Relaxing the Separation Between On-Fault and Off-Fault Seismicity Using an Empirical Mw–Deformed Area Relation

Abstract

Integrating fault‐based and distributed seismicity in earthquake rupture forecasts (ERFs) is one of the challenges in probabilistic seismic hazard analysis. Traditional approaches impose a sharp distinction between on-fault and off-fault earthquakes, even though real seismicity systematically violates this separation. Moreover, combining fault and distributed sources often introduces double counting in the overlapping magnitude range and produces discontinuities in spatial earthquake rates. Here we propose a physically motivated method to relax the boundary between fault and off-fault seismicity by exploiting an empirical relation between moment magnitude (M_w) and the areal extent of permanent ground deformation derived from interferometric synthetic aperture radar observations. Using a published global dataset of 96 earthquakes, we recompute log-linear regressions between M_w and deformed area for different faulting styles and use the resulting footprint as a deformation-based buffer around each modeled seismogenic structure. Within this buffer, distributed seismicity rates for magnitudes above the minimum magnitude of the fault source are reduced using a distance-dependent power-law taper and removed entirely for points lying within the surface projection of the fault. The method produces a smooth, physically interpretable transition between fault and off-fault activity, avoids double counting, and naturally accounts for overlapping influence zones of nearby faults. This deformation-based taper provides a reproducible and easily implementable framework for improving ERFs in regions where both fault sources and distributed seismicity contribute to the overall hazard.