Modeling frictional precursory phenomena using a wear-based rate- and state-dependent friction model in the laboratory

•Rate and state friction models used a worn surface to prescribe frictional heterogeneity.•Simulations produce complex precursory features (e.g. burst-like seismic and slow slip).•Toughness heterogeneity affected the transition to the nucleation phase of the seismic cycle.•Dominant brittle/worn patches had the ability to produce a rate-dependent cascade response. We examine numerical models that employ the rate-and-state frictional (RSF) framework to investigate earthquake sequences using laboratory driven descriptions of heterogeneous frictional properties. Using previously obtained experimental measurement of roughness, we observed that wear produced a bimodal Gaussian distribution of surface heights, which we hypothesized produced spatial heterogeneity of the critical slip distance Dc. In this numerical study, the fault surface was binarized into discrete smooth or rough sections, producing a barcode style version of frictional heterogeneity. The fault was predominantly rough except for two dominant asperities (A1 and A2) representative of larger polished sections. We simulated the resistive effect of increasing the fracture energy (toughness) of the rough barriers while maintaining constant properties of the embedded brittle/smooth asperities. Our numerical simulations generated burst-like seismic events and aseismic transients throughout the interseismic phase. At the late interseismic phase, bursts of seismicity (foreshocks) interacted with the accelerating preslip region at the transition to the preseismic (nucleation) phase. At lower levels of toughness heterogeneity, the slip rate increase was roughly inversely proportional to the time-to-failure tf for larger events. As fault toughness was increased, the dominant asperities initiated nucleation and thus force deviations of the fault from the smooth 1/tf acceleration observed for the homogeneous case, producing a rate-dependent cascade response. The calculations were validated by comparing two independently measured metrics from the experiments: (1) The expansion rate of slow ruptures during the interseismic and preslip phase and (2) the scalar seismic moment and source dimensions. While our study does not address the scaling problem, these results help to understand laboratory experiments that investigate transition to the preseismic (nucleation) phase during complex earthquake sequences.