Foreshocks and Aftershocks of Large-event Clusters in Simulated Catalogs

Abstract:

We employ the 3D boundary element code RSQSim with a California fault model, based on the UCERF3 fault model, to generate synthetic catalogs spanning from tens of thousands, up to a million years. The simulations incorporate rate-state fault constitutive properties, and the catalogs include foreshocks, aftershocks and occasional clusters of large events. The large-event clusters, which usually consist of event pairs, but occasionally consist of three or more large events, develop spontaneously on continuous fault segments, but clustering rates are higher in areas of structural complexity. We observe several potential indicators of impending large, secondary events comparable in magnitude to the prior mainshocks. When a large-event cluster occurs, the primary event has an especially productive aftershock sequence, roughly double the normal aftershock rates for non-clustered mainshocks; and the aftershock locations of the primary event in a cluster appear to correlate with the location of the next large event in the cluster. We find that aftershock rate is a proxy for the stress state of the faults. Also, aftershock sequences transition into foreshock sequences, wherein the average event rate increases prior to the impending large event in the cluster. These increased event rates prior to the second event in a cluster follow an inverse Omori’s law, which is characteristic of foreshock sequences in nature. Finally, the aftershock/foreshock locations migrate toward the point of nucleation of the next large event in the cluster. Empirical clustering probabilities based on aftershock rates have better predictive power than those obtained from Omori aftershock and Gutenberg-Richter magnitude frequency laws, which suggests that the high aftershock rates indicate near-critical stresses for failure in a large earthquake.