Informing Rupture Directivity Modeling with CyberShake Simulations

Jeff Bayless, & Norman A. Abrahamson

Submitted September 10, 2023, SCEC Contribution #12866, 2023 SCEC Annual Meeting Poster #177 (PDF)

In this study, a database of near-fault CyberShake simulations, consisting of many earthquake sources with multiple hypocenters and rupture realizations, is used to evaluate rupture directivity effects in the simulated ground motions. CyberShake ground motion residuals are calculated from the Meng et al. (2023; Mea23) ground motion model (GMM) to explicitly quantify the directivity effects. The overall performance of the Bayless et al. (2020; Bea20) directivity model as compared with the simulation residuals is promising, but wide-ranging. There are many instances of source and hypocenter location with residuals matching Bea20 quite well, and there are many instances which do not match as well. This is the same observation Bea20 made with respect to the recorded data used to develop their model.

Peak amplitudes of mean simulation residuals are found to be generally lower than the mean Bea20 predictions, and the variance of the simulation residuals are found to be generally larger than Bea20. Nonetheless, modeling improvements gained by incorporating Bea20 are quantified through residual variance reductions. At T=5 sec, residual variance reductions of between 0.05 and 0.09 are found. This reduction is larger than the empirically derived reduction from Bea20, which is based on a relatively sparse dataset, and represents about a 12% reduction in one component of the aleatory variability, which is large enough to be impactful in seismic hazard applications.

Key Words
directivity, cybershake, GMM

Citation
Bayless, J., & Abrahamson, N. A. (2023, 09). Informing Rupture Directivity Modeling with CyberShake Simulations. Poster Presentation at 2023 SCEC Annual Meeting.

Related Projects & Working Groups
Ground Motions