Assessing Improvement in Ground Motion Prediction Accuracy Provided by Distributed Slip Models

Date: 4/25/2019

Time: 06:00 PM

Room: Grand Ballroom

Earthquake Early Warning (EEW) systems aim to warn users of impending shaking before strong ground motion reaches them. Many EEW algorithms compute ground motion by using real-time source models derived from seismic and/or geodetic data as input to ground motion prediction equations (GMPEs). The ShakeAlert system currently includes two seismic algorithms. One characterizes the earthquake as a point source; the other uses a line source representation, providing the location, strike, and rupture extent. Source-to-site distance is a fundamental input for ground motion prediction (GMP), and including finite fault effects via a line source enables more accurate GMP than using a point source when rupture dimensions are comparable to or exceed source-site distances (Boese et al., 2017). However, the extent to which GMP might be further improved by instead using distance measures derived from 3D source geometries and distributed slip models is less clear. We first explore this question by comparing predicted ground motion obtained using point, line, and 3D distributed slip models (DSMs) for idealized synthetic cases. To further explore best-case scenarios, we compare observed ground motion for earthquakes in the Next Generation Attenuation Relationships for the Western U.S. (NGA-West2) database to that predicted using published DSMs. We do the same using the line source representation of each DSM that best fits the observed ground motion for the corresponding earthquake in order to evaluate whether using the full DSM gives substantially better results. To assess the relative ability of GMPEs to leverage the added information provided by DSMs we compare results using four NGA-West2 GMPEs, for which the source-site distance is determined by designating a minimum slip magnitude that defines the rupture area, and a new approach developed by Thompson and Baltay (2018) that incorporates weighting based on spatially variable slip in the distance calculation.

Presenting Author: Jessica R. Murray

Authors