Bayesian seismic hazard analysis

The Bayesian approach aims to obtain an inference utilizing indirect data as the prior information and direct data (or samples) as the observation. This paper presents a new Bayesian approach to estimate the annual rate of PGA exceedance (or SA: spectral acceleration), based on several sources of indirect data (e.g., ground motion prediction equations, seismic source models, and earthquake catalogs) and the directly recorded PGA data. In addition to the methodology, three case studies in Taiwan were also conducted. Like other Bayesian calculations, the case studies show that the Bayesian estimate is in between the estimate sorely based on the indirect data and the estimate sorely based on the PGA records (direct data), demonstrating the new Bayesian approach for estimating the annual rate of PGA exceedance (also known as seismic hazard) is fundamentally robust.