A Study On The Ground Motion Prediction Equations Of H/V Response Spectral Ratio For Earthquakes In Subduction Zone

Ground-motion prediction equations (GMPEs) are an important component in earthquake engineering and probabilistic seismic hazard analyses (PSHA). But now, the investigations forced on the subduction zone and horizontal-to-vertical (H/V) response spectral ratio are still lack of the literature. On this account, the study has a relatively high value to the theoretical innovation and engineering application.The KiK-net and K-net station is the largest strong motion network and strong motion seismographs are installed at the ground surface and at the bottom of borehole. Also this two stations have gained thousands of strong motion records, which contain three earthquake categories, a range of magnitude, fault-top depth and source distance. This means the data used in the study is representative.Resorting to random effect model, GMPEs of three categories of H/V response spectral ratios for subduction zone is developed, consisting of three categories of shallow crustal and upper mantle, subduction interface and subduction slab. The source of residuals and their standard deviations are analyzed and compared with that of GMPEs of horizontal response spectra developed by another seismologist who used the same data. Based on those work, the predicted H/V response spectral ratios are given to analyse factors affecting ground motion.With the analyses of the residuals and deviations, it is concluded that the uncertainty of the GMPEs is mainly caused by site effect and source effect, merely affected the difference of earthquakes. According to the comparison with the GMPEs of horizontal response, the hypothesis that this GMPEs have smaller deviations and can improve the goodness of fit is verified. By the discuss of the predicted H/V response spectral ratio, the conclusion that the peak of H/V response spectral ratio is mainly affected by the site condition and has little connection with the magnitude and source distance. All of the conclusions could provide a reference for PSHA.