EARTHQUAKE PERIODICITY, SYNCHRONIZATION, AND CLUSTERING IN A GEOMETRICALLY SIMPLE FAULT SYSTEM

Paleoseismic time series show that earthquakes tend to cluster in time and that neighboring faults can synchronize. Previous work has hypothesized that such synchronization and clustering can emerge from changes in the tectonic loading rate, time-variable fault strength, time-dependent loading from ductile shear zones, and phase locking of faults in their late interseismic stage. Here, we use open-source quasi-dynamic earthquake simulator Quake-DFN to model fault loading and release in time-scales comparable to those recorded in the stratigraphic sequence of paleoseismic trenches. We model a two-fault system governed by rate and state friction, where both faults are identical in dimensions and frictional parameters, with complete overlap, and we vary the spacing between the two faults and the nucleation size (∝Dc). Our models demonstrate the interevent time and magnitude variability that emerges from the elastic interactions between the faults, controlled by the fault spacing, and from the stress heterogeneity on each fault, controlled by the nucleation size. These two competing effects result in several distinct regimes: steady-state creep, quasiperiodic slow slip events on alternating faults, synchronous and asynchronous periodic full ruptures, and chaotic partial ruptures. We characterize the burstiness and memory of each regime and discuss the implications for interpreting earthquake clustering and synchronization, and their responsible mechanisms, from the paleoseismic record.