| Summary: | 碩士 === 國立臺灣海洋大學 === 河海工程學系 === 93 === Digital elevation model has received large attention because of its convenience of obtaining watershed geomorphic factors for watershed hydrological simulation. Although high-resolution digital elevation data set can provide detail description of the watershed topographic characteristics, it nevertheless requires more computation time for hydrological simulation. On contrary, low-resolution data set only requires less time on computing, but it usually results in bias simulations because of lacking watershed detail topographic information. Most previous researches were concentrated on the change of the basin area or the variation of the model parameter’s value due to data resolution changes. The objective of this study is to investigate the difference of watershed boundary and runoff simulation results while the resolution of the digital elevation data set is reduced.
The raster grid resolution used in this study are 40m×40m, 80m×80m, 120m×120m, and 160m×160m. The Analytical results show that the extracted stream network is changed and the part of the watershed area is “missed” in most of the cases while the resolution is reduced, and the ratio of the missed watershed area to the watershed total area can maximally reach to about 11%. To avoid the area-missed situation, a method is proposed in this study by only choosing different left-corner point for extraction, and then the accurate stream network can be extracted for further geomorphic factors calculation. A distributed non-inertia wave model is developed to investigate the influence of data resolution change on runoff simulation and model parameter calibration. The results indicate that good simulations can be obtained by choosing appropriate roughness coefficients.
|
|---|
