| The probability analysis and calculation of hydrological events is an important non engineering means of disaster prevention and mitigation in the field of water conservancy engineering,and the reliability and effectiveness of the analysis and calculation results depend on the number of hydrological observation data and the selection of methods.There are often multiple hydrological regions in the same basin.Due to the similarity and correlation of climate and terrain,the joint probability of hydrological events in multiple hydrological regions can be established and simulated randomly by analyzing the similarity of multiple hydrological regions and using the statistical correlation of hydrological data in adjacent regions to increase the sample size.In this paper,the Copula function is used to construct the joint distribution of monthly rainfall among stations,and the monthly rainfall for many years is generated by random simulation.Through the analysis and research of the simulation value,it provides a reference for rainfall correlation analysis,probability simulation and long-term series research of multiple hydrological intervals.This paper takes the monthly rainfall data of Suqian sluice station(upstream),Liulaojian station(midstream)and Xinyuan sluice station(downstream)in the Ancient Yellow River Basin of Suqian City,Jiangsu Province as the research object.Firstly,the P-? distribution,Gamma distribution,Exponential distribution and GEV distribution are selected as the candidate edge distribution functions,and the intuitionistic graphic method and root mean square error method are used for optimization;Then,three kinds of Archimedean Copula Functions are constructed as the alternative joint distribution functions,and the optimization is carried out by graphic calculation and analysis method and AIC information criterion;Finally,the two-dimensional Copula function combined with Gibbs sampling method is used to realize the random simulation of monthly rainfall of three stations,and the three-dimensional Copula function combined with Gibbs sampling method is used to simulate the monthly rainfall of three stations.Based on the mean,mean square error and coefficient of variation of the simulated and measured values of the three stations obtained by the above two methods,the relative errors of the three statistical characteristics are analyzed to verify the effectiveness of the stochastic simulation method in rainfall correlation analysis,probability simulation and long-term series research of multiple hydrological intervals.Through analysis and calculation,the optimal marginal distribution of monthly rainfall of the three stations is mostly P-? distribution function.The simulation results based on two-dimensional and three-dimensional Copula function are affected by the Copula function type,rainfall correlation between stations,rainfall prominent points and the fitting of rainfall and edge distribution.By comparing the simulated value with the measured value,it is found that the two methods can achieve relatively high accuracy,and generally the simulated value can keep the statistical characteristics close to the measured value.Among them,the error of the method based on two-dimensional Copula function is less than 10%,which has a good simulation effect.Although the three-dimensional Copula function based stochastic simulation method of monthly rainfall between three stations can achieve better simulation effect,the overall error is less than 10%,but the error is larger than the two-dimensional Copula function.The results show that Copula function has a good effect in the stochastic simulation of monthly rainfall among multi stations in Suqian Ancient Yellow River Basin,and basically retains the original statistical characteristics of rainfall events,which can provide reference for rainfall correlation analysis,probability simulation and long-term sequence research of multi hydrological intervals in Suqian Ancient Yellow River Basin. |
PDF Full Text Request