Coupled flow and geomechanics modeling of fracture reactivation and induced seismicity in the Basel geothermal field

Triggered and induced seismicity is at the cornerstone of discussions surrounding a wide range of subsurface technologies, e.g. unconventional hydrocarbon recovery, geologic carbon sequestration, underground gas storage, and geothermal energy extraction. We revisit the geothermal experiment in Basel, Switzerland, in which over 11 thousand cubic meters of water were injected into deep fractured crystalline rock. The injection took place in December 2006 and was performed at a single injection well, in three stages with increasing injection rate. Seismicity in the region increased during the experiment—with most of the seismic events occurring in the month following injection—and declined slowly after, but with seismic events still being recorded years after injection. The increase in seismicity is caused by reactivation of the pre-existing fractures: an increase in pore pressure decreases the normal effective stress across the fracture, which according to the Mohr-Coulomb failure criterion are therefore more prone to slip. The underlying processes, however, may be more complex. It is unclear the role that enhanced hydraulic connectivity in the fracture network plays on triggered seismicity, and on the emergence of seismicity clusters in space. It is also unclear what determines the delay between injection and recorded seismicity, and whether it can be explained by means of pressure propagation and/or dynamic weakening of fractures due to a drop in the friction coefficient as a result of decreased roughness from fracture slip. Here, we employ a computational model of coupled flow and geomechanics to quantitatively assess the impact of fluid injection on the recorded seismicity. We develop a simulation model that incorporates more than ten fractures, whose location, rake and dip are consistent with clusters of seismicity from a relocation of hypocenters and focal mechanisms. We adopt a multiscale description of flow (representing these fractures planes explicitly, and flow through the fracture network as an equivalent continuum medium) and a rate- and state-dependent friction law for the fractures. We conduct Bayesian model inversion from well data and seismological data, and perform a post-mortem assessment that allows us to identify plausible mechanisms responsible for the observed seismicity.