Validation of Broadband Ground Motion From 3D Dynamic Rupture Simulations: Towards Fine-Tuning Gmpes and Supplementing Observed Ground Motions in Data-Sparse Regions

Date: 4/25/2019

Time: 06:00 PM

Room: Grand Ballroom

We numerically model 3D dynamic ruptures of earthquakes along complex fault-geometry and simulate ground motion time series at surface stations, computationally accurate at frequencies up to ~5-10 Hz. Our simulations focus on modeling ruptures with realistic initial conditions (matching that of empirical observations) of both friction and stress, as well as constraints arising from datasets of fault geometry. In addition, we include nonlinear effects in our model via Drucker-Prager plasticity in a 1D layered velocity structure. By including these complexities in our model, we are able to realistically simulate ground motions at frequencies across a broad bandwidth of relevance for structural engineering purposes. Here, we focus on a dip-slip event matching the geometry of the Salt Lake City segment of the Wasatch Fault Zone and investigate the role that the hypocentral location, free-surface topography, and other rupture-related components have on ground motion. We compare our synthetically generated data with GMPEs, separating ground motion as a function of both median and variability. We analyze features that can be isolated to better understand the ability of our model to fit certain characteristics observed in empirical data: directivity, the hanging wall-effect, and distance dependence. We find good comparisons with 4 leading GMPEs models applied to near-source distances from ~Mw 7 events, but see localized variations dependent on hypocentral depth (and location along strike), in addition to influences from both small and large-scale features of fault geometry. The inclusion of off-fault plasticity is seen to directly affect ground motions across the fault plane, in particular with reduction on the hanging-wall. This suggests that we may be better able to predict future ground motions (and their possible range) from modeling instances of rupture, specific to regional and localized variation in observed fault geometry and stress conditions. Next, we plan to include 3D variations in the velocity model to study path/site effects on ground motion.

Presenting Author: Kyle Withers

Authors