Numerical Modeling of Injection-Induced Earthquakes Using Laboratory-Derived Friction Laws

Abstract

Modeling injection-induced earthquakes requires coupling porous media flow, rock mechanics, and fault friction. Highly nonlinear laboratory-derived constitutive laws for fault friction pose a major challenge for computational models that couple flow and geomechanics. We present a finite element formulation to simulate injection-induced earthquake sequences in rate-and-state faults embedded in poroelastic media. We simulate all phases of the stick-slip cycle: from fault reactivation as pressure accumulates near the fault, to earthquake nucleation phase, coseismic rupture propagation and interseismic periods. Our simulations are quasi-dynamic: we neglect inertia, and adopt the so-called radiation damping approximation. We perform validation and verification tests based on a simple spring-block analog that allows straightforward comparison between our 2-D finite element model and the single-degree-of-freedom dynamics. We also verify our frictional contact algorithm by simulating injection-induced earthquakes on a slip-weakening strike-slip fault. We finally study the impact of different rate-and-state laws (aging and slip laws), as well as the role of the degree of poroelastic coupling, by varying the Biot coefficient. We characterize the undrained pressure response triggered by the fast propagation of rupture fronts. Undrained pressure changes during rupture act as an additional coseismic weakening mechanism, controlling the propagation or arrest of the rupture fronts. We find that this feedback between pore pressure and slip propagation, which is absent in uncoupled simulations, leads to distinctively asymmetric rupture patterns in induced earthquakes in poroelastic media. Our results show that capturing the coupling between fault frictional processes and rock poroelastic behavior requires well-resolved and fully coupled simulations.