Implications of the 2011 M9.0 Tohoku Japan earthquake for the treatment of site effects in large earthquakes

Abstract

Site response in Japan is characterized using thousands of surface and borehole recordings from events of moment magnitude $$(\mathbf{M}) > 5.5$$ collected by the KiK-net network, including the 2011 M9.0 Tohoku earthquake. Site amplification is defined by the ratio of motions at the surface to those at depth (within the borehole), corrected for the depth effect due to destructive interference using a technique based on cross-spectral ratios between surface and down-hole motions. Site effects were particularly strong at high frequencies, despite the expectation that high-frequency response may be damped by nonlinear effects. In part, the large amplitudes at high frequencies are due to the prevalence of shallow soil conditions in Japan. We searched for typical symptoms for soil nonlinearity, such as a decrease in the predominant frequency and/or amplification, using spectral ratios of weak to strong ground motions. Localized nonlinearity occurred at some recording sites, but was not pervasive. We developed a general empirical model to express site amplification for the KiK-net sites as a function of common site variables, such as the average shear-wave velocity in the uppermost 30 m ( $$\text{ V}_\mathrm{S30})$$ and the horizontal-to-vertical (H/V) spectral ratio. We use the model to estimate site-corrected ground-motions for the Tohoku mainshock for a reference site condition; these motions are in reasonable agreement with the predictions of some of the published ground motion prediction equations for subduction zones.