中文摘要：

目前国际上设计波高的常用计算方法有三种：年最大波高法（Annual Maximum method）、波高阈值法（Peak—Over—Threshold method）和年N大波法（Annual N-Largest method）。利用澳大利亚悉尼观测站连续16年的实测资料和Gumbel概率分布函数对这三种方法进行分析和比较，选出一种比较好的计算方法。结果表明，由年最大波高法计算得到的设计波高值偏大，线性回归系数较小，均方误也较大；波高阈值法计算所得结果与阈值本身取值关系并不大，线性回归相关性也较好；年N大波法计算得到的设计波高与每年大波数量取值无关（N≥5时），均方误较小，线性回归相关性较好。因此，在这三种方法中，年N大波法更适合用来计算设计波高，建议每年大波数取5-6个。

英文摘要：

Three commonly-used methods for calculation of design wave height, i.e. the Annual Maximum method, the Peak - Over- Threshold method and the Annual N-Largest method, have been compared based on the 16-year storm peak wave height data collected off Sydney in Tasman Sea, Australia. The Gumbel distribution function and the least-square fitting method are used to calculate design wave heights. It has been found that the Annual N-Largest method （ N 〉 5） is in general more accurate than the Peak - Over - Threshold（POT） method, but much more accurate than the Annual Maximum method in terms of the correlation coefficient R2. This study also finds that the POT method is insensitive to the threshold wave height when the average number of storm peak wave heights per year, λ, is larger than 8, while the Annual N-largest method becomes insensitive to the annual number of storm peak wave heights, N, when N≥5.