Integrated Theoretical Modeling and Empirical Analysis of Injection-Induced Fault Slip and Seismic Hazard: From Parameter Sensitivity to Real-World Case Studies

Injection-induced seismicity is a complex, multi-physics phenomenon driven by coupled pore-pressure diffusion, poroelastic stress changes, and static stress transfer. With subsurface fluid injection increasingly central to energy production and waste disposal, understanding how these processes influence fault slip and seismic hazard has become critical. This thesis integrates numerical and empirical investigation of injection-induced fault behaviour, combining high-resolution rate-and-state friction models, dynamic stress simulations, and real-world case studies to identify key controls on seismic and aseismic slip. The thesis first explores how injection rate and cumulative volume jointly shape fault behavior using a long-term dynamic rupture framework. Thousands of simulated earthquake cycles show that these parameters govern transitions between delayed rupture, aseismic transients, and immediate seismic clustering. Notably, over 99\% of the total released moment occurs aseismically, underscoring the stabilizing role of slow slip and the enduring perturbation of the seismic cycle, even from short-lived injections. A novel functional approximation framework is developed to model spatiotemporally heterogeneous pressure diffusion along faults. It reproduces finite-difference outputs at a fraction of the computational cost (\string~94\% reduction), enabling thousands of simulations without loss of realism. Spatial heterogeneity raises the pore-pressure threshold for induced failure and heightens sensitivity to injection rate. These tools are applied to a digital reconstruction of the 2016 Mw 5.8 Pawnee earthquake, the largest known case of injection-induced seismicity in Oklahoma. By coupling MODFLOW-based hydrogeological modeling with poroelastic and Coulomb stress analyses, the study shows the mainshock did not nucleate on a critically stressed fault. Instead, pore-pressure diffusion activated foreshocks on the optimally oriented Labette Fault, which then transferred sufficient static stress to rupture the subcritically stressed, previously unmapped Sooner Lake Fault. Together, these findings show that induced seismicity arises from non-linear interactions across multiple spatial scales, driven by aseismic slip, fault-to-fault stress transfer, and evolving injection conditions. They expose the limits of mitigation strategies based on simple magnitude or pressure thresholds and argue for physics-based hazard frameworks that account for spatiotemporal stress evolution, aseismic processes, and inter-fault interactions. This thesis integrates numerical modeling with empirical validation to strengthen the scientific basis for risk management in fluid injection settings.

Récupéré en direct depuis OpenAlex et désinversé. Les résumés ne sont pas conservés dans cette base de données : les index inversés représentent 8,6 Go des 9,3 Go de texte de la base, et le serveur dispose de 13 Go libres.