Abstract
A probabilistic fault displacement hazard analysis(PFDHA) is a methodology that assesses an annual rate/probability of exceedance that an amount of displacement of a surface earthquake fault exceeds a certain quantity. To perform the analysis, displacements due to a principal fault and distributed faults are indispensable to evaluate in a probabilistic manner. For principal faults, an abundance of data are accumulated by making use of the past surface earthquake faults in Japan, while for distributed faults, the data are not necessarily sufficient in reality.Although a past study shows that an occurrence probability of distributed fault depends on a grid size employed in the analysis, no study using the Japanese domestic data for distributed faults has been reported.Taking this into account, the present study tried to improve the reliability of evaluation formula by analyzing the dependency of the grid size. In addition, model experiments and numerical analyses based on the discrete element method were applied to compensate the lack of data of the distributed faults. Furthermore, we adopted a logic tree methodology that can consider epistemic uncertainties and demonstrated its application example.
