Constraining On- and Off-Fault Nonlinear Dynamic Rupture Parameters via Hierarchical Bayesian Inversion of GNSS and Satellite Data for the 2019 Mw 7.1 Ridgecrest Earthquake
Alice-Agnes Gabriel, Zihua Niu, Maximilian Kruse, Linus Seelinger, Nico Schliwa, Heiner Igel, & Yehuda Ben-ZionSubmitted September 7, 2025, SCEC Contribution #14774, 2025 SCEC Annual Meeting Poster #TBD
Dynamic rupture simulations can incorporate physical laws and complex subsurface structures to better constrain earthquake processes. For example, the interactions between 3D co-seismic off-fault damage, seismic radiation, and rupture dynamics, which may facilitate rupture cascading across faults, linking delay times to damage rheology and fault zone evolution. However, despite increasing availability of high-quality surface observations, parameter inference with uncertainty quantification based on physics-based dynamic rupture simulations remains challenging due to computational costs. Typically, 3D simulations require thousands of CPU hours per model realization.
Here, we present the first Bayesian inversion of complex 3D dynamic rupture simulations with off-fault plasticity, jointly resolving on- and off-fault nonlinear parameters, constrained by fault-parallel surface offsets from satellite imagery, high-rate GNSS time series, and static GNSS displacements. We employ a multilevel delayed acceptance (MLDA) algorithm, reducing the number of costly simulations needed for uncertainty quantification by combining fast approximate models with fewer fully resolved simulations.
Off-fault plasticity critically influences seismic hazard by affecting rupture dynamics and ground motions. Our inversion quantifies uncertainties and correlations among on- and off-fault dynamic rupture parameters for the 2019 Mw 7.1 Ridgecrest earthquake, using multidisciplinary surface deformation data and over four million CPU hours of simulations. We identify a strong correlation between on-fault frictional weakening and off-fault plasticity, showing increased inelastic deformation can compensate for stronger velocity-weakening frictional behavior.
Preferred rupture models feature spatially varying friction parameters (a-b), increasing from northwest to southeast, reducing velocity-weakening effects, and improving agreement with observed surface offsets. This variation aligns with along-strike changes in fault maturity. Lower off-fault plastic cohesion further enhances the match to observed surface deformation, consistent with geophysically imaged shallow damage zones.
We infer a shallow slip deficit (SSD) of 13.1% ±5.1%, aligning with previous kinematic inversion estimates. Our results highlight the feasibility of integrating 3D dynamic rupture simulations and multilevel Bayesian inversion to rigorously characterize earthquake sources and quantify uncertainties.
Citation
Gabriel, A., Niu, Z., Kruse, M., Seelinger, L., Schliwa, N., Igel, H., & Ben-Zion, Y. (2025, 09). Constraining On- and Off-Fault Nonlinear Dynamic Rupture Parameters via Hierarchical Bayesian Inversion of GNSS and Satellite Data for the 2019 Mw 7.1 Ridgecrest Earthquake. Poster Presentation at 2025 SCEC Annual Meeting.
Related Projects & Working Groups
Fault and Rupture Mechanics (FARM)
