Fully-coupled hydrogeophysical inversion of surface deformation measurements for the monitoring of geological CO2 storage

The In Salah project in Algeria has shown that CO2 injection into deep saline aquifers leads to measurable and transient deformation of the surface. Time-series measurements of surface deformation with PS-InSAR and GPS are a promising monitoring tool for geological CO2 storage. These measurements have to be integrated with other observations to extract quantitative information about the properties of the reservoir and to constrain the evolution of the pressure distribution in the subsurface. This data integration requires a fully-coupled hydrogeophysical inversion for the reservoir parameters, based on a geomechanical and hydrological process model. As a first step, we formulate a fully-coupled hydrogeophysical inverse problem to infer the permeability distribution in a quasi-static poroelastic model. In this approach, the misfit between model prediction and observed surface deformation and hydrological data is minimized under the constraint given by the poroelastic equations. The resulting least-squares optimization problem is solved using a Newton method, which uses derivatives computed efficiently through the adjoint poroelastic equations. Both state and adjoint equations are solved with a discretely consistent fully-coupled continuous Galerkin spatial discretization and implicit time integration. The state equation has been benchmarked against the analytic solution to the Mandel-Cryer problem and a modified Mandel-Cryer problem is used to test our inverse formulation. Finally, we discuss the ability of surface deformation measurements to constrain the permeability distribution in a CO2 storage reservoir. In a numerical study we analyze how well lateral permeability variations can be retrieved from a combination of surface deformation and hydrological data.