Downloaded from http://sp.lyellcollection.org/ by guest on October 6, 2021

PROCESSES OF COLLISION Downloaded from http://sp.lyellcollection.org/ by guest on October 6, 2021 Downloaded from http://sp.lyellcollection.org/ by guest on October 6, 2021

Shortening of continental : the of Eastern a young collision zone

J.F. Dewey, M.R. Hempton, W.S.F. Kidd, F. Saroglu & A.M.C. ~eng6r

SUMMARY: We use the of Eastern Anatolia to exemplify many of the different aspects of collision tectonics, namely the formation of plateaux, thrust belts, foreland flexures, widespread foreland/hinterland deformation zones and orogenic collapse/distension zones. Eastern Anatolia is a 2 km high bounded to the S by the southward-verging Bitlis Thrust Zone and to the N by the Pontide/Minor Caucasus Zone. It has developed as the surface expression of a zone of progressively thickening beginning about 12 Ma in the medial and has resulted from the squeezing and shortening of Eastern Anatolia between the Arabian and European Plates following the demise of the last oceanic or quasi- oceanic tract between Arabia and Eurasia. Thickening of the crust to about 52 km has been accompanied by major strike-slip faulting on the rightqateral N Anatolian Transform (NATF) and the left-lateral E Anatolian (EATF) which approximately bound an Anatolian Wedge that is being driven westwards to override the oceanic lithosphere of the Mediterranean along zones from Cephalonia to Crete, and Rhodes to Cyprus. This neotectonic regime began about 12 Ma in Late Serravallian times with uplift from wide- spread littoral/neritic marine conditions to open seasonal wooded savanna with coiluvial, fluvial and limnic environments, and the deposition of the thick Kythrean in the Eastern Mediterranean. hypocentres are scattered throughout the region but large are concentrated mainly on the major faults and are mostly shallow, supporting the idea of a brittle elastic lid with hypocentres concentrated towards its base with more ductile deformation in the middle and lower crust. Neotectonic magmatic suites are nepheline- hypersthene normative alkali of origin, and silicic/intermediate/mafic calc- alkaline suites, both suites occurring in pull-apart basins in strike-slip regimes and along N-S extensional fissures, and both suites showing a strong change to central activity in the . Upper-crustal strains appear to be discontinuous in space and time, with zones of strong shortening representing shoaling of crustal detachment zones flattening between 5 and 10 km. Approximately NW- (dextral) and NE- (sinistral) trending bound less deformed wedges (low relief seismically 'dead' areas) and vary from simple strike-slip faults to com- plicated braided transform-flake boundaries with pull-apart and compressional segments (N and E Anatolian Transform Faults). Volcanoes lie in on N-S 'cracks' that extend into the Arabian Foreland and in transcurrent pull-aparts. Major extensional basins lie at plate () and flake (Karliova) triple junctions and result from compatibility problems.

Apart from impacts and possible but difficult involves the progres- to evaluate sub-lithospheric influences, the sive impingement of buoyant or highstanding geology of the is the con- with subduction zones. All scales and sequence of the flexure, stretching and variations exist on this theme between the shortening of the lithosphere. Rapid stretching collision of seamounts and seamount chains and shortening of the lithosphere produce with arcs through the collision of oceanic isothermal thinning and thickening respec- plateaux and microcontinents with arcs to the tively, with consequent basins and moun- collision of large continental masses. The scale tains. Thermal relaxation generates further of collision dictates the. style, duration and or uplift, enhanced, respectively, intensity of the resulting strain systems and by and . Conse- sequences (Dewey 1977). Colliding continental quently, most vertical motions leading to all margins are irregular and strain sequences are the subtleties and complexities of strati- usually diachronous along great strike lengths graphic development are the result of litho- along zones (Dewey & Burke 1973). spheric deformation. Continental collision is Prior to terminal continental collision, one or one of the principal mechanisms leading to both continental margins may have had a long lithospheric/crustal thickening and and complex history of exotic assembly building and is an appropriate topic for a (Coney et el., 1980). William Smith thematic meeting honouring Continental convergent plate boundaries Robert Shackleton, that prime observer and such as the Alpine/Himalayan System (Fig. 1), interpreter of rocks in continental deformation are wide, diffuse and complicated zones where zones. relative plate displacements are converted into

From COWARD, M. P. & RtES, A. C. (eds), 1986, Collision Tectonics, Geological Society Special Publication No. 19, pp. 3-36. Downloaded from http://sp.lyellcollection.org/ by guest on October 6, 2021

4 J.F. Dewey et al.

<. .=, r-i

r., E~.. ~ E

t-- © © ©:.~

~ ~ ~.~,~ ~-_0 >-

_= 0=h: = >

| • . "~ .~ ..o

| 0



"~ .~ ,-.,.-

-~ ..~ .,-, "-' ~ ~_ ~ ~ ~[.--

~_o..o" " Z o,= o~ E ~: .=. =

, ~ I;;;; ¢~ Downloaded from http://sp.lyellcollection.org/ by guest on October 6, 2021

Neotectonics of E Anatolia 5 complex and variable strains and smaller the thickened crust at an Argand Number block-bounding displacements. This contrasts (ratio of caused by crustal thickness dif- with oceanic plate boundaries, which are ference and stress needed to continue con- generally narrow, relatively simple zones in , England & McKenzie 1982) of 3. which only a small portion of relative plate This model has the particular merit of explain- motion is converted into strain and smaller dis- ing E-W lateral wedging and extension as a placements (McKenzie 1972). This contrast is late stage consequence of crustal thickening. probably due to the relative weakness and The is the higher and larger buoyancy of and the relative strength of the two major plateaux in the Alpine/ and negative buoyancy of olivine as the prin- Himalayan system (Fig. 1), the other being the cipal phases in the and Turkish/Iranian Plateau, a zone of lithospheric oceans respectively. Also, the great inhomo- horizontal shortening about 2 km above sea geneity and anisotropy of the continental crust, level ($eng6r & Kidd 1979). Such plateaux, riddled with zones of low strength, generated with roughly constant mean elevation, are one and modified by many varied mechanisms, of five tectonic components in collisional contrasts with the relative homogeneity of the systems (Figs 1 and 2), namely plateaux, thrust oceanic lithosphere generated by plate accre- belts, foreland lithospheric flexures, wide- tion with. zone modifications (Dewey spread foreland/hinterland deformation zones 1982). and orogenic collapse/distension zones. We A basic problem of collisional tectonics is here define foreland and hinterland to mean how relative plate displacement directions and those regions exterior to the outermost major rates are converted in strains and strain rates overthrust belts in the direction and away from within the convergent plate boundary zone the direction, respectively, of principal oroge- (Fig. 1). Our understanding of this problem, nic vergence. Thrust belts and foreland flexures although incomplete, has progressed greatly are common to all collisional systems, where- since Argand (1924) first explained the as plateaux, widespread foreland/hinterland Himalayan Orogen as a result of simple under- deformation and collapse zones may or may thrusting of beneath , a mechanism not be present in a particular portion of the advanced today by Powell & Conaghan (1973) orogen. Barazangi & Ni (1982), Ni & York (1978) and Thrust belts develop principally where the Ni & Barazangi (1984), among others, to thinned continental crust of a rifted margin is explain the thick Tibetan crust (Chen & progressively restacked and thickened toward Molnar 1981; Molnar & Chen 1983). Two the foreland. This commonly involves thrust further competing models have been suggested rejuvenation of old listric normal faults to explain the Tibetan Plateau (Fig. 1). Molnar (Jackson 1980) and thrust shortening is usually & Tapponnier (1975, 1977a,b, 1978, 1979, initially below before the crust is 1981), Tapponnier & Molnar (1976, 1977) and restacked to 30 km. The oldest, highest, Tapponnier et al (1981, 1982) have advanced a internal, -cored are generally horizontal plane strain slip-line solution to discontinuous along orogenic strike, whereas explain the pattern of strike-slip wedging and younger, exterior foreland thinner-skinned E-W extension in Tibet generated by penin- -thrust belts are continuous and highly sular India behaving as a rigid indenter. cylindroidal (Fig. 3). Where detachment Horizontal plane strain alone, however, cannot occurs along a weak horizon within a foreland explain the thickened Tibetan crust (> 80 km). sequence, rocks above the detachment can An alternative view is that the Tibetan litho- shorten significantly independently of the base- sphere is shortened and thickened by vertical ment for foreland distances of 400 km (Geiser stretching (Dewey & Burke 1973). We do not & Engelder 1983). The higher internal nappes imply a particular mechanism for vertical commonly carry sub-continental mantle and stretching, only that the crust is thickened by complete, but thin, crustal sections (Fig. 4), mechanisms, perhaps subhorizontal shearing crustal thinning dating from the time of crustal and thrusting, operating on a smaller than extension and separation. Tectonic shaving crustal scale. England & McKenzie (1982) thins the highest nappes and the resulting slices have argued for a viscous continuum model (e.g. the Sesia Zone in the ) may be sub- whereby the progressive impingement of the ducted to the facies (Fig. 4). rigid Indian lithosphere has caused vertical Metamorphic patterns in thrust belts generally stretching over a progressively increasing area involve a localized blueschist overprint of older of Asia. In their model, vertical stretching is crustal patterns overprinted in turn by a buffered at about 80 km by lateral spreading of regional amphibolite/greenschist pattern (Fig. Downloaded from http://sp.lyellcollection.org/ by guest on October 6, 2021

© l



0 ...~ . I ,< 0) oc uJ o~ GO o ,..~- E ~.c:: n"

-r- if) -- o u~ o n,, 0 0 uJ n,- ~.~ x n,- .~_ .~© o~ E N ,,~

I-- 'T

E'~ I - .~ ~-

,-~ ,~ wi o..~ o g,

¢J3 Downloaded from http://sp.lyellcollection.org/ by guest on October 6, 2021

Neotectonics of E Anatolia 7




FIG. 3. Block diagram and plan views of the extension of the upper continental crust by listric normal faulting (a) followed by the progressive restacking of the crust (b,c) by thrust faults nucleated on the earlier extensional faults. Arrows indicate extension (a) and convergence (b,c) direction of foreland and hinterland.

4). The innermost highest nappes and adjacent Archaean lithosphere modified, in the ophiolitic or cryptic suture zones are usually Himalayan Thrust Belt, by Triassic rifting. The steepened to overturned in late-stage crustal Swiss Plain Foredeep and associated peripheral scale rotation or r#trocharriage zones (Roeder bulge of the Central//Black 1979) that may involve extensive backthrusting Forest/ are of shorter wave- over adjacent plateaux or hinterlands (Figs 2 length superposed on a Hercynian lithosphere and 4). Crustal low-velocity zones occur in modified by Jurassic extension and Eocene thrust belts (Alps: Rybach et al. 1980) and rifting. plateaux (Tibet: Chen& Molnar 1981) (Figs 2 The widespread foreland deformation of and 4) and are discussed in a later section. NW Europe (Dewey 1982) and hinterland Foreland flexure involves the springboard- deformation of Asia (Molnar & Tapponnier like downbending of the lithosphere by the 1975), in response to Alpine and Himalayan vertical load of the advancing thrust sheets, to collisions respectively (Fig. 1), show that the form a , with a corresponding stresses generated by continental convergence outer arch or peripheral bulge, the wavelength can affect continental portions of plates hun- and amplitude of which depends on the dreds, even thousands, of km from the thrust flexural rigidity and, in turn, thermal age of belt; these stresses may have been enhanced the lithosphere (Karner & Watts 1983) (Figs 2 by push from the Mid-Atlantic Ridge and and 4). The Indo-Gangetic Foredeep of the from the Indian Ocean Ridge. The degree and and the associated peripheral bulge extent to which forelands deform, or whether are long wavelength flexures superposed on the hinterland or foreland takes up part of the Downloaded from http://sp.lyellcollection.org/ by guest on October 6, 2021

J~ E I ÷1, < ~= "~ Z ~=~ 71

8 .~..~ ~ ~

= ~.~ © ~ ~ .~ ~ ~ .

•--- ~ =-~

~.~ ",~ ~,.~


~-~,- o.,=

Z .'",'.'.' I :-:i.i!: ~ -I -

i.:.#!,;~r ~.~.~ j ii Ii iii II ]) I i,j- 1 ii ~'1,1 i ~ I~ll~ i It~l I~11 ]ili'il~ll ( ',,~L~',I. ~.~-~ ~

i- Downloaded from http://sp.lyellcollection.org/ by guest on October 6, 2021

Neotectonics of E Anatolia 9 convergence, appears to depend upon the and Riverside, where the is thermal age and anisotropy of the lithosphere. in a transpressional locking orientation, upper- The preferential hinterland deformation of crustal flakes up to 15 km thick (hypocentres Asia, rather than the Indian Foreland, is pro- shallower than 15 km) have detached from the bably because the Tibetan lithosphere was lower crust to offset the San Andreas Fault in warmed and thinned by pre-collisional / cross-section and to cause southwestward over- Early Tertiary subduction and because the thrusting of the San Gabriel and Asian lithosphere is a complex inhomogeneous Transverse Range. Also, a Mohave 'wedge' is assemblage of Hercynian blocks, arcs and moving eastwards from the San Andreas/ accretionary wedges whereas the Indian litho- Garlock convergence. Rotation of crustal sphere has mainly an Archaean basement age. flakes up to 90 ° (Luyendyk et al. 1980) has Extensional collapse zones occur principally occurred at the edges of which upper-crustal in the Alpine/Mediterranean region (Aegean, basins, such as the Ridge Basin, have opened. Tyrrhenian and Pannonian Basins, Fig. 1). The sense of rotation along the Cordilleran These depressions may have a similar origin to margin of is clockwise (Beck each other. Royden et al. (1983) have invoked 1976), consistent with dextral motion along the subducting slab rollback (Dewey 1980) of the margin between Pacific and North American European Foreland for the extension of the Plates. The structural and palaeomagnetic Pannonian Basin, a back-arc mechanism that evidence appears to suggest that most, if not all, could also explain the Tyrrhenian and Aegean of these flakes have undergone rigid body rota- Basins. tion, i.e. external rotation rather than internal A notable feature of the Alpine/Himalayan viscous rotations. However, a critical question, convergent zone is that, although the con- which we address for Eastern Anatolia in a vergence rate varies from 10 to 50 mm yr -~, later section, is to what extent is strain, at and the width of the deforming zone, as various structural levels in the crust, continu- defined by earthquake distribution, ranges ous or discontinuous in space and time in from a few hundred to several thousand km, continental convergence zones and, where the spatial average convergent strain rate is fault-bounded blocks or flakes can be observed, roughly constant along the belt at about to what extent are they internally rigid with 1.5 x 10 -15 S-1. Ben-Avraham & Nur (1976) strain confined to narrow slip zones at their have suggested that the width of the zone is boundaries. proportional to the convergent displacement It is likely that such flakes are the surface rate, that is the rate at which material is fed expression of thin upper-crustal sheets above into the convergent zone. If the same cross- intracrustal d~collements rather than small sectional area of material was involved in the 'platelets' because the latter would be mechan- zone of shortening, with time the strain rate ically difficult and unlikely narrow lithospheric would rise exponentially. A constant width- 'spindles'. The kinematic theory of strains, shortening zone maintains a constant strain rotations and displacements in complex con- rate. The geological evidence from many tinental convergent plate boundary zones has orogens indicates that material enters conver- been addressed by McKenzie & Jackson (1983) gent strain zones in the shortening direction and will not be considered here. with time because thrusts prograded into The depth frequency of hypocentres hitherto undeformed crust. (Meissner & Strehlau 1982) is a guide to the thickness of an upper brittle or elastic layer in Crustal and detachment which stress accumulation and release by faulting and jointing occurs. Hypocentral fre- Many lines of evidence are gathering, that in quency, magnitude and highest stress drop convergent zones, the upper crust forms a reach maxima (Das & Scholz 1983) at depths high-strength layer from 5 to 15 km thick, that from as little as 5 km in extensional areas of may be thrust for hundreds of km as a thin high heat flow to as much as 20 km in con- plate or a series of stacked flakes. In the vergent areas of low heat flow (Fig. 5). This Southern Appalachians, a thin sheet involving indicates an inverse relationship between heat the Blue Ridge and Inner Piedmont has moved flow and hypocentral depth maxima which, in westwards over the foreland for at least areas, is related to crustal stabilization 300 km (Cook et al. 1979). Similar detach- or lithospheric thermal age (Chen & Molnar ments are deduced in Pakistan from - 1983). In some regions (Chen & Molnar 1983), quake hypocentres (Armbruster et al. 1978). In a seismic gap exists in the lower crust between California, in the region between Bakersfield upper- and middle- crustal hypocentres and Downloaded from http://sp.lyellcollection.org/ by guest on October 6, 2021

10 J.F. Dewey et al.

10 5 -'f'mox kb 5 10

. A, B c ! 10



40 km 50




90 i I i i ~ i i i i i i' l i i 500 T OC 500 Fl~. 5. strength of wet and dry quartz and olivine in extension and compression at a strain rate of 10 ~5 s ], at heat flows for geothermal gradients corresponding to 1 and 2 HFU, modified and extrapolated from Meissner & Strehlau (1982) and Kirby (1980). Dry and wet rock curves refer to the brittle shear strengths (Byerlee 1968). Depth frequency distribution of hypocentres; SW, Schwarzwald; Co, Coso; H, Haicheng (Meissner & Strehlau 1982); P, Pakistan Himalayas (Armbruster et al. 1978); D, continental convergence zones (open circles, mantle hypocentres); A-C, continental interiors (square gives mean, line gives range) with basement ages of 250-800 Ma, 800-1700 Ma and greater than 1700 Ma, respectively (Chen & Molnar 1983). Temperature scale refers to geothermal gradients (GG) with surface heat flows 1 and 2 in HFU. Typical stress-drops of only a few hundred bars indicate that these theoretical strength profiles must be regarded only as maxima, for flawless materials. upper-mantle hypocentres. The maxima is stronger than quartz in the upper greenschist, correlate well (Fig. 5) with strength predictions amphibolite and granulite facies. A great range from extrapolated experimental data. Byerlee of petrological crustal profiles exist (Dewey & (1968) showed that, at low temperatures, the Windley 1981) involving various combinations coefficient of and hence fracture and of hydrous and anhydrous quartz and feldspar sliding strength of most materials is propor- dominated assemblages in a wide range of tional to confining pressure and therefore to metamorphic grades. In Fig. 6, a wet quartz/ depth. Shearing strength for dry and wet rocks dry quartz/dry olivine profile, for a heat flow of in compression and extension are plotted as a 0.75 HFU and a strain rate of 10 -15 s -1, has function of depth in Fig. 5. However, the shear been chosen to roughly match the depth fre- strengths of rocks and at higher quency for earthquake hypocentres (Jackson & temperatures are extremely temperature-, McKenzie 1984) in the Turkish/Iranian Plateau. rather than pressure-, dependent. At any given This gives three strength maxima and minima temperature, the creep strength of quartz, the corresponding to wet and dry quartz as upper principal crustal phase, is very much less than and lower layers in a 50 km crust and dry that of olivine, the principal mantle phase olivine in the . The segmentation (Fig. 5). Also, hydrous mineral assemblages of the upper high-strength layer or elastic lid are weaker than anhydrous. For dry or wet, (Dewey 1982) is the likely origin of upper- quartz- or olivine-dominated mineral assem- crustal flakes within which a wide variety of blages, strength maxima exist at the fracture/ detachment surfaces are possible (Fig. 6). The creep envelope intersection (Fig. 5). The role strength minima are probably zones of ductile of feldspar, a possible stress-bearing phase in strain in which constructive metamorphic some lower crusts, is uncertain, except that fabrics are generated in rather high-pressure/ petrographic textural evidence indicates that it low-temperature metamorphic facies with no Downloaded from http://sp.lyellcollection.org/ by guest on October 6, 2021

Neotectonics of E Anatolia 11

5 10 if'max kb

_2.~ . 30 ~ ..../ ~ ~-_-~ ~ ..... ,~ fissuring ~ ~ / hydrotherm ...... ~'~ •~--~"~ ...... ' veining - - , /convectional" " \ ~- I 4,~ 10 / flest~x. " ~' - '~ L,-:~Jshehrfailure ~ /' / melting-" s"l superplostic 2o :3 2C "3-.., A/ -/ catac.os,,ciio~-~---~.~~~-~,ow ~ ' ~ ~ Z groinsize red uct io n ~'~m~...~..._..-'-- 1 shear heatin~ ~~~~ /,:/ ~-~--L/\/ ///,~-k \ ' t , - ,~"~, - ~ - , - , , 3o " ~//// / :- ~ T'~.', - -. :~ >

70 .~

0.75 _/! I I I I I I I I T°C 500 FIG. 6. Strength profile (enveloping obliquely striped area) for a 50 km continental crust in compression comprising a 27 km thick upper hydrous crust (wq, wet quartz), a lower anhydrous crust (dq, dry quartz), and a dry upper mantle at a strain rate of 10-~5 s ' and for a corresponding to a surface heat flow of 0.75 HFU. Stippled area: depth/frequency distribution of hypocentres from Jackson & McKenzie (1984). A, andalusite; AM, amphibolite; B, blueschist; G, greenschist; K, kyanite; S, sillimanite; wgm, wet melting.

FIG. 7. Schematic illustrations of the strains and displacements in a collisional zone. (a-c) Modes of detachment termination referred to in text. BT, basement thrust; CA, coaxial strain; CDZ, convergent deformation zone; CW, conjugate wrenching; F, flower structure; FF, foreland folding; FP, foreland pull-apart; FR, foreland ; FU, foreland uplift; G, granite; LR, lateral ramp; NCA, non-coaxial strain; P, pull-apart; R, ramp; RF, rigid flake; S, surge zone; T, transform; TP, ; small circles: hypocentre of large earthquakes. Ellipses indicate finite elongation/ shortening directions. In (c) open and filled circles indicate respectively motion away from and towards the observer. Downloaded from http://sp.lyellcollection.org/ by guest on October 6, 2021

12 J.F. Dewey et al. crustal partial melting (Fig. 6). This is con- blocks and accretionary prisms and is charac- sistent with field structural evidence that, terized by great structural complexity. The with increasing metamorphic grade (depth), present (neotectonic) phase of collisional deformation becomes more homogeneous on a convergence began about 12 Ma ago in which smaller scale. At high structural levels, strain is the northwards convergence of Arabia with partitioned and concentrated on narrow high Eurasia produced shortening and thickening of strain rate (10 -12 s -1, Barton & England 1979) the crust and caused a wedge-shaped Anatolian slip zones while, at deeper levels, strain is 'block' to migrate westwards over the subduct- more penetrative. In later sections, we take ing oceanic lithosphere of the Eastern Mediter- the profile of Fig. 6 as the basic model for the ranean. We have been studying Anatolian East Anatolian Convergent Zone. neotectonics for the past 10 and a Elastic lid or flake detachment, especially preliminary outline of our conclusions is where rotations about vertical axes occur, presented below. to complicated compatibility problems at flake margins. This is well seen in Anatolia as Coilisional assembly of Anatolia shown below and is particularly evident in the Transverse Ranges of California where sub- Four major tectonic subdivisions of stantial localized pull-apart basins and thrust were defined by Ketin (1966a). The 'Central zones occur on a restricted geographic scale at Anatolian Massifs' (Menderes Massif, Kir~ehir flake boundaries and triple junctions. Massif) of Central and Western Turkey and the Local termination of basal flake detachments 'Anatolides' are now regarded as Alpine struc- offers a possible explanation for a wide variety tures (Ketin 1966a; Durr et al. 1978) and not of localized crustal structures, which may be Hercynian or older Zwischengebirge (Brink- arranged in various of geometrical combi- mann 1976) within the Anatolian Orogen. nations. These may comprise listric extensional Several workers have argued that the Central detachments passing laterally into zones of Turkish Anatolides and the Southern Turkish lower-crustal stretching with resulting pop-ups Taurides formed a single palaeogeographic (Fig. 7a), with mid-crustal flat bases realm during the entire Mesozoic and Tertiary, (Lynn et al. 1981) passing laterally into thrust characterized mainly by Triassic- zones (Fig. 7b), and rotated flakes with shelf carbonates although, in the internal parts marginal flower structures above detachments (Anatolides), sedimentation ended earlier that relay transform motion laterally into wide (Ricou et al. 1975; Ozgul 1976; Durr et al. lower-crustal and mantle shear zones. Figure 7 1978). The non-metamorphic, mildly deformed summarizes schematically the range of crustal part of this carbonate is exposed structures characterizing collisional orogens beneath large composite nappes especially in and their forelands, most of which occur in the Western Taurus (Brunn et al. 1971; Eastern Anatolia and are described in a later Bernoulli et al. 1974; Delaune-Mayere et al. section. 1977). It has a great along-strike continuity Eastern Anatolia is a well-exposed and westwards into the external units of the accessible present-day collisional convergent Hellenides and eastwards into (Ricou et zone (Fig. 8) forming a plateau averaging some al. 1975). 2 km above sea level. It may be one of the best During the Late , part of the areas in the world to study the geometry and ocean floor separating the Anatolide/Tauride kinematics of continental convergence. The Platform from the Northern Turkish Pontides E Anatolian Convergent Zone (EACZ) is a was obducted onto the former (Ricou et al. region of frequent and widespread earthquakes 1975). The Pontides, connecting the Rhodopian (Fig. 9) ranging up to M = 8 (Ergin et al. 1967; Massif and the Srednogorie province of the Canitez & Ucer 1967b; Dewey 1976; Balkan Ranges (Hsii et al. 1977) with the Buyukasicoglu 1979; Jackson & McKenzie (Adamia et al. 1977) formed, 1984) and measurable deformations and dis- at this time, a S-facing arc S of an placements are visibly occurring throughout opening marginal basin. The the region. The region is an aggregate of arcs, Pontides have Late Cretaceous-Palaeogene

FIG. 8. (Opposite). Neotectonic map of the E Anatolia Convergent Zone (EACZ) and its Arabian Foreland showing (fine lines), mapped faults (thick lines), young basins (dotted), volcanoes and volcanic fields (areas limited by thick dotted lines, thrusts and folds that have been active during the past 12 Ma). Encircled dots and small black circles are the epicentres of very large (M>7) and large (M>6) earthquakes, respectively. Downloaded from http://sp.lyellcollection.org/ by guest on October 6, 2021

0 po

I \ J k ---~ 00

,'+ / +-t;,l~ ' \ ?+~+k,~ .

~m I I ~L • %\---.-.. ,"

~p- rO CX3 pc-) 1~ pc) FIG. 9. Epicentres of E Anatolian earthquakes of M>5 and selected nodal plane solutions from Buyukasicoglu (1979). B, Bingol; E, Erzincan; G, Goynuk; L, Lice; V, Varto. Downloaded from http://sp.lyellcollection.org/ by guest on October 6, 2021

14 J.F. Dewey et al. calc-alkaline (MTA 1962; Tokay existence since the Late Triassic (Dewey et al. 1973; Seymen 1975) and widespread Late 1973; Bein & Gvirtzman 1977). From Early Cretaceous-Palaeocene ophiolitic m61ange Cretaceous to the Middle Miocene, it was also accumulation (Tokay 1973: Gansser 1974; consumed along a N-dipping subduction zone Seymen 1975; Bergougnan 1976)• That both (Dewey et al. 1973; Hall 1976; Dewey & calc-alkaline magmatism and m61ange accu- Seng6r 1979) and was completely obliterated mulation in the Pontides continued long after by a Late Miocene collision along the Bitlis the Cretaceous abduction of the Suture. onto the Anatolide-Tauride Platform, rules In latest Serravallian, earliest Tortonian time, out a Late Cretaceous Pontide-Anatolide about 11.8 Ma ago, a fundamental palaeogeo- collision proposed by Ricou et al. (1975). The graphic, sedimentological and tectonic change uplift of the Pontide-Anatolide Suture occurred throughout Eastern Anatolia (Fig. 10) (Lutetian plant NE of Ankara; Tokay that, we believe, resulted from the final colli- 1973) and the initiation of major imbrication sional closure of the Bitlis oceanic tract and and S-vergent thrusting accompanied by the accompanied the beginning of widespread rapid formation and southerly migration of crustal shortening and the beginning of uplift flysch- troughs in the Anatolide- across Eastern Anatolia. Throughout the Tauride region during the Early Eocene region, Serravallian shallow marine (Delaune-Mayere et al. 1977), are evidence for give way to terrestrial sediments intermittently Eocene Pontide-Anatolide collision at least in and patchily deposited in a wooded seasonal the western section of Anatolia. In the E, the savanna environment. In the southern border collision may not have occurred until the thrust zone, the Tortonian Lice Flysch signified (Seymen 1975). After the collision, the beginning of thrust loading. Tortonian the Anatolide-Tauride realm was further Kythrean Flysch follows the Serravallian de- imbricated and stacked and high T/P meta- position of Globerigina marls in the Eastern morphism and anatexic granite plutonism Mediterranean, witnessing the emergence of a affected the internal Anatolides (Durr et al. nearby source area. The Central Anatolian ova a978). regime, characterized by large, rather equant, In SE Anatolia, the Bitlis Ocean had been in fault-bounded depressions (ovas) with Neogene

(a) (b) [ (c)

PLEISTOCENE " strong ~ 0.9 Lu a E sand -- --1.8 - delta . uplift progradation o colluvial -dE 3.2 PLIOCENE ~- alluvial g c fluvial :~ c~ .2 limnic E ~" catastrophic ~co'6 .£: . ..,,o,L ...... ,~ 5.1. ' floodingoo ~ ~gravel Misis evaponte~ % Thrust <~ >c~ "_."" - .--- ~u~•':-'' s~I~;D..:"''. . .:6.2 # z ~z --6.8 a conglorn ~. molas.sf ( =° ._c~ uJ o ;~-d breccia -9° o g sand ~ ° c ~: :Kythrean :_ ? ~ marl -- "2 ,° I < rr TORTONIAN .Flysch.-~ "~ ~¢ lignite oJc~ r i11 -a evapor. - Lice .... [3_ n •-- -i .flysch.-" ~o

~..-o " • ".,"_ ' ...... , ~? 10.C~ J

11.8 ...... • ' "2 LJ I Z neritic Globigerina littoral Midyat c~ SERRAVALL marls 9 and clays lagoonal carbonate and marl

FIG. 10. Summary of the medial Miocene to present-day events in the Eastern Mediterranean region. (a) Eastern Mediterranean; (b) Eastern Anatolia (sediments and tectonics); and (c) Eastern Anatolia (volcanism). Stratigraphic divisions and ages from Rogl & Steininger (1983)• Volcanic data mainly from Innocenti et al. (1976). Downloaded from http://sp.lyellcollection.org/ by guest on October 6, 2021

Neotectonics of E Anatolia 15 terrestrial sediments and, locally, young basaltic made throughout to Figs 8 and 9. We have volcanics began during the Late Miocene. much more data for the NATF than for the Thus, by Late Miocene times, the composite EATF, reflected in the much longer ensuing Anatolian Orogen had already been assembled section on the former. more or less in its present configuration. Since Late Miocene, there has been continuous con- tinental lithosphere between the converging North Anatolian Transform Fault (NATF) Arabian and Eurasian in Eastern The NATF is one of many large strike-slip Anatolia, whereas, in Western and Central faults, striking at low angles to the general Anatolia, the Anatolian Wedge has been facing trend of the Alpine-Himalayan system of the underthrusting oceanic lithosphere of the Eurasia and forming late in the orogenic . history of the segments they cut (e.g. Insubric Line of the Western Alps (Gansser 1968; Displacement and strain in Eastern Trumpy 1973; Laubscher 1971); Pustertal Line of the ; Kraistide-Vardar Linea- Anatolia ment of the Balkan-Hellenide Chain In a detailed study of earthquake epicentres (Laubscher 1971; Boncev 1974); Zagros Fault and solutions, McKenzie of SE Iran (Berberian 1976); Karakorum Fault (1972a,b) postulated that the convergence of of the W Himalaya-Karakorum (Molnar & Arabia northward into Eurasia is forcing the Tapponnier 1975)). These faults appear to be wedge-shaped Anatolian Block westward. This closely related to the disintegration of colliding movement is accommodated along the right- promontories and change in direction of rela- lateral N Anatolian Transform Fault (NATF) tive motion along the suture after collision. and its complement, the left lateral E Anatolian The seismically active NATF is a right- Transform Fault (EATF). To the E, lies the lateral fault zone taking up the relative motion E Anatolian Convergent Zone (EACZ). To between the Black Sea and the Anatolian avoid excessive thickening by shortening, the 'Block' and is sub-parallel to the Black Sea EACZ 'wedged-out' a considerable piece of coast of Anatolia running some 1200 km from the Anatolian Orogen along two new bound- Karliova in the E to the Gulf of Saros in the aries, the N and E Anatolian Transform Faults, W, thereby connecting the EACZ (McKenzie towards the more readily subductable oceanic 1972a,b; ~eng6r 1977; ~;eng6r & Kidd 1979) region of the Eastern Mediterranean (McKenzie with the western end of the 1972a,b; Dewey & Seng6r 1979), thereby giving through the complex plate boundary zone of birth to the Anatolian 'Block' or 'Wedge'. the Aegean (Dewey & ~;eng6r 1979). This Epicentre maps by Ergin et al. (1967) and definition is that of Allen (1969) and McKenzie Buyukasikoglu (1979), studies of - (1972a,b) and is not equivalent to the N - sequences of large Anatolian earthquake fault (Kuzey Anadolu earthquakes (Dewey 1976), airphoto and Deprem Fayi) of Ketin (1957) and the N Landsat interpretation (Allen 1969, 1975; Anatolian strike-slip fault (Nordanatolische McKenzie 1976), field studies (Wallace 1968; Horizontalverschiebung) of Pavoni (1961a) Arpat & Saroglu 1972) and regional geologic and Ketin (1969, 1976), who believed it to synthesis (Dewey & ~;eng6r 1979) support this continue into Iran. interpretation. suggests that The NATF, Nowack's (1928) Paphlagonis- suturing between the Arabian Platform and che Narbe, Salomon-Calvi's (1936a, 1940) the Anatolide/Tauride Platform within the Fortsetzung der Tonale-Linie and Pamir's Bitlis Suture Zone was completed in the Late (1950) Cicatrice Nord-Anatolienne, was be- Miocene (Dewey & ~;eng6r 1979). lieved to be the vertex of the Alpine Orogen The style and geometry of strain and dis- in Anatolia. Salomon-Calvi (1936a, 1940) placement in the EACZ is summarized in viewed it in the context of Wegener's theory Figs 8 and 9. Figure 8 shows the distribution of of and, following Argand neotectonic faults and folds within the EACZ (1924), regarded it as the suture between compiled mostly from field studies, interpreta- collided Gondwanan and Eurasian elements. tion of Landsat images and airphotos and from A series of disastrous earthquakes along the published maps. Figure 9 is a compilation and NATF, beginning with the Tercan earthquake replotting of earthquake epicentres and focal on 21 November 1939 and the Erzincan catas- mechanism solutions. In this section, we trophe on 28 December 1939, began a phase of describe the structure and history of the NATF, detailed observations (eg. Pamir & Ketin 1940, EATF and EACZ and reference should be 1941; Parejas et al. 1942; Pamir & Akyol 1943; Downloaded from http://sp.lyellcollection.org/ by guest on October 6, 2021

16 J.F. Dewey et al.

Blumenthal 1945a,b). Ketin (1948) concluded quakes E of Karliova have thrust components that the earthquakes occurred on an active in contrast to the pure strike-slip earthquakes right-lateral, strike-slip fault zone that along the NATF (McKenzie 1972a,b). From extended along the entire length of the Black Karliova, the NATF is continuous to Erzincan, Sea mountains of N Turkey and showed that where it jumps for about 10 km to the N across this structure post-dated the main orogenic the extensional Erzincan Plain, a typical pull- structure of Turkey. Ketin (1948) interpreted it apart basin (Crowell 1974), characterized by as a young feature along which an Anatolian young sediments and small basaltic volcanoes Block, S of the fault, is moving westwards with (Ketin 1976). From Erzincan to Resadiye, the respect to the Black Sea and argued that a trace of the fault zone is again continuous; the complementary left-lateral, strike-slip fault characteristic morphological features of this must bound the Anatolian Block to the S, segment are elongate sag ponds, springs some- citing the Kozan earthquake in SE Turkey as times associated with travertines, fault scarps supporting evidence. Ketin & Roesli (1953) cutting the alluvium in the floor (related compared the NATF with the San Andreas to the 1939 Erzincan earthquake) and deformed Fault of California. stream valleys (Seymen 1975). To the W of In several recent plate tectonic interpreta- Erzincan, recent within Pliocene tions of the Eastern Mediterranean, the NATF sediments have strong morphological ex- plays an important role (McKenzie 1972a,b; pression (Tatar 1975). Between Resadiye and Tapponnier 1977; Dewey & ~;eng6r 1979). The Erbaa, the continuity of the fault is again lost NATF and analogous strike-slip faults in and, whereas the trace coming from Erzincan collision environments are key elements of turns into an E-W orientation S of Amasya, a recent models of continental collision pro- new trace begins to the N of Resadiye. cesses (Molnar & Tapponnier 1975; ~;eng6r Between the southern and northern branches, 1976; Dewey 1977). Thus, the detailed tectonic a third branch appears to be a secondary evolution of the N Anatolian Transform is of extensional feature within a broad pull-apart great importance, especially its age, slip rate basin similar to the Erzincan Plain. Seymen and cumulative offset, to be able to test and (1975) has mapped this region in detail and improve such models. shown that this extensional feature is also the Like most of the large, active, strike-slip locus of Recent basaltic volcanism. He showed faults of the circum-Pacific region (Allen 1965) that, on both sides of the Kelkit Valley, which or Central Asia (Molnar & Tapponnier 1975), here follows the main southern branch of the the NATF has an extremely well-developed N Anatolian Transform Fault, the surface expression for most of its length. It is bounding the auxiliary stream valleys are bent defined by a sharp ~rift morphology' delineat- in a clockwise (dextral) sense. ing a broad fault zone up to 1 km wide The Niksar Basin is a Pliocene-Quaternary composed of numerous sub-parallel and located along the NATF anastomosing faults, offset, captured and NE of Ankara. The main trace of the NATF is dammed streams, sag ponds and elongate offset approximately 10 km across this pull- island-like hills within major valleys following apart basin. East of the basin, the fault occurs the course of the fault zone. Earthquakes have as a single, well-defined trace. This breaks into caused sizable ; formed by a complicated series of poorly defined faults damming, are common along the in the vicinity of the Niksar and adjacent course of the fault (Pamir & Ketin 1941). Erbaa Basins. Both the NE and SW walls of Morphologically, the NATF can be followed the Niksar Basin are characterized by a series as a fairly continuous, strike-slip fault zone of broad terraces which are fairly continuous from Karliova to about Mudurnu. East of the laterally and which generally dip gently (up to point where it joins the EATF, about 10 km E 20 °) on the SW wall towards the interior of the of Karliova, it is lost in the block fault and basin. The NW margin of the basin has an en thrust terrane of the EACZ. Although several ~chelon set of E-W steeply dipping to vertical earthquakes E of Karliova (e.g. at Varto: faults approximately parallel with the main Ambraseys & Zatopek (1968), and Caldiran: to the E. Motion on these faults Arpat & Iz (1977); Toksoz et al. (1977)) appears to have had both normal (S side down produced right-lateral, surface breaks, they as evidenced by the offset of an surface lack the continuity and uniformity of the between Early and Late Cretaceous carbonates) NATF breaks and resemble the irregular and and right-lateral, strike-slip components. discontinuous strike-slip faults of NW Iran Structures along the SW margin of the basin (Berberian 1976). Also, many of the earth- are similar, except that the sense of vertical Downloaded from http://sp.lyellcollection.org/ by guest on October 6, 2021

Neotectonics of E Anatolia 17 offset on faults is N side down. The apparent offset of the Pontide-Anatolide Generally, the NATF forms a wide belt of Suture between Amasya and Erzincan is about numerous, sometimes parallel, sometimes 85 km and Bergougnan's (1976) mapping anastomosing, strike-slip faults. Canitez (1962) revealed a similar offset. Seymen (1975) argues has shown, on the basis of seismic and gravity that, because the dip of the offset suture near observations, that the crust beneath the fault the fault zone is sufficiently steep and the ver- zone is thinner than normal. Within the fault tical motions along the fault zone are small zone, the local are usually exten- compared with the horizontal component of sively crushed and mixed; the low resistance of movement, the apparent offset closely approxi- these fault rocks to subaerial erosion seems to mates the real offset. Several other lines of be largely responsible for the 'rift morphology' indirect evidence suggest an offset of the order along the trace of the transform. This rift of 80-100 km and an average rate of motion morphology extends from Karliova to Mudurnu along the fault of 1-2 cm yr -I (Canitez 1973; with only two minor interruptions by the pull- Arpat & Saroglu 1972). Tokay (1973) and apart basins of Erzincan and Resadiye and Tatar (1975), provided several constraints on finally merges with the and the possible minimum and/or maximum regime of W Anatolia, W of Mudurnu. amounts of offset that bracket the cumulative Attempts to estimate the age, throw and offset between 50 and 100 km. Although a offset of the NATF have been mostly specula- Quaternary rate of 9 mm yr -~ along the tive. Ketin (1948) remarked that the feature is western E-W, segment of the NATF is close young, but did not propose a time of initiation. to the 8.9 mm yr -1 average rate deduced for Pavoni (1961a) thought that the fault may have the eastern NW-SE segment, the total post- originated in the Early Tertiary and estimated Tortonian slip on the western segment is only its offset to be of the order of 350-400 km. 25 km (Barka & Hancock 1984) compared Large amounts of field data have accumulated with a total post-Serravallian 85 km on the during the last 20 years to show that neither eastern segment. Thus, some of the remaining the age nor the offset along the fault is as great 55 km of right-lateral motion can be accounted as Pavoni initially believed. Erinc (1973) for by displacement that occurred before the pointed out that the original drainage network latest Tortonian while some may have been around the NATF was established during the taken up on other faultsl Possibly either or Late Miocene, since when it has been modified both the Kure (Bergougnan et al. 1978) and by activity on the transform and a new drain- Sungurlu Faults took up the motion, or the age system, in places following the trace of the Kure Fault was a precursor to the western crushed zone for considerable distances, has segment of the NATF, or the latter was a left- been formed. Ketin (1976) pointed out that, lateral structure during pre-Pleistocene times within the rift zone, no sediments older than (Hancock & Barka 1981) and formed the south- Middle Miocene have been found, indicating ern boundary to a W-moving wedge similar to, that at least the morphological expression of but smaller than, the present Anatolian Block the fault did not exist prior to this time. Just W (Hempton 1982). of Erzincan, Tatar (1975) mapped inactive The large scale tectonics of Anatolia and branches of the NATF now covered by Pliocene surrounding regions does not support the idea sediments and concluded that the fault orig- of a large (more than a couple of 100 km) inated in pre-Pliocene times. Therefore, the offset along the N Anatolian Transform Fault. geomorphological data constrain the origin of Anatolian tectonic zones are traceable into the fault to between late Early Miocene and the Hellenides (Bernoulli et al. 1974) and Pliocene. there is neither enough deformation in, nor By detailed mapping around Resadiye, enough room between, Anatolia and Greece Seymen (1975) showed that the ophiolitic suture to accommodate an offset more than 300 km between the Pontide and Anatolide regimes is along the NATF. On the other hand, a smaller cut and offset by the transform. In this area, offset, in the order of 80-100 km, would be the major overthrusting of the Pontides onto compatible with the known geology and post- the Anatolides took place during the Burdiga- tectonics of the Aegean area (Dewey lian and Seymen (1975) interpreted this as the & Seng6r 1979). An even smaller offset of manifestation of the terminal suturing between about 15 km along the EATF (Arpat & Saroglu the two tectonic provinces, therefore arguing 1972) supports this view. Further detailed geo- that the NATF must be of post-Burdigalian logical mapping of, and synthesis of data on, age in the segment between Amasya and the NATF is needed to establish the cumulative Erzincan. offset of the fault along various segments, to Downloaded from http://sp.lyellcollection.org/ by guest on October 6, 2021

18 J.F. Dewey et al. determine whether the offset changes in any earthquakes on the NATF have been investi- way along the fault and whether the fault is gated by Chinnery (1969) and Hanks & Wyss composed of different segments that formed at (1972). Canitez & Ezen (1973) concluded that different times and in response to different the stress drop on the NATF is between 10 and strain systems, as maintained by Ricou et al. 15 bar for M > 7 and does not depend on (1975). magnitude as expressed by Aki (1972). For M The NATF seems to have episodes of seis- < 7, however, they found the stress drop to be mic unrest, separated in time by quiescent less than 10 bar. periods of about 150 yr (Ambraseys 1970). The Fault plane solutions of earthquakes along most recent phase began with the 1939 Erzincan the NATF have been presented mainly by earthquake. A series of large earthquakes Canitez & Ucer (1967a,b), McKenzie (1972b), followed and outlined a general migration of Dewey (1976) and Jackson & McKenzie (1983). seismic activity from E to W along the fault, a These solutions give consistently dextral slip peculiar characteristic, first noted by Ketin with minor thrust components between about (1948). During this last cycle of seismic activity, Eskipazar and Karliova. Because all major the style of seismicity of the NATF has been shocks produced surface breaks, there is no similar to the behaviour of the San Jacinto nodal plane ambiguity. East of Karliova, strike- segment of the San Andreas Fault in California, slip movement continues on a line on strike characterized by frequent shocks with magni- with the NATF (Ambraseys & Zatopek 1968; tude 6 < M < 7 (Scholz 1977). Scholz (1977) Ketin 1969), but the fault plane solutions here argued that this type of behaviour characterizes indicate an increased amount of thrust com- those segments of large strike-slip faults that ponent (Mckenzie 1972b), consistent with the strike parallel with the regional slip vector left-lateral EATF (Arpat & Saroglu 1972; between two plates, resulting in low normal Seymen & Aydin 1972; McKenzie 1976) joining stresses across the fault plane. This is approxi- the NATF thus imposing a thrust component mately the case for the NATF, particularly for onto the segments E of Karliova. Strike-slip its well-defined, nearly pure strike-slip seg- faults E of Karliova are elements of the con- ment between Mudurnu and Karliova, if one vergent regime of the Turkish-Iranian Plateau uses McKenzie's (1972b) Anatolia/Black Sea (~}eng6r & Kidd 1979) and not continuations of pole of rotation located at lat. 18.8°N and the NATF. Strike-slip motion on these faults is long. 35°E. due to their oblique orientation with respect to Using the earthquake data for 1960-71, the Arabia/Eurasia convergence. Alptekin (1978) computed the magnitude To supplement observations in the fields of frequency relations for the NATF, and other and , high-precision seismic provinces in Turkey. He found b triangulation and trilateration measurements values of 0.73, 0.56 and 0.66 for western, were started by the MTA in 1972, in the central and eastern sections of the NATF, western sector of the NATF (Gerede-Cerkes respectively. Alptekin (1978) considered the region). Comparison of the 1946 and 1972 relatively low b-values found for the NATF as measurements showed a 75 cm horizontal dis- an indication of high strain accumulation, par- placement along the eastern end of the 1944 ticularly in the central section of the fault. Gerede-Bolu earthquake fault. The total Seismic risk estimates based on magnitude relative displacement in the western portion, frequency relations obtained by Alptekin (1978) however, was 20 cm for the same period (Ugur appear to be highest (68% for M > 8.0, and 1974). Ambraseys (1970) reported an average for a time period of 100 yr) for the central displacement of 90 cm for the entire fault since section of the fault. In contrast with the NATF, 1939. Alptekin (1978) obtained a much higher b value (0.87) for the EATF. East Anatolian Transform Fault (EATF) Using fault lengths and average dislocations observed in the field, Canitez & Ezen (1973) The seismically active and morphologically derived a total seismic moment for the period distinct EATF extends for 400 km from 1900-71 of 1.77 × 102s dyne cm. Using this Karliova in the E to Maras in the W and marks value and following Brune (1968), they cal- the southeastern left-lateral, strike-slip bound- culated the average slip rate for the period of ary between the Anatolian 'Block' and the interest for different fault depth assumptions. Syrian Foreland (McKenzie 1976). A Middle They found, for instance, 2.4 cm yr -1 for d = Miocene marker horizon near Golbasi and an 20 km, 1.6 cm yr -1 for d = 30 km, and 1.2 cm between Miocene and crystalline yr -1 for d = 40 km. Stress drops for some rocks near Goynuk, are offset in a sinistral Downloaded from http://sp.lyellcollection.org/ by guest on October 6, 2021

Neotectonics of E Anatolia 19 sense for 18 and 22 km respectively, and the deformed by transcurrent cataclasis. Barka & Euphrates for 15 km SW of Hazar. Hancock (in press) have demonstrated an Fault-controlled basins along the transform identical origin for deformed gravels along the contain Pliocene lignite, bracketing the initia- NATF. tion of the fault between the medial Miocene Detailed mapping in the Lake Hazar area and the Pliocene. The fault ends in two con- has revealed a 6 km wide fault zone composed tinental triple junctions where incompatibility of subparallel strike-slip, oblique-slip and has led to the formation of complex intracon- normal faults surrounding a young pull-apart tinental basins (Karliova and Adana). North- basin (Hempton 1980). Fault zone structure west-trending left-lateral, strike-slip faults and morphology is very well exposed and up- splay from the main trunk (e.g. the Elbistan lifted lacustrine sediments contain deformed Fault) and become part of the internal horizons, possibly triggered by seismic Anatolian regime. Along the fault there is a activity (Hempton & Dewey 1983; Hempton major locking segment near Bingol, a site of et al. 1983). frequent earthquakes (e.g. 22 May 1971), and Along the SW margin of Lake Hazar, fresh at Celikhan. scarps cut basement rocks, transpose displaced The best exposures of the EATF are in the segments of hill sides and dextrally offset Goynuk River gorge near Choban Tasha, streams and small delta fans. West of Lake about 40 km NE of Bingol. The fault occurs Hazar, subparallel strike-slip and normal faults within a narrow (about 0.25 km) zone of result in a 6 km wide asymmetric transform brecciation and shutter ridge (Allen 1965) valley with a stepped southern wall. Strike-slip topography. Palaeozoic marble is juxtaposed faulting dominates but parallel normal faulting along vertical shear zones against Miocene- occurs on the higher southern wall as fault Pliocene . A few exposures of basalt blocks descend into the eroded transform show narrow (approximately 15 cm), sharp valley. zones of cataclasis where phacoidal or trape- The NATF and EATF join in a complex zoidal clasts are aligned with their long axes , the Karliova Basin, a Pliocene- parallel with the , with a structural Quaternary intermontane sedimentary basin defined by the clasts. approximately 1750 m ASL. The Quaternary The Neogene sedimentary basin near Boran, surface of the basin is almost perfectly flat, yet which is economically important because of no drainage larger than small juvenile streams lignite deposits associated with the dominantly flows through the basin. Coarse clastic sedi- lacustrine sediments, appears to have been ments are limited to alluvial fans at the mouths structurally controlled by the EATF. The of intermittent streams near the basin margin. Boran Basin is strongly deformed, mainly by None of the major faults (NATF, EATF, faulting and, in its wider extents (as much as 5 Varto) enter the basin as single strands, but km), shows spectacular strike-slip valley occur as diffuse zones of brittle deformation. morphology. Angular fault blocks dissected by The arrowhead-shaped Karliova Basin appears stream valleys, are tilted in all directions, a to be a compatibility structure, a zone of feature typical of strike-slip zones. Near extension caused by the westward motion of Boran, a minor is localized on a the Anatolian Wedge, with respect to the lignite originally perhaps 1 m thick. The EACZ. lignite is converted to anthracite and high- Geometrically, the Karliova Basin occupies grade bituminous coal and occurs in discon- an FFT triple junction (McKenzie & Morgan tinuous lenses in which phacoidal and 1969). That the triple junction is contained asymmetric folds determine the sense of offset entirely within continental lithosphere greatly in the lignite. Elsewhere in the Boran Basin, complicates its evolution. If we allow the small thrusts of minor offset are indicated by EACZ to shorten by vertical plane strain and conjugate fractures and Riedel shear assem- the Anatolian Block to move westward on the blages. Between Bingol and Karliova, Recent NATF and EATF, a gap develops at the junc- gravel talus fans have no adjacent source and tion. As 'holes' in the lithosphere cannot exist, are clearly displaced from their original hillside either the crust will extend by plastic flow or sources by strands of the EATF. Several expo- by complex faulting to close the gap, or the sures show these gravels vertically shingled and gap will be filled by igneous rocks and be the deformed by cataclasis with underlying base- locus of intense volcanic activity. The complex ment rocks. Probably, gravels slipped into fault pattern around Karliova (Seymen & extensional fissures in the basement along the Aydin 1972) may be due to such a complica- fault and the fissures then closed and were tion. Where complex intracontinental strike- Downloaded from http://sp.lyellcollection.org/ by guest on October 6, 2021

20 J.F. Dewey et al. slip faults are generated in convergent environ- of large and plans for nuclear facilities. ments, as in Iran, Afghanistan and Central Three segments of the EATF are particularly Asia, such triple-junction holes may be respon- significant for future study: (i) a compressional sible for basin formation and basaltic volcanism. soon-to-break bend near Bingol, (ii) an exten- Most earthquakes on the EATF (including sional pull-apart sag partially occupied by Lake the 1971 Bingol earthquake) and surface faults Hazar, and (iii) a seismic gap SW of Celikhan. occur in the pronounced bend of the EATF Modern seismology has been applied to the around Bingol (Arpat & Saroglu 1972; Seymen for the last three decades. & Aydin 1972), but fault traces aligned with Historical records going back 2000 yr illustrate the other straighter parts of the EATF, occur discrepancies with the instrumental record. to the NW of Bingol, where sharp folding Both these records are inadequate when con- occurs in the Neogene volcanic rocks (Arpat & sidering the time scale involved in tectonic Saroglu 1972). Reconnaissance field work has processes. For more significant evaluations of shown this area to be characterized by many seismic potential a much longer record is recent fault scarps and surficial strain indica- needed. As Allen (1975) and Ambraseys (1978) tors. It appears that a compressional bend near have emphasized, the geological history of Late Bingol is being cut off by a straighter fault Quaternary faulting is a promising source of to the NW much as the compressional bend statistics on the frequency, location and charac- of the San Gabriel Fault was truncated by ter of surface faulting of large earthquake the straighter San Andreas Fault during the shocks. Ambraseys (1978) notes that almost all Pleistocene. shallow earthquakes of magnitude equal to, or Examination of the earthquake distribution greater than, about 6.7 have been associated and topography near Golbasi suggests that this with surface faulting. This surface faulting has area may represent a seismic gap. Epicentres occurred along faults that have or could have are absent for a distance of about 160 km been recognized prior to the earthquake. Thus, between Maras and the Celikhan compressional for a more thorough understanding of the segment S of Malatya, and fault traces are not number of large earthquake events, maximum distinct on Landsat images. We think it expectable magnitude, amount of displacement probable that this segment of the EATF is per event and recurrence interval along the accumulating strain to be released in EATF, it seems imperative to study its Late the near future as a shock sequence that will Quaternary geologic history. The benefits of propagate southwestwards from Celikhan. this approach have been demonstrated in Ambraseys (1971) has shown from historical important studies by Sims (1973, 1975), Wallace records that both the NATF and EATF have (1977, 1978), Sieh (1978) and Swan et al. been active during the period 100-1700 A.D. (1980). and, his time-distribution plots of damaging An additional potential source of much earthquakes suggest that movements along the important data is the detailed mapping of NATF and EATF are diachronous. From surface faults in zones. This is a 0-500 A.D., the NATF was active while the relatively unexplored avenue of earthquake EATF was quiescent. From 500-1100 A.D., research. It is reasonable to assume that the the EATF was active while the NATF was character of surface faults (width, length, con- quiet. In 1100 A.D. the pattern was reversed. tinuity, distribution, pattern, behaviour at Over the last century, the NATF has been depth, texture of fault rocks, etc.) may reveal more active. This accounts for its greater much about their causative faulting and earth- coverage in the literature, especially after its quake processes (Bonilla 1979; Sibson 1983) spectacular migration of large earthquakes and seismic risk. westward from 1939 to 1968 (Dewey 1976). However, there is unequivocal Quaternary East Anatolian Convergent Zone (EACZ) evidence for movement on the EATF (Allen 1975; McKenzie 1976; Hempton 1980). Its The EACZ is bounded to the S by a com- recent quiescence and past history of active- plicated southward-prograding and S-vergent quiescent interludes, suggest that it may system of shallow-dipping thrusts (Assyrides) presently be 'locked' and storing elastic strain involving thin sheets of basement (Bitlis and energy. In the near future, it may be as active Poturge Massifs), ophiolites and ophiolitic and dangerous as the NATF. This makes it a wildflysch, and thick Tortonian flysch sequences critical area for evaluating seismic potential (Lice Flysch). This thrust complex is at present and predicting the character of surface faulting moving southwards (Lice earthquake) over its and is underscored by the recent construction foreland flexural basin (Selmo Molasse) with a Downloaded from http://sp.lyellcollection.org/ by guest on October 6, 2021

Neotectonics of E Anatolia 21 peripheral bulge some 200 km from the thrust clearly rejuvenate older structures. However, front (Fig. 8). Both the foreland basin and there is a preponderance of N-S fractures, foreland platform sequence contain some S- some of which control the positions of vol- vergent folds and thrusts. The northern part of canoes (Nemrut), NE-trending left-lateral, the Selmo foreland basin is S-sloping and is strike-slip faults and NW-trending right-lateral, being uplifted and dissected much like the strike-slip faults. The calc-alkaline double- molasse of the Swiss Plain today. The Assyride peaked strato- of lies in Thrust Belt is undergoing present day strong a complex pull-apart graben on a wide zone of uplift as witnessed by the antecedent Euphrates dextral transcurrent motion. Parasitic cones Gorge that cuts through the Poturge Massif associated with the large volcanoes of Eastern and leaves the truncated beds of intermittent Anatolia show N-S (Nemrut) and NW streams stranded high on its walls. The foreland (Ararat) trends (Fig. 11). Analysis of the has several large basalt spreads, especially topography indicates that all elongate basins in immediately SW of Diyarbakir, which, from the EACZ overlie a strike-slip fault and that the distribution of small conical volcanic most of the elongate high areas are bounded centres, were erupted from approximately by strike-slip faults. It seems certain that some N-S fissures. These N-S fissures are part of a of the intra-continental convergent strain is widespread array of N-S fissures and grabens taken up along these faults although their role and conjugate fractures throughout the Arabian is probably minor compared to the NATF and Shield, that appear to have been caused by the EATF. East of long. 41°E, the strike-slip faults Neogene collision between Arabia and Turkey/ are closely spaced, shorter and enclose larger Iran. Analogous structures, trending NE, but angles about N-S bisectors, suggesting that not accompanied by volcanism, characterize they may have experienced external and/or the eastern part of the Arabian Platform, SW internal rotation. of the Zagros Front and have been interpreted Several regions (e.g. Keban) are seismically also as products of Iranian/Arabian collision by quiescent and topographically smooth with Hancock et al. (1984). relatively little relief and less fracturing and North of the Assyride Thrust Belt, the E may represent more rigid stronger flakes. The Anatolian Plateau has a variable and com- Munzur Mountain block appears to be a plicated pattern of deformation. Neotectonic thrust-bounded horst. Thrust and fold zones fractures (Fig. 8) box the compass and many are important E of long. 41°E where Neogene


FIG. ll. Morphology of the four principal volcanoes of Eastern Anatolia contoured in metres and showing distribution of parasitic cones and small centres. Downloaded from http://sp.lyellcollection.org/ by guest on October 6, 2021

22 J.F. Dewey et al.

~,~i0-I Z's-I _

I LVZ --

FIG. 12. Schematic strike-slip pull-apart basins involving (a) only the elastic lid (basement symbol), and (b) the whole lithosphere. Black: granite; fine stipple: ; fine random ornament: mafic igneous rocks of mantle origin. basalts are strongly folded as, for example, suggests that the precise location of the EATF along the northern margin of the Van-Mus is a flake boundary and coincides with a litho- Basin. Between these zones of strong folding spheric transfrom only in a general way. the basalts are almost flat-lying, suggesting that Seismic activity, characterized by strike-slip strain-free regions (flakes) are bounded by and thrust slip, occurs around the fault bounded zones of strong shortening (flake boundaries). edges of the flakes within the upper seismogenic The Upper Miocene basaltic andesites of the elastic layer. Faulting is probably discon- Solhan Formation (Fig. 10) show this variation tinuous in space and time, accounting for particularly well, whereas the Pleistocene the complex heterogeneous distribution of Holocene volcanics around Lake Van are faults and seismic activity. East of long. 41°E, almost unfolded. The Neogene volcanics of deformation has been more intense, where the Eastern Anatolia hold great promise as strain lithosphere has probably suffered more bulk guages for Anatolian deformation, for studies shortening and thickening than the region W of both regional strain homogeneity and local of long. 41°E, where some shortening seems to and regional strain rate with time. be accommodated by slip along the NATF and The fault, seismic, topographic and field EATF. data suggest a regionally inhomogeneous strain pattern~ in the EACZ. Crustal/lithospheric Basin development shortening appears to be accomplished by a complex interaction of horizontal plane strain Basins in Anatolia are of four types; the by strike-slip wedge tectonics and vertical Selmo foreland flexural basin, pull-aparts on plane strain by the detachment of 10-15 km major strike-slip faults (Erzincan, Hazar), thick thrust flakes. The transpressional bend in flake and plate triple-junction compatibility the EATF at Celikhan coincides with the inter- basins (Karliova and Adana) and elongate section with the Assyride thrust front and strike-slip/thrust-bounded basins (Van-Mus). Downloaded from http://sp.lyellcollection.org/ by guest on October 6, 2021

Neotectonics of E Anatolia 23

500 T°C 1000



kb 45



25 0.5 Fro. 13. Pressure/depth/temperature plot showing some possible relationships between the P/T paths of rocks in convergent orogens, associated igneous rocks and metamorphic facies. 0.5, 1, 2 and 4 refer to geothermal gradients appropriate to these heat flow values in HFU. Dotted line: geothermal gradient reconstructed for the Lepontine metamorphic core of the Alps (England 1978) at about 30 Ma; large square: kyanite in the Tauern fenster; DL, unrecovered geothermal gradient for crust delaminated at 65 km; DL + DN, unrecovered geothermal gradient for crust delaminated at 65 km and truncated by 5 km denudation; TP, unrecovered geothermal gradient for crust doubled by a single thrust xy: P/T paths of a point in a crust undergoing high strain rate, isothermal, horizontal stretching to [3 = 4. xz: P/T path of a point in a crust undergoing shortening by a value of 0.5 followed by denudational stripping. A, andalusite; AM, amphibolite; AO, andulusite overprint; B, blueschist; D, ; E, eclogite; GA, ; G, greenschist; GD, granodiorite; GR, granulite; J, jadeite; K, kyanite; P, pumpellyite; S, sillimanite; SO, sillimanite overprint; Z, zeolite; WGM, beginning of wet granite melting; WTM, beginning of wet tholeiite melting (Yoder & Tilley 1962; Boettcher & Wyllie, 1968).

An interesting contrast exists between pull- may be a pull-apart that affects only the elastic apart basins on the NATF and those on the lid (Fig. 12a). EATF. From the slip rates of the NATF and Shortening and thickening of the crust may EATF and the lengths of pull-apart basins along be accomplished by either or both vertical them, the Erzincan Basin has an extensional stretching and thrust imbrication. Vertical strain rate of 10 -13 s -1 while the Hazar Basin stretching and thermal re-equilibration yields a has one of 3.6 x 10 -15 s -~. Both pull-apart P/T path of the general form X ~ Z (Fig. 13). basins lie within the regionally thickened crust Much higher geothermal gradients (Fig. 13) of Anatolia, yet the Erzincan Basin contains can be achieved by localized zones of rapid volcanics whereas the Hazar Basin does not. (isothermal) stretching, lithosphere delamina- Strain rate may be the governing difference; in tion (Bird 1978; Houseman et al 1981) and regions of high extensional strain rate the very rapid uplift and denudation (England lithosphere may be thinning faster than it 1978; England & McKenzie 1982). Rapid thickens by regional shortening and by thermal Oligocene uplift and denudation was probably re-equilibration, thus allowing fertile mantle to responsible for the Lepontine high T/P meta- depths where partial melting is possible (Figs morphism in the Central Alps, whereas 12b and 13). Alternatively, the Hazar Basin and rapid lithospheric stretching Downloaded from http://sp.lyellcollection.org/ by guest on October 6, 2021

24 J.F. Dewey et al. in pull-aparts, also affords a rationale for zone at the edge of the Anatolian Plateau. It mantle partial melting beneath and mafic could be a major pull-apart in a complex left- igneous rocks in orogens (Fig. 13), perhaps lateral transform zone that continues the accounting for the Ne/Hy normative basalts of EATF trend, but its position at the Africa/ Eastern Anatolia. Silicic but especially mafic /Anatolia triple junction (Fig. 14) offers a melts are possible causes of both sillimanite better kinematic explanation. A vector triangle and andalusite overprints of Buchan-type derived from the known trends and slip rates (Fig. 13) and the position and of the EATF (1.7 mm yr -1) and the Syrian shape of intrusions may be controlled by segment of the Transform Zone compatibility holes and gaps caused by flaking (4 mm yr -1) yields an Africa/Anatolia slip of the elastic lid particularly beneath pull-apart vector of 3.4 mm yr -I (Fig. 15) which gives basins (Fig. 12). Rapid stretching of the litho- NNW extension in the Adana region close to sphere also affords an explanation for localized the Maras triple junction. A normal fault first zones of andalusite-facies regional metamor- motion derived NW-trending slip vector, for phism not obviously related to intrusion. an earthquake on the northern edge of the Isothermal stretching allows crustal rocks at Adana Basin (Canitez & Ucer 1967), is close appropriate levels to pass rapidly from silli- to this extensional azimuth. The Adana Basin manite or kyanite fields into the andalusite may be, therefore, simply a compatibility gap field (Fig. 13, X ~ Y). Thus, a spatial relation- resulting from the evolution of an FFF plate ship is theoretically possible between pull- triple junction. apart rift basins at high levels and andalusite growth and horizontal stretching fabrics at Balancing and the slip vector deeper levels. The Mus-Van Basin is an elongate structure We now consider possible relationships along the northern edge of the Bitlis Massif between shortening, crustal thickening, strike- forming a sediment plain in the Mus region slip faulting and relative plate motion. Figure and the deeper parts of Lake Van. In Lake 14 summarizes the tectonics of Eastern Van, reflection profiles show considerable but Anatolia, and Fig. 15 is a vector diagram that zonal fault-controlled sediment deformation, relates relative plate displacements, slip on the particularly in the northern part of the basin NATF and EATF, and strain in Eastern (Degens & Kurtman 1978). Similarly, along Anatolia. The average convergence slip rate the northern margin of the Mus segment, between the Arabian and European Plates for Miocene carbonates and basalts are strongly the last 14 Ma has been 15.3 mu yr -1 giving an folded and locally cleaved. The Mus segment average strain rate of about 2.0 × 10- 15 s- 1 m • has a southerly slope, which has controlled the the EACZ, the slip rate derived by adding the progressive southward migration of the Murat Dead Sea transform slip rate (AF/AR, 5.3) to River. Terraces along the Murat River, just N the Europe/Africa slip rate (EU/AF, 10). This of the basin, slope southwards from 20 and 3 m Europe/Arabia rate may be too high for to the Murat alluvial plain in a distance of Anatolia, because the average slip rate on the some 5 km and artesian conditions exist in the Dead Sea Transform, N of the Antilebanon, basin. Pliocene/Pleistocene alluvial gravels and has been 4 mm yr-1, suggesting motion between sands in the Mus segment are affected by small the Arabian and a smaller Syrian Plate across thrust, normal and wrench faults on a fairly the Antilebanon/Palmyran Zone, a zone of penetrative scale. A seismic reflection profile present-day weak seismic activity and recent kindly loaned by TPAO (Turkish shortening (A. Quennell, pers. comm.). If Company) shows that the Mus segment con- a Europe/Syrian (EU/SY) convergence rate tains up to 4 km of Late Miocene to Pleistocene of 14 mm yr -x is appropriate, the average sediment along its northern margin but thin- E Anatolian strain rate is slightly reduced. ning and overlapping southwards. The northern A Europe/Syria/Anatolia (ANX) vector margin appears to be a transform/thrust right- triangle (Fig. 15) is constructed from the lateral flower structure (a structure in which a length and trend of the Europe/Syria slip and narrow zone of strike-slip displacement at the trends of the NATF and EATF. This con- depth widens or flowers upwards into a zone of struction indicates that, were the Europe/Syria en ~chelon thrusting) and the Mus-Van Basin convergence taken up solely by lateral wedging may be regarded as a ramp/strike-slip basin of a rigid Anatolian Block, the slip rates on the along a flake margin. The volcanics of Nemrut NATF and EATF would be 18.5 mm yr -~ and appear to truncate the basin unconformably. 19.3 mm yr -1 respectively. An 85 km offset of The Adana Basin is a lithospheric extension a Miocene suture along the NATF (~;eng6r Downloaded from http://sp.lyellcollection.org/ by guest on October 6, 2021

Neotectonics of E Anatolia 25

/ /"



~ORM~J-~~.yq.Z.I.m~L~,, "--1"

/ FIG. 14. Simplified tectonic map of Central and Eastern Anatolia and the Arabian Foreland showing slip rates on major faults, focal mechanisms for the Varto, Lice and Adana earthquakes (dilational quadrants: white; compressional quadrants: black, see also Fig. 9), young basins (stippled) and major plates (each with separate lined ornament) bounding the eastern plate boundary zone.

1979) gives a slip rate of 8.9 mm yr -t and a 15 complex array of southward-verging thrusts in km offset of the antecedent Euphrates River a the Bitlis Zone, several zones of folding and slip rate of 1.7 mm yr -1 along the EATF. thrusting and NW-trending right:lateral faults. Therefore, only a small part of the Europe/ Thus, E of a N-S line AB (Fig. 14) through Syria convergent displacement is taken up by Maras, about a third of the convergence slip on the NATF and EATF. Slip on the between Syria and Europe is taken up by NATF and EATF may be combined to give an wedging and westward slip of the Anatolian approximation of the slip direction and rate Block while the rest is accommodated by on the Varto Fault, a steeply NE-dipping thrust thrusting in the Bitlis Zone, folding and thrust- immediately E of the Karliova junction; the ing on other zones of shortening, displacement coincidence with the slip direction derived on mainly right-lateral, strike-slip faults and from the fault plane solution for the Varto internal strain within the Anatolian Block. earthquake (Dewey 1976) is reasonably close. West of line AB, convergent plate motion The portion of the EU/SY convergence vector across the Anatolian Block is constrained by that is not taken up by wedging and slip on the the Europe/Africa motion of 10 mm yr-1; the Varto Fault is given by the SY/SY join, which AF/AN join of 10.5 mm yr -1 represents an represents possible combinations of strain unknown combination of slip on the Africa/ within the Anatolian Wedge between the Anatolian boundary and strain within the NATF and EATF and strain between the Bitlis Anatolian Block. If no internal Anatolian Thrust Zone and the Minor Caucasus. The strain is occurring, the AF/AN join gives the SY/SY join is close in azimuth to the slip Africa/Anatolia slip direction. vector on the NE-dipping Lice Thrust derived That internal Anatolian Block strain has from the fault plane solution for the Lice occurred both E and W of line AB, is shown earthquake. The Europe/Syria convergence is by the 2 km high plateau which indicates a taken up E of the Karliova junction by a crust thickened to about 52 km, assuming Downloaded from http://sp.lyellcollection.org/ by guest on October 6, 2021

26 J.F. Dewey et al.




AR "~ ' ' ' ' '

Fro. 15. Vector diagram for Eastern Anatolia. Slip rates in mm yr -1. AF, Africa; AN, Anatolia; AR, Arabia; EAT, E Anatolian Transform Fault; EU, Europe; NAT, N Anatolian Transform Fault; SY, Syria. approximate isostatic balance. Thus, in addi- 104 km 2 must be accounted for by a combina- tion to convergence accommodated by strike- tion of lateral flow from the section plane and slip motion, which alone, can allow only a perhaps by the thrust restacking of thinned horizontal plane strain, the crustal thickening crust in the Himalayas. A similar calculation indicates about 80% vertical stretching. How for Eastern Anatolia (Fig. 16) yields the very this vertical component is structurally accom- small missing value of 12 × 10 2 km 2 according modated is unclear, as is its temporal rela- with the low slip rates on the NATF and tionship with strike-slip faulting. A similar EATF. tectonic situation exists in Tibet where a 5 km Plane strain shortening and vertical stretch- plateau caps an 80 km crust probably thick- ing of the crust, followed by lateral flow and ened by vertical stretching (Dewey & Burke spreading in the overall extent of the deformed 1973; England & McKenzie 1982). zone once the crust has reached a thickness of Another approach to relating crustal about 80 km, has substantial implications for thickening and lateral flow is to compare the orogenic polyphase strain sequences and cross-sectional input into the convergent zone fabrics below the elastic lid and for fault from plate slip rates with the cross-sectional sequences and tectonic regimes in the elastic area present in the convergent zone (Fig. 16). lid (Fig. 17). With no denudation, constant For the Himalayas and Tibet, a convergence volume and assuming coaxial bulk strain, rate of 50 mm yr-l for 40 Ma has contributed vertical plane strain proceeds to 63% shorten- 6.2 × 104 km 2. We add this to the material ing, followed by a phase of horizontal plane originally in the 1000km long box (3.1 × 104 strain at 168% bulk stretching that yields an km2. About 20 km has been denuded from the oblate bulk strain at 86% shortening. Further High Himalayas, whereas the Tibetan Plateau shortening gives an oblate bulk strain at pro- has suffered little denudation except for gressively higher K values. If an Argand value perhaps as much as 5 km to expose of 3 buffers crustal thickness at about 80 km, along its southern border. This gives a maxi- plane strain at shortening values greater than mum of 3.25 x 103 km lost by denudation; 63% may be accomplished where denudation 7.755 × 104 km 2 is the cross-sectional area of accompanies shortening, where volume loss crustal material in Tibet. The missing 1.22 × occurs, or by non-coaxial strains in high strain Downloaded from http://sp.lyellcollection.org/ by guest on October 6, 2021

Neotectonics of E Anatolia 27

Himalayas -Tibet ,,-~m Indus Suture //2p~ ...... y /..-XI5+ -,,__ [----i .... I...... -r-'v', - ,3,1 1 [ 80 Hercynian Arcnaean I<_~1__ 1000_ i - ~,1 J,Tnegligible addition )-~ / ~ / lateral + Moho3~ stacking of extrusion 3.25 x 10 4xdenuded thinned crust ~ 1 2~2xI04 ]- ~' "'-'- x "^/-* I --] .... L • <-5-/./~ lu <-F-~.dxlU

Eastern Anatolia

31, t I Pan Africa~ arcs- accretionary prisms I Hercynian

/ Ioteml extrusion minor denudation stacking of ~ ~i.F~ 1--~~ thinned crust '<------1.2 xlO 1 2.12 x 104

Fro. 16. Crustal balancing diagrams for the Himalayas/Tibet and Eastern Anatolia. Cross-sectional areas in km 2 and heights and lengths in km. Area fed in by convergent plate motion underlined. Area within deformed zone in box. zones. This model predicts that lateral wedging eliminated, a much broader phase of bulk along major strike-slip faults with their asso- intra-continental shortening and thickening ciated pull-apart basins will be superimposed commences near the suture (e.g. Tibet), on the orogen at a late stage and accounts well accompanied by more distant foreland for the young N-S grabens of the Tibetan deformation characterized by large intracon- Plateau. tinental strike-slip faults laterally moving huge blocks of continental lithosphere (e.g. in Conclusions Eastern Asia). In this phase, convergent strain is distributed up to thousand of km from the After initial , -continent con- suture zone of initial collision. vergence results in two phases of deformation Molnar & Tapponnier (1975, 1977a,b, 1978, (Dewey & Seng6r 1979). Because of the 1979) have explained the pattern of large irregular shapes of continental margins charac- strike-slip faults in the Asian Foreland, com- terized by salients and embayments, collisional prising , Mongolia and the southeastern deformation begins at projection points of U.S.S.R., as being analogous to slip lines initial impingement, separated by oceanic developed experimentally in plastic materials embayments. Continued convergence results in when indented by a rigid die. This model for thrusting and strike-slip movement of con- continental crustal deformation breaks down tinental slivers over open boundaries of the when considered in detail. Not all strain is remaining (McKenzie 1972b; distributed in uniformly spaced strike-slip Dewey & Burke 1973). Most of the convergent faults. There is significant folding and thrusting strain is confined to the zone defined by the in and around the Tien Shah, Nan Shan, and extent of remaining oceanic lithosphere. When Lung Men Shan. Large areas of extension all the oceanic lithosphere in the system is occur where strike-slip faults sidestep close to Downloaded from http://sp.lyellcollection.org/ by guest on October 6, 2021

28 J.F. Dewey et al.

k=oO 20[-

x/.v I ." ....::. RW"~ •,,.,,,, 768 0

t y/z 5 10 20

FIG. 17. Log deformation plot with shortening lines at 63 and 86% and elongation line at 168%. Arrowed line is theoretical strain path for a crust originally 31 km thick shortened by vertical plane strain and thickened to 80 km (a), followed by horizontal plane strain by lateral extrusion (b). Dots: strain states replotted from Pfiffner & Ramsay (1982). Enveloped areas: H, Dalradian slides (Hutton 1979); RW, slates (Ramsay & Wood 1976); WI, Palaeozoic rocks of western Ireland (Dewey 1969). the 'open' oceanic boundary to the E as in the cautiously; the models apply only to two- Baikal Rift System, the Shansi Graben System dimensional instantaneous patterns whereas and the Yunnan Grabens. Movement on large the geological structures with which they are foreland strike-slip faults is significant during compared are three-dimensional objects orogeny and probably continues after super- evolving over millions of years. The pattern ficial folding as shown by parallel post-folding comparison should be made only where the strike-slip faults in the Hercynian of Northern detailed geological picture is known; this is not Africa (Mattauer et al. 1972; Arthaud & Matte yet the case for Anatolia. 1977) and the of the Canadian It is suggested the there is a causal relation- Shield. However, the motion of the Anatolian ship between the NATF, EATF and the Wedge, with respect to its northern and extensional ova regime of Central Anatolia. southern neighbours and the movement of the This provides a possible and testable solu- bounding transform faults with respect to the tion to the geometrical as well as temporal bisector of their dihedral angle, resembles aspects of the longstanding ova problem in greatly the situation encountered in plastic Anatolia. Similar pairs of large-scale strike-slip extrusion in a modified Prandtl Cell. The faults faults exist elsewhere in continental litho- that bound the Central Anatolian ovas are sphere; in the Mojave Desert of California, similar to the slip lines in the Prandtl Cell. their possible role in controlling the deforma- Similar models of plastic behaviour have been tion of continental lithosphere was noted by proposed to explain large-scale fault patterns Cummings (1976). More detailed field evidence in the Mojave Desert of California (Cummings will have to be obtained, especially to discover 1976) and in the Caribbean (Burke et al. 1978). whether or not there is any systematic change Such comparisons should be approached in the amount of offset along the NATF and Downloaded from http://sp.lyellcollection.org/ by guest on October 6, 2021

Neotectonics of E Anatolia 29

EATF and to determine the precise geometry vary considerably. Where large [5 values occur- of the Karliova 'triple junction and the fault red -- up to a maximum of about 4.5, Dewey system that formed the Anatolian ovas and the (1982) the mantle is brought near the age and subsidence history of the ovas. Micro- surface and is more likely to be involved in earthquake surveys within Anatolia may subsequent flake detachment. The width over provide further information as to whether or which stretching occurred may vary from very not the deformation in the Anatolia Wedge, narrow, as for the sharply defined Red Sea which appears from Quaternary geology to be margins, to over 100 km, as for the more active, is generally free of earthquakes and is diffuse Bay of Biscay margin. Hence, sub- thus perhaps proceeding in a non-brittle sequent collisional margins will correspondingly fashion. vary from the sudden impingement of sharp Slip-line field theory may be appropriate for ramparts to the more progressive restacking of describing deformation below the brittle thicker and thicker crust along listric faults that seismogenic zone associated with large strike- cut progressively deeper into the thickening slip faults (Earton et al. 1970; Ben-Menahem crust. The polarity of earlier listric normal 1976). Recent fault zone models consider faults appears, generally, to be 'down to ocean strain to be accommodated via creep, in zones or rift' so that thrust regeneration induces a that deform plastically (Prescott & Nur 1981). continent-ward thrust polarity. However, Differences in the pattern of large strike-slip reversed polarity arrays of normal faults occur faults in Eastern Asia and experimental for example in the Gulf of Suez, separated patterns of slip lines may be due to crustal from normal polarity arrays by transform heterogeneity as well as different boundary relays. Thrust remobilization of such arrays conditions. Perhaps we should consider upper- would generate early r~trocharriage basement crustal convergent deformation in terms of nappes. The time elapsed between stretching domains or flakes of ductility contrast at all and subsequent shortening determines the scales. Older crustal domains are cooler and thermal age of the continental margins, i.e. its thicker (Sclater & Francheteau 1970) and lithospheric thickness, geothermal gradient, therefore stronger than younger, hotter and strength profile and hence, the thickeness of weaker crustal domains. This is particularly the elastic lid and position of detachment well developed in Eastern Asia (Molnar & zones during collision. Thermally young v. old Tapponnier 1981). The shields, margins will have thin v. thick thrust flakes India, Tarim and Angara, are relatively un- respectively. deformed and have been little affected by the A second factor is the and history of India-Eurasia convergence since the Tertiary. the colliding crust. If the collisional zone Conversely, the Mesozoic and Tertiary terrane results from the collapse and of ultra- of Tibet has suffered extensive diffusely distri- (Alps), collisional, thrust stacking occurs buted deformation (Landsat imagery studies from the beginning of convergence. Orogens by J.F.D. and W.S.F.K.). Palaeozoic terranes with substantial strike continuity result, of Tien Shan, Altai and Khangal are deformed although individual basement nappes may be but in a less diffuse style than in Tibet. A laterally impersistent. Where collisional zones ductility contrast probably also explains the are the terminal culmination of a longer less deformed nature of the Lut Block in Iran history of oceanic closure, often of several (Mohajer-Ashjai et al. 1975). oceans (Himalayas/Tibet), continental margins Many factors affect the structural style and may consist of assemblages of collided and tectonics of continental collision zones, such as laterally transposed blocks, arcs, exotic ter- their width, amount of shortening, the role of ranes and subduction-accretion prisms (Coney lateral wedging, the extent to which plateaux et al. 1980), along whose boundaries tectonic are developed and degree of strike continuity, mobilization can occur during collision. among many others. We list and outline below Furthermore, a long subduction history will those factors that we believe are important in thermally weaken the overriding margin and determining collisional orogenic style and render it more susceptible to convergent which, in combination, determined the great deformation. The time since earlier stabiliza- range of styles and geometries in the Alpine/ tion of colliding cratons -- i.e. their thermal Himalayan system (Fig. 1). age -- plays a fundamental role not only in First, the amount, rate, style, duration and determining the amplitude and wavelength of timing of lithospheric stretching that generated foreland flexures (Karner & Watts 1983), but the continental margins and rifts involved in also in the extent and style of foreland/hinter- subsequent collision-induced shortening may land deformation. The Archaean shield of Downloaded from http://sp.lyellcollection.org/ by guest on October 6, 2021

30 J.F. Dewey et al.

India is little affected by collisionally-in- be induced in collisional orogens (Dewey & duced deformation whereas the younger more Burke 1973; Dewey & Kidd 1974). inhomogeneous Sinian hinterland is affected Fourthly, the direction, rate, duration and by thickening, wedging, folding and rifting for degree of convergence are a controlling in- several thousand km from the Tsangpo Suture. fluence. Very oblique convergence induces a Continental crust varies greatly in character strong strike-slip component (Dewey & from the anhydrous granulitic , typical Shackleton 1984). A high convergence rate of some Proterozoic orogens, to hydrous sub- feeds continental lithosphere into the con- duction accretion prisms (Dewey & Windley vergent system at high rates, which in turn may 1981). The consequent lateral and vertical generate a wider orogen (Ben-Avraham & variations in bulk rheology of crustal profiles Nur, 1976). Average strain rates appear to be and the positions of potential detachment remarkably constant at about 1.5 x 10 -15 s -1, surfaces, determined both by the brittle- which indicates a buffering mechanism, pro- ductile transition and by crustal inhomogeni- bably olivine flow laws (England 1983). The ties, controlled stratigraphically and petro- duration of post-collisional convergence pro- graphically (Fig. 5), will be important in bably results from the configuration and evolu- determining strains and displacements during tion of the other boundaries of the plates collision. involved in the collisional zone (Bott 1982; Thirdly, collisional geometry and history is Bott & Kusznir 1984). Where a long period of controlled to a large extent by the shapes of convergence occurs, shortening zones appear colliding margins. Continental margins vary to become wider (England & McKenzie 1982) from rather simple and straight (Chile) to and large amounts of shortening may impose a complex and irregular with promontories and strong cylindroidism on the orogen. embayments. Therefore, collision is likely to Fifthly, the age of collision may be be diachronous from promontories, from important. Average plate thickness and slip which lateral crustal flow and wedging occurs, rate has decreased with time (Dewey & to embayments that subduct beneath lateral Windley 1981). This is reflected in consider- wedges, which may expand above subduction able differences between earlier Precambrian zones (Aegean), and receive the erosional and later Phanerozoic orogens. Precambrian products of the colliding promontories and deformation zones are wider with shorter close last. Hence great lateral variations may wavelength flexural features, pervasive fore-

(a) direct plate slip direction-fabriccorrelation (c) flake motion X ~" ,r \ ",r v \ X ,v \ ,i, v \ • ,Iv (b) block margin complexities eod,ng le/ reto.on ; ] ...... ~-:.-.- ...... ~ w[ ~,.~~X\~

(f) irregular diochronous collision/ ¢" 1 2- ~11"~..V,,"\ 3 1' \\1 l''\,J

suture ~ ~ ~ ¥ ~/" ** transform ,~lJlllllllllll suture "-~/

Fro. 18. Relationships between plate boundary slip vectors faulting and fabrics. Discussion in text. Oceanic lithosphere indicated by vertical lines, black ellipses: schematic sections through strain ellipsoids. All plan view except el which is a section. Downloaded from http://sp.lyellcollection.org/ by guest on October 6, 2021

Neotectonics of E Anatolia 31 land deformation and thicker stratigraphic in strain axes and causing the reorientation of sequences. early formed structures by both internal and Displacement and strain within the whole external rotation (Fig. 18e3). The lateral plate boundary zone should integrate to equal wedging of flakes (Fig. 18c) causes great varia- the slip vector between the converging plates. tions in slip direction at their margins and the The strain continuum or strain/displacement lateral flow of blocks over oceanic tracts (Fig. homogeneity problem is of fundamental 18b) may cause (Aegean) extension in the importance to the basic and difficult question overall plate convergence direction. Lastly, of the relationship between and gravity spreading from zones of thickened to classical , i.e. the extent to unthickened crust will generate structures that which, and under what conditions, relative may bear little relationship to the plate slip plate slip vectors are directly expressed in direction. structural geometry and fabric and, critically, If relationships are to be established how plate slip vectors can be deduced from between structure and plate slip vector there- integrated structural studies. Along narrow fore, we should be concentrating probably on and well-defined oceanic plate boundaries, a aureoles and the earliest fairly direct and relatively simple relationship high strain zones in the blueschist facies. Later between slip vector and structure might be flaking may rotate many of these earlier expected at, for example, ridge/transform formed structures but systematic palaeomagne- systems in subduction-accretion prisms and tic work will be an important tool in discover- along ophiolite obduction zones. Perhaps a ing and sorting out these possible rotations. simple relationship (Fig. 18a) would be Of course the recognition and understanding expected also where the first feathering of continental collision depends upon our collisional contacts are made by thrust restack- ability to recognize and understand collisional ing in zones of thin continental crust below sea sutures, whose position is not always readily level and commonly in the blueschist facies apparent. Sutures may be exceedingly cryptic (Dewey, 1982), where body forces are not yet (Dewey & Burke 1973). Also oceans may close important. As collisional tightening and crustal along one boundary by transform motion thickening proceed, complexities increase, (Fig. 18g) which drags in slivers of oceanic particularly in the elastic lid, by the complex lithosphere along a transcurrent fault zone. As motions of crustal flakes (Fig. 18b) much of Robert Shackleton once remarked, 'a cryptic which are induced by diachronous irregular suture is one that exists and you can't see but a collision and controlled by older crustal hallucinosuture [Fig. 18f] is one that you boundaries and other inhomogeneities (Fig. think you can see but doesn't exist'. 18f).The earlier simpler relationship between structure and slip vector progressively breaks ACKNOWLEDGMENTS: We are deeply indebted to down. Flakes and blocks at many scales rotate Paul Hancock and John Platt who greatly improved about both horizontal (Fig. 18el) and vertical an early draft of this paper. (Fig. 18e2) axes inducing great local variations


ADAMIA, S. A., LORDIKIPANIDZE, M. B. & occurred in Turkey and its vicinity. Contr. No. ZAKARIADZE, G. S. 1977. Evolution of an 6, Karadeniz Teknik Univ., Dept Geophys., active as exemplified by the Trabzon, Turkey. Alpine history of the Caucasus. , AMBRASEYS, N. N. 1970. Some characteristic features 40, 183-99. of the Zone. Tectono- AKX, K. 1972. Earthquake mechanisms. Tectono- physics, 9, 143-65. physics, 13, 423-46. __ 1971. Value of historical records of earthquakes. ALLEN, C. R. 1965. Transcurrent faults in continental Nature, Lond. 232, 379-97. areas. Phil. Trans. R. Soc. London, A258, __ 1978. Middle East -- a reappraisal of the seismi- 82-9. city. Q. Jl Engineering Geol. 11, 19-32. __ 1969. Active faulting in northern Turkey. Contr. & ZAXOPEK, A. 1968. The Varto Ustukran No. 1577, Div. Geol. Sci., California Institute of (Anatolia) earthquake of 1966 August 19: a Technology, Pasadena. summary of a field report. Bull. Seism. Soc. __ 1975. Geological criteria for evaluating seismi- Am. 58, 47-102. city. Bull. geol. Soc. Am. 86, 1041-57. ARGAND, E. 1924. La tectonique de l'Asie. Rep. ALPTEKIN, O. 1978. Magnitude-frequency relations 13th Int. Geol. Congr. l, 171-372. and deformation release for earthquakes that ARMBRUSTER, J., SEEBER, L., & JACOB, K. H. 1978. Downloaded from http://sp.lyellcollection.org/ by guest on October 6, 2021

32 J.F. Dewey et al.

The northern termination of the Himalayan BERNOULLI, D., DE GRACIANSKY, P. C. & MONOD, mountain front: active tectonics from micro- O. 1974. The extension of the Lycian Nappes earthquakes. J. geophys. Res. 83, 269-82. (SW Turkey) into the southeastern Aegean ARPAT, E. & Iz, H. B. 1977. 1976 Caldiran Depremi. . Eclog. geol. Helv. 67, 39-90. Yeryuvari ve lnsan, 2, 29-41. BIRD, P. 1978. Initiation of intracontinental subduc- __ ~ SAROGLU, F. 1972. The East Anatolian Fault tion in the Himalaya. J. geophys. Res. 83, System: thoughts on its development. Bull. 4975- 87. miner. Res. Explor. Inst. Ankara, 78, 33-9. BLUMENTHAL, M. M. 1945a. Ladik deprem batti. ARTHAUD, F. & MATTE, P. 1977. Late Paleozoic MTA Mecmuasi, 1/33, 153-74. strike-slip faulting in southern Europe and 1954b. Die Kelkit-Dislokation und ihre northern Africa: result of a right-lateral shear tektonische Rolle. MTA Mecmuasi, 2/34, zone between the Appalachians and the Urals. 372-86. Bull. geol. Soc. Am. 88, 1305-20. BOETrCnER, A. L. & WYLLIE, P. J. 1968. Melting of ATAMAN, G., BUKET, E. &, CAPAN, U. Z. 1975. granite with excess to 30 kilobars pres- Could the North Anatolian Fault Zone be a sures. J. Geol. 76, 235-44. palaeo-Benioff Zone? Bull. miner. Res. Explor. BONILLA, M. G. 11979. Historic surface faulting-map Inst. Ankara, 84, 97-102. patterns, relation to subsurface faulting, and BARAZANGI, M. & Ni, J. 1982. Velocities and pro- relation to pre-existing faults. In: Proceedings of pagation characteristics of Pn and Sn beneath Conference VIII: Analysis of'Actual Fault Zones the Himalayan arc and Tibetan plateau: possible in Bedrock, Open File Rep. 79-1239, 36-65, evidence for underthrusting of Indian contin- U.S. Geological Survey, Menlo Park, California, ental lithosphere beneath Tibet. Geology, 10, U.S.A. 179-85. BoTr, M. H. P. 1982. Origin of the lithospheric BARKA, A. A. & HANCOCK, P. L. 1984. Neotectonic causing sedimentary basins. Phil. Trans. deformation patterns in the convex-northwards R. Soc. London, A305, 319-24. arc of the North Anatolian Fault Zone. In: __ & KUSZNIR, N. J. 1984. The origin of tectonic DIXON, J. E. & ROBERTSON, A. H. F. (eds) The stress in the lithosphere. Tectonophysics, 105, Geological Evolution of the Eastern Mediter- 1-13. ranean. Spec. Publ. geol. Soc. London, BRINKMANN, R. 1976. , 158 pp. pp. 763-74. Blackwell Scientific Publications, Ferdinand Enke, Stuttgart. Oxford. BRUNE, J. N. M. 1968. Seismic moment, seismicity BARTON, C. M. & ENGLAND, P. C. 1979. Shear and rate of slip along major fault zones. heating at the Olympus (Greece) thrust and the J. geophys. Res. 73, 777-84. deformation properties of carbonate at geo- BRUNN, J. H., DUMONI', J. F., DE GRACIANSKY, P. C., logical strain rates. Bull geol. Soc. Am. 90, GUTN1C, M., JUTEAU, T., MARCOUX, J., 483- 92. MONOD, O. & POISSON, A. 1971. Outline of the BECK, M. E. 1976. Discordant palaeomagnetic pole geology of the western Taurids. In: CAMPBELL, positions as evidence of regional shear in the A. S. (ed.) Geology and , western Cordillera of North America. Am. J. pp. 225-55. Petrological Exploration Society, Sci. 276, 694-703. Libya, Tripoli. BE1N, A. & GVlRTZMAN, G. 1977. A Mesozoic BONCEV, E. 1974. Generalities of the Kraistide- edge of the along the Levant Vardaride . In: MAHEL, M. (ed.) coastline and its bearing on the evolution of the Tectonics of the Carpathian Balkan Regions, eastern Mediterranean. In: BIJU-DUVAL, B. & pp. 330-2. Dionyz Stur, Bratislava. MONTADERT, L. (eds) Structural History of the BURKE, K., Fox, P. J. & SENGOR, A. M. C. 1978. Mediterranean Basins, pp. 95-110. Editions Buoyant ocean floor and the evolution of the Technip, Paris. Caribbean. J. geophys. Res. 83, 3949-54. BEN-AVRAHAM, A. & NUR, A. 1976. Slip rates and BUYUKASIKOGLU, S. 1979. Seismolojik Verilere Gore morphology of continental collision belts. Guney Anadolu ve Dogu Akdenizde Avrasya- Geology, 4, 661-4. Afrika Leuha Sinirinin Ozellikleri, 75 pp. BEN-MENAnEM, A. 1976. Tectonics, seismicity and Technical University, Istanbul. structure of the Afro-Eurasian junction -- the BYERLEE, J. D. 1968. Brittle-ductile transition in breaking of an incoherent plate. Phvsics earth rocks. J. geophys. Res. 73, 4741-50. Planet. Interiors', 12, 1-50. CANITEZ, N. 1962. Gravite ve sismolojiye gore BERBERIAN, M. 1976. Seismotectonic map of Iran, Kuzey Anadoluda Arz kabugunun yapisi. 1TU scale 1/2,500,000. In: BFRBERIAN, J. (ed.) Maden Fak. Yay., lstanbul, 87 pp. Contribution to the of lran. 1I. __ 1973. Yeni Kabuk hareketlarine iliskin calisma- Iran Geol. Surv. Rep. 39, 516 pp. lar ve Kuzey Anadolu Fay problemi: Kuzey BERGOUGNAN, H. 1976. Structure de la Chaine Anadolu Fayi ve Deprem Kusagi Simpozyumu. pontique dans le Haut-Kelkit (Nord-Est de MTA Yayinl., Ankara, 12, 35-58. l'Anatolie). Bull. Soc. gdol. France, 18,675-86. 1971. Focal mechanism and source depth of __, FOURAUIN, C. & RICOU, L. 1978. Les deux earthquakes from body- and surface-wave data. troncons et le double jeu de la faille nord- Bull. Seism. Soc. Am. 61, 1369-79. anatolienne. C.r. Acad. Sci., Paris, 287, __ & EZEN, U. 1973. Slip rate and stress drop 1183-6. along the North Anatolian Fault. In: Symposium Downloaded from http://sp.lyellcollection.org/ by guest on October 6, 2021

Neotectonics of E Anatolia 33

on Earthquake Statistics. UNESCO Balkan ducts of continental collision, J. Geol. 18, Project, Istanbul. 683-92.

__ & UCER, S. B. 1967a. Computer determinations __ & KIDD, W. S. F. 1974. Continental collisions for the fault-plane solutions in and near in the Appalachian/Caledonian : Anatolia. Tectonophysics, 4, 235-44. variations related to complete and incomplete __ & __ 1967b. A Catalogue of Focal Mechanism suturing. Geology, 2, 543-6. Diagrams for Turkey and Adjoining Areas. , PITMAN III, W. C., RYAN, W. B. F. & I. T. U. Maden Fak. Arz Fiziji Enst. Yay. 25, 111 BONNIN, J. 1973. Plate tectonics and the evolu- PP. tion of the Alpine System. Bull. Geol. Soc. Am. CHEN, R. & MOLNAR, P. 1981. Constraints on 84, 3137-80.

velocity structure beneath the __ & SEN6OR, A. M. C. 1979. Aegean and sur- Tibetan plateau and their tectonic implications, rounding regions: complex multi-plate and con- J. geophys. Res. 86, 5937-62. tinuum tectonics in a convergent zone. Bull.

__ & __ 1983. Focal depths of intracontinental geol. Soc. Am. 90, 84-92. and intraplate earthquakes and their implication __ • SHACKLETON, R. M. 1984. A model for the for the thermal and mechanical properties of the evolution of the Grampian Tract in the early lithosphere. J. geophys. Res. 88, 4183-214. Caledonides and Appalachians. Nature, Lond. CHINNERY, M., 1969. Theoretical fault models. In: A 312, 115-21. Symposium on Processes in Focal Region. Publ. __ & WINDLEY, B. F. 1981. Growth and differen- Dem. Obs., Ottawa, Canada. 7, 211-23. tiation of the continental crust. Phil. Trans. R. CONEY, P. J., JONES, D. L. & MONGER, J. W. 1980. Soc. London, A301, 189-206. Cordilleran suspect terranes. Nature, Lond. 288, DEWEY, J. W. 1976. Seismicity of northern Anatolia. 329-33. Bull. Seism. Soc. Am., 66, 843-68. CooK, F. A., ALBAUGH, D. S., BROWN, L. D., DURR, S., ALTHERR, R., KELLER, J., OKRUSCH, M. KAUFMAN, S., OLIVER, J. E. & HATCHER, & SEIDEL, E. 1978. The median Aegean crystal- R. D. 1979. Thin-skinned tectonics in the crystal- line belt: , structure, metamorphism, line southern Appalachians; COCORP seismic- magmatism. In: CLoss, H. (ed.) Mediterranean reflection profiling of the Blue Ridge and Pied- Orogens, pp. 48-72. E. Schweizerbart'sche mont. Geology, 7, 563-7. Vertlagsbach-handlung, Stuttgart. CROWELL, J. A. 1974. Origin of late basins EATON, J. P., LEE, W. H. K. & PAKISER, L. C. in southern California. Spec. Publ. Soc. econ. 1970. Use of microearthquakes in the study of Palaeont. Miner., Tulsa, 22, 19(i)-240. the mechanics of earthquake generation along CUMMINGS, D. 1976. Theory of plasticity applied to the San Andreas Fault in Central California. faulting, Mojave Desert, southern California. Tectonophysics, 9, 259-82. Bull. geol. Soc. Am. 87, 720-4. ENGLAND, P. C. 1978. Some thermal considerations DAS, S. & SCHOLZ, C. H. 1983. Why large earth- of the Alpine metamorphism- past, present quakes do not nucleate at shallow depths. and future, Tectonophysics, 46, 21-40.

Nature, Lond. 305, 621-3. __ 1983. Constraints on the extension of contin- DEGENS, E. T. & KURTMAN, F. 1978. The geology ental lithosphere. J. geophys. Res. 88, 1145-52.

of Lake Van. MTA Yaginl. Ankara, 169, 158 -- & MCKENZlE, D. P. 1982. A thin viscous sheet PP. model for continental deformation. Geophys. J. R. DELAUNE-MAYERE, M., MARCOUX, J., PARROT, astr. Soc. 70, 295-321. J. F. & POISSON, A. 1977. Modele d'evolution ERGIN, K., GUSLU, U. & Uz, Z. 1967. A catalogue Mesozoique de la paleo-marge tethysienne au of earthquakes of Turkey and surrounding area niveau des nappes radiolaritiques et ophioliti- (11 AD to 1964 AD). Maden Fak. Arz Fizigi ques du Taurus Lycien, d' et du Baer- Enst. Yayinlari, Istanbul, 24, 1-28. Bassit: In: BIJU-DUvAL, B. & MONTADERT, L. ERINC, S., 1973. Turkiyenin sekillenmesinde neotek- (eds) Structural History of the Mediterranean tonigin rolu ve jeomorfoloji-jeodinamik iliskileri. Basins. pp. 79-94. Editions Technip, Paris. Cumburiyetin 50. Yili yerbilimleri Kongresi DEWEY, J. F. 1969. Structure and sequence in the 1-12. paratectonic Caledonides. Amr. Ass. petrol. GANSSER, A. 1968. The Insubric Line, a major geo- geol. Mem. 12. 309-35. tectonic problem. Schweiz. Miner. Petrogr.

__ 1975. Finite plate evolution: some implications Mitt. 48, 123-43. for the evolution of rock masses at plate 1974. The ophiolitic m61ange, a world-wide margins. Am. J. Sci. 275-A, 260-84. problem on Tethyan examples. Eclog. Geol. 1977. Suture zone complexities: a review. Helv. 67, 479-507. Tectonophysics, 40, 53-67. GEISER, P. & ENGELDER, T. 1983. The distribution

__ 1980. Episodicity, sequence, and style at con- of large parallel shortening fabrics in the vergent plate boundaries. Geol. Ass. Can. Spec. Appalachian foreland of New York and Pen- Pap. 2, 553-76. nsylvania. Mem. geol. Soc. Am. 158, 161-75.

__ 1982. Plate tectonics and the evolution of the HALL, R 1976. Ophiolite emplacement and the evolu- . J. geol. Soc. London, 139, tion of the Taurus Suture Zone, south-eastern 371-412. Turkey. Bull. geol. Soc. Am. 87, 1078-88. __ & BURKE, K. C. A. 1973. Tibetan, Variscan HANCOCK, P. L. & BARKA, A. A. 1981. Opposed and Precambrian basement reactivation: pro- shear senses inferred from neotectonic meso- Downloaded from http://sp.lyellcollection.org/ by guest on October 6, 2021

34 J.F. Dewey et al.

fracture systems in the North Arabian Fault Turkiye Jeoloji Kurumu Bulteni, 10, 1-2. Zone. J. struct. Geol. 3, 383-92. 1969. Ober die nordanatolische Horizontal- __, AL KADHI, A. & SHA'AT, N.A. 1984. Regional verschiebung. Bull. miner. Res. Explor. Inst. sets in the Arabian Platform as indicators Ankara, 72, 1-28.

of intraplate processes. Tectonics, 3, 27-43. __ 1970. Bati Anadolu Neojen havzalarinin sismo- HANKS, T. C. & WYss, M. 1972. The use of body- tektonik durumu. Gediz Depremi Zimopozyumu, wave spectra in the determination of seismic- Tebligler,Tanismalar, Insaat Muhendisleri Odasi, source parameters. Bull. Seism. Soc. Am. 62, Ankara, pp. 8-16. 561-89. __ 1976. San Andreas ve Kuzey Anadolu Faylari HEMPTON, M. R. 1980. Structure and morphology of arasinda bir karsilastirma. Turkiye Jeoloji the East Anatolian Transform Fault Zone near Kurumu Bulteni, 19, 149-54. Lake Hazar. GSA Annual Meeting Abs. with & ROASLI, F. 1953. Maskroseismische Unter- Programs, 12, 445. suchungen uber das nordwest-anatolische Beben 1982. The North Anatolian Fault and com- vom 18. Marz 1953. Eclog. Geol. Helv. 46, plexities of continental escape. J. struct. Geol. 187- 208. 4, 502-4. KIRBY, S. H. 1980. Tectonic stresses in the litho- & DEWEY, J. F. 1983. Earthquake-induced sphere: constraints provided by the experimental deformational structures in young lacustrine deformation of rocks. J. geophys. Res. 85, sediments, East Anatolian Fault, southeast 6353-63. Turkey. Tectonophysics, 98, 7-14. LAUBSCHER, H. P. 1971. Das Alpen-Dinariden __, DUNNE, L. A. & DEWEY, J. F. 1983. Sedimen- Problem und die Palinspastik der sudichen tation in a modern strike-slip basin, southeastern Tethys. Geol. Rdsch. 60, 813-33. Turkey. J. Geol. 91, 401-12. LUYENDYK, B. P., KAMERLING, M. J. & TERRES, R.

__ & SENGOR, A. M. C. 1980. The East Anatolian 1980. Geometrical model for Neogene crustal transform fault: its age offset and significance in reactions in southern California. Bull. geol. Soc. the neotectonics of the eastern Mediterranean. Am. 91, 211-8. EOS, 61, 360. LYNN, H. B., HALE, L. D. & THOMPSON, G. A. HOUSEMAN, G. A., MCKENZIE, D. P. & MOENAR, 1981. Seismic reflections from the basal contacts P. 1981. Convective instability of a thickened of batholiths. J. geophys. Res. 86, 10633-8. boundary layer and its relevance for the thermal MATTAUER, M., PROUST, F. & TAPPONNIER, P. evolution of continental convergent belts. 1972. Major strike-slip fault of late Hercynian J. geophys. Res. 86, 6115-35. age in Morocco. Nature, Lond. 237, 160-2. Hs/2, K. J., NACHEV, I. K. & VUCHEV, g. T. 1977. MCKENZIE, D. P. 1972. Active tectonics of Mediter- Geologic evolution of Bulgaria in light of plate ranean region. Geophys. Jl, 18, 1-32. tectonics. Tectonophysics, 40, 245-56. 1972. Active tectonics of the Mediterranean HURON, D. H. W. 1979. The strain history of a Region. Geophys. J. R. astr. Soc. 30, 109-85.

Dalradian slide using pebbles with low fluctua- __ 1976. The East Anatolian Fault: a major struc- tions in axis orientation. Tectonophysics, 55, ture in eastern Turkey. Earth planet. Sci. Lett. 261-73. 29, 189-93. INNOCENTI, P., MAZZUOLI, R., PASQUARE, G., __ & JACKSON, J. A. 1983. The relationship be- RADICATI Ol BROZOLO, F. & VILLARE, L. 1976. tween strain rates, crustal thickening, palaeo- Evolution of the volcanism in the area of interac- magnetism, finite strain and fault movements tion between the Arabian, Anatolian and Iranian within a deforming zone. Earth planet. Sci. Lett. Plates (Lake Van, eastern Turkey). J. Volcan. 65, 182-202. Geotherm. Res. 1, 103-12. MEISSNER, R. & STREHLAU, J. 1982. Limits of JACKSON, J. A. 1980. Reactivation of basement faults stresses in continental crust and their relation- and crustal shortening in orogenic belts. Nature, ship to the depth-frequency of shallow earth- Lond. 283, 343-6. quakes. Tectonics, 1, 73-89.

__ & MCKENZIE, D. P. 1983. Active tectonics of the MINSTER, J. B., JORDAN, T. H., MOENAR, P. & Alpine-Himalayan Belt between western Turkey HAINES, E. 1974. Numerical modelling of in-

and Pakistan. Geophys. J. R. astr. Soc. 77, stantaneous plate tectonics. Geophys. J. R. 185-264. Astr. Soc. 36,541-76. KARNER, G. D. & WATrS, A. B. 1983. Gravity MOHAJER-ASHJA1, A., BEHZADI, H. & BERBERIAN, anomalies and flexure of the lithosphere at moun- M. 1975. Reflections on the rigidity of the Lut tain ranges. J. geophys. Res. 88, 449-60. block and recent crustal deformation in eastern KETIN, I., 1948. Uber die tektonisch-mechanischen Iran. Tectonophysics, 25,281-301. Folgerungen aus den grossen anatolischen MOLNAR, P. & CHEN, W. R. 1983. Focal depths and Erdbeben des letzten Dezenniums. Geol. Rdsch. fault plane solutions of earthquakes under the 36, 77-83. Tibetan plateau. J. Geophys. Res. 88, 1180-96. 1957. Kuzey Anadolu Deprem. Fayi. ITU __ & TAPPONNIER, P. 1975. Cenozoic tectonics of Dergisi, 15, 49-52. Asia: effects of a continental collision. Science, 1966a. Tectonic units of Anatolia. Bull. miner. 189,419-26.

Res. Explor. Inst. Ankara, 66, 23-34. __ & 1977a. Relation of the tectonics of

__ 1966b. 6 Ekim 1964 Manyas depremi espasinda eastern China to the India-Eurasia collision: zeminde meydana gelen tansiyon catlaklari. application of slip line field theory to large-scale Downloaded from http://sp.lyellcollection.org/ by guest on October 6, 2021

Neotectonics of E Anatolia 35

continental tectonics. Geology, 5,212-6. ROEDER, D. 1979. Continental collisions. Rev.

__ & __ 1977b. Active faulting and tectonics in Geophys. Space Phys. 17, 1098-109. China. J. geophys. Res. 82, 2905-30. VON ROTE, F. & STE1N1NaER, F. F. 1983. Vom Zerfall & . 1978. Active tectonics of Tibet. J. der Tethys zu Mediterran und Paratethys. Ann. geophys. Res. 83, 5361-75. Naturhist. Mus. Wien, 85/A, 135-63.

__ & 1979. Active faulting and Cenozoic ROYDEN, L., HORVATH, F. & RUMPLER, J. 1983. tectonics of the Tien Shan, Mongolia and Bayka Evolution of the Pannonian Basin System. I. regions. J. geophys. Res. 84, 3425-59. Tectonics. Tectonics, 2, 63-90.

__ & __ 1981. A possible dependence of tecto- RYBACH, L., MUELLER, S., MILNES, A. G., nic strength on the age of the crust in Asia. ANSORGE, J., BERNOULLI, D. & FREY, M. 1980. Earth planet. Sci. Lett. 52, 107-14. The Swiss Geotraverse Basel-Chiasser -- a MTA 1962. Explanatory Text of the Geological Map review Eclog. Geol. Helv. 73,437-62. of Turkey, 1/500,000, 75 pp. MTA Enstitusu, SALOMON-CALVl, W. 1936a. Die Fortsetzung der Ankara. Tonalelinie in Kleinasien. Yuksek Ziraat Enst. NI, J. • BARAZANGI, M. 1984. Seismotectonics of Calismalarindan, 9, 11-3.

the Himalayan collision zone: geometry of the __ 1936b. Die Entstehung der anatolischen Ova. underthrusting beneath the Yuksek Ziraat Enst. Calismalarindan, 30, 11 pp. Himalayas. J. geophys. Res. 89, 1147-63. 1940. Anadolu'num tektonik tarzi tesekkulu

__ & YORK, J. E. 1978. Late Cenozoic tectonics of hakklinda kisa izahat. MTA Mecmuasi, 18, the Tibetan Plateau. J. geophys. Res. 83, 35-47. 5377-48. SCHOLZ, C. H. 1977. Transform fault systems of NOWACK, E. 1928. Die wichtigsten Ergebnisse California and : similarities in their meiner anatolischen Reisen. Z dt geol.Ges. 80, tectonic and seismic styles. J. geol. Soc. London, B304-12. 133,215-29. OZGUL, N. 1976. Toroslar' in bazi temel jeoloji SCLATER, J. G. & FRANCHETEAU, J. 1970. The ozellikleri. Tuirkiye Jeoloji Kurumu Bulteni, 19, implications of terrestrial heat flow observations 65-78. on current tectonic and geochemical models of PAMm, R. N. 1950. Les seismes en Asie Mineure the crust and upper mantle of the earth. entre 1939 et 1944. La Cicatrice Nord- Geophys. J.R. astr. Soc. 20, 509-42. Anatolienne. Proc. Int. Gbol. Congr. Lond. 13, SENGOR, A. M. C., 1976. Collision of irregular con- 214-8. tinental margins: implications for foreland __ & AKYOL, I. H. 1943. Corum ve Erbaa depre- deformation of Alpine-type orogens. Geology, mieri. Turk Cografya Dergisi, 2, 1-7. 4, 779-82.

__ & KETIN, I. 1940. Das Erbdeben in der Turkei __ 1977. Post-Miocene tectonic evolution of eastern vom 27 und 28 Dezember 1939. Geol. Rdsch. Turkey: initial stages in the formation of a 31, 77-8. Tibetan-type high plateau. Geol. Soc. Am.,

__ & __ 1941. Das anatolische Erdbeben Ende Abst. Prog. 9, 1168-76.

1939. Geol. Rdsch. 32,279-87. __ 1979. North Anatolian Fault: its age offset and PAREJAS, E., AKYOL, I. H. & ALT1NII, E. 1942. Le tectonic significance. J. geol. Soc. London, 136, tremblement de terre d'Erzincan due 269-82. 27 Decembre 1939. Revue Fac.Sci. Univ. __& KIDO, W. S. F. 1979. Post-collisional tecto- Istanbul, BVI, 177-222. nics of the Turkish-Iranian plateau and a com- PAVONI, N. 1961a. Die Nordanatolische Horizontal- parison with Tibet. Tectono'physics, 55,361-76. verschiebung. Geol. Rdsch. 51,122-39. SEYMEN, I. 1975. Kelkit vadisie kesiminde Kuzey

__ 1961b. Faltung durch Horizontalverschiebung. Anadolu Fay Aonunun tektonik ozelligi, 192 pp. Eclog. geol. Helv. 54, 515-34. ITU Maden Fak Yay, Istanbul.

PFIFFNER, O. A. & RAMSAY, J. G. 1982. Constraints __ & AYDIN, A. 1972. The Bingol Earthquake on geological strain rates, arguments from finite Fault and its relation to the North Anatolian strain states of naturally deformed rocks. Fault Zone. Bull. miner. Res. Explor. Inst. J. geophys. Res. 87,311-21. Ankara, 79, 1-8. POWELL, C. McA. & CONAGHAN, P. J. 1973. Plate SIBSON, R. H. 1983. Continental fault structure and tectonics and the Himalayas. Earth planet. Sci. the shallow earthquake source. J. geol. Soc. Lett. 20, 1-20. London, 140,741-67. PRESCOtt, A. & NUR, A. 1981. The accommodation SIEH, K. E. 1978. Prehistoric large earthquakes of relative motion at depth on the San Andreas produced by slip on the San Andreas Fault at Fault System in California. J. geophys. Res. 85, Pallet Creek, California. J. geophys. Res. 83, 999-1004. 3907-39. RAMSAY, J. G. & WOOD, D. S. 1976. A discussion SIMS, J. D. 1973. Earthquake-induced structures in on natural strain and geological structure. Phil. sediments of Van Norman Lake, San Fernando, Trans. R. Soc. London, A283, 344 pp. California. Science, 182, 161-3. RICOU, L. E., ARGYRIADIS, I. & MARCOUX, J. 1975. 1975. Determining earthquake recurrence L'Axe Calcaire du Taurus, un alignment de intervals from deformational structures in young fenetres araboafricaines sous des nappes radio- lacustrine sediments. Tectonophysics, 29, 141- laritiques, ophiolitiques et metamorphignes. 52. Bull. Soc. geol. France 17, 1024-44. SWAN III, F. H., SCHWARTZ, D. P. & CLUFF, L. S. Downloaded from http://sp.lyellcollection.org/ by guest on October 6, 2021

36 J.F. Dewey et al.

1980. Recurrence of moderate to large magni- Kusagi Simpozyumu, pp. 12-29. MTA Yayini, tude earthquakes produced by surface faulting Ankara. on the Wasatch Fault Zone, Utah. Bull. seismol. ToKsoz, M. N., ARPAT, E. & SAROGLU, F. 1977. Soc. Am. 70, 1431-62. East Anatolian earthquake of 24 November 1976. TAPPONNIER, P. 1977. Evolution tcctoniquc du Nature, Lond. 270, 423-5. Systeme Alpin en Medtierran6e: poinconnement TRUMPY, R. 1973. The timing of orogenic events in et ecrasement rigide-plastique. Bull. Soc. g(ol. the Central Alps. In: DEJONG, K. A. & France, 19, 437-60. SCHOLTEN, R. (eds) Gravity and Tectonics, pp.

__ & MOLNAR, P. 1976. Slip-line field theory and 229-251. John Wiley and Sons, New York. large-scale continental tectonics. Nature, Lond. U~UR, E. 1974. Recent crustal movements in the 264, 319-21. Gerede-Cerkes region of the North Anatolian

__ & __ 1977. Active faulting and tectonics in Fault Zone by Geodetic Methods. Ph.D. thesis, China. J. geophys. Res. 82, 2905-30. I.T.U., Istanbul. __, MERCIER, J. L. ARMIJO, R. HAN TONGI.IN & WALLACE, R. E. 1968. Earthquake of August 19, ZHOU, J. 1981. Field evidence for active normal 1966, Varto area, eastern Turkey. Bull. Seism faulting in Tibet. Nature, Lond. 294, 410-4. Soc. Am. 58, 11-45.

__, PELTZER, G., LE DAIN, A. Y., ARMIJO, R. ~¢ __ 1977. Time-history analysis of fault scarps and CO~BOLI), P. 1982. Propagating extrusion fault traces -- a longer view of seismicity. In: tectonics in Asia: new insights from simple Proceedings of the 6th World Conference on experiments with plasticene. Geology, 10, 611- , New Delhi, India, 1, 6. 766-9.

TATAR, Y. 1975. Tectonic structures along the North __ 1978. Patterns of faulting and seismic gaps in Anatolian Fault Zone, north-east of Refahiye the Great Basin province. In: Proceedings of (Erzincan). Tectonophysics, 29, 401-10. Conference VI, Methodology for Identifying TCHALENKO, J. S. 1977. A reconnaissance of the Seismic Gaps and Soon-to-break Gaps. U.S. seismicity and tectonics at the northern border of geol. Survey Open File Rep. 78-943, 858-68. the Arabian Plate (Lake Van Region). Rev. YODER, H. S. & TILLEY, C. E. 1962. Origin of Geogr. Phys. Geol. Dyn. 19, 189-208. basalt : an experimental study of natural TOKAY, M. 1973. Kuzey Anadolu Fay Zonunun and synthetic rock systems. J. Petrol. 3, 342- Gerede ile Ilgaz arasindaki kisminda jeolojik 532. gozlemler. Kuzey Anadoluy Fayi ve Deprem

J. F. DEWEY, Department of Earth Sciences, University of Oxford, Oxford OX1 3PR, UK. M. R. HEMPTON, Bellaire Research Center, Shell Development Company, , , U.S.A. W. S. F. KIDD, Department of Geological Sciences, State University of New York, Albany, New York, U.S.A. F. SAROGLU, Maden Tetkik ve Arama, Ankara, Turkey. A. U. C. ~ENGOR, Faculty of Geology, Istanbul Technical University, Istanbul, Turkey.