Evaluating probabilities of earthquake fault jumps from 2D numerical simulation of seismic cycles

The efficiency of an earthquake to cross a barrier can be evaluated based on geometric and frictional properties of faults, and specific seismic parameters such as the stress drop during an earthquake. Numerical modelling of seismic cycles allows to generate thousands of seismic events and to explore the effect of the physical properties with respect to a barrier effectiveness criteria. The probability of an event passing a barrier can thus be evaluated on the basis of this barrier effectiveness criteria. Such approach has been used for frictional barriers and fault bends. In this study we focuses on earthquake fault jumps which has been observed on multiple occasions such as the latest 2024 M7.5 earthquake in Japan. We use the quasi-dynamic algorithm VEGA, which numerically simulates seismic cycles of 2D fault networks and is based on rate and state friction. The problem is simplified to two planar faults separated by a gap. Among other parameters, we explore the effect of the overlap, distance and angle between the two faults. The loading of a fault network can be done in multiple ways. We thus explore the impact on the dynamics of sequences of earthquakes either from a far-field stress loading or from imposing a back slip rate loading on each fault. We also look at the effect of adding creeping - velocity strengthening - sections at the borders of the faults. We finally compare our results with the statistics of jump probabilities from published observed seismic events. Our study allows for a rapid assessment of thousands of earthquake scenarios and is a promising approach to facilitate the integration of earthquake physics into seismic hazard.