【摘要】 本文介绍了以两种不同类型的样本求取多年一遇极值的推算方法。第一种类型是用长期资料的年极值为样本,分别依耿贝尔极值分市(矩法估算参数和最小二乘法估算参数)、第一极值分布、皮尔逊Ⅲ型分布,推算了多年一遇的风暴增水和水位;对所有这些分布的推算结果,用柯尔莫哥洛夫方法进行了测适度检验,给出了95%的可信度。第二种类型用短期连续资料力样本,以三个月的实测水位资料用Weibull分布推算了不同重现期的水位。用两种类型推算的结果是相同的。更多还原

【Abstract】 In this paper, two methods for calculating the extreme values with required periods are presented using different types of samples. For the first method, the annual extreme values of long-term data are taken as samples to estimate the parameters for Gumbel distribution, its first extreme disthibution and pearson-Ⅲ distibution; then the extreme storm setups and sea levels with required return periods are calculated. Finally the Kolmogorov method is used to make a goodness of fit test of the above derived distributions at 95% confidence level. For the second method, three months of consective sea level observation data are taken as samples to calculate the extreme sea levels with different return periods by the use of Weibull distribtion. The extreme sea levels obtained with the two methods are about the same.更多还原