ESTIMATION OF DROUGHT RISKS USING ARCHIMEDEAN COPULAS IN THE KARASU RIVER, TURKEY


KATİPOĞLU O. M., Can I.

FRESENIUS ENVIRONMENTAL BULLETIN, vol.27, pp.9799-9811, 2018 (SCI-Expanded) identifier

  • Publication Type: Article / Article
  • Volume: 27
  • Publication Date: 2018
  • Journal Name: FRESENIUS ENVIRONMENTAL BULLETIN
  • Journal Indexes: Science Citation Index Expanded (SCI-EXPANDED)
  • Page Numbers: pp.9799-9811
  • Erzincan Binali Yildirim University Affiliated: No

Abstract

In this article, the copula based bivariate drought severity and duration frequence analysis is employed in order to calculate recurrence intervals. Drought duration and severity is acquired by using the runs analysis. for each threshold level (q = 0.5 m(3)/s; q =0.3 m(3)/s; q= 0.1 m(3)/s), Drought severity and duration are presumed to be univariate gamma and exponential distributions, respectively. The Drought return period varies depending on the selected marginal distributions and copula functions. So that the best copula class should be determined. Among probable copulas, Gumbel copula has been found most suitable for drought analysis. The best copula class is determined thereby comparing Bayesian Information Criterion and Akaike Information Criterion of duration and severity of drought. Recurrence intervals for the dry terms lengths of historic and synthetic series are calculated using the selected copula class for each threshold level (q = 0.5, q = 0.3, q = 0.1 (m(3)/s)). Using these recurrence intervals, drought risks are obtained. Drought frequency analysis provides a detailed view point by considering duration and severity when managing water resource, assessing drought occurrence risks.