High-Precision Earthquake Locations Reveal Detailed Fault Geometries in the western Corinth Rift

Seismicity and fault networks often display a high degree of complexity and segmentation at multiple scales. At shallow crustal levels, fault geometries and rupture processes are typically characterized using seismic exploration methods and precise earthquake relocation, which delineates fault surfaces. However, standard absolute location methods can produce scattered event distributions, necessitating more advanced techniques. Here, we reconstruct the fine-scale complexity of the Corinth Gulf fault system with unprecedented detail by applying a new probabilistic relocation method, NLL-SSST-WC. This technique combines source-specific, station travel-time corrections (SSST) with waveform-coherence (WC) stacking of probabilistic locations for closely spaced events.

Our SSST-WC approach iteratively computes smoothed source-specific station travel-time corrections, which significantly improve relative relocation accuracy and event clustering by accounting for three-dimensional lateral heterogeneities in the Earth. Waveform coherence groups similar events, enhancing relative locations through stacking the corresponding probability density functions.

We analyzed more than 3,500 earthquakes (M_w0.5–5.4) that occurred during 2020 – 2021 and recorded by the 49 stations of HUSN (Hellenic Unified Seismic Network) located at distances less than 135 km from the study area. For each event, we used manually picked P- and S-phase arrivals (weighted by signal-to-noise ratio) and employed a regional 1D velocity model. During NLL-SSST relocations, we initially used a large smoothing distance (D = 999 km) to treat station residuals as static corrections (given the small source area compared to station distances). Subsequently, we iteratively reduced the smoothing distance to 2.5 km to capture finer-scale heterogeneities.

The resulting high-precision earthquake locations (with errors under 100 m) reveal detailed structures of the main fault segments, spanning roughly 10 km in an east–west orientation, dipping northward, and exhibiting extensional kinematics as indicated by focal mechanisms. The fault surface curves eastward, consistent with previous focal mechanism solutions for large earthquakes in the region. The spatiotemporal distribution of the events highlights multiple clusters and distinct migration patterns over time, revealing interactions among smaller faults adjacent to those associated with the main events, within a complex fault network identified in our analysis.

This refined earthquake imaging provides improved constraints on fault geometries, rupture initiation points, and kinematic properties, which are crucial for seismic hazard assessment. It also supports more accurate modeling of rupture processes by providing detailed input data, including precise hypocenter locations, fault orientations, and spatiotemporal clustering patterns.