Group B, Poster #242, Ground Motions

Evolution of ground motion characteristics over earthquake cycle timescales

Dunyu Liu, Yongfei Wang, & Christine A. Goulet
Poster Image: 

Poster Presentation

2022 SCEC Annual Meeting, Poster #242, SCEC Contribution #12513 VIEW PDF
Dynamic rupture models provide a critical link between earthquake physics, velocity structures, and ground motion. This physics-based simulation approach is particularly valuable to assess potential near-fault broadband ground motions affected by a specific fault in a critical area of interest, but poorly represented in empirical datasets. However, such applications suffer from significant uncertainties associated with pre-event assumptions, notably for the initial stress conditions. These are partially inherited from earlier surrounding earthquakes, meaning that the variability inherent to rupture processes and resulting ground motions may depend on the earthquake sequence history. The cons...ideration of the stress and strength evolution over rupture sequences help constrain the state of stress and strength just before a new event. To address this issue, we propose to simulate dynamic rupture and high-frequency (up to 3 Hz) ground motion over earthquake cycles that consider temporally evolving stress and strength heterogeneities. Earthquake cycles on a structurally rough strike-slip fault governed by rate- and state- friction with off-fault damage will be simulated. Key output ground motion metrics (e.g., spectral accelerations as functions of rupture distances and periods) will be validated against empirical ground-motion models (GMMs) over earthquake cycles. Technically, different phases of an earthquake cycle such as dynamic ruptures and interseismic phases will be organized as individual, but interconnected computing modules. A python-based coupling infrastructure will be developed to facilitate communications and resource allocation between modules. This distinct approach will provide a pathway to study ground motion in light of fault system evolution, which permits a comprehensive comparison of what characteristics are persistent and which evolve, and on what timescales. We intend to (1) define external two-way data flows that interface with the various existing modules without changing their internal configurations and (2) develop the potential for on-demand generation of datasets that can be validated against hazard-targeted observations such as fault displacements or ground shaking. In addition, this work bridges a gap between the SEAS TAG and the Dynamic Rupture Verification and Validation Group and contributes to validating a new physics-based simulation ecosystem that aims to improve and quantify uncertainties in seismic hazard analysis.