Paleomagnetic results from Western Anatolia: evidence of microblock rotations after emplacement of the lower Miocene Yuntda ğ volcanic rocks

The eastern Aegean region has undergone north dipping subduction in the Oligocene, followed by continental collision in the early Miocene and then Miocene-Pliocene extension, which was associated with widespread Miocene volcanism. The aim of this study is to assess evidence of crustal block rotations due to stress variations in the Dikili ( İ zmir) province, Western Anatolia, based on paleomagnetic data obtained from 35 independent sites in addition to data from 19 sites reported in earlier studies. The lower Miocene Yuntda ğ volcanic rocks were erupted in three different structural blocks, the Bergama, Zeytinda ğ and Dikili blocks. Clockwise rotation is found in the Dikili and Zeytinda ğ blocks that varies from R ( ± 𝛥 R) = 12.5° ( ± 7.4 ° ) in the west to R ( ± 𝛥 R ) = 35.6° ± (13.2 ° ) in the east of the Bak ı rçay graben, respectively. In contrast, a counterclockwise rotation of R ( ± 𝛥 R ) = ‐ 38.1° ( ± 6.4 ° ) is estimated for the Bergama block, to the north of the Dikili and Zeytinda ğ blocks. A scissor-like basin evolution is proposed during the opening of the Bak ı rçay graben which led to counterclockwise rotation of the Bergama block and clockwise rotation of the Dikili and Zeytinda ğ blocks after the early Miocene up to the Present. The rotation pattern derived from results of this study demonstrates that localized small scale deformation due to basin evolution besides regional affects must be considered as part of the deformation matrix in this area. our study is focused on an area of tectonic escape and subduction-related and NW–SE trending strike slip faults and E-W trending normal faults Sangu et 2020]. Our paleomagnetic results provide evidence of relatively small, contiguous blocks, which show both clockwise and counterclockwise rotations of different magnitude. They were deformed from the early Miocene to the Present as a result of extension, scissor-like basin evolution and faulting in the eastern Aegean region, in connection with the westward escape of Anatolia.


Introduction
The Alpine-Himalayan orogenic belt has experienced variable deformation along its length, reflecting both, the sizes of the colliding continental blocks and the presence of oceanic basins in proximity to orogens among other variables. The Eastern Mediterranean region is a typical example of orogenic collapse and tectonic escape.
In this region, the northward subduction of the African plate beneath the Aegean region resulted in extension of the upper crust since the Oligocene [Westaway, 1994;Faccenna et al., 2003;Taymaz et al., 2007] (Figure 1a).
When the Anatolian plate collided with the Arabian Platform during the early Miocene, the East Anatolian Plateau began uplifting as a result of crustal shortening [Dewey et al., 1986;McClusky et al., 2000]. In response to subduction and continental collision, the continental crust began to escape laterally to the west of the collision front in the late Miocene-Pliocene. This tectonic escape has been accommodated by displacement along the right-lateral North Anatolian Fault Zone (NAFZ) and the left-lateral Eastern Anatolian Transform Fault Zone (EAFZ) [e.g. McKenzie, 1978;Dewey and Şengör, 1979;Le Pichon and Angelier, 1981;Taymaz 1990;Taymaz et al., 1991;McClusky et al., 2000McClusky et al., , 2003.
The origin of Late Cenozoic magmatic activity in Western Anatolia has been debated. Two phases of volcanic activity have been documented. The older, dominantly intermediate calc-alkaline volcanic successions were followed by mafic alkaline magmatic assemblages [Yılmaz et al., 2000;Karacık et al., 2007;Dilek and Altunkaynak., 2009;Ersoy et a., 2011;Uzel et al., 2020]. Güleç [1991] showed that the lower-middle Miocene volcanic rocks were generated from a shallow mantle reservoir and modified by extensive crustal contamination, whereas the younger, upper Miocene to Quaternary volcanic rocks were generated by upwelling of an isotopically-depleted deeper mantle source, also associated with an extensional regime. Uzel et al. (2020) distinguished early Miocene calc-alkaline and shoshonitic rocks and middle to late Miocene alkaline rocks in the volcanism of the eastern Aegean region along the NE-SW striking İzmir-Balıkesir Transfer Zone (İBTZ). They reported a significant age gap in the Langhian magmatism. An alternative view suggests that volcanic activity was initiated in an extensional regime with N-S extension in the latest Oligocene-early Miocene [e.g. Scott, 1992, Seyitoğlu et al., 1997]. A number of authors [e.g. Yılmaz, 1989Yılmaz, , 2017 suggested that the N-S compressional regime in the early Miocene was replaced by N-S extension during the middle-late Miocene. This led to N-S trending grabens that were formed during E-W extension in the early Miocene [Yılmaz et al., 2000;Yılmaz, 2017], and E-W trending grabens that developed due to N-S extension in the late Miocene-Pliocene [Koçyiğit et al., 1999;Bozkurt, 2000Bozkurt, , 2003.
Several paleomagnetic studies deal with the reconstruction of crustal blocks due to the activity of the North Anatolian Fault over the past [Piper et al., 1996;2010;Gürsoy et al., 2011;Tatar et al., 2013]. Previous studies in western Anatolia reported different senses of rotations, either during or after the extensional phase during the Miocene [Kaymakçı et al., 2007;Uzel et al., 2017]. Counterclockwise rotations have been observed in upper Oligocene, lower-middle and upper Miocene volcanic rocks in Western Anatolia, in the area between Çanakkale, Foça and Bergama (Figure 1a,b) by Orbay et al. [1998] and İşseven et al. [1995]. These authors, however, infer no significant rotation in the north of Biga Peninsula (Figure 1a,b), but inclination discrepancies were postulated as a result of tilting related to listric faulting [Orbay et al., 1998;İşseven et al., 1995]. In Karaburun and Izmir, Lauer [1984] reported data suggesting clockwise and counterclockwise rotations, yet Orbay et al. [2001] reported data suggesting clockwise rotations in this region. Counterclockwise rotation of between 20° and 30° was inferred for the entire Anatolian plate in a study by Platzman et al. [1998] and Kissel et al. [2003]. Van Hinsbergen et al. [2010] used previously published paleomagnetic data to analyse the regional tectonic structures and proposed that oroclinal bending affected Western Anatolia. They documented a counterclockwise rotation of ~20° for the Lycia nappe and Beydağları autochthon between 16 and 5 Ma. This rotational phase was interpreted to coincide with the exhumation of the Central Menderes Massif [van Hinsbergen 2010; Figure 1a,b] in the late Oligocene-Miocene. In the area between Lesbos Island and Uşak, no significant rotation took place since the Miocene, although significant clockwise rotation of ~20° was reported between this area and Afyon. The study of Uzel et al. [2015], however, concluded that there were two phases of rotation, in the early Miocene and in the late Miocene, based on the lateral displacement, first sinistral and then reactivated as dextral, along the İzmir-Balıkesir Transfer Zone (İBTZ, Figure 1a,b) across Bergama and Söke [Uzel et al., 2017].
For this study, we sampled 35 sites in the lower Miocene volcanic rocks of the Bergama, Zeytindağ and Dikili subareas of Western Anatolia and combined our results with previous data from the study area to unravel possible block rotations. Although previous studies have interpreted any paleomagnetic data in the framework of large scale crustal deformation, the focus of our investigation concentrates on differential rotation and small block displacement that may have affected this area.

Geologic setting
The study area is characterized by NE-SW trending normal faults, displacement along which graben and horst blocks, such as the Bergama-Dikili graben and the Kozak Horst, formed [Karacık et al., 2007;Hou et al., 2015]. The active fault zones in the study area are the NE-SW oriented left lateral Zeytindağ-Bergama Fault zones ( Figure 2).
The NW-SE Dikili graben is the most important tectonic feature in the study area and represents the western extension of the V-shaped 60 km long Bergama graben [Yılmaz et al., 2000]. A series of E-W trending normal faults merge within the graben system, providing evidence of a N-S extension regime in the area.
Our research is concentrated on the crustal elements, the Bergama, Zeytindağ and Dikili blocks ( Figure 2). The Bergama block is confined along its southern margin by the NW-SE Dikili depression, by the NW directed Altınova Anatolia [Emre et al., 2018]. Sangu et al. [2020] designated the NE-trending Yenikent-Bergama and Çandarlı Fault Zones as major faults at a scale of less than 15 km around Bergama, Yuntdağ and Dikili and argued the Bergama and Kınık-Turgutalp fault zones -as normal dextral strike slip faults -to be the southernmost branches of the NAFZ. Westerweel et al. [2020] reported that faults observed within the Soma basin, which lies to the east of the study area, can be classified into three general groups: NE-SW trending dextral strike-slip faults with high dip angles, NW-SE trending sinistral strike-slip faults with a larger normal component, and E-W trending normal faults.
The stratigraphy of the study area is defined at the base by metamorphic Paleozoic rocks which were metamorphosed during the latest Triassic [Bingöl et al., 1973;Akyürek and Soysal, 1983;Koçyiğit, 1987;Genç and Yılmaz, 1995;Altıner and Koçyiğit, 1993;Altunkaynak, 1996]. Metamorphic, plutonic, volcanic and sedimentary rocks crop out north of the Kozak horst [Bingöl et al., 1982]. The volcanic and associated sedimentary rocks are of Neogene age (Figure 3). Volcanic rocks are the most common rocks exposed and include pyroclastic rocks, lava flows with the associated sedimentary or volcanoclastic rocks of the Ballıca/Soma formation and the upper Miocene-Pliocene Zeytindağ Group, which consists of rhyolitic domes and basaltic andesite-basalt lavas and dikes [Aldanmaz et al., 2000;Karacık et al., 2007;Sangu et al., 2020]. Yılmaz et al. [2000] assigned the lower-middle Miocene volcanic rocks in the study area to the Dikili Group, which comprises a 750 m thick sequence that overlies unconformably Paleozoic metamorphic rocks. Near the base, the sequence includes dark siltstones, purple mudstones, and bituminous shales. The volcanic rocks, spatially and temporally associated with the sedimentary rocks of the Dikili Group, are predominantly andesite, latite, dacite lavas and their pyroclastic equivalents [Karacık et al., 2007]. Numerical age data on the andesite and dacite lavas range from 19 to 15 Ma [Ercan et al., 1985;Altunkaynak and Yılmaz, 1998, and references therein]. Palaeontologic data from intercalated sedimentary rocks imply early-middle Miocene ages [Akyürek and Soysal, 1983;Ercan et al., 1985]. The upper Miocene-Pliocene Zeytindağ Group begins with internally-chaotic coarse grained detrital rocks that grade upwards into sandstones, siltstones, mudstones, marls and white lacustrine limestones. Andesitic volcanic activity had evidently waned before the development of the northeast-southwest trending grabens because the sedimentary fill is devoid of intermediate composition volcanic rocks. Only a few, scattered basaltic lava flows were erupted during this time. These are interbedded with lacustrine limestones of late Miocene Ballıca/Soma Formation [Nebert, 1978;Akyürek and Soysal, 1983;Ercan et al., 1985]. The lavas were fed from fissures, associated mainly with NE-SW graben-bounding faults, and their Rb/Sr age has been estimated at 9 to 6 Ma [Borsi et al., 1972;Ercan et al., 1985]. The contact between the Dikili and Zeytindağ groups has been shown to be an unconformity [Karacık et al., 2007].

Paleomagnetic sampling and methods
Sampling for the paleomagnetic study focused on lower Miocene volcanic rocks of the Yuntdağ sequence which are distributed over three subareas around Bergama, Zeytindağ and Dikili ( Figure 2). This volcanic sequence consists mainly of pyroclastic deposits and flows. The pyroclastic deposits are medium-sized to coarse-grained ignimbrites of andesitic and latitic composition. They are intercalated with a lower sedimentary sequence [Karacık and Yılmaz, 1998;Genç et al., 2001]. Lavas of the Dikili Group are of early to middle Miocene age [16.7 -18.5 Ma: Borsi et al., 1972;Benda et al., 1974; 15.2 ± 0.40 -15.5 ± 0.30: Aldanmaz et al., 2000]. We sampled andesitic volcanic rocks at 35 sites: twenty one sites (DM9,12,17,18, BG1-17) were obtained from the Bergama block (Figures 2, 3a), four sites (DM13-16) were sampled on the Zeytindağ block and ten sites (DM1-8,10,11) were collected in the Dikili block. In addition, outcrops were separated by tens to more than hundreds of metres at different elevations to avoid multiple sampling the same volcanic unit.
A motorized portable drill was used to collect orientated independent core samples. Drill cores of 2.5 cm diameter were cut into 2.2 cm long cylindrical specimens. Sample orientation was determined using both magnetic and sun compasses.
A total of 250 specimens was subjected to both stepwise thermal and alternating field (AF) demagnetization. The directions and intensities of the natural remanent magnetization (NRM) were measured using a JR-6 spinner magnetometer (AGICO) in the Yılmaz Ispir Paleomagnetism Laboratory of Istanbul University-Cerrahpaşa, Turkey.
Both thermal and alternating field demagnetizations were applied to isolate the characteristic remanent magnetizations (ChRM) in steps between 50° C and 700° C, using a Magnetic Measurements MTD-80 oven, or 2.5 and 150 mT, using a 2G600 AF demagnetizer, respectively. Magnetization components were identified and calculated using orthogonal vector projections [Zijderveld, 1967]. The directions of individual NRM components were calculated using principal component analysis [Kirschvink, 1980]. Fisher's [1953] statistics were used to calculate site mean directions and their angular dispersion. To test the angular standard deviation of the virtual geomagnetic poles of the data in order to check for adequate averaging of the geomagnetic field, we used the criteria developed by Deenen et al. [2011]. These criteria are statistically expressed by polar A 95 confidence limits (A 95 max and A 95 min), and the number of sites (N) per subarea. Fold tests after McElhinny [1964] and McFadden [1990] together with the reversal test of McFadden and McElhinny [1990] were applied to evaluate the ChRM ages in the context of the regional deformation. Temperature-dependent low field susceptibility of 35 representative samples was measured by heating in air, using an MS2 Bartington susceptibility meter. Isothermal remanent magnetization (IRM) was induced for one specimen per site by applying fields up to 1 T along the Z-axis with an ASC pulse magnetizer.
Coercivity spectra were tested by applying fields of 1 T along the sample Z-axis (hard component), 0.4 T (medium component) along the Y-axis, and 0.12 T (soft component) along the X-axis (Lowrie, 1990). Subsequently, these samples were thermally demagnetized to identify the magnetic carriers, based on their thermal unblocking behavior.
Anisotropy of magnetic susceptibility (AMS) was measured with a Bartington magnetic susceptibility meter (MS2 & MS2B system). Eigenvalues (κ 1 ≥ κ 2 ≥ κ 3 ) and principal axes directions of the AMS ellipsoid were calculated with the Bartington software package. Mean AMS eigenvectors and their 95% confidence limits were calculated using the methods of Jelinek [1977; 1978].

Rock magnetic results
Rock magnetic results suggest that the volcanic rocks can be divided into three types of magnetic assemblages. Lava flows tend to be foliated as recognized by well-grouped clustering of κ min axes either parallel or perpendicular to the flow contacts, and κ 1 axes typically are in the flow plane. In the P-U diagram [Jelinek, 1981], ellipsoid shapes are mainly oblate (T > 0; Figure 5 and Figure 6). Lineation data, as expressed by the L parameter (κ 1/ κ 2 ), range from 1.005 to 1.08, whereas foliation, as expressed by the F parameter (κ 2/ κ 3 ), ranges between 1.00 and 1.06. The P values (κ 1/ κ 3 ) are between 1.03-1.12, and T values are almost exclusively positive (Appendix A). The directions of maximum susceptibility axes (κ 1 ) may define the orientation of lava flow directions [Knight and Walker 1988;Archanjo et al., 2000;Herrero-Berrera et al., 2001;Gil et al., 2002;Ellwood, 1978;Canon-Tapia et al., 1996;Geoffroy et al., 2002].
The maximum susceptibility axes generally plunge with less than 45° relative to flow contacts, suggesting lava flow deformation by simple shear [Loock et al., 2008]. The regionally highly variable and unsystematically distributed AMS axes directions imply that the volcanic rocks have not been deformed internally after emplacement.  where is the angular distance between the ith VGP and the spin axis [McElhinny and McFadden, 1997].

Paleosecular Variation
The dispersion (S F ) produced by paleosecular geomagnetic field variations is derived by subtracting the within-flow circular standard deviation of the VGP (S W ), which is generally very small, from S P . It arises from the random errors in the paleomagnetic direction at the corresponding site and has been estimated by Cox [1969] as [cf. also PMAG Software, Tauxe, 1998]. If the A 95 angle calculated for a mean VGP lies between the lower (A 95 min) and upper (A 95 max) boundaries, the observed VGP population scatter is inferred to average out PSV. Our investigated lava flows were emplaced over a sufficient time interval of both normal and reverse geomagnetic field polarity and average out PSV on all of our three blocks studied ( Table 1). As noted above, outcrops were separated from tens of metres to more than hundreds of metres at different elevations to avoid sampling the same volcanic unit multiple times. All of our localities yield A 95 values that lie within the reliability envelope (Table 1).  (H corresponds to horizontal position), declination D g(s) and inclination I g(s) describe the mean directions in geographic (before tilt correction) and stratigraphic coordinates (after tilt correction), respectively; k the precision parameter [Fisher, 1953]

Paleomagnetic Mean Directions
Our paleomagnetic results and the previous dataset reported by van Hinsbergen et al. [2010] were compiled to evaluate the possibilities of different senses of rotations (Figure 2b) among the three inferred structural areas.
These data are used to calculate a group mean direction for each area, referred to as Groups 1, 2 and 3.
Group 2 (Table 1 and Figure 8c). Several sites have not been considered in the calculation of a mean direction of this group because no interpretable results could be identified during demagnetization (DM8, 10, 11) or because these exhibit very high dispersion (sites DI5,6,8,9,11,15,18,21,23,25). Seven site mean directions from normal polarity sites (resultant vector of length R = 6.5 and k = 11.6) and fifteen of reverse polarity (resultant vector of length R = 14.2 and k = 16.7) show an observed angular distance γ o = 16.5°, which is less than the critical angle γ c = 18.1° [cf. McFadden, 1990], thus showing a positive reversal test with "C" classification. A reliable fold test could not be performed on data from Groups 2 and 3, because tilted lavas -bedding varying between 0° and 48° -were included only in the present study whereas the previous study only sampled horizontal lava flows.

Discussion
The three observed group mean directions are compared with directions calculated from Eurasian Global Apparent Wander Path reference poles after Torsvik et al. [2012] using R. Enkin's unpublished PMGSC [2004] software. The lower-middle Miocene volcanic rocks of Group 1, separated from Groups 2 and 3 by SE-NW and ENE-WSW trending faults to the west and south, respectively (i.e. the Altınova and Bergama Fault Zones), exhibit a distinct counterclockwise rotation of R ± R = -38.1° ± 6.4°. A clockwise rotation of R ± R = + 35.6° ± 13.2° is estimated from Group 2 data in the tectonic block defined by the Zeytindağ area. The Group 3 block in the Dikili area, separated from the Zeytindağ Block by a NE-SW striking fault, experienced modest clockwise rotation by R ± R = + 12.5° ± 7.4°. Uzel et al. [2015] inferred paleomagnetism-based rotations along a transfer zone, named the SSW-NNE trending İBTZ (Figure 1), and proposed a two-stage extension model. Clockwise rotation of 22 ± 11° in the lower Miocene and a counterclockwise rotation of 25 ± 14° in middle to late Miocene time along the transfer zone were attributed to an orthogonal deformation controlled by detachment and strike-slip faults.

Van Hinsbergen et al. [2010] reported a rotational difference between the northern and southern Menderes
Massif of ∼25° -30°. The authors considered the different senses of vertical axis rotation to be compatible with extension related to exhumation of the Menderes Metamorphic Core Complex, which formed during back-arc extension in the eastern part of the Aegean region. Platzman et al. [1998] and Kissel et al. [2003], however, interpreted their results in terms of a rotation of the entire Anatolian plate in a counterclockwise sense by ~20° -30° since the middle Miocene. In contrast, data from lower Miocene volcanic rocks from the Chios area by Kondopoulou et al. [2011] implied a chaotic rotation pattern, and Uzel et al. [2015] reported a poorly defined result from the Karaburun península with a large error envelope (−8 ± 30°). Westerweel et al. [2020] presented early Miocene paleomagnetic clockwise rotations of 29° ± 5° in the Soma Basin which they ascribed to the İBTZ as a dextral shear zone, whereas counterclockwise rotation of 21° ± 12° in the late Miocene was related to widespread extension as a result of the subduction of the African plate below the Aegean region. Neogene tectonics of the Aegean region are characterized by continental collision between the Arabian and Eurasian Plates and westward escape of the Anatolian block along the North and East Anatolian fault zones [McKenzie, 1989;Barka and Reilinger, 1997]. The driving mechanism in upper-plate extension in the Aegean region has been assigned to NNE subduction of the African plate beneath the Aegean trench and slab rollback along the Hellenic trench since 30 Ma [Royden, 1993;Jolivet and Faccenna, 2000;Faccenna et al., 2003;van Hinsbergen et al., 2005].
Our new paleomagnetic results, based on a relatively limited dataset, reveal rotations of statistically significant opposite sense among three different small crustal blocks, suggesting crustal rotations on a rather local scale.
Complex magmatic activity in Western Anatolia took place during two cycles, when lavas and pyroclastic products initially were emplaced during the early to middle Miocene ( Figure 9a) and secondly shoshonitic volcanic rocks (basaltic to trachy-andesitic) were emplaced in the middle Miocene. In the Dikili and Zeytindağ blocks, clockwise rotations of about 12° and about 36° are obtained for this time period. Farther north, in the Bergama block, counterclockwise rotation of about 38° is observed (Figure 9b). We suggest that scissors-like opening of the Bakırçay graben is responsible for counterclockwise rotation of Block 1 (Bergama) and clockwise rotations of Blocks 2 and 3 (Zeytindağ, Dikili) (Figure 9 b, c). Each block was in the latest tectonic phases surrounded by dextral faults to the north and south lasting into the Quaternary. Sangu et al. [2020] recently named these structures as the right lateral Altınıova, Çandarlı, Bergama and Gülbahçe oblique faults (Figure 9c). Further displacement along these faults may have resulted in varying amounts of rotations which may escape paleomagnetic resolution.
On a larger scale, the pattern of differential block rotation could be accommodated by the N-S extension of the western Anatolian area, as several smaller fragmented crustal blocks were rotated with respect to one another to result in the present distribution of extensional basins. From the early to late Miocene, the overall region experienced about 45° of CW rotation, yet during the same time interval the Menderes part to the south experienced a compatible 48° CCW rotation Uzel et al., 2015]. These large magnitudes of counterclockwise and clockwise rotations, however, could not only be explained by basin evolution over a small scale area. Therefore the counterclockwise rotation of Anatolia during its westwards escape must have also had a significant control in this rotational pattern. Uzel et al. [2015] suggested that the displacement along the İBTZ, which lies SE of the study area, should also affect areas that are outside this transfer zone. If this hypothesis is reasonable, it may explain why the Zeytindag block, which is situated immediately within the zone, experienced a greater amount of clockwise rotation, compared to the Dikili block. In addition, the contribution of the westward escape of the Anatolian block and the slab rollback along the Hellenic trench also needs to be considered, together with displacement along the NNE-SSW trending İBTZ during tectonic deformation. Therefore, the influence of smaller scale differential rotations resulting from relative displacement along grabens bounding faults should not be neglected when considering tectonic rotations in the context of the regional Aegean region deformation.

Conclusions
Paleomagnetic data from lower to middle Miocene volcanic rocks in the Dikili area and adjacent structural blocks provide evidence for the presence of three separate structurally detached crustal elements that are delineated from each other by NW-SE and E-W oriented normal faults. Group 1, which is defined as the Bergama Block, yields data suggesting counterclockwise rotation of R ± R = -38.1° ± 11.1°. On the other hand, clockwise rotations of R ± R = 35.6° ± 13.2° and R ± R = 12.5° ± 7.4° are obtained for Group 2, (Zeytindağ Block), and Group 3, (Dikili Block), which are separated by a ENE-WSW trending graben and fault system (cf. Bergama, resp. Zeytindağ graben in Figure 2).
Considering regional tectonics, our study is focused on an area of tectonic escape and subduction-related extension [e.g. McKenzie, 1970;Dewey and Şengör, 1979;Le Pichon and Angelier, 1979], that is locally delimited by NE-SW and NW-SE trending strike slip faults and E-W trending normal faults [Westerweel et al., 2020;Sangu et al., 2020]. Our paleomagnetic results provide evidence of relatively small, contiguous blocks, which show both clockwise and counterclockwise rotations of different magnitude. They were deformed from the early Miocene to the Present as a result of extension, scissor-like basin evolution and faulting in the eastern Aegean region, in connection with the westward escape of Anatolia. c) The influence of dextral faults in the north led to counterclockwise rotation of the Bergama block, while the evolution of the Bakırçay graben provoked clockwise rotation in the Dikili and Zeytindağ blocks.