Fault Reactivation in Response to Saltwater Disposal and Hydrocarbon Production: Site-Specific Geomechanical Models of Induced Earthquake Sequences in the Dallas-Fort Worth Basin, Texas

Induced seismicity in response to saltwater disposal and hydrocarbon production is an unresolved and persistent issue. Multiphysics geomechanical simulations that fully encapsulate the coupled poroelastic reservoir response to injection, production, and fluid diffusion have the potential of establishing firm cause-effect relationships that are needed for effective hazard mitigation. We conducted fully coupled three-dimensional poroelastic finite element simulations for site-specific geomechanical analysis to assess the potential for reactivation of basement-rooted faults in response to saltwater disposal and hydrocarbon production in the Dallas-Fort Worth Basin near Azle and Venus, Texas. Earthquake activity with magnitudes as high as M=4 primarily occurred in the basement section of normal faults extending from basement across the Ellenburger disposal reservoir and into the overlying gas-producing Barnett Shale. At both locations, we find that using best estimates of in situ stress and fault orientation and fault and reservoir parameters do not predict fault reactivation. Fault slip is predicted assuming the most favorable, yet still admissible, fault and stress orientation, and by expanding the parameter space. Of fundamental importance is the correct selection of the model domain size and the number of injection and production wells to properly account for regional changes in formation pressure. Overall, these simulations demonstrate that neglecting coupled poroelastic effects may significantly over- or underestimate the potential for fault reactivation and thus for induced seismicity. While the implementation of interacting physical processes in these coupled simulations provides firm cause-effect relationships they are sensitive to the selection of input model parameters, increasing the demand on well-constrained field data. Because of the high computational expense, these fully coupled multiphysics geomechanical models are not suitable to stochastically explore the full parameter space thus requiring integration with computationally faster hydrogeologic and reservoir engineering approaches. Lastly, our simulations suggest that faults in the Dallas-Fort Worth basin are less prone to reactivation than previously surmised based on Coulomb failure stress assessments.