Research And Application Of Flood Forecasting And Warning Method Considering Rainfall Uncertainty

Flood forecasting and warning work is an important part of non-engineering measures for flood control.Early and effective forecasting and warning is of great significance for reservoir operation and flash flood warning.The sources of uncertainty in the flood process and the causes of mountain torrents disaster are extremely complicated,which exacerbates the difficulty of flood forecasting and warning work.In response to this problem,this article takes the basin above the Yingnahe reservoir as the research background,firstly,based on the improved DHF model,re-rate the model parameters and formulate flood forecasting plan;Secondly,analyze and quantify the impact of rainfall uncertainty on flood forecasting,explore the flood forecasting method that takes into account the rainfall uncertainty;Finally,apply the flood forecasting method considering rainfall uncertainty to the basin flash flood warning.The main research contents and results are as follows:(1)For the current situation of poor application of Yingnahe reservoir forecasting program based on the original DHF Model,by improving the DHF model and classifying the degree of pre-soil moisture to re-study the Yingnahe reservoir flood forecasting program.The results show that compared with the original flood forecasting scheme,the overall forecast accuracy of the newly formulated flood forecasting scheme has been improved,but the forecasting accuracy of partial floods is still relatively low.(2)In response to the relatively low accuracy of partial floods'forecasting,based on the new flood forecasting method,using fuzzy set theory and Monte-Carlo method,considering the impact of rainfall centers and cumulative rainfall,to quantify and analyze the impact of uncertainty,such as magnitude,time-allocation and spatial distribution of rainfall,on flood forecasting.The results show that the uncertainty of the rainfall level plays a major role in the impact of the flood process,and the uncertainty in the time-course distribution and spatial distribution of the rainfall is second;The median flow rate process is less affected by the spread of rainfall uncertainty,basically consistent with the actual flow process,can be used for flood forecasting.(3)Based on the results of the rainfall uncertainty analysis,take various estimated rainfall processes as improved DHF model input conditions,to simulate floods of Yingnahe reservoir.According to the conditions set by the model,take median of the flow rate process that meet the requirements,and then propose a flood forecasting method considering the uncertainty of rainfall.This method was applied to the"20170804"flood simulation of Yingnahe reservoir,the simulated peak height increased from 2925.7m~3/s to 5020.6m~3/s,and the peak time changed from unqualified to qualified.Meanwhile,this method was also applied to flood simulation in the Gengwangzhuang river basin,and the simulation accuracy was improved compared with the original scheme.It can be seen that this method has good applicability and can be applied to flood forecasting in other river basins.(4)The critical rainfall is deduced based on the flood forecasting method that takes into account the uncertainty of rainfall,and the grey correlation coefficient method is used to quantify the impact of pre-soil moisture content,cumulative rainfall and rainfall intensity on the critical rainfall,and then formulate dynamic critical rainfall curve of flash flood warning based on flood forecasting method which considering rainfall uncertainty.The results show that the correlation between critical rainfall and the pre-soil moisture content,cumulative rainfall and rainfall intensity is significant,and there is a nonlinear relationship;this method is better than the warning method without considering rainfall uncertainty.Finally,summarized the full text and pointed out the work that needs to be resolved and improved.