1 Introduction

The Eastern Anatolian region is a geologically very complex and intensely deformed high plateau, is seismically active and is a continent–continent collision region located at the collision point of the Arabian and Eurasian Plates within the Alpine–Himalayan fold-thrust fault belt (Şengör and Kidd 1979; Şengör et al. 1985; Yılmaz et al . 1987; Pearce et al. 1990). For that reason, crust in eastern Türkiye is weak and hot. Focal depths of the earthquakes in this region do not exceed depths of 25 km (Reilinger et al. 1997).

In this region, a number of studies (Seber et al . 2001; Al-Lazki et al. 2003; Gok et al. 2003; Şengör et al. 2003; Zor et al . 2003; Sandvol and Zor 2004; Gelişli and Maden 2006; Angus et al. 2006; Barazangi et al. 2006; Pamukcu et al. 2007; Maden et al. 2009a, b; Gök et al. 2011; Maden 2012b) have been performed to examine the crustal structure of the Eastern Anatolia and surrounding regions. Seber et al. (2001) found that the crust thickness of the Turkish–Iranian plateau varies between 45 and 50 km. Al-Lazki et al. (2003) and Gok et al. (2003) studied P n tomographic imaging of mantle velocity and anisotropy, and regional S n wave propagation.

Şengör et al. (2003) determined the crustal thickness of the East Anatolian High Plateau to be only 45 km by using seismic data collected via a network of 29 seismograph stations. Zor et al. (2003) examined the crustal structure of the Anatolian plateau in Eastern Türkiye from teleseismic recordings of a 29-station broadband PASSCAL temporary network using receiver functions. They found that crustal thickness varies from 42 km near the southern part of the Bitlis suture zone to 50 km along North Anatolian fault, and increases from 40 km in the northern Arabian plate to 46–48 km in the middle of the Anatolian plateau.

Tomography studies by Sandvol and Zor (2004) showed that the low velocity zone is also underlain by a slightly lower velocity in the upper mantle beneath the northern Arabian plate and the eastern-most portion of the Anatolian plate. High Bouguer gravity anomalies support the presence of asthenospheric material underlying the Moho in this region (Ateş et al. 1999). Barazangi et al. (2006) showed that the Anatolian plateau lithosphere in eastern Türkiye has no lithospheric mantle and computed the average thickness of the crust as 45 km. Angus et al. (2006) observed three teleseismic S-to-P converted crustal phases which are the Moho with depths ranging between 30 and 55 km, pointing to variable tectonic regimes within this continental collision zone. Pamukcu et al. (2007) evaluated that the Moho depths increase from 38 to 52 km in the Eastern Anatolia region.

Maden et al. (2009a) determined the average Moho, Conrad and basement depth of the eastern Pontides as 35.7, 26.5 and 4.6 km, respectively. Also, they showed that the Moho depths fluctuate between 33.9 and 42.6 km, as obtained from the power spectrum method, and increase from 30.1 to 43.8 km, as determined from gravity inversion studies. Maden et al . (2009b) investigated the tectonic and crustal structure of the eastern Pontides orogenic belt and concluded that the Moho depth varies from 29.0 km in the north to 47.2 km in the south by using the power spectrum method applied to the gravity data. Gök et al. (2011) indicated that the Moho depth reaches 52 km in the southeastern part of the eastern Pontides orogenic belt and northern Armenia. Maden (2012b) estimated the Moho thickens from north (27.2 km) to south (40.8 km) along the central Pontides.

Many researchers (Ateş et al . 2005; Aydın et al. 2005; Dolmaz et al. 2005; Bektaş et al . 2007; Maden et al. 2009b; Maden 2009, 2010, 2013) published maps of Curie point depths of the Türkiye mainland and Black sea. Aydın et al. (2005) presented a Curie point isotherm map of all of Türkiye from aeromagnetic data with a spectral analysis technique. While the maximum Curie point depths are between 20–29 km in the Eastern Pontides and Western Taurus belt, the minimum Curie point depths in the Central Anatolian and Aegean region are 6–10 km. Dolmaz et al . (2005) determined the Curie point depth for Western Anatolia ranges between 8.2 and 19.9 km. Curie point depths of central Anatolia vary from 7.9 and 22.6 km. The shallowest Curie point depths (7.9 km) were observed around the Cappadocia and Erciyes volcanic complexes in Central Anatolia (Ateş et al. 2005). Bektaş et al. (2007) showed Curie point depths in the Eastern Anatolia region vary from 12.9 to 22.6 km. Maden (2009) estimated the depth to Curie point surface thickens from south (14.8 km) to north (21.8 km) along the central Pontides and determined the average Curie point depth as 20.4 km. Maden et al. (2009b) found that the Curie point depths fluctuate between 14.3 km in the south and 27.9 km in the north by using magnetic data of the Eastern Pontides. Maden (2010) computed two Curie point depths as 11.4 and 13.7 km, reflecting the high geothermal potential of the Cappadocia and Erciyes volcanic region. Maden (2013) determined the mean Curie point depth value of 25 km, which is thinner than the Moho depth value in the continental areas.

The first heat flow map for Türkiye was prepared by Tezcan and Turgay (1989) with constant thermal conductivity data gathered from exploration wells. Tezcan (1995) revised this map by using temperatures measured in oil and coal wells mostly drilled in the southeastern Türkiye and Thrace basin. Additionally, İlkışık (1992) made a heat flow map of Türkiye based on the silica and the chemical content of thermal springs. In this map, the average heat flow values are estimated in the western, the central, and the eastern Anatolia as 110.72, 102.78, and 112.81 m Wm−2, respectively. Later, Tezcan (1995) stated that the heat flow values over the Anatolia should be much higher than those over the Mediterranean and the Black Sea.

Maden (2009, 2010, 2012a, b, 2013) studied the heat flow values for the Black sea, Eastern Pontides, Central Pontides, and Central Anatolia regions. In the Central Pontides, the heat flow values computed from the Curie point depth values vary from 94.1 m Wm−2 in the south (back arc) to 63.9 m Wm−2 in the north (arc), indicating that the southern part of the investigated region has a relatively high geothermal potential with respect to the northern part of the region (Maden 2009). Maden (2010) reported heat flow values in the Erciyes region of 88.8 and 106.5 m Wm−2 based on silica data and conventional heat flow methods, respectively. Maden (2012a) determined that the surface heat flow values vary between 66.5 and 104.7 m Wm−2 and computed Moho temperature as 590 ± 60 °C, indicating the presence of a brittle-ductile transition zone along the southern Black sea coast. Maden (2012b) revealed that the Moho temperatures decrease from 992 °C in the south (back-arc) to 415 °C in the north (arc) of the Central Pontides. Maden (2012b) also suggested that the Eurasia plate had moved from north to south under the Anatolia plate along the south Black sea coast. More recently, Maden (2013) determined the Moho temperature increases from 367 °C in the trench (Black sea) to 978 °C in the arc region (Eastern Pontides). The heat flow values at the Moho surface change between 16.4 m Wm−2 in the Black sea basin and 56.9 m Wm−2 in the Eastern Pontides.

In this study, we used two different methods to calculate the crustal thickness of the region, the power spectrum method and the empirical equation method. In addition, we evaluated the Curie point depth, thermal gradient, and surface heat flow values of the Erzurum-Horasan-Pasinler to propound the thermal structure of the study region.

2 General Geology and Tectonics

Eastern Anatolia, where the thermal activity is high, is a continent–continent collision region situated in the Alpine–Himalayan fold-thrust fault belt. It is observed in many parts of East Anatolia that ophiolitic and young volcanic rocks are widespread. Two systems of strike-slip faults control the main tectonics of the Plateau. These are NW–SE trending dextral strike-slip faults paralleling the North Anatolian Transform Fault Zone (NATFZ) and NE–SW trending sinistral strike-slip faults paralleling the East Anatolian Transform Fault Zone (EATFZ). In response to the compressional regime of East Anatolia, the North and East Anatolian fault systems were formed in the region (Zor et al. 2003; Koçyigit et al . 2001).

The Erzurum-Horasan-Pasinler basin, which surrounds Miocene rocks, is oriented E–W and located in Eastern Anatolia (Türkiye). The Pontides bound this basin in the north and the Bitlis Mountains are to the south (Fig. 1). The sedimentary sequence in the region is Eocene to Quaternary. The ophiolite and ophiolitic mélange from the Cretaceous age in the Palandöken and Kop mountains form the basement of the sedimentary sequence. These sedimentary patches overlapping unconformably with the Late Cretaceous-Middle Eocene are covered with Upper Oligocene-Lower and Middle Miocene sedimentary sequences. In the epoch between the Upper Miocene and the present time, the youngest sedimentary patches of the studied area have been developed. Within this, patches of sedimentary sequences, accumulated since the Upper Oligocene era, play important roles as source beds, reservoir beds, and covered beds. In addition, due to oil seepage and geological data indicating hydrocarbons, geophysical studies have gradually increased in the region (Pelin et al. 1980).

Fig. 1
figure 1

Simplified geological and Bouguer anomaly map of the study region (Gelişli and Maden 2006). The contour interval of the Bouguer anomaly map is 10 mGal

Full size image

According to Şengör and Yılmaz (1981), in the late Paleocene along the Pontide arc, the Neotethys was subducted beneath the southern margin of Eurasia and was completely consumed. However, recent geologic and geophysical studies (Şengör et al. 2003) showed that, during the middle Miocene, a small branch of the Neotethys subducted northward beneath eastern Türkiye. Extensive melting and collisional volcanism formed the regional uplift in the east Anatolian plateau (Yılmaz 1993; Angus et al. 2006).

3 Data and Methods

The gravity data used, provided by the General Directorate of Mineral Research and Exploration (MTA) of Türkiye, was acquired at 500 m intervals and with a precision of 0.01 mGal. Gravity values were tied to MTA and General Command of Mapping base stations related to the Potsdam 9,81,260.00 mGal absolute gravity value, which is accepted by the International Union of Geodesy and Geophysics in 1971. Latitude correction was applied according to the 1967 International Gravity Formula. Bouguer reduction and topographic correction to a distance of 167 km were calculated, assuming a terrain density of 2.67 gcm−3. A gravity anomaly map of the region is shown in Fig. 1.

The aeromagnetic data gathered between the years 1978 and 1989 were surveyed by the MTA with a 600-m flight altitude above the ground surface. Flight lines vary between one and five kilometers depending on the expected mineral, geothermal, or other resource in the region; sample spacing along the flight lines was approximately 70 m. The flight direction was perpendicular to the regional geologic formations and tectonic structures. The IGRF (1982.5) was removed from the data for preparing the aeromagnetic anomaly map of Türkiye. The year 1982 was chosen for the removal of IGRF since it was the mid-date of the survey span and the larger sectors were overflown during this year. The total component of the geomagnetic field was measured along N–S trending profiles, then measured values were reduced to 1982 with daily variation and direction error corrections. The aeromagnetic anomaly map (Fig. 2.) grid interval of 2.5 km used in this study was obtained from the MTA (Aydın and Karat 1995; Ateş et al. 1999; Aydın et al. 2005).

Fig. 2
figure 2

Low-pass filtered map of magnetic anomaly data of the study region for computing Curie-point depths. The contour interval is 50 nT

Full size image

Spectral analysis, a statistical-approaching method, has been widely used in order to determine the depth of anomaly sources (Bhattacharya 1965, 1966; Spector and Bhattacharya 1966; Naidu 1970; Spector and Grant 1970; Treitel et al. 1971; Mishra and Naidu 1974; Curtis and Jain 1975; Cianciara and Marcak 1976; Hahn et al . 1976; Blakely 1995; Hofstetter et al. 2000; Maden and Gelişli 2001; Rivero et al. 2002; Büyüksaraç et al. 2005; Bansal et al. 2006; Gelişli and Maden 2006; Bektaş et al. 2007; Maden 2009; Maden et al. 2009a, b, 2010, 2012a). This method provides a relationship between anomaly spectrums and the depth of sources in the wavenumber domain. This method does not require a priori knowledge about the geometry and density contrasts of the source bodies. Spector and Grant (1970) have shown a simple relation between the power spectrum and depth of prismatic magnetic sources. Dimitriadis et al. (1987) generalized the application of the spectral statistical method to gravity data observed at the surface for estimating the mean depth of ensembles of magnetic sources (Treitel et al. 1971). Gravity data is transformed into the wavenumber domain to analyze the wavenumber content of the data. A plot of the logarithm power spectrum versus wavenumber usually shows several straight-line segments that decrease in slope with increasing wavenumber (Fig. 3). The slopes of various segments provide an estimate of the depths to the various levels of anomalous sources. The slope of the low wavenumber parts of the power spectrum curve yield the deep sources, while the slope of the high wavenumber region imply the depth to the shallow gravity sources. The cut-off wavenumber separates the power spectrum curve of the Bouguer gravity data into high and low wavelength. The high and low wavenumber domains are associated with shallow and deep structures, called residual and regional anomalies, respectively.

Fig. 3
figure 3

Power spectrum of the Bouguer gravity data of the study region (Gelişli and Maden 2006). It is a plot of the logarithm of the energy versus wavenumber of the gravity data collected in the Erzurum-Horasan-Pasinler basins and surroundings. Estimate h1 represents the Moho depth; h2, h3, and h4 represent the depth estimates for the main physical interfaces within the crust

Full size image

Following Pal et al. (1979), the depth to a given ensemble of prismatic bodies d, can be estimated from the power spectrum of the associated anomaly, by applying:

$$E(k,0) = E(0,0)e^{{\left( {\frac{2\pi }{N}d - m} \right)k}} ,$$
(1)

where k is the wavenumber, m is the the slope of the average spectrum, and N is the number of gravity data. The depth (d) to the top of the basement can be easily calculated, since the gradient of \(\left( {\frac{2\pi }{N}d - m} \right)\) should be zero. This technique is a fast and affordable method for estimating the depth of disturbing anomaly sources and gives reliable depth estimates (Nwogbo 1998). When the power spectrum method is applied to the gravity data, the cut off wavenumber is fixed in regional-residual separation and, when applied to the magnetic data, the Curie depth can be estimated (Maden and Gelişli 2001).

Several empirical studies on crustal structure, as determined from the relation between a Bouguer gravity anomaly (Δg) and seismically-determined crustal thickness (H) performed by several authors (Wollard 1959; Wollard and Strange 1962; Ram Babu 1997; Hofstetter et al. 2000; Rivero et al. 2002), have shown that they are linearly related. A brief account of the theory on the linear relationship between Δg and H is described below.

Let H 1 and H 2 be the Moho depths at two locations P 1 and P 2, respectively, with corresponding gravity anomalies Δg 1 and Δg2 (Fig. 4). Let σ c and σ m be the average densities of the crust and upper mantle, respectively. Since H 2 − H 1 , H 1 and H 2 ≪ x, where x is the horizontal extent of the structure involved, the maximum gravity anomaly caused by the plate of thickness H 2 − H 1 is given by.

Fig. 4
figure 4

Schematic cross section of a gravity anomaly profile over a density surface (after Ram Babu 1997)

Full size image
$$\Delta g_{1} - \Delta g_{2} = 2\pi G(\sigma_{m} - \sigma_{c} )\left( {H_{2} - H_{1} } \right),$$
(2)

where G is the universal gravitation constant. It follows from Eq. (2) that

$$H_{1} = H_{2} + M(\Delta g_{2} - \Delta g_{1} ),$$
(3)

where M = 1/2πG(σ m − σ c ). If H 0 is the Moho depth corresponding to Δg = 0, Eq. (3) can be generalized as

$$H = H_{0} + M\Delta g,$$
(4)

A number of deep seismic sounding profiles in the study region must be available to calculate the Moho depths from Eq. (4). This equation represents a straight line with slope M. The M and H 0 constants are determined from graphs of seismic Moho depths against Bouguer values observed in the same stations (Ram Babu 1997). Ram Babu (1997) showed that the empirical relationships between a Bouguer anomaly and seismically-determined crustal thickness are linearly related. In this study, we used

$$H = 32 - 0.08\Delta g$$
(5)

and

$$H_{c} = 18.6 - 0.031\Delta g,$$
(6)

formulas (Wollard 1959; Demenitskaya 1967) to estimate the Moho and Conrad depths of the study area. In these formulas, Δg, H, and H c are gravity anomaly, the Moho depth, and the Conrad depth, respectively. Because the empirical relations are computed in distinct regions, it is important to consider the geology and tectonics of the region if one decides to use the relationship for the gravity data.

In general, sedimentary basins are associated with low gravity values due to low-density sediment fill (Athy 1930; Howell et al. 1966). Densities of sedimentary rocks increase with depth mainly due to compaction. For the interpretation of the sedimentary basin gravity anomaly, the generally stacked prism model of Bott (1960) is used as an initial model. In Bott’s ( 1960) method of gravity interpretation, the mathematical geometries of a sedimentary basin were modelled with a series of juxtaposed vertical prisms having equal widths. In the absence of the parameters of such basins, Murthy and Rao (1989), Leão et al. (1996), and Barbosa et al. (1997, 1999) developed methods using uniform density contrast to trace the basement interfaces from the observed gravity anomalies. On the other hand, in order to determine the sedimentary basin configuration, Cordell (1973) used the exponential depth-density relation, Rao (1990) proposed a quadratic density function, Chai and Hinze (1988) developed an iterative algorithm in the frequency domain using an exponential density function, and Visweswara Rao et al. (1993, 1994) utilized the parabolic density function. Chakravarthi et al. (2002) and Chakravarthi and Sundarajan (2004) represented the sedimentary basins with a series of 3D-juxtaposed vertical prisms whose upper surface is the earth surface and bottom surface is at the bottom of the basin, and whose density contrast is constant.

We apply the gravity inversion method to the residual gravity data of the study region to find out the basement depth of the sedimentary layer in the study area. The initial depth of the sedimentary layer is calculated by using Bott’s (1960) method. The Bott (1960) method of gravity interpretation involves an estimation of a sedimentary basin by a series of 2D-juxtaposed blocks. The inversion results of the last iteration that perform the best fit between the observed and calculated gravity values have been used for the sedimentary basement model of the region.

Spector and Grant (1970) examined statistical properties of magnetic anomaly patterns to obtain the average depths to the top of magnetized bodies. This method gives a connection between the 2D FFT power spectrum of the magnetic anomalies and the depth of magnetic sources by transforming the spatial data into the wavenumber domain. Bhattacharyya and Leu (1975, 1977) and Okubo et al. (1985, 1989) provided a two-step method to define the bottom depth (Z b ) of the deepest magnetic sources using the spectral analysis method designed by Spector and Grant (1970). The first step is the computation of the depth to the centroid of a magnetic body (Z 0) from the slope of the longest wavelength part of the spectrum. The second step is estimation of the top boundary (Z t ) of the magnetized source from the slope of the second longest wavelength spectral segment. The basal depth, inferred from the Curie point depth, Z b is then obtained from these two depth values by using the formula: Z b  = 2Z 0 − Z t (Tanaka et al. 1999; Okubo et al. 1985).

Heat flow can be computed using the formula

$$q = k\left[ {\frac{\theta }{{Z_{b} }}} \right],$$
(7)

where q is the heat flow in m Wm−2, k is the coefficient of thermal conductivity in Wm−1 K−1, θ is the Curie temperature in  °C, and Z b is the Curie point depth in km. In this equation, it is assumed that the surface temperature is 0 °C and no heat sources exist between the Earth’s surface and the Curie point depth. In this equation, the Curie point depth is inversely proportional to the heat flow (Turcotte and Schubert 1982; Tanaka et al. 1999; Stampolidis et al. 2005).

On the other hand, surface heat flow density for the region can be computed from the equation of Sharma et al. (2005) as follows:

$$Q_{s} = 640 H^{ - 0.68} ,$$
(8)

where Q s is the heat flow density in m Wm−2 and H is the Curie temperature depth in km. Typical surface heat flow densities are about 30–40 m Wm−2 in fore-arc regions, more than 100 m Wm−2 in the volcanic front, and about 70–80 m Wm−2 in back-arc regions (Hyndman and Lewis 1999). The thermal structure of the crust can be approximated by steady state solution of the heat conduction equation as follows:

$$\nabla^{2} T = - \frac{A}{k},$$
(9)

where T is the temperature (°C), A is the radioactive heat production in μ Wm−3, and k is the thermal conductivity in Wm−1 K−1. In case of constant heat production, the analytical solution to the heat equation with depth can be given with the following equation of Cermak et al . (1991):

$$T(z) = T_{0} + \frac{{Q_{s} }}{k}z - \frac{{A_{0} }}{2k}z^{2} ,$$
(10)

where T is the temperature at depth z in  °C, T 0 is the annual mean temperature at the Earth’s surface (5.6 °C obtained from Turkish State Meteorological Service for the region), k is the thermal conductivity in Wm−1 K−1, and A 0 is the heat production value in μ Wm−3 at the surface. He et al. (2009) used the heat production values for the upper, middle, and lower crusts and the mantle as 1.10, 0.83, 0.37, and 0.24 μ Wm−3, respectively. In addition, Jokinen and Kukkonen (1999) used the heat production values for the upper, middle, and lower crust and the lithospheric mantle as 1.8, 0.6, 0.2, and 0.002 μ Wm−3, respectively.

Turcotte and Schubert (1982) suggested that the geothermal conductivity (k) of basalts and granites are 1.3–2.9 and 2.4–3.8 Wm−1 K−1, respectively. Tezcan (1979) assigned a geothermal conductivity value of 2.1 Wm−1 K−1 to Neogene clayey formations. Rao et al. (1970) suggested using a thermal conductivity value for the sediments of 3.14 Wm−1 K−1. Correia and Jones (1995) used the thermal conductivity values of the upper, middle, and lower crust as 2.7, 2.5, and 2.1 Wm−1 K−1, respectively. Rimi ( 1999 ) suggested that the thermal conductivity value of the lower crust is 2.6 Wm−1 K−1. In this study, the thermal conductivity values of the upper and middle crust are taken from Correia and Jones (1995) and Rimi (1999) as 2.7 and 2.6 Wm−1 K−1, respectively.

4 Crustal Structure of the Erzurum-Horasan-Pasinler Basins

The Bouguer gravity map of the Erzurum-Horasan-Pasinler basin shows closed anomalies because of the low-density sedimentary units. Especially, the Pasinler basin produces large negative gravity anomalies (Fig. 1). It is shown that Bouguer gravity anomaly counters lay parallel to the main tectonic elements of the region. The high frequency magnetic anomalies near the Pasinler, Çobandede, and Çimenli region appear to be associated with volcanic rocks (Fig. 2). Regional geological features, topography of the region, and magnetic core fields produce long wavelength magnetic anomalies and could affect the centroid depth estimates and probably create erratic estimates of the Curie point depths (Okubo et al. 1985). For that reason, the magnetic data of the research area were reduced to the pole and were low-pass filtered to prevent the effects of topography.

The power spectrum method (Spector and Grant 1970) has been performed by taking a 2D Fourier transformation of the Bouguer gravity data of the region. To compute the average depth component of the gravity sources in the study area, logarithmic amplitude spectra of the radial wavenumber were drawn (Fig. 3). In this graphic, power spectrum analysis reveals three distinct linear segment lines that give three source depths. The first one, which is the low wavenumber region between 0 and 0.1 km−1 and respects contributing long wavelength sources, is considered due to density difference in the Moho discontinuity (43.0 km). The second one, which is an intermediate region in the interval 0.15–0.25 km−1, is associated with the density difference between the upper and lower crust called the Conrad interface (20.9 km). Sediment filling (5.2 km) cause the third region to have a high wavenumber between 0.3–0.5 km−1. The last one, which is the larger wavenumbers region (0.5–1.0 km−1), is considered to be the youngest sedimentary patche (2.4 km).

Various empirical relations are generally based on knowledge of the Moho depth through seismic methods. These empirical relationships between the Bouguer anomaly (Δg) and seismically-determined crustal thickness (H) reveal that they are linearly related (Wollard 1959; Wollard and Strange 1962; Ram Babu 1997; Hofstetter et al. 2000; Rivero et al. 2002). In this study, we employ the Eqs. (5) and (6) suggested by Wollard (1959) and Demenitskaya (1967) to estimate the Moho and Conrad depths of the study region, respectively. Eq. (5) proposed by Wollard (1959) was suggested for the whole earth instead of a specific region. However, the depths computed by the Wollard equation give, in general, the information on the crustal thickness in Eastern Türkiye. In addition, it provides an opportunity to compare the results obtained from other methods. By using the Wollard equation, we fixed and mapped the Moho depth of the study region. In this map, the Moho depth ranges between 41.0 and 44.5 km (Fig. 5). In the same region, we estimate the average Moho depth as 43.0. When we take into account two depth values, there is almost a 2.0 km difference between the power spectrum method and the Wollard equation. On the other hand, the depth values of the Conrad discontinuities are determined to vary between 22 and 26 km (Fig. 6) by using Eq. (6) (Demenitskaya 1967). In the same way, the maximum difference is aproximately 5.1 km between the power spectrum method and empirical equations.

Fig. 5
figure 5

Moho depth of the study area calculated from the regional gravity anomaly by using empirical equations. Contour interval is 0.5 km

Full size image
Fig. 6
figure 6

Conrad depth of the study area calculated from the regional gravity anomaly by using empirical equations. Contour interval is 0.2 km

Full size image

We used a 3D gravity inversion method to estimate the basement depth of the Erzurum-Horasan-Pasinler sedimentary basins. Initial depths of the sedimentary basin are derived from the Bouguer slab formula and subsequently updated iteratively based on the differences between the observed and theoretical gravity anomalies until the modeled gravity anomalies resemble the observed ones. The residual gravity data shown in Fig. 7a was computed by subtracting the regional trend surface from the Bouguer gravity field. The calculated gravity field and the basement relief maps of the basin are shown in Fig. 7b and c, respectively. The closeness of fit between the observed (Fig. 7a) and calculated (Fig. 7b) gravity anomalies shows the validity of the results. The maximum basement depth of the study region is found to be 6 km by using 3D gravity inversion of the residual gravity data (Fig. 7c). This depth value coincides fairly well with the basement depth value (5.2 km) determined from power spectrum curves. The basement depths of sedimentary layer counters are parallel to the tectonic units and maps obtained for Moho (Fig. 5) and Conrad (Fig. 6) discontinuities.

Fig. 7
figure 7

Observed (a) and calculated (b) gravity fields, and basement relief (c) maps of the Erzurum-Horasan-Pasinler basins. Basement relief map of the region determined from observed residual gravity using a constant density contrast of −0.45 gcm−3. Basement relief map contour interval is 0.5 km. Contour interval of the observed and calculated residual gravity maps is 5 mGal

Full size image

5 Thermal Regime of the Region

The Curie point depth estimates the average depth of magnetic sources and it is believed they reflect the thermal structure of the region. Partial melting of the lower crust is thought to have contributed to the present thermal structure of the crust. Previous studies show that the Curie point depth varies greatly according to the geological context. It is shown that the Curie point depths are shallower than about 10 km within volcanic and geothermal areas, 15–25 km in island arcs and ridges, deeper than 20 km in plateau regions, and deeper than 30 km in trench areas (Tanaka et al. 1999).

In this paper, we utilize the available aeromagnetic data over the Erzurum-Horasan-Pasinler region to calculate the Curie isotherm depths. Curie point depth estimation requires the deepest magnetic sources. However, reduced-to-pole magnetic anomaly data contain both long and small wavelength anomalies. It is important to take into account the fact that the magnetic data must be reduced-to-the-pole before low pass filtering to remove the effects of topography (Tanaka et al. 1999). The low pass filtered magnetic anomaly map (Fig. 2) was used for the Curie point depth estimation.

We determined the depth to the tops and centroids of the magnetic sources from an azimuthally-averaged power spectrum of the low pass filtered magnetic anomaly data to be 9.6 and 13.8 km, respectively (Fig. 8a, b). The average Curie point depth is then computed from those two depth values as 18.0 km for this region. Bektaş et al. (2007) calculated the Curie point depths vary between 17 and 19 km in Eastern Anatolia, near Erzurum. The geothermal gradient (\(\frac{{{\text{d}}T}}{{{\text{d}}Z}}\)) between the Earth’s surface and the Curie point depth (Z b ) can be calculated by using the formula \(\frac{580}{{Z_{b} }}\) giving 32.2 °C km−1 in the region. In this calculation, the Curie point temperature was assumed to be 580 °C.

Fig. 8
figure 8

Power spectrum curve of the aeromagnetic data. a top boundary of magnetic layer (Z t ); b centroid of a magnetic layer (Z 0) derived from the slope of the power spectrum curve (Gelişli and Maden 2006)

Full size image

Typical surface heat flow density values are about 30–40 m Wm−2 in fore-arc regions, more than 100 m Wm−2 in the volcanic front, and about 70–80 m Wm−2 in back-arc regions (Hyndman and Lewis, 1999). Two surface heat flow density values are estimated as 90.3 and 87.9 m Wm−2 from the Curie temperature depth of the region by using the Eqs. (7) and (8), as suggested by Sharma et al . (2005). The average heat flow value for these values is 89.1 m Wm−2.

Pollack and Chapman (1977) empirically established that 60 % of the surface heat flow density is attributed the mantle heat flow density, whereas 40 % is the contribution of radiogenic sources in the upper crust. Hence, 53.5 m Wm−2 is the contribution from the mantle and the remainder (35.6 m Wm−2) is from the crust. High mantle heat flow density values (53.5 m Wm−2) should be related to the melting, and collisional volcanism formed the regional uplift in the east Anatolian plateau. as suggested by Yılmaz (1993) and Angus et al. (2006).

Geotherm calculations reveal that the Moho temperature (Fig. 9) is 1,028.0 °C based on the crustal model, as determined from power spectral analysis for the study region. High heat flow values might be related to melting of the lithospheric mantle caused by upwelling of the asthenosphere. The high heat flow density values are attributed to the tectonic activities, as suggested by Şengör and Yılmaz (1981) and Bozkurt (2001). The depth of Curie temperature is found to be 21 km, which is more or less consistent with the Curie temperature depth value given by Bektaş et al. (2007) and Gelişli and Maden (2006). In addition, this depth coincides fairly well with the Curie point depth estimated using the power spectrum method (18 km). The geothermal gradient value can be computed by using this depth value as 27.7 °C km−1. The difference between the geothermal gradient values obtained using the power spectrum method and the geotherm calculation is estimated as 4.5 °C km−1. The mantle heat flow value (Fig. 10) determined from the geotherm calculations is 48.7 m Wm−2 in the Erzurum-Horasan-Pasinler basin. Table 1 shows the parameters used in the geotherm calculation for the region.

Fig. 9
figure 9

Temperature as a function of depth for the Erzurum-Horasan-Pasinler basins based on spectrum analysis studies

Full size image
Fig. 10
figure 10

Heat flow versus depth profile for the Erzurum-Horasan-Pasinler basins. Crustal thickness values are obtained from power spectral analysis

Full size image
Table 1 Depth values for the crustal model, thermal conductivity, and heat production values used for the geotherm calculation
Full size table

6 Discussions

We applied the power spectrum method to Bouguer gravity data and found the Moho, Conrad, and basement of the sedimentary layer as 43.0, 20.9, and 5.2 km, respectively. On the other hand, by applying the empirical equation of Wollard (1959), we determined the Moho depth values vary between 41.0 and 44.5 km. It is shown that these values derived from the power spectrum method and empirical equations are consistent with the results computed by Seber et al . (2001), Al-Lazki et al. (2003), Gok et al. (2003), Zor et al. (2003), Sandvol and Zor (2004), and Angus et al. (2006). Pamukcu et al. (2007) estimated a Moho depth of 47 km near the Erzurum-Horasan-Pasinler region. This depth is only 2.5 km different from our Moho depth value computed using the Wollard equation.

Tanaka et al. (1999) showed that Curie point depths are shallower than about 10 km in volcanic and geothermal areas, 15–25 km at island arcs and ridges, deeper than 20 km at plateaus, and deeper than 30 km at trenches. We determined the Curie point depth to be 18.0 km in this region by using the power spectrum method applied to the magnetic data. On the other hand, the Curie point depth value for the region is computed as 21 km by using geotherm computations. This study show that the maximum difference for the Curie point depth values between this study and Aydın et al. (2005) is 2 km in the study region. In the Eastern Anatolia region, Bektaş et al. (2007) found that the Curie point depth values change from 12.9 to 22.6 km, conforming reasonably well with the results obtained in this study. In addition, our study results are not too dissimilar from the results of Tanaka et al. (1999) found in the plateaus.

We determined the maximum depth of the sedimentary basin basement in the research area to be 6 km by using the gravity inversion method; using the power spectrum method, the basement depth was calculated to be 5.2 km. The difference between the power spectrum method and inversion results is 0.8 km.

In the Erzurum-Horasan-Pasinler basin, the Moho temperature is found to be 1,028.0 °C, based on the crustal model determined via power spectral analysis applied to the gravity data. The mantle heat flow value in the study region is estimated as 48.7 m Wm−2. The average Curie point depth for the region, as determined from a spectral analysis technique applied to magnetic data, was found to be 18.0 km. The depth of the Curie temperature (21 km) obtained from geotherm calculations corresponds to the crustal magma chambers, as previously identified by Keskin et al. (1998, 2006), Bektaş et al. (2007), Karslı et al. (2007), Eyuboglu et al. (2011a, b), Özdemir et al. (2011), and Moghadam et al. (2014). The thermal gradient values are 32.2 and 27.7 °C km−1. Surface heat flow value is estimated at 89.1 m Wm−2.

The high heat flow value (89.1 m Wm−2) found in the east Anatolian plateau is accompanied by the values estimated by Hyndman and Lewis (1999). Thin magnetic crust found in this region accounts for tectonically active regions also associated with higher heat flow values. It is also possible that partial melting of the crust caused from the magma rising contributed to the current high geothermal potential in the study area.

7 Conclusions

The following conclusions have been drawn from the present study:

  1. 1.

    Moho depth values change between 41.0 and 44.5 km by using empirical equations for the region.

  2. 2.

    The Moho, Conrad, and sedimentary basement depths are estimated as 43.0, 20.9, and 5.2 km, respectively, as determined using the power spectrum method.

  3. 3.

    The maximum depth of the sedimentary basin basement is 6 km, as derived by using the 3D gravity inversion method.

  4. 4.

    The Curie point depth and the geothermal gradient values are computed as 18.0 km and 32.2 °C km−1, respectively, by using the power spectrum method.

  5. 5.

    The Curie point depth and the geothermal gradient values are 21.0 km and 27.7 °C km−1, respectively, as determined via geotherm computations.

  6. 6.

    The average surface heat flow density value is evaluated as 89.1 m Wm−2.

  7. 7.

    The Moho temperature and mantle heat flow values in the Erzurum-Horasan-Pasinler basin are 1,028.0 °C and 48.7 m Wm−2, respectively.

  8. 8.

    High heat flow values are possibly related to crustal magma chambers, the ophiolitic and young volcanic rocks of which are observed in many parts of Eastern Anatolia continent–continent collision region situated in the Alpine–Himalayan fold-thrust fault belt; the high heat values might also be related to melting of the lithospheric mantle caused by upwelling of the asthenosphere.