Dynamic Early Warning Mode of Flash Flood Disaster Based on Rainfall Process Identification
Aiming at the problem of"empty report"and"missed report"caused by ignoring the randomness and di-versity of rainfall based on the traditional design of rainfall patterns,this paper proposed a dynamic early warning model for flash floods based on rainfall process identification.Firstly,the WRF model was used to forecast the total rainfall,and the rainfall pattern database was constructed by using the K-means clustering analysis method.The similarity formula was set on the basis of"shape similarity"and"volume similarity"to realize the dynamic identification of rainfall processes under the condition of incomplete rainfall information.Secondly,the HEC-HMS model of small and medium-sized watersheds was constructed.The rainfall-runoff process was simulated based on the identified rainfall processes.The disaster flow in the control section of the disaster prevention object was compared with the simulated flood peak to is-sue the dynamic warning of flash flood.The Xinxian sub-basin in northern China was taken as an example.The results show that the simulated rainfall process by the method is more suitable for the actual rainfall.With the increase of infor-mation completeness,the NSE exceeded 0.8 in the third period,which improved the accuracy of early warning and effec-tively alleviated the problems of empty and missed warnings.At the same time,the early warning mode was issued 2 h before the end of rainfall,which prolongs forecast lead time compared with the traditional mode,which provides the theo-retical support for the prevention and relief of flash floods in hilly watersheds.
flash flood disasterrainfall patternrainfall identificationHEC-HMS modeldynamic early warning mode
万方数据
维普
