Relationship among Probability of Failure, Landslide Susceptibility Index, and Rainfall Parameters

Abstract:

Common or basic susceptibility may exist in a region. This common feature has been tested to be true in several drainage basins in Taiwan in recent years. A multi-temporal landslide inventory was commonly used in building a landslide susceptibility model for a region. An event landslide inventory and triggering factors were also used to build an event landslide susceptibility model in recent years. An event-based landslide susceptibility model is dependent on that event, and different event models exhibit different susceptibility pattern. However, if we extract the triggering factors from the event model, then the model would be event-independent and can represent the basic susceptibility of the region. We found that different event-independent susceptibility model for the same region are similar in pattern, and is similar also to a susceptibility model built by a multi-temporal landslide inventory at that region. On this basis, we can use an event-independent susceptibility model to represent the basic susceptibility of the region with confidence. We tested the relationship between probability of failure and rainfall intensity, as well as total rainfall at each basic susceptibility bin. It was found that the relation is good; the probability of failure increases with an increase in the rainfalls and also an increase in the susceptibility. A fitting surface of probability of landslide failure using a rainfall parameter and the basic susceptibility as two independent variables was done and the example will be presented.