Study Of Flood Forecasting Methods At Small Watershed And Application

Due to the fact that the flood forecasting characteristics of small watersheds are varied in mountain area, and the measured hydro logical data is relatively scarce, it is difficult to establish the flood forecasting scheme with the impact of human activities. Currently, although there are lots of flood forecasting models, it's not the fact that every model is applicable to an arbitrary small watershed. So in practical application, the summary in applicable conditions of various hydrological models, and how to select the forecasting model according to characteristics of small basins are vital parts of the flood forecasting.To solve these problems, firstly, the paper briefly analyzed the characteristics and application conditions of several domestic flood forecasting models. And Bingyugou watershed and Dengshahe watershed in northern semi-arid and semi humid area were chosen as case studies, the paper put forward the forecasting models of each basin. Then, historical flood simulation study of the two small watersheds were carried out based on the selected models. And the simulation results were compared to analyze the applicability of different models in the two small watersheds. At last, the flood forecasting models suitable for small watershed were analysed, and the rule of model's parameters at small watersheds was summarized. Through the study results, the paper can provide kinds of ideas for the flood forecasting of other small watersheds similar to hydrological and meteorological conditions, and provide certain reference basis for parameters value of other small basin with missing data. The main results of this paper are as follows:(1) The importance of study on flood forecasting methods of small watershed was demonstrated by analysis of the characteristics and difficulties of small basin in flood forecasting and the research status of flood forecasting methods, and the necessity of the applicability analysis of forecasting models was clarified. Then, the flood forecasting models of Bingyugou watershed and Dengshahe watershed were selected respectively.(2) In the case of Bingyugou watershed in semi-humid and semi-arid area, DHF model, XAJ model. TOPMODEL and HEC-HMS model were selected to simulate the seven historical floods. The simulation results suggested that, the simulated runoff pass rate of four models was 100%, and the results of simulated peak discharge and peak time were qualified but "20080731" flood. So there was no obvious difference between the four models, and all of them had good applicability in Bingyugou watershed. But in contrast, the model based on saturation runoff generation method simulated the peak time a little poorly. So if we can get more of the observed data, infiltration excess runoff generation model is applied or saturation runoff generation model is combined to establish flood forecasting scheme. But the HEC-HMS model is of great promotive value in the basins with missing data.(3) Compared to Bingyugou watershd, Dengshahe watershed is easy to occur infiltration excess runoff. DHF model and HEC-HMS model were selected to carry out the simulation study on 11 historical floods. In addition, the simulation result of XAJ model was to test the rationality of the selected model. The results showed that, the simulated runoff pass rate of DHF model and HEC-HMS model were 100%, and the pass rate of simulated peak discharge and peak time were 81.8%, and the simulation effect of Nash-Sutcliffe coefficient was good. So the results can be used to guide the flood forecast and early warning of small watersheds. But the mean error of the peak flow in DHF model was smaller than that of HEC-HMS model, so in practical work, the parameter values of HEC-HMS model should be seriously considered. While the XAJ model had low applicability at Dengshahe small watershed. Meanwhile, it showed that the selected forecasting model of the watershed is reasonable.(4) In view of the integrated simulation results of the two small watersheds, for the small watershed in the semi-humid and semi-arid area of the northeast region, the models based on the method of infiltration-excess runoff were more suitable for flood forecasting at small watershed. The rules of parameter values in DHF model were:Generally, for the small watershed similar to hydrological and meteorological conditions, the evapotranspiration parameters of every month varied between 3.4 and 5.5, and the values in May, June and August were large. The parameter K was between 0.4 and 0.7, and the parameter N was between 3.0 and 5.0; however, the parameter N was greater than 4.0 if the vegetation was well. For the small watershed with good vegetation, the parameter SO was greater than 40; but SO was less than 40 for the watershed with poor vegetation. In the case of dry soil for most of the time, the parameter U0 was greater than 90, but U0 was less than 90 if the soil was wet.