AR233-EA33-02.tex AR233-EA33-02.sgm LaTeX2e(2002/01/18) P1: IKH

82 S¸ENGOR¨ ET AL. by California Institute of Technology on 01/09/13. For personal use only. Annu. Rev. Earth Planet. Sci. 2005.33:37-112. Downloaded from www.annualreviews.org Reliably measured cumulative offsets along the North Anatolian Fault (NAF). Offsets of markers a-a, c-c, d-d, e-e, f-f, and Figure 17 j-j are our measurements onfrom Herece features, & some Akay of (2003). Offset whichcorroborated k-k by had is Bergougnan been from (1976). pointed Armijo et out al. earlier (2002) by and others l-l is (see from text). Le Offsets Pichon b-b, et al. g-g, (2001). h-h, s-s and is i-i Seymen’s (1975) are offset, 18 Mar 2005 19:32 AR AR233-EA33-02.tex AR233-EA33-02.sgm LaTeX2e(2002/01/18) P1: IKH

THE NORTH ANATOLIAN FAULT 83

13 ± 1km(see the 1:100,000-scale geological map in Herece & Akay, 2003, appendix 7). But Herece & Akay’s matching also is not reliable because on both sides of the NAF, similar units are folded in a similar style and it is arbitrary with which “southern fold” one wishes to correlate a “northern fold.” This shows the great dangers of mapping from space images and only cursory field checks, without producing detailed geological maps. Similarly, Hubert-Ferrari and colleagues’ (2002) estimate of 65 to 95 km offset of the Filyos (sensu lato) is unlikely owing to the complexity of river capture history there (see Erin¸cetal. 1961a). Below, we present our own evaluation of the available data on cumulative offsets, including the data of Herece & Akay (2003). The following offsets are the ones we consider the most reliable. In the following list, G indicates offset of geological markers and M indicates geomorphological markers. 1. Offset of the Elmalı-Peri¸cay System (M) (Figure 17, a-a): This tributary of the Euphrates is deflected right-laterally for some 60 km between Kümbet (38◦54N, 39◦55E) in the east and Akımlı (39◦26N, 40◦19)inthe west (Anonymous 1977, sheet 340-A, Erzurum). Hubert-Ferrari et al. (2002) have here found a similar offset (65 km) resembling the previous estimates (Barka &Gülen 1989, Gaudemer et al. 1989). Because the river system is Pliocene in age here (Erin¸c 1953), the deflection represents a minimum offset. 2. The Yedisu Offset (G) (Figure 16, b-b): Here, the Yedisu Fault, the easternmost segment of the NAF, offsets a thrust contact between an Aptian-Lower Senonian ophiolitic mélange and an Upper Senonian-Palaeogene volcanic and volcaniclastic unit consisting of agglomerates, andesites, basalts, dacites, trachytes, and conglomerates for 50 km right-laterally (Herece & Akay 2003, appendix 13). This is a minimum offset for the NASZ, for south of the NAF there are a number of fault splays that take up further displacement. Unfortu- nately, in this region such splays are entirely within Pliocene volcanics and it is as yet not possible to assess the amount of displacement they accomplish (see Herece & Akay, ibid.).

by California Institute of Technology on 01/09/13. For personal use only. 3. Offset of the Karasu River (Euphrates tributary) (M) (Figure 17, c-c): Barka &G¨ulen (1989) pointed out that the Karasu was displaced right-laterally for

Annu. Rev. Earth Planet. Sci. 2005.33:37-112. Downloaded from www.annualreviews.org 50 km across the Erzincan Basin. However, they ignored the bending of the river into the fault zone (most probably along a number of parallel faults). When that bending is taken into account, the morphological offset increases to some 70 km (Anonymous 1977, sheet 340-A, Erzurum). 4. Turhal-Amasya Plain deflection of the Ye¸silırmak (M) (Figure 17, d-d): Between the town of Turhal and the Amasya Plain the Ye¸silırmak is deflected right-laterally for some 30 km. This deflection is on strike with narrow Albian to Middle Campanian ophiolitic mélange units recognized amid Paleo-Tethyan mélange units of pre-Liassic age (Bozkurt et al. 1997), which seem to have been emplaced along young strike-slip faults, expressed in the morphology as prominent parallel ridges (Anonymous 1977, sheet 324-D, 18 Mar 2005 19:32 AR AR233-EA33-02.tex AR233-EA33-02.sgm LaTeX2e(2002/01/18) P1: IKH

84 S¸ENGOR¨ ET AL.

Samsun). We interpret these faults to be related to the Sungurlu Fault some 20 km to the north (and thus parts of the NASZ). 5. Amasya Plain-Lâdik deflection (M) (Figure 17, e-e): This major deflection of the Ye¸silırmak sums the offsets (50 km) between the Sungurlu Fault to the south and the main strand of the NAF to the north, and hence gives us a part of the offset along the NASZ (Anonymous 1977, sheet 324-D, Samsun). If the scenario outlined in Hubert-Ferrari et al. (2002) is correct, this offset may be as much as 75 km. 6. The Kargi offset of the Kızilırmak (M) (Figure 17, f-f): The Kızilırmak River is displaced for some 40 km right-laterally along the main strand of the NAF. The river here is probably early Pliocene in age, so the offset is probably close to the true cumulative displacement along the fault. The suggestion by Hubert-Ferrari (2002) that the offset is greater (approximately 80 km) and should be measured between the Soruk Stream and Hacihamza (41◦05N, 34◦27E), where the sharp bend of the Kızılırmak is located, is implausible because of the absence of terraces in the Soruk valley (Tüysüz 1985). By contrast, there are terraces of the Kızılırmak, where the river now turns north at A¸sıkbükü (41◦08N, 34◦47E). 7. Mudurnu Ç ayi offsets (G) (Figure 17, g-g, h-h, and i-i): These three sets of offset geological markers (Herece & Akay 2003, appendices 3 and 4) are among the best constrained along the entire NAF. The g marker is a late Cretaceous ophiolitic mélange thrust over the Lower to Middle Eocene sandstones. The h marker is also a steep fault zone juxtaposing Albian to Lower Senonian sandstones, shales, agglomerates, lavas and red clayey limestones, and tuffites with Lower Devonian arkosic conglomerates, violet-colored siltstones, and fossiliferous siltstones. The i marker is a Devonian tectonic slice sitting in late Cretaceous (?) ultramafics. All of these are offset for some 50 km right-laterally along the Southern Strand of the NAF. The 110◦ clockwise rotation inferred by S¸engör et al. (1985) and measured paleomagneti- cally by Sarıbudak et al. (1990) seems not to have affected this part of the

by California Institute of Technology on 01/09/13. For personal use only. Almacik “ﬂake.” 8. The Pamukova river diversion (M) (Figure 17, j-j): Ko¸cyi˘git (1988) pointed

Annu. Rev. Earth Planet. Sci. 2005.33:37-112. Downloaded from www.annualreviews.org out that in Pamukova (Figure 7, no. 22) the Sakarya River is deflected right- laterally for 22 km. We have measured a similar deflection of 26 km. 9. The dextral displacement for 4 km of the northwest-southeast-trending fold of the Central High (M) (see Figure 12 and Figure 17, k-k): Armijo et al. (2002) first reported this offset, which is corroborated by Le Pichon et al. (2003). 10. The displacement of the western margin of the Central Basin of the Sea of Marmara (M) (see Figure 12 and Figure 17, l-l): Le Pichon et al. (2001) pointed out that the eastern margin of the Central Basin in the Sea of Marmara has been displaced for some 4 km since 200 ka ago, which is the age of the Northern Strand of the NAF here. 18 Mar 2005 19:32 AR AR233-EA33-02.tex AR233-EA33-02.sgm LaTeX2e(2002/01/18) P1: IKH

THE NORTH ANATOLIAN FAULT 85

Nowhere else along the Northern Strand could a cumulative offset be measured reliably. The 75 ± 5kmreported by Armijo et al. (1999, 2000) along the Gelibolu Peninsula has been disputed (Yaltırak et al. 2000) and could not be corroborated by us in the ﬁeld (X. Le Pichon, M. Sakın¸c&A.M.C. S¸eng¨or, unpublished observations) or by others (A.I.˙ Okay, personal communication, 2003) either.

THE AGE AND EVOLUTION OF THE NORTH ANATOLIAN FAULT

One of the main themes of this review is the recognition that the NAF, i.e., the main, through-going right-lateral strike-slip fault depicted with heavier lines in Figure 2, is only a member of a much larger right-lateral shear zone in northern Turkey consisting of a large number of dextral shear-related structures. Conse- quently, the age and evolution of the NAF cannot be discussed independently of the entire shear zone. We have seen that the oldest basins related to the shear evolution are located in the easternmost part of the fault and the youngest in its northwesternmost part, although along its Southern Strand the basins are also as old as late Miocene in age. The oldest basins are medial to late Miocene in age, whereas the youngest are hardly older than the Pleistocene (Figures 8A,B). Judging from the basins directly associated with it, the NAF clearly becomes younger as we go westward [Figure 8B(b)]. This is corroborated by observations in the Sea of Marmara, where the Main Mar- mara Fault is not much older than 200 ka and the ﬁlls of the main basins are not much older than Pleistocene. Along the Southern Strand, the shear-related basins are late Miocene in age, as mentioned above, and there are no observations to constrain the ages of the individual faults making up the Southern Strand. All we have to go on is the 26-km dextral offset since a certain “Plio-Quaternary” in the Pamukova Basin (no. 22 in Figure 7) and the fact that the Susurluk farther west is not similarly offset (Figure 6). However, S¸eng¨or et al. (1985) pointed out, and Le Pichon et al. (2003) concurred, that the “fragmented” nature of the Southern Strand shows that by California Institute of Technology on 01/09/13. For personal use only. a through-going main fault has yet to materialize, at least west of Pamukova. When we consider all the basins in the NASZ, this discrepancy in age and

Annu. Rev. Earth Planet. Sci. 2005.33:37-112. Downloaded from www.annualreviews.org offset between the east and the west becomes much smaller, if not entirely nonex- istent. Basins along the Southern Strand of the fault south of the Sea of Marmara (Figure 9B, nos. 19, 20) are all medial to late Miocene in age, as mentioned above. One might think that the westernmost basins may have developed in relation to the Aegean north-south extension independently of the NASZ and only became incorporated into the shear zone as it developed [Figure 8B(b,c)]. But there are other basins of similar age outside the area of inﬂuence of the Aegean extensional regime along the NASZ. Even the C¸ erke¸s and the Tosya basins (Figure 7, nos. 7 and 8) may belong to that group. From the age and similarity of style in basins all along the NASZ, we surmise that the total offset along the shear zone also most likely remains constant. Clearly, this is only an inference bereft of direct proof. 18 Mar 2005 19:32 AR AR233-EA33-02.tex AR233-EA33-02.sgm LaTeX2e(2002/01/18) P1: IKH

86 S¸ENGOR¨ ET AL.

It looks as if the Southern Strand evolved slowly with the rest of the NASZ, but that the Northern Strand formed very fast (since a certain late Pliocene date) probably after the NAF was well deﬁned east of Adapazarı (40◦47N, 30◦25E) and in response to the inability of the south-concave Southern Strand to accommodate the increased offset and the rate of motion along the NAF (Le Pichon et al. 2003). It therefore appears that the NASZ as a whole is medial to late Miocene in age, but not the NAF. Why is that so? We believe the answer lies in the way through-going strike- slip faults develop from much wider shear zones as shear progresses (Figure 18). Figure 19 shows how a certain magnitude of displacement across a shear zone is taken up differently by shear zones of differing width. At the displacement given by California Institute of Technology on 01/09/13. For personal use only. Annu. Rev. Earth Planet. Sci. 2005.33:37-112. Downloaded from www.annualreviews.org

Figure 18 Shear evolution in a strike-slip zone and the structures generated. For discussion and references, see S¸eng¨or (1995). 18 Mar 2005 19:32 AR AR233-EA33-02.tex AR233-EA33-02.sgm LaTeX2e(2002/01/18) P1: IKH

THE NORTH ANATOLIAN FAULT 87 by California Institute of Technology on 01/09/13. For personal use only. Annu. Rev. Earth Planet. Sci. 2005.33:37-112. Downloaded from www.annualreviews.org Three squares of different size (A, B, and C) deformed by the same amount of dextral ery different. Terminology and structures are from Tchalenko (1970). Figure 19 displacement applied to them. Notev that the resultant shear strain and the structures resulting from it are 18 Mar 2005 19:32 AR AR233-EA33-02.tex AR233-EA33-02.sgm LaTeX2e(2002/01/18) P1: IKH

88 S¸ENGOR¨ ET AL.

in Figure 19, shear zone A is still at the prepeak structure stage (sensu Tchalenko 1970; see Figure 18). Individual structures looked at without combining them into a general picture show hardly an indication of the shear zone. In the same Figure, shear zone C has already reached the preresidual structure stage (Tchalenko 1970; Figure 18) with a clear through-going strike-slip fault. Shear zone B has the same displacement as the other shear zones and it has developed Riedel shears at the postpeak structure stage I (Tchalenko 1970) that have not yet coalesced into a single through-going system. At this stage it is not possible to say which of the Riedel shears in the sheared square B will join up to form a major strike-slip system (Figure 19). Some will no doubt be left outside the final through-going fault. If the displacements along these abandoned structures (including the extensional structures, such as dykes and normal faults, and shortening structures, such as folds and thrusts) are not taken into account while measuring the total displacement, one would always underestimate the cumulative offset of the entire shear zone. We have selected five stations to study representative slices of the NASZ during its evolution (Figure 20). Because the shear zone widens westward, these sections cannot be squares as depicted in Figure 19, but must be more complex quadrilaterals (Figure 21). Strain in these quadrilaterals is not homogeneous simple shear, but must be inhomogeneous. Figures 21A,B show the geometric properties of such quadrilaterals and the strain fields in them. Where the NASZ has pure strike-slip, the geometry and kinematics shown in Figure 21A apply. We assume that this represents the situation in stations C, D, and E (Figures 20 and 22). By contrast, regions west of Bolu already have some transtensional component and we have imposed the same transtension on stations A and B. We are aware, as discussed above, that the situation along the NASZ is more complicated than these assump- tions, but we feel that they are adequate to give us a first-order feeling of how the NASZ and NAF might have evolved. Figure 22 shows the evolution of the five quadrilateral slices we have selected along the NASZ during six time intervals. The parameters of this evolution are tabulated in Table 2. The following discussion shows how the six intervals were selected. We empha- size that they can be selected in any arbitrary way, so long as they are distributed

by California Institute of Technology on 01/09/13. For personal use only. in time in a representative way. Figure 23 is a “speedogram” of the NASZ. It plots cumulative offset against

Annu. Rev. Earth Planet. Sci. 2005.33:37-112. Downloaded from www.annualreviews.org time since the origin of the NASZ deduced from its associated basins (Figures 8A,B). Also plotted is the present rate of motion of the fault, which is approximately 2.5 cm/year (Figures 15 and 16). If we project the present rate linearly backward in time, we see that the present cumulative offset could have accumulated in 3.5 Ma. The NASZ (and the NAF) would have formed in the early Pliocene between Zanclean and Piacenzian times. This is geologically unlikely: We know from the Karnos Basin (Figure 8B, no. 15) that the NASZ originated some 13 to 11 Ma ago in the east (where the NASZ and NAF are almost coincident), not 3.5 Ma ago. We therefore connect the present rate of motion with the rate of motion at the time of origin, which had to be 0 cm/year. We make the connection by means of a smooth curve and explore its implications. 18 Mar 2005 19:32 AR AR233-EA33-02.tex AR233-EA33-02.sgm LaTeX2e(2002/01/18) P1: IKH

THE NORTH ANATOLIAN FAULT 89 by California Institute of Technology on 01/09/13. For personal use only. Annu. Rev. Earth Planet. Sci. 2005.33:37-112. Downloaded from www.annualreviews.org Five stations selected to study the possible shear history of the North Anatolian Fault (NAF) olution. Figure 20 according to the shearev model depicted in Figure 18. The ﬁve quadrilaterals are used to track the shear 18 Mar 2005 19:32 AR AR233-EA33-02.tex AR233-EA33-02.sgm LaTeX2e(2002/01/18) P1: IKH

90 S¸ENGOR¨ ET AL. by California Institute of Technology on 01/09/13. For personal use only.

Figure 21 The geometrical and strain properties of the ﬁve representative quadrilat- Annu. Rev. Earth Planet. Sci. 2005.33:37-112. Downloaded from www.annualreviews.org erals shown in Figure 19. (A) Case for simple shear: I. The strain quadrilateral: β is the angle between the two sides of the shear zone, γ is the displacement, is the shear angle of the right (eastern) side, and is the shear angle for the left (western) side of the representative quadrilateral. Lines ad and bc do not change their length during the deformation. All other lines of the quadrilateral do change their lengths. II. The distribution of strain zones within the quadrilateral: Zone 1, all lines lengthen at all times during the deformation; zone 2, lines ﬁrst shorten then lengthen; and zone 3, all lines shorten at all times during the deformation. Black area is the family of lines that do not change their length at all during the deformation. (B) Case for transtension: I. The strain quadrilateral [symbols are the same as in (A)I, except here is displacement]. II. The distribution of strain zones within the quadrilateral [symbols same as in (A)II]. 18 Mar 2005 19:32 AR AR233-EA33-02.tex AR233-EA33-02.sgm LaTeX2e(2002/01/18) P1: IKH

THE NORTH ANATOLIAN FAULT 91

Figure 22 Shear evolution of the North Anatolian Shear Zone (NASZ) as illustrated by the shear strain undergone by the ﬁve selected quadrilaterals, here shown to have undergone purely ductile deformation to illustrate the variation in shear strain along the NASZ.

We ﬁrst assume, for the sake of argument, that the NASZ had a uniform width of some 100 km and a length of 1200 km. This is about the same width to length ratio as that of the shear zone that Tchalenko (1970) used in his clay cake experiments, illustrating the evolution of through-going strike-slip faults (see his ﬁgure 5). In

by California Institute of Technology on 01/09/13. For personal use only. these experiments, 13% of the total displacement across the shear zone is accomplished at what Tchalenko termed the prepeak strength deformation stage

Annu. Rev. Earth Planet. Sci. 2005.33:37-112. Downloaded from www.annualreviews.org (Figure 18). If the experiment corresponds to the development in nature (although it necessarily ignores dynamic rupture in earthquakes), that stage would have been accomplished along a uniform width NASZ approximately 6 Ma ago, with a total offset of some 11 km. The rate of motion would have been 0.44 cm/year. At that stage, no through-going fault would have formed, and shear would have been taken up by a series of Riedel (R) and anti-Riedel (R) shears plus some tension gashes at approximately 135◦ (measured anticlockwise from the ordinate taken parallel with the shear zone) to the shear zone and folds (and/or thrust faults) at 45◦.At the stage called “peak structure” by Tchalenko (1970), 26% of the total offset is accomplished. At this stage, the R shears usually rotate and lock, and other features continue to evolve. There is as yet no through-going fault. Some tension gash segments may begin to link some R shears. Our hypothetical NASZ would have 18 Mar 2005 19:32 AR AR233-EA33-02.tex AR233-EA33-02.sgm LaTeX2e(2002/01/18) P1: IKH

92 S¸ENGOR¨ ET AL. Phase of shear Shear zone development % by California Institute of Technology on 01/09/13. For personal use only. Annu. Rev. Earth Planet. Sci. 2005.33:37-112. Downloaded from www.annualreviews.org Evolution of the North Anatolian Shear Zone for six times at stations A to E in Figure 20 11.05 0.4444.2 6.3 1.4 8.585 12.4 18.3 41.6 23.7 2.5 0.11 31 0.15 41.7 0.22 40.4 53 0.33 48.5 0.89 74.3 59.5 11 0.44 68.7 0.6 81.7 14.8 0.85 0.89 22.2 1.13 1.33 1.7 3.55 33.3 44.4 2.56 88.75 6.87 59.2 pP 85.3 88.8 pP 113.8 133 170.7 pP 256 355 pP 682.7 Pp1 pP Pp1 P Pp2 Pp2 Pp1 pR Pp2 R pR me Rate of Shear angle (degree) Shear strain BP Offset motion Ti (Ma) (km) (cm/year) A B C D E A B C D E A B C D E A B C D E 4.33.3 22.10 0.9 29.750.8 1.2 62.9 12.4 16.7 2 16.7 23.7 22 33 31 60.7 32.2 42 0.22 40 0.3 67.3 51.6 0.44 0.3 62.2 0.65 0.4 78.7 1.87 0.63 0.6 22.2 0.84 0.9 29.6 1.26 1.9 2.4 44.4 5 29.7 66.6 177.5 39.8 63.2 pP 59.7 84.2 pP 89.6 126 239 pP 189.5 P 505 pP pR pP P P Pp1 Pp2 Pp1 pR pR pR ABLE 2 T 16 2 3 42 5 60 pP: pre-Peak; P: Peak; Pp1: post-Peak I; Pp2: post-Peak II; pR: pre-Residual; R: Residual. 18 Mar 2005 19:32 AR AR233-EA33-02.tex AR233-EA33-02.sgm LaTeX2e(2002/01/18) P1: IKH

THE NORTH ANATOLIAN FAULT 93

Figure 23 The “speedogram” of the North Anatolian Shear Zone (NASZ). See text for discussion.

reached this stage approximately 4.2 Ma ago and would have accumulated by that time an offset of approximately 22 km. Rate of displacement would have been 0.9 cm/year. At the ﬁrst part of the “post peak structure stage,” R shears considerably lengthen and the linking tension gash segments may develop into true pull-apart basins. Approximately 35% of the total offset is already accomplished at by California Institute of Technology on 01/09/13. For personal use only. this stage. This might have corresponded with a 31 km total offset along the NASZ approximately 3.4 Ma ago, when the rate of motion might have been 1.2 cm/year.

Annu. Rev. Earth Planet. Sci. 2005.33:37-112. Downloaded from www.annualreviews.org This would have been a time of increased basin subsidence along pull-aparts when the activity of very much lengthened R shears that may have begun to crowd into the fairly narrow zone of a future through-going rupture. As yet there was probably still no through-going main strand. The second part of the postpeak structure stage is the time when ﬁnally a through-going single strand originates by the appearance of P-shears that link the already much lengthened and overlapping R shears. Just more than half the total offset is already accomplished by the time these structures establish themselves. The NASZ, given the hypothetical conditions we imposed on our discussion, might have reached this stage some 2 Ma ago, i.e., in the beginning of the Pleistocene, with a total offset of some 45 km and a rate of motion of 1.4 cm/year. Finally, the through-going fault strand becomes stabilized 18 Mar 2005 19:32 AR AR233-EA33-02.tex AR233-EA33-02.sgm LaTeX2e(2002/01/18) P1: IKH

94 S¸ENGOR¨ ET AL.

when what Tchlenko called “the preresidual stage” is reached with 74% of the total offset accomplished. In our hypothetical case, this would have happened at some 800 ka ago and the rate of motion would have reached 2 cm/year. In the residual stage, 100% of the current offset is already accomplished and the fault zone is well established. This scenario, however, clearly does not fit the NASZ except at its westernmost station (A in Figures 20, 22, and 24), where the width of the shear zone is indeed 100 km. Eastward, the shear zone becomes narrower, in harmony with the narrow- ing width of the Tethyside accretionary complexes, and, correspondingly, shear strain increases! Figure 22 shows five arbitrarily chosen stations along the NASZ, for each of which, and for five times in the past and for the present, we have computed the shear angle and the shear strain and shown in which of Tchalenko’s stages they would be found at any given time (Table 2). Figure 24 shows theoretical structural evolution for each of the quadrilaterals shown in Figures 20 and 22 for the times selected. One feature of the evolution thus displayed that is immediately obvious is the westward propagation of the main through-going fault strand. For instance, a through-going strike-slip system was already well-established at station E approximately 4 Ma ago (Figures 22 and 24, stage 2), whereas hardly a single strike-slip fault had formed at station A. Some 850 ka ago (Figures 22 and 24, stage 5), a major strike-slip fault at station E had already accumulated 63 km of cumulative offset, whereas at station A there was as yet no through-going main strand, although a number of subordinate R shears must have been active and shared a part of the total offset. Another feature that presents itself to our observation is that measuring cumulative offset at the eastern stations (D and E) would be much easier than at the central and western stations because of the clean nature of the principal zone of displacement, even if the fault zone had not extended parallel with the predominant paleotectonic strike-lines in the central and western parts. The fact that the fault zone is parallel with the main trend-lines in its central and western parts naturally greatly compounds the difficulty of measuring the cumulative offset along the

by California Institute of Technology on 01/09/13. For personal use only. NASZ and along the NAF. In any case, the theoretical scenario discussed above and depicted in Figures 22

Annu. Rev. Earth Planet. Sci. 2005.33:37-112. Downloaded from www.annualreviews.org and 24 now looks very familiar to students of the North Anatolian Fault Zone. It explains why the fault becomes younger westward and, if we take the data displayed in Figures 8B(b) and (c)atface value, suggests a propagation velocity of some 11 cm/year. It also explains why the fault zone is so much narrower in the east than it is in the west and why the shear-related structures occupy such a wider zone west of station C than to its east. Almost everyone who has worked on the NAF has obtained the impression that the cumulative offset along it becomes smaller from east to west, which is corroborated by the most recent observations, as we discussed above. The evolutionary model presented here provides a rationale for that inference: The offset increases eastward because the main through-going fault had formed earlier there than it did in locations to the west. If our model is correct, the offset along 18 Mar 2005 19:32 AR AR233-EA33-02.tex AR233-EA33-02.sgm LaTeX2e(2002/01/18) P1: IKH

THE NORTH ANATOLIAN FAULT 95 by California Institute of Technology on 01/09/13. For personal use only. Annu. Rev. Earth Planet. Sci. 2005.33:37-112. Downloaded from www.annualreviews.org Theoretical shear evolution according to the model illustrated in Figure 22 at ﬁve selected stations (A, B, C, D, E; see Figure Figure 24 20) along the Northstation, Anatolian corresponding Shear with Zone the amount (NASZ)strain of at parameters, shear six see strain Table time undergone 2. intervals according to (1–6). the Here, scheme theoretical shown structural in Figure evolution 18. is For shown the at various every shear and 18 Mar 2005 19:32 AR AR233-EA33-02.tex AR233-EA33-02.sgm LaTeX2e(2002/01/18) P1: IKH

96 S¸ENGOR¨ ET AL.

the fault will have to be found to decrease systematically and continuously from east to west. And so it seems according to the most recent observations. We are aware, however, that nature is more complex than our presentation im- plies. The shear zone pinches and swells, and in the pinched regions the NAF may have formed earlier than in the swelled regions to their east and west. This would generate a complex evolution and would mar the simple picture we present. The greatest advantage of the model presented here lies, however, in the very pre- cise predictions it makes, as shown in Table 2. All the parameters there shown can be tested by careful field-mapping. Our model also emphasizes that the NAF cannot be understood properly unless the entire NASZ is considered. The disparate estimates of the age and total offset of the fault have generally resulted from individual studies that considered a segment of the fault without taking the evolution of the entire NASZ into account. The very disparate evolutionary histories in and around the Sea of Marmara at the western end of the NAF from those in its eastern parts have been perhaps one of the most valuable lessons learned from the past four years’ studies. The one thing that our model cannot predict is the forking of the NAF west of Bolu. That bifurcation is probably the result of the existence of structures already established in the west as a result of the Aegean extension. It seems that here, the NAF in its westerly propagation exploited individual rifts that were already in place. This is supported by the observation that the Northern Strand is located in the Northern Marmara Trough and the Southern Strand has several imperfectly formed branches extending along the transtensional structures south of the Sea of Marmara. Along both the Northern Strand and the Southern Strand, the Aegean extensional structures in turn mostly followed the older paleotectonic fabric (cf. S¸engör & Yılmaz 1981, S¸engör et al. 1985). In the north, the location of the Intra- Pontide suture (S¸engör & Yılmaz 1981), exactly coinciding with the course of the Main Marmara Fault, further facilitated the latter’s nucleation and has led to asymmetric strain across a bimaterial fault.

by California Institute of Technology on 01/09/13. For personal use only. CONCLUSIONS

The NAF is a diachronous structure that formed by progressive strain localization Annu. Rev. Earth Planet. Sci. 2005.33:37-112. Downloaded from www.annualreviews.org in a westerly widening right-lateral shear zone in northern Turkey along mostly a bimaterial interface juxtaposing subduction-accretion material of the Tethysides and older and stronger continental basement to its north. It has formed since approximately 11 Ma ago in the east, near Erzincan, and may have propagated westward at a rate of 11 cm/year if the assumption of the continuously widening width of NASZ westward is correct. If not, irregularities will be seen locally in the rate and direction of propagation. NAF reached the Sea of Marmara no earlier than 200 ka ago, although the NASZ-related deformation there commenced in the late Miocene. The fault has a very distinct morphological expression and is seismically active. Since the seventeenth century, it has seemed to display a 18 Mar 2005 19:32 AR AR233-EA33-02.tex AR233-EA33-02.sgm LaTeX2e(2002/01/18) P1: IKH

THE NORTH ANATOLIAN FAULT 97

cyclical seismic behavior, with century-long cycles commencing in the east and migrating westward, but the record does not lend itself to a unique interpretation before the twentieth century. For earlier times, reaching back to the third century BC, the record is less satisfactory, although clearly indicating a lively seismicity. The twentieth century record has been successfully interpreted in terms of a Coulomb failure model, whereby every earthquake concentrates the shear stress at the western tip of the broken segments leading to westward migration of the large earthquakes. It is believed that the August 17 and November 12, 1999, events have loaded the Marmara segment of the fault and that a major, M ≤ 7.6 event is to be expected in the next half century with an approximately 50% probability on this segment. Now, the strain in the Sea of Marmara region is highly asymmetric with greater strain to the south of the Northern Strand. This is conditioned by the geology and it is believed that this is generally the case for the entire NAF. What is now needed is a careful revision, and enlargement to cover the entire NASZ, of the Geological Atlas of the NAF (Herece & Akay 2003) in the light of the discussions developed here. More intensive paleontological and magnetostrati- graphic studies in the NASZ basins to date them more precisely and the collection of more paleomagnetic data along the entire NASZ will be needed. Once such a data set becomes available, we will be in a better position to assess the history of the entire NAK.

ACKNOWLEDGMENTS We thank S¸. Can Gen¸c, Cenk Yaltırak, Omer¨ Emre, and G¨ulsen U¸carku¸s for help in pulling the data together. John F. Dewey discussed with us the tectonics of transtensional zones. We are particularly grateful to Erdal Herece and Erg¨un Akay, both of the General Directorate of Mineral Research and Exploration (MTA) in Ankara, Turkey, for having placed their great Atlas of the NAF (scale 1:100,000) at our disposal before publication. Ross Stein and Serkan Bozkurt kindly supplied Figure 13 and gave permission to include it in this paper. Korhan Ertura¸c prepared

by California Institute of Technology on 01/09/13. For personal use only. Figure 1. We thank Kevin Burke for a superb review and Jennifer Jongsma for excellent editorial work. S¸engör and Görür are grateful for the support of the

Annu. Rev. Earth Planet. Sci. 2005.33:37-112. Downloaded from www.annualreviews.org Turkish Academy of Sciences.

The Annual Review of Earth and Planetary Science is online at http://earth.annualreviews.org

LITERATURE CITED

Akartuna M. 1968. Armutlu Yarımadasının Je- Akartuna M, Atan OR. 1981. Geyve-Ikramiye-˙ olojisi. Istanbul˙ Univ.¨ Fen Fak. Monogr. Fındıksu (Sakarya) dolayının jeolojisi. (Tabiˆı Ilimler˙ Kısmı), No. 20. 105 pp.+3 fold- Selc¸uk Univ.¨ Fen Fak. Derg., No. 1, Ser. A outs (Yerbilim.), pp. 25–46 18 Mar 2005 19:32 AR AR233-EA33-02.tex AR233-EA33-02.sgm LaTeX2e(2002/01/18) P1: IKH

98 S¸ENGOR¨ ET AL.

Akkan E. 1964. Erzincan Ovası ve Çevresinin ern Anatolia, Turkey. J. Volcanol. Geotherm. Jeomorfolojisi. Ankara Univ.¨ Dil Tarih- Res. 102:67–95 Cograf.˘ Fak. Yayın., No. 153. 104 pp.+1er- Allen CR. 1969. Active faulting in northern rata page+11 foldouts Turkey. Contrib. No. 1577, Div. Geol. Sci., Akkan E. 1970. Bafra Burnu-Delice Kav¸sa˘gı Calif. Inst. Technol. 32 pp. Arasında Kızılırmak Vadisinin Jeomorfolo- Allen CR. 1982. Comparisons between the jisi. Ankara Univ.¨ Dil Tarih-Cograf.˘ Fak. North Anatolian Fault of Turkey and the San Yayın., No. 191. VI+158 pp.+17 foldout Andreas Fault of California. See I¸sıkara & plates Vogel 1982, pp. 67–85 Aksu AE, Calon TJ, Hiscott RN, Ya¸sar D. 2000. Alpar B, Yaltırak C. 2002. Characteristic fea- Anatomy of the North Anatolian Fault Zone tures of the North Anatolian Fault in the east- in the Marmara Sea, western Turkey: exten- ern Marmara region and its tectonic evolu- sional basins above a continental transform. tion. Mar. Geol. 190:329–50 GSA Today, June, pp. 3–7 Altınlı IE.˙ 1968. Geologic investigation of the Aksu AE, YaltırakC, Hiscott RN. 2002. Quater- Izmit-Hereke-Kurucada˘˙ g area. Bull. Miner. nary paleoclimatic-paleoceanographic and Res. Explor. Inst. Turk., No. 71, pp. 1–28 + tectonic evolution of the Marmara Sea V plates and environs. Mar. Geol. 190[ii]+552 Altınlı IE.˙ 1975. Geology of the middle Sakarya pp. River (Turkey). In Congress of Earth Sci- Aktar M, Org¨¨ ulüG.2001. Study of strong ences on the Occasion of the 50th Anniversary aftershocks of Izmit˙ and Düzce earth- of the Turkish Republic, Papers, pp. 161–96 quake sequence using regional moment ten- + 4 foldout plates. Ankara: Miner. Res. Ex- sor inversion. In Scientific Activities 2001 plor. Inst. ITU Faculty of Mines: Symp. Seismotec- Altınok Y, Alpar B, Yaltırak C. 2003. S¸arköy- tonics North-Western Anatolia-Aegean Re- Mürefte 1912 earthquake’s tsunami, exten- cent Turk. Earthq., Symp. Book, pp. 72–77. sion of the associated faulting in the Marmara Istanbul:˙ Istanbul˙ Tech. Univ. Sea, Turkey. J. Seismol. 7:329–46 Aktimur HT, Ate¸sS¸, Oral A. 1986. Altunel E, Akyüz S, Guest S, eds. 2003. A Bolucevresindeki ¸ deprem zararlarının special issue commemorating Aykut Barka. azaltılmasına yönelik sismik zonlama. Turk.¨ Turk. J. Earth Sci. 12(No. 1). 156 pp. Jeol. Kurumu Bul.¨ 29:43–49 Altunel E, Barka A, Akyüz S, Orhan A, Akyol IH.˙ 1940. Erzincan zelzelesi ve son eds. 2001. Aktif Tektonik Ara¸stırma Grubu feyezanlar dolayısıyla. Ulk¨ u¨ 14(No. 84):499; DördüncüToplantısı, Osmangazi Univ.¨ (No. 85):17; (No. 86):113 16–17 Kasım 2000, Eski¸sehir, ATAG-4, by California Institute of Technology on 01/09/13. For personal use only. Akyüz HS, Hartleb R, Barka A, Altunel E, Makaleler, Eski¸sehir, [II]+ 171 pp. Sunal G, et al. 2002. Surface rupture and slip Ambraseys NN. 1969. Some characteristic fea-

Annu. Rev. Earth Planet. Sci. 2005.33:37-112. Downloaded from www.annualreviews.org distribution of the 12 November 1999 Düzce tures of the North Anatolian Fault Zone. earthquake (M 7.1), North Anatolian Fault, Tectonophysics 9:143–65 Bolu, Turkey. Bull. Seismol. Soc. Am. 92:61– Ambraseys NN. 1975. Studies in historical 66 seismicity and tectonics. In Geodynamics Akyüz S, Barka A, Altunel E, Hartleb R, Sunal Today—A Review of Earth’s Dynamic Pro- G. 2000. Field observations and slip distribu- cesses, pp. 7–16. London: R. Soc. tion of the November 12, 1999 Düzce earth- Ambraseys NN. 2002a. The seismic activity of quake (M = 7.1), Bolu—Turkey. See Barka the Marmara Sea region over the last 2000 et al. 2000b, pp. 63–70 years. Bull. Seismol. Soc. Am. 92:1–18 Aldanmaz E, Pearce JA, Thirlwall MF, Mitchell Ambraseys NN. 2002b. The earthquake of 1509 JG. 2000. Petrogenetic evolution of late in the Sea of Marmara, Turkey, revisited. Cenozoic, post-collision volcanism in west- Bull. Seismol. Soc. Am. 91:1397–416 18 Mar 2005 19:32 AR AR233-EA33-02.tex AR233-EA33-02.sgm LaTeX2e(2002/01/18) P1: IKH

THE NORTH ANATOLIAN FAULT 99

Ambraseys NN, Finkel CF. 1987. The gölüne morfolojik mü¸sahadeler. Istanbul˙ Saros-Marmara earthquake of 9 August Univ.¨ Cograf.˘ Enst. Derg. No. 10, p. 146 1912. Earthq. Eng. Struct. Dyn. 15:189– Ardel A. 1964. Batı Karadeniz Bölgesinde mor- 211 folojik mü¸sahadeler. Istanbul˙ Univ.¨ Cograf.˘ Ambraseys NN, Finkel CF. 1991. Long-term Enst. Derg. 8(No. 14):62–75 seismicity of Istanbul˙ and of the Marmara Ardel A. 1965. Anadolu havzalarinin te¸sekkül Sea region. Terra Nova 3:527–39 ve tekâmülü hakkında dü¸sünceler. Istanbul˙ Ambraseys NN, Finkel CF. 1995. The seismic- Univ.¨ Cograf.˘ Enst. Derg. 8(No. 15):60–73 ity of Turkey and adjacent areas—a histor- Ardel A. 1967. I¸˙c Anadoluda, Kızılırmak yayı ical review, 1500–1800. Eren Istanbul.˙ 240 i¸cinde kalan sahada co˘grafˆımü¸sahadeler. pp. Istanbul˙ Univ.¨ Cograf.˘ Enst. Derg. 8:9–17 Ambraseys NN, Jackson JA. 1998. Faulting Armijo R, Meyer B, Hubert-Ferrari A, Barka A. associated with historical and recent earth- 1999. Westward propagation of North Ana- quakes in the Eastern Mediterranean region. tolian Fault into the Northern Aegean: timing Geophys. J. Int. 133:390–406 and kinematics. Geology 27:267–70 Ambraseys NN, White D. 1997. The seismicity Armijo R, Meyer B, Hubert-Ferrari A, Barka A. of the Eastern Mediterranean region 550–1 2000. Westward propagation of North Ana- BC: a re-appraisal. J. Earthq. Eng. 1:603–32 tolian Fault into the Northern Aegean: tim- Andrews DJ, Ben-Zion Y. 1997. Wrinkle-like ing and kinematics: Reply. Geology 28:188– pulse on a fault between different materials. 89 J. Geophys. Res. 102:5530–71 Armijo R, Meyer B, Navarro S, King G. 2002. Anonymous. 1973. Kuzey Anadolu Fayı ve De- Slip partitioning in the Sea of Marmara pull- prem Ku¸sa˘gı Simpozyumu (29, 30, 31 Mart apart: a clue to propagation processes of the 1972—Ankara), Tebli˘gler: M. T. A. Enst., North Anatolian Fault. Terra Nova 14:80–86 Ankara, 170 pp. Arpat E, S¸aro˘glu F. 1972. The east Anatolian Anonymous. 1977. Yeni Türkiye Atlası. M. S. fault system: thoughts on its development. B. Harita Genel Mud¨ url¨ u¨g˘u,¨ Ankara. Unpag- Bull. Miner. Res. Explor. Ins. Turk., No. 78, inated pp. 33–39 Anonymous. 2001. Türkiye Kuvaterneri Ç alı- Arpat E, S¸aro˘glu F. 1975. Türkiye’de bazı stayı,¸ Makaleler ve Ozetler:¨ Istanbul˙ Tek. onemli¨ gen¸c tektonik olaylar: Turk.¨ Jeol. Ku- Univ.¨ Avrasya Yerbilim. Enst., 21–22 Mayıs, rumu Bul.¨ 18:91–101 Istanbul.˙ 158 pp. Atakan K, Ojeda A, Meghraoui M, Barka AA, Anonymous. 2003a. Int. Workshop North Ana- Erdik M, Bodare A. 2002. Seismic hazard in tolian, East Anatolian and Dead Sea Fault Istanbul˙ following the 17 August 1999 Izmit˙ by California Institute of Technology on 01/09/13. For personal use only. Systems, Abstr., Recent Progress Tecton- and 12 November 1999 Düzce earthquakes. ics and Paleoseismology and Field Train. Bull. Seismol. Soc. Am. 92:466–82

Annu. Rev. Earth Planet. Sci. 2005.33:37-112. Downloaded from www.annualreviews.org Course Paleoseismology. Middle East Tech. Ataman G, Buket E, Ç apan UZ. 1975. Could Univ. (METU), Ankara, Turk. XV+184 pp. North Anatolian Fault Zone be a paleo- Anonymous 2003b. Türkiye Kuvaterneri Benioff Zone? Bull. Miner. Res. Explor. Inst. Ç alı¸stayı IV, Istanbul˙ Tek. Univ.¨ Avrasya Turk., No. 84, pp. 97–102 Yerbilim. Enst., 20–30 Mayıs, Istanbul.˙ 207 Aytun A. 1982. Creep measurements in the pp. Ismetpa¸˙ sa Region of the North Anatolian Ansal A, ed. 2001. Lessons learned from recent Fault Zone. See I¸sıkara & Vogel 1982, pp. strong earthquakes. Proc. 15th Int. Conf. Soil 279–92 Mech. Geotech. Eng., Earthq. Geotech. Eng. Aydın A, Nur A. 1982. Evolution of pull-apart Satell. Conf., Istanbul,˙ Turk. Maya Basın basins and their scale independence. Tecton- Yayın Ltd. S¸ti, Istanbul.˙ VIII+382 pp. ics 1:91–105 Ardel A. 1959. Izmit˙ körfezinden Iznik˙ Bargu S, Sakın¸cM.1989/1990. Izmit˙ Körfezi 18 Mar 2005 19:32 AR AR233-EA33-02.tex AR233-EA33-02.sgm LaTeX2e(2002/01/18) P1: IKH

100 S¸ENGOR¨ ET AL.

ile Iznik˙ Gölü arasında kalan bölgenin je- Barka A, Kozacı O,¨ Akyüz S, Altunel E, olojisi ve yapısalozellikleri. ¨ Istanbul˙ Univ.¨ eds. 2000b. The 1999 Izmit˙ and Duzce¨ Muhendis.¨ Fak. Yerbilim. Derg. 6:45–76 Earthquakes: Preliminary Results. Istanbul:˙ Barka A. 1985. Kuzey Anadolu Fay Zonundaki Istanbul˙ Tech. Univ.¨ X+349 pp. bazı Neojen-Kuaterner havzalarının jeolojisi Barka AA. 1981. Seismo-tectonic aspects of ve tektonik evrimi. Ketin Simp. Kitabi, Turk.¨ the North Anatolian Fault. PhD thesis. Univ. Jeol. Kurumu, Ankara, pp. 209–27 Bristol. [XX]+335 pp. Barka A. 1992. The North Anatolian Fault zone. Bartlett WL, Friedman M, Logan JM. 1981. Ann. Tecton. 6:164–95 Experimental folding and faulting of rocks Barka A. 1993. Erzincan Baseni,cevresinin ¸ under confining pressure: part IX. Wrench tektoni˘gi ve 13 Mart 1992 depremi. In faults in limestone layers. Tectonophysics Ulusal Deprem Muhendisli¨ gi˘ Konf. 10–13 79:255–77 Mart, TMMOB Ins˙ ¸aat Muhendisleri¨ Odası, Ba¸sH.1975. Petrologische und Geochemis- Istanbul˙ S¸ubesi, Deprem Muhendisli¨ gi˘ Turk.¨ che Untersuchungen an Subrezenten Vulka- Milli Komitesi, I.T.˙ U.¨ Yapı Deprem Uygu- niten der Nordanatolischen Storungszone¨ lama Aras¸tırma Merkezi, pp. 259–70 (Abschnitt: Erzincan-Niksar), Turkei.¨ PhD Barka A. 1996. Slip distribution along the North thesis. Univ. Hamburg, Ger., 116 pp. Anatolian Fault associated with the large Bellier O, Over¨ S, Poisson A, Andrieux J. 1997. earthquakes of the period 1939 to 1967. Bull. Recent temporal change in the stress state Seismol. Soc. Am. 86:1238–54 and modern stress field along the North Ana- Barka A. 1997. Neotectonics of the Marmara tolian Fault Zone (Turkey). Geophys. J. Int. Region. See Schindler & Pfister 1997, pp. 131:61–86 55–87 Benda L, Meulenkamp JE. 1979. Biostrati- Barka A. 1999. The 17 August 1999 Izmit˙ earth- graphic correlations in the eastern Mediter- quake. Science 285:1858–59 ranean Neogene. 5th calibration of sporo- Barka A, Akyüz S, Altunel E, Sunal G, Ç akır morph associations, marine microfossils and Z, et al. 2000a. The August 17, 1999 Izmit˙ mammal zones, marine and continental earthquake, M = 7.4, Eastern Marmara re- stages and the radiometric scale. Ann. Geol.´ gion, Turkey: study of surface rupture and Pays Hell. Ser.´ 1:61–70 slip distribution. See Barka et al. 2000b, Ben-Zion Y. 2001. Dynamic ruptures in recent pp. 15–30 models of earthquake faults. J. Mech. Phys. Barka A, Akyüz HS, Altunel E, Sunal G, Ç akır Solids 49:2209–44 Z, et al. 2002. The surface rupture and slip Ben-Zion Y, Andrews DJ. 1998. Properties distribution of the 17 August 1999 Izmit˙ and implications of dynamic rupture along by California Institute of Technology on 01/09/13. For personal use only. earthquake (M 7.4), North Anatolian Fault. a material interface. Bull. Seismol. Soc. Am. Bull. Seismol. Soc. Am. 92:43–60 88:1085–94

Annu. Rev. Earth Planet. Sci. 2005.33:37-112. Downloaded from www.annualreviews.org Barka A, Akyüz HS, Cohen HA, Watchorn Bergougnan H. 1976. Structure de la Chaˆıne F. 2000c. Tectonic evolution of the Niksar pontique dans le Haut-Kelkit (Nord-Est de and Ta¸sova-Erbaa pull-apart basins, North l’Anatolie). Bull. Soc. Geol.´ Fr. 18:675–86 Anatolian Fault Zone: their significance for Bilgin T. 1969. Biga Yarımadasının Güneybatı the motion of the Anatolian Block. Tectono- Kısmının Jeomorfolojisi. Istanbul˙ Univ.¨ physics 322:243–64 Yayın No. 1433; Cograf.˘ Enst. Yayın No. 54, Barka A, Gülen L. 1989. Complex evolution Istanbul,˙ IX+273 pp. of the Erzincan Basin (eastern Turkey). J. Bilgin T. 1984. Adapazarı Ovası ve Sapanca Struct. Geol. 11:275–83 Olu˘gunun Alüvyal Morfolojisi ve Kuater- Barka A, Kadinsky-Cade K. 1988. Strike-slip nerdeki Jeomorfolojik Tekâmülü: Istanbul˙ fault geometry in Turkey and its influence on Univ.¨ Edeb. Fak. Yayın No. 2572, XV+199 earthquake activity. Tectonics 7:663–84 pp.+8 foldouts in pocket 18 Mar 2005 19:32 AR AR233-EA33-02.tex AR233-EA33-02.sgm LaTeX2e(2002/01/18) P1: IKH

THE NORTH ANATOLIAN FAULT 101

Bingöl E. 1989. Türkiye Jeoloji Haritası/ ments in the interval 1989–1997. Geophys. Geological Map of Turkey, scale 1: J. Int. 135:195–214 2,000,000. Ankara: Maden tetkik ve Arama Ç orlu V, ed. 1999. Cogito Deprem Ozel¨ Sayısı. Genel Müdürlü˘gü. One sheet Cogito, No. 20. 375 pp. Blumenthal M. 1945. La ligne sismique de Crowell JC. 1974. Origin of late Cenozoic Lâdik. Maden Tetk. Arama Mecmuası, No. basins in Southern California. Soc. Econ. Pa- 1/33, pp. 162–74 leontol. Miner., Spec. Publ. 22:190–204 Blumenthal MM. 1948. Bolu Civarı ile Crowell JC. 1982. The tectonics of Ridge Basin, A¸sa˘gı Kızılırmak Mecrası Arasındaki Kuzey southern California. In Geologic History of Anadolu Silsilelerinin Jeolojisi. Maden Tetk. Ridge Basin Southern California, ed. JC Arama Enst. Yayın, Ser. B, No. 13, Ankara, Crowell, MH Link, Book 22:25–41. Los 265 pp.+54 photographic plates+ numerous Angeles: Pac. Sect. Soc. Econ. Paleontol. unnumbered foldout plates Mineral. Bozkurt E, Guest ed. 2001. Special issue Neo- Da˘gtekin H. 1981. Genel Tarih Atlası. Inkıl˙ apˆ tectonics of Turkey. Geodin. Acta 14(No. 1– Aka, Istanbul˙ , III+64+27 pp. 3) 212 pp. Deli F, Pérouse J-F. 1999. Le tremblement Bozkurt E, Holdsworth BK, Ko¸cyi˘git A. 1997. de terre de Yalova-Izmit-˙ Istanbul—Premi`˙ ers Implications of Jurassic chert identified in the El´´ ements d’Appréciation: Les dossiers de Tokat Complex, northern Turkey. Geol. Mag. l’IFEA, sér: la Turquie aujourd’hui, no. 134:91–97 1, Inst. Fran¸cais d’Etudes´ Anatoliennes Bozkurt E, Ko¸cyi˘git A. 1995. Almus Fault Georges Dumézil, Istanbul,˙ 40 pp. Zone: Its age, total offset and relation to the Demirba˘gE,Rangin C, Le Pichon X, S¸engör North Anatolian Fault Zone. Turk. J. Earth AMC. 2003. Investigation of the tectonics of Sci. 4:93–104 the Main Marmara Fault by means of deep- Bozkurt E, Ko¸cyi˘git A. 1996. The Kazova towed seismic data. Tectonophysics 361:1– basin: an active negative flower structure on 19 the Almus Fault Zone, a splay fault system Dewey JF, Helman ML, Turco E, Hutton DHW, of the North Anatolian Fault Zone, Turkey. Knott SD. 1989. Kinematics of the west- Tectonophysics 265:239–54 ern Mediterranean. Geol. Soc. London, Spec. Bozkurt E, Oberhänsli R. 2001. Menderes Mas- Publ. 45:265–83 sif (western Turkey): structural, metamor- Dewey JF, S¸engör AMC. 1979. Aegean and phic and magmatic evolution—a synthesis. surrounding regions: complex multiplate and Int. J. Earth Sci. 89:679–708 continuum tectonics in a convergent zone. Bozkurt E, Winchester JA, Piper JDA, eds. Geol. Soc. Am. Bull. 90:84–92 by California Institute of Technology on 01/09/13. For personal use only. 2000. Tectonics and magmatism in Turkey Dirik K. 1993. Geological history of the north- and the surrounding area. Geol. Soc. London ward arched segment of the North Anatolian +

Annu. Rev. Earth Planet. Sci. 2005.33:37-112. Downloaded from www.annualreviews.org Spec. Publ. 173, xvii 521 pp. Transform Fault Zone. Geol. J. 28:251–66 C¸a˘gatay N. 2003. Chronostratigraphy and sed- Do˘gan E, Kurter A, eds. 2000. Marmara imentology of the Marmara Sea over the last Denizi’nin Jeolojik O¸sinograﬁsi. Istanbul˙ 40 kyrs. Geophys. Res. Abstr. Vol.5, EAE03- Univ.¨ Deniz Bilim. Is˙¸letmeciligi˘ Enst., A-01883 Istanbul˙ , VII+504 pp. Canıtez N, U¸¨ cer SB. 1967. A Catalogue of Fo- Durukal E, U¸ckan E, S¸e¸setyan K, Demircio˘glu cal Mechanism Diagrams for Turkey and Ad- MB. 2002. MARSH Earthquake Risk As- joining Areas. IT˙ U¨ Maden Fak., Arz Fizigi˘ sessment for Industrial Facilities in Istanbul, Enst. Yayın No. 25. 111 pp. coordinated by M. Erdik: Bo˘gazi¸ci Univer- Clarke PJ, Davies RR, England PC, Parsons sity Kandilli Obs. Earthq. Res. Inst., Dep. B, Billiris H, et al. 1998. Crustal strain in Earthq. Eng., Istanbul.˙ 303 pp. central Greece from repeated GPS measure- Egeran EN. 1947. Tectonique de la Turquie 18 Mar 2005 19:32 AR AR233-EA33-02.tex AR233-EA33-02.sgm LaTeX2e(2002/01/18) P1: IKH

102 S¸ENGOR¨ ET AL.

et Relations entre les Unités Tectoniques Erdik M, Aydıno˘glu N, Fahjan Y, S¸e¸setyan K, et les Gˆıtes Métallifères de la Turquie: Demircio˘glu M, et al. 2003. Earthquake risk Georges Thomas, Nancy, IX+197 pp.+16 assessment for Istanbul Metropolitan area. photographs+4 foldout plates Earthq. Eng. Eng. Vib. 2(No. 1):1–25 Egeran N, Lahn E, compilers. 1942–1946. Erdik M, Yüzügüllü O,¨ Yılmaz Ç, Akka¸sN. Türkiye Jeolojik Hartası scale 1:800,000. 1992. 13 March 1992 (Ms = 6.8) Erzincan Maden Tetk. Arama Enst., Ankara. 8 sheets earthquake: a preliminary reconnaissance re- Egeran N, Lahn E. 1944. Note sur la carte sis- port. Soil Dyn. Earthq. Eng. 11:279–310 mique de la Turquie au 1:2.400.000. Maden Erin¸cS.1953. Do˘gu Anadolu Co˘grafyası: Tetk. Arama Enst. Mecmuası, year 9, No. Istanbul˙ Univ.¨ Yayın No. 572, Istanbul˙ Univ.¨ 2/32, pp. 279–89+one foldout map Edeb. Fak. Cograf.˘ Enst. Yayin No. 15, Eisenlohr T. 1997. The thermal springs of the Istanbul,˙ [III]+124 pp. Armutlu Peninsula (NW Turkey) and their Erin¸cS.1956. Yalova civarında bahri Pleistosen relationship to geology and tectonic. See tara¸caları ve depoları. Turk¨ Cograf.˘ Derg., Schindler & Pfister 1997, pp. 197–228 No. 15–16, pp. 188–90 Emre O,¨ Awata Y, Duman TY, eds. 2003. Erin¸cS.1973. Türkiye’ninsekillenmesinde ¸ 17 A˘gustos 1999 Izmit˙ Depremi Yüzey neotektoni˘gin rolüvejeomorfoloji-jeodina- Kırı˘gı/Surface Rupture Associated with the mik ili¸skileri. Jeomorfol. Derg., No. 5, pp. 17 August 1999 Izmit˙ Earthquake. Gen. Dir. 15–25 Miner. Res. Explor., Spec. Publ. 1, Ankara, Erin¸cS,Bilgin T, Bener M. 1961a. Gerede [VII]+280 pp.+ 1 foldout map in pocket cıvarında akarsusebekesi. ¸ Istanbul˙ Univ.¨ Emre O,¨ Duman TY, Do˘gan A, Ate¸sS¸, Ke¸cer Cograf.˘ Enst. Derg. 6:90–99 M, et al. 1999. 12 Kasim Düzce Depremi Erin¸cS,Bilgin T, Bener M. 1961b. Ça˘ga De- Saha Gözlemleri ve On¨ De˘gerlendirme Ra- presyonu ve Bo˘gazı. Istanbul˙ Univ.¨ Cograf.˘ poru. M.T.A. Genel Mud¨ url¨ u¨g˘u,¨ Jeol. Etutleri¨ Enst. Derg. 7:170–73 Dairesi, 16 Kasim, Ankara, [I]+27 pp. Eser Teknik Sondaj ve Ticaret AS¸.2000. Düzce Emre O,¨ Erkal T, Kazancı N, Görmü¸sS,Görür Kenti—I. Nolu Bölge (3444 Hektar–MIA N, et al. 1997. Güney Marmara’nın Neojen Yakın Çevresi) 12 Kasım Depremi Sonrası ve Kuaterner’deki Morfotektoni˘gi. In Guney¨ Yeniden Yapılanma Kapsamında Jeolojik- Marmara Bolgesinin¨ Neojen ve Kuaterner Jeoteknik Inceleme˙ Raporu, Ankara. 61 pp. Evrimi, TUB¨ ITAK˙ YDABÇAG-426/G Proje Eyido˘gan H, Utku Z, Gü¸clüU,De˘girmenci Raporu, pp. 36–68 (Open File Rep.) E. 1991. Türkiye Büyük Depremleri Makro- Emre O,¨ Erkal T, Tchepalyga A, Kazancı N, Sismik Rehberi (1900–1988). Istanbul˙ Tek. Ke¸cer M, Unay¨ E. 1998. Do˘gu Marmara Univ.¨ Maden Fak., Jeofizik Muhendisli¨ gi˘ by California Institute of Technology on 01/09/13. For personal use only. bölgesinin Neojen-Kuaternerdeki evrimi. Bol.¨ , [II]+198 pp. Maden Tetk. Arama Derg., No. 120, pp. 289– Farinacci A, Ager DV, Nicosia U, eds. 1991.

Annu. Rev. Earth Planet. Sci. 2005.33:37-112. Downloaded from www.annualreviews.org 314 Geology and Paleontology of Western Pon- Emre O,¨ Ku¸s¸cu I,˙ Do˘gan A, Duman TY, tides, Turkey—Jurassic-Early Cretaceous Ozalp¨ S, et al., eds. 2002. Aktif Tektonik Stratigraphy, Tectonics and Paleogeographic Ara¸stırma Grubu Altinci Toplantisi, Bildiri Evolution. Geol. Romana,Vol. 27, VIII+447 Ozetleri.¨ M.T.A. Genel Mud¨ url¨ u¨g˘u¨, 21–22 pp. Kasım, Ankara, iii+125+[8] pp. Fayon AK, Whitney DL, Teyssier C, Garver JI, Ercan T, Satir M, Steinitz G, Dora A, Dilek Y. 2001. Effects of plate convergence Sarıfakıo˘glu E, et al. 1995. Biga Yarımadasi obliquity on timing and mechanism of ex- ile G¨ok¸ceada, Bozcaada ve Tav¸san humation of a mid-crustal terrain, the central Adalarındakı (KB Anadolu) Tersiyer Anatolian crystalline complex. Earth Planet. volkanizmasınınozellikleri. ¨ Maden Tetk. Sci. Lett. 192:191–205 Arama Enst. Bul.¨ , No. 117, pp. 55–86 Fuenzalida H, Dorbath L, Cisternas A, 18 Mar 2005 19:32 AR AR233-EA33-02.tex AR233-EA33-02.sgm LaTeX2e(2002/01/18) P1: IKH

THE NORTH ANATOLIAN FAULT 103

Eyido˘gan H, Barka A, et al. 1997. Mecha- Mediterranean, ed. JE Dixon, AHF Robert- nism of the 1992 Erzincan earthquake and its son, Geol. Soc. London Spec. Publ. 17:467– aftershocks, tectonics of the Erzincan basin 82 and decoupling on the North Anatolian Fault. Görür N, Papadopoulos AG, Okay N, eds. 2002. Geophys. J. Int. 129:1–28 Integration of Earth Science Research on the Gaucher E. 1993. Modèle en profondeur et en Turkish and Greek 1999 Earthquakes. NATO vitesse du basin d’Erzincan. DEA Rep. 2, Sci. Ser. IV. Earth Environ. Sci. Vol. 9. Dor- Inst. Phys. Globe Paris, Jussieu, pp. 215–46 drecht: Kluwer. XVI+207 pp. Gaudemer YP, Tapponnier P, Turcotte DL. Görür N, S¸engör AMC, Akkök R, Yılmaz Y. 1989. River offsets across active strike-slip 1983. Pontidlerde Neo-Tetis’in a¸cılmasına faults. Ann. Tecton. 3:55–76 ili¸skin sedimentolojik veriler (Sedimento- Gautier P, Bozkurt E, Hallot E, Dirik K. 2002. logical evidence for the opening of the north- Dating the exhumation of a metamorphic ern branch of Neo-Tethys in the Pontides). dome: geological evidence for pre-Eocene Turk.¨ Jeol. Kurumu Bul.¨ 26:11–20 unroofing of the Nigde Massif (central Ana- Greber E. 1996. Stratigraphic evolution and tolia, Turkey). Geol. Mag. 139:559–76 tectonics in an area of high seismic- Gen¸c SC. 1993. Iznik-˙ Ineg˙ ol¨ (Bursa) ity:Akyazı/Adapazarı (Pontides, Northwest- Arasındaki Tektonik Birliklerin Jeolojik ern Turkey). Turk. J. Earth Sci. 5:63–79 ve Petrolojik Incelenmesi.˙ PhD thesis. Greber E. 1997. Stratigraphic evolution and Istanbul˙ Teknik Univ.,¨ Fen Bilim. Enst. 522 tectonics in an area of high seismic- pp. ity:Akyazı/Adapazarı (Pontides, Northwest- Gökmen V, Memiko˘glu O, Da˘glı M, Oz¨ D, ern Turkey). See Schindler & Pfister 1997, Tuncalı E. 1993. Türkiye Linyit Envanteri. pp. 141–60 Maden Tetk. Arama Genel Mud¨ url¨ u¨g˘u¨, Guidoboni E, Comastri A, Traina G. 1994. Ankara, [V]+356 pp. Catalogue of Ancient Earthquakes in the Gökten E, Varol B, Seyito˘glu G, Emre O,¨ Mediterranean Area up to the 10th Century. Ozaksoy¨ V, Iler˙ O,¨ eds. 2001. Aktif Tektonik Roma: Ist. Naz. Geofis. 504 pp. Ara¸stırma Grubu Be¸sinci Toplantısı Ankara Gürbüz C, Aktar M, Eyido˘gan H, Cisternas A, Univ.¨ Mühendis. Fak. Jeol. Mühendisli˘gi Haessler H, et al. 2000. The seismotectonics Bölümü 15–16 Kasım, Ankara, ATAG-5, of the Marmara region (Turkey): results from Bildiri Ozetleri,¨ Ankara, [III]+42+3 pp. a macroseismic experiment. Tectonophysics Görür N. 2002. TUB¨ ITAK’ın˙ Marmara 316:1–17 Denizi’nde Koordine Etti˘gi Deprem Ara¸stır- Helvacı C, Griffin WL. 1984. Rb-Sr geochro- maları. TUB¨ ITAK˙ Marmara Aras¸tırma nology of the Bitlis Massif, Avnik (Bingöl) by California Institute of Technology on 01/09/13. For personal use only. Merkezi, Gebze, Kocaeli, 59 pp. area, S. E. Turkey. Geol. Soc. London Spec. Görür N. 2003. Marmara Denizi’nin Deprem Publ. 17:403–13

Annu. Rev. Earth Planet. Sci. 2005.33:37-112. Downloaded from www.annualreviews.org Potansiyeli. Ankara: Türk. Bilim. Akad. Hempton M, Dunne LA. 1983. Sedimentation Yayın. 60 pp. in pull-apart basins: active examples in east- Görür N, Ça˘gatay MN, Sakın¸cM,Sümengen ern Turkey. J. Geol. 92:513–30 M, S¸entürk K, et al. 1997. Origin of the Sea Herece E, Akay E. 2003. Kuzey Anadolu of Marmara as deduced from Neogene to Fayı (KAF) Atlası/Atlas of North Anato- Quaternary paleogeographic evolution of its lian Fault (NAF). Maden Tetk. Arama Genel frame. Int. Geol. Rev. 39:342–52 Mud¨ url¨ u¨g˘u,¨ Ozel¨ Yayın. Ser. 2, Ankara, Görür N, Oktay FY, Seymen I,˙ S¸engör [IV]+61 pp.+13 appendices as separate AMC. 1984. Palaeotectonic evolution of the maps Tuzgölü basin complex, central Turkey: sed- Hewsen RH. 2001. Armenia—A Historical imentary record of a Neo-Tethyan closure. Atlas. Chicago: Univ. Chicago Press. xviii+ In Geological Evolution of the Eastern 341 pp. 18 Mar 2005 19:32 AR AR233-EA33-02.tex AR233-EA33-02.sgm LaTeX2e(2002/01/18) P1: IKH

104 S¸ENGOR¨ ET AL.

Hill ML, Diblee TW Jr. 1953. San Andreas, Turkey and Pakistan. Geophys. J. R. Astron. Garlock, and Big Pine Faults, California. Soc. 77:185–264 Geol. Soc. Am. Bull. 64:443–58 Karaalio˘glu B. 1973. Merzifon-Gümü¸shacıköy Hodgson JH, Wickens AJ. 1965. Computer- Ovası Hidrojeolojik Etüt Raporu: Türkiye determined P-nodal solutions for the larger Cumhuriyeti Enerji ve Tabii Kaynaklar earthquakes of 1959–1962. Publ. Dom. Obs., Bakanlı˘gı, Devlet Su I¸˙sleri Genel Ottawa 31:123–43 Müdürlü˘gü, Ankara, 50 pp. Harvard Univ. 1998. Dep. Earth Planet. Sci., Karaca M, Ural D, eds. 1999. ITU-IAHS Int. Cambridge, MA. http://www.seismology. Conf. Kocaeli Earthq. 17 August—A Scien- harvard.edu tific Assessment and Recommendations for Hubert-Ferrari A, Armijo R, King GCP, Meyer Re-Building. Istanbul˙ Tech. Univ., Istanbul,˙ B, Barka A. 2002. Morphology, displace- [IV]+382 pp. ment, and slip rates along the North Ana- Kaya O. 1988. A possible early Cretaceous tolian Fault, Turkey. J. Geophys. Res. 107: thrust origin for the ancestral North Anato- 2235 lian Fault. METU J. Pure Appl. Sci. 21:105– Imren˙ C, Le Pichon X, Rangin C, Demirb˘g 26 E, Ecevito˘glu B, Görür N. 2001. The North Kaypak B. 2002. Erzincan Havzası 3-B Hız Anatolian Fault within the Sea of Marmara: Yapısının Yerel Deprem Tomografisi ile anew interpretation based on multi-channel Belirlenmesi. PhD thesis. IT˙ U¨ Fen Bilim. seismic and multi-beam bathymetry data. Enst., Istanbul.˙ 216 pp. Earth Planet. Sci. Lett. 186:143–58 Kazancı N. 1993. Kuzey Anadolu Fayıuzerinde ¨ Inandık˙ H. 1952/1953. Adapazarı Ovası ve tektonik kontrollü depolanmaorne˘ ¨ gi: Su¸se- cevresinin¸ jeomorfolojik etüdü. Istanbul˙ hri Havzası (Alt Pliyosen-Holosen), Türkiye. Univ.¨ Cografya˘ Enst. Derg. 2(No. 3–4):109– Doga-T˘ urk¨ Yerbilim. Derg. 2:89–102 38+ 3 photographic plates Ketin I.˙ 1948. Uber¨ die tektonisch-mechani- Int. Seismol. Centre. 2003. ISC Catalogue schen Folgerungen aus den grossen ana- 1904–2000, Edinburgh. CD-ROM tolischen Erdbeben des letzten Dezenniums. Irrlitz W. 1971. Neogene and older Pleistocene Geol. Rund. 36:77–83 intra-montane basins in the Pontic region of Ketin I.˙ 1957. Kuzey Anadolu Deprem Fayı. Anatolia. Newsl. Stratigr. 1:33–36 IT˙ U¨ Derg. 15:49–52 Irrlitz W. 1972. Lithostratigraphie und tektonis- Ketin I.˙ 1968. Relations between general tec- che Entwicklung des Neogens in Nordost- tonic features and the main earthquake re- anatolien (Känozoikum und Braunkohlen in gions in Turkey. Bull. Miner. Res. Explor. der Türkei. 6.). Beih. Geol. Jahrb., No. 120, Inst. Turk. 71:63–67 by California Institute of Technology on 01/09/13. For personal use only. 111 pp.+10 plates Ketin I.˙ 1969. Uber¨ die nordanatolische Hori- I¸sıkara AM, Vogel A, eds. 1982. Multidisci- zontalverschiebung. Bull. Miner.Res. Explor.

Annu. Rev. Earth Planet. Sci. 2005.33:37-112. Downloaded from www.annualreviews.org plinary Approach to Earthquake Prediction. Inst. Turk. 72:1–28 Braunschweig/Wiesbaden: Friedrich Vieweg Ketin I.˙ 1976. San Andreas ve Kuzey Anadolu & Sohn Fayları arasında bir kar¸sılastırma. Turk.¨ Jeol. Izbırak˙ R. 1948. Akda˘g ile Güney Çevresinde Kurumu Bul.¨ 19:149–54 Jeomorfoloji Gözlemleri ve Bu Çevrenin Ketin I,˙ Rösli F. 1953. Makroseismische Un- Bayındırla¸sması Hakkında Bazı Dü¸sünceler. tersuchungenuber ¨ das nordwest-anatolische Ankara Universitesi¨ Dil ve Tarih-Cog˘grafya˘ Beben vom 18 März 1953. Eclogae Geol. Fak. Yayımları No. 59, Cografya˘ Enst. Helv. 46:187–208 No. 3. 82 pp. +7 maps+17 sections+19 Kinder H, Hilgemann W. 1982. Atlas zur photographs+1 graph Weltgeschichte—Von den Anfangen¨ bis zur Jackson J, McKenzie DP.1984. Active tectonics Gegenwart. München: Piper. IX+287 pp. of Alpine-Himalayan Belt between western King GCP, Hubert-Ferrari A, Nalbant SS, 18 Mar 2005 19:32 AR AR233-EA33-02.tex AR233-EA33-02.sgm LaTeX2e(2002/01/18) P1: IKH

THE NORTH ANATOLIAN FAULT 105

Meyer B, Armijo R, Bowman D. 2001. East Tech. Univ. (METU), Ankara, Turk. Coulomb interactions and the 17 August 27 pp. 1999 Izmit,˙ Turkey earthquake. C.R. Acad. Ko¸cyi˘git A, Beyhan A. 1998. A new intra- Sci. II Ser. A 333:557–69 continental transcurrent structure: The Cen- Kiratzi AA. 1993. A study of the active crustal tral Anatolian Fault Zone, Turkey. Tectono- deformation of the North and East Anatolian physics 284:317–36 fault zones. Tectonophysics 225:191–203 Ko¸cyi˘git A, Rojay B, Cihan M, Ozacar¨ A. 2001. Kober L. 1914. Die Bewegungsrichtung der The June 6, 2000 Orta (Ç ankırı, Turkey) Alpinen Deckengebirge des Mittelmeeres. earthquake: sourced from a new antithetic Petermanns Geogr. Mitt. Jg. 60(Part 2):pp. sinistral strike-slip structure of the North 250–56 Anatolian Fault system, the Dodurga Fault Kober L. 1921. Der Bau der Erde. Gebrüder Zone. Turk. J. Earth Sci. 10:69–82 Borntraeger. Berlin [II]+324 pp.+1 foldout Ko¸cyi˘git A, Yılmaz A, Adamia S, Kuloshvili S. plate 2001. Neotectonics of East Anatolian Plateau Kocaefe S, Ataman G. 1976. Anadolu’da sis- (Turkey) and Lesser Caucasus: implications motektonik olaylar: Antalya-Fethiye-Denizli for transition from thrusting to strike-slip u¸¨cgeni i¸cinde yer alan bölgenin incelenmesi. faulting. Geodin. Acta 14:177–95 Hacettepe Univ.¨ Yerbilim. Derg. 2:55–70 Kopp K-O. 1964. Geologie Thrakiens II: Die Ko¸cyi˘git A. 1988. Tectonic setting of the Geyve Inseln und der Chersones. Neues Jahr. Geol. Basin: Age and total displacement of the Palaontol.¨ Abh. 119:172–214 Geyve Fault Zone. METU J. Pure Appl. Sci. Kröner A, S¸engör AMC. 1990. Archean and 21:81–104 Proterozoic ancestry in late Precambrian to Ko¸cyi˘git A. 1989. Su¸sehri basin: an active fault- early Paleozoic crustal elements of southern wedge basin on the North Anatolian Fault Turkey as revealed by single zircon dating. Zone, Turkey. Tectonophysics 167:13–29 Geology 18:1886–90 Ko¸cyi˘git A. 1990. Tectonic setting of the Kropschot SJ, ed. 2000. Implications for Earth- Gölova Basin: total offset of the North Ana- quake Risk Reduction in the United States tolian Fault Zone, E Pontide, Turkey. Ann. from the Kocaeli, Turkey, earthquake of Au- Tecton. Spec. issue IV:155–70 gust 17, 1999. U.S. Geol. Surv. Circ. 1193. Ko¸cyi˘git A. 1991. An example of an accre- 64 pp. tionary forearc basin from northern central Le Pichon X, Chamot-Rooke N, Huchon P, Anatolia and its implications for the history Luxey P. 1993. Implications des nouvelles of subduction of Neo-Tethys in Turkey. Geol. mesures de géodésie spatiale en Grèce et en Soc. Am. Bull. 103:22–36 Turquie sur l’extrusion latérale de l’Anatolie by California Institute of Technology on 01/09/13. For personal use only. Ko¸cyi˘git A. 1996. Superimposed basins and et de l’Egée. C.R. Acad. Sci. Ser. II 316:983– their relations to the Recent strike-slip fault 90

Annu. Rev. Earth Planet. Sci. 2005.33:37-112. Downloaded from www.annualreviews.org zone: a case study of the Refahiye super- Le Pichon X, Chamot-Rooke N, Rangin C, imposed basin adjacent to the North Ana- S¸engör AMC. 2003. The North Anatolian tolian Transform Fault, northeastern Turkey. Fault in the Sea of Marmara. J. Geophys. Res. Int. Geol. Rev. 38:701–13 108(B4):2179 Ko¸cyi˘git A. 2003. The Gerede-Kaynaslı Fault Le Pichon X, S¸engör AMC, Demirba˘gE,Ran- Zones of the North Anatolian Fault System gin C, Imren˙ C, et al. 2001. The active (NAFS)-Turkey–A Post-International Work- Main Marmara Fault. Earth Planet. Sci. Lett. shop Excursion Guidebook. Int. Workshop 192:595–616 North Anatolian Dead Sea Fault Systems. Leuchs K. 1940. Das jüngste Grossbeben in Recent Progress in Tectonics and Paleoseis- Anatolien. Geol. Rund. 31:70–76 mology, and Field Training Course in Pa- Linneman S. 2002. Quaternary volcanism of the leoseismology,31Aug.–12 Sept., Middle Erzincan Basin, Eastern Turkey: an example 18 Mar 2005 19:32 AR AR233-EA33-02.tex AR233-EA33-02.sgm LaTeX2e(2002/01/18) P1: IKH

106 S¸ENGOR¨ ET AL.

of pull-apart basin volcanism. In The Tec- Nowack E. 1932. Kreide-Entwicklung und tonics of Eastern Turkey and the Northern Großtektonik in Nord-Anatolien. Cent. Arabian Plate, Int. Workshop, 23–25 Sept., Miner.,Geol. Palaontol.¨ Jg. 1932(Part B, No. Erzurum, Turkey,p.19 6):286–99 Mallet R. 1862. Great Neapolitan Earthquake Ocal¨ N. 1966. Geometrical solutions of fault- of 1857—The First Principles of Observa- plane problem of the destructive earthquakes tional Seismology. Vol. II. London: Chap- occurred in Anatolia in the period 1938– man & Hall. viii+399 pp.+3 foldout plates 1955. Z. Geophys. Jg. 32(Heft 5/6):293– (Reprinted 1987 Storia Geofisica Ambiente, 309 Bologna) Okay A, Tüysüz O. 1999. Tethyan sutures McClusky S, Balassanian S, Barka A, Demir of northern Turkey. In The Mediterranean C, Ergintav S, et al. 2000. Global Positioning Basins: Tertiary Extension within the Alpine System constraints on plate kinematics and Orogen, Geol. Soc. London, Spec. Publ. 156, dynamics in the eastern Mediterranean and ed. B Durand, L Jolivet, F Horváth, M Caucasus. J. Geophys. Res. 105:5695–719 Séranne, pp. 475–515 McKenzie DP. 1972. Active tectonics of the Okay AI,˙ Demirba˘gE,Kurt H, Okay N, Ku¸s¸cu Mediterranean region. Geophys. J. R. Astron. I.˙ 1999. An active, deep marine strike-slip Soc. 30:109–85 basin along the North Anatolian Fault in McKenzie D, Jackson J. 1983. The relation- Turkey. Tectonics 18:129–47 ship between strain rates, crustal thickening, Okay AI,˙ Ka¸slılar-Ozcan¨ A, Imren˙ C, Boztepe- palaeomagnetism, finite strain and fault mo- Güney A, Demirba˘gE,Ku¸s¸cu I.˙ 2000. Ac- ments within a deforming zone. Earth Planet. tive faults and evolving strike-slip basins in Sci. Lett. 65:182–202 the Marmara Sea, northwest Turkey: a mul- Meade BJ, Hager BH, McClusky SC, Reilinger tichannel seismic reflection study. Tectono- RE, Ergintav S, et al. 2002. Estimates of seis- physics 321:189–218 mic potential in the Marmara Sea region from Okay AI,˙ S¸ahintürk O.¨ 1997. Geology of the block models of secular deformation con- Eastern Pontides. In Regional and Petroleum strained by Global Positioning System mea- Geology of the Black Sea and Surrounding surements. Bull. Seismol. Soc. Am. 92:208– Region, Am. Assoc. Pet. Geol. Mem. 68, ed. 15 AG Robinson, pp. 291–311 Meri¸cE,ed. 1995. Izmit˙ Korfezi¨ Kuvaterner Okay N, Okay AI.˙ 2002. Tectonically in- Istifi.˙ Kocaeli Valili˘gi Çevre Koruma Vakfı, duced Quaternary drainage diversion in the Kocaeli. XIV+354 pp. northeastern Aegean. J. Geol. Soc. London Mihailovi¸cJ.1923. Le mécanisme des mou- 159:393–99 by California Institute of Technology on 01/09/13. For personal use only. vements séismiques de la Mer de Marmara. Org¨¨ ulüG,Aktar M. 2001. Regional moment Glas Srpske Krjalevske Akad. Nauk 108:18– tensor inversion for strong aftershocks of the ˙ =

Annu. Rev. Earth Planet. Sci. 2005.33:37-112. Downloaded from www.annualreviews.org 63 August 17, 1999 Izmit earthquake (Mw Mihailovi¸cJ.1927. Les grandes catastrophes 7.4). Geophys. Res. Lett. 28:371–74 séismiques sur la Mer de Marmara. Glas Over¨ S. 1996. Analyse tectonique et Etat de Srpske Krjalevske, Akad. Nauk, posebna iz- Contrainte cenozo¨ıque dans la Zone cen- danja 65 (+prirod.-mat. sp. 16), Belgrad. trale de la Faille Nord Anatolienne (Turquie). 303 pp. Thèse. Univ. Paris XI. 294 pp. Mittwede SK, Bozkurt E, guest eds. 2001. Over¨ S, Bellier O, Poisson A, Andrieux J. 1997. Special issue on Turkey. Int. Geol. Rev. Late Cenozoic stress state changes along the 43(No. 7–8):577–770 central North Anatolian Fault Zone. Ann. Nowack E. 1928. Die wichtigsten Ergebnisse Tecton. 11:75–101 meiner anatolischen Reisen. Z. Dtsch. Geol. Ozalaybey¨ S, Ergin M, Aktar M, Tapırdamaz C, Ges. 80(Monatsbericht, No. 8–10):304–12 Bii¸cmen F, Yörük A. 2002. The 1999 Izmit˙ 18 Mar 2005 19:32 AR AR233-EA33-02.tex AR233-EA33-02.sgm LaTeX2e(2002/01/18) P1: IKH

THE NORTH ANATOLIAN FAULT 107

earthquake sequence in Turkey: seismologi- JH. 2000. Heightened odds of large earth- cal and tectonic aspects. Bull. Seismol. Soc. quakes near Istanbul:˙ An interaction-based Am. 92:376–86 probability calculation. Science 288:661–65 Ozalaybey¨ S, Karabulut H, Ergin M, Aktar Pavoni N. 1961. Die Nordanatolische Horizon- M, Tapırdamaz C, et al. 2003. Seismogenic talverschiebung. Geol. Rund. 51:122–39 zones of the Sea of Marmara: recent seis- Pérouse J-F. 2001. Turquie: l’après séismes: micity and focal mechanism solutions. EGS- Les dossiers de l’IFEA, sér.: la Turquie au- AGU-EUG Jt. Assembly, Nice, Fr.,Vol. 5, jourd’hui, no. 7, Inst. Fr. d’Etudes´ Ana- pp. 12,680 tolinnes Georges Dumézil, Istanbul,˙ 52 pp. Ozeren¨ MS. 2002. Crustal structure and forces Pfannenstiel M. 1944. Diluviale Geologie des in continental deformation. PhD thesis. Univ. Mittelmeergebietes: Die diluvialen Entwick- Cambridge, Fitzwilliam Coll., 255 pp. lungsstadien und die Urgeschichte von Dar- Ozt¨¨ urk A. 1979. Lâdik-Destek yöresinin strati- danellen, Marmarameer und Bosporus. Ein grafisi. Turk.¨ Jeol. Kurumu Bul.¨ 22:27–34 Beitrag zu den klimatisch bedingten eu- Ozt¨¨ urk A. 1980. Lâdik-Destek yöresinin tek- statischen Spiegelschwankungen des Mit- toni˘gi. Turk.¨ Jeol. Kurumu Bul.¨ 23:31–38 telmeeres. Geol. Rund. 34:334–42 Pamir HN. 1944a. Une ligne sismogénique en Philippson A. 1918. Kleinasien. In Handbuch Anatolie septentrionale. Rev. Fac. Sci. Ser. 3, der Regionalen Geologie, ed. G Steinmann, 9:144–58 OWilckens, Vol. V, Part 2, 22nd issue. Pamir HN. 1944b. Kuzey Anadolu’nun deprem- Heidelberg: Carl Winters Univ. 183 pp.+3 ler yapan bir beresi. Türk FizikˆıveTabiˆı Il- foldout plates imler Sosyetesi Yıllık Bildirileri ve Ar¸sivi, Pınar N. 1943. Marmara denizi Havzasının sayı 12, s. 4 Sismik Jeoloji ve Meteorolojisi/Geologie´ et Pamir HN, Ketin I.˙ 1940. Das Erdbeben in der Met´ eorologie´ Sismique du Bassin de Mer Türkei vom 27./28. Dezember 1939. Geol. Marmara. PhD thesis. Inst. Géol., Inst. Rund. 31:77–78 Phys. Gén. Univ. Istanbul,˙ Kenan Matbaası. Pamir HN, Ketin I.˙ 1941. Das anatolische Erd- 63 pp.+1 foldout map beben Ende 1939. Geol. Rund. 32:279–87 Pınar N. 1953. Etude géologique et macrosis- Papadopoulos GA. 2002. The Athens, Greece, mique du tremblement de terre de Kur¸sunlu earthquake (Ms 5.9) of 7 September 1999: (Anatolie septentrionale) du 13 août 1951. an event triggered by the Izmit,˙ Turkey, 17 Rev. Fac. Sci. Univ. Istanbul,˙ Ser.´ A 18: August 1999 earthquake? Bull. Seismol. Soc. 142 Am. 92:312–21 Piper JDA, Moore JM, Tatar O, Gürsoy H, Park Papadopoulos GA, Kondopoulou DP, Leven- RG. 1996. Palaeomagnetic study of crustal by California Institute of Technology on 01/09/13. For personal use only. takis GA, Pavlides SB. 1986. Seismotec- deformation across an intracontinental trans- tonics of the Aegean region. Tectonophysics form: the North Anatolian Fault Zone in

Annu. Rev. Earth Planet. Sci. 2005.33:37-112. Downloaded from www.annualreviews.org 124:67–84 northern Turkey. In Palaeomagnetism and Paréjas E, Akyol IH,˙ Altınlı E. 1942. Le trem- Tectonics of the Mediterranean Region, Geol. blement de terre d’Erzincan du 27 Décembre Soc. London Spec. Publ. 105, ed. A Morris, 1939 (secteur occidentale). Rev. Fac. Sci. DH Tarling, pp. 299–310 Univ. Istanbul,˙ Ser.´ B 6(Nos. 3–4):187– Piper JDA, Tatar O, Gürsoy H. 1997. De- 222 formational behaviour of continental litho- Parke JR, Minshull TA, Anderson G, White RS, sphere deduced from block rotations across McKenzie D, et al. 1999. Active faults in the North Anatolian Fault Zone in Turkey. the Sea of Marmara, western Turkey, imaged Earth Planet. Sci. Lett. 150:191–203 by seismic reflection profiles. Terra Nova Platzman ES, Platt JP, Tapırdamaz C, San- 11:223–27 verM,Rundle CC. 1994. Why are there Parsons T, Toda S, Stein R, Barka A, Dieterich no clockwise rotations along the North 18 Mar 2005 19:32 AR AR233-EA33-02.tex AR233-EA33-02.sgm LaTeX2e(2002/01/18) P1: IKH

108 S¸ENGOR¨ ET AL.

Anatolian Fault Zone? J. Geophys. Res. PhD thesis. Istanbul˙ Univ.,¨ Fen Bilim. Enst., 99:21705– 15 Istanbul.˙ 240 pp.+7 foldouts Rangin C, Demirba˘gE,Imren˙ C, Crusson A, S¸aro˘glu F. 1988. Age and offset of the North Normand A, et al. 2001. Marine Atlas of the Anatolian Fault. METU J. Appl. Sci. 21:65– Sea of Marmara (Turkey). IFREMER, Paris, 79 11 plates S¸aro˘glu F, Emre O,¨ Boray A. 1987. Türkiye’nin Reilinger RE, McCLusky SC, Oral MB, King Diri Faylari ve Depremsellikleri: Maden RW,Toksoz MN. 1997. Global Position- Tetk. Arama Genel Mud¨ url¨ u¨g˘uJ¨ eol. Etud.¨ ing System measurements of present-day Dairesi Bas¸kanlıgı,˘ Ankara, III+394 pp.+11 crustal movements in the Arabia-Africa- maps in pocket (Open File Rep.) Eurasia plate collision zone. J. Geophys. Res. S¸aro˘glu F, Emre O,¨ Ku¸s¸cu I.˙ 1992. Türkiye Diri 102:9983–99 Fay Haritası (Active Fault Map of Turkey), Rockwell T, Barka A, Dawson T, Akyuz S, scale 1:2,000,000, one sheet. Maden Tetk. Thorup K. 2001. Paleoseismology of the Arama Genel Mud¨ url¨ u¨g˘u,¨ Ankara Gaziköy-Saros segment of the North Anato- Schindler C, Pfister M, eds. 1997. Active lian Fault, nortwestern Turkey: comparison Tectonics of Northwestern Anatolia—The of the historical and paleoseismic records, Marmara Poly-Project. A Multidisciplinary implications of regional seismic hazard, and Approach by Space-Geodesy, Geology, models of earthquake recurrence. J. Seismol. Hydrogeology, Geothermics and Seismology. 5:433–48 Zürich: ETH Sakın¸cM,Yaltırak C. 1997. Güney Trakya S¸enel M, ed. 2002. Türkiye Jeoloji Haritası/ sahillerinin denizel Pleyistosenc¨ ¸okelleri ve Geological Map of Turkey, scale 1:500,000. paleoco˘grafyası. Maden Tetk. Arama Enst. Maden Tetk. Arama Genel Mud¨ url¨ u¨g˘u,¨ Derg. No. 119, pp. 43–62 Ankara. 19 sheets Sakın¸cM,Yaltırak C, Oktay FY. 1999. S¸engör AMC. 1979a. The North Anatolian Palaeogeographical evolution of the Thrace Transform Fault: its age, offset and tectonic Neogene Basin and the Tethys-Paratethys re- significance. J. Geol. Soc. London 136:269– lations at northwest Turkey (Thrace). Palaeo- 82 geogr. Palaeoclimatol. Palaeoecol. 153:17– S¸engör AMC. 1979b. Mid-Mesozoic closure of 40 Permo-Triassic Tethys and its implications. Salomon-Calvi W. 1936. Die Fortsetzung der Nature 279:590–93 Tonalelinie in Kleinasien. Yuksek¨ Ziraat S¸engör AMC. 1984. The Cimmeride orogenic Enst. Ç alıs¸malarından, No. 9, pp. 11–13 system and the tectonics of Eurasia. Geol. Salomon-Calvi W. 1940a. Les tremblements Soc. Am. Spec. Pap. 195. xi+82 pp. by California Institute of Technology on 01/09/13. For personal use only. de terre d’Erzincan du 21 novembre et 27 S¸engör AMC. 1989. The Tethyside orogenic décembre 1939. Rev. Etude´ Calam. 3:178– system: An introduction. In Tectonic Evolu-

Annu. Rev. Earth Planet. Sci. 2005.33:37-112. Downloaded from www.annualreviews.org 80 tion of the Tethyan Region, ed. AMC S¸engör, Salomon-Calvi W.1940b. Kurze Ubersicht¨ uber ¨ pp. 1–22. Dordrecht: Kluwer den tektonischen Bau Anatoliens. Maden S¸engör AMC. 1990a. Plate tectonics and Tetk. Arama Enst. Mecmuası Year 5, No. orogenic research after 25 years: A 1/18, pp. 35–74 Tethyan perspective. Earth Sci. Rev. 27:1– Sarıbudak M, Sanver M, S¸engör AMC, Görür 201 N. 1990. Palaeomagnetic evidence for sub- S¸engör AMC. 1990b. A new model for the late stantial rotation of the Almacık Flake within Palaeozoic-Mesozoic tectonic evolution of the North Anatolian Fault Zone, northwest Iran and implications for Oman. Geol. Soc. Turkey. Geophys. J. Int. 102:563–68 London Spec. Publ. 49:797–831 S¸aro˘glu F. 1985. Dogu Anadolu’nun Neotek- S¸engör AMC. 1995. Sedimentation and tecton- tonik Donemde¨ Jeolojik ve Yapısal Evrimi. ics of fossil rifts. In Tectonics of Sedimentary 18 Mar 2005 19:32 AR AR233-EA33-02.tex AR233-EA33-02.sgm LaTeX2e(2002/01/18) P1: IKH

THE NORTH ANATOLIAN FAULT 109

Basins, ed. CJ Busby, RV Ingersoll, pp. 53– sif, Western Turkey: implications for tec- 117. Oxford: Blackwell tonic evolution and evidence for Pan-African S¸engör AMC. 1996. Kuzey Anadolu Fayı’nın basement in Turkey. Tectonics 3:693– Ke¸sfi ...: Bilim Tek., No. 338, pp. 8–15 707 S¸engör AMC, Burke K, Dewey JF. 1982. Tec- S¸engör AMC, Tatar O, eds. 1996. Special is- tonics of the North Anatolian Transform sue on Turkey. Int. Geol. Rev. 38(No. 8): 91– Fault. See I¸sıkara & Vogel 1982, pp. 3– 786 22 S¸engör AMC, Tatar O, Ernst GW, eds. 1996. S¸engör AMC, Büyüka¸sıko˘glu S, Canıtez N. Special issue on Turkey. Int. Geol. Rev. 1983. Neotectonics of the Pontides: implica- 38(No. 9):787–882 tions for ‘incompatible‘ structures along the S¸engör AMC, Yılmaz Y. 1981. Tethyan evo- North Anatolian fault. J. Struct. Geol. 5:211– lution of Turkey: a plate tectonic approach. 16 Tectonophysics 75:181–241 S¸engör AMC, Canıtez N. 1982. The North Ana- S¸engör AMC, Yılmaz Y, Ketin I.˙ 1980. Rem- tolian Fault. In Alpine-Mediterranean Geo- nants of a pre-late Jurassic ocean in northern dynamics, Geodynamics, ed. H Berekhemer, Turkey: fragments of Permo-Triassic Paleo- KHsü, Ser. 7, pp. 205–16. Boulder, CO: Tethys? Geol. Soc. Am. Bull. 91(Part I):599– Geol. Soc./Washington, DC: Am. Geophys. 609 Union S¸engör AMC, Yılmaz Y, Ketin I.˙ 1982. Rem- S¸engör AMC, Demirba˘gE,Tüysüz O, Kurt H, nants of a pre-late Jurassic ocean in northern Görür N, Ku¸s¸cu I. 1999. A preliminary note Turkey: fragments of Permo-Triassic Paleo- on the structure of the Gulf of Izmit:˙ impli- Tethys? Reply. Geol. Soc. Am. Bull. 93:932– cations for the westerly prolongation of the 36 North Anatolian Fault. See Karaca & Ural Seymen I.˙ 1975. Kelkit Vadisi Kesiminde 1999, pp. 25–37 Kuzey Anadolu Fay Zonunun Tektonik S¸engör AMC, Görür N, S¸aro˘glu F. 1985. Strike- Ozelli˘¨ gi: Dr. Eng., Istanbul˙ Tek. Univ.,¨ slip faulting and related basin formation in Maden Fak., XIX+ 192 pp. +2 foldout maps zones of tectonic escape: Turkey as a case Seymen I.˙ 1985. Kır¸sehir Masifi metamorfit- study. In Strike-slip Deformation, Basin For- lerinin jeoloji evrimi. In Ketin Simp. Kitabı, mation, and Sedimentation, Soc. Econ. Pale- Turk.¨ Jeol. Kurumu, Ankara, ed. T Ercan, ontol. Miner. Spec. Publ. 37 (in honor of J.C. MA Ça˘glayan, pp. 133–48 Crowell), ed. KT Biddle, N Christie-Blick, Seymen I,˙ Aydın A. 1972. The Bingöl earth- pp. 227–64 quake fault and its relation to the North Ana- S¸engör AMC, McKenzie D. 1997. Memorial to tolian Fault Zone. Bull. Miner. Res. Explor. by California Institute of Technology on 01/09/13. For personal use only. Ihsan˙ Ketin. Geol. Soc. Am. Memorl. 28:31– Inst., No. 79, pp. 1–8 35 Sieberg A. 1940. Die türkische Erdbebenkatas-

Annu. Rev. Earth Planet. Sci. 2005.33:37-112. Downloaded from www.annualreviews.org S¸engör AMC, Natal’in BA. 1996. Palaeotec- trophe unter geodynamischen Gesichtspunk- tonics of Asia: Fragments of a synthesis. In ten. Umschau 44:49–52 The Tectonic Evolution of Asia, Rubey Collo- Sipahio˘glu S. 1984. Kuzey Anadolu Fayının quium, ed. A Yin, M Harrison, pp. 486–640. Yapısı Abant-Dokurcun arasındaki yerle¸simi Cambridge: Cambridge Univ. Press ve fizyografik-jeomorfolojikozellikleri. ¨ S¸engör AMC, Ozeren¨ S, Gen¸cT,Zor E. Kuzey Anadolu Fay Zonunun Izmit-˙ 2003. East Anatolian high plateau as a Kastamonu Kesiminde Depremleri Onceden¨ mantle-supported, north-south shortened do- Belirleme Projesi Neotektonik Ç alıs¸ma mal structure. Geophys. Res. Lett. 30: Grubu, Istanbul˙ Univ.,¨ Muhendislik¨ Fak., 8045 Jeofizik Muhendisli¨ gi˘ Bol¨ um¨ u,¨ Istanbul.˙ 50 S¸engör AMC, Satır M, Akkök R. 1984. Tim- pp.+1 foldout map (Open File Rep.) ing of tectonic events in the Menderes Mas- Smith AD, Taymaz T, Oktay F, Yüce H, 18 Mar 2005 19:32 AR AR233-EA33-02.tex AR233-EA33-02.sgm LaTeX2e(2002/01/18) P1: IKH

110 S¸ENGOR¨ ET AL.

Alpar B, et al. 1995. High-resolution seis- terdam: Elsevier. Tectonophysics 29:401– mic profiling in the Sea of Marmara (north- 9 west Turkey): Late Quaternary sedimenta- Tatar O. 1996a. Determination of extension tion and sea-level changes. Geol. Soc. Am. directions in naturally deformed limestones Bull. 107:923–36 within strike-slip fault zones using mi- Soysal H, Sipahio˘glu S, Kol¸cak D, Altınok Y. crofractures: Niksar basin (Turkey) as a case 1981. Türkiye ve Çevresinin Tarihsel De- study. Turk. J. Earth Sci. 5:45–53 prem Katalo˘gu (M. O.¨ 2100-M.S. 1900). Tatar O. 1996b. Neotectonic structures indi- Turk.¨ Bilim. Tek. Aras¸tırma Kurumu, Mat.- cating extensional and contractional strain Fiz. Biyol. Bilim. Aras¸tırma Grubu, Istanbul.˙ within Pliocene deposits near the NW margin 87 pp. of the Niksar pull-apart basin, Turkey. Turk. Spindler I, Andrusov N, Ostroumov A. 1896. J. Earth Sci. 5:81–90 Marmornoe More. Ekspeditsia Imperatorsk- Tatar O, Kavak KS¸, Ozden¨ S, eds. 2000. Aktif ago Russkago Geograficheskago Obshestva Tektonik Ara¸stirma Grubu 3. Toplantısı, v 1894 godu. Zapiski Ekspeditsia Impera- Cumhuriyet Univ.¨ 4–5 Kasim 1999, Sivas, torskago Russkago Geogr. Obshestva Ob- ATAG-3, Makaleler, Sivas, [II]+120 shei Geogr.,Vol. 33, No. 2, VII+180 pp.+V pp. foldout plates Tatar O, Piper JDA, Graham PR, Gürsoy H. Stchepinsky V. 1940. Rapport sur la Mission 1995. Palaeomagnetic study of block rota- Sismique d’Erzincan-Gümü¸shane-Sivas. tions in the Niksar overlap region of the Maden Tetk. Arama Enst.,12pp.+1 map North Anatolian Fault Zone, central Turkey. (Open File Rep.) Tectonophysics 244:251–66 Stchepinsky V, Fırat H, Ulusan C, Pamir HN. Tatar Y. 1978. Kuzey Anadolu Fay Zo- 1940. Erzincan-Kelkit Zelzele Mıntakasında nunun Erzincan-Refahiye arasindaki bölümü Yapılan Tetkiklere Ait Rapor. Maden Tetk. uzerinde¨ tektonik incelemeler. Yerbilimleri Arama Enst. 23 pp.+1 map (Open File Rep. 4:201–36 1097) Taymaz T, ed. 2001. Symposia on Seismotecton- Stein RS, Barka AA, Dieterich JH. 1997. Pro- ics of the North–Western Anatolia—Aegean gressive failure on the North Anatolian fault and Recent Turkish Earthquakes: Scientific since 1939 by earthquake stress triggering. Activities 2001. IT˙ U-Fac.¨ Mines, Istanbul˙ Geophys. J. Int. 128:594–604 Tech. Univ., Istanbul,˙ XXIII+113 pp. Suess E. 1901. Das Antlitz der Erde v. III/1: Taymaz T, Jackson J, McKenzie D. 1991. Ac- Tempsky, Wien, IV+508 pp.+1 foldout plate tive tectonics of the north and central Aegean Suzanne P, Lyberis N, Chorowicz J, Nurlu M, Sea. Geophys. J. Int. 106:433–90 by California Institute of Technology on 01/09/13. For personal use only. Yürür T, Kasapo˘glu E. 1990. La géometrie Taymaz T, Tan O. 2001. Source parameters de faille nord-Anatoliennea ` partir d’images of June 6, 2000 Orta-Ç ankırı and Decem-

Annu. Rev. Earth Planet. Sci. 2005.33:37-112. Downloaded from www.annualreviews.org Landsat-MSS. Bull. Soc. Geol.´ Fr. Ser.´ 8,6: ber 15, 2000 Sultanda˘gı-Ak¸sehir earthquakes 589–99 (Mw = 6.0) obtained from inversion of tele- Tapırdamaz MC, ve YaltırakC. 1996. Trakya’da seismic body wave forms. See Taymaz 2001, Senozoyik Volkaniklerinin Paleomanyetik pp. 96–107 Ozellikleri¨ ve B¨olgenin Tektonik Evrimi. Tchalenko JS. 1970. Similarities between shear Maden Tetk. Arama Enst. Derg., No. 119, zones of different magnitudes. Geol. Soc. pp. 27–43 Am. Bull. 81:1625–40 Tatar Y. 1975. Tectonic structures along the Tillotson E. 1940. The earthquake in Turkey. North Anatolian Fault Zone, northeast of Re- Nature 145:13–15 fahiye (Erzincan). In Recent Crustal Move- Tokay M. 1973. Kuzey Anadolu Fay Zonunun ments, Developments in Geotectonics, ed. Gerede ile Ilgaz arasındaki kısmında jeolojik NPavoni, R Green, pp. 401–9. Ams- g¨ozlemler. In Kuzey Anadolu Fayi ve Deprem 18 Mar 2005 19:32 AR AR233-EA33-02.tex AR233-EA33-02.sgm LaTeX2e(2002/01/18) P1: IKH

THE NORTH ANATOLIAN FAULT 111

Kus¸agı˘ Simp., 29–31 Mart 1972, Ankara, the Malatya-Ovacık fault zone. Geodin. Acta pp. 12–29. Tebli˘gler: M. T. A. Enst., Ankara 14:103–31 Toksöz MN, Guest ed. 2002. Dedicated issue. Wong HK, Lüdmann T, Ulu˘gA,Görür N. 1995. The Izmit,˙ Turkey, earthquake of 17 August The Sea of Marmara: a plate boundary sea 1999. Bull. Seismol. Soc. Am. 92:1. iv+526 in an escape tectonic regime. Tectonophysics pp. 244:231–50 Toprak V. 1988. Neotectonic characteristics of Yalcın ¸ T. 1997. Hydrogeological investigation the North Anatolian Fault Zone between of the Gönen and Ek¸sidere thermal waters Koyulhisar and Su¸sehri (NE Turkey). METU (northwest Turkey). See Schindler & Pfister J. Pure Appl. Sci. 21:155–66 1997, pp. 275–300 Türkecan A, Din¸cel A, Hep¸sen N, Papak I,˙ Yalcınlar ¸ I.˙ 1946. Manyas Havzasının Mor- Akba¸sB,etal. 1991. Bolu-Ç ankırı (Köro˘glu folojik Etüdü. Istanbul˙ Univ.¨ Edeb. Fak. Da˘gları) arasındaki Neojen ya¸slı volkanit- Cograf.˘ Enst. Yayın., No. 9, 110 pp.+1 lerin stratigrafisi ve petrolojisi. Turk.¨ Jeol. errata page+13 unnumbered photographic Kurultayı Bul.¨ 6:85–103 plates Tutkun SZ, Inan˙ S. 1982. Niksar-Erbaa (Tokat) Yaltırak C. 1995. Gaziköy-Mürefte (Tekirda˘g) yöresinin jeolojisi. Karadeniz Univ.¨ Bilim. arasının sedimanter ve tektonikozellikleri. ¨ Derg. 2:51–67 Turk.¨ Petrol Jeol. Dern. Bul.¨ 6:93–112 Tüysüz O. 1985. Kargı Masifi ve Dolayındaki Yaltırak C. 1996. Ganos Fay Sistemi’nin tek- Tektonik Birliklerin Ayırdı ve Ara¸stırılması tonik tarihi. Turk.¨ Petrol Jeol. Dern. Bul.¨ (Petrolojik Inceleme).˙ TUB¨ ITAK˙ TBAG 8:137–56 521 Proje Raporu, 431 pp. (Open File Yaltırak C. 2002. Tectonic evolution of the Mar- Rep.) mara Sea and its surroundings. Mar. Geol. Tüysüz O. 1990. Tectonic evolution of a part 190:493–529 of the Tethyside orogenic collage: the Kargı Yaltırak C, Alpar B. 2002a. Kinematics and Massif, northern Turkey. Tectonics 9:141– evolution of the northern branch of the North 60 Anatolian Fault (Ganos Fault) between the Tüysüz O. 1993. Karadeniz’den Orta Anado- Sea of Marmara and the Gulf of Saros. Mar. lu’ya bir jeotravers: Kuzey neo-Tetis’in tek- Geol. 190:351–66 tonik evrimi. Turk.¨ Petrol Jeol. Dern. Bul.¨ Yaltırak C, Alpar B. 2002b. Evolution of the 5:1–33 middle strand of the North Anatolian Fault Unay¨ E, De Bruijn H. 1998. Plio-Pleistocene ro- and shallow seismic investigation of the dents and lagomorphs from Anatolia. Meded. southeastern Marmara Sea (Gemlik Bay). Ned. Inst. Toegepaste Geowetenschappen Mar. Geol. 190:307–27 by California Institute of Technology on 01/09/13. For personal use only. TNO 60:431–66 Yaltırak C, Sakın¸cM,Aksu AE, Hiscott RN, Unay¨ E, Emre O,¨ Erkal T, Ke¸cer M. 2001. The Galleb B, Ulgen¨ UB. 2002. Late Pleistocene

Annu. Rev. Earth Planet. Sci. 2005.33:37-112. Downloaded from www.annualreviews.org Rodent fauna from the Adapazarı pull-apart uplift history along the southwestern Mar- basin (NW Anatolia): its bearing on the age mara Sea determined from raised coastal de- of the North Anatolian Fault. Geodin. Acta posits and global sea-level variations. Mar. 14:169–75 Geol. 190:283–305 Weertman J. 1980. Unstable slippage across a Yaltırak C, Sakın¸cM,Oktay FY. 2000. fault that separates elastic media of different Westward propagation of North Anatolian elastic constants. J. Geophys. Res. 85:1455– Fault into the northern Aegean: Timing 61 and kinematics. Comment. Geology 28:187– Wegener A. 1915. Die Entstehung der Konti- 89 nente und Ozeane. Braunschweig: Friedrich Yılmaz Y. 1990. Allochthonous terranes in Vieweg&Sohn. V+94 pp. the Tethyan Middle East: Anatolia and Westaway R, Arger J. 2001. Kinematics of the surrounding regions. In Allochthonous 18 Mar 2005 19:32 AR AR233-EA33-02.tex AR233-EA33-02.sgm LaTeX2e(2002/01/18) P1: IKH

112 S¸ENGOR¨ ET AL.

Terranes, ed. JF Dewey, IG Gass, GB Curry, Grubu Ba¸skanlı˘gı, Ankara, IV+[I]+459 pp. NBW Harris, AMC S¸engör, pp. 155–68. Unpublished rep. London: R. Soc. Yılmaz Y, Karacık Z. 2001. Geology of the Yılmaz Y, Gen¸cS¸C,Gürer F, Bozcu M, Yılmaz northern side of the Gulf of Edremit and its K, et al. 2002. When did the western Ana- tectonic significance for the development of tolian grabens begin to develop? In Tecton- the Aegean grabens. Geodin. Acta 14:31–43 ics and Magmatism in Turkey and Surround- Yılmaz Y, Tüysüz O, Yi˘gitba¸sE,Gen¸cS¸C, ing Area, ed. E Bozkurt, JA Winchester, JDA S¸engör AMC. 1997b. Geology and tectonic Piper, pp. 353–84. Geol. Soc. London, Spec. evolution of the Pontides. In Regional and Publ. No. 173 Petroleum Geology of the Black Sea and Sur- Yılmaz Y, Gen¸cS¸C, Gürer OF,¨ Elmas A, rounding Region, Am. Assoc. Pet. Geol. Mem. Karacık Z, et al. 1997a. Ayvalık-Dikili- 68, ed. AG Robinson, pp. 183–266. Tulsa, Ç andarlı-Bergama ara¸sında (Edremit-Ber- OK: Am. Assoc. Pet. Geol. gama grabenlericevresinde) ¸ gen¸c magma- Youd TL, Bardet J-P, Bray JD, Tech. eds. 2000. tizmanin jeolojik ve petrolojik ara¸stırıl- Kocaeli, Turkey, earthquake of August 17, ması: TUB¨ ITAK˙ Projesi, Proj. No. YDAB 1999. Earthq. Spectra, Suppl. A to Vol. 17, GAÇ -228/G, 420/G, 74 pp. (Open File pp. xiv+461 Rep.) Yücemen S, Coordinator. 1992. Tübitak ekip- Yılmaz Y, Gürpinar O, Yi˘gitba¸sE,Yıldırım lerinin Erzincan deprem bölgesicalı¸ ¸ smalari M, Gen¸cS¸C,etal. 1993. Tokat Masifi ve (ön inceleme raporu): TUB¨ ITAK-Ankara.˙ 76 Yakıncevresinin ¸ Jeolojisi, T.P.A.O. Arama pp. by California Institute of Technology on 01/09/13. For personal use only. Annu. Rev. Earth Planet. Sci. 2005.33:37-112. Downloaded from www.annualreviews.org HI-RES-EA33-Sengor.qxd 3/29/05 10:38 AM Page 1

THE NORTH ANATOLIAN FAULT C-1 by California Institute of Technology on 01/09/13. For personal use only. Annu. Rev. Earth Planet. Sci. 2005.33:37-112. Downloaded from www.annualreviews.org Digital elevation model of northern Turkey, derived from the GTOPO30 data of the U.S. Geological Survey, showing the North data of the U.S. Geological Survey, derived from the GTOPO30 Turkey, Digital elevation model of northern Figure 1 Figure Anatolian Fault (NAF) and related neotectonic structures forming the North Anatolian Keirogen (NAK). Anatolian Fault (NAF) and related neotectonic structures forming the North HI-RES-EA33-Sengor.qxd 3/29/05 10:38 AM Page 2

C-2 SENGO‹ R ET AL. by California Institute of Technology on 01/09/13. For personal use only. Annu. Rev. Earth Planet. Sci. 2005.33:37-112. Downloaded from www.annualreviews.org HI-RES-EA33-Sengor.qxd 3/29/05 10:38 AM Page 3

THE NORTH ANATOLIAN FAULT C-3

Figure 13 Progressive failure of the North Anatolian Fault (NAF) during the twentieth century earthquake cycle by stress concentration at the tips of failed segments. Red regions are where the stresses are high, representing likely places where the next break will take place. Courtesy of Ross Stein and Serkan Bozkurt. by California Institute of Technology on 01/09/13. For personal use only. Annu. Rev. Earth Planet. Sci. 2005.33:37-112. Downloaded from www.annualreviews.org P1: KUV March 23, 2005 16:53 Annual Reviews AR233-FM

Annual Review of Earth and Planetary Sciences Volume 33, 2005

CONTENTS

THE EARLY HISTORY OF ATMOSPHERIC OXYGEN:HOMAGE TO ROBERT M. GARRELS, D.E. Canﬁeld 1 THE NORTH ANATOLIAN FAULT:ANEW LOOK, A.M.C. S¸engor,¬ Okan Tuys¬ uz¬ , Caner Im˙ ren, Mehmet Sakõnc¸, Haluk Eyidogˇan, Naci Gor¬ ur,¬ Xavier Le Pichon, and Claude Rangin 37 ARE THE ALPS COLLAPSING?, Jane Selverstone 113 EARLY CRUSTAL EVOLUTION OF MARS, Francis Nimmo and Ken Tanaka 133 REPRESENTING MODEL UNCERTAINTY IN WEATHER AND CLIMATE PREDICTION, T.N. Palmer, G.J. Shutts, R. Hagedorn, F.J. Doblas-Reyes, T. Jung, and M. Leutbecher 163 REAL-TIME SEISMOLOGY AND EARTHQUAKE DAMAGE MITIGATION, Hiroo Kanamori 195 LAKES BENEATH THE ICE SHEET:THE OCCURRENCE,ANALYSIS, AND FUTURE EXPLORATION OF LAKE VOSTOK AND OTHER ANTARCTIC SUBGLACIAL LAKES, Martin J. Siegert 215 SUBGLACIAL PROCESSES, Garry K.C. Clarke 247 FEATHERED DINOSAURS, Mark A. Norell and Xing Xu 277 MOLECULAR APPROACHES TO MARINE MICROBIAL ECOLOGY AND THE MARINE NITROGEN CYCLE, Bess B. Ward 301 EARTHQUAKE TRIGGERING BY STATIC,DYNAMIC, AND POSTSEISMIC by California Institute of Technology on 01/09/13. For personal use only. STRESS TRANSFER, Andrew M. Freed 335 EVOLUTION OF THE CONTINENTAL LITHOSPHERE, Norman H. Sleep 369 Annu. Rev. Earth Planet. Sci. 2005.33:37-112. Downloaded from www.annualreviews.org EVOLUTION OF FISH-SHAPED REPTILES (REPTILIA:ICHTHYOPTERYGIA) IN THEIR PHYSICAL ENVIRONMENTS AND CONSTRAINTS, Ryosuke Motani 395 THE EDIACARA BIOTA:NEOPROTEROZOIC ORIGIN OF ANIMALS AND THEIR ECOSYSTEMS, Guy M. Narbonne 421 MATHEMATICAL MODELING OF WHOLE-LANDSCAPE EVOLUTION, Garry Willgoose 443 VOLCANIC SEISMOLOGY, Stephen R. McNutt 461

ix P1: KUV March 23, 2005 16:53 Annual Reviews AR233-FM

x CONTENTS

THE INTERIORS OF GIANT PLANETS:MODELS AND OUTSTANDING QUESTIONS, Tristan Guillot 493 THE Hf-W ISOTOPIC SYSTEM AND THE ORIGIN OF THE EARTH AND MOON, Stein B. Jacobsen 531 PLANETARY SEISMOLOGY, Philippe Lognonne« 571 ATMOSPHERIC MOIST CONVECTION, Bjorn Stevens 605 OROGRAPHIC PRECIPITATION, Gerard H. Roe 645

INDEXES Subject Index 673 Cumulative Index of Contributing Authors, Volumes 23Ð33 693 Cumulative Index of Chapter Titles, Volumes 22Ð33 696

ERRATA An online log of corrections to Annual Review of Earth and Planetary Sciences chapters may be found at http://earth.annualreviews.org by California Institute of Technology on 01/09/13. For personal use only. Annu. Rev. Earth Planet. Sci. 2005.33:37-112. Downloaded from www.annualreviews.org