Long-Term Prediction Of Positive And Negative Storm Surge Elevation In Qingdao Area

Qingdao is the important economic center and coastal city of Shandong. With economic rapid growth of Qingdao, the density of economic worth in coastal zone becomes larger and larger. Severe storm surges usually cause enormous economic losses. With the development of economy, disaster-prevention engineering should be constructed to protect reduce economic losses to a minimum.Basing on the observational data of Qingdao Port, this paper analyze meteorological origins which cause storm surge calamities in Qingdao and proposes a nonlinear method to fit long term distribution -- Genetic Algorithms. The proposed procedure is compared with traditional estimation methods, such as method of moment, least square method, maximum likelihood method and probability weighted moment method. This paper also presents a Combined Distribution Model, which is suitable for storm surge frequency analysis. Considering the influence of the storm surge frequency, Genetic Algorithms is applied to calculating the return values of positive and negative storm surge elevation.The major work of this paper is as follows:1. To collect the storm surge observational data of Qingdao hydrological station from 1950 to 1979 and analyze the meteorological causes of storm surges.2. To compare with different methods of estimating the extreme distribution parameters in hydrology statistics. In the long-term statistical analysis of the storm surge, this paper puts forward a kind of nonlinear methods to fit distribution -- Genetic Algorithm. This approach can be applied to the extreme distribution with several unknown parameters. Genetic Algorithm is a kind of optimization algorithms based on the rule of "survival of the fittest". It offers a kind of general frame that can solve complex systems. Independent on the specific systems,. Genetic Algorithm has stronger robustness for the problem solution. This method is simple and leads to less labor of calculation.3. To summarize the univariate extreme value distribution models and put forward a Combined Distribution Model, which is suitable for statistical analysis of storm surge in typhoon-effected area.4. Data samples are selected from 30-year observed storm surge series. Considering the occurrence frequency of typhoon, this paper calculates the return values of storm surge for disaster-prevention structure design.