1. Introduction
The Bitlis–Zagros orogenic belt of western Asia and the related wide area of deformation within the European foreland to the north (Fig. 1) are regarded as one of the best examples of ongoing continental collision in the world. Present-day reduction of surface area within the collision zone is estimated at 31 × 103 km2 Ma−1 (Reilinger et al. Reference Reilinger, McClusky, Vernant, Lawrence, Ergintav, Cakmak, Ozener, Kadirov, Guliev, Stepanyan, Nadariya, Hahubia, Mahmoud, Sakr, Arrajehi, Paradissis, Al-Aydrus, Prilepin, Guseva, Evren, Dmitrotsa, Filikov, Gomez, Al-Ghazzi and Karam2006). Most of the decrease in surface area is being accommodated by coherent lateral transport of Anatolia out of the collision zone (c. 70%) and by shortening along the Bitlis–Zagros and Greater Caucasus orogenic wedges (c. 15%). The remaining decrease in surface area is distributed across the Anatolian–Iranian plateau and the Lesser Caucasus (Fig. 2).
Figure 1. Tectonic map of Asia Minor and Transcaucasia (modified from Sosson et al. Reference Sosson, Rolland, Müller, Danelian, Melkonyan, Kekelia, Adamia, Babazadeh, Kangarli, Avagyan, Galoyan, Mosar, Sosson, Kaymakci, Stephenson, Bergerat and Starostenko2010).
Figure 2. Digital elevation model of the Eastern Mediterranean and the Middle East. The arrows indicate the GPS-derived velocities with respect to a stationary Eurasia (modified from Reilinger et al. Reference Reilinger, McClusky, Vernant, Lawrence, Ergintav, Cakmak, Ozener, Kadirov, Guliev, Stepanyan, Nadariya, Hahubia, Mahmoud, Sakr, Arrajehi, Paradissis, Al-Aydrus, Prilepin, Guseva, Evren, Dmitrotsa, Filikov, Gomez, Al-Ghazzi and Karam2006; Copley & Jackson, Reference Copley and Jackson2006); the dots indicate the epicentres of earthquakes M > 4.8 (depth of hypocentres: orange dots 0–33 km; yellow dots: 33–70 km) (1973–2012 data from USGS/NEIC PDE online catalogue).
The age of the initial collision between Arabia and Eurasia has been the topic of much debate, with estimates of Late Cretaceous (Hall, Reference Hall1976; Berberian & King, Reference Berberian and King1981; Alavi, Reference Alavi1994), Late Eocene – Oligocene (35–25 Ma, Jolivet & Faccenna, Reference Jolivet and Faccenna2000; Agard et al. Reference Agard, Omrani, Jolivet and Mouthereau2005; Vincent et al. Reference Vincent, Morton, Carter, Gibbs and Barabadze2007; Allen & Armstrong, Reference Allen and Armstrong2008), Miocene (Şengör, Görür & Şaroğlu, Reference Şengör, Görür, Şaroğlu, Biddle and Christie-Blick1985; Dewey et al. Reference Dewey, Helman, Turco, Hutton, Knott, Coward, Dietrich and Park1989; Yılmaz, Reference Yilmaz1993; Robertson et al. Reference Robertson, Parlak, Rizaoğlu, Ünlügenç, İnan, Tasli, Ustaömer, Ries, Butler and Graham2007) and Pliocene (Philip et al. Reference Philip, Cisternas, Gvishiani and Gorshkov1989; Avdeev & Niemi, Reference Avdeev and Niemi2011). The only available low-temperature thermochronological dataset for the Bitlis collision front points to an episode of fast exhumation in the Middle Miocene (Okay, Zattin & Cavazza, Reference Okay, Zattin and Cavazza2010), in agreement with the stratigraphy of the adjacent foreland basins.
In this paper, the first low-temperature thermochronological dataset for the Eurasian foreland north of the Bitlis collision zone suggests that the tectonic stresses related to the Arabian indentation were transmitted efficiently over large distances, focusing preferentially at rheological discontinuities along the eastern Black Sea coast and in the Lesser Caucasus. Since the late Middle Miocene a new tectonic regime has been active, as the westward translation of Anatolia (e.g. McKenzie, Reference McKenzie1972; Dewey & Şengör, Reference Şengör1979; Şengör, Reference Şengör1979; Şengör et al. Reference Şengör, Tüysüz, İmren, Sakinç, Eyidoğan, Görür, Le Pichon and Rangin2005) is accommodating most of the Arabia–Eurasia convergence, thus precluding efficient northward stress transfer.
2. Geological setting
The study area represents the region of maximum indentation between Arabia and Eurasia (Fig. 1). From north to south, four major geological provinces are present: (i) the eastern Black Sea (EBS), (ii) the eastern Pontides (EP) portion of the Sakarya Zone, (iii) the Anatolide–Tauride block (ATB), and (iv) the Arabian platform (AP).
(i) The more than 2000 m deep EBS is partly floored by quasi-oceanic crust and represents the remnant of a composite Paleocene – Middle Eocene back-arc basin, which developed on the Eurasian upper plate during N-dipping subduction of the Neotethys (e.g. Spadini, Robinson & Cloetingh, Reference Spadini, Robinson, Cloetingh and Robinson1997; Stampfli & Borel, Reference Stampfli, Borel, Cavazza, Roure, Spakman, Stampfli and Ziegler2004; Edwards et al. Reference Edwards, Scott, Shillington, Minshull, Brown and White2009).
(ii) The EP are the easternmost segment of a W–E-trending composite mountain belt traceable for more than 1200 km from Thrace to the Adjara–Trialeti region of the Lesser Caucasus of Georgia (Fig. 1). The EP are part of the Sakarya Zone, a continental fragment of Laurasian affinity (Okay & Tüysüz, Reference Okay, Tüysüz, Durand, Jolivet, Horváth and Séranne1999; Cavazza et al. Reference Cavazza, Federici, Okay and Zattin2012).
(iii) The ATB forms the bulk of southern Turkey (Fig. 1) and can be traced to the east in Transcaucasia and Iran. In contrast to the Pontides, the ATB shows a stratigraphy similar to the AP (Okay & Tüysüz, Reference Okay, Tüysüz, Durand, Jolivet, Horváth and Séranne1999). The Palaeogene İzmir–Ankara–Erzincan suture marks the boundary between the ATB and the Sakarya Zone to the north.
(iv) The southern portion of the study area is characterized by the Gondwanian terranes of the AP flexurally bent towards the Bitlis–Zagros orogenic front to the north.
3. Apatite fission-track data and thermal modelling
We collected samples for apatite fission-track (AFT) analysis across a wide swath of territory from the EP and Adjara–Trialeti region to the north to the Bitlis collision zone to the south. The samples were taken from a variety or rock types, comprising Cretaceous and Palaeogene granitoids, gneisses and metasandstone of the Bitlis and Pütürge massifs, and deeply buried Palaeogene sandstones (Table 1). Very few localities were suitable for sampling in the eastern Anatolian Plateau since most of this region is covered by a thick pile of Plio-Quaternary volcanic and volcaniclastic rocks. Procedures for sample preparation and analysis are those described in Zattin et al. (Reference Zattin, Landuzzi, Picotti and Zuffa2000). Apatite grains from 60 samples were sent for irradiation. However, only 26 samples yielded apatite grains suitable for fission-track analysis.
Table 1. Apatite fission-track data
MCTL – mean confined track length. Central ages calculated using dosimeter glass CN5 and ζ-CN5 = 336.34 ± 16.24 (analyst I. Albino). ρs – spontaneous track densities (× 105 cm−2) measured in internal mineral surfaces; Ns – total number of spontaneous tracks; ρi and ρd – induced and dosimeter track densities (× 106 cm−2) on external mica detectors (g = 0.5); N i and N d – total numbers of tracks; P(χ2) – probability of obtaining χ2-value for ν degrees of freedom (where ν = number of crystals − 1); a probability > 5% is indicative of a homogenous population. Samples with a probability < 5% have been analysed with the binomial peak-fitting method. (*) Ages from Okay, Zattin & Cavazza (Reference Okay, Zattin and Cavazza2010) (analyst M. Zattin).
Despite the lithological and age diversity, AFT results have a consistent geographic distribution, with younger ages (18–12 Ma; Early–Middle Miocene) in the Bitlis orogen and in the easternmost Pontides along the Black Sea, and Palaeogene ages in the Anatolian Plateau and Adjara–Trialeti region (Table 1; Fig. 3). There is no relationship between AFT ages and sample elevations.
Figure 3. Geographic distribution of apatite fission-track ages. See Table 1 for complete dataset. The two dotted areas include all FT ages between c. 12–15 Ma and between c. 15–20 Ma. NAF – North Anatolian Fault; EAF – East Anatolian Fault; IAES – Izmir–Ankara–Erzincan suture; KTJ – Karliova triple junction.
Modelling on samples containing a statistically significant number of confined tracks constrained further the thermochronological evolution of the study area (Fig. 4). Sample TU274 (Late Cretaceous granodioritic body of the EP magmatic arc) shows a phase of fast cooling (average cooling rate c. 22°C Ma−1) between c. 16 and 14 Ma. Considering a geothermal gradient of 25–30°C km−1, based on heat flow (Tezcan, Reference Tezcan, Gupta and Yamano1995), the average exhumation rate in the easternmost Pontides during this period of cooling is 0.7–0.9 km Ma−1. Cooling/exhumation in the EP mirrors the evolution of the Bitlis–Pütürge massif along the Arabia–Eurasia collision zone where sample TU149 (Pan-African augen gneiss) shows a rapid increase in exhumation at c. 12 Ma (Fig. 4; see also Okay, Zattin & Cavazza, Reference Okay, Zattin and Cavazza2010, p. 37).
Figure 4. Time–temperature paths obtained from inverse modelling of AFT data using the HeFTy program (Ehlers et al. Reference Ehlers, Chaudhri, Kumar, Fuller, Willett, Ketcham, Brandon, Belton, Kohn, Gleadow, Dunai and Fu2005), which generates the possible T–t paths by a Monte Carlo algorithm. Predicted AFT data were calculated according to the Ketcham, Donelick & Carlson (Reference Ketcham, Donelick and Carlson1999) annealing model and the Donelick, Ketcham & Carlson (Reference Donelick, Ketcham and Carlson1999) c-axis projection. Parameters (model and measured age, model and measured mean length) related to inverse modelling are reported. GOF (goodness-of-fit) values give an indication of the fit between observed and predicted data (values close to 1 are best). Shaded areas mark envelopes of statistically acceptable fit (GOF > 0.5) and the thick lines correspond to the most probable thermal histories. Thermal paths out of the partial annealing zone (dotted) are largely inferential as fission-track data cannot give reliable information out of this temperature range. Range and scale of x- and y-axes are identical in all diagrams to facilitate comparison.
Sample TU255 is an early Oligocene sandstone turbidite at the base of the Muş basin, a foreland basin located north of the Bitlis suture and associated with northward subduction of the Arabian plate (Hüsing et al. Reference Hüsing, Zachariasse, Van Hinsbergen, Krijgsman, Inceöz, Harzhauser, Mandic, Kroh, van Hinsbergen, Edwards and Govers2009). Following deposition, this sample was progressively buried and entered the apatite partial annealing zone (PAZ) at about 23 Ma. A rapid phase of cooling/exhumation began at 19 Ma (late Early Miocene), likely the result of the progressive incorporation of the basin southern margin into the growing Bitlis orogenic wedge. Post-depositional burial of sample TU255 was not deep enough to completely erase the thermochronological record of the sediment source rocks, showing a Late Cretaceous – Palaeogene episode of cooling/exhumation correlatable with widespread deformation in the area related to the closure of the İzmir–Ankara–Erzincan ocean (Okay & Tüysüz, Reference Okay, Tüysüz, Durand, Jolivet, Horváth and Séranne1999).
Sample TU279 (Eocene granodiorite intruding volcanic/volcaniclastic rocks in the Adjara–Trialeti zone of western Georgia) shows very rapid cooling at 36–35 Ma (latest Eocene), in line with thermochronological data from the western Greater Caucasus (Vincent et al. Reference Vincent, Carter, Lavrishchev, Rice, Barabadze and Hovius2011). The sample then underwent progressive heating during most of the Miocene and cooled definitively outside the apatite PAZ (120–60°C) in the Late Miocene, likely the result of orogenic-wedge dynamics in the Adjara–Trialeti northward-verging nappe stack facing the flexural foreland basin to the north.
4. Discussion and conclusions
Our thermochronological dataset shows that exhumation of Cretaceous and Eocene granitoids along the easternmost Pontides occurred in the Middle Miocene. The notion of a discrete and relatively rapid mid-Miocene episode of exhumation/erosion in the region is also supported by unpublished fission-track data from the composite Kackar batholith (Cretaceous – Late Eocene) immediately west of our study area (R. Jonckheere, pers. comm., 2012). The previously unrecognized exhumation/erosion episode along the Black Sea coast documented here mirrors the age of maximum tectonic coupling between the Eurasian and Arabian plates along the 2400 km long Bitlis–Zagros suture zone, c. 250 km to the south: exhumation ages along the easternmost Pontides are virtually identical to those obtained by Okay, Zattin & Cavazza (Reference Okay, Zattin and Cavazza2010) along the Bitlis suture. We argue that tectonic stresses generated along the Bitlis collision zone were transmitted northward across eastern Anatolia and focused at the rheological boundary between the Anatolian continental lithosphere and the (quasi)oceanic lithosphere of the Black Sea.
Mechanical coupling of a collisional orogen and its forelands can induce far-field tectonic stresses and significant compressional structures at distances > 1500 km from a collision front (e.g. Ziegler, Cloetingh & Van Wees, Reference Ziegler, Cloetingh and Van Wees1995; Dickerson, Reference Dickerson2003). Localization of compressional deformations far from the collision zone is controlled by spatial and temporal strength variations of the lithosphere (Ziegler, Van Wees & Cloetingh, Reference Ziegler, Van Wees and Cloetingh1998; Cloetingh et al. Reference Cloetingh, van Wees, Ziegler, Lenkey, Beekman, Tesauro, Förster, Norden, Kaban, Hardebol, Bonté, Genter, Guillou-Frottier, Ter Voorde, Sokoutis, Willingshofer, Cornu and Worum2010). Passive continental margins – like the Black Sea coast of the study area – mark the largest compositional and rheological contrast within the lithosphere (Niu, O'Hara & Pearce, Reference Niu, O'Hara and Pearce2003) and are, therefore, preferential loci of deformation. Synchronous deformation at the opposing ends of the Anatolian continental plateau documented here mirrors the results of recent studies that argue for deformation at the northern margin of the Tibetan Plateau synchronous with the early stage of India–Asia collision (e.g. Yin et al. Reference Yin, Dang, Zhang, Chen and Mcrivette2008; Clark et al. Reference Clark, Farley, Zheng, Wang and Duvall2010). The area affected by faulting increased very little through time as the northern margin of Tibet was established early; deformation has propagated northward by only a minor amount during the entire period of collision. Dayem et al. (Reference Dayem, Molnar, Clark and Houseman2009) showed that with certain boundary and initial conditions, thin viscous sheet calculations can yield significant strain rates in the northern part of Tibet soon after continental collision begins. Such conditions include (i) pre-existing topography in southern Tibet, (ii) short distances between India and northern Tibet at the time of collision, (iii) the presence of strong lithospheric blocks such as the Tarim Basin, and (iv) pre-existing weaknesses in the Asian lithosphere. It is worth noting that several of these pre-conditions were met in our study area during the Arabia–Eurasia collision, as discussed in this section.
Cooling at temperatures below the apatite PAZ in the Anatolian Plateau and in the Lesser Caucasus (Adjara–Trialeti region of western Georgia) occurred instead in the Palaeogene (with a cluster of ages in the Middle – Late Eocene; Fig. 3; Table 1), coevally with the development of the İzmir–Ankara–Erzincan suture (e.g. Okay & Tüysüz, Reference Okay, Tüysüz, Durand, Jolivet, Horváth and Séranne1999). The successive uplift of the Anatolian Plateau did not exhume a new PAZ and thus is not recorded by the AFT data.
The GPS-derived velocity field for eastern Turkey, Transcaucasia and NW Iran (Fig. 2) shows that continental material north of the Bitlis suture appears to move around the oceanic lithosphere of the EBS. Vectors in eastern Anatolia point coherently to the west, defining the apparent ‘extrusion’ of the Anatolian plate, whereas east of the Karliova triple junction they show a progressive rotation to the east (McClusky et al. Reference McClusky, Balassanian, Barka, Demir, Ergintav, Georgiev, Gurkan, Hamburger, Hurst, Kahle, Kastens, Kekelidze, King, Kotzev, Lenk, Mahmoud, Mishin, Nadariya, Ouzounis, Paradissis, Peter, Prilepin, Reilinger, Sanli, Seeger, Tealeb, Toksöz and Veis2000; Reilinger et al. Reference Reilinger, McClusky, Vernant, Lawrence, Ergintav, Cakmak, Ozener, Kadirov, Guliev, Stepanyan, Nadariya, Hahubia, Mahmoud, Sakr, Arrajehi, Paradissis, Al-Aydrus, Prilepin, Guseva, Evren, Dmitrotsa, Filikov, Gomez, Al-Ghazzi and Karam2006). Similarly, the two areas are characterized by different deformation patterns. West of the triple junction the Anatolian plate is moving as a single entity bounded by the North and East Anatolian Fault systems, whereas east of it deformation is distributed along a complex system of strike-slip and thrust faults (Copley & Jackson, Reference Copley and Jackson2006; Adamia et al. Reference Adamia, Zakariadze, Chkhotua, Sadradze, Tsereteli, Chabukian and Gventsadze2011). The different deformation patterns can be explained by the different boundary conditions imposed on these two regions: westward motion of the Anatolian plate is favoured by slab retreat along the Hellenic trench (Jolivet, Reference Jolivet2001) whereas eastern Turkey and Transcaucasia are caught between the Bitlis collision zone and the rheologically stronger (quasi)oceanic crust of the Black Sea to the northwest and the Eurasian continental crust to the northeast.
The analysis of present-day crustal dynamics and the thermochronological data presented in this paper provide a comparison between short- and long-term deformation patterns for the entire eastern Anatolian–Transcaucasian region. Two successive stages of Neogene deformation of the northwestern foreland of the Arabia–Eurasia collision zone can be inferred. (i) During the Early and Middle Miocene continental deformation was concentrated along the Arabia–Eurasia (Bitlis) collision zone but tectonic stress was transferred northward across eastern Anatolia, focusing along the EBS continent–ocean rheological transition. The Black Sea (quasi)oceanic lithosphere is fundamentally stronger than the polydeformed continental lithosphere to the south and therefore represented a ‘backstop’ resisting deformation and deviating the impinging continental lithosphere (McClusky et al. Reference McClusky, Balassanian, Barka, Demir, Ergintav, Georgiev, Gurkan, Hamburger, Hurst, Kahle, Kastens, Kekelidze, King, Kotzev, Lenk, Mahmoud, Mishin, Nadariya, Ouzounis, Paradissis, Peter, Prilepin, Reilinger, Sanli, Seeger, Tealeb, Toksöz and Veis2000). (ii) Since late Middle Miocene time the westward translation of Anatolia and the activation of the North and East Anatolian Fault systems (Dewey & Şengör, Reference Şengör1979; Şengör, Reference Şengör1979; Şengör et al. Reference Şengör, Tüysüz, İmren, Sakinç, Eyidoğan, Görür, Le Pichon and Rangin2005) have reduced efficient northward stress transfer. In this new tectonic regime – still active today – most of the Arabia–Eurasia convergence has been accommodated by the westward motion of Anatolia whereas the EP have been mechanically decoupled from the foreland of the Bitlis collision zone, as shown by the absence of significant seismicity in the area (Fig. 2). The following regional-scale topographic uplift of the Anatolian Plateau has not exhumed a new PAZ and thus is not recorded by the apatite fission tracks.
Acknowledgements
Reviews by S. Vincent and an anonymous reviewer greatly improved the manuscript. W. Cavazza and I. Albino were supported by a grant from MIUR (Italian Ministry of University and Research). A. I. Okay was supported by a grant from TÜBA (The Turkish Academy of Sciences).
