New Ground-Motion Model With Long-Period Non-Ergodic Path Effects From the Cybershake Simulations in the Southern California Region
Description:
Sung et al. (2023) developed a methodology to integrate the 3-D path effects in CyberShake simulations (Graves et al., 2011) into the Abrahamson et al. (2014) (ASK14) ground-motion model (GMM), creating a non-ergodic GMM that captured site-specific path effects caused by the 3-D crustal model using the varying coefficient model (VCM), but the example application only used a subset of 600 scenarios from the v15.4 CyberShake simulations and only modeled the 3-second response spectral values. In this study, we utilized a much larger dataset from the v15.4 CyberShake simulations, which includes 565,000 events and 117 million seismograms at 336 stations for spectral periods of 2, 3, 4, 5, 7.5, and 10 seconds. Additionally, we applied a new form of the spatial correlation (Lacour et al., 2024) that considers the azimuths differences between ray paths that better reflects the spatial correlation in the CyberShake data. For most ray paths, the larger data set shows a strong correlation between the path effects for moderate and large earthquakes with the same closest point to a site (correlation coefficient of 0.7 to 0.85), indicating that the path effects of large-magnitude earthquakes can be approximated by using the path effects of smaller earthquakes. The modified ASK14 GMM with CyberShake site and path effects at periods of 2-10 s, and it has reduced aleatory variability. Compared to an ergodic GMM representation of the CyberShake data, for the period range of 2-10 s, the total standard deviation is reduced from 0.565 to 0.636 for the ergodic GMM to 0.435 to 0. 544 for the non-ergodic GMM. An additional advantage of using the larger data set is that it reduces the potential for overfitting as compared to the previous subset. We present examples of hazards calculation for T = 3 s and T = 5 s for the ergodic and non-ergodic GMMs and for the hazard computed using the CyberShake data directly. The resulting modified ASK14 GMM can effectively capture the path effects observed in the CyberShake simulations, including both the median and the aleatory standard deviation.
Session: The 2023 USGS National Seismic Hazard Model and Beyond - III
Type: Oral
Date: 5/1/2024
Presentation Time: 02:45 PM (local time)
Presenting Author: Chih-Hsuan
Student Presenter: No
Invited Presentation:
Authors
Chih-Hsuan Sung Presenting Author Corresponding Author karensung@berkeley.edu University of California, Berkeley |
Norman Abrahamson abrahamson@berkeley.edu University of California, Berkeley |
Maxime Lacour maxlacour@berkeley.edu University of California, Berkeley |
Xiaofeng Meng xiaofenm@usc.edu Statewide California Earthquake Center |
|
|
|
|
|
New Ground-Motion Model With Long-Period Non-Ergodic Path Effects From the Cybershake Simulations in the Southern California Region
Category
The 2023 USGS National Seismic Hazard Model and Beyond