General Description for the Shihmen Reservoir
Shihmen reservoir is located in Taoyuan and Hsinchu Counties, the design and initial effective storage capacities (May, 1963) were 3.0 x 108 and 2.4 x 108 m3, respectively. The contribution catchment of the Shihmen reservoir is 763 km2 and the average annual rainfall is 2,350 mm which is concentrated in summer. Due to the heavy rainfall and vulnerable geological conditions, landslide activities occur frequently in the Shihmen watershed and produce massive amount of soil sediments. Subsequently, it decreases the reservoir storage capacity and threatens the water supply capacity. For example, the 2004 Typhoon Aere flushed 2.7 x 107 m3 of soil into the Shihmen reservoir. Moreover, it also shut down the water supply system with a period of 18 days due to the drastically rise in raw water turbidity. In order to assist the government departments to establish the suitable sustainable strategies for the Shihmen reservoir, such as watershed conservation, reservoir desilting, and water supply system improvement, the DPWE builds a cross-disciplinary team and focuses on the three most urgent issues: landslide, sediment transportation, and water supply system.
Simulation of Reservoir Sediment Transport
In order to provide useful information for devising suitable desilting plans, the horizontal 2-D hydraulic and sediment model (RESED2D) is adopted to simulate the behavior of sediment transportation in the Shihmen reservoir. The simulation results demonstrates that cohesive sediment with the particle size < 0.0625 mm mostly deposits around the dam. Moreover, the vortexes have been found in several regions, which might reduce the desilting efficiency.
Reservoir Deposition Risk Analysis
With the integration of the three approaches: (1) watershed landslide risk analysis, (2) in-situ sediment concentration monitoring, and (3) hydraulic simulation, the DPWE evaluates the reservoir deposition risk associated with the probability density function for annual deposition volume. Results show that the expected annual deposition volume is 3.2 x 106 m3. Moreover, if the efficiency of sediment flushing could be improved to 40%, the annual deposition volume will be reduced to 2.27 x 106 m3.
Probabilistic Analysis of Landslide Potential
Based on the 1-D Richard's equation and infinite slope theorem, the DPWE develops a rainfall-triggered shallow landslide model. Furthermore, incorporating the uncertainties of hydrological and geological properties, the Rosenblueth's point estimation and Monte Carlo simulation are adopted to analyze the landslide risk map for Shihmen reservoir watershed under various rainfall scenarios. The risk analysis results show that the areas with higher landslide probabilities have a dramatic increase if the total rainfall exceeds 600 mm, the expected landslide volume approximates 11 x 106 - 14 x 106 m3, which indicates that the landslide and mudflow disasters in the Shihmen reservoir watershed are mostly triggered by extreme rainfall events.
Sediment Monitor System
In order to control the behavior of sediment transportation and deposition in the Shihmen reservoir during flood events, an extensive automatic suspended sediment concentration monitoring program is established at the reservoir inflow and outflow locations. In addition, the monitoring station at the Cross Section 24 has a unique feature that could provide the multi-depth water temperature and sediment concentration measurements. Based on the monitored data during several flood events, the efficiency of sediment flushing for the Shihmen reservoir approximates 13% - 24%.
Water Shortage Risk Analysis
During extreme flood events, the turbidity of raw water in the Shihmen reservoir drastically rise to tens of thousands units which far exceeded the treatment capacities of local water treatment plants. According to the monitored hydrological data and the water supply system in Taoyuan County, the DPWE incorporates the uncertainties for sub-components of water supply system to analyze the water shortage risk due to high raw water turbidity. The analysis results show that the shortage risk could be reduced to below 5% under the installation of additional pumping facilities for drawing the reservoir water at higher levels. Moreover, the shortage probability could be reduced if the backup water supply is planned.