Seismotectonics and Seismic Source Parameters of the Mid-Eastern Iraq-Western Iran Using Moment Tensor Inversion Technique

The study area is encompassed by the 33.59-34.93°N latitudes and 45.44-46.39°E longitudes and divided into four groups with respect to earthquake event locations. We determined fault plane solutions, moment magnitudes, focal depths, and trend of slip with the direction of the moment stress axes (P, N, and T) for 102 earthquakes. These earthquakes had a local magnitude in the range between 4.0 and 6.4 for the time period from January 2018 to the end of August 2019, with focal depths ranged between 6 and 17 km. Waveform moment tensor inversion technique was used to analyze the database constructed from seismic stations on local and neighboring country networks (Iraq, Iran, and Turkey). We separated the studied events into four regional subsets (circles). The types of the obtained fault plane solutions are predominantly thrust fault and strike-slip, with the focal depths ranging from 8 to 21 km. A new scaling relation between local magnitude (Ml) and the estimated moment magnitude (Mw) has been developed utilizing a linear regression. Good match results obtained in the present research good match with both seismic trends concluded from earthquake locations and mapped faults. Generally, direction shows NW–SE striking focal planes corresponding with the tectonic framework of the Arabian–Eurasian continental collision zone. The anticlockwise rotation of the Arabian plate that appears accountable for strike-slip displacements on fault surfaces.


Introduction
Earthquakes are generally concentrated on the boundaries of tectonic plates [1].Important information about the tectonic features and structural styles of the subsurface geological structures could be considered from the analysis of earthquakes. Most of the Iraqi previous studies of the seismicity and seismotectonic parameters indicated that the levels of seismicity are moderate to high at northern and northeastern boundaries while decreasing towards the southern and southwestern [2,3]. Focal mechanisms are the most effective tool in seismology to study the relative magnitudes and orientation of the stresses emitted during earthquakes. Seismologists refer to the direction of slip in an earthquake and the orientation of the fault on which it occurs as the focal mechanism. Different types of focal mechanisms can be described, which are displayed graphically by the so-called "beach ball" diagram. One of the fundamental techniques in studying active continental motions is seismic source analysis along with earthquake focal mechanism evaluation. These focal mechanisms could assist to determine the state and orientations of tectonic motions and to give a better understanding of the current mode of tectonic deformation in a certain region. The study area is located in the collision tectonics between the Arabian and the Iranian plate (Figure-1) between latitudes 33.59-34.93°N and longitudes 45.44-46.39°E (Figure-2). These earthquakes have a moment magnitude of 4.0≤ MW≤6.4 that occurred in the region on the beginning January 2018 to the end of August 2019. Most events are located in the mid-eastern Iraq and western part of Iran, near to the borders. The reasons for most events occurring in the region are probably the stresses generated by the Arabian plate movement to the north and northeast and its collision with the Iranian-Turkish plateau, also in addition to the impact of neo tectonic activation of the upper crust [4]. Database from the region was subdivided into four groups; AA (Belula-Ezgeleh), BB (Khanaqin-Qasre Shirin-Sarpol Zahab), CC (Mandali-Samoor), and DD (Qolqoleh-Halol), according to the spatial distribution of the events (Figure-3). The aim of this study is to use moment tensor inversion technique to determine fault plane solutions, moment magnitudes, and focal depths according to a previously described approach [9].

Tectonic and Structure Setting
Iraq is located at the extreme northeastern part of the Arabian Plate, which is in collision with the Eurasian Plate and surrounded by regions of relatively high seismicity. This collision still continues forward and has caused alignment of the gradually developed structures in NW-SE trend mainly, particularly in the northern, northeastern and eastern sides of Iraq. However, many transversal linear elements of the NE-SW trend, represented through rivers, streams, valleys, playas, anticlines and offsets, are developed, in a parallel trend to the major compressional forces created by the aforementioned collision [10]. In recently active tectonic areas, the topography of the land surface is an accurate reflection of the underlying structure and its interaction with surface processes [11]. Large red and blue colored symbols simplify divergence and convergence movements with overall amount and age, respectively. Smaller green arrows show present-day GPS values with respect to fixed Europa from Iran and white arrow from Oman [5]; triangles represent geographic distribution of seismic stations (Table-1  2-Regional tectonic subdivision of Middle Eastern Iraq -western part of Iran. Names within the parentheses are known in the Iranian part of Zagros (modified after [6]); [7,8]. The black rectangle represents the study area plotted in Figure-3, and the magnitude of the 102 earthquakes is indicated by the diameter of the colored circles. Generally, the Zagros Fold-Thrust Belt (including the study area) is one of the most seismically active belts in Asia. The SW-Vergence fold-thrust belt and circumferential foreland basin were formed as a result of the Arabia-Eurasia collision during final closure of the Neotethys Ocean [12].
The intensity of deformation, age of the sedimentary units and amount of shortening in the study region showed a reduction from north and northeast (Zagros Suture Zone) in the direction of SW (Foreland region) [13].The NW portion of the Zagros Fold-Thrust Belt is subdivided into zones depending on the topography and deformational style, with the zones from NE to SW are Zagros Suture Zone, Imbricated Zone, High Folded Zone, and Foothill Zone ( Fig. 2 in [14]). One of the morphologically generality distinguished, structural elements of the Zagros Fold-Thrust Belt is the Mountain Front Flexure (MFF), which separates the High Folded Zone and the Foothill Zone (these zones are known as a part of the Zagros Simply Folded Belt and Zagros Foredeep in Iran, respectively; Fig.2). These include the folding of sedimentary rocks which are expressed at the surface by a series of anticlines and synclines which dominate the range physiography [15]. The Foothill Zone is divided by the Kirkuk Fault into the Chemchamal and the Hemrin Subzone [14]. Major NW-SE trending and SW dipping back thrust faults are also one of the feature structures in this zone [14,16,17,18,19]. Tectonic research reported that the Iranian plateau features a high density of energetic and recent faults along with reverse faulting with large strike-slip component (i.e., dextral-reverse oblique slip near the impinging zone). Pure thrust faulting tends to dominate the tectonics of the region toward central Iran [20]. Territorial shortening along the major thrust systems, that are accountable for producing strong earthquakes in northeast Iran, is about 1-2 mm/year, using optically stimulated luminescence (OSL) dating of river deposits [21]. The GPS data denote that the average of north-south shortening from Arabia to Eurasia is nearly 2-2.5 cm/year [22]. Nevertheless, the continental convergence between Arabia and Eurasia is stationary at 2 to 3 cm/year since 56 Ma [23].

Data set preparation and processing
From January 2018 to August 2019, the IRSC catalogue has 102 earthquakes in the study area at depths of 06-17 km with 4.0 ≤ Ml ≤ 6.4. The waveform data, gathered from five local seismic stations and contiguous networks including 39 stations, belong to the Iraqi Meteorological Organization and Seismology (IMOS), North Iraq Seismographic Network (NISN), Iraq Seismological Laboratory of University of Basrah (SLUB), Iranian Seismological Center (IRSC), and Kandilli Observatory and Earthquake Research Institute (KOERI)-Turkey ( Fig.1 and Table-1). Most broadband stations usually record 100 samples per second. A perfect azimuthal coverage is taken into consideration around the source, being very desirable and critical to such focal mechanism solution studies. The epicentral distances between the stations and seismic events were less than 600 km. The origin parameters of the events were chosen from the seismic catalogue of the Iranian Seismological Center (IRSC) and listed in Table-2. The original seismic data format is in MiniSEED or SEISAN format (i.e., IMOS, IRSC and NISN) and needed to be transformed into SAC and then GSAC formats. The header file for all data files required preparation before the execution of the analysis software. An instrument response correction was applied for each seismic event during the processing steps, using deconvolution response file for each seismometer. Three components (Vertical Z, North-South N, and East-West E directions) were presented for each recorded event. These three components needed to be rotated from a station-based recording system (ZNE) to an event-based recording system (ZRT). The ZRT rotation system is a 2D rotation where the Z component is still pointing in the same direction as in the original ZNE recording, while the two horizontal components N and E are rotated into the radial R and tangential T components, respectively.

Methodology
The Computer Programs in Seismology (CPS) version 3.30 was used in the present work as previously described [9]. LINUX platform was also applied, which includes 149 programs that can be run using scripts. In general, the requirements of performing the data and source inversion routines in seismology are waveform data, instrument response, station locations, earthquake location, velocity model, Green's functions and results assessment. O. Time (UTC) original time in UTC, Lat latitude in degree, Log longitude in degree, RD reported depth, Ml reported magnitude, CD calculated depth measured in kilometer (km), M W moment magnitude, Mo seismic moment measured in dyne-cm, S strike, D depth measured in km, R rake angle, P, N, and T compressional, normal, and tensional moment stress axes, respectively, PL plunge angle, AZ azimuth.

Green's functions (GF)
is created and described based on model assumptions or empirically gained data and includes wave propagation from the seismic source to a receiver, in response to specific source excitation. In the framework of source inversion problems, they are used to build synthetic seismograms, which are compared to real surface data records in the moment tensor inversion [24]. The use of the suitable regional velocity model is significant not only to correspond to the waveforms but also to define the moment magnitude of the earthquake, because the theoretical amplitudes at high frequencies depend very highly on the velocity model. Synthetic waveforms were calculated using a 1D-velocity model (Table 3) which is modified from the global IASPEI91 model in crustal part. This synthetic waveform was used later for calculating Green's functions needed for moment tensor inversion [25] Figure-4.

Moment Tensor Inversion of Seismic Waveforms
Moment tensor is a mathematical characterization of seismic source that depends on wave propagation, earth model and synthetic seismograms. Moment tensor solution method is used to give a better assessment of seismic moment energy released from earthquakes and to identify the stress regime system and faults orientation for different seismic sources in the region. Moment magnitude (Mw), Seismic Moment (Mo) and data for seismic source process are acquired by this procedure. Least square fitting of amplitude and/or waveform data can be derived from seismograms moment tensor components [26]. The waveform inversion method was used, which is based on the grid search method that involves the inspection of over all possible focal mechanism solutions. In general, the steps of performing moment tensor inversion are shown in the flowchart (Figure-5). The steps are explained in detail in appendix B of a previous study [27]. The moment tensor solutions have a quality factor assigned by the number of stations used during the inversion and the goodness of fit between synthetic and observed data. The inversion technique was applied as in the example in Figure-(6a). Seismic stations were used for the moment tensor inversion method solution of the event 2018/01/11 08:00:39 UTC. A grid search over the strike, dip and rake angles were used for every depth from 0.5 to 39 km in increments of one km, as shown in Figure-(6c). Figure-(6b) indicates the correlation and percentages between the observed (red traces) and predicted (blue traces) values.  The three components of the seismic record are R (Radial), Z (Vertical), and T (Transverse). Waveforms were plotted using the same scale. Peak amplitudes were indicated by the numbers to the left of each trace and contain pair of numbers that indicate the time shift required for maximum correlation between the observed and predicted traces. Percentage of variance reduction was used to characterize the individual goodness of fitness (100% indicates a perfect fit). To improve the signal-tonoise ratios (SNRs), the waveforms were filtered in frequency bands that were individually adjusted according to the data characteristics and quality. All traces in Figure-(6b) represent ground velocity (m/s) filtered in the 0.05-0.15 Hz band. Some traces were also removed because of noise, which may be due to site effects and instrumentation (i.e., in KGS1, AMR2 and KAR2 stations, only the vertical component remained, while traces R and T components in Figure-(6b) were rejected. Correlation and percentages between observed (red) and predicted (blue) traces. The three components of the seismic record are R (Radial), Z (Vertical), and T (Transverse). Each observed-predicted component is plotted using the same scale and the numbers to the left of each trace indicate peak amplitudes. A pair of numbers is given in black at the right side of each predicted traces indicate: (i) the upper number, the time shift required for maximum correlation between the observed and predicted traces and (ii) percentage of variance reduction to characterize the individual goodness of fit. (c) The best fit as a function of depth sensitivity. The best fit value is 0.5560 and indicates a depth of 18 km. (d) Waveform inversion focal mechanism, the beach-ball shows the best solution at this depth.

Classification of Focal Mechanism Solutions
Detailed information on the resulting focal mechanisms are listed in (Table-2) and include the strike orientation, dip angle, rake angle of both nodal planes, the direction of slip, and the orientations b d of the moment stress axes. These are calculated by a moment tensor inversion technique for 102 events. The beach balls were plotted for each group using color coding according to the focal mechanism type: red for normal with strike slip faulting, green for strike slip, blue for thrust pure or with strike slip faulting, and gray for unknown or oblique type that indicates the maximum horizontal stress azimuth that is not defined. Figure-7 shows the beach balls for the group AA (Belula-Ezgeleh) plots for northeastern Iraq which have a dominance of thrust with strike slip (TS) and a few strike-slip (SS) focal mechanisms, while the plots for the groups BB (Khanaqin-Qasre Shirin-Sarpol Zahab) and CC (Mandali-Samoor) show a predominance of thrust (TF), thrust with strike slip (TS) and some strike-slip (SS) focal mechanisms. Finally, all focal mechanisms with a group DD (Qolqoleh-Halol) have a relatively strike-slip (SS), except for one event with a normal with strike slip (NS). In general, the largest number of earthquakes with the highest energy happens in the upper 40 Kilometers of the earth's crust [28]. Regarding the focal mechanism depths solution of earthquakes in the study region, it ranged from 8 to 21 km, while the reported focal depths ranged between 6 and 17 km ( Table 2).  Table 2). The lines colors indicate the type of faulting: red, normal with strike slip (NS); blue, thrust or thrust with strike-slip (TF or TS respectively); and green, strike-slip.

Moment stress axes
The moment stress axes (P, N, and T for, respectively, maximum shortening, neutral axes and maximum extension) are perpendicular to each other from both the fault and the auxiliary planes. The same is correct for σ1, σ and σ3, only in the state of new fracture generation in a homogeneous isotropic medium [29]. These are calculated for focal mechanism solutions of each seismic event. The P axis bisects the compressional dihedron (45 from both focal planes) [30]. Table-2 shows moment stress axes depending on their azimuth and Figure-7 shows that, in the study area, the major compressional orientation was mostly is NE-SW and W-E, with NW-SE in northern and northeastern parts of the region.

Moment magnitude (Mw) versus local magnitude Ml
Magnitude is used as an expression for energy releases from an earthquake. Most of the instrumental data around the world use the local magnitude (Ml) scale that covers the entire magnitude range from potentially less than zero and up to 6 or more. The suitable conversion from local magnitude (Ml) to moment magnitude (Mw) is a significant prerequisite for any seismic-hazard assessment [31]. Local magnitude is the primary magnitude scale calculated in the department of Seismology in Iraqi Meteorological Organization and Seismology for the 102 events in this study, which ranged from 4.0 to 6.4. Moment magnitude derived from moment tensor techniques, ranged from 3.67 to 6.11 ( Table 2).
The developed relation between local magnitude (Ml) and moment magnitude (Mw) is derived by a linear relation, as shown in Figure-8 and expressed as follows: =0.8877) where: a = 0.962, b = 0.0534 and R² represents the correlation coefficient.
In most of the events, the value of Mw from moment tensor inversions was less than the reported Ml. The average difference was in the order of 0.33 magnitude units. There is a feasible acceptance for events with magnitudes of 4 to 6.5, indicating that the Ml and magnitude body wave (mb) are equal or more than Mw [32].

Discussion
In this paragraph, we will discuss the results of the moment tensor inversion of all events, their tectonic context, and the active tectonic structures. For each event, moment tensor inversion was used to determine source mechanisms from the regional waveforms. The use of individual focal mechanisms for the assessment of the tectonic stress is not a direct solution, since the degree that they sample intrinsic tectonic episodes varies. This is due to several uncertainties related to inherent ambiguities in the definition of the fault plane and slip direction or to the accuracy of the individual focal mechanisms and the respective inversion methods' assumptions. Previous studies proposed that the earthquakes in the study area likely happen on the Mountain Flexure Fault (sometimes referred to as Main Front Fault, noted MFF in Figure-2) on blind reverse faults buried under or within a thick, folded sedimentary cover [33]. Along the major part of the Zagros belt, the MFF follows a NW-SE axis with a ∼ 120• azimuth and corresponds to many topographic lineaments. The next sections will debate the focal mechanism or the fault plane solutions, moment magnitudes, and the focal depths in each sub region. 1. Belula-Ezgeleh (group AA): The Northern part of the study region is located on the border of Iraq-Iran within the Zagros fold and thrust belt, which delimits the continental collision between the Arabian and Eurasian Plates. Only six focal mechanisms exist for this sub-region. The resulting fault plane solutions for all the studied earthquakes demonstrate thrust faulting with a strike-slip (5events) and strike slip pure (one event). Stress tensor suggests that compression is homogeneous in a nearly horizontal NE-SW direction and the combination of thrust with strike-slip faulting. 2. Khanaqin-Qasre Shirin-Sarpol Zahab (group BB). This group represents the largest number of events (45 events). Most of them are inside Iran and a few on the Iraq-Iran border, located within the Zagros fold and thrust belt, which is a very seismically active region. A tectonically effective region that accommodates crustal shortening resulting from the collision between the Arabian plate and the Eurasian plate is observed as a differential tectonic movement divided between different types of faults [34]. Local structural complexities may explain why the Qasre Shirin sequence broke in earthquakes series of moderate events rather than as a single thoroughgoing larger rupture with NW striking segments, which is parallel to the regional trend in the fold axes [35]. Thrust faulting with almost all compression axes in the northeastern to southwest orientations characterizes the solutions, including a larger proportion of thrust and thrust with strike slip focal mechanisms, while a few pure strike slip and unknown or oblique slips were present (Table-2 (Table-2).

DD (Qolqoleh-Halol):
The Qolqoleh-Halol is located in the western part of the study area. All the 22 seismic events are located inside Iran between High Zagros Fault and Main Recent Fault. This is a major NW-SE, 800-km-long right lateral strike-slip fault that accommodates some of the strain [36] ( Figure-2).These events form a right lateral strike-slip focal mechanism but are located more than 100 km west of the Main Recent Fault. This group is completely different from the previous groups with its distinction, as 21 focal mechanisms are dominantly strike-slip, with only one event that has a normal faulting. All of these mechanisms reveal a general trend of NW-SE orientation.

Conclusions
The present application of the CPS software package allowed us to compute seismic moment tensors for 102 earthquakes by using broadband waveform inversion with an Ml between 4.0 and 6.4. The most distinguishing characteristic in the focal mechanism distributions in the study area is their depth dependence, thus, all of the earthquakes according to the results have crustal depths ranging from 8 to 21 km. The major stress compressional (P) axes orientations are determined from moment tensor solutions of the locally recorded earthquakes that occurred in the study area. Deformation style of the groups (AA, BB, and CC) suggests predominant thrust faulting, thrust with strike-slip faulting component, and some unknown or oblique faulting that indicates compressional or transpressional tectonic environment. However, strike-slip earthquakes (SS) predominate to the west (group DD). Nevertheless, earthquakes are distributed in the study area of faulting in the Zagros range, which is amongst the most rapidly deforming and seismically active fold-and-thrust belts on earth. The majority of seismicity are occurring on the blind reverse faults buried under or within a thick, folded sedimentary cover. Therefore, the distribution of earthquakes provides vital information about the location of active faulting at depth.
Good matching results were obtained in the current study with both seismic trends (concluded from earthquake locations) and mapped faults. These matching results are corresponding with the tectonic framework of the Arabian-Eurasian continental collision zone and the anticlockwise rotation of the Arabian plate that appears accountable for strike-slip displacements on fault surfaces. Also, a relation between moment and local magnitude was obtained appropriately for the investigated area and magnitude range.