Font Size: a A A

Study On Early Warning Mode Of Flash Flood Disaster By Coupling Rainfall Forecast And Rain Pattern Recognition

Posted on:2023-01-29Degree:MasterType:Thesis
Country:ChinaCandidate:S H ZhangFull Text:PDF
GTID:2531306620489054Subject:Water Resources and Hydropower Engineering
Abstract/Summary:PDF Full Text Request
With the technological innovation and the rapid development of society,the early warning of flash flood has higher requirements,and the forecast is developing in a more accurate and finer direction.In recent years,the rainfall numerical forecast has developed rapidly,but in the work of flash flood disaster early warning,the application of rainfall numerical forecast technology is less,and it still belongs to the exploration stage.And because of the low time resolution of rainfall numerical forecast,it can’t meet the time precision requirement of flash flood disaster warning in small watershed.In order to solve this problem,this paper studies the accuracy of different rainfall numerical forecast models on the basin rainfall forecast,and makes regression correction on the rainfall forecast values in combination with meteorological factors.On this basis,based on the cluster analysis theory,the characteristic rainfall patterns of the area are determined,and the rainfall process of each session is determined by combining the relevant principles of "shape similarity" and "quantity similarity" and some measured rainfall information.By establishing HEC-HMS model of small watershed,combining with rainfall forecast,the flood forecast of watershed was established,and the flash flood warning model was studied considering the danger degree of flood formation.Taking Xinxian,Tengzhou and Luo Fu Fort small watershed as examples,the application was demonstrated.The main research contents and achievements are as follows:(1)The optimization of rainfall numerical forecast based on forecast score and the study of rainfall forecast correction considering meteorological factors.The applicability of common rainfall numerical forecast models ECMWF,NCEP and JMA in three typical watersheds was evaluated by CIS and Bias.The results show that the maximum value of ECMWF ensemble forecast has the best applicability in different basins,and its forecast ability is better.On this basis,makes regression correction on the rainfall forecast values in combination with meteorological factors..Compared with before correction,the accuracy is higher and the forecast performance is more stable,and the error is reduced to less than 20%.It shows that the result of this correction method is good,and it is feasible to correct rainfall forecast by combining meteorological factors.(2)Analysis of watershed characteristic rainfall pattern based on cluster analysis principle and identification of rainfall process based on similarity principle.Based on the principle of cluster analysis,a rain pattern database of rainfall characteristics is constructed.At the same time,according to the principles of "shape similarity" and "quantity similarity",combined with rainfall numerical forecast technology and some measured rainfall information,a similarity formula is constructed to identify and correct rainfall events.According to some measured rainfall information,the whole rainfall process is estimated,and the rationality,reliability and applicability of this method are verified by comparing with the actual rainfall process.By using this method,the identified rainfall process can be assigned to the rainfall forecast values,and the time resolution of rainfall forecast can be improved to meet the time scale requirements of flash flood warning.At the same time,the rainfall process can be used in flash flood warning to improve the warning accuracy,and three typical small watersheds are taken as examples for calculation and analysis.(3)Study on the early warning mode of flash flood based on different warning levels.EW early warning indicators are put forward according to the different flood flows,and the early warning modes of flash flood with different warning levels are established.The HEC-HMS model is used to establish flood forecast and to carry out early warning test on selected flood events in the control sections of three small watersheds.The results show that this early warning model can effectively give early warning to flash flood caused by rainfall,and it has good applicability in the early warning of flash flood in small watersheds of hilly areas.
Keywords/Search Tags:Flash flood in small watershed, Rainfall numerical forecast, Determination of rainfall process, HEC-HMS hydrological model, Flash flood early warning mode
PDF Full Text Request
Related items