The devastating February 6, 2023, earthquake doublet in Turkey, with moment magnitudes (Mw) of 7.8 and 7.5, caused widespread destruction and significant loss of life. These earthquakes ruptured along the East Anatolian Fault Zone (EAFZ), extending fault lengths of approximately 368 km and 133 km respectively, making them among the longest recorded continental strike-slip ruptures. The EAFZ is a major left-lateral strike-slip fault responsible for numerous large earthquakes throughout history. The 2023 events initiated near the junction of the Pazarcık and Amanos segments of the EAF, with the Mw 7.8 mainshock originating on a splay fault, the Narlı Fault (NF). The Mw 7.5 aftershock occurred on the Sürgü Fault. Understanding the rupture kinematics and dynamics of these events is crucial for assessing seismic hazards in the region and for informing earthquake early warning systems. This study utilizes multiple datasets and advanced analytical techniques to investigate the rupture processes of the Kahramanmaraş earthquake sequence, with a specific focus on rupture speeds and the potential implications for seismic hazard assessment.

Previous studies on the East Anatolian Fault (EAF) have documented its history of large earthquakes, including events in 995, 1114, 1789, 1893, and 1905. Fault mapping reveals a complex geometry with segments and splay faults. Recent research has focused on the seismotectonics of the EAF, examining fault segmentation, slip rates, and the potential for future large earthquakes. The 2020 Elazig earthquake, while significant, was not considered to significantly increase the Coulomb stress on the Pazarcık segment, which ruptured in the 2023 mainshock. Studies using different methodologies have produced varying results regarding rupture speeds of past earthquakes, highlighting the complexities inherent in such analyses. Debates exist in literature regarding the extent and significance of supershear ruptures during large events.

This study integrates multiple datasets including seismic waveforms from dense arrays in China and Alaska, strong motion recordings from stations near the fault, high-rate Global Navigation Satellite System (GNSS) measurements, and Synthetic Aperture Radar (SAR) imagery from Sentinel-1, ALOS-2, and LuTan-1 satellites. Two primary methodologies were employed: Slowness Enhanced Back-Projection (SEBP) and joint Finite Fault Inversion (FFI). SEBP was applied to both the Mw 7.8 mainshock and Mw 7.5 aftershock to image the spatiotemporal evolution of the rupture. FFI was used to model the slip distribution on the NF and EAF, utilizing seismic waveforms from both close-fault and teleseismic stations, as well as GNSS and SAR data. The SAR data processing involved speckle tracking for Sentinel-1 and InSAR/MAI for ALOS-2 and LuTan-1 to measure coseismic surface deformation. The combined geodetic data were used to further constrain the FFI. The SEBP analysis used slowness correction terms derived from nearby aftershocks to mitigate the effects of 3D wave propagation. To assess the rupture speed, a Mach wave analysis of far-field Rayleigh waves was performed. Synthetic tests were conducted to evaluate the effects of bilateral rupture and rupture kinks on the observed cross-correlation coefficients and amplitude ratios. Finally, the historical seismicity data were analyzed to understand the potential for earthquake supercycles on the EAF, and this analysis included computations of moment accumulation rates, utilizing both observed historical seismicity and the Gutenberg-Richter law to estimate the contribution from smaller unobserved events.

The SEBP and joint FFI analyses revealed a complex rupture process for the Mw 7.8 mainshock. The rupture initiated on the NF and propagated bilaterally onto the EAF. The overall rupture speeds were subshear, with an average speed of approximately 3.05 km/s to the northeast and 3.11 km/s to the southwest. Mach wave analysis of far-field Rayleigh waves confirmed the subshear nature of the rupture, showing no evidence of persistent supershear propagation. Three large slip asperities were identified on the EAF, located on the Pazarcık, Erkenek, and Amanos segments. The largest slip of 9.2 meters occurred on the Pazarcık segment. The analysis of historical seismicity suggests a supercycle behavior on the EAF, with a potential return period of at least 900 years, based on the last major earthquake (M≥7.8) occurring in 1114. Moment accumulation calculations, considering both historical events and background seismicity estimated via the Gutenberg-Richter law, support this long supercycle timescale, although uncertainty remains in the estimates due to the limitations of historical earthquake data and the complexities of the long-term fault slip rate estimates. The rupture initiation on the splay NF shares similarities with other major earthquakes that initiated on secondary faults before propagating onto the main fault. Geometric similarities were found between the EAF system and the San Andreas Fault (SAF) system, particularly the San Jacinto Fault (SJF), with similar main fault-splay fault geometries. The notably lower slip rates on the EAF compared to those observed on the SAF and SJF suggests that the potential for a great earthquake on the SAF is substantial.

The findings highlight the complexities of rupture propagation during large earthquakes, particularly those involving multiple fault segments and splay faults. The subshear rupture speeds, despite the substantial magnitude of the events, challenge some existing models. The observed supercycle behavior on the EAF provides critical insights into the long-term seismic hazard potential of this region. The similarities between the EAF and SAF systems raise important implications for the seismic hazard assessment of the San Andreas Fault. The lack of large earthquakes (M>7) on the southern SAF segment since 1857, coupled with high slip rates, raises concern about a potential future M8 event, drawing important parallels to the rupture mechanisms of the 2023 Kahramanmaraş events. The study underscores the need for continued monitoring and research to refine seismic hazard models and improve earthquake preparedness strategies. Further analysis of the stress interaction between the NF and EAF, as well as the specific dynamic triggering mechanisms involved in initiating and propagating the rupture across multiple fault segments are areas deserving future research.

This study provides a comprehensive analysis of the 2023 Mw 7.8 Kahramanmaraş earthquake sequence, revealing a complex, overall subshear rupture propagating across multiple fault segments. The results support the hypothesis of a millennial-scale earthquake supercycle on the EAF. The remarkable similarities between the EAF and SAF systems, particularly the SAF and SJF, highlight the potential for a major earthquake on the SAF. Future research should focus on refining models of earthquake supercycles, investigating the dynamic interactions between main faults and splay faults, and improving seismic hazard assessments for strike-slip fault systems worldwide.

The study acknowledges uncertainties inherent in interpreting historical earthquake data, particularly regarding magnitudes and precise locations of events predating the instrumental period. The assumptions made in moment accumulation calculations, such as the completeness of historical catalogs and the use of Gutenberg-Richter relationships to estimate background seismicity, introduce potential biases. The 1-D velocity model used in some analyses might not fully capture the complexities of 3D wave propagation. The interpretation of rupture speeds relies on various assumptions and approximations and the potential for localized and transient supershear ruptures cannot be entirely excluded.