Ground Motion Simulation and Seismic Loss Estimation Using the Stochastic Finite-Fault Method in the Western Quebec Seismic Zone, Canada

The stochastic finite-fault (SFF) method is a powerful approach for generating time-series ground motions, particularly in data-sparse regions such as Eastern North America (ENA). This study applies an extended SFF framework that improves upon the conventional SFF method by incorporating spatially variable stress rather than a uniform stress parameter. In this approach, slip models are converted to spatially variable stress distributions following the method of Ripperger and Mai (2004). The framework is applied to the Timiskaming Fault within the Western Quebec Seismic Zone, Canada.

The fault geometry is constrained using Coulomb stress change analysis, and three scenario magnitudes (Mw 6.0, 6.5, and 7.0) are considered based on regional seismic history. For each magnitude, two rupture configurations are modeled: a fixed hypocenter at the fault center and a floating hypocenter distributed across the fault plane. To capture rupture variability, 50 stochastic slip distributions with prescribed spatial correlation are generated for each scenario. Ground motions are simulated at the locations of buildings along the fault using ENA-specific empirical models for additional SFF input parameters.

The simulated ground motions are compared with predictions from ground-motion prediction equations (GMPEs), including the NGA-East models. Using both SFF- and GMPE-based motions, seismic losses are estimated for wooden buildings in the study area, which represent approximately 90% of the regional building stock. Ground motions serve as the primary input for the seismic risk assessment. The results highlight key differences between stochastic simulations and empirical predictions, providing valuable insights for improving seismic hazard and risk assessment in data-limited regions.