
Seismic Hazard Assessment for Mainland China Based on Spatially Smoothed Seismicity and Projected Sets of Ground Motion Models
Abstract
The latest version of the Chinese seismic hazard map (CSHM) (i.e., fifth-generation CSHM) was developed based on the delineated seismic source model, which is defined based on geological and seismological information but also with some subjectivities. The results of the CSHM were presented in terms of peak ground acceleration (PGA) and peak ground velocity. No maps for spectral acceleration (SA) are given. This is partly due to the ground motion models (GMMs) to predict SA for mainland China are unavailable. The unavailability of the GMMs for SA results in the lack of uniform hazard spectra (UHS) that is important for structural seismic design.
There is a gap in the development of the new GMMs and seismicity models for Chinese seismic hazard mapping. This thesis is focused on the evaluation of the seismic hazard and the development of the uniform hazard spectra for mainland China. For the evaluation, a set of GMMs applicable to mainland China is developed by applying the so-called projection method. This method projects the GMMs developed for a reference region with a large number of historical records to a target region where the actual ground motions are scarce. For the projection, the NGA-West2 GMMs developed for California are considered for the reference region, and different regions in mainland China are considered as the target regions.
Rather than using the delineated seismic source model, smoothed seismic source regions based on the historical catalogue and spatial smoothing techniques are considered. Moreover, an analysis is carried out to assess the completeness of the historical Chinese earthquake catalogue. Two smoothed seismic hazard models for mainland China are obtained. The first one is based on cluster analysis and spatial smoothing by considering that a seismic magnitude-recurrence is applicable to a cluster, so the smoothing is carried out for the annual earthquake occurrence rate. The second one considers that the magnitude-recurrence relation is spatially varying, which is obtained by carrying out the smoothing by considering the earthquake magnitude (i.e., the smoothing the cumulative event count as a function of magnitude).
The CSHMs for mainland China are assessed using the newly projected GMMs, which could be used to predict the PGA and SA and the spatially smoothed seismicity models. A parametric investigation is carried out by considering different combinations of GMMs, magnitude-recurrence relations, and smoothed source models. The logic tree approach is used to represent the combinations. The newly developed CSHMs are presented in terms of PGA and SA. In addition, the uniform hazard spectra are developed for different locations within mainland China. A comparison of the newly developed seismic hazard maps to that of the fifth-generation CSHM indicates that they exhibit similar trends, although there are differences in the estimated return period values of PGA. A comparison of the normalized UHS to the standardized design spectrum in Chinese design codes is presented, indicating that the standardized design spectrum is conservative for short and long natural vibration periods.