基于极值拓展模型的广东汕尾增水推算

Calculation of Surge in Shanwei of Guangdong Province Based on Extreme Value Extension Model

为解决传统极值分布的尾部趋于零的速度快而导致的对海洋观测极值数据的尾部拟合不足的问题,本文基于组合模型的思想构建了一系列用于推算海洋环境设计参数的极值拓展模型。该系列模型利用积分变限函数融合了对数据尖峰拟合较好的经典极值分布和对数据厚尾拟合良好的Pareto分布的特征,可以同时拟合数据的“尖峰”和“厚尾”。利用该系列模型对广东汕尾1990—2012年间的增水数据进行分析,通过与传统极值模型的拟合效果进行对比,得到拟合数据的最佳模型为Gumbel-Pareto极值拓展模型。应用最佳拟合模型推算汕尾不同重现期下的增水值,为广东汕尾的海洋工程建设与安全分析提供参考。

In order to solve the problem that the tail of the traditional extreme value distribution tends to zero quickly,which leads to the insufficient tail fitting of Marine observation extreme value data,this paper constructs a series of extreme value extension models for calculating the design parameters of Marine environment based on the idea of combinative model.This series of models use integral variable limit function to integrate the characteristics of classical extreme value distributions which fit the peak of data well and Pareto distribution which fits the thick tail of data well,which can fit the“peak”and“thick tail”of data simultaneously.This series of models was used to analyze the surge data from 1990 to 2012 in Shanwei of Guangdong Province.By comparing the fitting effect with the traditional extreme value model,the best model of fitting data is Gumbel-Pareto extreme value extension model.The optimal fitting model is used to calculate the surge value of Shanwei in different recurrence periods,which provides reference for the construction and safety analysis of ocean engineering.