Ground-Motion Prediction Equations for Shallow-Crustal and Upper-Mantle Earthquakes in Japan Using Site Class and Segmented Geometric Attenuation Functions for the Horizontal Component

ABSTRACT We present a set of ground-motion prediction equations for the horizontal components of the strong-motion records from shallow-crustal and upper-mantle earthquakes in Japan. We used a reasonably large dataset from earthquakes with a moment magnitude (Mw) over 4.9 and a reliable earthquake category determined by the tectonic location of earthquakes up to the end of 2012. The modeling of source and site effects in this study is similar to that by Zhao, Zhou, et al. (2016); a two-segmented model for the magnitude scaling function hinged at Mw 7.1 and site-period-based site classes were used as the site effect proxy that accounts for the effect of the nonlinear soil response. We used two types of multisegmented geometric attenuation functions with the distance break points based on the distributions of within-event residuals versus source distance. We found that the geometric attenuation functions with three distance break points (70, 130, and 220 km) for the shallow-crustal earthquakes and with two distance break points (90 and 150 km) for the upper-mantle events improved the model performance significantly. The first two break points for the shallow-crustal events are similar to those for the attenuation of the Fourier spectrum due to Moho reflection, but the corresponding coefficients between these spectra differ significantly. For the upper-mantle events, the geometric attenuation coefficient is somewhat surprisingly close to −1.0 for the first segment, 0.0 for the second segment, and −0.5 for the third segment at many periods, because these values are close to the classic attenuation rates of the Fourier spectrum for shallow-crustal events but having different distance break points. Moho reflection is not usually present in the ground motions from deep events, and we can only provide a plausible explanation; a future study is warranted for a theoretical explanation.