Fault Reactivation by Fluid Injection: Controls From Stress State and Injection Rate

We studied the influence of stress state and fluid injection rate on the reactivation of faults. We conducted experiments on a saw cut Westerly granite sample under triaxial stress conditions. Fault reactivation was triggered by injecting fluids through a borehole directly connected to the fault. Our results show that the peak fluid pressure at the borehole leading to reactivation increases with injection rate. Elastic wave velocity measurements along-fault strike highlight that high injection rates induce significant fluid pressure heterogeneities, which explains that in such cases, the onset of fault reactivation is not determined by a conventional Coulomb law and effective stress principle, but rather by a nonlocal rupture initiation criterion. Our results demonstrate that increasing the injection rate enhances the transition from drained to locally undrained conditions, where local but intense fluid pressures perturbations can reactivate large faults, and contribute to continuing seismicity beyond the period of injection. Plain Language Summary One of the most worrisome picture of the recent years in geophysics corresponds to the exponential increase of the seismicity in Oklahoma since the beginning of deep wastewater injections. In order to reduce seismic hazard, regulators have planned a 40% reduction in the injection volume per day. While the reactivation of fault due to fluid pressure has been extensively studied, the influence of injection rate on fault reactivation remains poorly documented. In this study, we present state of the art experimental results regarding the influence of the state of stress and of the injection rate on the onset of fault reactivation. Our results demonstrate that an increase of the stress acting on the fault and/or of the injection rate induce the transition from a drained system where the classical reactivation theory is respected, toward an undrained system where the onset of fault reactivation is not determined by conventional Coulomb law and effective stress principle. Our results suggest that in such conditions, the reactivation of fault is a function of the size of the fault patch affected by the fluid pressure, that is, the diffusion of the fluid along the fault, rather than a function of the initial state of stress.