An alternative approach for extreme value statistics of climatologic dependent variables in small samples


Sen Z.

THEORETICAL AND APPLIED CLIMATOLOGY, vol.58, pp.169-173, 1997 (Journal Indexed in SCI) identifier identifier

  • Publication Type: Article / Article
  • Volume: 58
  • Publication Date: 1997
  • Doi Number: 10.1007/bf00865017
  • Title of Journal : THEORETICAL AND APPLIED CLIMATOLOGY
  • Page Numbers: pp.169-173

Abstract

Extreme event analysis of meteorological variables by conventional statistical methods does not meet the necessary mathematical assumptions such as the independence of successive occurrences. An alternative method is proposed for analyzing extreme values of dependent variables by considering all the extreme values above a threshold value in finite samples. The probability for the maximum event in a given finite sequence of dependent meteorological variables is derived explicitly.