The Application Of Bivariate Compound Extreme Distribution In Hydrological And Disaster Prevention Engineering

It is one of the important problems to design the flood-prevention water level of the estuarine cities. Because of the special geographic location of the estuarine cities, they are affected by river ocean and meteorology environmental loads and so on. So how to determine the flood-prevention water level in the area is a question to need to study into it.This paper proposes a method to estimate the hydrological disaster prevention design criteria for the estuarine city Shanghai, the biggest estuarine city in China. Considering the combined effect of typhoon storm surge, flood peak from upper river Yangtze and astronomical spring tide, one type of the Bivariate Compound Extreme Distribution: the Poisson-Gumbel Mixed Compound Extreme Distribution is adopted to predict the joint probability of storm surge and flood in the condition of spring tide. The stochastic simulation technique based on the Importance Sampling Procedure is also used to predict joint probability of storm surge, flood and spring tide. In comparison with the traditional method botn of them give more reasonable design criteria for coastal defence of Shanghai.This paper also proposes the Poisson-Gumbel Mixed Compound ExtremeDistribution model to analyze flood frequency--the joint probabilityof flood peaks and statistically corresponding volumes. The model istested and validated using observed flood data from the hydrological station Datong of river Yangtze.