| Summary: | 碩士 === 國立中央大學 === 應用地質研究所 === 103 === Regional landslide susceptibility is commonly predicted via a statistical approach based on multi-temporal landslide inventories and environmental factors. If multi-temporal landslide inventories are not available, we could use an event-based landslide inventory and triggering factors to build an event-based susceptibility model. If we extract the triggering factor from the model, then the model can represent the basic susceptibility of the region, we call it event-independent susceptibility model. It is a key issue to find out whether different event-independent susceptibility models for the same region are similar in pattern.
In the present study, we chose four event-landslide inventories to build four single-event susceptibility models and one combine event model. It is confirmed that different event-independent susceptibility models for the same region are similar in pattern. This basic landslide susceptibility map is also similar with the susceptibility map build by multi-temporal landslide inventories. After establishing a regional basic landslide susceptibility model, we could analysis the relationship among probability of failure, landslide susceptibility and rainfall.
|
|---|
