The overall-subshear and multi-segment rupture of the 2023 Mw7.8 Kahramanmaraş, Turkey earthquake in millennia supercycle

Two large earthquakes (Mw 7.8 and Mw 7.5) struck south-central Turkey and northern Syria on 6 February 2023, producing exceptionally long continental strike-slip ruptures (~368 km and ~133 km). The events occurred within the East Anatolian Fault Zone (EAFZ) and near the northern end of the Dead Sea Fault Zone. The study seeks to resolve the source processes, rupture kinematics, and speeds (subshear vs. supershear) of the Mw 7.8 mainshock and the Mw 7.5 aftershock, to understand fault interactions and long-term behavior (supercycles) on the EAF. Using seismic, geodetic, and SAR observations, the authors aim to determine where the rupture initiated, how it propagated across multiple segments, whether supershear occurred, and what the implications are for seismic hazard on analogous systems such as the San Andreas–San Jacinto faults.

The EAFZ has hosted numerous historical earthquakes (e.g., 995 M7.4; 1114 M≥7.8; 1789 M7.4; 1893 M7.1; 1905 M6.8) and more recent damaging events (e.g., 1998 M6.3 Adana-Ceyhan; 2003 M6.4 Bingöl; 2011 M7.1 Van; 2020 M6.7 Elazig). Prior to 2023, the 2020 Elazig earthquake ruptured the Pütürge segment but was not estimated to increase Coulomb stress on the Pazarcık segment. Contemporary studies reported differing rupture speeds for the 2023 mainshock: some found overall subshear speeds (~2–3.4 km/s), others suggested transient or significant supershear (up to 5–6 km/s) on parts of the rupture. Supershear interpretations based solely on near-field component amplitude ratios remain debated. The concept of earthquake supercycles—clusters of events with quiescence intervals and occasional multi-segment ruptures larger than characteristic events—has been documented on the San Andreas, Sumatra, Cascadia, and Tohoku systems, informing expectations for the EAF.

The study integrates two main techniques: Slowness-Enhanced Back-Projection (SEBP) and joint Finite Fault Inversion (FFI). SEBP: Broadband teleseismic P-wave data from two large-aperture arrays (China: >700 stations in 1–4 Hz; Alaska: ~200 stations in 0.5–2 Hz) are used. Aftershock-based slowness corrections mitigate 3D path effects and sharpen high-frequency radiator imaging. Sliding windows (10 s, 1 s step) track rupture fronts. Local strong-motion stations (33) provide validation by comparing theoretical S-wave arrivals from selected radiators with observed high-energy pulses. FFI: Six vertical fault planes (S1–S6) constructed from SAR-derived surface traces and aftershock distributions are divided into subfaults (9 km × 3 km). A joint inversion simultaneously uses 20 teleseismic P, 13 S, 23 Rayleigh, 17 Love waveforms (filtered), 57 near-field strong-motion components, 24 high-rate GNSS displacement waveforms, and static ground displacements from Sentinel-1 speckle tracking, ALOS-2 InSAR and MAI, and LuTan-1 InSAR. Wavelet-domain inversion with Laplacian smoothing constrains slip amplitude, rupture time, and rise time; rupture velocity bounds 1.5–4.9 km/s with reference 3.0 km/s; rake ±30° around 11°. Green’s functions use an f-k integration approach with a 1-D local crustal model. Ten inversions with different random seeds assess model uncertainty, selecting the minimum-misfit model. SAR processing: Sentinel-1 speckle tracking (ISCE-2) with oversampling, large correlation windows, filtering and masking; ALOS-2 InSAR and MAI (ionosphere correction via split-spectrum; Goldstein filtering; unwrapping; geocoding); LuTan-1 InSAR (Gamma + ISCE-2); corrections for troposphere (ERA5/PyAPS) and solid Earth tides (PySolid). Three-component displacement fields are estimated with weighted uncertainty propagation. Mach wave analysis: Far-field Rayleigh waves (15–25 s) at 46 stations are cross-correlated with a nearby M5.3 aftershock (EGF). Azimuthal distributions of cross-correlation and amplitude-to-moment ratios are compared to theoretical predictions for supershear vs. subshear (near-Rayleigh) rupture. Synthetic tests model bilateral rupture with varying speeds and a kink to evaluate Mach cone signatures. Moment accumulation analysis: Compute build-up rates using rigidity, segment lengths, seismogenic depth (20 km), and long-term slip rates; releases include historical M≥7 events and background Mw 1–7 seismicity via Gutenberg–Richter with two b-value scenarios to bracket accumulation since 1114.

- Rupture initiation and path: The Mw 7.8 mainshock nucleated on the Narlı Fault (splay), then jumped to the EAF at ~15–20 s and propagated bilaterally across Pazarcık, Amanos, and Erkenek segments. About 92% of moment was released on the EAF segments. - Rupture speeds: Overall subshear propagation with average speeds of ~3.05 km/s (NE, ~120 km extent) and ~3.11 km/s (SW, ~200 km extent). These are close to the Rayleigh wave speed (0.92Vs with Vs ≈ 3.39 km/s, implying ~3.12 km/s). - Far-field validation: Rayleigh-wave Mach-wave analysis shows CC and amplitude ratio peaks aligned with rupture directions, consistent with rupture speeds near Rayleigh and inconsistent with expected supershear Mach-cone patterns. Synthetic tests reproduce observed azimuthal patterns for near-Rayleigh bilateral rupture. - Source parameters: Preferred FFI model moment M0 = 7.67×10^20 Nm (Mw 7.85); average slip ~3.1 m; peak slip ~9.2 m on Pazarcık; largest average slip at ~4 km depth (~4.4 m); rupture duration ~100 s; total rupture length ~370 km, matching field/SAR estimates (~368 km). - Slip heterogeneity: Three large slip asperities north and south of the hypocenter, including near the NF–EAF junction, ~70 km NE, and on south Pazarcık/north Amanos; relatively uniform shallow slip on Amanos (~3.1 m average). - Secondary short-period burst: A secondary BP power peak ~120–140 s corresponds to radiators offshore near the Karasu fault; timing relative to last on-fault radiator and S-wave travel time suggests an immediately triggered aftershock rather than continued mainshock rupture. - Mw 7.5 Sürgü Fault event: Bilateral rupture ~130 km; speeds ~2 km/s eastward and ~3.6 km/s westward; parts of westward branch may have been supershear locally (consistent with other studies), though main Mw 7.8 event remained overall subshear. - Supercycle evidence: Very large slip on Pazarcık and Erkenek occurred despite relatively short intervals since last large events, favoring a segmented-fault supercycle model. Moment accumulation since 1114 indicates a ≥900-year supercycle lower bound: low-accumulation scenario leaves ~32% of 2023 moment deficit (equivalent Mw ~7.52), high-accumulation ~90% (Mw ~7.82). - Hazard implications: EAF geometry resembles the SAF–SJF system. Given higher slip rates on SAF/SJF and the absence of M>7 events on ~500 km of southern-central SAF since 1857, a scenario analogous to Kahramanmaraş could produce ~M8 if rupture initiates on SJF and branches bilaterally onto SAF.

The combined SEBP, FFI, strong-motion validation, and far-field Rayleigh-wave analyses coherently indicate that the Mw 7.8 mainshock propagated overall at subshear speeds near the Rayleigh velocity on both branches. This reconciles conflicting reports by demonstrating the absence of Mach cone signatures expected for long-lived supershear and by showing agreement between independently estimated speeds and near-Rayleigh predictions. While local transient supershear cannot be excluded, it did not dominate the main rupture. The heterogeneous, multi-segment rupture that initiated on a splay (NF) and released most moment on main EAF segments supports a supercycle framework wherein multi-segment ruptures episodically occur after sequences of segment-limited events and quiescence. Moment accounting since 1114 suggests that an ~900-year or longer cycle is plausible, though uncertainties remain. The mechanistic insights—how nucleation on a splay can trigger bilateral propagation across a main fault system—bear directly on hazard for analogous geometries like SAF–SJF, where higher slip rates and long elapsed times since the last great events imply potential for a larger, possibly ~M8 multi-segment rupture.

This study integrates dense teleseismic, near-field seismic, GNSS, and multi-mission SAR data with SEBP and joint FFI to resolve the complex, multi-segment, overall subshear rupture of the 2023 Mw 7.8 Kahramanmaraş earthquake. Key contributions include: (1) demonstrating bilateral, near-Rayleigh rupture speeds validated by far-field Mach-wave analysis; (2) quantifying heterogeneous slip with peak ~9.2 m and total moment Mw 7.85; (3) identifying a likely immediately triggered offshore aftershock near the Karasu fault; and (4) providing quantitative support for a ≥900-year supercycle on the EAF. The geometric and kinematic analogies to the SAF–SJF system highlight the plausibility of a large, potentially ~M8, multi-segment rupture in California. Future work should: refine moment-budget estimates with improved interseismic coupling and aseismic slip constraints; investigate the dynamic conditions enabling splay-to-main-fault triggering and overcoming potential stress shadows; and further test for localized transient supershear with combined near-field and dynamic modeling.

- Supershear assessment relies on far-field Rayleigh waves at 15–25 s periods; short, localized supershear segments comparable to fault width may not generate identifiable Mach signatures at these periods. - Historic earthquake catalogs (magnitudes, locations, completeness) and temporal variability in geodetic slip rates introduce significant uncertainty into moment accumulation estimates. - Some northern strong-motion stations lacked data in the relevant time windows; bilateral interference limits early SW back-projection imaging resolution. - Inversions assume 1-D velocity structure for Green’s functions and vertical fault planes; geometric and rheological complexities may not be fully captured. - Joint FFI non-uniqueness is mitigated by multiple runs, but model trade-offs (e.g., rise time vs. slip) remain.