We have significantly advanced the understanding of the triggering mechanisms that induce seismicity during and after the stop of injection. We have demonstrated that the classical conceptualization in which pore pressure controls induced seismicity may serve to control post-injection seismicity when injecting into a single fracture/fault, but it fails for fracture networks, which is the typical case in the subsurface. As a result, a change in the paradigm is necessary, in which not only pore pressure changes should be considered, but also poromechanical stresses, cooling-induced stress changes, static stress transfer due to seismic and aseismic slip, and deformation-induced pore pressure changes.
We have shed light to the intriguing and counterintuitive phenomenon of post-injection seismicity. Some faults undergo stability improvement during injection because of compressional poromechanical stress induced by the expansion of fractures and intact rock caused by pressurization. After the stop of injection, pore pressure rapidly drops around the injection well, causing poromechanical stress relaxation and thus vanishing the stabilizing effect, which may lead to reactivation of those faults. As a result, bleeding-off the well may be contraindicated because the more pronounced pressure drop around the well further destabilizes distant faults.
We have identified geologic carbon storage as a technology with low induced seismicity risk, while deep geothermal systems have a larger potential to induce moderate earthquakes. The reason for this is that, in conventional geologic carbon storage, CO2 is injected into sedimentary rock, which are relatively deformable, whereas stiff, brittle crystalline rock is typically found at the depths necessary to find temperature high enough to generate electricity. The stiff crystalline basement accumulates more stress than the softer shallow sedimentary rock, making it critically stressed and prone to induce seismicity. By developing a state-of-the-art numerical model, we have been able to reproduce the spatio-temporal evolution of the monitored seismicity at the enhanced geothermal system of Basel. By developing a hybrid forecasting model that uses the effective stress changes computed numerically to estimate the seismicity rate and using statistical seismology, we have identified stimulation protocols that would have prevented the largest-magnitude earthquake after the stop of injection. We have also found that poromechanical stress changes significantly influenced the induced earthquake at Pohang, and that limiting the injection pressure would have improved fault stability in the post-injection period. To derisk geo-energies, we have advanced subsurface characterization of both porous and fractured rock interpreting laboratory experiments and by designing a long-term field test at the Mont Terri underground rock laboratory.
