Established Vs30 Grids Data for Taiwan Areas by Weighted Kriging

Established Vs30 Grids Data for Taiwan Areas by Weighted Kriging

Although the Taiwan strong motion instrumentation program (TSMIP) has compiled many reliable site-effect parameters in the Engineering Geological Database for TSMIP (EGDT), the drilling survey stations are not uniformly distributed. Since resolution is not high enough to represent characteristics of local areas, this study attempted to establish Vs30 grid data for the Taiwan area using weighted kriging. Weighted kriging considers not only distance factors, but also confidence of different Vs30 methods. It can control grid values to yield higher confidence values. In this study case, we chose data points where distance was less than 6 kilometers for interpolation purposes. We also evaluated interpolation results five times using the cross validation method. First, we randomly removed 5% of samples from each Vs30 method by program code, then re-interpolated the data. Finally, we analyzed errors between sampled and interpolated results. Our statistical analysis showed high reliability in plain areas and large errors in mountainous areas or areas with rapidly altered terrain. Obtaining more Vs30 data in mountainous areas in the future will allow further control points to improve reliability and accuracy for ground motion predictions and disaster applications.