Rapid basin-scale hydro-geomechanical modeling of fluid injection for induced seismicity analysis

Causal assessment, hindcasting, and forecasting of injection-induced seismicity hinge on subsurface pore pressure and stress tensor as inputs. Such inputs are usually obtained from hydro-geomechanical (HM) modeling calibrated against observations such as ground deformation. While classical numerical methods such as the finite element method remain an important avenue to HM modeling, they are expensive and impractical for basinscale injection problems. Here, I develop an alternative method balancing physics, details, and speed for modeling basin-scale pressurization of an injection zone and the induced full stress tensor and displacement vector everywhere. To this end, I simplify the injection zone as a poroelastic layer embedded within a 3D elastic half-space and the injectors as arbitrarily distributed line sources with arbitrary mass rates. I conduct the hydrological modeling based on two classic solutions. One is a slight improvement of a pore pressure diffusion solution, and the other is adapted from a poroelastic solution. I then generalize both solutions to include hydraulic full anisotropy and prove their equivalence irrespective of constant or variable injection history. For the geomechanical modeling, I derive from the quasi-static force balance law three pore pressure-related source terms driving stress changes. They are two (equivalent) body forces including the incremental fluid specific weight and the negative lateral pore pressure gradient, and a surface force arising from the discontinuous pore pressure across the injection zone top and bottom. I then discretize the injection zone into numerous sufficiently small source cells and compute from these source terms their discrete point forces. Next, I employ the 3D half-space elasto-static displacement and stress Green's functions associated with unit point forces, and perform numerical integration over the injection layer, to produce geomechanical perturbations at arbitrary receiver locations. For the integral involving the pore pressure gradients, I outline three computational strategies. Lastly, I demonstrate the method over a multi-decadal, dual-zone saltwater disposal problem in the Northern Delaware Basin-currently one of the more seismically active areas in North America-and model in-zone pore pressure, subsurface (e.g., basement) stresses, and ground deformation. When pore pressure solutions are available from external reservoir models, the geomechanical model can also be used for other geo-energy applications such as hydrocarbon production and carbon storage.