Effects of fluid-induced aseismic slip on earthquake cycles and earthquake source processes

Abstract:

Understanding the underlying physics of faults under fluid perturbation is pivotal to effective seismic hazard mitigation. Through numerical modeling of rate-and-state faults, we investigate the effect of fluid-induced aseismic motions of the fault on the timing and source processes of earthquakes following pore pressure perturbations at different times during a selected seismic cycle. Our 3-D fault model features a velocity-weakening (VW) patch embedded in a large velocity-strengthening (VS) region. For small pore-pressure perturbations (up to 25% of the average event stress drop), we observe a wide range of fault aseismic responses that lead to either the advance or delay of the next seismic event. Earthquakes can sometimes be delayed even when pore pressure perturbation occurs late in the seismic cycle. There is a positive correlation between the aseismic moment released during the perturbed interseismic period and the time change of induced mainshocks (positive time change indicates delay). In terms of source processes, we find that aseismic motions triggered on the fault significantly change the nucleation process of induced events, with the nucleation size varying from 40% to 160% of the nucleated size of unperturbed events. While the variation of magnitudes of induced earthquakes in our model space remains small, there is a clear negative relationship between nucleation size and moment magnitude. Apart from mainshocks that rupture the entire VW region, the aseismic transients occasionally trigger seismic precursors. Pore pressure perturbations may also have lasting effects in subsequent seismic cycles such as cycle-lengthening.

Our ongoing work is directed toward quantifying the properties of the triggered aseismic transients within the seismogenic zone, i.e. their spatial and temporal evolution with respect to the onset of pore pressure perturbation. We aim to better categorize aseismic motions that can bring drastically different responses on the fault. We will also incorporate specific fault frictional properties and fluid injection histories from field observation to provide better constraints on fault conditions favorable for earthquake triggering.