| Abstract |
| Distributed model is applied to actual rainfall in this study, then compared and analyzed, and the results are as follows; AGNPS model theory and algorithm are applied to actual rainfall event by constructing a number of mashes considering homogeneity for estimation of the spatially varying palrameters from the surface factors like soil characteristics, topography, plantation, and so on, and meteorological factors of the watershed. Dongkok, Koro, Misung, Byeongchun, Hyoryeong and Musung are constructed with 12, 30, 45, 76, 46 and 126 coarse grids respectively, and subdivided into 8, 24, 24, 44, 12 and 64 fine grids respectively for detailed channel direction and slope shape of the stream. Simulation result using AGNPS shows that the values of the peak flow have relative errors of 1.0~25.0%, 4.0~27.0%, 7.0~29.2%, 2.0~23.9%, 3.0~25.0%, 3.6~21.0% for each rainfall event at Dongkok, Koro, Misung, Byeongchun, Hyoryeong and Musung, respectively when the peak flows were compared with the measured values. Analysis result from AMC Ⅱ shows similarity to the measured, while the results from AMC Ⅰ or from AMC Ⅲ shows relatively small or great values, respectively. Modified curve numbers are reorganized to modify and complement the CN values according to AMC conditions suggested by the SCS so they fit to our circumstance, and the estimation result of the coefficient in the modified curve number method shows mostly coincident pattern with recently addressed research results. Estimation result by proposed modified method is almost similar to the measured. The relation to estimate CN for runoff computation by AGNPS is derived using antecedent precipitation, soil porosity, and topographic factor that is gradient of each cell, and the analysis result shows that there is the closest correlation between CN and the antecedent precipitation. |
|
|
| Key Words |
| AGNPS-UH model, antecedent precipitation, soil porosity, topographic facter |
|
|
|
|
|
PDF