Seismotectonics, Geomorphology and Paleoseismology of the Doroud Fault, a Source of Seismic Hazard in Zagros

Abstract

In this study, the active tectonics, paleoseismicity, and seismic hazards of the Doroud Fault are examined through high-resolution satellite image interpretations, field investigations, outcrop and trench excavations, and the dating of geochronology samples. The Doroud Fault (DF), one of the essential segments of the Main Recent Fault in the northern margin of the Zagros mountain range, has a historical and instrumental background of high seismicity. We present the first constraints from tectonic geomorphology and paleoseismology along the Doroud Fault near the capital city of Dorud. Detailed observations from satellite imagery, field investigations, real-time kinematic (RTK) measurements, paleoseismological trenching, the radiocarbon (C14), and optically stimulated luminescence (OSL) as ages allowed us to map the fault in detail, describe and characterize its kinematics, and document its recent activity and seismic behavior (cumulative displacements, paleoseismicity, and magnitude, as well as recurrence interval) relevant to the recent seismic activity of the Doroud Fault during the late Holocene as one of the most important seismogenic faults in Zagros. Modern alluvial terraces of gullies and loess accumulations are systematically deflected and/or offset with co-seismic rupture, landslides, and scarps, indicating that the Doroud Fault has been active in the late Quaternary and is characterized by dextral strike–slip movements with a normal component. In addition, our findings provide a comprehensive analysis of the fault displacement, the timing of paleoearthquakes, and the right-lateral slip rate of the Doroud Fault. The late Holocene slip rate of the Doroud Fault using the OSL dating the gully is as follows: the minimum and maximum horizontal slip rates are estimated to be 1.82 and 2.71 mm/yr, and vertical slip rates of 1.03 and 1.53 mm/yr are calculated for the past 4600 ± 900 years in the middle segment of the fault. This study focused on a paleoseismological trench within the archeological sites of Darbe-Astaneh. The central portion of the fault has historically hosted more than nine earthquakes in the last 66 ka years, according to the study’s findings. According to paleoseismology studies, the Doroud Fault has the seismic capability to cause earthquakes with a magnitude of more than 7.4 and a total slip rate of about 3.83 ± 0.1 m. The average recurrence interval for the identified paleoearthquakes is approximately 104 ± 7 years.

1. Introduction

Paleoseismology studies with a geomorphologic viewpoint on tectonically active areas are essential for modern seismic hazard assessment and determining the seismic potential of seismogenic faults over much more extended periods than instrumental measurements [1]; moreover, these studies provide tightly constrained data for the location, size, and recurrence of large earthquakes associated with surface rupture in recent geological time, which are critical for the assessment of seismic potential and hazards along active faults [2,3,4,5]. Strike–slip faults have played a critical role in the development of paleoseismology for several reasons. First, strike–slip faults are often the longest faults on continental landmasses and typically have conspicuous geomorphic expressions [6,7]. Second, many of these faults have long records of seismicity because they pass through populated continental regions and have experienced surface ruptures during large and great historical earthquakes. Third, because coseismic deformation along strike–slip faults is horizontal, subsequent earthquakes do not deeply bury or expose erosion or traces of earlier events [5]. Using earthquake geology, paleoseismology, and archaeology in regions with long historical records, we can determine earthquake construction, potentially slippery conditions, and slip allocation over multiple earthquake cycles [8]. Paleoseismic research aims to ascertain past earthquake activity by locating fault movements, rupture sites, and slip/ground shaking strength distributions. One of the most fundamental principles of paleoseismology is reconstructing the sequence of past earthquakes in relatively young geological units [9]. One of the primary goals of earthquake geology and seismology is to determine the earthquake recurrence interval, which permits probabilistic predictions of the time and magnitude of future earthquakes along a fault [10,11,12,13]. Iran, located in the Alpine-Himalayan belt, is one of the most seismically active areas on the planet. Its seismic activity is distributed along the same active fault lines (Figure 1), with individual faults rupturing every few hundred to thousands of years during strong earthquakes with a magnitude higher than 7.4 [14,15,16,17,18,19,20,21,22]. Iran experienced 145 earthquakes with a magnitude higher than 5.5 between 1900 and 2022, which resulted in 190,300 fatalities. More than 1000 people were killed in 17 of these earthquakes, 11 of which had a magnitude of 7 or higher [23]. The magnitude and time of the last seismic events are crucial to figuring out the earthquake’s recurrence. Understanding a fault’s past behavior can help predict how it might behave in the future, which can help with loss mitigation and seismic hazard estimation.

The Zagros mountain range (Figure 1a), one of the main structural zones of SW Iran, which includes a large number of Iran’s devastating earthquakes, is a fold-and-thrust belt at the leading edge of the Arabia–Eurasia continental collision zone, which accounts for about one-third of the collision zone’s overall 20 to 30 mm/year convergence rate [25,26]. The perceptive impact of large-magnitude earthquakes and the intervening destructive episodes of historical and antique cultural heritage sites located in earthquake-prone areas near active faults (especially in the Zagros) are of scientific, architectural, engineering, cultural, and social significance. The Main Recent Fault (MRF), located on the northwestern side of the Zagros Mountains (Figure 1b), is the most active fault system in the region where the convergence of the Arabian and Eurasian plates is oblique to the range’s trend [27,28,29]. The oblique convergence is consummated by the spatial separation of the strike–slip and reverse components on parallel faults [30,31]. It seems that a significant portion of the strike–slip mechanism is spent along the MRF [32]. The trend of the MRF along its length changes from ~330° in the northwest, between Nahavand and Borujerd, to ~300° in the central part and to ~315° in the Doroud–Borujerd basin, where it ends in frequent splay faults. Based on the geomorphology along with the northwestern and southeastern general segments of the MRF, as well as the focal mechanism of earthquakes, some researchers (e.g., [30]) suggest that the normal slip vector is approximately parallel to the middle segment (Doroud Fault), which has experienced destructive earthquakes (Figure 2a). The Doroud Fault, throughout its length, has cut through Quaternary sedimentary units, the Bakhtyari formation (Pliocene–Pleistocene), Khanehkat–Neyriz formations (Triassic–Jurassic), Gerey and limestone rock units (Triassic–Jurassic), and Dalan (Permian) and Mila formations (Cambrian–Late Ordovician) (Figure 2b). Within and on the border of the rock units, the Doroud Fault trace with a deformation of the geomorphic features can be clearly seen. In this paper, the geomorphology and paleoseismology of the Doroud Fault, as well as its geometry and kinematics, are investigated. The primary goal of this research is to improve seismic hazard assessments, thereby providing the means to mitigate the eventual devastation caused by large earthquakes resulting from the activity of the Doroud Fault. A crucial factor in determining the seismic potential of active faults is the slip rate. The lateral slip rate of strike–slip faults is estimated by reconstructing the offset of linear features with known ages. Geomorphological analysis of surface features, such as offset channels [33,34,35,36,37,38], can be used to figure out the offsets, or paleoseismological trenching can be used to measure the offsets from linear features that are buried, such as paleochannels [39,40,41,42,43,44]. In this study, paleoseismology and geomorphology were used appropriately to determine the slip rate of the Doroud Fault in order to understand its seismic behavior and accurately assess the earthquake risk in SW Iran.

2. Tectonic Setting and Seismicity

The Zagros mountain range in southwest Iran (Figure 1a) expanded from ~250 km to ~400 km in width and ~1500 km in length from SE Turkey (in the west) to the Hormuz Strait (in the east) (Figure 1). This range was formed in a series of tectonic events related to the collision of the Arabian plate with the Eurasia plate in the Miocene [45,46]. The Main Zagros Reverse Fault (MZRF) and MRF delineate the Arabia–Eurasia suture that limits the northern extension of the Zagros range (Figure 3) [16].

The main structural features of the Zagros and its evolution history are known by various researchers [16,45,47]. During the Arabia–Eurasia collisional process, the sedimentary rock units in the northern domain of the Arabian plate have been folded and detached from its Pan-African basement into a long belt (Zagros fold-thrust belt). To the north of the Zagros fold-thrust belt and parallel to it is the Sanandaj–Sirjan Zone, and the Urumieh–Dokhtar magmatic arc, which consists of imbricated sedimentary and metamorphic rocks, and Andean-type volcanic rocks, respectively [47].

The obduction of the Neo-Tethys oceanic crust onto the northern margin of the Arabian plate in the late Cretaceous has been the first tectonic event in the development of the Zagros orogenic belt. In the late Campanian, the obduction finished while remains of the Neo-Tethys oceanic crust continued to subduct beneath Central Iran in the Oligocene–Miocene. Based on the NUVEL-1A plate motion model [28], currently, the collisional stage is active with a convergence rate of ~30 mm/yr in the north–south direction. However, the GPS proposed a velocity of ~20 mm/yr of the Arabian plate related to Central Iran [25]. The GPS indicates about half of the convergence is consumed by Zagros. The rates of convergence along the Zagros belt are seen to be continuously decreasing from SE to NW, from ~9 mm/yr in SE Zagros to ~6 mm/yr and ~4 mm/yr in Central and N Zagros, respectively.

The multisegmented Zagros Main Recent fault is a major NW–SE-trending right-lateral strike–slip seismic fault zone of more than 800 km long located between the Zagros fold-and-thrust belt in the southwest and Central Iranian range-and-basin in the northeast [16,20,32]. The fault trace in the northern margin of the Zagros range depicted by Ambraseys and Melville [48] is equivalent to what is now known as the MRF, but there was no field evidence for the strike–slip component; only vertical displacements were observed (north-east side down). The MRF is divided into five major segments (from Kamyaran to Arjanak), including the Doroud, Nahavand, Garun, Sahneh, and Morvarid segments [32] (Figure 3). In the last century, several destructive earthquakes, such as Silakhorin 1909.01.23 with Mw7.3; Razan 1955.12.04, UTC: 14:02, epicenter: 33.50 N–48.80 E Ms 6.0; Heydarabad 1961.10.28, UTC: 10:46:33.54 N–48.49 E and Ms 5.0; Sahneh 1958.08.14, UTC: 15:26 and Ms 5.5; Nahavand 1958.08.16, Mw 6.6, UTC: 19:13 and epicenter: 34.25 N–47.84 E; Dinevar 1958.09.21, UTC: 16:18 and Mw 5.2; 1987.05.29, epicenter: 34.06 N–48.28 E, Ms 6.5 and Night/Wednesday; Farsineh 1957.12.13 with Mw 6.5; Nevahand 1958.08.16 with Mw = 6.7; Karkhaneh 1963.03.24, UTC: 12:44, epicenter: 34.43 N–47.89 E and Mw 5.8; and Chalencholan 2006.03.31 with Mw = 6.1, have occurred on these segments (Figure 3) [25]. The most destructive was the 1909 earthquake (Ms 7.4) associated with a fault rupture over 45 km long [32]. It was felt in a large area where 128 villages were damaged and it caused more than 8000 people to be killed [18,48].

The event Chalencholan 2006.03.31 with Mw~6.1 (with 68 deaths and ~1400 injured) had a consistent focal mechanism of the right-lateral strike–slip action, and it ruptured the Doroud segment of the MRF. According to relocation models, the dextral slip of up to 90 cm was recorded at a depth of ~4 km in the 2006 Chalencholan event. Based on Peyret et al. [49], the maximum horizontal displacement near Chalencholan city is suggested to be ~12 cm. In any case, geomorphological evidence shows 50–70 km of right-lateral displacement along the MRF [30].

3. Method and Data

To determine the slip rate of the Doroud Fault, the offsets of tectonic features were analysed using field observations, topographic data, Google Earth satellite images, and OSL dating results [50,51,52]. The topographic data were obtained using real-time kinematic measurements (RTK). Given that the horizontal uncertainties for RTK data is <0.5 m, measurement errors on the displacements mainly arise from the determination of markers and construction of fitting lines [13,53]. We thus chose potential research sites based on the interpretation of a digital elevation model (DEM) and Google Earth satellite images; moreover, before digging trenches, Holocene and Pleistocene rock units were carefully separated, the fault trace and surface geomorphic features around the fault were mapped in the field, and topographic cross-sections from RTK data were drawn (using The Environment for Visualizing Images (ENVI)) software across the fault scarps at the site. To study paleoseismology, we excavated a trench (see its location in Figure 4; TC (Trench)) using both hand and machine tools across the fault scarp and pond, and mainly along the middle segment of the Doroud Fault. We chose the pond to excavate a trench because it is a location of accumulation of recent sediments that certainly record seismic events, and this pond was formed due to the action of the fault in the hanging wall of the Doroud Fault and is the closest location to the epicenter of the 1909 earthquake. A string grid defining 1 m × 1 m panels was strung along the length of the trench walls after the walls were scraped and cleaned. Then, each grid panel was photographed with four images that slightly overlapped with one another. The sedimentary layers in each trench were classified as separate units based on color, particle size, sorting, texture, bedding thickness, and other characteristics, such as the colluvium wedge. The sequence of ground-rupturing events was then determined based on strata cross-cutting relationships, sediment thickness variations, soft sediment deformation, and fissures. The radiocarbon (C14) and optically stimulated luminescence (OSL) ages of sediment samples collected from trenches were utilized to constrain the ages of the stratigraphic units and, consequently, the interpreted events.

4. Geomorphology and Slip Rate of the Doroud Fault

The use of morphological offsets to determine the long-term slip rate on an active fault requires datable morphologic features, which were first displaced by the fault motions and then preserved over the observable displacement [54,55,56]. We used the satellite image of the satellite interpretations of the digital elevation model (DEM) at 10 m, SASPlanet satellite images (http://www.sasgis.ru), and the field observations by reconstructing the satellite image and measuring in the field; we came to the conclusion that the Doroud Fault has caused the maximum cumulative horizontal (dextral) offsets in the Pliocene–Quaternary units of the rivers by ~775 m and ~30 m (Figure 5 and Figure 6). The observed displacements are associated with the southeastern and middle segments of the Doroud Fault, which ruptured during the 1909 Silakhor earthquake [16,18,32,48]. Geomorphic evidence, such as dextral surface ruptures including beheaded gullies (Figure 6c) and shutter ridges (Figure 6d), is visible within the mezoseismal area of the 1909 earthquake, which demonstrates the fault zone in this region (Figure 6 and Figure 7) and its relationship to the late Pleistocene–Holocene activity of the Doroud Fault. However, the surface expression of the strike–slip fault is poorly exposed in the study area, with only a few possible nominations. These offsets, ranging from several meters to a kilometer, are observed from southeast to northwest.

Slip rate studies are landform-based paleoseismic studies, the aim of which is to calculate a long-term fault slip rate from landforms offset by multiple events. They are usually reconnaissance studies on poorly known faults and precede trenching studies. Slip rate studies can be performed at widely varying spatial and temporal scales [5]. These different sizes of offset features propose an excellent opportunity to better document the seismic behavior and the occurrence of past large earthquakes on the Doroud Fault. Several displaced gullies close to the TC trench site (Figure 4) afford an opportunity to establish the fault slip rate. We document offset terraces, and river channels at three sites using the resulting high-resolution topographic map allow one to restore the offset channels to simple linear patterns along the fault, illustrating cumulative dextral offsets along the fault trace on the order of about 30 m (Figure 6).

It is difficult to determine the long-term slip rate along the Doroud Fault because of the intense human activity (e.g., agriculture fields and constructions) which has changed the original geomorphology in many places. However, a few geomorphologic features can be used to analyze the tectonic activity of the fault. In the central part of the fault, 3 km east of Doroud city, close to the Darbe-Astaneh village, the features have been displaced by multiple earthquakes. Therefore, the offsets represent the cumulative amounts of surface movements. We used kinematic GPS (real-time kinematic/GNSS) to obtain accurate displacement measurements, and in order to obtain the time of sedimentation from the gully in Figure 7c, alluvial sediments of the present age consisting of gravel, sand, and silt were sampled, which were analyzed by the OSL, and the age of 4600 ± 900 was obtained. The obtained values were used for the calculation of fault parameters and the preparation of digital elevation models. Our analysis was focused on three shifted crest lines in the central part of the fault (S1, S2, and S3 in Figure 7a,b). Using the trigonometric functions Vf = V/sinδ and ρ = artg (Vf/H), where H is the horizontal component, V is the vertical component, Vf is the vertical component projected along the fault, δ the dip, and ρ the rake (see Figure 7e) [57], we estimated a rake angle for the slip vector comprised between 20 and 40 degrees. We constrained the kinematics of the fault at a smaller scale within its southeastern part. The offset drainages were measured both in the field with a tape and, on the high-resolution topographic map obtained by a real-time kinematic (RTK) survey devoted to a systematic survey of offset drainages (Figure 7d), S1 and S2 in Figure 7a,b have cumulative horizontal offsets of ~22 m and ~14 m and cumulative vertical offsets of ~14.35 m and ~6.8 m; the cumulative horizontal and vertical offsets of S3 are ~10.5 m (A’B’ in Figure 7d) and ~5.7 m (C’D’ in Figure 7f), respectively (

Table 1. The mean of the displacements obtained in the gully area due to Doroud Fault activity. H(m) = cumulative horizontal displacement; Vmax (m) = cumulative displacement of the slope on the ground; Vf (m) = displacement of the slope collapse on the fault surface; R (Radian) = rake angle; NS (m) = cumulative displacement fault slip (net slip).

Number	H (m)	Vmax (m)	Vf (m)	R (Radian)	NS (m)
1	21.59	6.82	6.48	17	22.58
2	13.94	14.34	13.48	44	19.40
3	10.05	5.92	5.57	29	11.49

). In order to calculate the amount of slip rate, we used the data of S3.

In Figure 8, we presented a schematic model of the Doroud Fault, which cuts the gully in the long fault in S3. We discussed the ages of these geomorphic markers using OSL dating (Dt1 Post IRSL gully) to derive the slip rate averaged over the late Pleistocene–Holocene. Based on the formula (Average slip rate = slip distance/slip time), slip rates were calculated thus [58]:

10.5/4600 ± 900 = 2/28 ± 0.42 mm/yr

5.7/4600 ± 900 = 1.29 ± 0.23 mm/yr

The results indicate the horizontal slip rate of the Doroud Fault, in the period of the last 4600 ± 900 years, is 2.28 ± 0.42 mm/yr (between 2.7–1.86 mm/yr) and the vertical slip rate is ~1.29 ± 0.23 mm/yr (between 1.52–1.6 mm/yr).

5. Palaeoseismic Trenching along the Doroud Fault

Previous studies related to the analysis of the seismic activity of the Doroud Fault have been carried out to investigate the pre-shocks and aftershocks of the two major earthquakes in 1909 (Ms 7.4) and 2006 (Ms 6.1) using Coulomb stress and InSAR data [49]. Investigating the fault ruptures and seismic gaps in the northern part of the fault [32] indicates the fault has the ability to have a magnitude 7.1 earthquake. Therefore, the Doroud plain (or Silakhor plain) is considered to be highly active in the duration of tectonic activity, according to seismic catalog data [59,60,61]. To retrieve the chronology of surface-rupturing paleoearthquakes, we conducted the first paleoseismological study along the Doroud Fault at a site called Darbe-Astaneh (Figure 4 and Figure 9). This site was selected based on geomorphological observations performed from a 3-D model obtained from the GPS (RTK) surveying and field surveys (Figure 9). The trench was excavated perpendicular to the general trend of the fault (Figure 9a,b), where our DEM map and topography profile show a fault scarp (Figure 9c–e).

The Darbe-Astaneh trench was excavated across the fault within the wet pond (Figure 9). The sag pond’s best depositional environments for preserving paleoearthquake evidence are relatively low-energy environments where sediments accumulate episodically in thin strata, separated by weathering profiles, organic soils, or peats. The stratigraphy in the sag pond is dominated by the continuous quiet-water deposition of finely laminated clayey silt with some interbeds of well-sorted sand. The depositional environment and the general absence of bioturbation because of the high groundwater conditions permit the cm-scale resolution of individual units [5]. The trench strikes N40° E and has a length of 30 m, a depth of 4.5–5 m, and a width of around 1.5 m. Once the trench was excavated, the walls were cleaned to remove trench smears and gouges created by the backhoe, and field photographs were shown at the same time as the excavation of the trench (Figure 9f,g). The excavation exposes a succession of fine-grained, well-bedded layers interbedded with a few gravelly units. The almost fine-grained deposits are primarily horizontal and composed of laminated silts, sands, and gravel. Based on the grain size and the location of trench excavation within the pond, 27 units are defined within the trench. Many potteries and other artifact fragments were found in the stratigraphic units after careful logging, including pieces of charcoal, which were collected for radiocarbon C₁₄ analysis (Figure 10), and the place was determined with sediments containing quartz and feldspar for sampling and dating by the OSL method (collected from units 2 and top 19 and colluviums wedge C1) (

Table 2. Values used to calculate luminescence ages (OSL).

Sample ID	De (Gy)	±	Total (Gy/ka)	±	Age (ka)
Dt1 Post IRSL river	14.91042	2.74	3.23	0.08	4.61
Dt1 IRSL 2	231.5025	16.55	3.48	0.09	66.62

). Several faults were highlighted in zones A and B of the trench (Figure 10 and Figure 11), between meters 1 to 4 and 18 to 22.

Paleoearthquakes are typically recognized in exposures of strike–slip faults from six general types of evidence: (1) upward termination of fault displacement, (2) abrupt changes in the vertical separation of strata as faults are traced upsection or downsection, (3) abrupt changes in the thickness of strata or of facies across a fault, (4) fissures and sand blows in the stratigraphic sequence, (5) angular unconformities produced by folding and tilting, and (6) colluvial wedges shed from small scarps [5]. To decipher the diverse paleoseismic events, we found evidence 1, 2, 3, and 6 to identify the events in the trench. In our paleoseismological investigation of the trench photomosaic and logs (Figure 11 and Figure 12), the main fault plane was easily identifiable because the sediments display distinctly different colors as the hanging wall appears, allowing us to propose at least nine paleo-earthquakes from the top to base of the trench in two fault zones (Figure 13 and Figure 14). They are named Ev1′ thought Ev9 from youngest to oldest, which can be identified as the event that occurred between about 1225 and 1275 years and 71 to 61 ka years, respectively (

Table 3. Geometrical and seismic parameters of the Doroud Fault (obtained from TC studies).

U Event	Age Cal BP	Event Horizon	Units	V (m)	H (m)	Ds (m)	Ss (m)	Net Slip (m)	MW (AD) SS	MW (MD) SS	MW (AD) N	MW (MD) N	MW (AD) G	L (Km)	F (Km)	A (Km2)	M0g (dyne cm)	MW
I’	Before 1250–1360	Base of Unit 2	2	3.83 ± 0.10	3.35 ± 0.1	1.76 ± 0.1	11.57 ± 0.1	3.83 ± 0.1	7.55	7.03	7.15	7.03	7.4	49	15	7.3512	1.3127	7.4
II’	Before 1360–3440	Base of Unit 2	2	Fissure size similar to Ev1′	3.35 ± 0.1	1.76 ± 0.1	11.57 ± 0.1	3.83 ± 0.1	7.55	7.03	7.15	7.03	7.4	49	15	7.3512	1.3127	7.4
III	After 1560	Base of Unit 22	22	0.093 ± 0.1	0.14	0.08 ± 1	0.02 ± 0.1	0.19 ± 0.1	6.4	6.03	6.31	6.22	6.33	49	15	7.3512	1.8125	6.1
IV	Before 1560–1720	Base of Unit 22	22	Fissure size similar to Ev1	0.16 ± 0.1	0.08 ± 0.1	0.02 ± 0.1	0.19 ± 0.1	6.4	6.03	6.31	6.22	6.33	49	15	7.3512	1.8125	6.1
V	Before 1720–2385	Base of Unit 16-a	16-a	0.194 ± 0.10	0.34 ± 0.1	0.18 ± 0.1	0.12 ± 0.1	0.4 ± 0.1	6.68	6.26	6.5	6.39	6.6	49	15	7.3512	4.6325	6.6
VI	Before 2385–3235	Base of C1	C1	-	-	-	-	-	-	-	?	?	?	?	15	?	?	?
VII	Before 3235–8375	Base of Unit 4	4	1.04 ± 0.10	1.87 ± 0.1	0.98 ± 0.1	3.61 ± 0.1	2.14 ± 0.1	7.33	6.84	6.99	6.85	7.2	49	15	7.3512	1.3127	7.2
VIII	Before8375–66620	Base of Unit 3	3	0.63 ± 0.10	1.13 ± 0.1	0.59 ± 0.1	1.31	1.29 ± 0.1	7.13	6.68	6.85	6.71	7.02	49	15	7.3512	3.1626	7
IX	Before 66620	Base of Unit 2	2	0.5 ± 0.10	0.9 ± 0.1	0.47 ± 0.1	0.84 ± 0.1	1.03 ± 0.1	7.05	6.6	6.78	6.59	6.94	49	15	7.3512	2.1226	6.9

The quantities derived from measurements on three trenches and their restored logs. V means vertical separation; Ds, dip slip = (V.Sin^–1δ); Ss, strike slip; Sn, net slip = (Ds² + Ss²)0.5; pitch or rake equals Arc Cos (Ss/Sn); L, rupture length; D, focal depth; A, rupture Area = L × F; Mw, energy magnitude or moment magnitude that equals 2/3 LogM_o^g–10.7, where the seismic or geologic moment equals μ × A × Sn, and here is the shear or rigidity module = 3 × 10¹¹ (dyne/cm²). E.g.M_o^g = 3 × 10¹¹ (dyne/cm²) × (49 × 10⁵ (cm) × 15 × 10⁵ (cm) × 0.21× 10² (cm) = 4.63 × 10²⁵ (dyne cm); Mw = 2/3Log4.63 × 10²⁵–10.7 = 6.41.

).

The parameters related to the activity of the fault on the surface and subsurface calculated applying empirical relationships (Equations (1)–(6)) for strike–slip faults are listed in

Table 4. Parameter survey results for AD, SRL, RLD, RW, RA, and u using equations (Equations (1)–(6)).

Event	Mw	AD (m) [59,60,61,62]	SRL (km)	RLD (km)	RW (km)	RA (km2)	U (m)
I’, II’	7.4	2.35	70.79 ± 0.21	95 ± 0.7	29.5 ± 0.35	2137 ± 10	0.74
III, IV	6.1	0.089	12.88 ± 0.21	14 ± 0.7	7.9 ± 0.35	117 ± 10	0.71
V, VI	6.6	0.19	19.49 ± 0.21	22 ± 0.7	10.96 ± 0.35	234 ± 10	0.72
VII	7.2	1.65	61.65 ± 0.21	81 ± 0.7	25.7 ± 0.35	1479 ± 10	0.74
VIII	7	0.81	41.68 ± 0.21	53 ± 0.7	19.05 ± 0.35	870 ± 10	0.73
IX	6.9	0.54	33.88 ± 0.21	41 ± 0.7	16.21 ± 0.35	602 ± 10	0.73

.

Average displacement (m) [62]

M = 7.04 ± (0.05) + 0.89 ± (0.09) × log (AD)

(1)

Surface rupture length (km) [63]

Mw = a × log (L) + b, a = 1.67, b = 4.24

(2)

Rupture area (km²) [62,64]

M = 3.98 ± (0.07) + 1.02 ± (0.03) × log (RA)

(3)

Subsurface rupture length (km) [62]

M = 4.33 ± (0.06) + 1.49 ± (0.05) × log (RLD)

(4)

Subsurface rupture width (km) [62]

M = 3.8 ± (017) + 2.59 ± (0.18) × log (RW)

(5)

Maximum slip (m) [65]

Log (u) = 4.38 + 1.49 × log (Ls)

(6)

M: Magnitude, normally Mw or equivalent, Mw: Moment magnitude (Mw = 2/3 log M0 − 10.07), [66] L: Fault rupture length (km), a and b: Coefficients and standard error, RA: Rupture Area (km²), RLD: Subsurface rupture length (km), RW: Subsurface rupture width (km), u: Maximum slip, Ls: surface rupture length.

While our study focused on the transtenstive Doroud Fault in the Zagros region, we acknowledge that blind faulting may also play a role in the partitioning and accommodation of plate movement in the area. Blind faults are faults that do not reach the surface, and their existence and activity can be difficult to detect without direct geological or geophysical evidence. In some cases, blind faults can contribute significantly to the total slip rate and seismic hazard of a region.

In the Zagros region, blind faults have been identified and studied in some areas, and their potential impact on seismic hazard and tectonic deformation has been discussed in previous research. However, the exact extent and significance of blind faulting in our study area are not well-known, and further investigation would be necessary to fully understand its role in the tectonic evolution of the region.

While our study did not explicitly consider the impact of blind faulting, we acknowledge that it could be an interesting topic for further investigation, particularly in relation to the partitioning and accommodation of plate movement among different structures. We believe that this point should be discussed in our manuscript to provide a more comprehensive context for our study and to highlight the potential avenues for future research in the area.

The recognition of the most recent earthquake event (Ev1′) is based on three independent lines of evidence. First, we have the abrupt upward terminations of two fault strands (F1′ and F4′), capped by the most recent pond deposit, in Log fault zone A (latest event) (Figure 13a,b). These faults are younger than 25 ± 1250 Cal BP, the age obtained for sample DT1-C10 collected in Unit 2. However, note that the base of unit 2 is displaced 6 ± 180 cm along the faults (F1′ and F4′); it may be a cumulative clastic deposited by different events, hence the fault units 1-1′ and 1-2′ cutting and fault strands at the base of unit 1-2′. These two fault strands, one near 0′ meter and the other near 3′ meter, and the event horizon are at the base of unit 2. which is overlain by units 3 and 4 (Figure 15a,b). This event occurred after the deposition of units 1-2. Therefore, using the parameters of the fault plane (Table 3), the magnitude of the earthquake that caused this event (according to [51]) is obtained as Mw~7.4 (Table 4 and

Table 5. The seismic catalog of the Doroud Fault correlated by results of paleoseismology. Historical and instrumental data for earthquakes with magnitude more than MW > 5.5.

Mw	Time	Number
6.1	2006.03.31	1
7.4	1909.01.23	2
7.4	740 ± 37 AD	3
7.4	654 ± 9 AD	4
6.1	488 ± 42 AD	5
6.6	488 < EQ < 3158 AD	6
-	1497 ± 28 < EQ < 470 ± 53 BC	7
7.2	7448 ± 57 BC < EQ < 1497 ± 28 BC	8
7	EQ > 7448 ± 57 BC	9
?	EQ > 16019 BC	10
?	20186 < EQ < 16019 BC	11
6.8	EQ > 66620 BC	12

). In earthquake rupture, the average length of the surface rupture and the average subsurface length were 79.79 and 87.09 km, respectively; the average width of the subsurface rupture was 29.5 km, the average area of the rupture was 2137 km², and the average slip was 0.74 m (according to [51]).

The recognition of a second paleoearthquake is located in Log fault zone A, a fissure fill localized at the base of units 1-2′, and shows the occurrence of an earthquake before the deposition of units 1-2′ (Figure 13c,d). This event is well-characterized by two faults F2′ and F3′, which are younger than 1360 ± 35 BP, which is the age obtained for sample DT1-C9 collected in unit 1-1′; therefore, this event postdates 1360 ± 35 BP and predates 3440 ± 25 Cal BP. The event horizon of this event is the base unit 1-2′ (Figure 15a,b); the earthquake magnitude of this event and other parameters related to the seismic activity of the fault is similar to the event Ev_1′, which is given in Figure 10.

A third event, Log fault zone B, can be interpreted along the fault F2 ruptured unit 22. This event occurred after the deposition of unit 22, which affected the youngest-dating unit 22. Hence, it postdates 1560 ± 255 Cal BP. This fault is younger than 3976 ± 152 cal BP, the age obtained for sample number DT1-C4, collected from unit 22 (Figure 10). Therefore, the event horizon of this event is the base of units 22 and 21 (Figure 13e (post-event), f (pre-event) and Figure 15c,d). The calculated moment magnitude is Mw ~6.1 (Table 3).

The fourth event is identified with faults F4, F9, and F10, a tiny fracture in Log fault zone B. This event can be interpreted along fault F4 and the fissure filled with sedimentary deposits of unit 22 and the faults F10 and F9, the abrupt upward terminations of two fault strands sealed by the surface soil of this fissure with the maximum depth and width, which are 34 cm and 7.27 cm, respectively (Figure 13h). These faults and fractures are younger than 1720 ± 30 Cal BP, which is the age obtained for sample DT1-C7 collected in unit 16-a. The event horizon of this event is the base of the 16-a unit (Figure 15c,d). Moreover, this event indicates the occurrence of an earthquake after the time of the deposited unit 16-a (Figure 13g (post-event), h (pre-event)). The calculated moment magnitude is Mw ~6.1 (Table 3), and the average surface and average subsurface rupturing lengths were 12.88 and 14 km, respectively. The average subsurface rupture width was 7.9 km, the rupture area was 117 km², and the average slip was 0.71 m (according to [51]) (Table 4).

The fifth event can be interpreted with the cutting units 4 and C1 and rupturing along F2 (Figure 14a (post-event), b (pre-event)) in Log fault zone B. The event horizon of this event is the base of unit 22. This event occurred after the deposition of unit C1. Hence, it postdates 2385 ± 35 Cal BP, the age obtained for sample DT1-C6 collected in unit C1, and predates 1720 ± 30 Cal BP, which is the age obtained for sample DT1-C7 collected in unit 16-a (Figure 10) (Figure 15c,d). Its calculated moment magnitude is Mw~6.6 (Table 3). Moreover, the average surface and average subsurface rupture lengths were 19.49 and 22 km, respectively, the average subsurface rupture width and average rupture area were 10.96 km and 234 km², and the average slip was 0.72 m (according to [51,58,59] (Table 4).

The sixth event was identified with colluvial wedge unit C1 in Log fault zone B, which overlies and has eroded into the underlying units 1, 2, 3, and 4, as a colluvial wedge developed following a surface rupture in Event 6. Unit 4 is cut by fault F2. The colluvial wedge deposits resulting from the degradation of a fault scarp erosion which formed as a result of the normal component of the F2 fault (Figure 14c (post-event) and d (pre-event)); therefore, the deposits of wedge C1 are the result of the erosion and sedimentation of a mixture of units 4 and 16-a indicating the earthquake occurrence before the deposited unit C1 and after unit 4 (Figure 15c,d). We evaluated the maximum wedge thickness of 21.74 cm, which shows a vertical displacement of 19.43 cm at the top of unit 4 (Table 3). The event horizon of this event is the base of the C1 unit. This event occurred after the deposition of unit 4, which is a colluvial wedge sample DT1-C6, dated 2385 ± 35 Cal BP. Hence, it postdates 2385 ± 35 Cal BP (Figure 10). The seventh event is identified with faults F1, F2, F3, and F4 in Log fault zone B; the faults affect units younger than unit 4. These faults are younger than 3235 ± 25 Cal BP, which is the age obtained for sample DT1-C5 collected in unit 4 (Figure 14e (post-event) and f (pre-event)). This event occurred after the deposition of unit 4; hence, it postdates 3235 ± 25 Cal BP (Figure 10). Several distinct abruptly terminating upward fault strands, with the most demonstrative ones observed between 18 and 20 m, provide evidence for event Ev7 (Figure 16a,b). The faults F1, F2, F3, and F4 cut through unit 4 and abruptly terminate at the base of unit 16-a. Unit 4 is associated with average vertical displacements along the faults F1, F2, F3, and F4 with a range of 100 ± 4 cm (Table 3). The event horizon has therefore been set at the base of unit 4. The calculated moment magnitude is Mw ~7.2 (according to [51]) (Table 4 and Table 5), and other calculated seismic parameters are given in Table 4.

The eighth event is evidence of faults abruptly terminating upward and cutting units along the faults in Log fault zone B. This event was found only in the southwestern part of zone B and well-characterized by the abruptly terminating upward fault F11 at the base of units 3 and F10, cutting through the lower units 3, 2, and 1 (Figure 14g (post-event) and h (pre-event)). Unit 3 is associated with vertical displacements along the F10 that are about 60 ± 3 cm (at the base of unit 3) (Table 3). Therefore, the event horizon of the seven paleoearthquakes is set at the base of unit 3 (Figure 16c,d). This event occurred after the deposition of unit 3; hence, it was younger than 8375 ± 35 Cal BP, which is the age obtained for sample DT1-C20 collected in unit 3 (Figure 10). Using the fault plane parameters (Table 3) and the same equation linking them to the earthquake magnitude obtains Mw ~7, according to [51] (Table 4), and other calculated seismic parameters are given in Table 4.

The ninth event is indicated by unit 2 being offset by faulting in Log fault zone B. This event can be interpreted as a fault termination, with the fault F5 cutting through unit 2 and faults F6, F7, and F8 abruptly terminating upward at the base of unit 2. This event is associated with vertical displacements in unit 2 along the F5, which was calculated to be an average of 50 ± 0.1 cm (at the bottom of unit 2) (Table 3) (Figure 14i (post-event) and Figure 14j (pre-event)). Therefore, the event horizon is set at the base of unit 2 (Figure 15c,d). This event occurred after the deposition of unit 2. Hence, it is younger than 66.62 ± 5 ka, which is the age obtained for sample Dt1 IRSL 2 collected in the middle of unit 2 (Table 2). Using the fault plane parameter Table 4 and Table 5 and the same equation linking them to the earthquake magnitude obtains Mw~6.8 (according to [51]) (Table 3).

6. Discussion

The geomorphologic features of the strike–slip fault system along the MRF zone and their effects on the outcrop patterns of lithological units, rivers, and gullies show the recent right-lateral strike–slip motion and activity of the MRF. Active tectonics of this fault have also caused large earthquakes. Considering recent activities along the DF, as described above, it is important to conduct an accurate evaluation of the slip rate and offset on this active fault; regarding the geological markers, rivers, and gullies, the maximum offset on the DF is about 775 m, which may be observed as a right-lateral motion displacement of the river in south Bagh-Lotfian in the southern segment of DF. However, there is clear morphotectonic evidence related to the late Pleistocene–Holocene activity of the Doroud Fault on satellite images. So far, many studies have been carried out to estimate the slip rate on the Main Recent Fault: Talebian and Jackson [30] and Walpersdorf et al. [67] suggested 10–17 and 4–6 mm/yr, respectively; Bachmanov et al. [68] 10 mm/yr; Copley and Jackson [69] and Vernant et al. [25] 2–5 and 3 mm/yr slip rate, respectively; Authemayou et al. [70] 3.5–12.5 mm/yr; Nankali [71], using GPS velocities, expanded a three-dimensional mechanical model containing pre-existing faults with frictional properties and suggested a long-term slip rate for the MRF (~2.3 mm/yr); and Alipoor et al. [59] estimated 1.6–3.2 mm/yr. Our work is the first slip rate study along the Doroud Fault. To determine the slip rate in the middle segment of the Doroud Fault, GPS RTK data were utilized (discussed in Section 4). We conclude that the average cumulative dextral offset, horizontal and vertical, on the Doroud Fault are about 10.5 m and 5.7 m, respectively, leading to a slip rate, horizontal and vertical, of about 2.28 ± 0.42 mm/yr and 1.29 ± 0.23 mm/yr, respectively.

Our work is the first paleoseismological study along the Doroud Fault. The Darbe- Astaneh trench site is located near the epicenter of the 1909 earthquake with the mapped rupture (Figure 4). This site was selected based on remote sensing and field observations for its relatively high sedimentary potential. It delivered well-expressed surface deformation and adequate deposits for age determinations. This paleoseismic investigation along the Doroud Fault demonstrates the occurrence of several large, infrequent, and irregular earthquakes. Indeed, in our trench, nine paleoearthquakes were identified; the fault hosted at least eight large (Mw ~5.5) earthquakes within the last 8375 ± 35 Cal BP, and one older earthquake probably before some 66.62 ± 5 ka, and the youngest occurred at 1250 ± 25 Cal BP. Indeed, the fault hosted four earthquakes of more than Mw ~7 and four of more than Mw ~6 within the last 66.62 ± 5 ka. The occurrence of seven seismic events within a short time interval may indicate that the seismic behavior of the Doroud Fault can be characterized by the clustering of earthquakes. Figure 17 is a time diagram that summarizes the observations from our trenching studies. The occurrences of seven large earthquakes during the last ~3.3 ka, with a magnitude of more than an Mw which can be attributed to ~6, suggest that the Doroud Fault could be characterized by a clustering of events (A). The two oldest events of the nine events identified occurred at 8350 ± 35 to 66.62 ± 5 ka, respectively, in the absence of macroseismic activity in the time interval between 3260 Cal BP to 8350 Cal BP, and in the time interval between 8350 Cal BP and 66 ka, two seismic gaps are observed, which can be attributed to a high rate of erosion, sedimentation, or unconformity. In paleoseismology, the non-preservation of sedimentary records is a recognized issue that can affect the identification of seismic events. Seismic gaps identified in our study may be due to the non-preservation of sedimentary layers that would have recorded the event rather than the absence of seismic activity. However, our results still demonstrate a bimodal behavior of seismic activity in the area, as illustrated in Figure 17, which shows a higher frequency of seismic events during the recent interval and a lower frequency of events during the past interval.

To account for the potential impact of non-preservation on our results, we carefully examined the sedimentary records from the intervals we studied, and we found no evidence of seismic activity during those periods. While non-preservation may still bias our interpretation of the seismic gaps, our results remain significant as they demonstrate a clear difference in the frequency of seismic activity between the two intervals. We acknowledge the limitations of our study due to non-preservation and encourage future studies to account for this issue by carefully examining the sedimentary record and considering alternative methods of seismic event identification.

From the youngest paleoearthquake related to fault zone A with the age of 1250 ± 25 BP (calendar year BP) to the 1909 earthquake, which is about 700 ± 37 years old, according to paleoseismic results and historical earthquakes from the Berberian book [12], an earthquake with even Mw~5.5 has not been reported along the Doroud Fault, so in the 14th to 20th centuries, the Doroud Fault had no seismic activity, while other segments of the MRF had seen seismic activity. Therefore, seismic migration occurred from the southeast to the northwest. There is evidence of this migration of a cluster of events destroying Dinor in 913.08.07 and 1008.04.27: 34.61 N–47.50 E, Ms~7 and time: night/Sunday, and 1107.08.22, center: 34.60 N–47.50 N, leaving a twentieth-century seismic gap to the southeastern part of the Nahavand segment, which itself was ruptured during the 1316.01.05, center: 34.08 N–48.40 E and Ms~6.5 earthquake [62]. Then, during the twentieth century, the 1909.01.23, center: 33.38 N–49.28 E, UTC: 02:48, with Mw~7.4 Silakhor earthquake along the Doroud segment of the Zagros Main Recent Fault was followed by the 2006.03.31, Mw~6.1, UTC: 01:01cent: UTC: 33.58 N–48.79 E 23 January 1909 Mw 7.4 31 March 2006 Mw 6.1 Chalanchulan earthquake; the latter filled the gap left on the northwestern segment of the Doroud segment by the 1909 event. The two major events in 1909 and 2006 on the Doroud Fault in a time interval of approximately 100 years indicates the beginning of a new seismic cluster. According to the obtained results, the seismic behavior of the Doroud Fault is a cluster, and the recurrence interval of this behavior is about 700 ± 37 years obtained using historical and instrumental earthquakes (Figure 17). Using paleoseismic and seismic data can suggest an almost complete seismic catalog for events with a magnitude greater than 6 for the Doroud Fault (Figure 18 and Table 5). However, the geometrical, kinematical, and seismic behavior characteristics of the Doroud Fault show that this fault has the ability to create earthquakes with Mw~7.4 and surface ruptures.

7. Conclusions

Our analysis of the Doroud Fault in southwest Zagros has produced new insights into its activity; in general, the fault is a hidden strike–slip, which also provides the first estimate of the slip rate on the Doroud Fault and suggests that large earthquakes (6.6 ≥ Mw ≥ 7.4) have occurred along this fault in the past as a result of its activity. Minimum and maximum horizontal slip rates of 1.82 and 2.71 mm/yr and minimum and maximum vertical slip rates of 1.53 and 1.03 mm/yr for the past 4600 ± 900 BP are estimated. The fault’s relatively high slip rate indicates its significance in accommodating dextral displacement and crustal shortening. Trenching investigations along the Doroud Fault indicate approximately nine paleoearthquakes that may have occurred in the last 66.62 ± 5 ka years. We estimate that, from the results of paleoseismology, the average recurrence interval of earthquakes with a magnitude greater than 5.5 is ~104 ± 7 yr. Considering the earthquake that occurred in 2006, the next quake is expected to occur within the next ~90 years up to the first decades of the 22nd century AD. According to paleoseismology studies, the Doroud Fault has the seismic capability to cause earthquakes with a magnitude of more than 7.4 and a total slip rate of about 3.83 ± 0.1 m, and another earthquake magnitude and slip rate are listed in Table 3. An extensive surface rupture with an approximate length of 45 km has been predicted to result from large earthquakes on the Doroud Fault.