SOME CRITICAL REMARKS ON THE USE OF EXTREME-VALUE STATISTICS IN CLIMATOLOGY

被引:10
作者
GERSTENGARBE, FW
WERNER, PC
机构
[1] Department of Research, German Weather Service, Potsdam, Telegrafenberg
关键词
D O I
10.1007/BF00865546
中图分类号
P4 [大气科学(气象学)];
学科分类号
0706 ; 070601 ;
摘要
Conventional statistical methods for the description of extreme events do not meet all mathematical conditions with respect to the analysis of meteorological parameters. Examples are presented illustrating the problems of using conventional extreme-value statistics. Inter alia, non-stationarity, extrapolation and estimation of the errors are discussed. Some ideas on how to solve these problems are presented, e.g. by the use of marginal distributions, spline functions and the estimation of the magnitude of non-stationarity.
引用
收藏
页码:1 / 8
页数:8
相关论文
共 9 条
[1]  
BOYACK CF, 1985, METEOROL MAG, V114, P100
[2]   Limiting forms of the frequency distribution of the largest or smallest member of a sample [J].
Fisher, RA ;
Tippett, LHC .
PROCEEDINGS OF THE CAMBRIDGE PHILOSOPHICAL SOCIETY, 1928, 24 :180-190
[3]  
FLOHN H, 1989, MANNH FOR 88 89 BOEH, P135
[4]  
GERSTENGARBE FW, 1989, Z METEOROL, V39, P224
[5]  
GERSTENGARBE FW, 1987, Z METEOROL, V37, P299
[6]  
GUMBEL EJ, 1964, SCHWEIZER ARCH ANGEW, V30
[7]  
Hupfer P., 1989, Zeitschrift fur Meteorologie, V39, P327
[9]  
OLBERG M, 1972, GERL BEITR GEOPHYS, V81, P57