← Back to Sessions

A Nonergodic Ground-Motion Model for Japan

Description: 

Spatially varying coefficient (SVC) models are prominent tools for developing a nonergodic ground-motion model (NGMM) that can explicitly account for the repeatable source, path and site effects. Such models allow us to extrapolate the coefficients to points in the region where the data are not available and also to track the uncertainties. In this study, we use Gaussian Process Regression to model the systematic source, path, and site effects for the Japan region using the residuals from the Abrahamson and Gulerce (2020) ground-motion model for subduction earthquakes, including both forearc and backarc stations. For ray paths crossing the volcanic arc with rupture distances greater than 100 km, there is a much steeper attenuation with distance seen for backarc stations than for forearc stations in Japan. This large difference in attenuation across the volcanic arc is not seen in other subduction zones. We investigate the influence of the volcanic arc and the geologic and tectonic structures of Japan on the varying coefficients of the NGMM. We find that the typical non-ergodic linear distance scaling approach used to define the non-ergodic anelastic attenuation can perform poorly in capturing the anisotropy in the distance scaling of the ground motion from the forearc to the backarc in Japan. That is, the rapid attenuation in the distance scaling across the volcanic arc cannot be explained by a proxy for Q only. Additional effects due to the 3-D velocity structure need to be considered. We consider an additional path term capturing the effect of the 3-D velocity structure and discuss its implication for probabilistic seismic hazard analysis.

Session: Advances in Probabilistic Seismic Hazard Analysis and Applications

Type: Oral

Date: 4/19/2023

Presentation Time: 11:15 AM (local time)

Presenting Author: Esra Zengin

Student Presenter: No

Invited Presentation: 

Authors