Theory of multivariate compound extreme value distribution and its application to extreme sea state prediction  

作　　者：LIU Defu WANG Liping PANG Liang 

机构地区：[1]Disaster Prevention Research Institute, Ocean University of China, Qingdao 266003, China [2]Department of Mathematics, Ocean University of China, Qingdao 266071, China [3]College of Engineering, Ocean University of China, Qingdao 266071, China

出　　处：《Chinese Science Bulletin》2006年第23期2926-2930,共5页

基　　金：This work was supported by the National Natural Science Foundation of China(Grant No.50379051).

摘　　要：In this paper, a new type of distribution, multivariate compound extreme value distribution (MCEVD), is introduced by compounding a discrete distribution with a multivariate continuous distribution of extreme sea events. In its engineering application the number over certain threshold level per year is fitting to Poisson distribution and the corresponding extreme sea events are fitting to Nested Logistic distribution, then the Poisson-Nested logistic trivariate compound extreme value distribution (PNLTCED) is proposed to predict extreme wave heights, periods and wind speeds in Yellow Sea. The new model gives more stable and reasonable predicted results.In this paper, a new type of distribution, multivariate compound extreme value distribution （MCEVD）, is introduced by compounding a discrete distribution with a multivariate continuous distribution of extreme sea events. In its engineering application the number over certain threshold level per year is fitting to Poisson distribution and the corresponding extreme sea events are fitting to Nested Logistic distribution, then the Poisson-Nested logistic trivariate compound extreme value distribution （PNLTCED） is proposed to predict extreme wave heights, periods and wind speeds in Yellow Sea. The new model gives more stable and reasonable predicted results.

关 键 词：多变量混合极端价值分布 嵌套-后勤模型 极端海洋状态 阈值 

分 类 号：P7[天文地球—海洋科学]