Path and Site-Effects Revealed by Source-Normalized Intensities from a Suite of Three-Dimensional Simulations of M7.0 Hayward Fault Ruptures Resolved to 5 Hz

Session: Numerical Modeling of Rupture Dynamics, Earthquake Ground Motion and Seismic Noise
Type: Oral
Date: 4/29/2020
Time: 11:45 AM
Room: 230 + 235
Description: 

Path- and site-effects due to three-dimensional (3D) wave propagation cause systematic biases in earthquake ground motion intensity measures (IMs) relative to ergodic empirical ground-motion models (GMMs). If unmodelled, these biases contribute to large standard deviations of the IMs in GMMs and confound seismic hazard estimates. Fully deterministic 3D wave-propagation simulations can be used to characterize the bias in path- and site-specific ground motion IMs due to Earth structure. We explored the bias due to path- and site-effects using SW4 ground motion simulations resolved 0-5 Hz for a suite of moment magnitude (M_W) 7.0 earthquake ruptures on the Hayward Fault in northern California. We kept fault geometry fixed and computed the response for a suite of 50 ruptures in both 3D and average plane-layered (one-dimensional, 1D) Earth models. The 3D model included surface topography, a minimum shear wavespeed of 500 m/s and a minimum grid spacing of 12.5 m. The source-normalized site term, dz, was formed as the difference of natural logarithm IM from the 3D and 1D Earth models for the same rupture model (z_3D and z_1D, respectively): dz = z_3D - z_1D. We show that dz is a robust estimate of path- and site-effects for the specific rupture geometries considered and is independent of rupture directivity. The dz values for long-period spectral accelerations and peak ground velocity are correlated with site-specific parameters such as depths to the shear-wave velocities of 1.0 and 2.5 km/s (Z_1.0 and Z_2.5, respectively), although deviations from average behavior reveal how simulations capture more complex basin response than Z_1.0 and Z_2.5. Source-normalized site terms provide a model-based calibration of path- and site-effects that can be used to adjust ground motions computed for the assumed 1D Earth model to account for wave-propagation effects from a 3D Earth model. Non-circular application of the site term to ground motions from the 3D model can reduce the standard deviation of IMs for a specific rupture by 30-50% compared to the ergodic approach.

Presenting Author: Arthur J. Rodgers

Authors