Accounting for the stochastic occurrence of landslides when predicting sediment yields

Landslides may lead to sedimentation of streams and rivers in tropical environments over time periods of decades. Hence, it is important to account for the stochastic occurrence of landslides when predicting sediment yields, particularly in areas undergoing deforestation. A hillslope-stability simulation model is developed based on applying an infinite slope stability analysis and the groundwater model of Lida (1984) to a population of synthetic landslide sites. The decline of apparent root cohesion over time due to vegetation removal is included, and the random occurrence of landslides is generated by a rainfall model. The model is used to predict landslides and sediment yields in mature and second-growth forests, and in clearcuts in a basin in Washington State, U.S.A. The model is best suited for large basins with hundreds to thousands of well-defined landslide sites.