The Inconsistent Flood-frequency Analysis Method Considering The Annual Jump Variation And The Seasonal Distribution Variation

Flood-frequency analysis is a key step in the process of design flood.And the design quantiles calculated in the analysis are the important indicators for design of hydro projects as well as the significant basis to determine the plan of the flood control and the protection against drought and the reservoir regulation.However,it is clear that the formation regimes of flood in many watersheds have changed in varying degrees due to the global climate and underlying surface change impact,resulting the loss of annual consistency in the flood series.Moreover,there are different causes of floods occurring in different periods in one year,which means that the flood series is inconsistent even within a year.To ensure the safety and economic plausibility of hydro projects,a flood-frequency analysis method in which both annual jump variations and seasonal distribution variations are considered about,is proposed in this thesis,aiming to improve the accuracy and rationality of the design quantiles.There are three parts in this method,including the ensemble method for detection of change points in hydrological time series,the hydrological staging method based on change point detection and the flood analysis method combining the decomposition synthesis theory and the additive rule of probability.In this thesis,the main rivers in Guangdong Province,China are taken as objects of study.Comparing with the conventional method and the inconsistent method of decomposition synthesis theory,the rationality and feasibility of the new method called the inconsistent flood-frequency analysis method considering the annual jump variation and the seasonal distribution variation(IFACAS)are verified.The main findings are as follows:(1)From the results of the Monte-Carlo statistical experiment,it is found that the efficiency of the 12 methods considered has obvious difference.And a more obviously jump variation means that the change point could more easily be detected.(2)The ensemble method for detection of change points performs well in the application of the hydrological season division and the annual variation diagnosis.And this ensemble method could solve the problem that the change points detected by a certain method are often not as reliable as expected and the results also vary with several methods used.(3)The hydrological staging method based on change point detection is used to divide the hydrological season of the main rivers in Guangdong Province.The results echo the previous conclusion that frontal rain and tropic cyclonic rainfall cause floods in the early flood season and the later flood season respectively.(4)The IFACAS processes data mutating annually with the theory of decomposition synthesis and data mutating seasonally with the additive rule of probability.In the example of Boluo Station which is the outlet control point of the East River,the result of the IFACAS not only fits better and have a better correlation to.the empirical points than that of both conventional one and the decomposition synthesis theory,but also could show the hydrologic characteristics caused by the change of seasons.