Effect of Antecedent Precipitation for Rainfall-Induced Landslide Warning System

Rainfall is widely recognized as an important factor in the landslide disaster. Therefore, rainfall-induced warning systems are the most commonly used landslide warning systems and has been established for many years. In the rainfall-based landslide warning system, most researchers used the empirical methods to consider the effect of antecedent rainfall and antecedent soil moisture for occurrence of the landslide, named the Antecedent Precipitation Index (API). In Taiwan, we use rainfall intensity product effective accumulated rainfall for landslide warning. As effective accumulated rainfall, we consider antecedent rainfall and use the same decay constant in Taiwan. However, for the real cases, the decay constant should base on the different soil type. In this study, a rainfall infiltration model was developed based on the one-dimensional Richards equation and combined with the infinite slope stability analysis. By the model, the time-dependent change of pore water pressure and volumetric water content can be calculated and can let us know the relationship between the antecedent moisture and the antecedent rainfall. Furthermore, with the different hydraulic conductivity, when the deterministic model gives the landslide warning, the effective accumulated rainfall should be equal to the critical rainfall value. Based on this concept, we can find out the best decay constant for the Antecedent Precipitation Index in Shenmu village, Taiwan.