Research On Flash Flood Warning Technique Coupled With Rainfall Ensemble Forecast

Flash flood is a highly destructive disaster, and it has brought serious harm. The complex causes of flash flood make it very difficult to warn it. Flash flood warning, as an effective non-engineering measure, has played an important role in the prevention of flash flood. This paper take Yingnahe basin as the studied basin, fiestly, analyze the availability of rainfall ensemble forecast information in the field of hydrology; secondly, explore the flash flood warning method of Yingnahe basin and extend the method to small basin; lastly, research the coupling method of flash flood warning and rainfall ensemble forecast in order to prolong the forecast period of flash flood warning. The main research contents and results are as follows:(1) Analyzing the availability of rainfall ensemble forecast information. According to the classification standard of the meteorological department, grading the 0-6h,0-12h and 0-24h rainfall forecast information and analyzing the accuracy of different forecast rainfall grade. It turns out that the accuracy of no rain and light rain forecast are higher, which can provide reliable information for flash flood warning; Forecast rainfall at the grade of heavy rain or above has a tendency to be smaller, and its accuracy needs to be improved.(2) Taking Yingnahe basin as the studied basin to explore the flash flood warning method of medium basin. This paper proposes a flash flood warning method which based on hydrological model to calculate the critical precipitation. This method consider four factors (antecedent precipitation, accumulated precipitation, rainfall intensity and rainfall distribution) which had common influence on flash flood and establish relationships between critical precipitation and every factor for three time-scales(1h,3h and 6h). The method was applied to 18 floods of Yingnahe basin, result shows that all the correct rate of three time-scales are 94.4%, which demonstrats that early warning effect is very good and the method is reasonable.(3) Taking Yingnahe basin as the studied basin to analyze the applicability of the flash flood warning method based on hydrological model in small basin. The dynamic critical precipitation curves are established by using API model and Xin'anjiang model respectively and applied to 8 floods, result indicates that the correct rates of flash flood warning based on two hydrological models all is 100%, which demonstrats that the method can be applied in samll basin and the dynamic critical precipitation curve can be established by different hydrological models.(4) Researching the coupling method of flash flood warning and rainfall ensemble forecast. Selecting two typical floods of Yingnahe basin to exlore the coupling method. Comparing the critical precipitation at each moment with the forecast rainfall to judge the possibility of flash flood disadter, but the result is not ideal. The coupling method is revised after analyzing the reasons, result shows that the revised coupling method can warn flash flood timely and accurately and prolong the forecast period of flash flood warning.In the end of this paper, summarizing the research contents and point out the problem in the study and the work to be improved.