Based on our previous study, we developed ground motion prediction equations (GMPEs) with two site proxies, that is, with one GMPE using the site period (TS, four times the travel time of shear waves down to the bedrock) and the other one using a pseudosite period (TVS30, four times the travel time of shear waves down to 30 m depth). We used two reasonably simple site-effect models based on TS or TVS30, and each site model has four parameters with each parameter having a physical meaning and miming the theoretical process in site amplification calculations. We found that only the linear magnitude terms in the site class (SC) model from our previous study need to be modified to accommodate the change in the site response parameters, suggesting that the source effect interacts with the site effect. At short periods up to about 0.3 s, the performances of these site models are largely similar. However, at spectral periods over 0.3 s, the performances of these site models are strikingly different, and the rank of the three models in the site model performance is (from the best to the worst) TS, SC, and TVS30 site models. It is a surprise that the SC model with four classes can be better than the TVS30 site model, which does not account for the effect of the soil or rock layers between the 30 m and the bedrock depth. These results may mean that the inclusion of these deep soil layers is more important than the limitations of using a simple step function for the SC model. The GMPE using TS presented in this study improved the prediction accuracy significantly, especially for sites with TS over 0.3 s at spectral periods over 0.3 s. The GMPE using TS in this study can be used whenever TS is available.

No abstract

A large dataset from earthquakes up to the end of 2012 with a reliable earthquake tectonic category was used to develop a ground motion prediction equation (GMPE) for the vertical strong motions from the subduction interface earthquakes in Japan. Guided by the increase in the maximum log likelihood of the regression model, a minor extension on the maximum usable period was applied so that more records can be used at long spectral periods. A bilinear magnitude scaling function hinged at a moment magnitude (Mw) of 7.1 was adopted, and the scaling rate for events with Mw>7.1 is less than that for the other events. Site period-based site classes were used as the site parameter without a nonlinear site term. We used two datasets with the first dataset including strong-motion records from large events at stations with inferred site classes, and these records were removed in the second dataset so that all sites have a measured shear wave velocity profile down to the engineering bedrock. We found that excluding these records from sites with inferred site classes improved the model fit significantly. We modeled the effect of volcanic zones using the slant distance passing a set of volcanic zones. We also found that the between-site standard deviations for the vertical component in this study were generally less than those for the horizontal component at nearly all spectral periods. The total model standard deviations from this study are generally less than or similar to those from the GMPEs for the vertical ground motions for other earthquake categories. The distributions of between-site residuals with site period or TVS30 (120/VS30 in m/s) are not significantly biased, suggesting that the site term coefficients can be interpolated between that site class terms if a continuous predicted spectrum with site period or TVS30 is required.