As a result of the geological weakness in Taiwan, the landslides often occur during heavy rainfall. Consequences of landslides can be serious silting in reservoirs, mudslides to destroy building and facilities, and escalating economic losses. Assessment on landslide potential in the watershed by heavy rainfalls has been successfully analyzed by the physical and numerical model. However, phenomena on landslides induced by the pressure in unsaturated porous media or the impact of rainfall-runoff process have not yet been appropriately predicated. Therefore, this study is to establish a two-dimensional, high-resolution model focused on rainfall-triggered unsaturated landslide in limited region with consideration on rainfall infiltration, surface runoff, and the variation of pore pressure in both saturated and unsaturated soil.
The annual report of this research project shows a one-dimensional model on rainfall-triggered unsaturated landslide established by employing Richards equation and infinite slope stability theory, and consequently the effects of rainfall pattern and the geological parameters have been discussed. With the application on uncertainty of parameters, the landslide potential around the preserved regions in Shihmen Reservoir is established by the first-order variance estimation (FOVE) method. Finally in this year, this study has preliminarily established the framework incorporated with the simulation on the mechanism of landslide and statistical analysis, for evaluating outcomes of the preservation in Shihmen watersheds. In the result from previous researches, Shihlei, Taiyao and Paling sub-catchment may possibly be major areas with high landslide potential around the preserved regions. In the evaluation on watersheds preservation, show that the reduction rate of annual landslide sediment yield on overall catchment may reach 2.38% after completing the first stage of preservation project. The estimated reduction rate of annual landslide sediment yield Yu Feng sub-catchment may probably reach 6.86%, which is the highest in evaluated regions.