| Summary: | 博士 === 國立成功大學 === 水利及海洋工程學系 === 103 === The removal of old dams is increasingly being considered; however, little progress has been made with regard to the modeling of geomorphic responses following the removal of such dams. This issue is particularly important in areas with strongly seasonal hydrology, high flow rates and sediment yields, and moderate to steep gradients. In this study, we sought to identify geomorphic responses in mountain channels following the removal of dams and understand the applicability of sediment transport models. Based on data obtained in the field, we investigated geomorphic responses following the removal of Chijiawan Check Dam and evaluated the applicability of NETSTARS (quasi-2D) in the case of the Barlin Dam as well as the SRH-1D ＆ SRH-2D models in the case of the Chijiawan Dam.
Following the removal of the Chijiawan Dam, we observed a rotating knickpoint with migration rates of up to 22 m/day along the reach. Following dam removal, changes in the channel of a mountain river appear to be driven as much by the timing of dam removal as by the morphology of the channel, the distance from the dam, and the degree of variability in the hydrology. Modeling of the failure of the Barlin Dam indicates that the model is more sensitive to active layer thickness and sediment transport function than to the number of stream tubes, and development of dam removal models are likely to benefit from varying the active layer thickness in time, In modeling the removal of the Chijiawan Dam, the lowest root mean square error values from the results of SRH-1D and SRH-2D modeling during low stream power events were 0.87 m and 1.18 m, respectively. SRH-1D failed to simulate any of the 17 m of lateral erosion that actually occurred 48 m upstream of the dam. SRH-2D in conjunction with a bank erosion model predicted lateral erosion; however, it underestimated the amount of erosion that occurred. Based on the results of this study, although 1D, quasi-2D, and 2D models simulate good fits for the observed changes in bed elevation, the models cannot accurately simulate the local feature. These findings underline the importance of field surveys.
|
|---|
