Regional Effects on Style-of-Faulting Ratios in Ground Motion Models

Date: 4/25/2019

Time: 06:00 PM

Room: Grand Ballroom

An unremarkable number of ground-motion prediction equations (GMPEs) include the style-of-faulting (SoF) as an additional predictor variable to the invariably used magnitude, distance and site classification terms. This study focuses on regional effects on the SoF-ratios while using a large dataset of previously compiled strong ground motion recordings of the Western United States and Taiwan (NGA-West2), Europe (Engineering Strong Motion Database, ESM) and Japan (KiK-Net Database). Only records of shallow crustal earthquakes (Mw 3.5 - 8.0) with a max. depth of 35km and a maximum source-to-site distance of 300km (Rjb) are used to calculate the SoF-ratios for each region. Records with unknown SoF have been ignored. All stations are classified by measured average shear-wave velocities of the top 30m soil (V_s30). Firstly, the ground-motion amplitudes are scaled to reference rock site conditions (V_s30 = 760 m/s) with a nonlinear soil amplification model. The functional form of the predictive model also includes a trilinear-hinged magnitude scaling factor, magnitude dependent geometrical spreading and an anelastic attenuation term. The SoF-effects are computed for ground-motion amplitudes of reverse-to-strike-slip (R:SS) and normal-to-strike-slip ratios (N:SS). Since the main purpose is to show regional effects, a regression analysis for each mentioned database has been performed separately. Our results show that the R:SS are generally higher than unity for the NGA-West2 and ESM datasets, whereas for Japanese data, reverse earthquakes produce lower amplitudes with respect to strike-slip events. On the other hand estimations made the for NGA-West2 database show that N:SS are below unity and ratios of European and Japanese datasets are significantly higher. Finally, when we merge the datasets to one large database and calculate the ratios again, the resulting trends show major differences. Thus, flags were added for each region to recompute the SoF scaling coefficients.

Presenting Author: Elif Tuerker

Authors