科研之友
微信
新浪微博
Facebook
分享链接
摘要
Understanding the variation of seismic site response plays a crucial role in mitigating a future earthquake disaster. The spatial variation of site response is caused primarily due to varying amplification of the seismic wave by the ground surface layers. In practice, the spatial variation of site response is projected on maps. Site responses are correlated with some parameters and the variation of these parameters are mapped in advance of preparing the site response map. However, in conventional mapping, the observations recorded over events at a site are generally reduced to a single averaged value. Although the site responses at two sites with some difference in their average value are visually considered to be different, the statistical significance of this difference is directly ungraspable without any information on the distribution of the data. This study addresses the issue of site response uncertainty in mapping. A mapping methodology termed as uncertainty projected mapping (UPM) has been proposed which reflects the data insignificance on the map resolutions, based on the uncertainty. In UPM, a hierarchical Bayesian model has been adopted for the spatial distribution of site responses as it incorporates uncertainties and allows modelling variabilities as parameters. In this paper, the theory of UPM is introduced at first. The UPM methodology is then validated with certain numerical experiments. The validated UPM is then applied to make a site response map for a case-study area in Furukawa, Japan. The results are then discussed in comparison to a conventional mapping method.
- 出版日期2018-9
- 单位防灾科技学院
