1 Introduction

The East Anatolia Fault Zone (EAFZ) is identified as a left-lateral strike-slip fault system, arising from the interactions among the Anatolian Plate, Eurasian Plate, African Plate, and Arabian Plate [1, 2]. This fault zone consists of the primary East Anatolia Fault (EAF) located on the southern branch and the Surgu-Misis Fault (SMF) situated on the northern branch (refer to Fig. 1).

The dynamics of the EAFZ are influenced by the relative motion of the Anatolian Plate, Arabian Plate, and Eurasian Plate, with the Anatolian Plate advancing southwestward at a rate of 15 mm/year relative to the Arabian Plate along its southeastern boundary [3]. This movement contributes to the left-lateral strike-slip kinematic behavior of the EAFZ [2] and has precipitated a series of significant seismic events, including notable earthquakes such as Mw7.8 (1114, 1822), Mw7.4 (1513), Mw7.8 (1822), Mw7.2 (1866), Mw7.1 (1874), Mw7.1 (1893), Mw6.8 (2020), and the double earthquake sequence of Mw7.8-Mw7.6 in 2023 [4,5,6,7,8,9,10]. The Mw7.8-Mw7.6 double earthquake event that occurred on February 6, 2023, stands as the most significant seismic event in the region since the 20th century. This event was characterized by a mere 9-hour interval between the two major earthquakes, a distance of less than 100 km between their epicenters, and a rupture scale extending hundreds of kilometers, resulting in intense ground shaking. This occurrence marks a notable deviation from the relative absence of significant seismic activity in the EAFZ since the 20th century and has attracted considerable attention from the academic community.

Fig. 1
figure 1

Geological structure and historical strong earthquake distribution map of the study area. The blue dashed line represents the plate boundary, the white lines represent fault distribution, and the red lines represent the main faults of the EAFZ (EAF: F1: Karlıova, F2: Ilıca, F3: Palu, F4: Pütürge, F5: Erkenek, F6: Pazarcık, F7: Amonos) and SMF (F8: Sürgü, F9: Cardak, F10: Savrun) [1, 5]; the green circles represent historical strong earthquake events, and the beach balls represent the focal mechanism solutions of three strong earthquake events since the 20th century, with red and green beach balls representing the focal mechanism solutions of the 2023 Mw7.8-Mw7.6 double earthquake, and the gray beach ball representing the focal mechanism of the Mw6.7 earthquake in 2020 (USGS). The gray arrows represent GPS deformation [3], and the black arrows represent the relative motion directions of the Anatolian Plate and the Arabian Plate with respect to the Eurasian Plate; the study area is indicated by the red dashed box on the right, and the green lines represent the rupture areas of historical strong earthquakes [9]

Full size image

Geological, seismological, and paleomagnetic studies indicate that the EAFZ displays significant segmentation characteristics [11]. The southern branch of the EAFZ, known as the EAF, extends approximately 580 km and comprises seven segments, each ranging from 31 to 112 km in length, with a strike orientation between approximately 35 degrees and 75 degrees eastward, all classified as Holocene faults. In contrast, the northern branch, referred to as the Sinjar Fault (SMF), spans roughly 380 km and consists of nine Holocene fault segments, predominantly exhibiting left-lateral motion [1, 11]. The double earthquake sequence of 2023, with magnitudes Mw7.8 and Mw7.6, exemplifies the intricate fault activity resulting from the interactions of three tectonic plates at the Hatay triple junction [12], with the Mw7.8 event occurring along the EAF and the Mw7.6 event along the SMF [13].

Prior to this double earthquake event in Türkiye, seismic activity within the EAFZ region typically exhibited characteristics of seismic gaps, localized clustering, and widespread dispersion [6]. Bulut et al. [4] conducted an analysis of the spatiotemporal characteristics of earthquakes in the northeastern segment of the EAFZ (south of latitude 37 degrees) using precise relocation data and focal mechanism solutions from 2000 to 2007, revealing a gradual interaction between the primary fault of the EAF and secondary structures. A regional-time-magnitude analysis of seismic and tectonic parameters identified an anomalous area characterized by low b-values and high Z-values in the central EAFZ, indicating a potential for significant seismic events in this region [14, 15]. Güvercin et al. [6] further analyzed seismic tectonic activity in the EAF utilizing precise relocation catalogs and 160 focal mechanism solutions from 2007 to 2020, concluding that the seismic activity pattern of the EAF is influenced by pronounced geometric irregularities, uneven coupling, and complex plate motions.

Following the double earthquake in Türkiye in 2023, numerous seismologists and geophysicists initiated a series of studies on this significant earthquake sequence [13], revealing the migration patterns of aftershocks and characteristics of seismic activity, which indicate a transition from locking to creep in the EAF, alongside notable morphological differences between the EAF and SMF.

By synthesizing data from multiple sources, including far-field seismic waveforms, GNSS/GPS observations, strong motion records, and SAR imagery, kinematic and co-seismic slip models have been developed for the Mw7.8-Mw7.6 double earthquake that occurred in Turkey [9, 16]. The findings suggest that the Mw7.8 event demonstrates super-shear propagation along the primary rupture segment, while the Mw7.6 earthquake is initiated by a combined dynamic-static mechanism. The underlying factors contributing to the high magnitude and destructive capacity of the Mw7.8-Mw7.6 double earthquake include initial branch triggering, collaborative rupture across multiple fault segments, and super-shear propagation.

Research on the b-value derived from historical earthquake catalogs indicates a significantly low b-value characteristic in the epicentral region of the main shock in Turkey, which suggests a high stress state [17,18,19]. Furthermore, the area exhibiting a low b-value spatially coincides with subsequent high slip zones [19], while local regions continue to display a low b-value during the aftershock period, indicating that stress has not been entirely released [17]. Ongoing monitoring of the persistently low b-value areas and the notably quiet Z-value regions is essential [18]. Yin and Jiang [20] assessed the probability distribution of migration risk associated with the Mw7.8-Mw7.6 double earthquake activity by analyzing b-value changes through cumulative image migration methods. A general decline in b-values is observed in the vicinity of the source region prior to the occurrence of the double earthquake events, with a pronounced high probability image migration distribution near the earthquake nucleation point, suggesting an increase in differential stress within the crust that elucidates the nucleation process of the double earthquake.

Wang et al. [21, 22] conducted an analysis of the seismic environment associated with the Mw7.8-Mw7.6 double earthquake in Turkey, employing a multi-physical parameter approach through seismic tomography. Their findings indicate that the mechanisms underlying these double earthquake events are influenced by variations in deep structural features. To further investigate the spatiotemporal evolution of aftershocks, site amplification effects, and the structural characteristics of the East Anatolian Fault Zone (EAFZ) in relation to the Mw7.8-Mw7.6 double earthquake, Zor et al. [23] undertook extensive seismic array observations. Current research predominantly emphasizes the accurate localization of aftershocks, the processes of rupture, and the deep structural attributes of the source region associated with the Mw7.8-Mw7.6 double earthquake sequence. While previous studies have examined the b-value of seismic activity before and after the occurrence of the double earthquake, these analyses have primarily relied on the original earthquake catalog. There remains a notable gap in comprehensive analyses of the seismic activity characteristics within the EAFZ region, particularly those based on high-precision earthquake catalogs derived from long-term sequences.

This research aims to conduct a precise relocation of seismic events in the EAFZ region from January 1, 2010, to July 31, 2024, employing a multi-stage positioning strategy that integrates absolute positioning methods (HypoInverse) [24] and relative positioning methods (HypoTD) [25]. The distribution of b-values in the study area was calculated, and the stress field distribution was inverted utilizing 1,065 focal mechanism solution data surrounding the EAFZ. Through an analysis of long-term seismic activity, b-value distribution, and stress field distribution, this study elucidates the spatiotemporal characteristics of seismic activity in the EAFZ region, the relationship between earthquakes and fault structures, and the characteristics of the stress field, thereby providing a reference for a more profound understanding of the fault activity characteristics, precursory environment, and dynamic characteristics of the Turkish double earthquake sequence.

2 Tectonic background

The Anatolian Plate operates as a tectonic convergence zone, engaging with the Eurasian Plate, African Plate, and Arabian Plate (refer to Fig. 1). Its kinematic behavior is predominantly shaped by the ongoing northward compression from the Arabian Plate and the oblique subduction of the African Plate [3], which has resulted in significant regional tectonic activity [26]. This intricate, multi-directional tectonic stress field has facilitated the emergence of two primary strike-slip fault systems along the plate boundaries: the North Anatolia Fault Zone (NAFZ) and the East Anatolia Fault Zone (EAFZ) [27]. The EAFZ, located at the boundary between the Anatolian Plate and the Arabian Plate, extends southwestward, linking with the Mediterranean subduction zone and the Dead Sea transform fault [28]. It serves as the principal structural feature that delineates the boundary between the Anatolian and Arabian Plates, characterized by left-lateral strike-slip motion [29, 30]. The formation of the EAFZ is closely associated with escape tectonics resulting from the collision between the Arabian Plate and the Eurasian Plate [31, 32]. The complex tectonic interactions in this region have led to significant stress accumulation, creating conditions favorable for the occurrence of substantial earthquakes.

The EAFZ is acknowledged as one of the most seismically active regions in Türkiye. Research indicates that the segmentation and geometric configuration of the faults within the EAFZ, along with variations in the stress field, significantly influence the seismic activity in the area [33]. Historical seismic records demonstrate that this region has experienced multiple significant earthquakes, each with magnitudes of 7 or greater [28]. Since the 20th century, seismic activity has predominantly concentrated in the northern segment of the EAFZ, and following the Mw 6.7 earthquake sequence in 2020, there has been a notable southwestward extension of seismic activity. A recent significant seismic event was the Mw 7.8-Mw 7.6 double earthquake that occurred on February 6, 2023, resulting in considerable loss of life and property in Türkiye and Syria [9, 34]. This earthquake produced approximately 300 km of surface rupture along the EAFZ [29], and the double earthquake event has led to substantial changes in the regional stress distribution, potentially increasing the seismic risk in adjacent fault zones [35].

3 Seismic relocation

3.1 Data selection

This study compiled seismic phase data related to 105,690 earthquake events recorded by the Türkiye Disaster and Emergency Management Authority (AFAD) (https://deprem.afad.gov.tr/event-catalog) within the EAFZ and surrounding areas. The geographic scope of the analysis was defined by the coordinates 35.00˚E–41.00˚E and 35.75˚N–39.25˚N, covering the period from January 1, 2010, to July 31, 2024 (see Fig. 2). The dataset comprised two events with magnitudes of M7.0 or greater, three events with magnitudes ranging from M6.0 to 6.9, 74 events between M5.0 and 5.9, 861 events from M4.0 to 4.9, 6,451 events within the M3.0 to 3.9 range, and 98,299 events below M3.0. Overall, the analysis included 1,059,848 P-wave phases and 848,629 S-wave phases, with the majority of initial earthquake source depths concentrated in the 5–10 km range, resulting in an average travel time residual of 0.42 s.

Fig. 2
figure 2

Distribution of earthquake events and stations. (a) Distribution of earthquake epicenters, the points represent earthquakes, and the color of the points indicates earthquake depth, with the color scale shown in the lower right corner; white triangles represent stations, and white lines represent fault distribution; (b) Distribution of earthquake source depths; (c) Magnitude-time distribution of earthquake events; (d) Statistical distribution of earthquake source depths

Full size image

To improve the precision of earthquake relocation, the data selection criteria stipulated that the distance from the epicenter to the recording stations should not exceed 400 km, and that a minimum of four stations must have recorded each event. The distribution of recording stations effectively covered the range of earthquake occurrences, leading to a final dataset of 105,341 earthquake events and 126 seismic stations (as illustrated in Fig. 2). This refined selection included 1,051,048 P-wave phases and 841,463 S-wave phases (shown in Fig. 3a). The velocity model utilized for earthquake localization was developed based on previous studies conducted in the vicinity of the East Anatolian Fault Zone [6, 13] (depicted in Fig. 3b), featuring a P-wave to S-wave velocity ratio of 1.73.

Fig. 3
figure 3

Travel Time-Source Distance Distribution and One-Dimensional Velocity Model for Relocation. (a) Travel time-source distance distribution, where black points represent filtered phase data, and gray points represent excluded phase data; (b) One-dimensional velocity model used for relocation

Full size image

3.2 Multi-stage earthquake location

To improve the accuracy of earthquake location determinations, this study implemented a methodology that combines HypoInverse and HypoTD for the relocation of seismic events within the specified study area.

  1. (1)

    HypoInverse Absolute Location

HypoInverse is an absolute earthquake location technique that employs the least squares algorithm. This command-line based method provides multiple weight constraints, various configurations of velocity models, and a range of input-output format options. It includes functionalities such as the determination of earthquake locations, evaluation of location quality, magnitude calculation, and correction of station delays. Since its inception in 1978 [36], the HypoInverse method has undergone significant development and refinement, resulting in its extensive application [37].

Fig. 4
figure 4

Error Distribution of HypoInverse Absolute Location. (a) Horizontal error; (b) Vertical error

Full size image

In the process of HypoInverse absolute location, the weight assigned to P-wave phases was set at 1.00, while the weight for S-wave phases was established at 0.75. The initial horizontal direction utilized data from stations within a 120 km radius for location purposes. In subsequent iterations, stations within a 50 km radius were employed to constrain the depth of the location, with a minimum requirement of four station records for successful location of earthquake events. Ultimately, a total of 103,891 earthquake events were accurately located, accounting for over 98% of the total number of earthquakes analyzed. The error distribution following the location process is depicted in Fig. 4, which indicates an average horizontal location error of 1.50 km, a vertical average error of 2.87 km, and a root mean square (RMS) travel time residual of 0.30 s.

  1. (2)

    Relative Relocation

To improve the precision of earthquake event localization, we conducted a relative location analysis on the earthquake catalog subsequent to the absolute location process. This investigation employed the relative earthquake location methodology known as HypoTD [25], which exhibits reduced reliance on the velocity model, to re-locate the earthquake events following the absolute location. HypoTD is an advancement of the double-difference location algorithm HypoDD [38, 39] and the seismic tomography algorithm TomoDD [40]. The fundamental principle of the HypoDD method is to utilize travel time difference data from earthquake events for localization calculations, thereby diminishing dependence on the seismic location velocity model and minimizing errors associated with the common path of seismic rays. The HypoDD approach employs event pairs for travel time difference and cross-correlation time difference data to derive an inversion location, achieving a least squares solution by iteratively adjusting the vector differences between earthquake pairs to minimize the residuals between observed and theoretical travel time differences, thus facilitating accurate earthquake localization. The HypoTD method enhances HypoDD by incorporating station pairs and event pairs-station pairs data, enabling the utilization of diverse data sources for localization, including seismic travel times, event catalogs, and travel time differences from event pairs, station pairs, and event pair-station pairs. The constraints derived from event pair-station pair travel time difference data can concurrently enhance both absolute and relative location accuracy [25]. Furthermore, the 3-D pseudo-bending algorithm from the TomoDD method [41] was employed for ray tracing calculations and the determination of damping factors, further refining the relative location outcomes of earthquake events. This methodology has been successfully implemented in various regions, including the San Andreas Fault [25], the East Pacific Rise GTF [42], and the boundary regions of Northern and Southern California [43].

Fig. 5
figure 5

Error Distribution of HypoTD Relative Location. (a) X-direction location error; (b) Y-direction location error; (c) Z-direction location error

Full size image

In this study, we executed relative location analysis on the earthquake events following the HypoInverse absolute location, utilizing the HypoTD algorithm. The relative location process of HypoTD addressed 105,341 earthquake events categorized into 10 groups. In the construction of earthquake pairs, the distance from the event pair to the station was established at 250 km, with both the minimum number of observing stations and the minimum connection number set to 4, and a maximum source distance of 30 km, resulting in a total of 11,027,380 P-wave phases and 8,047,509 S-wave phases constructed. Given the substantial number of earthquakes in each relocation sequence, the conjugate gradient method was selected for location calculations, with P-wave and S-wave travel time weights assigned values of 1.0 and 0.75, respectively, and the location calculation was iterated five times. The inversion damping for location was determined through trade-off analysis. Ultimately, 101,966 earthquakes were relatively located, representing 98.15% of the total number of located earthquakes. The root mean square (RMS) travel time residual for location was reduced to 0.28 s, and the distribution of location uncertainties is illustrated in Fig. 5, with average location errors of 0.65 km in the X direction, 0.63 km in the Y direction, and 0.94 km in the Z direction.

3.3 Relocation results

The findings from the earthquake relocation analysis are illustrated in Fig. 6. Specifically, Fig. 6(a) presents the distribution of earthquake epicenters within the Eastern Anatolian Fault Zone (EAFZ), while Figs. 6(b) and 6(c) depict the earthquake distribution along the Eastern Anatolian Fault (EAF) and the South Marmara Fault (SMF) in terms of depth, with varying colors indicating different time periods in years.

In summary, the earthquakes within the EAFZ region are predominantly aligned along the EAF and SMF faults, with source depths primarily confined to within 20 km, indicating a predominance of shallow earthquakes. These events exhibit notable clustering characteristics in both temporal and spatial dimensions. Spatially, along the principal fault, the EAF, earthquakes are concentrated in segments F4 to F7, particularly on the northern side of the fault in the eastern segments F4 and F5. Furthermore, a southwest-northeast trending cluster of earthquakes intersects with the EAF on the northern side of segment F6 (EF), while segments F1 to F3 display a relatively sparse distribution of seismic activity.

In the northern branch of the EAFZ, specifically along the SMF, earthquakes in segments F9 and F0 are primarily located along the northern side of the fault, mirroring the overall shape of the fault distribution. Segment F8 exhibits minimal seismic activity, indicating a pronounced seismic gap; however, a cluster of earthquakes is present on the northern side of segment F8 (IJ), forming a band-like distribution parallel to segment F5 of the main EAF.

Regarding temporal characteristics, there are significant differences in seismic activity between the SMF and EAF. The seismic activity along the main fault, EAF, demonstrates a clear migratory pattern from northeast to southwest over time, along with distinct segmentation features. The aftershocks of the Mw6.8 earthquake in 2020 were predominantly concentrated in segment F4, whereas the aftershocks of the Mw7.8 earthquake in 2023 were primarily active in the area west of segment F4. In contrast, seismic activity surrounding the SMF did not exhibit a migratory trend over time, with earthquake occurrences primarily aligned along the rupture zone of the Mw7.5 earthquake in 2023, and the SMF displayed relatively low seismic activity prior to 2023.

Figure 6 illustrates the distribution of earthquakes along various profiles in the depth dimension, with the locations of these profiles indicated in Fig. 6(a). Specifically, Fig. 6(b) depicts the profile along the primary fault, the East Anatolian Fault (EAF), while Fig. 6(c) represents the profile along the northern branch fault, the Sinjar Mountain Fault (SMF). Each profile delineates the characteristics of earthquake distribution in the depth direction, utilizing earthquake events occurring within 2.5 km to the left and right of the profile line. The data indicates that seismic activity along both the EAF and SMF is predominantly shallow, with source depths concentrated between 5 km and 17.5 km. Notably, earthquake occurrences on the EAF are concentrated within fault segments F3 to F7, whereas segments F1 and F2 exhibit minimal activity. The distribution of earthquakes within segments F3 to F7 reveals both spatial and temporal gaps; specifically, there is a spatial gap between earthquakes in segments F5 and F6, and a temporal gap between earthquakes in segments F4 and F5, with segment F4 primarily characterized by events from 2020, while segment F5 reflects events occurring predominantly after 2022.

Fig. 6
figure 6

Relocated Earthquake Distribution Map. (a) Epicenter distribution map, where points represent earthquake epicenters, colors represent the time of occurrence, with the color scale in the upper right corner; black lines represent faults, and red lines represent the EAF and SMF; the histogram in the lower right corner shows the statistical distribution of earthquake source depths after precise location. (b) Earthquake distribution profile along the EAF; (c) Profile along the SMF; profile EF location is shown in Fig. 6(a), with profiles displaying earthquake events within 2.5 km to the left and right of the profile line, and white pentagons representing earthquakes of magnitude M5.0 and above

Full size image

On the SMF, seismic activity is primarily concentrated in segments F9 and F10. In the horizontal dimension, as shown in Fig. 6(a), a notable seismic gap is observed in the eastern segment of the SMF, adjacent to the EAF (F8). Overall, the profiles of the EAF and SMF exhibit distinct segmentation characteristics in both spatial and temporal dimensions, with seismic activity on the EAF demonstrating a migration pattern from northeast to southwest over time. The pronounced spatiotemporal distribution characteristics of earthquakes in the East Anatolian Fault Zone (EAFZ) further underscore the complexity of the geometric structures of the EAF and SMF faults. The interactions and stress transfer between these intricate structures result in localized stress concentrations and uneven distributions. It is posited that the segmented nature of seismic activity on the EAF and SMF may be associated with the coupling strength between faults in this region.

Research analyzing earthquake activity rates derived from the earthquake catalog indicates a decreasing seismic activity rate on the EAF from east to west, which aligns with geodetic measurements that reveal higher strain accumulation in the eastern segment of the EAF compared to the western segment [6, 44]. Earthquake ruptures along a single fault or a system of interconnected faults are typically regarded as continuous [45]. However, the observed trend of the EAF from east to west does not exhibit continuous characteristics in either temporal or spatial dimensions. The precise locations of earthquakes demonstrate significant temporal and spatial variability within the EAF region. In segments F2, F3, and F4 of the EAF, the temporal seismic activity correlates closely with the fault segmentation features, while spatially, there is a lack of significant earthquake activity in adjacent segments, likely due to static or dynamic stress alterations resulting from larger seismic events. For instance, the Mw6.8 earthquake in 2020 and its aftershocks were confined to segment F4. Conversely, notable differences are evident in segments F5, F6, and F7 of the EAF, where the Mw7.8 strong earthquake rupture in 2023 traversed segments F5, F6, and F7, with aftershocks extending both northeast and southwest along the EAF, concluding in segment F5 on the northeastern side. The high-precision earthquake catalog further elucidates the differences in segmentation characteristics of the EAF from both temporal and spatial perspectives, suggesting that these discrepancies may reflect variations in rock properties influenced by the degree of coupling within the EAF fault system and the strength of the fault composition as affected by the regional stress field.

4 b-value estimation

The b-value, derived from the Gutenberg-Richter (G-R) relationship [46], serves as a critical parameter for assessing seismic activity. This parameter plays a significant role in estimating stress levels within fault zones, analyzing the characteristics of seismic activity, and evaluating seismic hazards [47, 48]. On a global scale, the b-value ranges from 0.3 to 2.0 [49], with an average value of approximately 1.0 [50]. Its variability is influenced by several factors, including tectonic features, the distribution of stress fields, and deep velocity structures [51]. Consequently, the b-value serves as an indirect indicator of the seismic structure and geological characteristics of the area under study. Numerous studies have established an inverse relationship between the b-value and stress levels in seismically active regions [52,53,54]. Furthermore, the b-value is closely associated with the minimum completeness magnitude (Mc) of earthquake catalogs, leading to the development of various methodologies for calculating the b-value based on Mc [55,56,57]. Maximizing the number of earthquakes included in b-value calculations is essential for achieving high-quality results; however, this is constrained by the completeness magnitude Mc, which is affected by factors such as the seismic activity of the study area and the monitoring capabilities of the seismic network. Temporal and spatial variations can influence statistical outcomes, and the relative distribution of large and small magnitudes may affect b-value estimation [58]. Therefore, minimizing the dependence of the b-value on the completeness magnitude Mc is vital for accurate b-value estimation.

In this research, we employed the B-Positive method, which is based on the continuous magnitude difference of seismic events as proposed by Van der Elst [57], to estimate the b-value in the EFAZ region. The B-Positive method is recognized as a robust technique for estimating b-values from earthquake catalogs [59]. This method capitalizes on the observation that the magnitude differences of earthquakes conform to a Laplace distribution, with its statistical parameters aligning with the b-value of the traditional G-R relationship. Importantly, it does not rely on the completeness magnitude Mc and mitigates the impact of magnitude incompleteness by focusing on positive magnitude differences. Its primary advantage lies in its low sensitivity to variations in the completeness of the earthquake catalog. In contrast to the traditional maximum likelihood estimation method, which imposes stringent requirements on the completeness of the earthquake catalog, the B-Positive method circumvents this issue by avoiding direct reliance on the completeness threshold (Mc) [60]. Given the influence of seismic monitoring capabilities and mainshock-aftershock sequences in the EFAZ region during the study period, the application of the B-Positive method based on positive magnitude differences is deemed more advantageous for determining the b-value in this region.

In this study, we utilized a high-precision earthquake catalog that had been relocated to perform b-value calculations in both temporal and spatial domains using SeismoStats [61]. For the temporal analysis, we employed a sliding window approach, calculating the b-value in chronological order with each interval containing 500 earthquakes and a sliding window length of 150. For each interval, we conducted 50 random samples, each time selecting 100 earthquake magnitudes for magnitude difference calculations, while excluding samples with magnitude differences less than 0.1 and retaining those with quantities exceeding 25 for b-value and standard deviation calculations. The average value derived from the 50 iterations was taken as the b-value and its associated error for the interval. Regarding spatial distribution, we tested multiple grid configurations, ensuring that the b-value imagery encompassed major areas such as the EAF and SMF, while also guaranteeing that each grid contained a sufficient number of earthquake samples for b-value calculation. Ultimately, we opted to divide the horizontal direction into units of 0.125 degrees, with a radius of 0.25 degrees, selecting earthquake samples exceeding 120 for b-value calculations at each grid point. The b-value calculations for each grid point were consistent with those for each interval in the temporal analysis. Additionally, to evaluate the reliability of the b-value, we conducted an assessment of b-value uncertainty [62].

Fig. 7
figure 7

Comparison of b-value sensitivity to completeness magnitude. The black dotted line represents the b-value from the MLE method, while the red, green, and blue dotted lines represent the b-values from the B-Positive method with magnitude differences of 0.1, 0.2, and 0.5, respectively. The shaded areas of each color represent the 95% confidence interval of b-value uncertainty

Full size image

To validate the stability of the B-Positive method in b-value calculations, we compared the b-values obtained using this method with magnitude difference thresholds of 0.1, 0.2, and 0.3 against the maximum likelihood estimation method [55] (MLE) utilizing the earthquake catalog from the EAFZ region. Figure 7 illustrates the variation of b-values with completeness magnitudes ranging from 0.5 to 4.5. The comparative results indicate that as the completeness magnitude increases, the b-value tends to rise. Notably, within the completeness magnitude range of 0.5 to 2.7, the B-Positive method demonstrates relatively stable b-value variations across different magnitude differences, particularly when the completeness magnitude is between 1.5 and 2.7, where it is minimally influenced by magnitude. Beyond a completeness magnitude of 2.7, the b-value exhibits significant changes, accompanied by a marked increase in uncertainty. In contrast, the MLE method displays greater variability in b-values in response to changes in completeness magnitude, suggesting a higher dependence on the completeness magnitude. Consequently, the utilization of the B-Positive method effectively mitigates the unreliability of b-values resulting from inaccurate estimates of the minimum completeness magnitude Mc.

Figure 8 illustrates the temporal and spatial distribution of b-values surrounding the East African Fault Zone (EAFZ). In this investigation, we computed the b-value distributions over time for the EAFZ, East African Fault (EAF), and Southern Main Fault (SMF) regions. A comparative analysis of b-value variations across these three regions reveals that the average b-values for both the EAFZ and SMF regions are approximately equivalent, at around 0.83, whereas the EAF region exhibits a comparatively lower b-value of 0.78. The b-values, along with their calculated standard deviations for each region, are presented in Table 1. Notably, the b-value calculation results for the EAFZ in this study align closely with those reported by Akar [17], who utilized the Maximum Likelihood Estimation (MLE) method to derive a b-value of 0.86 at a magnitude cutoff (Mc) of 2.7.

Fig. 8
figure 8

Temporal and spatial distribution of b-values. (a) Distribution of b-values over time; the black circles represent earthquake magnitudes, and the black, red, and blue lines represent b-values for the EAFZ, EAF, and SMF regions, respectively. The white pentagons represent two significant earthquake events since the 20th century; (b) Spatial distribution of b-values, with the b-value range color scale shown in the lower right corner. The white pentagons represent three significant earthquake events since the 20th century, and the red lines represent the EAF and SMF; (c) Distribution of b-value uncertainty, with the uncertainty range shown in the lower right corner. The black lines represent fault distributions, and the red lines represent the main fault EAF in the EAFZ

Full size image
Table 1 Statistics of b-values and mean standard deviations in the study area
Full size table

Figure 8(a) depicts the temporal distribution of b-values from 2010 onwards, revealing a gradual decline in b-values over time for the EAFZ, EAF, and SMF regions. The changes in b-values within the EAFZ region generally correspond with those observed in the primary fault EAF region. A comparison of the b-value distributions between the main fault EAF and the northern branch SMF indicates that the b-value in the SMF region is typically higher than that in the EAF region. In the EAFZ region, the distribution of earthquake magnitudes and b-values demonstrates that prior to the Mw 6.8 earthquake in 2020, the b-value in the EAF region experienced a rapid decline from approximately 0.8 to 0.5. Similarly, preceding the Mw 7.8 and Mw 7.6 earthquakes in 2023, the b-values for the EAF and SMF faults decreased from around 1.0 and 0.7 to approximately 0.3, respectively. This trend reflects a reduction in b-values to varying extents prior to the occurrence of significant seismic events, thereby reinforcing the hypothesis that b-values typically diminish before large earthquakes [63]. Since the 20th century, the EAFZ has generally exhibited low seismic activity until the Mw 6.8 earthquake in the northern segment of the EAF in 2020. By correlating this characteristic of low activity with the observed decreasing trend of b-values in the EAFZ region over time, we infer that this alteration in b-value signifies the ongoing accumulation of stress along the EAFZ fault and the establishment of strong coupling.

Figure 8(b) presents the distribution of b-values in the EAFZ region along the horizontal axis, while the uncertainty associated with the b-value distribution is depicted in Fig. 8(c). Overall, the b-value uncertainty surrounding the EAF and SMF regions is determined to be within 2.5%. The b-values within the EAF and SMF fault system regions exhibit considerable lateral inhomogeneity, with lower b-values (0.5–0.6) identified in the middle segment (F5, F6) and northern segment (F3) of the EAF, as well as in the middle (F9) and northern (F8) segments of the SMF. Conversely, higher b-values (0.65–0.80) are observed in areas such as the southwestern segment of the EAF and the southwestern segment of the SMF. Furthermore, the sources of several significant earthquakes since the 20th century display distinct b-value characteristics. The b-value distribution indicates that the source area of the Mw 6.8 earthquake in 2020 exhibited a higher b-value, while the source area of the Mw 7.6 earthquake in 2023 is situated in a transitional zone characterized by both high and low b-values, and the source area of the Mw 7.8 earthquake in 2023 exhibited a low b-value. We hypothesize that the uneven distribution of b-values in the EAFZ region may reflect the collision of the Arabian Plate moving northwestward, as well as the spatial movement characteristics and deep structural differences between the EAF and SMF.

5 Stress field inversion

The inversion of stress tensors based on focal mechanism solutions is a widely utilized approach in the study of regional tectonic stress fields [64, 65]. This inversion process allows for the estimation of the orientations of the three principal stresses (σ1, σ2, σ3) as well as the associated stress magnitude factor R, which ranges from 0 to 1 [64]. The R value is essential for the analysis of tectonic stress fields [65].

In order to investigate the characteristics of the stress field in the EFAZ region, this study compiled focal mechanism solution data from 1,065 seismic events from AFAD and employed the MSATSI method [66] for stress field inversion in the EAFZ region, as illustrated in Fig. 9. The MSATSI method is a sliding direction fitting technique based on damped linear inversion. It systematically divides the study area into uniform grids and allocates focal mechanism solutions to the respective grids, utilizing a least squares approach with a damping factor to concurrently invert the stresses within each grid while minimizing stress discrepancies between adjacent grids.

The uncertainty associated with the inversion is assessed using the bootstrap method [67], which effectively mitigates model uncertainty and overfitting, thereby enhancing the spatial reliability and stability of the stress model. In this inversion process, the grid size was established at 0.25 degrees, with a minimum of five focal mechanism solutions per grid. The bootstrap sampling frequency for evaluating inversion uncertainty was set to 2,000 iterations, with a confidence level of 95%. Figure 9b presents the curve used to select the optimal damping coefficient, ultimately determining the final optimal damping parameter to be 1.2.

Figure 9a illustrates the tectonic stress field derived from the focal mechanism solutions. The distribution of the stress field indicates that the East Anatolian Fault (EAF) exhibits a pronounced segmented characteristic in the orientation of the principal stress σ1, which extends from the southwest to the northeast. The southwestern segment (F7, F6) and the northeastern segment (F1, F2, F3) of the EAF demonstrate an almost north-south alignment, with R values around 0.5, signifying distinct orientations for the three principal stress axes σ1, σ2, and σ3, as well as a well-defined stress environment. Conversely, the middle segment (F4, F5) of the EAF trends approximately north-northwest to south-southeast, with an R value of approximately 0.8, indicating a clear direction for the principal stress σ1, while σ2 and σ3 rotate within the plane perpendicular to σ1. The Southern Marmara Fault (SMF) exhibits a principal stress σ1 direction that trends approximately north-northwest to south-southeast from west to east, with R values in the western and middle segments of the SMF generally exceeding 0.8, indicating a clear orientation for the principal stress σ1, while σ2 and σ3 also rotate in the plane perpendicular to σ1. In the eastern segment of the SMF, the R value approaches 0.5, suggesting that the orientations of the three principal stress axes in this region are consistent with those in the northeastern segment of the EAF, indicating a uniform stress environment.

Fig. 9
figure 9

Distribution of the stress field in the EAFZ region. (a) Black and white short lines represent the maximum and minimum principal compressive stresses, respectively. The direction and length of the short lines indicate the direction and dip angle of the stress. The blue lines represent the EAF and SMF, and the colored circles represent R values, with the R color scale shown in the upper left corner. The red arrows represent GPS deformation [3]; (b) Curve showing the trade-off between data fitting errors and model length in damped linear inversion. The circles represent results calculated with different damping coefficients, and the crosses indicate the best damping coefficient, with the numbers above indicating the optimal damping value

Full size image

When considered alongside surface deformation observations, this analysis reflects that the three principal stress axes in the southwestern and northeastern regions of the East Anatolian Fault Zone (EAFZ) are generally in a stable state. The interaction between the Arabian Plate and the Anatolian Plate, influenced by the collision with the northeastern Eurasian Plate, results in the EAF and SMF exhibiting left-lateral strike-slip characteristics. This interaction leads to north-northwest to south-southeast compression in the middle segment of the EAF and the middle-west segment of the SMF, while a near north-south tension is observed on the southwestern side of the EAF. Additionally, a stronger compressive stress distribution is evident in the northeastern segment of the EAF compared to the southwestern segment.

6 Results and discussion

6.1 Comparison of earthquake location

The residual statistics of the original earthquake catalog (illustrated in Fig. 10a), the absolute location catalog (depicted in Fig. 10b), and the combined absolute and relative location catalog (shown in Fig. 10c) are presented herein. Following the relocation process, the travel time residual for the earthquake events was reduced from an initial value of 0.42 s to 0.28 s. In terms of spatial distribution, the overall characteristics of the earthquake epicenter distribution remain consistent with the initial distribution, predominantly aligning along the East Anatolian Fault (EAF) and the Southern Main Fault (SMF), with certain localized areas exhibiting clustering tendencies, particularly on the southern side of the EAF and the northern side of the SMF. Transitioning from the HypoInverse absolute location to the HypoTD relative location, the horizontal average location error for the earthquake events decreased from 1.5 km to approximately 0.6 km.

Fig. 10
figure 10

Multi-stage earthquake location residual distribution. (a) HypoTD relative location residual; (b) HypoInverse absolute location residual; (c) Initial earthquake catalog residual

Full size image

Significant alterations were observed in the earthquake source depths when compared to the initial earthquake catalog; specifically, 80% of the initial earthquake events were concentrated at depths ranging from 5 to 10 km, with a reduced number of events occurring at other depths. Following the integration of absolute and relative location methodologies, earthquakes at depths of 0 to 5 km constituted about 10%, those at 5 to 10 km accounted for approximately 25%, and events between 10 and 15 km represented around 40%, with the frequency of other earthquake events diminishing progressively with increasing depth. The average depth error was also reduced from 2.87 km to 0.97 km.

The combined location approach resulted in a decrease in both travel time residuals and location errors, leading to a more stable error distribution. The multi-stage location process indicates that the accuracy of earthquake location based on travel time is influenced not only by the distribution of seismic stations, the timing of data picking, and the selection of velocity models, but also significantly by the chosen location strategy.

In this study, we conducted a comparative analysis of the multi-stage location results (illustrated in Fig. 11b) with the relocation catalog generated using the NLL-SSST method as described by Lomax [68] (depicted in Fig. 11a). Lomax provided a comprehensive location of the AFAD catalog covering the period from January 1, 2020, to March 4, 2023; consequently, we selected the multi-stage location catalog for the same timeframe for our comparison. Lomax identified a total of 26,602 earthquake events, yielding an average travel time residual of 0.38 s and an average source depth of 10.86 km. In contrast, our multi-stage location identified 25,906 earthquake events, with an average travel time residual of 0.30 s and an average depth of 13.81 km. The comparative analysis indicated that, notwithstanding the variations in location methodologies and velocity models employed, the overall spatial distribution of the results exhibited a notable degree of similarity. Furthermore, with respect to location residuals, the multi-stage location catalog from this study demonstrated a more convergent distribution pattern, suggesting that the parameter settings utilized during the multi-stage location process were appropriate and that the resulting locations are reliable.

Fig. 11
figure 11

Comparison of the multi-stage location catalog of this study (b) with the Lomax catalog [68] (a). The white lines in the figure represent fault distributions, and (d) and (e) show the earthquake distribution along the ABCD profile. The profile location is indicated by the blue dashed lines in (a) and (e). The depth and RMS statistics of the two earthquake catalogs are shown in (b), (c), (f), and (g)

Full size image

6.2 Türkiye double earthquake sequence and its consequences

On February 6, 2023, eastern Türkiye experienced two significant seismic events, measuring Mw7.8 and Mw7.6, which are among the rare instances of strong earthquakes recorded in the East Anatolian Fault Zone (EAFZ) since the 20th century. These powerful earthquakes resulted in extensive and complex ruptures within the EAF and the Sinjar Fault (SMF) systems. Following the Mw7.8 earthquake, which occurred first, a total of approximately 12,268 aftershocks were recorded by February 28. This total included 2 events with magnitudes between 6 and 6.9, 45 events between 5 and 5.9, 462 events between 4 and 4.9, and 11,759 events below magnitude 4.0.

To enhance the understanding of the co-seismic rupture development associated with these two significant earthquakes, Fig. 12 illustrates the temporal and spatial distribution of the seismic events and their aftershocks, following precise geolocation. The analysis indicates that the epicenter of the Mw7.8 earthquake (36.99˚E, 37.25˚N) is situated on the southeastern side of the junction of the southern branch of the EAF, specifically near faults F6 and F7, at a depth of 4.18 km. Conversely, the epicenter of the Mw7.6 earthquake (37.23˚E, 38.07˚N) is located in the central section of the SMF, with a depth of 9.92 km.

The temporal distribution of the seismic sequences from both earthquakes reveals a distinct migratory pattern. Following the Mw7.8 earthquake, the aftershock sequence rapidly transitioned to the main fault of the EAF, with aftershocks migrating in both the northeast and southwest directions within a span of 9 h. Approximately 9 h later, leading up to the Mw7.6 earthquake, the aftershock sequence shifted towards the SMF. Within roughly 3 h post the Mw7.6 earthquake, seismic activity predominantly migrated along the SMF, extending both eastward and westward from the epicenter of the Mw7.6 event. Notably, 12 h after the Mw7.8 earthquake, aftershocks were primarily concentrated along both the EAF and SMF.

Figure 13 illustrates the distribution of seismic events and b-value distribution surrounding the East Anatolian Fault Zone (EAFZ) both prior to and following the Mw7.8 earthquake that occurred in 2023. Specifically, Figs. 13(a) and 13(b) depict the epicenter distribution before and after the aforementioned earthquake, while Figs. 13(c) and 13(d) present the b-value distribution for the same periods

Fig. 12
figure 12

Temporal and spatial distribution of Türkiye’s double earthquakes. (a) Earthquake distribution within 9 h after the 2023 Mw7.8 earthquake; (b) Earthquake distribution from 9 to 12 h after the 2023 Mw7.8 earthquake; (c) Earthquake distribution 12 h after the 2023 Mw7.8 earthquake. The red lines in the figure represent the main faults EAF and SMF of the EAFZ, and the red and green beach balls represent the source mechanism solutions of the 2023 Mw7.8-Mw7.6 double earthquakes

Full size image

Additionally, the b-value uncertainty was assessed in the calculations [63], revealing that the b-value uncertainty for the EAF, SMF faults, and surrounding areas remained within 5% both before (Fig. 13e) and after (Fig. 13f) the Mw 7.8 earthquake in 2023.

Prior to the 2023 Mw7.8 earthquake, seismic activity in the EAFZ region (Fig. 13a) was predominantly concentrated in the northeastern segment F4 of the EAF, particularly around the epicenter of the 2020 Mw6.8 earthquake, with other regions exhibiting minor clustering characteristics. The b-values in segments F3, F5, and F6 of the EAF were relatively low, approximately 0.55, whereas the junction area of segments F4, F5, and F6, along with the central and western sections of the SMF, displayed significantly higher b-values, around 0.7 (Fig. 13c). Following the 2023 Mw7.8 earthquake, the seismic sequence in the EAF demonstrated a northeast-southwest distribution, extending northeast towards F5 and southwest towards the southern side of F7, while the SMF exhibited a west-east distribution, primarily affecting the northern sides of segments F9, F10, and F8. This overall pattern suggests that the 2023 Mw7.8 earthquake occurred in a region characterized by low seismic activity and low b-values, indicating a high-stress state at the time of the event.

Fig. 13
figure 13

Distribution of seismic activity and b-value changes before and after the Türkiye Mw7.8 earthquake. (a), (c), (e) show the distribution of earthquakes, b-values, and b-value uncertainties in the EAFZ region before the Türkiye Mw7.8 earthquake; (b), (d), (f) show the distribution of earthquakes, b-values, and b-value uncertainties in the EAFZ region after the Türkiye Mw7.8 earthquake. The black lines in the figure represent the main faults EAF and SMF of the EAFZ, and the pentagrams represent the epicenter locations of three strong earthquake events since the 20th century

Full size image

From a tectonic perspective, the collision of the Arabian Plate moving northwestward with the Anatolian Plate, influenced by the Eurasian Plate to the northeast, has resulted in the accumulation of significant stress along the EAF, which serves as the northwest collision front, prior to the 2023 Mw7.8 earthquake. The oblique collision system is capable of accumulating elevated shear stress [69], leading to varying degrees of low b-value locking (coupling) across the distinctly segmented sub-faults of the EAF. This phenomenon establishes a backdrop of super-shear stress within the EAF, culminating in rupture within a brittle environment characterized by low porosity and low fluid saturation [22], ultimately resulting in the super-shear-triggered Mw7.8 earthquake [9].

It is noteworthy that the Mw7.6 earthquake, which occurred on segment F9 in the central section of the SMF, was located less than 100 km from the Mw7.8 earthquake and transpired approximately nine hours later. The source area of the Mw7.6 earthquake exhibited a contrastingly higher b-value characteristic. Analysis of the deep velocity structure indicates that the source area of the Mw7.6 earthquake possesses plastic characteristics, including high porosity and high fluid saturation. By correlating the time series of seismic activity with changes in b-values, we hypothesize that the co-seismic effects of the 2023 Mw7.8 earthquake led to an increase in pore pressure within the high porosity fault environment, due to the rapid migration of deep fluids and other plastic materials, thereby triggering the Mw7.6 earthquake.

Fig. 14
figure 14

Differential distribution of b-values before and after the 2023 Mw7.8 earthquake. The black lines in the figure represent the main faults EAF and SMF of the EAFZ. The pentagrams represent the epicenter locations of three strong earthquake events since the 20th century

Full size image

Significant seismic events, such as strong earthquakes, typically manifest in regions characterized by stress accumulation within fault zones, particularly in areas exhibiting convex shapes or robust locking mechanisms [55, 70]. Following an earthquake, the stress distribution surrounding the fault undergoes alterations. Research indicates that variations in b-values, as delineated by the Gutenberg-Richter (G-R) law, can serve as valuable indicators of fault stress, surface roughness, and disturbances in pore pressure [71]. This parameter is applicable for medium- to long-term assessments of seismic hazards. The double earthquakes of 2023, with magnitudes Mw7.8 and Mw7.6, significantly modified the stress field within the Eastern Anatolian Fault Zone (EAFZ), resulting in a substantial number of aftershocks and the activation of adjacent faults [72]. The distribution of b-values in the vicinity of the EAFZ following this double earthquake event (Fig. 13d) exhibited notable changes. An analysis of b-value variations before and after the Mw7.8 earthquake (Fig. 14) revealed a decrease in b-values within the triangular area delineated by the EAF and the Sinjar Fault (SMF) in the EAFZ region. Specifically, b-values diminished in transitional zones F5-F6, the southwestern section of F4 adjacent to F5, the central region of F7, and the northwestern area of F9 in the SMF, indicating an increase in stress in these locations subsequent to the Mw7.8-Mw7.6 double earthquakes. Although this stress increase did not reach a threshold sufficient to trigger a major earthquake, it is a significant factor contributing to the persistent occurrence of numerous aftershocks in the region following the double earthquake event. This phenomenon reflects the processes of afterslip or fault creep during the stress adjustment period post-Mw7.8-Mw7.6 double earthquakes [73], akin to observations made during the 2022 Mw6.7 Menyuan earthquake [74].

In comparison to the b-value distribution prior to the 2023 Mw7.8 earthquake, it is evident that while the stress levels in the EAFZ region underwent adjustments following the Mw7.8-Mw7.6 double earthquakes, the transitional areas of F5-F6 and the northwestern side of F3 continued to exhibit elevated stress levels that warrant careful monitoring. Although b-values serve as indicators of fault stress levels, it is widely accepted that b-values tend to decrease in the lead-up to significant earthquakes [63], with changes in b-values being influenced by the rupture mode of the fault. This rupture is governed by a multitude of factors, including effective stress, the properties of deep-seated materials, and fluid migration [71]. Consequently, when analyzing the mechanisms underlying the earthquake preparation process and assessing seismic hazards through statistical parameters such as b-values, it is imperative to thoroughly consider the tectonic movement patterns and the structural properties of deep materials within the study area.

7 Conclusions

This study presents a comprehensive analysis of 105,690 earthquake events occurring in the East Anatolian Fault Zone (EAFZ) from January 1, 2010, to July 31, 2024, resulting in the development of a high-precision earthquake catalog through a multi-stage joint location strategy. By integrating the spatiotemporal evolution of b-values, stress field inversion based on 1,056 source mechanism solutions within the study area, and an examination of the activity surrounding the 2023 Mw7.8 and Mw7.6 double earthquakes in Türkiye, we have systematically elucidated the tectonic control mechanisms and characteristics of the earthquake preparation environment in the EAFZ. The principal findings are as follows:

  1. (1)

    Earthquake occurrences in the EAFZ are predominantly aligned in a belt-like configuration along the southern branch of the East Anatolian Fault (EAF) and the northern branch of the Sinjar Fault (SMF), with source depths primarily within 20 km, concentrated between 5 km and 17.5 km, indicating a predominance of shallow earthquakes. The seismic activity along the EAF exhibits a migratory pattern from northeast to southwest, with the spatial distribution of earthquakes aligning closely with the segmented features of the EAF. The SMF has remained in a prolonged state of seismic quiescence, which was disrupted by the stress induced by the 2023 Mw7.8 earthquake.

  2. (2)

    Prior to the Mw6.8 earthquake in 2020, the b-value in the EAF region experienced a rapid decline from approximately 0.8 to 0.5. Similarly, preceding the Mw7.8 and Mw7.6 earthquakes in 2023, the b-values for the EAF and SMF faults decreased from around 1.0 and 0.7 to approximately 0.3, respectively. This trend indicates a notable reduction in b-values prior to significant seismic events, corroborating the hypothesis that b-values typically diminish before major earthquakes, thereby suggesting a short-term precursor indicative of instability in the region’s critical stress state.

  3. (3)

    The principal compressive stress axis (σ1) in the central section of the EAF and the western section of the SMF is oriented NNW-SEE, reflecting the combined influences of the northward subduction of the Arabian Plate and the resistance posed by the Eurasian Plate. This interaction results in left-lateral strike-slip motion along the fault, creating localized areas of compressive stress concentration in the central section, which provides a dynamic backdrop for the potential occurrence of significant earthquakes.

  4. (4)

    The Mw7.8 and Mw7.6 double earthquakes of 2023 transpired in two distinct environments characterized by similar b-value states. The Mw7.8 earthquake occurred in a region exhibiting low seismic activity and a low b-value (~ 0.55), whereas the Mw7.6 earthquake took place in an area of low seismic activity but with a relatively higher b-value (~ 0.70). It is tentatively suggested that there are variances in the earthquake preparation environments for these two seismic events. Furthermore, the super-shear effect associated with the 2023 Mw7.8-Mw7.6 double earthquakes induced stress adjustments within the EAFZ, resulting in a decreased b-value in the triangular area delineated by the EAF and SMF in the central EAFZ, which continues to exhibit an enhanced locking state, indicating a potential risk for future strong earthquakes.