热带气旋资料长度对风暴潮危险性评估结果的影响

Effects of the tropical cyclone data set length on the result of risk assessment of storm surge

热带气旋历史样本数不足一直困扰着风暴潮风险评估研究,本文基于西北太平洋62a(1949-2010年)历史观测热带气旋事件集资料和用随机模拟方法构造的1 000a模拟热带气旋随机事件集,以福建省连江县为例,开展了资料长度对风暴潮灾害危险性评估结果的影响分析。文中用ADCIRC模型模拟了两种数据集强迫下的风暴潮增水,采用极值I型分布法得到了典型重现期的风暴潮增水,经过对计算结果分析发现典型重现期的风暴潮增水计算结果与所用数据资料长度有着密切相关性,数据资料越长,结果越稳定。对于1 000a一遇的风暴潮增水值,使用500a长度的资料已经趋于稳定,并接近用1 000a资料计算得到的结果。在进行风暴潮危险性评估时,相比用几十年尺度的热带气旋历史数据集,1 000a的热带气旋模拟数据集的计算结果更具实际意义。

The reasonable storm surge risk assessment is always limited by the number of the historical Tropical Cyclone（TC）samples.In this paper the influence of the data set length to the storm surge risk assessment result is studied in Lianjiang County of Fujian Province,which is based on the historical TC observations（1949-2010）and stochastic TC events（1 000a）.The two datasets were applied to force the ADCIRC model to simulate storm surges,which is used to determine the surge heights of typical return periods with the method of extreme value type I.The experiment result indicates that the surge heights of typical return periods have a close correlation with data sets length,i.e.,the longer the data sets length,the more stable of the results.As for the surge heights of 1 000 years return period,the results from 500 alength data sets can give a rather stable result,which is close to the results generated from the 1 000 alength data set.When carrying out storm surges hazards analysis,the results are more reasonable by using TC stochastic events in 1 000 years than those from historical TC events in decades of years.