Study On Uncertainty Of Critical Rainfall Of Flash Flood Based On Random Rainfall Pattern

At present,in the early warning and forecast of mountain torrents in China,the critical rainfall is the most commonly used early warning index.When determining the critical rainfall,only the regional design rainfall pattern is used for rainfall time distribution,ignoring the uncertainty of critical rainfall caused by the randomness of rainfall.With the change of natural and social factors,critical rainfall is not the only value.however,at present,fixed values are still used to issue early warning information in most early warning work of mountain torrents,which reduces the accuracy of early warning.In view of the uncertainty of critical rainfall caused by the randomness of rainfall,this paper deeply explores the structural characteristics and distribution law of rainfall time history distribution,and proposes a stochastic rain pattern generation method based on parameter control.Based on the random rain pattern,the critical rainfall threshold space set of typical disaster prevention objects is determined.In addition,a critical rainfall risk early warning model based on random rainfall pattern is proposed.through the similarity identification of rainfall series,the possible critical rainfall is estimated and the corresponding early warning information is issued.It not only provides a reference value for improving the accuracy of critical rainfall early warning of mountain torrents,but also provides a new way for the early warning model of critical rainfall of mountain torrents.The main contents and research results of this paper are as follows:(1)Uncertainty analysis of critical rainfall.Firstly,we analyzed the impact of different rainfall patterns on the critical rainfall,and then analyzed the impact of different early impact rainfalls on the critical rainfall.As a result,it was found that the impact of rain patterns on the critical rainfall was huge.It should be considered in determining the critical rainfall.(2)The method of determining random rain pattern based on parameter control.According to the regional rainfall characteristics,the rainfall pattern information matrix is established,and the optimal distribution function of rainfall time history distribution ratio in each period is selected.With the boundary value as the constraint condition,the random rain pattern set in the study area is selected through the rain peak position coefficient and peak multiple ratio,and the 6 h critical rainfall threshold space of typical disaster prevention objects is calculated.the results show that the random rainfall pattern has good application in solving the uncertainty of critical rainfall.Finally,we further analyzed the relationship between the characteristic parameters of the rain pattern and the critical rainfall is,and quantified the magnitude of the influence between them.(3)The risk early warning model of critical rainfall based on random rainfall.Through the random rain pattern generation area,the corresponding critical rainfall threshold space is calculated,and the risk zoning of the critical rainfall is carried out according to the characteristic rainfall in the disaster period.At the same time,according to the similarity of the real-time rainfall information,the critical rainfall is called to release the early warning information.The application results in Xiawan formation show the superiority of this early warning model,and provide reference value for perfecting the critical rainfall early warning and forecasting system of mountain torrent disaster.