Comparison Of Flood Forecast Models For Semi-arid And Semi-humind Regions

Shanxi Province located in the semi-arid and semi-humind region of the north of China. The flood disaster occurred several times over the years. It is difficult to meet the needs of the current flood control situation and production practice of the flood forecasting results in this region. The valley of Shanxi Province as the basis, the semi-arid and semi-humid area of four hydrological models, focuses on the analysis of differences in four hydrological models structure, and the majority of Shanxi area in the basin a similar classification, for typical watershed model is chosen to give reference.The main work is as follows:(1) Analysis of flood forecast in semi-arid and semi-humind regions, combined with the principle structure of the four models and the difference of runoff yield and concentration are analyzed in detail with the Shuangchao hydrological model, Xinanjiang model, Shanbei model and Hebe Rainfall-flood model. For traditional optimization algorithm only according to flood a characteristic value for parameter optimization and calibration, this paper multi-objective optimization algorithm, taking the peak flow of qualified rate and peak time qualified rate as the index of model accuracy, flood forecast in the valley of Shanxi Province for technical support.(2) According to the characteristics of small watershed in Shanxi area, extraction physical characteristics of all small watershed, with the self-organizing maps neural network, combined with the relevant physical parameters in the areas for similar classification, divided into 13 categories, selects the four typical basin used to simulate flood forecasting to solve the problem of lack of information.(3) The four model applied to the selected typical watershed of simulation analysis, Taohe River Basin, Honganjian basin, Jiang Basin using Shuangchao hydrological model effect best, Beishi Basin using Hebe Rainfall-flood model effect best.