Abstract
It is important to estimate the occurrence probability of deep-seated catastrophic (rapid) landslides (DCL) in order to plan measures to counter future disasters. We attempted to estimate the occurrence frequency at two sites using three methods. Firstly, we estimated the occurrence frequency based on a literature survey. Secondly, we estimated the occurrence frequency based on the result of the interpretation of multiple aerial photographs taken at different times to clarify the timing of DCL occurrence. Thirdly, we used a relationship between rainfall magnitude and DCL to evaluate the return period of rainfall-magnitude-induced DCL. This method also used the interpretation of multiple aerial photographs to clarify rainfall-magnitude-induced DCL. Here, we show that these three methods could estimate the occurrence probability of DCL, although the frequency based on the literature survey was underestimated by one order of magnitude. The rainfall magnitude and frequency relationship is effective for areas for which there is limited information about DCL occurrence.
