Statistical Modelling of Extreme Rainfall in Taiwan
In this study,the annual maximum daily rainfall from 1911 to 2010 are modeled for 6 stations in Taipei,Taichung,Tainan,Hengchun,Hualien,and Taitung.We fit the rainfall data both with stationary and non-stationary generalized extreme value (GEV) distribution,and further estimated future patterns based on the best fitted model.The non-stationary model used the location and scale parameter in the GEV distribution as a function of time to detect the temporal trend of the maximum rainfall.The results suggest that non-stationary GEV and Gumbel distribution provide best fit for Haulien and Taitung,respectively.The return levels of the stationary models for 4 stations and their 95% confidence intervals (CI) are provided.In addition,the return period of Typhoon Morakot is also calculated based on stationary models in this study.
Extreme Value Theory Extreme Rainfall Return Level Typhoon Morakot
Michael McAleer Lan-Fen Chu Ching-Chung Chang
Econometric Institute, Erasmus, School of Economics, Erasmus University Rotterdam;Tinbergen Institut National Science and Technology Center for Disaster Reduction(NCDR), Taiwan Institute of Economics, Academia Sinica, Taiwan
国际会议
太原
英文
1389-1392
2012-12-08（万方平台首次上网日期，不代表论文的发表时间）