During storm events, huge amounts of sediments damage the storage and service life of Shihmen reservoir. The major portion of sediment yield was contributed from landslide caused by several relatively extreme storm events. The purpose of this study is to develop a landslide risk assessment model which enables the determination of landslide probability. Furthermore, by integrating the behavior for surface erosion and sediment delivery, the risk-based framework of sediment yield evaluation will be established. With the risk assessment model combining sediment yield evaluation framework, one should be able to outline the management priorities for the regions with different level of landslide potential.
Based on the infinite slope stability theory, the deterministic model is first established. The one-way sensitivity analysis is adopted to specify the uncertain parameters including hydraulic conductivity, cohesion, and friction angle. By integrating the uncertain parameters, the deterministic model can be excluded as the so-called stochastic model through the first-order variance estimation (FOVE) method. Both the deterministic and stochastic models enable the determination of safety factor and landslide probability for specified slope, respectively. Through the use of sediment delivery ratio (SDR) with the quantification of landslide and surface erosion by developed model and MUSLE empirical formula, respectively, the risk-based framework of sediment yield evaluation, then, is established.
The feature of the developed risk assessment model is based on mechanics with considering parameters uncertainties. Thus, the model can be applied to any kinds of watershed without limitations. The capabilities of developed risk assessment model are well demonstrated through application studies on Shihmen reservoir. In addition, the application results could be useful for engineers to propose the proper risk-based conservation strategies in Shihmen watershed.