An empirical- and simulation-based ground-motion model for Southern California

Morgan P. Moschetti, Eric M. Thompson, Nicolas Luco, Thomas H. Jordan, Peter M. Powers, Allison M. Shumway, Mark D. Petersen, Robert W. Graves, Scott Callaghan, Christine A. Goulet, Kevin R. Milner, Philip J. Maechling, Feng Wang, & John Rekoske

Submitted August 15, 2019, SCEC Contribution #9820, 2019 SCEC Annual Meeting Poster #022

We report progress on the development of a long-period (T>=2 s) ground-motion model (GMM) for Southern California based on an empirical GMM and the CyberShake simulations. Ground-motion prediction within sedimentary basins is hampered by complex wave propagation phenomena, which preclude the use of simple analytical models in the standard regressions for empirical GMMs. As a first step in implementing results from CyberShake into probabilistic seismic hazard models, we focus on the average basin amplifications that result from analyses of the CyberShake suite of simulations. The joint empirical-simulation-based GMM uses the CyberShake simulations to adjust the ground motions from the Abrahamson et al. (2014; ASK14) GMM for active crustal regions within the sedimentary basins of Southern California. Basin amplification is implemented through site maps (“b-maps”), which were computed by the averaging-based factorization method of Wang and Jordan (2014) for the GMMs of the NGA-West-2 project. The site maps provide spatially varying, and period-dependent, amplifications with respect to a mean ground-motion level. We adjust the site maps through a linear mixed-effects regression to a database of ground motions from M>=3.5 earthquakes from the Los Angeles region. The database contains good azimuthal coverage and multiple recordings per station, primarily from small- to moderate-sized earthquakes, thereby minimizing finite-fault effects and path-dependent amplifications. The resulting model--the “joint empirical-simulation-based GMM”--combines empirical ground-motion scaling with distance, magnitude, fault geometry, and sense of slip from ASK14 together with the average basin amplifications modeled by CyberShake. The study provides one framework for bringing the 3D-simulation efforts into the U.S. National Seismic Hazard Model (NSHM), for quantifying and reducing seismic risk in the Los Angeles region. Future work will test the accuracy of the resulting GMM against records from larger-magnitude earthquakes. Implementation in the NSHM may also consider concomitant reductions in aleatory variability, which are likely to have large effects on low-probability ground motions that affect engineering design in the region.

Key Words
ground motion prediction, Southern California, CyberShake, 3D simulations

Citation
Moschetti, M. P., Thompson, E. M., Luco, N., Jordan, T. H., Powers, P. M., Shumway, A. M., Petersen, M. D., Graves, R. W., Callaghan, S., Goulet, C. A., Milner, K. R., Maechling, P. J., Wang, F., & Rekoske, J. (2019, 08). An empirical- and simulation-based ground-motion model for Southern California. Poster Presentation at 2019 SCEC Annual Meeting.


Related Projects & Working Groups
Ground Motions