Abstract
We examine the 6 February 2023 Türkiye–Syria earthquakes using an extensive SAR dataset, addressing some limitations of previous studies. Large surface displacements caused significant loss of coherence in the Sentinel-1 Differential SAR Interferometry (DInSAR) results and prior analyses using Pixel Offset Tracking (POT) were limited by the poor azimuthal resolution of the available Sentinel-1 and ALOS-2 SAR images. For the first time, we present high-azimuth-resolution displacement measurements obtained thanks to the SAOCOM-1 sensors. Azimuth information is important considering that the main movements occurred were horizontal and, in some areas, with an important N-S component. By exploiting multi-frequency Sentinel-1, ALOS-2, and SAOCOM-1 SAR data and applying the DInSAR and POT techniques, where appropriate, we derived a detailed displacement field and retrieved an elaborated fault model comprising 22 segments; this model accurately characterizes the geometry and kinematics of the two main faults. Maximum slip reaches ∼15 m for both faults, and the total seismic moment corresponds to Mw 7.9. A finite-fault ShakeMap generated from this source model shows improved agreement with near-field ground motions relative to point-source formulations, while on-fault static stress changes identify low-slip areas that remained unbroken during rupture. The three-dimensional displacement field reveals a broad uplifted region and opposing horizontal motions between the two main ruptures, indicating distributed deformation within an interfault block that accommodates part of the Arabia–Anatolia convergence. This off-fault deformation has not been documented previously for the 2023 sequence and provides new constraints on strain partitioning and future seismic hazard along the East Anatolian Fault Zone.
1 Introduction
The Eastern Mediterranean region was struck by a major seismic sequence on 6 February 2023 in southeast Türkiye (Figure 1). The first and strongest event of magnitude Mw 7.8, may have ruptured within a previously identified seismic gap () and was followed 9 hours later by an Mw 7.5 earthquake. These earthquakes rank among the most severe in the region in recent decades and caused catastrophic impacts, including more than 57,000 fatalities () and widespread infrastructure damage such as building collapses and gas pipeline failures ( and references therein). The sequence occurred along the East Anatolian Fault Zone, a major plate boundary approximately 600 km in length (Figure 1). Given that only a few continental strike-slip earthquakes exceeding Mw 7.5 have been documented globally in recent years (), this sequence drew international scientific attention, resulting in a substantial number of studies (e.g., ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ).
FIGURE 1
The earthquakes produced large-scale surface deformation, prompting extensive analysis of co-seismic data using Differential SAR Interferometry (DInSAR; ; ; ). Initial studies relied mainly on Sentinel-1 (S-1) C-band data (5.6 cm wavelength), which are freely available under the COPERNICUS program open-access policy (). The Interferometric Wide Swath (IWS) capability of TOPSAR provides spatial coverage of up to ∼250 km, advantageous for mapping the widespread deformation of the Türkiye–Syria sequence. However, the large displacements near the faults led to severe decorrelation and very high fringe rates, which limited phase unwrapping and restricted the ability of interferometric analysis to fully resolve near-field permanent displacements (e.g. ).
Such limitations are quite clear in Supplementary Figure S1; these results, along with many other for earthquakes that have accrued globally, are generated through the EPOSAR tool (; ; ), developed to automatically compute co-seismic DInSAR products, which are available via the EPOSAR TCS catalogue (https://www.ics-c.epos-eu.org). Moreover, the application of the Pixel Offset Tracking (POT) technique (e.g., ; ) to the Sentinel-1 image amplitudes, has significant limitations. Indeed, although it can be beneficial for what concerns the retrieval of the large across-track (range) displacements, due to the relatively high ground range resolution (on average 5 m), it is extremely limited concerning the (along-track) azimuth component, due to the rather poor azimuth resolution (about 20 m), characterizing the Sentinel-1 TOPSAR acquisition mode. Assuming that the accuracy of the retrieved displacement, with the POT technique, is approximately 1/10 resolution (; ), it turns out that only the horizontal displacements in range directions can be effectively mapped. Azimuth information is important considering that the main movements occurred were horizontal and, in some areas, with an important N-S component. For this reason, it was crucial to integrate the Sentinel-1 data with other SAR data, characterized by different wavelengths and better spatial resolutions.
We therefore included L-band SAR data from the ALOS-2 satellite acquired by JAXA (Japan Aerospace Exploration Agency) and the SAOCOM-1 constellation operated by CONAE (Comisión Nacional de Actividades Espaciales, Argentina). ALOS-2 data were available in ScanSAR mode (15 m range, 50 m azimuth resolution), while SAOCOM-1 data were acquired in Stripmap mode (10 m range, 5 m azimuth resolution). The finer azimuth resolution of SAOCOM-1 allowed us to generate a highly detailed deformation map of the surface displacements induced by the February 2023 sequence.
The manuscript is organized as follows: We first describe the Earth Observation datasets used, the methodological framework, and the DInSAR and POT results for displacement mapping. We then detail the DInSAR analysis of ALOS-2 data and the POT analysis of SAOCOM-1 data, which together provide the basis for our displacement field. Presented here for the first time, SAOCOM-1 allowed us to obtain an unprecedented, detailed deformation map of the surface displacements, caused by the two earthquakes. Thanks to this dataset we were able in many cases to avoid making a priori assumptions about the parameters of the fault segments and to estimate a detailed slip distribution model. We also generate an updated finite-fault ShakeMap and assess off-fault deformation patterns to explore how the earthquakes relate to the broader Arabia–Anatolia plate boundary system.
2 Satellite earth observation
2.1 Satellite data and DInSAR, POT results
A central aim of this study is to jointly utilize the most reliable information derived from DInSAR and POT applied to the full set of available multi-frequency SAR data acquired by the Sentinel-1, ALOS-2, and SAOCOM-1 sensors. In Table 1, their main characteristics, and those of the derived products are shown. Moreover, the SAR sensors acquisition footprints are presented in Figure 2, where Sentinel-1, ALOS-2, and SAOCOM-1 tracks are highlighted in blue, orange, and red colour, respectively.
TABLE 1
| Sensor | [cm] | [m] | [m] | Acquisition mode | DInSAR and POT pairs | [days] | [m] | Exploited products | ||
|---|---|---|---|---|---|---|---|---|---|---|
| DInSAR phase | POT range | POT azimuth | ||||||||
| S-1 ASC | 5.6 | TOPS | 04022023_28022023 | 24 | 108.28 | NO | YES | NO | ||
| 28012023_09022023 | 12 | −174.17 | NO | YES | NO | |||||
| S-1 Desc | 5.6 | TOPS | 29012023_10022023 | 12 | 105 | NO | YES | NO | ||
| ALOS-2 ASC | 23.6 | ScanSAR | 05092022_20022023 | 168 | 19.55 | YES | NO | NO | ||
| ALOS-2 DESC | 23.6 | ScanSAR | 16092022_17022023 | 154 | 48 | YES | NO | NO | ||
| SAOCOM-1 ASC | 23.6 | StripMap | 02052022_14022023 | 288 | −576.11 | NO | NO | YES | ||
| 21072022_14022023 | 208 | 1468.7 | ||||||||
| 27042022_22042023 | 360 | 0.4983 | ||||||||
| 11022022_02032023 | 384 | −10.078 | ||||||||
| 27022022_02032023 | 368 | −23.66 | ||||||||
| 29052022_30042023 | 336 | −219.83 | ||||||||
| 14062022_25022023 | 256 | −1546 | ||||||||
| 30062022_25022023 | 240 | −677.71 | ||||||||
Main characteristics of the SAR images and data pairs used in the integrated DInSAR/POT analysis (Δrg denotes ground range resolution).
FIGURE 2
For Sentinel-1, we used six C-band Single Look Complex (SLC) images from both ascending (Tracks 14 and 116) and descending (Track 21) orbits of Sentinel-1A. From these data, two ascending and one descending interferograms were generated (Table 1). As shown in Supplementary Figure S1, interferometric coherence is strongly degraded near the ruptured faults due to the metre-scale displacements caused by the high-magnitude earthquakes. In addition, dense fringe rates in several zones hampered conventional phase unwrapping (; ). As a result, Sentinel-1 DInSAR maps provided only limited Line-of-Sight (LoS) displacement information. To overcome this, we applied the POT technique to the corresponding amplitude images, enabling the retrieval of displacement fields in both LoS and azimuth directions (). In particular, to retrieve the POT results we exploited the AMPCOR routine (https://github.com/isce-framework/isce2/tree/main/contrib/PyCuAmpcor), which performs a cross-correlation amplitude analysis between coseismic SAR image pairs, allowing to generate two main products: across-track and along-track displacement maps, respectively. Moreover, we remark that for each deformation measurement we also compute the covariance value along both the azimuth and range directions, highlighting that a small value of this covariance corresponds to a high accuracy of the retrieved displacement. Accordingly, the covariance values represents a quality indicator of the amplitude based retrieved measurements, similarly to the interferometric coherence for what attains the phase difference based computed displacements ().
It is also worth noting that the accuracy of the achieved POT results depends mainly on the resolution of the exploited SLC SAR images. Therefore, in the case of the used Sentinel-1 TOPS data, the ground range resolution is about 5 m, which permits us to achieve an accuracy level of the displacement maps on the order of 0.5 m (i.e., about 1/10 of the resolution, as mentioned above). Conversely, the azimuth resolution of the Sentinel-1 TOPSAR data is about 20 m, corresponding to an accuracy level of the along-track direction POT products on the order of 2 m, which is inadequate to accurately retrieve the deformation field. Accordingly, we focused on the LoS deformation component, retrieved by applying the POT technique to the Sentinel-1 TOPSAR data. Figure 3 shows the POT deformation maps along the corresponding radar LoS directions obtained through the data collected from both ascending (T14 and T116 tracks, Figure 3a) and descending (T21 track, Figure 3b) orbits. We also remark that only the pixels characterized by a covariance value less than a selected threshold (we typically assume 0.04) are considered reliable because they allow to achieve the expected accuracy of 1/10 of the (azimuth/range) spatial resolution (; ).
FIGURE 3
For what concerns the Sentinel-1 POT results, relevant to the range displacements, they are characterized by a rather high level of accuracy (on the order of 0.5 m, as discussed above), although they clearly do not reach the one achievable through the DInSAR technique, that corresponds to a fraction of the radar signal wavelength (). Thus, to obtain additional information on the LoS displacement component, we benefited from the availability of the SAR data acquired by lower frequencies satellite systems operating at L-band, which helped to mitigate the issues related to the coherence loss and high fringe rate that, as previously discussed, strongly affect the Sentinel-1 C-band DInSAR products particularly near the ruptured faults. Specifically, we took advantage of the L-Band ALOS/PALSAR-2 SAR data. In particular, we exploited four SLC ALOS-2 images (as said, with a resolution of about 15 m and 50 m in the ground range and azimuth directions, respectively) acquired through the ScanSAR mode (similar to the Sentinel-1 IWS TOPSAR mode) from ascending and descending orbits, and covering an area of approximately 350 km × 350 km. These data enable the application of phase unwrapping procedures and the retrieval of displacement maps along the LoS direction with good spatial coverage, making them particularly suitable for analysing large-area deformation phenomena, as those characterizing the investigated earthquakes.
The corresponding LoS DInSAR displacement maps were created by exploiting the interferometric pairs listed in Table 1 and are shown in Figure 4 (wrapped interferograms can be found in the Supplementary Material). We retrieved the LoS displacement field of the overall frame with centimetric accuracy, except for a few areas located near the fault traces. Indeed, within these zones, the coherence losses due to the misregistration errors, due to the large magnitude of the earthquakes, were very significant, thus leading to masking the pixels of these zone.
FIGURE 4
We further remark that in order to retrieve the full 3D displacement field related to the investigated events, it is necessary to also estimate the displacement component along the azimuth direction with high accuracy. Unfortunately, similar to the Sentinel-1 TOPS case, also the ALOS-2 ScanSAR data suffer from the poor azimuth resolution (50 m) characterizing their acquisition mode. Accordingly, such a component cannot be retrieved with a sufficient level of accuracy by applying the POT technique. To overcome this issue, we benefited from L-band SAR data obtained through different acquisition modes, as for the StripMap one that can guarantee good spatial resolution (in the order of few meters) in the azimuth direction.
In particular, we made large use of the SAR data acquired by the Argentinean SAOCOM-1 constellation, operated by CONAE, which consists of two L-band SAR twin satellites SAOCOM-1A and B, launched in 2018 and 2020, respectively. This system ensures a DInSAR-oriented acquisition plan in StripMap mode over the Italian Space Agency (ASI) exclusivity region that includes the area affected by the investigated seismic events. For our analysis we used the available SAR data characterized by very good coverage, which is relevant to two tracks (198 and 197) and five different swaths from ascending orbits, using a total of 13 SLC images paired in eight couples (Table 1). In particular, the azimuth resolution is of about 5 m, allowing the retrieval of ground displacements with high spatial density and, even more remarkable, avoiding the extreme interferometric fringes undersampling effects that may arise when the DInSAR technique is used in cases of very large magnitude events. This is evident in Figure 5a where the SAOCOM-1 DInSAR results are shown; indeed, in this case, a clear interferometric fringe pattern is visible in the coherent areas. However, the spatial extent of each SLC image is approximately 40 km in the range direction, so the phase unwrapping technique may not be straightforwardly applied, because the large displacement variation across the fault element causes drastic phase unwrapping errors. Nevertheless, as mentioned before, thanks to the good azimuth resolution (5 m) it is possible to retrieve the azimuth component of the displacement through the POT technique. The merged deformation map (Figure 5b), resulting from the exploitation of the SAR data pairs listed in Table 1, is achieved with an accuracy of approximately 0.5 m, which permits us to better retrieve a co-seismic ground deformation component that is unique with respect to the signals retrieved through the previously analysed Sentinel-1 (TOPS) and ALOS-2 (ScanSAR) data.
FIGURE 5
Starting from the retrieved results obtained by applying the DInSAR and POT techniques to the available Sentinel-1, ALOS-2, and SAOCOM-1 SAR images, it was then possible to compute the complete 3D co-seismic displacement field for the investigated earthquakes, as discussed in the following Section.
2.2 Joint exploitation of DInSAR and POT measurements for the 3D-displacement field retrieval
The measurements obtained through the DInSAR and POT techniques, applied to the phase and amplitude of the exploited SAR image pairs, respectively, represent the projections along the radar LOS and the along-track directions of the ground displacements associated with the 6 February 2023 Türkiye–Syria seismic events. In order to provide a clearer understanding of the occurred ground motions, a 3D deformation field can be retrieved by jointly exploiting these heterogeneous displacement measurements available in a common geographic grid. To obtain this result, different approaches can be considered. In particular, the Strain Model and Variance Component Estimation (SM-VCE) method permits to estimate the 3D deformations by properly exploiting the spatial correlation of adjacent points (; ), while the previously developed solution described in is based on the Least Squares (LS) technique. In the following, we exploit the latter approach that is rather effective and easy to implement.
Before presenting the obtained results, let us first detail the rationale of the 3D deformation field retrieval. In particular, in Figures 6a,b, the simplified acquisition geometries along the East/vertical plane and the East/North plane are reported, respectively. Specifically, the look angle θ is the angle between the nadir and the radar LoS, while the satellite heading angle α represents the angle between the North direction and the projection of the sensor path on the ground. Regarding the heading angle, it is important to remark that the available SAR satellites follow nearly polar orbits, implying a value of α that is of about 345° for Sentinel-1 data, as well as 347° and 353° for ALOS-2 and SAOCOM-1 images, respectively, for what concerns the ascending passes. Moreover, it is about 195° for Sentinel-1 data and 192° for ALOS-2, related to the descending passes.
FIGURE 6
Furthermore, in Figure 6, represents the actual displacement of a specific pixel on the ground both in East-Up and East-North planes; (the rationale is the same for both ascending and descending components) and represents the one-dimensional measurement due to the projection of the deformation vector, say , along the radar LoS (range) (e.g., ; ). Conversely, (the rationale is the same for both ascending and descending components) is a one-dimensional measurement that derives from the projection of the deformation vector, say along the azimuth direction; it is important to highlight that the azimuth vector and, consequently, the deformation component , belong to the East-North plane, so it is not sensitive to vertical deformation component.
In this framework, we can express the deformation components along the range and azimuth directions, for the ascending and descending orbits, as follows ():
In the previous equations, the limitation of the LoS deformation measurements retrieved through the typical space-borne DInSAR technology is clear. Indeed, due to the quasi-polar orbits, the LoS displacement measurements are not very sensitive to the displacement component along the North direction (being very close to zero), while they are highly sensitive to the vertical and horizontal (East-West) deformation components. On the contrary, the azimuth displacement measurements retrieved through the POT technique may regain sensitivity to the North component, thanks to the fact that the polar orbit followed by the satellite SAR system entails the azimuth direction that is close to the North one (as a consequence, is very close to one).
Accordingly, it is evident that only for the pixels that are imaged by at least three non-coplanar measurements (specifically, one ascending range, one descending range, and one azimuth measurement) it is possible to perform a 3D deformation field retrieval. Due to the fact that in our study the azimuth displacement maps are available only for SAOCOM-1 acquisitions (see Table 1), it is clear that the area where the 3D displacement field can be achieved, corresponds to the overall SAOCOM-1 footprints represented, in Figure 2 by the area coloured in red.
We also remark that, for the investigated 3D displacement field retrieval, only a subset of “reliable” pixels is considered. In particular, with reference to the DInSAR products, the coherence threshold for the pixel selection is equal to 0.2 (; ), while the covariance threshold for the selection of the pixels exploited in the POT analysis is 0.04 (; ).
We underline that for most of the investigated pixels of the common geographic grid the system described in Equation 1 is over-determined (more than three independent equations are available) and, consequently, it can be solved through a least square method. By applying this inversion method to the azimuth and range displacement measurements (say, see in Equation 1), we can easily demonstrate the efficacy of this data integration and how much this system is well constrained by evaluating the condition number (cond) of the matrix A (; ). The condition number is essentially a measure of the sensitivity to errors in the data relevant to a linear equations system, where cond(A) ≫ 1 is an indicator for an ill-conditioned linear system (; ). As shown in Equation 2, the condition number of the square matrix A in Equation 1 can be obtained as the ratio between the maximum and the minimum singular values vector, say [w]:
In our case the obtained results have a mean value of 1.57 and a standard deviation value of 0.18 thus confirming that the system is well conditioned and, consequently, that the errors on the obtained solutions are comparable with those of the DInSAR and POT measurements. An image of the condition number map, retrieved on a pixel by pixel basis, is shown in Figure 7.
FIGURE 7
Concerning the expected errors of the retrieved 3D displacements, we remark that the uncertainties of the available DInSAR and POT measurements “propagate” into those of the retrieved 3D displacement components as follows:where, for each investigated pixel, represents the standard deviation vector relevant to the displacements along the North (N), East (E), and Up (U) directions, while represents the diagonal matrix operator and is the covariance matrix given by:
Note that the matrix A in (1) contains, for each exploited SAR dataset, the cosine projections of the displacements vector onto the North, East, and Up directions, and it is structured as reported in Equation 1; moreover, is the corresponding pseudoinverse matrix of A, and represents a diagonal matrix containing the variance of the displacement measurements. Once is computed, the standard deviations of the three direction components can be easily obtained by extracting the square root of the diagonal elements, as shown in Equation 3. Thus, by considering the typical accuracies of the exploited measurements, which is of approximately 1/10 of the resolution for the POT results (corresponding to about 50 cm for both the Sentinel-1 and SAOCOM-1 products, see Table 1) and λ/4 for each single DInSAR displacement map (corresponding to about 6 cm for the ALOS-2 products) and applying the covariance propagation formula in Equation 4, the resulting uncertainties on the geodetic components can be estimated. It important to highlight that the resulting values differ from pixel to pixel, depending on the number of available displacements measurements (equations) and on the accuracy of the involved techniques (DInSAR or POT). The resulting mean standard deviation values are of about 20 cm for the retrieved vertical and east-west displacements, and of about 50 cm for what concerns the north-south deformation component. These findings are not surprising because the retrieved north-south displacements are predominantly constrained by the accuracy of the along-track information provided by the POT technique, while the vertical and east-west components significantly benefit from the exploited DInSAR measurements contributions in addition to those available through the POT. We also remark that the exploited 3D retrieval approach is rather effective and easy to implement although more advanced solutions, based on maximizing the positive impact of the DInSAR measurements higher accuracy, where available, are worth for future analyses.
The results obtained through the above-mentioned procedure are reported in Figure 8. This analysis allows us to observe that the maximum deformation occurred along the East-West direction with a displacement reaching approximately 530 cm toward the West and about 350 cm along the East, respectively. About the North-South direction we identified that the maximum deformation was about 390 cm toward the South and 360 cm along the North. Finally, the Vertical component results indicate a maximum uplift of about 185 cm together with a maximum subsidence value of 200 cm.
FIGURE 8
3 Seismology
3.1 Fault segmentation and inversion strategy
Amongst the main goals of our study was the definition of a detailed fault model. To this end, the first step was to define the fault segmentation. As constraints we used focal mechanisms derived from seismic waveforms, relocated earthquake hypocenters (), mapped fault traces (AFEAD database, ) and discontinuities observed in the SAR-derived displacement fields. For the modelling we used as input the DInSAR data from ALOS-2 and Pixel Offset Tracking (POT) of Sentinel-1, and SAOCOM-1 satellites.
The important contribution of SAOCOM-1 data can be seen in Figure 9, where there is the comparison of all the sensors, shown for different areas of the faults. Comparing the different SAR-based deformation results, SAOCOM-1 has the best overall coverage; based on that, it enabled us to define the trace of the segments (the red line in the map in Figure 9) in very high detail. SAOCOM-1 is the only sensor that provided clear high quality and high resolution of both on-fault and near-fault information, making it a unique dataset to study the displacement of the February 2023 Türkiye-Syria earthquakes.
FIGURE 9
Fault modelling was performed using the analytical elastic dislocation formulation of . The procedure consists of a two-step inversion: a non-linear inversion with uniform slip to determine the primary geometrical and kinematic parameters, minimizing a chi-square misfit function through the Levenberg–Marquardt algorithm (; ; ), followed by a linear inversion to estimate the slip distribution, imposing non-negativity to avoid unphysical back-slip ().
In the case of the Türkiye-Syria faults, characterized by a huge number of segments, we had to adopt a specific strategy, to retrieve by inversion the highest possible number of parameters, without inverting them simultaneously. This approach is aimed at suppressing tradeoffs that make the solution highly non-unique. We firstly defined the segments required to approximate the fault trace, then we focused on the modeling of every single segment separately after an ad hoc local subsampling (some examples are shown in Figure 10). Thus, we built up, step-by-step, the uniform slip model by exploiting in each step the previous step’s optimum solution; for few specific cases we had to a priori constrain some parameters to achieve the optimization convergence (Supplementary Table S1).
FIGURE 10
In the same way and based on these uniform slip solutions we then proceeded to the linear inversions to estimate the slip distributions, after subdividing the fault into patches of 2 × 2 km. More details of the inversion procedure adopted can be found in the Supplementary Material.
Our preferred model includes 22 fault segments. The rupture associated with the first mainshock (Mw 7.8) extends over ∼310 km, while that of the second event (Mw 7.5) is ∼150 km long. For reference, Figure 11 labels the northern segments as “N” and the southern ones as “S”. The results show that many of the causative faults are not vertical (Supplementary Table S1; Figure 12). Along the northern segments, the dip varies systematically: starting at ∼50° in the east, steepening westward, and returning to ∼50° near the western termination. The dominant kinematics are left-lateral strike-slip with additional dip-slip components on some segments. The southern segments (that in some previous studies were assumed to be vertical), exhibit more variable dip values, ranging from ∼50° to vertical.
FIGURE 11
FIGURE 12
The maximum slip, 15.1 m (±1.03), is located on the N segments associated with the second mainshock (Mw 7.5), rather than the first event. This may reflect either afterslip captured within the SAR acquisition window or the sharper moment release of the second rupture (). The latter interpretation suggests that the northern fault system may have been close to failure and was dynamically triggered by the rupture of the southern segments and its aftershocks ().
On the southern faults, slip reaches up to 14.7 m (±0.8) at ∼5 km depth, with several patches extending to the surface. The slip distribution is more heterogeneous than in the north, with five to six principal asperities distributed across multiple segments. Segment S4 shows the highest slip values. In contrast, the northern system is dominated by three to four slip patches, with the largest on segment N2, which also hosts the second mainshock. Slip uncertainties and resolution values can be found in Supplementary Figure S10, 11.
The seismic moment estimated of the northern rupture is 3.23 × 1020 N m (Mw 7.6), while the southern rupture released 5.21 × 1020 N m (Mw 7.8). These results indicate that both systems contributed significantly to the overall energy release, but with contrasting rupture styles and slip heterogeneity. Table 2 summarizes the results for the northern and southern segments of the rupture.
TABLE 2
| Rupture system | Length (km) | Maximum slip (m) | Depth of max slip (km) | Slip distribution | Seismic moment (N·m) | Equivalent Mw |
|---|---|---|---|---|---|---|
| Northern (N) | ∼150 | 15.1 | Near surface to ∼5 | 3–4 main patches, dominant on N2 (Mw 7.5 event) | 3.23 × 10^20 | 7.6 |
| Southern (S) | ∼310 | 14.7 | ∼5 | 5–6 patches, highest on S4, more heterogeneous | 5.21 × 10^20 | 7.8 |
Comparative source parameters for the northern (N) and southern (S) rupture systems of the 6 February 2023 earthquakes.
3.2 Shakemap
ShakeMap computations were performed for the Mw 7.8 event, consistent with its dominant contribution to recorded ground shaking. The finite-fault rupture of the Mw 7.8 event exhibits strong along-strike variability in segment length, dip and rake. To model the spatial decay of ground shaking, we implemented the multi-segment geometry in the OpenQuake engine (). For comparison, a point-source representation with the Ground Motion Model (GMM) is shown in Figure 13a, where several observations exceed the two-standard-deviation threshold. This reflects the limitations of point-source metrics in cases of extended rupture.
FIGURE 13
A more consistent representation is obtained when rupture geometry is incorporated explicitly. Using the GMM with the RJB distance metric improves the match with observations, particularly within the near field (Figure 13b). The BO14 model was therefore adopted for the ShakeMap generation, due to its suitability for large magnitudes and its extended distance range. The earthquake magnitude was fixed at Mw 7.8, and coefficients appropriate for strike-slip rupture were used. Basin depth was treated as unknown, and the corresponding basin adjustment in the GMM was disabled, as recommended by the authors. PGA values were calculated by combining model predictions on a dense grid with recorded values from strong-motion stations. For each grid point, RJB was computed from the rupture surface; given the local Vs30 value, PGA was then evaluated using BO14.
Observed ground motions, originally reported as peaks of the two horizontal components, were converted to the rotationally independent GMRotD50 metric () using the relationships of . Predicted PGA values were calculated on a grid based on the USGS Vs30 dataset (; https://earthquake.usgs.gov/data/vs30/), considering all points within ∼400 km of the rupture, consistent with the upper applicability limit of BO14. To reduce computational cost, the grid was downsampled by a factor of 10, lowering the resolution from 0.0083° to 0.0833° (∼9 km) without loss of fidelity. Observed data (262 stations) and synthetic predictions (11,817 points) were merged, yielding a total of 12,079 points.
Figure 14 shows the distribution of PGA and the associated residuals. The finite-fault configuration captures the main patterns of strong shaking, with only a few stations located ∼20 km from the rupture exceeding model predictions. Some overestimation at larger distances (>50 km) is evident, consistent with known variability in long-period attenuation.
FIGURE 14
Stress redistribution along the fault planes (Supplementary Material), computed following , reveals zones of coseismic stress release that coincide with areas of maximum slip, particularly in the southeastern part of segment N2 during the second event. In contrast, extensive unruptured or low-slip portions of the faults retain elevated shear stress and may represent potential locations for future seismic activity.
4 Discussion
The model derived from combined DInSAR and Pixel Offset Tracking (POT) analysis of Sentinel-1, ALOS-2 and SAOCOM-1 datasets indicates slip of approximately 15 m for both southern and northern fault structures. Compared with other published slip-distribution models, report peak slips of 6 m for the first event and 8 m for the second. Typical peak slip values in the literature are ∼8 m for the Mw 7.8 earthquake and ∼10–12 m for the Mw 7.6 earthquake. , estimate slip exceeding 7 m for one event and 9 m for the other. report maximum slips of 8 m and 12 m for the first and second events, respectively, while present a peak slip of approximately 10 m. found maximum slip values of 8.1 m for the first event and 11 m for the second, with slip uncertainties spanning 0–2 m. Among previous studies, we could say that our results are closer to those of , who propose a maximum slip of about 15 m for the Mw 7.6 event. It is possible that the larger values retrieved here can be explained by the improved on-fault displacement data, provided by SAOCOM-1.
The joint seismic moment of the events is 8.44 × 1020 N m, corresponding to a moment magnitude (Mw) of 7.9. Along-strike variability in slip distribution is pronounced, particularly in southern segments (segments S3, S7–S9), where reduced slip occurs at fault bends. This variability illustrates the influence of fault geometry on rupture propagation, consistent with the notion that geometrical complexities act as barriers or asperities, affecting the distribution and intensity of ground shaking (). The rupture width remains shallower than 20 km, consistent with rupture-width saturation in large interplate strike-slip earthquakes and with the local Moho depth (∼32–38 km; ), supporting empirical scaling relationships (). The maximum slip in the S segments (in S4) occurred close to a segment junction. It’s possible that this high slip at the exact location could be the result of a strong focusing of rupture energy due to the junction.
As mentioned in , no discrete throughgoing faulting appears to have been activated in the intervening region between the N and S segments. This is indicating that rupture transfer likely occurred through dynamic or static stress interactions rather than along a continuous fault plane. We can see that in Figure 15 we have both horizontal and vertical movements. A distinctive feature of the 3D displacement field is the presence of opposing horizontal motions and significant uplift within the region between the southern (Mw 7.8) and northern (Mw 7.5) ruptures (Figures 15–17). North–south displacements show northward motion south of the southern rupture and alternating zones of southward and northward motion within the interfault region. East–west displacements reveal the expected sinistral pattern, with westward motion north of the southern rupture and a narrow zone of eastward motion in the interfault region (Figure 16).
FIGURE 15
FIGURE 16
FIGURE 17
Except the expected movement change defined by the causative strike-slip faults, possibly the local geology and secondary faults also play a role in the expression of the horizontal components in the East-West direction (see the areas within the orange rectangle in Figure 16). We should note here that, in the interfault area, the forefront of the N-S opposing movements and the forefront of the E-W opposing movements (dashed black lines in Figure N–S and Figure E–W), coincides.
Vertical displacement shows a broad uplifted zone between the two ruptures, bounded by the mapped fault traces but not clearly associated with specific surface geological discontinuities (Figures 17–19). Profile AA′ displays a relatively smooth uplift of more than 1.5 m across a distance of ∼60 km, whereas profile BB′ exhibits sharper gradients consistent with fault-controlled deformation. These observations might suggest that both local fault structure and broader regional tectonics influence the uplift pattern.
FIGURE 18
FIGURE 19
This region lies at the boundary between the Arabian and Anatolian plates, where convergence is transferred to the EAFZ. The patterns of uplift and opposing horizontal motions are consistent with deformation within a restraining block situated between the two major ruptures (Figure 20). This block appears to rotate and uplift during the coseismic phase, accommodating shortening in a transpressional setting. Field surveys have documented localized thrust-like features and localized transtension and transpression, on different sites mainly on the locations of the causative structures (e.g., ). Other InSAR studies have presented 3D displacement patterns (e.g., ) here, we focus on and analyse this large off-fault area inbetween the two faults. Our analysis indicates that plate convergence in this part of the EAF, expressed by the two major earthquake ruptures, is accommodated through the rotation and uplift of a compressional micro-block.
FIGURE 20
Regarding the style of the previously analyzed activity in the area but also in a broader framework, the overall collisional mechanism is complex. The activity might be related with pre-existing structures and on how the stress acts in the brittle upper crust and by the geology of the area. Also, elevated pore pressure changes along pre-existing structures might have produced alteration in the frictional behavior contributing to this expression. We note that, it has been reported that shallow but also deep processes have played a role in the postseismic displacement expression of the 2023 February Türkiye-Syria events (). At the same time, the fault zone creation could have also been caused by variations in strength in the lithosphere (). At the study area, the S-wave velocities are relatively slow in the Anatolian plate, the crust has a thickness of 40–45 km and the lithosphere is thin, about 60 km; on the contrary the Arabian plate is having a thinner crust (30 km) and a thicker lithosphere (60–70 km) (; ; ; ). reported that at the area, the Anatolian plate is weaker (there is ductile deformation related to its thinner lithosphere) than the Arabian plate (it has a strong lithospheric mantle). The Arabian plate underthrusts beneath the Anatolian (Figure 21), with the specific rheological differences. This process might be a factor that contributes to the concentration of stress, transmitted to the surface, to the seismicity expression and the faulting style (). Part of the area in-between the S and N segments that hosts this rotational micro-block, lies within a broad deformation zone with unstable tectonics, since it is exactly at the collisional front. It’s possible that together with the shallow processes, there are also lithospheric controls that might had influenced the expression and deformation style of the 2023 Türkiye-Syria seismic events. The available data do not allow us to distinguish whether the uplift originates primarily from shallow fault geometry or deeper lithospheric structure; both interpretations remain plausible within the resolution of the current dataset.
FIGURE 21
Overall, the 2023 Türkiye–Syria earthquake sequence demonstrates how distributed deformation, geometric complexities and variable fault kinematics combine to shape rupture propagation and off-fault deformation. The uplifted and rotating block between the two ruptures represents a significant expression of strain partitioning at the Arabia–Anatolia boundary and may influence future patterns of seismic activity along this part of the East Anatolian Fault Zone.
5 Conclusion
The February 2023 Kahramanmaraş events demonstrated a rather unexpected rupture propagation across multiple fault segments. This study provides high-resolution constraints on fault behaviour, coseismic deformation, and stress redistribution for one of the most significant recent earthquake sequences in the Eastern Mediterranean. The combination of SAR-derived displacement fields, detailed slip modelling, stress analysis, and rupture-aware ground motion simulations, enhances the understanding of the mechanics of large interplate strike-slip earthquakes, identifies fault segments and structural complexities that may govern future seismic hazard and sheds light into the collisional processes between Arabian and Anatolian plate.
We presented and exploited for the first time SAOCOM-1 results for the Türkiye-Syria February 2023 seismic events, overcoming the limitations of the use of Sentinel-1 satellite whose azimuth POT products were of low quality. SAOCOM-1 results provided a better spatial coverage, important input for a detailed definition of the fault trace and on-fault/near-fault displacement measurements. Based on DInSAR and Pixel Offset Tracking dataset of ALOS-2, Sentinel-1 and SAOCOM-1 we were able to create a detailed slip model. The slip distribution shows marked along-strike variability, with reduced slip at major bends and concentrated patches near segment junctions. The rupture width remains less than 20 km, consistent with the regional crustal structure.
The finite-fault configuration improves the representation of near-field ground shaking compared with point-source formulations. The updated ShakeMap, based on the Ground Motion Model, captures the observed spatial pattern of strong shaking and illustrates the importance of incorporating rupture geometry when modelling ground motions from large strike-slip earthquakes. Static stress changes derived from the slip model, presented in the Supplementary Material, highlight low-slip areas that remained unbroken and may influence future seismic activity.
The three-dimensional displacement field reveals a broad uplifted region between the two ruptures, accompanied by opposing horizontal motions. These observations indicate that deformation is distributed within an interfault block situated at the Arabia–Anatolia plate boundary. The uplifted and rotating character of this block suggests a transpressional environment where shortening is accommodated during major ruptures. This feature has not been documented previously for the 2023 sequence and provides new constraints on strain transfer along the East Anatolian Fault Zone.
The kinematics of this compressional block, together with previously documented rheological contrasts between the Arabian and Anatolian lithosphere, may influence how strain is partitioned across this part of the plate boundary (as proposed by ). Improved understanding of this structure is important for assessing long-term deformation patterns and future seismic hazard. Further work, including continuous geodetic monitoring, detailed field surveys and paleoseismological investigations, is needed to evaluate its persistence and role in accommodating collision-related deformation.
Statements
Data availability statement
The original contributions presented in the study are included in the article/Supplementary Material, further inquiries can be directed to the corresponding author.
Author contributions
NS: Conceptualization, Data curation, Formal Analysis, Investigation, Methodology, Resources, Validation, Visualization, Writing – original draft, Writing – review and editing. PS: Data curation, Formal Analysis, Methodology, Resources, Visualization, Writing – original draft, Writing – review and editing. SA: Conceptualization, Data curation, Formal Analysis, Methodology, Software, Validation, Writing – original draft, Writing – review and editing. MB: Formal Analysis, Writing – review and editing. NV: Data curation, Formal Analysis, Methodology, Visualization, Writing – original draft, Writing – review and editing. CT: Data curation, Formal Analysis, Writing – review and editing. AK: Conceptualization, Funding acquisition, Writing – original draft, Writing – review and editing. FC: Formal Analysis, Methodology, Software, Writing – review and editing. CB: Writing – review and editing. CD: Writing – review and editing. MP: Investigation, Writing – review and editing. MF: Formal Analysis, Writing – review and editing. AA: Writing – review and editing. MM: Writing – review and editing. FM: Resources, Writing – review and editing. YB: Writing – review and editing. RL: Conceptualization, Funding acquisition, Writing – review and editing.
Funding
The author(s) declared that financial support was received for this work and/or its publication. This work was supported in part by the Italian Civil Protection Department (DPC) within the framework of the IREA-DPC (2022–2024) agreement; the contents do not necessarily reflect the official opinions or policies of the DPC. Additional support was provided by the European Union–Next Generation EU (PNRR-M4C2) under Project CN-MOST (CN00000023), Project ICSC–CN-HPC (CN00000013), Project MEET IR (IR0000025), and Project GeoSciences IR (IR0000037), as well as by the National Operational Program Infrastructures and Networks 2014–2020 of the Italian Ministry of Infrastructure and Transport under Project GRINT (PIR01_00013) and Project IBiSCo (PIR01_00011). Support from the Geo-INQUIRE project (GA 101058518) is also acknowledged. AK acknowledges support from EPOS-ON project (GA 101131592).
Acknowledgments
Strong motion recordings were retrieved from AFAD (Disaster and Emergency Management Presidency, https://deprem.afad.gov.tr) and KOERI (Kandilli Observatory and Earthquake Research Institute, http://www.koeri.boun.edu.tr). Seismic Data can be also accessed via the EPOS Platform https://www.epos-eu.org/dataportal. Geological map is from 1:500,000 scale Geology Map of Türkiye, prepared and published by the Institute of Mineral Research and Exploration (MTA), Ankara - Turkey, 1961. Downloaded from https://www.dotaltai.org/documents/geology-map-of-turkiye. Active fault mapping used in this study incorporates the Active Faults of Eurasia Database (AFEAD; ), which is openly available at https://doi.org/10.5194/essd-14-4489-2022. The authors thank the Italian Space Agency (ASI) for providing SAOCOM-1 data under the ASI-CONAE SAOCOM-1 License to Use Agreement. This article includes modified Copernicus Sentinel data 2024; the contributions of the European Commission, Copernicus Program, and ESA in supporting systematic Sentinel-1 acquisitions and open data policy are gratefully acknowledged. ALOS-2 data were provided by JAXA under the framework of the third Research Announcement on Earth Observations Collaborative Research Agreement (PI nr. ER3A2N543). The DEM of the study area was obtained from the NASA SRTM archive. Some figures were produced using GMT (https://www.generic-mapping-tools.org/) and Python (Python Software Foundation, https://www.python.org/). The authors thank all institutions whose data and processing infrastructures supported this work. The slip distribution models created in this study are available in the supplementary material of this paper and the finite source database of the Istituto Nazionale di Geofisica e Vulcanologia – INGV site, https://terremoti.ingv.it/en/finitesource.
Conflict of interest
The author(s) declared that this work was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
The authors CB, AK declared that they were an editorial board member of Frontiers at the time of submission. This had no impact on the peer review process and the final decision.
Generative AI statement
The author(s) declared that generative AI was not used in the creation of this manuscript.
Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.
Publisher’s note
All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.
Supplementary material
The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/feart.2025.1717560/full#supplementary-material
References
1
AbgarmiB.DelphJ. R.OzacarA. A.BeckS. L.ZandtG.SandvolE.et al (2017). Structure of the crust and African slab beneath the central Anatolian plateau from receiver functions: New insights on isostatic compensation and slab dynamics. Geosphere13, 1774–1787. 10.1130/GES01509.1
2
AkkarS.SandıkkayaM. A.BommerJ. J. (2014). Empirical ground-motion models for point-and extended-source crustal earthquake scenarios in Europe and the Middle East. Bull. Earthquake Engineering12, 359–387. 10.1007/s10518-013-9461-4
3
AnQ.FengG.HeL.XiongZ.LuH.WangX.et al (2023). Three-dimensional deformation of the 2023 Turkey Mw 7.8 and Mw 7.7 earthquake sequence obtained by fusing optical and SAR images. Remote Sens.15, 2656. 10.3390/rs15102656
4
ArtemievaI. M.ShulginA. (2019). Geodynamics of Anatolia: lithosphere thermal structure and thickness. Tectonics38, 4465–4487. 10.1029/2019TC005594
5
AtzoriS.HunstadI.ChiniM.SalviS.TolomeiC.BignamiC.et al (2009). Finite fault inversion of DInSAR coseismic displacement of the 2009 L'Aquila earthquake (Central Italy). Geophys. Res. Lett.36, L15305. 10.1029/2009GL039293
6
AtzoriS.AntonioliA.TolomeiC.NovellisV.LucaC.MonterrosoF. (2019). InSAR full-resolution analysis of the 2017–2018 M>6 earthquakes in Mexico remote sens. Environ234, 111461. 10.1016/j.rse.2019.111461
7
BaiC.XuW.ZhaoL.SunK.XieL. (2025). 3D coseismic deformation and fault slip model of the 2023 Kahramanmaraş earthquake sequence constrained by GPS, ALOS-2 and Sentinel-1 data. J. Earth Sci- Ence36 (2), 812–822. 10.1007/s12583-024-0146-5
8
BarbotS.LuoH.WangT.HamielY.PiatibratovaO.JavedM. T.et al (2023). Slip distribution of the February 6, 2023 Mw 7.8 and Mw 7.6, Kahramanmaraş, Turkey earthquake sequence in the east anatolian fault zone. Seismica2 (3). 10.26443/seismica.v2i3.502
9
BayraktarA.VenturaC. E.YangT. Y.HökelekliE.TaşY. (2024). Observed damage behavior of Earth dams during the 2023 Kahramanmaraş, Turkey, earthquakes. Geotech. Geol. Eng.42, 7025–7040. 10.1007/s10706-024-02911-4
10
BooreD. M. (2010). Orientation-independent, nongeometric-mean measures of seismic intensity from two horizontal components of motion. Bull. Seismol. Soc. Am.100, 1830–1835. 10.1785/0120090400
11
BooreD. M.KishidaT. (2017). Relations between some horizontal‐component ground‐motion intensity measures used in practice. Bull. Seismol. Soc. Am.107, 334–343. 10.1785/0120160250
12
BooreD. M.StewartJ. P.SeyhanE.AtkinsonG. M. (2014). NGA-West2 equations fo0r predicting PGA, PGV, and 5% damped PSA for shallow crustal earthquakes. Earthq. Spectra30, 1057–1085. 10.1193/070113eqs184m
13
BöseM.CeylanS.AndrewsJ.MassinF.ClintonJ.SaundersJ. K.et al (2024). Rapid finite-fault models for the 2023 Mw 7.8 Kahramanmaraş, Türkiye, earthquake sequence. Seismol. Res. Lett.95, 2761–2778. 10.1785/0220230426
14
BürgmannR.RosenP.FieldingE. J. (2000). Synthetic aperture radar interferometry to measure earth’s surface topography and its deformation. Annu. Rev. Earth Planet. Sci.28, 169–209. 10.1146/annurev.earth.28.1.169
15
CasuF.ManconiA. (2016). Four-dimensional surface evolution of active rifting from spaceborne SAR data. Geosphere12 (3), 1–9. 10.1130/GES01225.1
16
CasuF.ManconiA.PepeA.ManzoM.LanariR. (2010). Advances in the generation of deformation time series from SAR data sequences in areas affected by large dynamics. IEEE Int. Geoscience Remote Sens. Symposium, 2618–2621. 10.1109/IGARSS.2010.5651926
17
CasuF.ManconiA.PepeA.LanariR. (2011). Deformation time-series generation in areas characterized by large displacement dynamics: the SAR amplitude pixel-offset SBAS technique. IEEE Trans. Geoscience Remote Sens.49 (7), 2752–2763. 10.1109/TGRS.2010.2104325
18
CetinK. O.SoylemeB. z.GuzelH.CakirE. (2024). Soil liquefaction sites following the February 6, 2023, Kahramanmaraş-Türkiye earthquake sequence Bulletin of Earthquake Engineering10.1007/s10518-024-01875-3
19
CheloniD.De NovellisV.AlbanoM.AntonioliA.AnzideiM.AtzoriS.et al (2017). Geodetic model of the 2016 Central Italy earthquake sequence inferred from InSAR and GPS data. Geophys. Res. Lett.44 (13), 6778–6787. 10.1002/2017gl073580
20
ChenJ. (2024). Zhou Yu, Coseismic slip distribution of the 2023 earthquake doublet in Turkey and Syria from joint inversion of Sentinel-1 and Sentinel-2 data: an iterative modelling method for mapping large earthquake deformation. Geophys. J. Int.237 (1), 636–648. 10.1093/gi/ggae066
21
ChenJ.ZhouY. (2024). Coseismic slip distribution of the 2023 earthquake doublet in Turkey and Syria from joint inversion of Sentinel-1 and Sentinel-2 data: an iterative modelling method for mapping large earthquake deformation Geophys. J. Int.237, 636–648. 10.1093/gji/ggae066
22
ConvertitoV.TramelliA.GodanoC. (2024). B map evaluation and on-fault stress state for the Antakya 2023 earthquakes. Sci. Rep.1596 (14), 2045–2322. 10.1038/s41598-023-50837-3
23
CurlanderJ. C.McDonoughR. N. (1992). Synthetic aperture radar: systems and signal processing.
24
DelphJ. R.AbgarmiB.WardK. M.BeckS. L.ÖzacarA. A.ZandtG.et al (2017). The effects of subduction termination on the continental lithosphere: linking volcanism, deformation, surface uplift, and slab tearing in central Anatolia. Geosphere13, 1788–1805. 10.1130/GES01478.1
25
DelphJ. R.DarinM. H.WhitneyD. L.CoscaM. A.TeyssierC.KaymakcıN.et al (2024). Deep lithospheric controls on surface deformation and seismicity around the East Anatolian Fault Zone and A3 Triple Junction. GSA Today34, 4–12. 10.1130/GSATG584A.1
26
De LucaC.ZinnoI.ManuntaM.LanariR.CasuF. (2017). Large areas surface deformation analysis through a cloud computing P-SBAS approach for massive processing of DInSAR time series. Remote Sens. Environ.202, 0034–4257. 10.1016/j.rse.2017.05.022
27
DelouisB.van den EndeM.AmpueroJ.-P. (2023). Kinematic rupture model of the 6 February 2023 Mw 7.8 Türkiye earthquake from a large set of near-source strong-motion records combined with GNSS offsets reveals intermittent supershear rupture. Bull. Seismol. Soc. Am.114, 726–740. 10.1785/0120230077
28
DemmelJ. W. (1997). Applied numerical linear algebra. University City, PA: Society for Industrial and Applied Mathematics.
29
DingH.ZhouY.GeZ.TaymazT.GhoshA.XuH.et al (2023). High-resolution seismicity imaging and early aftershock migration of the 2023 Kahramanmaraş (SE Türkiye) MW7.9 and 7.8 earthquake doublet. Earthq. Sci.36 (Issue 6), 417–432. 10.1016/j.eqs.2023.06.002
30
ElachiC. (1988). Spaceborne radar remote sensing: applications and techniques. New York: IEEE Press, 72–77.
31
ErgintavS.VernantP.TanO.KarabulutH.ÖzarpacıS.FloydM.et al (2024). Unexpected far-field deformation of the 2023 Kahramanmaraş earthquakes revealed by space geodesy Science. Sci.386 (6719), 328–335. 10.1126/science.ado4220
32
FornaroG.FranceschettiG.LanariR.RossiD.TesauroM. (1997a). Interferometric SAR phase unwrapping using the finite element method. IEE Proceedings-Radar, Sonar Navigation144, 266–274. 10.1049/ip-rsn:19971259
33
FornaroG.FranceschettiG.LanariR.SansostiE.TesauroM. (1997b). Global and local phase-unwrapping techniques: a comparison. J. Opt. Soc. Am. A14, 2702–2708. 10.1364/josaa.14.002702
34
FotiouK.ArgyriouA.AlatzaS.TheocharidisC.LoupasakisC.ProdromouM.et al (2023). Impact assessment of the catastrophic earthquakes of 6 February 2023 in Turkey and Syria via the exploitation of satellite datasets. Proc. SPIE 12786, Ninth Int. Conf. Remote Sens. Geoinformation Environ. (21), 127861C. 10.1117/12.2682926
35
GoldbergD. E.TaymazT.ReitmanN. G.HatemA. E.Yolsal-ÇevikbilenS.BarnhartW. D.et al (2023). Rapid Characterization of the February 2023 Kahramanmaraş, Türkiye, Earthquake Sequence. Seismic Rec.3 (2), 156–167. 10.1785/0320230009
36
HeL.FengG.XuW.WangY.XiongZ.GaoH.et al (2023). Coseismic kinematics of the 2023 Kahramanmaras, Turkey earthquake sequence from InSAR and optical data. Geophys. Res. Lett.50, e2023GL104693. 10.1029/2023GL104693
37
HeathD. C.WaldD. J.WordenC. B.ThompsonE. M.SmoczykG. M. (2020). A global hybrid VS30 map with a topographic slope–based default and regional map insets. Earthq. Spectra36 (3), 1570–1584. 10.1177/8755293020911137
38
HuJ.LiuJ.LiZ.ZhuJ.WuL.SunQ.et al (2021). Estimating three-dimensional coseismic deformations with the SM-VCE method based on heterogeneous SAR observations: selection of homogeneous points and analysis of observation combinations. Remote Sens. Environ.255, 0034–4257. 10.1016/j.rse.2021.112298
39
HuJ.LiuM.TaymazT.DingL.IrmakT. S. (2025). Characteristics of strong ground motion from the 2023 Mw 7.8 and Mw 7.6 Kahramanmaraş earthquake sequence. Bull. Earthq. Eng.23, 1225–1254. 10.1007/s10518-023-01844-2
40
HuaJ.FischerK. M.WuM.BlomN. A. (2020). New approaches to multifrequency Sp stacking tested in the Anatolian region. J. Geophys. Res. Solid Earth125, 11. 10.1029/2020JB020313
41
HussainE.KalaycıoğluS.MillinerC. W. D.ÇakirZ. (2023). Preconditioning the 2023 Kahramanmaraş (Türkiye) earthquake disaster. Nat. Rev. Earth Environ.4, 287–289. 10.1038/s43017-023-00411-2
42
KarabulutH.GüvercinS.HollingsworthJ.KoncaA. (2023). Long silence on the East Anatolian Fault Zone (Southern Turkey) ends with devastating double earthquakes (6 February 2023) over a seismic gap: implications for the seismic potential in the Eastern Mediterranean region. J. Geol. Soc.180, jgs2023-021. 10.1144/jgs2023-021
43
KobayashiT.MunekaneH.KuwaharaM.FuruH. (2024). Insights on the 2023 Kahramanmara s ¸Earthquake, Turkey, from InSAR: fault locations, rupture styles and induced deformation Geophys. J. Int.236, 1068–1088. 10.1093/gji/ggad464
44
LeonardM. (2010). Earthquake fault scaling: self-consistent relating of rupture length, width, average displacement, and moment release. Bull. Seismol. Soc. Am.100 (5A), 1971–1988.
45
LevenbergK. (1944). A method for the solution of certain problems in least squares. Q. Appl. Math.2, 164–168. 10.1090/qam/10666
46
LiS.WangX.TaoT.ZhuY.QuX.LiZ.et al (2023). Source model of the 2023 Turkey earthquake sequence imaged by Sentinel-1 and GPS measurements: implications for heterogeneous fault behavior along the east Anatolian fault Zone. Remote Sens.15, 2618. 10.3390/rs15102618
47
LiuC.LayT.WangR.TaymazT.XieZ.XiongX.et al (2023). Complex multi-fault rupture and triggering during the 2023 earthquake doublet in southeastern Türkiye. Nat. Commun.14, 5564. 10.1038/s41467-023-41404-5
48
LiuJ.HuangC.ZhangG.ShanX.KorzhenkovA.TaymazT. (2024). “Immature characteristics of the East Anatolian Fault Zone from SAR,” in GNSS and strong motion data of the 2023 Türkiye–Syria earthquake doublet. 10.1038/s41598-024-61326-6
49
LiuX.XuG.ChenM.ZhangT. (2025). Factors influencing the effectiveness of SM-VCE method in solving 3D surface deformation. Geodesy Geodyn.16 (1), 55–66. 10.1016/j.geog.2024.06.001
50
LomaxA. (2023). Precise, NLL-SSST-coherence hypocenter catalog for the 2023 Mw 7.8 and Mw 7.6 SE Turkey earthquake sequence.
51
LuckmanA.QuinceyD.BevanS. (2007). The potential of satellite radar interferometry and feature tracking for monitoring flow rates of Himalayan glaciers. Remote Sens. Environ.111 (2007), 172–181. 10.1016/j.rse.2007.05.019
52
MaZ.LiC.JiangY.ChenY.YinX.AokiY.et al (2024). Space geodetic insights to the dramatic stress rotation induced by the February 2023 Turkey-Syria earthquake doublet. Geophys. Res. Lett.51, e2023GL107788. 10.1029/2023GL107788
53
MagenY.BaerG.ZivA.InbalA.NofR. N.HamielY.et al (2024). Fault coalescence, slip distribution, and stress drop of the February 2023 Southeast Türkiye earthquakes from joint inversion of SAR, GNSS, and burst overlap interferometry. Seismol. Res. Lett.XX, 1–17. 10.1785/0220230271
54
MaiM.AspiotisT.AquibT.Valero CanoE.Castro‐CruzD.Espindola‐CarmonaA.et al (2023). The Destructive Earthquake Doublet of 6 February 2023 in South‐Central Türkiye and Northwestern Syria: Initial Observations and Analyses. Seismic Rec.3 (2), 105–115. 10.1785/0320230007
55
MarquardtD. (1963). An algorithm for least-squares estimation of nonlinear parameters. SIAM J. Appl. Math.11 (2), 431–441. 10.1137/0111030
56
MassonnetD.FeiglK. L. (1998). Radar interferometry and its application to changes in the Earth's surface. Rev. Geophys.516 (36), 441–500. 10.1029/97rg03139
57
MengJ.KuskyT.MooneyW. D.BozkurtE.BodurM. N.Lu WangL. (2024). Surface deformations of the 6 February 2023 earthquake sequence, eastern Türkiye. SCIENCE383 (6680), 298–305. 10.1126/science.adj3770
58
MonterrosoF.BonanoM.De LucaC.De NovellisV.LanariR.ManuntaM.et al (2019). Unsupervised and automatic generation of DInSAR Co-seismic displacement maps by means of Sentinel-1 data, IGARSS 2019 - 2019. IEEE Int. Geoscience Remote Sens. Symposium2019, 9658–9661. 10.1109/IGARSS.2019.8898772
59
MonterrosoF.BonanoM.LucaC. D.LanariR.ManuntaM.ManzoM.et al (2020). A global archive of coseismic DInSAR products obtained through unsupervised Sentinel-1 data processing. Remote Sens.12 (19), 3189. 10.3390/rs12193189
60
MonterrosoF. (2022). “Automatic generation of Sentinel-1 DInSAR co-seismic maps at global scale,” in 2022 IEEE 21st mediterranean electrotechnical conference (MELECON), Palermo, Italy, 2022, 460–463. 10.1109/MELECON53508.2022.9843041
61
NoflD.DarwisheH.ChaabanF.MohammadA. (2024). Mapping surface displacements after the 6 February 2023 earthquake in Syria and Turkey using DInSAR and GIS techniques. Spat. Inf. Res.32, 231–251. 10.1007/s41324-023-00552-y
62
OgdenC. S.BastowI. D. (2022). The crustal structure of the Anatolian Plate from receiver functions and implications for the uplift of the central and eastern Anatolian plateaus. Geophys. J. Int.229, 1041–1062. 10.1093/gji/ggab513
63
OkadaY. (1985). Surface deformation due to shear and tensile faults in a half-space. Bull. Seism. Soc. Am.75 (4), 1135–1154. 10.1785/bssa0750041135
64
OkuwakiR.YagiY.TaymazT.HicksS. P. (2023). Multi-scale rupture growth with alternating directions in a complex fault network during the 2023 south-eastern Türkiye and Syria earthquake doublet. Geophys. Lett.50, e2023GL103480. 10.1029/2023GL103480
65
ÖverS.DemirciA.ÖzdenS. (2023). Tectonic implications of the February 2023 earthquakes (Mw7.7, 7.6 and 6.3) in south-eastern Türkiye. Tectonophysics866, 230058. 10.1016/j.tecto.2023.230058
66
ÖzkanA.SolakH. I.Tiryakioğluİ.MuratM. D.AktuğB.GezginC.et al (2023). Characterization of the co-seismic pattern and slip distribution of the February 06, 2023, Kahramanmaras (Turkey) earthquakes (Mw 7.7 and Mw 7.6) with a dense GNSS network. Tectonophysics866 (2023), 30041. 10.1016/j.tecto.2023.230041
67
ProvostF.KarabacakV.MaletJ. P.Van der WoerdJ.MeghraouiM.MassonF.et al (2024). High-resolution co-seismic fault offsets of the 2023 Türkiye earthquake ruptures using satellite imagery. Sci. Rep.14, 6834. 10.1038/s41598-024-55009-5
68
PucciS.CaciagliM.AzzaroR.Di MannaP.BlumettiA. M.PoggiV.et al (2025). Examples of rupture patterns of the 2023, Mw 7.8 kahramanmaraş surface-faulting earthquake Türkiye. Geosciences15, 252. 10.3390/geosciences15070252
69
ReitmanN. G.BriggsR. W.BarnhartW. D.Thompson JobeJ. A.DuRossC. B.HatemA. E.et al (2023). Fault rupture mapping of the 6 February 2023 Kahramanmaras, Türkiye, earthquake sequence from satellite data. U.S. Geol. Surv. Data Release. 10.5066/P985I7U2
70
RenC.WangZ.TaymazT.HuN.LuoH.ZhaoZ.et al (2024). Supershear triggering and cascading fault ruptures of the 2023 Kahramanmaraş, Türkiye, earthquake doublet. Science383, 305–311. 10.1126/science.adi1519
71
RippergerJ.MaiP. M. (2004). Fast computation of static stress changes on 2D faults from final slip distributions. Geophys. Res. Lett.31, L18610. 10.1029/2004GL020594
72
RosenP. A.HensleyS.JoughinI.LiF.MadsenS.RodriguezE.et al (2000). Synthetic aperture radar interferometry. Proc. IEEE88 (3), 333–382, March. 10.1109/5.838084
73
SilvaV.CrowleyH.PaganiM.MonelliD.PinhoR. (2014). Development of the OpenQuake engine, the global earthquake Model’s open-source software for seismic risk assessment. Nat. Hazards72, 1409–1427. 10.1007/s11069-013-0618-x
74
StrozziT.LuckmanA.MurrayT.WegmullerU.WernerC. (2002). Glacier motion estimation using SAR offset-tracking procedures. IEEE Trans. Geoscience Remote Sens.40 (11), 2384–2391. 10.1109/TGRS.2002.805079
75
SultanK.SinemÇ.NazliÖ. T.KarakasG. K.VeyselE. K.CandanG. (2025). Landslides triggered by the 6 February 2023 Kahramanmaraş earthquakes (Türkiye). Turkish J. Earth Sci.34 (1). 10.55730/1300-0985.1945
76
TarantolaA. (2005). Inverse problem theory and methods for model parameter estimation. University City, PA: Society for Industrial and Applied Mathematics.
77
TokerM.YavuzE.UtkucuM.UzuncaF. (2023). Multiple segmentation and seismogenic evolution of the 6th February 2023 (Mw 7.8 and 7.7) consecutive earthquake ruptures and aftershock deformation in the maras triple junction region of SE-Anatolia, Turkey. Phys. Earth Planet. Interiors345, 0031–9201. 10.1016/j.pepi.2023.107114
78
TorresR.SnoeijP.GeudtnerD.BibbyD.DavidsonM.AttemaE.et al (2012). GMES Sentinel-mission. Remote Sens. Environ.120 (2012), 9–24. 10.1016/j.rse.2011.05.028
79
TrikhunkovY. I.ҪelikH.LomovV. S.TrifonovV. G.BachmanovD. M.KarginogluY.et al (2024). Geological position, structural manifestations of the elbistan earthquake and tectonic comparison of two strongest seismic events 06.02.2023. East. Türkiye Geotekt.3, 108–126. 10.31857/S0016853X24030054
80
TungS.SipplC.ShirzaeiM.TaymazT.MasterlarkT.MedvedI. (2024). Structural controls on fault slip models of the 6 February 2023 Kahramanmaraş, Turkiye earthquake doublet with finite element analyses. Geophys. Res. Lett.51, e2023GL107472. 10.1029/2023GL107472
81
UnalE. O.KocamanS.CandanG. (2024). Impact assessment of geohazards triggered by 6 February 2023 Kahramanmaras earthquakes (Mw 7.7 and Mw 7.6) on the natural gas pipelines. Eng. Geol.334, 107508. 10.1016/j.enggeo.2024.107508
82
WangZ.ZhangW.TaymazT.HeZ.XuT.ZhangZ. (2023). Dynamicrupture process of the 2023 Mw 7.8Kahramanmaraş earthquake (SE Türkiye):Variable rupture speed and implicationsfor seismic hazard. Geophys. Res.50, e2023GL104787. 10.1029/2023GL104787
83
WuF.XieJ.AnZ.LyuC.TaymazT.IrmakT. S.et al (2023). Pulse-like ground motion observed during the 6 February 2023 MW7.8 Pazarcık Earthquake (Kahramanmaraş, SE Türkiye). Earthq. Sci.36 (4), 328–339. 10.1016/j.eqs.2023.05.005
84
XuL.MohannaS.MengL.JiC.AmpueroJ. P.YunjunZ.et al (2023a). The overall-subshear and multi-segment rupture of the 2023 Mw7.8 Kahramanmaraş, Turkey earthquake in millennia supercycle. Commun. Earth Environ.4, 379. 10.1038/s43247-023-01030-x
85
XuL.AokiY.WangJ.CuiY.ChenQ.YangY.et al (2023b). The 2023 Mw 7.8 and 7.6 earthquake doublet in southeast Türkiye: coseismic and early postseismic deformation, faulting model, and potential seismic hazard. Seismol. Res. Lett.XX, 1–12. 10.1785/0220230146
86
ZeleninE.BachmanovD.GaripovaS.TrifonovV.KozhurinA. (2022). The active faults of eurasia database (AFEAD): the ontology and design behind the continental-scale dataset. Earth Syst. Sci. Data14, 4489–4503. 10.5194/essd-14-4489-2022
87
ZhangY.TangX.LiuD.TaymazT.EkenT.GuoR.et al (2023). Geometric controls on cascading rupture of the 2023 Kahramanmaraş earthquake doublet. Nat. Geosci.16, 1054–1060. 10.1038/s41561-023-01283-3
88
ZhaoJ.-J.ChenQ.YangY.-H.XuQ. (2023). Coseismic faulting model and post-seismic surface motion of the 2023 Turkey–Syria earthquake doublet revealed by InSAR and GPS measurements. Remote Sens.15, 3327. 10.3390/rs15133327
89
ZhouJ.XuY.ZhangY.FengW.TaymazT.ChenY. T.et al (2025). Geometric barriers impacted rupture processes and stress releases of the 2023 Kahramanmaraş, Türkiye, earthquake doublet. Commun. Earth Environ.6, 56. 10.1038/s43247-025-02004-x
Summary
Keywords
2023 Türkiye–Syria earthquake sequence, coseismic displacement, east anatolian fault zone, fault slip modelling, InSAR deformation analysis, pixel offset tracking
Citation
Svigkas N, Striano P, Atzori S, Bonano M, Vavlas N, Tolomei C, Kiratzi A, Casu F, Bignami C, De Luca C, Polcari M, Franzese M, Antonioli A, Manunta M, Monterroso F, Belen Roa YL and Lanari R (2026) Insights into the Arabia–Anatolia plate collision from integrated SAR analysis and detailed modelling of the 2023 Türkiye–Syria earthquakes. Front. Earth Sci. 13:1717560. doi: 10.3389/feart.2025.1717560
Received
02 October 2025
Revised
05 December 2025
Accepted
11 December 2025
Published
28 January 2026
Volume
13 - 2025
Edited by
Paolo Capuano, University of Salerno, Italy
Reviewed by
Bhaskar Kundu, National Institute of Technology Rourkela, India
Jamal Doski, Duhok Polytechnic University, Iraq
Updates
Copyright
© 2026 Svigkas, Striano, Atzori, Bonano, Vavlas, Tolomei, Kiratzi, Casu, Bignami, De Luca, Polcari, Franzese, Antonioli, Manunta, Monterroso, Belen Roa and Lanari.
This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
*Correspondence: Nikos Svigkas, nikos.svigkas@ingv.it
Disclaimer
All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article or claim that may be made by its manufacturer is not guaranteed or endorsed by the publisher.