Abstract

Eastern Anatolia, Turkey, is a part of the Alpine-Himalayan collisional belt where continental crust is relatively thin for a collisional belt. The region contains part of the Zagros suture zone, which formed during collision of the Arabian and Anatolian plates in the Miocene. It is underlain by a low-velocity zone associated with asthenospheric flow in the uppermost mantle.

We constructed gravity models of the crust and upper-mantle structures to assess the driving mechanism of asthenospheric flow and the isostatic state of Eastern Anatolia. Our density models are based on terrestrial and satellite-derived gravity data, and they are constrained by receiver function and seismic tomography. The gravity models show significant lithospheric thickness variations across the Anatolian and Arabian plates. The lithospheric mantle in Eastern Anatolia is thinner (∼62–74 km) than the Arabian plate (∼84–95 km), indicating that part of the Anatolian mantle lithosphere might have been removed by delamination. The lithospheric removal process might have occurred following the detachment of the Arabian slab in the Miocene. Widespread Holocene volcanism and high heat flow in Eastern Anatolia can be considered as evidence of lithospheric delamination and slab break-off. The upward asthenospheric flow and subsequent asthenospheric underplating beneath Eastern Anatolia might have been induced by both delamination and slab break-off. These two processes may account for the rapid uplift of the Anatolian Plateau. There is a residual topography of ∼1.7 km that cannot be explained by crustal roots. Based on our gravity models, we suggest that part of the eastern Anatolian Plateau is dynamically supported by asthenospheric flow in the upper mantle.

INTRODUCTION

The Eastern Anatolia region, Turkey, with an average elevation of 2 km above sea level, is a classic example of a young continental collision zone (Fig. 1). The geodynamic evolution of the region involved major ocean closures (e.g., Şengör and Yılmaz, 1981; Okay and Tüysüz, 1999; Okay et al., 2010) due to subduction of oceanic lithosphere and continental collision of the Arabian and Eurasian plates (e.g., Keskin, 2003; Faccenna et al., 2006; Göğüş and Pysklywec, 2008; Koulakov, 2011). The complex combination of these processes resulted in the present-day crustal structure of Eastern Anatolia and surrounding regions. These structures include the Zagros fold-and-thrust belt, the north, northeastern, and east Anatolian fault zones, the Anatolian Plateau, and east Anatolia volcanic centers (Fig. 1).

During the 1970s and 1980s, crustal thickening due to continental collision was proposed by several models to explain the geodynamic evolution of the eastern Anatolian Plateau and high topography (e.g., Şengör and Kidd, 1979; Şengör and Yılmaz 1981; Dewey et al., 1986; McKenzie and Bickle, 1988). To test the proposed geodynamic models, the Eastern Turkey Seismic Experiment Project was conducted in the 1990s (Sandvol et al., 2003). The results of the project suggested that mantle lithosphere is either absent or extremely thin beneath the eastern Anatolian Plateau (Al-Lazki et al., 2003; Gök et al., 2003; Sandvol et al., 2003). The thickness of the lithospheric mantle in the Eastern Anatolian region is ∼60 km (e.g., Pearce et al., 1990; Al-Lazki et al., 2003; Gök et al., 2003; Sandvol et al., 2003). This is much less than the 100–125 km thickness of the cold and stable mantle lithosphere in the Arabian Shield and Iranian Plateau (e.g., Angus et al., 2006). Consequently, delamination and slab break-off models have been proposed to explain the thin lithosphere (e.g., Al-Lazki et al., 2003; Gök et al., 2003; Keskin, 2003; Şengör et al., 2003; Faccenna et al., 2006; Lei and Zhao, 2007; Göğüş and Pysklywec, 2008; Toksöz et al., 2010; Biryol et al., 2011; Koulakov, 2011; Fichtner et al., 2013; Bartol and Govers, 2014). The rapid topographic uplift in Eastern Anatolia between the late Miocene and early Pliocene might be attributed to the dynamic and isostatic effects of delamination, slab break-off, and a compressional regime between the Arabian and Eurasian plates (Keskin, 2003; Şengör et al., 2003; Faccenna et al., 2006; Göğüş and Pysklywec, 2008).

The upper-mantle structure of Eastern Anatolia has been imaged in various body and surface wave tomography experiments (Al-Lazki et al., 2003; Gök et al., 2003; Lei and Zhao, 2007; Toksöz et al., 2010; Biryol et al., 2011; Salaün et al., 2012; Koulakov, 2011; Fichtner et al., 2013; Delph et al., 2015). The lithosphere in Eastern Anatolia is underlain by a low-velocity zone (Pn velocity = 7.6–7.9 km/s), and this has been interpreted as anomalously hot asthenosphere in the uppermost mantle (Toksöz et al., 2010; Biryol et al., 2011; Salaün et al., 2012; Koulakov, 2011; Fichtner et al., 2013; Delph et al., 2015). The hot asthenospheric flow in the upper mantle might have affected the crustal and lithospheric structure in Eastern Anatolia (e.g., Keskin, 2003). The presence of low-velocity structures within the lower crust might be attributed to the low-velocity zone in the uppermost mantle (e. g., Pamukçu and Akçığ, 2011; Warren et al., 2013; Delph et al., 2015).

The presence of widespread Miocene to Pleistocene volcanism across the region and the existence of hot asthenospheric material in the upper mantle suggest that delamination and slab break-off might have occurred in the late Miocene (Al-Lazki et al., 2003; Gök et al., 2003; Keskin, 2003; Şengör et al., 2003; Faccenna et al., 2006; Lei and Zhao, 2007; Göğüş and Pysklywec, 2008; Toksöz et al., 2010; Biryol et al., 2011; Koulakov, 2011; Fichtner et al., 2013; Bartol and Govers, 2014). Distinguishing between the two processes is not easy, since both processes equally explain the genesis of widespread volcanism and the dynamic topography in the region. Although several studies have been carried out in Eastern Anatolia to understand the crust and upper-mantle structure, the driving mechanism of asthenospheric flow in the uppermost mantle beneath Eastern Anatolia and isostatic state are not well understood (Şengör et al., 2003; Keskin, 2007; Pamukçu and Akçığ, 2011).

In this paper, we assessed the lithospheric structure and isostatic state of Eastern Anatolia based on gravity data modeling. The main purpose of this study was twofold: (1) to determine the detailed lithospheric structure of Eastern Anatolia down to a depth of 250 km, and (2) to interpret the gravity model to determine the driving mechanism of asthenospheric flow and residual topography in the region. Our density model is based on gravity data from the European Improved Gravity Model of the Earth (EIGEN-6C4; Förste, et al., 2015). To reduce ambiguity inherent in potential field interpretations, the gravity model was constrained using results from seismic-reflection, seismic-refraction, and earthquake tomography studies in Eastern Anatolia. In this paper, we present four 2.5-dimensional (2.5-D) gravity models, representative of the eastern, central, and western parts of Eastern Anatolia, and we discuss the dynamic implications of the models.

GEOLOGIC OVERVIEW

Although Eastern Anatolia has a long geological history since the fragmentation of Rodinia in the late Proterozoic, only the neotectonic evolution of the region will be discussed here. The reader is referred to Şengör and Yılmaz (1981), Yılmaz (1993), Keskin (2007), and Şengör et al. (2008) for a comprehensive discussion of the geology of Eastern Anatolia and surrounding regions.

Neotectonic deformation in the eastern Anatolian Plateau was initiated during the Arabian-Eurasian collision in the early Miocene. The collision formed the thrust faults of the Zagros fold-and-thrust belt and initiated contraction and shortening across Eastern Anatolia due to the northward motion of the Arabian plate (e.g., Şengör and Yılmaz, 1981; Perinçek and Çemen, 1990; Yılmaz, 1993; Faccenna et al., 2006). The collision and associated plate indentation increased the accumulation of stress across Eastern Anatolia and led to the formation of the North and East Anatolian fault zones in the late Miocene (Şengör et al., 1985; Perinçek and Çemen, 1990; Çemen et al., 1992; Yılmaz, 1993; Faccenna et al., 2006). Both the North and East Anatolian faults are responsible for the westward lateral motion of the Anatolian plate (Figs. 1 and 2). Contraction and strike-slip tectonics across the region are still active (e.g., Şengör et al., 2003; Yılmaz, 2017). However, few extensional basins, probably associated with the rapid uplift of the topography and subsequent gravitational forces, are controlled by normal faults (Göğüş and Pysklywec, 2008). After the rapid uplift of the region between the late Miocene and early Pliocene, the eastern Anatolian Plateau experienced widespread Pliocene–Quaternary volcanism (Fig. 2), represented by calc-alkaline to alkaline volcanic rock sequences (Pearce et al., 1990; Keskin, 2003; Şengör et al., 2003). Geochemical studies indicate that the volcanic centers in the region consist of enriched asthenospheric material (Keskin et al., 1998; Keskin, 2003).

The eastern Anatolian Plateau consists of different tectonic units accreted during the Late Cretaceous to Early Tertiary (Fig. 2). The Eastern Rhodope–Pontide metamorphic massif is located in the northern part of Eastern Anatolia (Şengör and Yılmaz, 1981). The massif is overlain by a thick volcano-sedimentary rock sequence (Yılmaz, 1993). The Eastern Anatolian accretionary complex is located in the middle of Eastern Anatolia and trends northwest-southeast. The complex consists of remnants of a subduction-accretion complex, including a Late Cretaceous–age ophiolitic mélange and Paleogene-age flysch sequences (e.g., Şengör and Yılmaz, 1981). The flysch sequences in the north are older than the ones in the south, indicating that the area gradually became shallower from north to south (e.g., Şengör and Yılmaz, 1981; Yılmaz, 1993). The Bitlis-Poturge Massif is exposed in the southernmost portion of Eastern Anatolia (Fig. 2) and is composed of medium- and high-grade metamorphosed sediments and igneous rocks formed between Late Cretaceous and middle Eocene time (Okay and Tüysüz, 1999; Yılmaz, 1993). The Miocene Bitlis-Zagros suture zone marks the closure of the southern branch of the Neotethys Ocean (Fig. 2). Shallow-marine deposits and collision-related subaerial volcanic units are Neogene to Quaternary in age, and the volcanic units become younger from north to south/southeast (Keskin, 2003).

In Eastern Anatolia, global positioning system (GPS) measurements (Fig. 3) clearly show north-south shortening, coherent westward escape of the Anatolian plate between the North and East Anatolian fault zones, and northeast motion of the Anatolian plate northeast of the Bitlis-Zagros suture zone (Reilinger et al., 2006). Most of the shortening in the eastern Anatolian Plateau is being accommodated by lateral motion of the Anatolian plate along the North and East Anatolian fault zones (Reilinger et al., 2006, 2010). The southern part of the eastern Anatolian Plateau moves faster than the northern part, and the rate of motion gradually decreases from south to north (Reilinger et al., 2006, 2010; Şengör et al., 2008; Yılmaz, 2017).

DATA AND METHODS

Gravity Database

The gravity data used for this study are based on the European Improved Gravity Model of the Earth (EIGEN-6C4; Förste et al., 2015). The EIGEN-6C4 geopotential model is a spherical harmonic representation of the gravitational field of Earth up to degree and order of 2190 (spatial resolution ∼ 8 km). The model is constrained by terrestrial and satellite gravity data from the LAGEOS (Laser Geodynamics Satellites), GRACE (Gravity Recovery and Climate Experiment), and GOCE (Gravity Field and Steady-State Ocean Circulation Explorer) satellite missions (Förste et al., 2015). The spherical representation is based on the World Geodetic System 1984 (WGS1984) reference system.

The spatial resolution of the Earth Gravitational Models (e.g., EGM2008, EIGEN-6C4) depends on the availability of high-quality land gravity data (Köther et al., 2012; Gutknecht et al., 2014; Godin and Harris, 2014). We would like to emphasize that the land gravity data in the EIGEN-6C4 model are based on the Earth Gravitational Model (EGM2008; Pavlis et al., 2012), which includes all available land gravity data. Existing land gravity data in Eastern Anatolia are available at 2–5 km station intervals (Ates et al., 1999; Ekinci and Yiğitbaş, 2015). The land gravity data sets from Eastern Anatolia are provided to the National Geospatial-Intelligence Agency (NGA) by external organizations in Turkey without any restriction and are included in the EGM2008 (Data Repository Fig. DR11; Pavlis et al., 2012) and EIGEN-6C4 models (Förste et al., 2014). In some areas, the available land gravity data sets are included with restrictions (e.g., Himalaya). Their use is limited to a resolution corresponding to 15 arc-minute area-mean value (Fig. DR1 [see footnote 1]; Pavlis et al., 2012).

To develop the 2.5-D gravity models of the deep crust and upper-mantle structure of Eastern Anatolia, we downloaded the free air anomaly of the region between 38°E and 37°N and 44°E and 41°N from the data portal of the International Centre for Global Earth Models (ICGEM: http://icgem.gfz-potsdam.de/ICGEM/). Then, we computed the complete Bouguer anomaly of the study area using Gravity Terrain Correction code (GTeC; Cella, 2015). The complete Bouguer anomaly (Fig. 4) is based on spherical cap and terrain corrections up to a radius of 168 km. The corrections are based on elevation data from the Shuttle Radar Topography Mission (SRTM; Jarvis et al., 2008) and a standard reduction density of 2.670 g cm–3.

Initial Model and Data Constraints

In this paper, we discuss four 2.5-D gravity models, representative of the western, central, and eastern parts of the Eastern Anatolia region. The models depict the deep crust and upper-mantle structure of the Anatolian and Arabian plates at 38°E, 41°E, 42°E, and 44°E longitude and span from 37°N to 41°N (Fig. 1).

To reduce the ambiguity inherent in potential field interpretations, the densities and geometric structures of the sediments, crust, lithospheric mantle, and asthenosphere were constrained by velocity models from receiver function analysis and seismic tomography (e.g., Örgülü et al., 2003; Piromallo and Morelli, 2003; Zor et al., 2003; Reiter and Rodi, 2006; Lei and Zhao, 2007; Özacar et al., 2008; Gans et al., 2009; Toksöz et al., 2010; Biryol et al., 2011; Koulakov, 2011; Salah et al., 2011; Fichtner et al., 2013; Pasyanos et al., 2014; Delph et al., 2015).

The crustal thickness in Eastern Anatolia, as determined from receiver function and seismic tomography, ranges from 30 to 55 km (e.g., Zor et al., 2003; Angus et al., 2006; Özacar et al., 2008; Gök et al., 2011; Gökalp, 2012; Tezel et al., 2013; Vanacore et al., 2013; Pasyanos et al., 2014; Schildgen et al., 2014; Delph et al., 2015), and the sediment thickness ranges from 0 to 5 km, based on the global sediment map determined by seismic data (Laske and Masters, 1997).

Different lithospheric thickness values obtained from receiver function analysis show that the Arabian lithosphere is thicker, ranging from 75 to 160 km, than the Eastern Anatolian lithosphere, which ranges from 40 to 90 km (e.g., Angus et al., 2006; Özacar et al., 2008; Zor, 2008; Pasyanos et al., 2014; Kind et al., 2015). Low Pn velocities (7.6–7.9 km/s) dominate the uppermost mantle of Eastern Anatolia, and these are interpreted as indicative of asthenospheric underplating beneath Eastern Anatolia (Al-Lazki et al., 2003; Gök et al., 2003; Lei and Zhao, 2007; Toksöz et al., 2010; Biryol et al., 2011; Koulakov, 2011; Fichtner et al., 2013).

Table 1 shows the velocities of sediments, crust, lithospheric mantle, and asthenosphere in Eastern Anatolia, along with the corresponding densities. The density values of major tectonic units were derived from empirical relationships between P-wave velocities and densities of rocks at relevant pressure and temperature conditions (Sobolev and Babeyko, 1994; Nafe and Drake, 1957). The velocity and geometric structures were used to constrain the 2.5-D gravity models of Eastern Anatolia.

In order to determine the deep crust and upper-mantle density structure of Eastern Anatolia, we applied two modeling approaches: First, 2.5-D forward modeling of the Bouguer anomaly was performed using GM-SYS Gravity and Magnetic Modeling Software (Geosoft Oasis Montaj, 2017), which makes use of the method of Talwani et al. (1959) and Talwani and Heirtzler (1964) to calculate gravity anomalies. We would like to emphasize here that the strike lengths of the geological units in the study area are not long enough to assume 2-D modeling. The strike length of a linear structure should be at least 4–5 times the width of the geologic features for 2-D modeling. The 2-D and 2.5-D gravity modeling techniques are generally applicable to profiles nearly perpendicular to linear structures. The main difference between 2-D and 2.5-D modeling is that a 2-D model has an infinite strike length, whereas a 2.5-D model has a finite strike length.

In the second step, to improve the fit of the calculated gravity to the observed data, the density values of all tectonic units along each of the cross sections were inverted using a ridge-regression algorithm, taking all available geological information into consideration. The errors in the misfits were minimized in a least-square sense. The density values of the final gravity models and their average tolerable variations are given in Table 2.

The 2.5-D gravity models were set perpendicular to geologic strike, and the strike length was constrained by the physical limits of geologic units. The modeling space, and hence the mass therein, is a small fraction of the entire Earth. This has an effect on the forward gravity modeling, because the levels of the observed and modeled gravity values are not identical and can result in offsets between the measured and modeled gravity values. To account for the effect of the surrounding mass on the forward gravity modeling, the dimension of the modeling space was extended to 10,000 km beyond the dimension of the 2.5-D gravity models.

RESULTS AND DISCUSSION

Gravity Anomalies and Analysis

The Bouguer gravity map of Eastern Anatolia contains long- and short-wavelength gravity anomalies (Fig. 4). The gravity map reveals broad regional negative and positive Bouguer anomalies over the Anatolian and Arabian plates, respectively. A negative Bouguer anomaly as low as −219 mGal coincides with the highest elevation in Eastern Anatolia, and the anomaly increases with decreasing elevation to the NW and SW of the eastern Anatolian Plateau.

To better understand the gravity anomalies of Eastern Anatolia, we applied upward continuation and wavelength filtering. The upward continuation process enhances long-wavelength anomalies and attenuates short wavelengths. The low-pass or long-wavelength filter suppresses short-wavelength anomalies and enhances regional long-wavelength gravity anomalies. Practically, both low-pass filtering and upward continuation accentuate long-wavelength anomalies of deep origin. In the first step, the Bouguer gravity anomalies were continued upward to various heights. The purpose of this process is to help us select the cutoff wavelength for the low-pass filtering. We found that most of the short wavelengths (<200 km) were attenuated at a height of 50 km above the surface. The gravity anomalies at heights >50 km are regional in nature. In the second step, the Bouguer gravity anomalies were filtered using a low-pass filter for various cutoff values. The cutoff wavelength (300 km) was selected based on similarity of amplitudes of gravity anomalies obtained using the upward continuation and wavelength filtering methods.

The regional gravity anomaly map of Eastern Anatolia (Fig. 5) does not suggest direct spatial correlation between the locations of volcanic centers and the broad long-wavelength regional gravity anomaly, although the centers are associated with rapid asthenospheric upwelling beneath Eastern Anatolia. This implies that the regional gravity anomalies of Eastern Anatolia are mainly due to density variations in the deep crust and upper-mantle structures.

Three distinct positive and negative anomalies can be identified on the regional gravity map (Fig. 5): The negative anomaly in the northeastern part of the map area (north of the Bitlis-Zagros suture zone) correlates with the eastern Anatolian Plateau, and the two positive anomalies in the north and southwestern sections of the map coincide with Black Sea and Arabian plates, respectively. The transition from high- to low-gravity anomalies (gravity gradients) in Eastern Anatolia is marked by four major tectonic structures (Figs. 4 and 5). This indicates that the deformation associated with major tectonic structures in Eastern Anatolia most probably affects the deepest part of the lower crust. The residual gravity anomalies of Eastern Anatolia, as obtained from high-pass filtering, are shown in Figure 6. The residual gravity anomalies range from 93 mGal to –64 mGal (Fig. 6) and appear to correlate inversely with the topography, indicating a crustal root beneath the eastern Anatolian Plateau that causes a low in the gravity anomaly.

Isostatic State of Eastern Anatolia

The region beneath Eastern Anatolia may not have a sufficient crustal root to explain the observed topography (cf. e.g., Zor et al., 2003; Vanacore et al., 2013). Thus, part of the eastern Anatolian Plateau may not be isostatically compensated. To assess the isostatic state and compensation mechanism of Eastern Anatolia, we determined residual topography of the region (Fig. 7) from differences between observed and calculated isostatic topography based on the Airy isostasy model. The observed topographic data were obtained from the SRTM (Jarvis et al., 2008). The isostatic topography was determined based on crustal thickness derived from receiver function analysis (Tezel et al., 2013) and the Crust1.0 model (Laske et al., 2013). We assumed a global mean crustal thickness of 31.2 km for the Airy isostatic model (Watt, 2015). The assumed crustal and mantle densities for the Airy model were 2.750 g cm–3 and 3.250 g cm–3, respectively.

The reliability of the calculated residual topography depends on the accuracy of crustal thickness and density data. To quantify the effects of density variation in the calculated residual topography and determine the corresponding uncertainty, we used a range of density values for the crust (2.7–2.8 g cm–3) and upper mantle (3.2–3.35 g cm–3). The effect of density variation of the residual topography is in the order of 50 m. We did not include the contribution of mantle lithosphere to isostatic compensation, because the available information on the depth of the lithosphere-asthenosphere boundary beneath Eastern Anatolia is limited (cf. e.g., Angus et al., 2006; Kind et al., 2015). However, this limitation did not affect the results of our analysis, because the contribution of the continental crust to isostatic compensation is more than the lithospheric mantle (Faccenna et al., 2014). This should be due to the contrasting composition between continental crust and lithospheric mantle.

The residual topography map (Fig. 7) shows negative and positive topographic anomalies (residuals) over Eastern Anatolia where negative and positive residuals indicate over- and undercompensation conditions. The high topography in Eastern Anatolia is characterized by positive residual values (∼1.7 km), indicating that the eastern Anatolian Plateau may be undercompensated. This further implies that the crustal thickness beneath the eastern Anatolian Plateau may not be sufficient to isostatically support the observed topography. This is in agreement with some of the recent studies in Eastern Anatolia (e.g., Faccenna et al., 2014; Komut, 2015). Thus, the Anatolian plateau may be partly compensated by another mechanism, most probably by asthenospheric flow in the uppermost mantle. Other possible compensation mechanisms include density heterogeneity in the crust and upper mantle, as well as thickness of the lithosphere.

Lithospheric Structure and Driving Mechanism of Asthenospheric Flow in Eastern Anatolia

Our gravity models (Fig. 8) along longitude 38°E, 41°E, 42°E, and 44°E are representative of the western, central, and eastern sectors of Eastern Anatolia and show the crust and upper-mantle structure of the Arabian and Anatolian plates. As shown in Figure 8, the long-wavelength gravity anomaly of Eastern Anatolia is well explained in terms of a thin lithosphere and anomalous asthenosphere in the uppermost mantle. There are significant variations in the lithospheric structure of Eastern Anatolia. The lithospheric mantle beneath Eastern Anatolia is thinner (∼62–74 km) than the Arabian plate (∼84–95 km), indicating that the mantle lithosphere beneath Eastern Anatolia may have been delaminated. This is in agreement with results of S-wave receiver function studies in Eastern Anatolia (Fig. 9; cf. e.g., Angus et al., 2006; Kind et al., 2015). The lithospheric instability and the subsequent delamination of the Anatolian mantle might have occurred following slab break-off in the region. Several receiver function studies have determined the presence of the detached Arabian slab at the base of the transition zone (Lei and Zhao, 2007; Özacar et al., 2008; Zor, 2008; Biryol et al., 2011). Thus, the upward asthenospheric flow and subsequent asthenospheric underplating in the uppermost mantle beneath Eastern Anatolia might be attributed to both detachment of the Arabian slab in the Miocene and delamination of the Anatolian mantle in the late Miocene.

The lithospheric delamination of the Anatolian continental mantle and slab break-off might have induced asthenospheric flow beneath Eastern Anatolia. The NE-SW–oriented fast polarization direction of the upper-mantle seismic anisotropy in Eastern Anatolia may be considered as evidence of the direction of recent asthenospheric flow (Sandvol et al., 2003; Biryol et al., 2010; Yolsal-Çevikbilen, 2014; Vinnik et al., 2016). The polarization direction is nearly parallel to plate motion and may have provided driving forces in a NE-SW direction (e.g., Vinnik et al., 2016). Thus, the asthenospheric flow and subduction in the Caucasus region may have driven northeastward plate motion. The asthenospheric flow beneath Eastern Anatolia accounts for rapid topographic uplift in the late Miocene and early Pliocene, high heat flow, and widespread Holocene volcanism. The highest heat-flow value (105 mW m–2) was observed in the Eastern Pontides orogenic belt, associated with Neogene and Quaternary volcanism (Maden and Öztürk, 2015). The exact contributions of delamination and slab break-off to the overall upward asthenospheric flow, and hence to the dynamic topography and volcanism in eastern Anatolia, is not easily determined, because the composition of magma derived from asthenospheric flow induced by delamination and slab break-off may be the same. Moreover, delamination and slab break-off could occur during and after subduction.

The crustal structure of Eastern Anatolia is segmented, much like the underlying lithospheric mantle. The crustal thickness in Eastern Anatolia increases with increasing elevation and ranges from 33 km in the west to 46 km in the east (Fig. 8), similar to previous crustal thickness estimates for the eastern Anatolian Plateau (Fig. 9 [Angus et al., 2006; Motavalli-Anbaran et al., 2016] and Fig. 10 [cf. e.g., Delph et al., 2015]). The crust becomes thicker north of the Bitlis-Zagros suture zone. This might be attributed to the ongoing collision between the Arabian and Eurasian plates. The crust might have been affected by the hot asthenospheric material in the uppermost mantle. The densities of the upper and lower crust in the central and eastern sectors of Eastern Anatolia are less than the average density of the crust in the region, indicating that the crust may be thermally weakened (Fig. 8). The receiver function analysis by Delph et al. (2015) also suggested that the shear wave velocity of the crust beneath the Anatolian plate is slower than the global average crustal shear wave velocity of 3.4 km s–1. We interpret the low densities in the upper and lower crust beneath the central and eastern sectors of Eastern Anatolia to indicate a thermally altered crust.

CONCLUSIONS

Our 2.5-D gravity models (Fig. 8) in the Eastern Anatolian region provide a possible explanation of the density and geometric structures and the driving mechanism of asthenospheric flow that resulted in the rapid topographic uplift and widespread Holocene volcanism in the region. The models, based on EIGEN-6C4 data, show that the lithospheric mantle in Eastern Anatolia is thinner (∼62–74 km) than the Arabian plate (84–95 km) north of the Bitlis-Zagros suture zone. These data support the hypothesis that the Anatolian lithosphere is delaminated and might have induced asthenospheric flow beneath Eastern Anatolia. The lithospheric instability and the subsequent delamination probably occurred following slab break-off in the region in the late Miocene. The asthenosphere probably ascended to the base of the thin lithosphere following the slab break-off and delamination. The ascension of asthenosphere accounts for the rapid topographic uplift and extensive melting that resulted in widespread Holocene volcanism across the region.

The densities of the lower crust (2.71–2.88 g cm–3) beneath the central and eastern parts of Eastern Anatolia are less than the average density of the crust in the region (2.80–2.95 g cm–3). This indicates that the crustal rocks are probably thermally affected by asthenospheric underplating. Based on our gravity models, we suggest that the hot asthenospheric material in the uppermost mantle might have induced thermal weakening in the overlying crust.

Our residual topographic model shows that part of the eastern Anatolian Plateau is not isostatically supported. There is residual topography up to ∼1.7 km that cannot be explained by a crustal root. The residual topography in Eastern Anatolia is probably supported by rising hot asthenosphere in the uppermost mantle. However, this interpretation does not exclude the possibility that the Anatolian Plateau may partly be compensated by density heterogeneity in the crust and upper mantle, as well as thickness of the lithosphere.

ACKNOWLEDGMENTS

This work was supported by a College Academy of Research, Scholarship, and Creative Activity (CARSCA, a unit of the College of Arts and Sciences, University of Alabama) grant to Mahatsente. We thank the Turkish Petroleum Corporation (TPAO) for supporting Gökay Önal during his M.S. studies at the Department of Geological Sciences, University of Alabama. We also thank the two anonymous reviewers for thoughtful and constructive comments, which were of great help in the preparation of the final version of this paper. The data used in this study are available via file transfer protocol (FTP) by contacting the authors.

1GSA Data Repository Item 2018111, Figure DR1: Land and ocean gravity anomaly data used to develop the EGM2008 model: (a) data availability, and (b) data source identification (Pavlis et al., 2012), is available at http://www.geosociety.org/datarepository/2018, or on request from editing@geosociety.org.
Gold Open Access: This paper is published under the terms of the CC-BY-NC license.