No abstract

No abstract

No abstract

No abstract

Many investigators have attempted to define the threshold of landslide failure, that is, the level of the selected climatic variable above which a rainfall‐induced landslide occurs. Intensity–duration (I–d) relationships are the most common type of empirical thresholds proposed in the literature for predicting landslide occurrence induced by rainfall. Recent studies propose the use of the kinetic power per unit volume of rainfall (J m−2 mm−1) to quantify the threshold of landslides induced by rainfall. In this paper, the relationship between rainfall duration and kinetic power corresponding to landslides triggered by rain was used to propose a new approach to define the threshold for predicting landslide occurrence. In particular, for the first time, a kinetic power per unit volume of rainfall–duration relationship is proposed for defining the minimum threshold needed for landslide failure. This new method can be applied using commonly used relationship for estimating the kinetic power per unit volume of rainfall and a new equation based on the measured raindrop size distribution. The applicability of this last method was tested using the data of rainfall intensity, duration and median volume diameter for 51 landslides in Taiwan. For the 51 landslides, the comparison between the measured pairs' kinetic power–duration and all selected relationships demonstrated that the equation based on the measured raindrop size distribution is the best method to define the landslide occurrence threshold, as it is both a process‐oriented approach and is characterized by the best statistical performance. This last method has also the advantage to allow the forecasting of landslide hazard before the end of the rainfall event, since the rainfall kinetic power threshold value can be exceeded for a time interval less than the event duration.