Induced seismicity hazard forecasting in the San Juan Basin: Stresses, faults, and computation

Quantifying the seismicity associated with fluid injection into the subsurface and the resulting seismic hazard is crucial for evaluating the safety of geologic carbon storage. To better understand the risks of induced seismicity, it is essential to assess the likelihood of fault rupture due to pressure perturbations from injection. This allows operators to gauge the potential seismic hazard in a given region of interest. We present a three-step workflow developed for the San Juan Basin CarbonSAFE project, which is transferable to other geologic carbon storage or fluid injection sites worldwide. Our workflow begins with evaluating three distinct methods for approximating the state of stress. The Simpson's coefficient method provides a first-order approximation, generating a simple model with minimal input data. The State of Stress Analysis Tool employs Bayesian statistics to produce both a model of the stresses and the associated uncertainty. The hybrid approach integrates petrophysical correlations to determine which of the first two methods should be applied at specific depths. The next step involves using machine learning or ant-tracking techniques to automatically identify fault sizes, shapes, locations, and orientations from a three-dimensional seismic image. The final step computes the Coulomb failure function - a measure of geomechanical stability - by applying the state of stress to a kriged surface through the fault detected. Our results provide a hazard map that highlights the seismic risk along faults. This map can guide operators in assessing the geomechanical safety of fluid injection at different reservoir locations, helping to mitigate the risk of induced seismicity.