Modeling seismic and aseismic fault reactivation in deep geothermal systems: impacts of pore-pressure, poroelastic, and thermal stresses.

Injection-induced seismicity is commonly attributed to fluid diffusion, poroelastic stress transfer, and stress loading associated with aseismic creep. In deep geothermal systems, thermal stresses generated by fluid–rock heat exchange constitute an additional mechanism that may significantly influence seismicity. Modeling studies suggest that pore pressure diffusion plays a dominant role in fault reactivation during the early stages of geothermal operations (months to a few years), whereas the contribution of thermal stress may become significant over longer timescales (years to decades). However, the relative contributions of these triggering mechanisms in both aseismic and seismic fault reactivation remain poorly constrained, and the influence of fault properties and operational strategies is still largely unexplored.

In this work, we investigate how thermal stress changes drive aseismic and seismic fault reactivation, as well as their contribution relative to pore pressure and poroelastic stress redistribution. We introduce the effective normal and shear stress changes calculated from a 3D Thermo-Hydro-Mechanical (THM) model into a 3D fault slip sequence model in the rate-and-state frictional framework to simulate fault slip responses to geothermal operation scenarios. We explore two background fault slip scenarios, a seismically active fault undergoing multiple seismic cycles, and a fault that accommodates periodically recurring aseismic deformation.

Overall, our results indicate that the timing of stress perturbation relative to the background slip cycle exerts the primary influence on clock advance of both aseismic and seismic fault slip behaviour during injection operations. All model configurations and variations in the timing of applied stress perturbations lead to an initially aseismic response in both the seismic and aseismic faulting scenarios. However, in the seismic-cycle scenario, aseismic slip accelerates to seismic slip within the injection period when perturbations are introduced late in the background seismic cycle, that is, when the fault is already close to failure. The aseismic-cycle scenario exhibits similar behaviour under conditions where the fault experiences a substantially larger (a factor of 2) stress perturbation. Pore pressure changes preferentially control both the timing and extent of coseismic fault rupture, primarily due to the larger spatial region over which they affect the fault plane. Thermal stress changes are significant in magnitude, yet they exert a comparatively minor influence on the overall aseismic and coseismic slip distribution on the fault because of the limited extent of the fault surface over which they act.

From an operational perspective, we find that cyclic injections generate larger pore-pressure changes compared to a constant injection rate, which may promote earlier seismic reactivation or facilitate a transition from aseismic to seismic slip during the injection period. Finally, our results suggest that a simple doublet configuration could significantly reduce the risk of seismic fault reactivation during injection and production. In such cases, fault architecture (conduit vs. barrier) and relative positioning of injection and production wells play a critical role.