Group B, Poster #212, Computational Science (CS)

Recent Enhancements to the CyberShake PSHA Platform

Scott Callaghan, Philip J. Maechling, Fabio Silva, Kevin R. Milner, Kim B. Olsen, Robert W. Graves, Christine A. Goulet, Karan Vahi, Ewa Deelman, Albert R. Kottke, Yifeng Cui, Thomas H. Jordan, & Yehuda Ben-Zion
Poster Image: 

Poster Presentation

2023 SCEC Annual Meeting, Poster #212, SCEC Contribution #13216 VIEW PDF
The SCEC CyberShake platform performs physics-based probabilistic seismic hazard analysis using 3D wave propagation simulations with reciprocity to produce a rich suite of site-specific and regional hazard data products. We highlight two recent major CyberShake developments: (1) the calculation of Study 22.12, a broadband CyberShake hazard model for Southern California, and (2) the release of the CyberShake data access tool, designed to streamline the process of obtaining CyberShake data and metadata for the SCEC community.

In 2023, we completed CyberShake Study 22.12, a Southern California broadband study including both deterministic low-frequency and stochastic high-frequenc...
y components. For this study, we modified the 3D velocity model CVM-S4.26.M01 by applying a merged taper, in which a Vs30-based taper is applied to the upper 700m at locations for which it will reduce Vs. This reduces near-surface velocities outside of the basins and has been found to yield better fits with observations (Hu, Olsen, and Day, 2022). Additionally, we updated the rupture generator code to the version used in the SCEC Broadband Platform (v22.4) (Pitarka et al. 2021), increased hypocentral sampling to better capture variability, and added sampling of variability in rupture velocity. We leveraged the Pegasus workflow management system to automate the execution of this 3-month study on OLCF Summit. We will present results from this study, including comparisons with empirical ground motion models and previous CyberShake studies. Compared to previous studies, we found generally higher ground motions outside of the basins, likely due to the taper. We also found a trend towards higher ground motions at shorter periods but lower at longer, consistent with changes to the rupture generator which increased short-period energy.

To improve access to data products, we have created and released the CyberShake data access tool. This Python-based tool prompts the user for the hazard model of interest, data product to retrieve, and filters to apply, and then uses a CyberShake database and data files hosted at USC to deliver the requested data. Currently, site metadata, event metadata, intensity measures, and seismograms from Study 22.12 are accessible using the software. The goal is to reduce barriers for SCEC community members to use CyberShake results in their own research. We will describe the tool and show examples of how it can be used to obtain CyberShake data for common use cases.