摘要
以浙江温州站点为例,对5种极值模型模拟台风降水极值累积概率分布的方法进行了详细阐述,并分析不同极值模型的差异;对全国748个气象观测站点的1951—2005年所有台风事件的降水及概率进行了系统分析,结果表明Weibull分布模拟中国台风极值降水整体检验效果最佳;定量探讨了Weibull分布与其他分布模型在台风降水强度-概率分布上的定量差异及采用不同分布模型可能造成的台风降水概率分布不确定性.
Rainfall is a major typhoon hazard,causing direct economic loss.Rainfall may induce secondary disasters such as floods,landslides,and mudslides.Five extreme value models were used to simulate cumulative probability of typhoon rainfall in the present work,using Wenzhou station in Zhejiang province as an example,with the differences among the five models analyzed.Based on surface observed rainfall from 748 stations（1951—2005）,typhoon rainfall probability of 402 stations impacted by typhoons was calculated.Significance test showed that Weibull distribution was the best model to fit typhoon rainfall probability in China,though other models achieved relatively good performance also.Rainfall intensity difference or uncertainty caused by selecting different models was quantitatively mapped.
出处
《北京师范大学学报（自然科学版）》
CAS
CSCD
北大核心
2011年第5期493-501,共9页
Journal of Beijing Normal University(Natural Science)
基金
国家科技支撑计划资助项目(2008BAK49B04
2008BAK50B07)
关键词
台风降水
概率分布
极值理论
不确定性
中国
typhoon rainfall
probability distribution
extreme value theory
uncertainty
China
