A layered segmentation method for fault geometry reconstruction: integrating surface traces and aftershock sequence

Well-constrained fault geometry is crucial for understanding the fault-rupture process. However, uncertainties in non-linear inversion methods and spatial discrepancies between aftershock distributions and surface rupture traces pose challenges in resolving irregular fault geometry. In this study, we propose a novel Layered Segmentation Method (LSM), which integrates aftershock sequences and surface rupture traces to construct more reasonable fault geometry. The method segments aftershocks into discrete clusters and fits polylines to these clusters using the differential evolution algorithm, thereby overcoming limitations of conventional approaches. We validate the LSM using synthetic datasets for both high-angle strike-slip and low-angle dip-slip fault cases, demonstrating its ability to reliably reconstruct fault surfaces. In application to the 2021 Mw 7.4 Maduo and the 2022 Mw 6.7 Menyuan earthquakes, the LSM effectively reconciles the spatial discrepancies between aftershock sequences and surface ruptures, resulting in fault geometry that aligned well with observed aftershock distributions. Coseismic slip inversion based on these geometries shows that the predicted surface displacements and slip patterns are consistent with geodetic and geological observations. Compared to conventional methods, the LSM offers a more robust and physically grounded representation of fault geometry, highlighting its critical role in controlling coseismic stress and slip distributions.