Paleo and modern stress regimes of central North Tabriz fault, Eastern Azerbaijan Province, NW Iran

0. INTRODUCTION

Scarcity of focal-mechanism solutions and insufficiency of information about stress states around the North Tabriz fault (NTF) (Taghipour, 2004; Berberian and Arshadi, 1976) in northwestern Iran are impediments to kinematic studies on this fault. Though the last major destructive earthquake on the fault was in 1854, the city of Tabriz, with a current population of nearly two million people, has been flattened numerous times according to historical sources (Berberian and Arshadi, 1976). Resolution of the stress states might shed light on the seismotectonic study of this fault, as well as on the geodynamics of the Arabia-Eurasia collision zone. To address this vacuum of knowledge, we studied the local stress states along the central part of the NTF, from its intersection with the North Misho fault (Sufyan) to the NW slope of Takalti Mountain (Qizilja-Meydan) (Fig. 1).

Figure  1.  Geologic map of the central part of the NTF. F1, F2 and F3. See the first paragraph of section Moro Mountain (subarea 1); L. lineament; NMF. North Misho fault; NM2F. North Moro fault; NTF. North Tabriz fault; SF. Shirinje fault; SMF. South Misho fault. The cherrywood brown coloured lines indicate faults delineated in this study. The open rectangle denotes the lineament in subarea 1. The open ellipses define the subareas 1, 2, and 3. A and B show the origin and end point of swath profile along the NTF, respectively.

DownLoad: Full-Size Img PowerPoint

Many studies have determined the stress states in different tectonic settings by assuming parallelism between slip direction and resolved maximum shear stress (first proposed by Wallace (1951) and Bott (1959)) and by using some kinematic indicators (for instance, the attitudes of a fault plane and the associated striae) for determining the slip direction on a fault surface (for example, Doblas, 1998; Hancock, 1985; Carey and Brunier, 1974). Different statistical manipulations of kinematic data has led to development of a variety of methods proposed to estimate paleo-stress states from measured fault-slip data (e.g., Yamaji, 2000; Fry, 1999; Nemcok and Lisle, 1995; Hardcastle and Hills, 1991; Galindo-Zaldívar and González-Lodeiro, 1988; Huang, 1988; Lisle, 1988, 1987; Simón-Gómez, 1986; Angelier, 1984, 1979; Armijo et al., 1982; Etchecopar, 1981). The results of stress inversion analysis are limited to a reduced stress tensor (Angelier, 1989) composed of four variables (i.e., the orientations of the principal stress axes (σ₁, σ₂, σ₃) and the stress ratio Φ (Bishop's definition used in this study)). A comparison of different normalised stress tensors can be carried out using parameters such as the stress tensor difference D, average tensor σ_M, and dispersion of stress results τ_octM (Orife and Lisle, 2003; Lisle and Orife, 2002).

We applied multiple stress inversion method to the fault-slip data collected from different segments of the NTF. The results of the inversion analysis, together with stratigraphic, geometric, and structural information, are used to reveal (1) the temporal and spatial variations of the stress states and (2) the tectonic regimes along the different segments of the NTF.

1. GEOLOGIC SETTING

The NTF is a major brittle structure with an NW-SE trend which is located in the NW part of the Alborz-Azerbaijan (AL-AZ) sedimentary-structural sub-domain of the Iran Plateau (Aghanabati, 2004; Nabavi, 1976; Eftekhar-nezhad, 1975; Stöcklin, 1968). Golonka (2004) considers the NW part of this sub-domain (i.e., the Azerbaijan Plateau) to be a different block and limits it by the Lahijan fault (on the southeast) and the Aras fault (on the northwest) from the Alborz and Lesser Caucasus blocks, respectively (Figs. 2a and 2b).

Figure  2.  (a) Paleoenvironment of the circum-Caspian area during Messinian; plates position at 6 Ma (redraw from Golonka, 2004). Al. Alborz; LC. Lesser Caucasus; SCM. South Caspian microcontinent; SS. Sanandaj-Sirjan; Tl. Talysh (Azerbaijan Plateau); Tr. Transcaucasus. (b) Kinematic map of NW Iran, eastern Turkey and the Caucasus. Black thicker lines: Major active faults; AF. Aras fault; EAF. East Anatolian fault; LF. Lahijan fault; MRF. main recent fault; NAF. North Anatolian fault; NTF. North Tabriz fault. White and grey dots in the main figure and black dots in the inset are earthquake epicentres; gray arrow. the direction of Arabia-Eurasia convergence at 34°N, 42°E (modified after Talebian and Jackson, 2002). Open thick and dashed arrows: shortening and right-lateral directions and related deformation rates, respectively, which are measured with GPS (Vernant et al., 2004).

DownLoad: Full-Size Img PowerPoint

The division and northward drifting of the Lesser Caucasus, Azerbaijan, Alborz, Sanandaj-Sirjan, and Lut blocks and the Arabian Plate from the NE margin of Gondwana and their collisions with the Eurasian Plate have induced roughly N-S oriented compressions at different times (Golonka, 2004). The compressional-extensional neotectonic regime that Koçyiğit et al. (2001) proposed for the eastern Anatolian Plateau and the Lesser Caucasus can be extended to the Azerbaijan Plateau in terms of the prevalence of transcurrent deformation, which according to Vernant et al. (2004), is mainly accommodated on the NTF zone (Fig. 2b).

The central part of the NTF, between Sufyan and Qizilja-Meydan, is composed of multiple segments with slightly varying azimuths (119°–122°) (Fig. 1). The NTF's general dip is considered vertical (Aghanabati, 2004; Berberian and Arshadi, 1976), although according to Moradi et al. (2009) it has a SW-ward vergence in this central part. The displacements along different segments of this fault zone are a combination of dextral and reverse or thrust components (Karakhanian et al., 2004; Vernant et al., 2004; Hessami et al., 2003; Berberian and Yeats, 1999; Nabavi, 1976). The Shirinje fault (SF) to the northeast is sub-parallel to the NTF, and there are elongate folds between the NTF and SF with axes sub-parallel to the NTF. Different segments of the NTF form several restraining bends and step-overs, and a strike-slip splay has formed near the southeast end of the fault.

The central part of the NTF has Precambrian–Cenozoic lithologic units on the NE-side and Plio-Quaternary sediments on the SW-side. At the northwest end of the NTF, Precambrian to older Cenozoic lithologic units have been thrust over younger Neogene and Quaternary units. Toward the southeast, Neogene lithologic units are prevalent. The creation of local unconformities within the fault zone is an indication of interchange between contractional and extensional regimes as well as restraining and relaxing bends or step-overs.

According to the Geological Survey of Iran(2006, 1993), tectonic activities during Silurian–Carboniferous (Caledonian), Carboniferous–Permian (Hercynian), Jurassic–Cretaceous (Early Cimmerian), Early–Late Cretaceous (Austrian), Cretaceous–Miocene (Laramian), Miocene–Pliocene (Atikan), and Pleistocene–Quaternary (Pasadenian) influenced the geology of the study area.

2. DATA COLLECTION AND PROCESSING

Data consisting of fault plane orientations and associated slip lineations have been collected at 88 locations along the central part of the NTF. The data collection sites were chosen to sample (ⅰ) large faults within the fault zone; (ⅱ) faults in fault tip zones, bends, and step-overs; and (ⅲ) within the raised area of Moro Mountain. In addition to field observations, geologic maps of the study area were used to determine the relative motions of faults using overprinting and cross-cutting relationships.

Yamaji's multiple inverse method (Yamaji, 2000) (MIM, main processor: Version 6) was used to determine the significant stress states. The differences between the recognised stress states were determined by the 'STRESSTAT' program (Lisle and Orife, 2002). Two criteria, geometric and mechanical, were used to study the correlation between the stress states driven from inversion and relative fault-slip data. For the geometric criterion, three critical misfit angles (i.e., 10°, 20°, and 30°) were chosen. For the mechanical criterion, the Mohr point of faults, which satisfy τ≥μσ_n, was adopted, where μ is the coefficient of internal friction (Žalohar and Vrabec, 2007). Considering that fault-slip data are collected mainly from sandstones, 0.51 is chosen as the value of μ (Schellart, 2000). A stress ratio-misfit angle diagram (Yousefi-Bavil, 2014) is used to show resulted stress inversion statistics based on the criteria (Figs. 3c, 5c, and 6c).

Figure  3.  Results of stress inversion in subarea 1. (a) The attitudes of faults and associated striae used in the inversion; (b) the stress tensors recognised by Yamaji's multiple inverse method; (c) phi-misfit diagram for the dataset after the exclusion of data corresponding to stress state C; the black and red numbers at the three critical misfit angles indicate the number of faults, which meet the geometric and mechanical criteria, respectively; the open circles denote the faults meeting only the geometric criterion, while the red solid circles denote the faults meeting both criteria; (d) the symbols star, plus and triangle indicate, respectively, the orientations of σ₁, σ₂ and σ₃ for the determined stress state(s); (e) the correlation between the stress states and sampled fault planes.

DownLoad: Full-Size Img PowerPoint

3. RESULTS

The results of field observations and stress inversions in subareas 1, 2, and 3 (Fig. 1), as well as the results of the study of geologic maps, were as follows.

3.1 Folds

The orientations of fold axes were studied over a large area bounding the NTF, to gain insight into the general orientation of the maximum compressive stress at a larger scale. In the structurally less competent Cretaceous and Miocene lithologic units northeast of the NTF (Fig. 1), there is a preferential NW-SE orientation of those fold axes with length greater than 3 km, though the orientations of shorter fold axes are more variable. This axial preference of the longer folds can be seen in axes hosted in both pre-Miocene and syn- to post-Miocene lithologic units. Because the mean azimuths of the total population of fold axes is 117°–297°, the possible azimuths of the regional maximum compressive stress is inferred to be 27°–207° (Fig. 4a).

Figure  4.  (a) Rose diagram of the orientation of fold axes. σ_1f is the maximum compressive stress determined from the study of the fold axes; (b) elevation differences along the NTF in subarea 2, obtained from a swath profile. The origin and end points of the profile (i.e., A and B) are shown in Fig. 1. The solid and dashed lines indicate the elevation range (maximum elevation-minimum elevation) and mean elevation, respectively. The average value of the elevation range is 254 m; (c) orientation of the determined stress states around the central part of the NTF.

DownLoad: Full-Size Img PowerPoint

3.2 Moro Mountain (Subarea 1)

The deformation zone in the Moro Mountain area is limited by the NTF to the southwest, the Shirinje fault to the northeast, and the North Misho and South Misho faults to the northwest (Fig. 1). The inner eastern boundary of this zone is defined by the North Moro fault (NM2F), an arcuate fault that consists of E-W and NW-SE-trending faults, having common reverse displacement. At the south of this zone, Paleozoic units and Quaternary sediments are juxtaposed along fault F1 with an ESE-WNW trend. Toward the north along the other E-W-trending fault (F2), Cambrian red sandstones and micaceous shales have been thrust over younger pre-Jurassic lithologic units. In the central part of the zone, there is an NW-SE-trending dextral fault (F3) whose northwest extension approaches the conjunction of the NTF and the North Misho fault. Between faults F1 and F2, the Cambrian sandstones and shales have been folded, with the axial plane extending into NE-SW azimuths and dipping towards the NW (with the general orientation 223°/87°). In the east, the lithologic units are displaced by generally NE-SW-trending sinistral faults. The field observations and the map study show that, generally, the NE-SW-trending faults have been cut by the NW-SE-trending faults. Also, small-scale faults with NW-SE and NE-SW trends crosscut both WNW-ESE and NW-SE trending larger faults. A positive flower structure is suggested for the Moro deformation zone, mainly on the evidence of stratigraphic relationships of lithologic units (i.e., the overriding of older units on younger ones) and the uplift of the Quaternary sediments next to the strike-slip NTF.

Fault-slip data for 41 of the 88 total study samples were collected from subarea 1, measured on Paleozoic and Mesozoic units. On one fault, the dip and lateral slips of the first motion have opposite sense with respect to those of the latest motion. In the case of three faults the superimposed motion has the same plunge and lateral slip sense with respect to those of the first motion.

At the first attempt of the inversion on these 41 fault-slip data three high density clusters were obtained. According to the terms proposed for a qualitative description of stress difference (D) between two stress tensors (Orife and Lisle, 2003), two significant stress states were recognised. However, there was another stress state with a significant cluster density, but because of the value of D between it and the most similar significant stress tensor (nominated colour for Φ=0.5 in MIM's colour scale), they were classified as 'similar' stress states. Assigning a stress tensor for each nominal interval of Φ, where possible, 10 fault-slip data were found to be incompatible with the stress tensors determined. These incompatible data were excluded from the original dataset, and applying the inversion to them showed a stable stress tensor in which only one fault had a misfit angle greater than 30°.

Repeating the procedure, after the exclusion of the incompatible faults, two stress states were still recognisable, and the misfit angles were less than 30° for all datasets (i.e., 31 fault-slip data) and less than 20° for 28 faults. There is a slight increase in the value of D for compressive stress tensors (Φ=0–0.4) and a slight decrease for extensional stress tensors (Φ=0.5–0.8). According to the mechanical criterion, stress states with Φ=0–0.1 in comparison to the rest cause more unstable faults. There is a similar distribution of unstable faults through stress states with Φ=0.2–0.8 (Figs. 3a, 3b and 3c; Table 1). In the stress stereograms of Fig. 3, the positions of maximum and minimum principal stresses in extensional stress tensors are more concentrated in comparison to the more compressive tensors.

Table  1.  Results of stress inversion in the three subareas along

	Recognized stress states					Determined significant stress states
	σ1	σ3	Φ	Max D		σ1	σ3	Φ	DMIM	Θ°	Stress state
Subarea 1	017/02	272/81	0.03	0.683 6		011/2	279/47	0.05	0.457	26.676	A
	019/03	122/77	0.09
	015/04	187/86	0.09
	179/02	270/19	0.25
	354/25	261/05	0.46	0.830 6		347/27	257/00	0.7	0.621	36.55	B
	354/33	261/04	0.57
	001/46	263/07	0.85
	346/39	079/04	0.87
	328/18	063/14	0.85
	156/03	066/00	0.94
	157/27	065/04	0.91
	001/82	141/07	0.64	1.669 1		001/82	141/07	0.64	0.262	15.081	C
Subarea 2	170/02	079/03	0.04	0.869 6		175/00	084/12	0.32	0.347	20.045	A
	172/00	082/17	0.06
	172/02	082/4	0.28
	359/01	089/18	0.40
	357/02	088/13	0.50
	357/04	088/15	0.62
	182/05	092/11	0.73
	182/09	089/16	0.80
	198/53	086/17	0.91	1.018 9		198/53	085/17	0.91	0.128	7.327	B
Subarea 3	193/42	077/26	0.30	0.629 6		176/40	076/13	0.72	0.317	18.299	A
	164/22	069/13	0.40
	182/35	078/19	0.59
	181/44	076/15	0.72
	181/57	073/12	0.72
	181/42	080/13	0.79
D indicates tensor differences determined by STRESSTAT. The maximum value of D (max D) for individual stress tensors in each subarea indicates its maximum difference with respect to the other recognised tensors of the same subarea. The parameters DMIM and Θ refer to the mean stress difference and mean angular stress difference, respectively, determined using Yamaji's MIM software.

 | Show Table

DownLoad: CSV

3.3 Aynali Mountain (Subarea 2)

Subarea 2 is limited by the NTF zone to the south and the Shirinje fault to the north (Fig. 1). There are several push-up structures that are the result of restraining bends between different segments of the NTF, and a horsetail splay at the southeast end of the subarea. The youngest lithologic units of the Miocene, which cover this area, have low dip angles (ranging from 4° to 25°). Fold axes are oriented sub-parallel to the NTF and the Shirinje fault (WNW-ESE). Right-lateral displacement along this strand of the NTF is conspicuous from the deflection of drainage channels on the Quaternary sediments southeast of Anakhatun. Faults with both NW-SE and NE-SW trends cross-cut and displace each other; however, the cross-cutting of NE-SW trending faults by NW-SE trending faults is more frequent.

Fault-slip data for 37 of the 88 total study samples were collected from subarea 2. No superposition was found on the sampled faults. At the first attempt of inversion, there was one cluster with high density, which generally consists of compressive stress tensors. However, there were sparse stress tensors with different orientations and extensional values of Φ. The study of the values of D showed two 'different' stress tensors overall. There were seven fault-slip data that were not compatible with any of the stress tensors determined, and their separate inversions did not show a stable stress tensor. Therefore, these data were considered spurious and were excluded from the dataset. The results of the secondary inversion, on 30 fault-slip data, did not change dramatically, so the stress tensors defined are still oriented almost coaxially. Except for one fault, with a misfit angle between 20° and 30°, the others have a misfit angle less than 20° (Figs. 5a, 5b and 5c; Table 1). According to the mechanical criterion, more than 70% of the faults are suitable for instability with respect to the tensors of stress state A. In spite of the decrease in the density of extensional stress tensors (Φ=0.9), there are a remarkable number of faults that can be considered unstable with respect to these stress tensors.

Figure  5.  Results of stress inversion in subarea 2. (a), (b), (d) and (e) see the same items in Fig. 3; (c) Phi-misfit diagram for the dataset after the exclusion of incompatible data.

DownLoad: Full-Size Img PowerPoint

3.4 Qizilja-Meydan (Subarea 3)

Subarea 3 involves a small segment of the NTF that is limited by a restraining bend, fault splays at its northwest end and a restraining bend at southeast end (Fig. 1). The latter is engaged in the raised area of Takalti Mountain, which is probably a pop-up structure. At the northeast side of the NTF there is a low relief area in which Miocene lithologic units are partially covered by Quaternary sediments. Close to the southeast restraining bend and adjacent to the NTF—where the Miocene units are uplifted—the layers facing the NTF have high-angle dips. On the opposite side of the uplifted area, there is a scarp along which Miocene layers have high-angle dips. We did not find striae on planes parallel to the scarp, but the displacement and high-angle dips of beddings are indications of uplift presumably by faults with normal components. Generally, faults with a NE-SW trend are sinistral in this subarea.

The superposition of striae was found only on one fault plane, in which secondary motion with a normal component is overprinted on a first motion with reverse component. There is no preferential trend in the cross-cutting relationships of the faults.

Fault-slip data for 10 of the 88 total study samples were collected from subarea 3. In the first inversion attempt, only one fault-slip datum was incompatible with the stress tensor defined, and its exclusion did not change the stress tensors. The recognised stress tensors are classified as 'very similar' tensors according to their D values. According to the mechanical criterion, in comparison to compressional stress tensors, extensional stress tensors are more significant because they identify a greater number of unstable faults (Figs. 6a, 6b and 6c; Table 1).

Figure  6.  Results of stress inversion in subarea 3. (a), (b), (d) and (e) see the same items in Fig. 3; (c) phi-misfit histogram for the dataset.

DownLoad: Full-Size Img PowerPoint

4. DISCUSSION

4.1 Temporal Relationship

The locations of the sampling sites provide the opportunity for a comparison of the possible differences in the stress states along the NTF before and after the Miocene, because—with the exception of subarea 1—the subareas are covered mainly by Miocene lithologic units.

4.1.1   Subarea 1 temporal solutions

Among the three determined stress states in subarea 1, the first (stress state A, Fig. 3d) is an almost axial compressional stress state and is compatible for reverse and thrust faulting. The thrust of Paleozoic lithologic units over Mesozoic units along the WNW-ESE trending faults is probably a result of this stress regime, where their displacements by NW to NE trending small-scale faults—which are the youngest faults in regard to their temporal relationship—indicates the presence of a pre-modern deformation episode. Given that this axial compressional stress regime does not exist in the other subareas and that the orientation of σ₁ is close to the orientation of maximum compressive stress, which is approximately inferred from the orientations of the folds axes (hereafter σ_1f), we conclude it is independent of the modern tectonic regime around the NTF.

With regard to the modern stress regime, the results of fault plane solution of earthquakes along the NTF indicate a right-lateral strike-slip setting for the NTF (Moradi et al., 2009). Therefore, the axial compressional stress we find in subarea 1 is attributed to a paleo-stress regime after the Cretaceous and before the Miocene. Considering the depositional gap during Paleocene–Oligocene, this compressional paleo-stress is most likely related to the tectonic activities of the Laramian structural stage.

In the stress state C of subarea 1 (Fig. 3d), σ₁ is sub-vertical and the azimuth of its intermediate stress axis is 232°. This orientation is sub-parallel to the orientation of lineaments with an NE-SW trend within and outside of subarea 1. These lineaments are manifested as long-scale and wide rivers, approximately sub-parallel to σ_1f, and NE-SW-trending strike-slip faults which according to Golonka (2004) have been developed on Cretaceous–Quaternary units during Paleogene and Neogene. The lineaments are usually displaced by NW-SE-trending faults, so we ascribe them to a paleo-stress regime. Within subarea 1, there is also a lithologic lineament that divides the northwest part—with its north and northeast areas covered by the oldest Miocene lithologic units—from the southeast part, which is dominated by pre-Cenozoic units (Fig. 1). Along this lineament the thickness of the Miocene units varies and is thicker on the NW-side than the SE-side. Here, the outcrop of Cambrian lithologic units through the Miocene units and their unconformity indicates an uplift history prior to the deposition of the older units of the Miocene. We ascribe stress state C (subarea 1) specifically to the paleo-stress event that resulted in deposition of Late Miocene units in an extensional regime after the uplifting processes and a non-depositional period during the Paleocene–Oligocene.

The stress state B in subarea 1 (Fig. 3b) has the highest cluster density, with a σ₁ in a favourable direction for the dextral displacements along the NTF—determined from the deflections of the NTF––crossing drainage channels on Quaternary sediments, also from the results of fault plane solutions along the NTF (Moradi et al., 2009). The orientation of σ₁ in stress state B (subarea 1) is consistent with the regional N-S shortening axis determined from GPS measurements for present-day deformation in the NW of Iran (Vernant et al., 2004), which is also similar to stress state A from subarea 2. Therefore, the stress state B (subarea 1) is considered the youngest stress regime.

On the whole, the temporal relationships in subarea 1 indicate that the area has experienced three stress regimes: the pre-Early Miocene compressional regime was followed by the Early Miocene extensional regime, which was followed by the modern transcurrent stress regime. Perhaps stress regime A (subarea 1) is related to the convergence and collision processes between the Sanandaj-Sirjan Plate and the Transcaucasus-Talysh system at the Late Campanian–Early Paleocene (Golonka, 2004). Stress regime C (subarea 1) probably is a local stress regime, which caused an extension along pre-existing NE-SW-trending weaknesses in a regional compressional stress regime. Stress regime B (subarea 1) probably is a result of counterclockwise rotation of the Arabian Plate with respect to the Eurasian Plate, and collision of the Indian continent and Lut Plate with the Eurasian Plate, which according to Golonka (2004) caused the development of NW-SE-trending strike-slip faults.

4.1.2   Subarea 2 temporal solutions

Stress state A of subarea 2 (Fig. 5d) is similar to stress state B of subarea 1, in terms of the orientation of σ₁. Based on the age of the lithologic units hosting the sampled faults (i.e., youngest Miocene units), stress state A (subarea 2) is attributed to the post-Miocene structural stage (i.e., Pasadenian).

Given that a fold axis on the northeast side of the NTF is covered by Early Quaternary sediments, the extensional stress state B (subarea 2) might be active before stress state A (subarea 2) or, more likely, is the result of local stress perturbation.

4.1.3   Subarea 3 temporal solutions

Stress state A of subarea 3 (Fig. 6d) similar to stress state A of subarea 2 is the result of sampling from young Miocene lithologic units; therefore, it is constrained to a post-Miocene structural stage (i.e., Pasadenian).

4.2 Spatial Distribution of Stress States and Deformation Regimes along the NTF

4.2.1   Subarea 1

Subarea 1 has experienced three stress regimes (Figs. 3b and 3d; Table 1). The result of stress state A is a compressional contractional deformation regime. The maximum compressive stress is horizontal and sub-perpendicular to the NTF, while the steepest principal axis is σ₃. The uplift of Moro Mountain is probably a result of this tectonic regime.

Stress state C (subarea 1), with a sub-vertical σ₁ and a sub-horizontal σ₂, created an extensional stress regime. The attitude of the σ₁σ₂ plane is 231°/84°, which is sub-parallel to the trend of σ_1f and is quite close to that of the lithologic lineament with orientation of 50°–230° (Fig. 1). We correlate stress state C with the lineament on the basis of this orientational parallelism. In this regard, it is probable that in subarea 1 under stress regime C the location of pre-existing weaknesses (NE-SW-trending strike-slip faults) was favourable for the development of extensional structures and the deposition of thicker Miocene units on the NW-side of the lineament.

Stress state B (subarea 1) is characterised by an extensional stress ratio. The orientation of σ₁ is sub-horizontal, while its steepest axis is σ₂, which is favourable for a strike-slip setting as manifested in high dip angles (i.e., more than 70°) of most sampled fault planes.

The acute angle between the orientations of σ₁ from stress state B (subarea 1) and the NTF—as the southern boundary of this deformation zone—is approximately 45°. In addition, the orientations of sampled faults evoke the Riedel pattern, such that their acute angles with respect to the trend of the NTF (with azimuths of approximately 120°–300°) are 10° and 70° (Fig. 3e), which indicates the presence of a simple shear zone. The reverse and thrust faults (Fig. 1) are sub-parallel to the NTF, which are sub-perpendicular to σ_1f and almost perpendicular to the σ₁ in stress state A (subarea 1). This difference in gross orientation indicates the different deformation histories of the strike-slip faults compared to the reverse and thrust faults.

An extensional stress state (i.e., B of subarea 1) is in contradiction with the active transpressional deformation and the postulated positive flower structure in this deformation zone. Three explanations are proposed for this contradiction: (1) the sampling of the fault-slip data may have been biased to faults with normal dip-slip components and did not cover all deformations; (2) the extensional stress state might be an effect of a classification applied to the recognised stress tensors; and (3) it is probable that the youngest stress is an extensional regime, and the area indeed experiences an extensional strike-slip deformation. In the case of the third explanation, the stress regimes would show change from a compressional deformation—inherited from the older stress regime (i.e., stress state A of subarea 1)—to a modern extensional deformation in which pre-existent weaknesses and local perturbation of far-field stress state are probably effective factors in this deformational transition.

With regard to local perturbation, in subarea 2, which has simpler deformation history and experiences a compressional stress regime, the maximum compressive axis of stress state A (subarea 2) has a similar trend and a lower plunge compared to stress state B of subarea 1. It is possible that in a recent, perturbed, stress regime the location of a given pre-existent discontinuity would favour local extensional deformations.

In general, according to stress state B (subarea 1) and considering the right-lateral component of the NTF's displacement (Geological Survey of Iran, 2006, 1993; Karakhanian et al., 2004; Taghipour, 2004; Hessami et al., 2003), the NTF zone faults with slight vergence towards the NE and SW would be expected to have normal and reverse components, respectively.

4.2.2   Subarea 2

The significant stress state in subarea 2 is characterised as a compressional stress regime, with σ₁ and σ₂ being horizontal and sub-vertical, respectively (Fig. 5d; Table 1). The strike-slip setting is manifested in the high dip angles of the sampled faults, of which almost 50% have dip angles greater than 80° (Fig. 5e). Based on this information and considering that the NTF is a major structure and that the layers of elevated Miocene units at the NE-side of the fault have low dip angles, we conclude that the deformation in Aynali Mountain was mainly accomplished by strike-slip faults and in rather brittle conditions.

The average rake and dip of the faults with azimuths between 110°–130° and 290°–310° is 18° and 80°, respectively, and the acute angle between the azimuth of σ₁ and the trend of the NTF is approximately 55°. The mean of the elevational difference along the NTF, which divides the heights of Aynali Mountain on the northeast side from the low relief area of the Tabriz Plain, is approximately 254 m (Fig. 4b). The rake and dip averages, azimuth angles, and mean elevational differences were used to make a rough estimation of the proportion of vertical to lateral displacements along the NTF. The calculated ratio on the plane 300°/80° is 0.32, and the total lateral displacement is 794 m. This ratio indicates that the strike-slip component is the main component of displacement along this part of the NTF.

4.2.3   Subarea 3

In subarea 3, the stress state is characterised by the extensional stress ratio, and the steepest principal axis is σ₂ (Fig. 6d; Table 1). The extensional strike-slip setting along this segment of the NTF is the result of NE-SW-trending sinistral faults with normal dip-slip components, which are parallel to the trend of σ_1f (Fig. 6e). The depositional condition on the northeast side of the NTF, the creation of a scarp, and the high-angle dips of the Miocene units along this scarp are morphotectonic evidence of this tectonic regime. According to the orientation of σ₁ and σ₂, and considering the right lateral displacements along the fault zone (which can be inferred from the deflection of stream channels and the presence of shutter ridges), it is expected that the faults with a slight vergence towards the NE and SW have reverse and normal components, respectively. The uncertainty about the stress regime in this subarea is due to the lower number of fault slip data, but the morphotectonic evidence is supportive of our interpretation.

4.3 Correlation between the Subareas

The correlation of the determined stress tensors indicates that—in the case of stress states with sub-horizontal σ₁—the stress states tend to be compressional as their σ₁ trends approach σ_1f (Figs. 3d, 3e, 5d, and 5e). The youngest extensional stress regime in subarea 1 is replaced by a compressional regime in subarea 2, where, in comparison with subarea 1, the trend of σ₁ results in a higher angle with respect to the NTF by an 8° clockwise rotation. With a similar azimuth of σ₁, the stress regime becomes extensional in subarea 3 in comparison with subarea 2 (Fig. 4c). The correlation also indicates the variation of the deformation regime along the different segments of the NTF, though the deformations are accommodated on strike-slip faults.

The maximum value of D for the youngest stress states (i.e., stress states B (subarea 1), A (subarea 2) and A (subarea 3)) is 0.749 3, which indicates 'similar' tensors (Table 1). The parameters of their mean stress tensors are σ₁=170°/01°; σ₂=269°/82°; σ₃=080°/08°; and Φ=0.67; these values represent the characteristics of a strike-slip setting with an extensional stress ratio, which indicates that the deformation along the central part of the NTF is mainly accommodated on strike-slip faults. These stress inversion results are consistent with the compressional-extensional regime proposed by Koçyiğit et al. (2001), for the eastern Anatolian Plateau and the Lesser Caucasus. The transcurrent regime and the trend of maximum compressive stress we find are consistent with the results of the fault plane solutions of Talebian and Jackson (2002) and Moradi et al. (2009), which show approximately N-S-trending compressive stress, and the NW-SE-trending deformation zone that extends along the North Anatolian fault and the NTF (Fig. 2b). Our results are also in agreement with the N-S shortening trend and the right-lateral displacements deduced from GPS measurements of Vernant et al. (2004).

5. CONCLUSIONS

Three tectonic episodes have been recognised along the central part of the North Tabriz fault.

(1) The compressional regime (subarea 1: stress A) is configured by σ₃ as the steepest principal stress and a horizontal σ₁, and its time span is limited to post-Cretaceous and pre-Early Miocene tectonic activities (Laramian structural stage). The uplift of Moro Mountain is mainly the result of this stress regime.

(2) The Early Miocene extensional regime (subarea 1: stress C) is characterised by a sub-vertical σ₁ and horizontal σ₂ and σ₃.

(3) The youngest, post-Miocene, tectonic regime (Pasadenian structural stage) appears as different stress regimes along different segments of the NTF: (a) the extensional regime (subarea 1: stress B) is characterised by σ₁, σ₂, and σ₃ as sub-horizontal, steepest, and horizontal principal stresses, respectively; (b) the compressional regime (subarea 2: stress A) is characterised by σ₁, σ₂, and σ₃ as horizontal, sub-vertical, and sub-horizontal principal stresses, respectively; and (c) the extensional regime (subarea 3: stress A) is characterised by σ₂ and σ₃ as steepest and sub-horizontal principal stresses, respectively.

On the whole, according to the fault-slip data collected from 88 localities, inversion analysis of the fault-slip data, and stratigraphic, geometric, and structural information used in this study, transcurrent deformation is prevalent in the NTF and adjacent areas and is generally accomplished by predominant NW-SE-trending dextral and NE-SW-trending sinistral faults.