Hydrologic Drought Frequency Analysis in Karkhe Basin Based on Bivariate Statistical Analysis

Abstract

In recent years, the joint distribution properties of drought including severity and duration have been widely evaluated using copula. Few studies, however, have worked on drought modeling based on stream flow, especially in semi-arid regions such as the southern regions of Iran. This study followed two purposes. The first purpose was to find the appropriate marginal distribution function for hydrologic drought duration and severity, and the other one was to find appropriate copulas. First, the severity and duration of hydrological droughts in the hydrometery stations in the Karkhe basin were extracted by the run’s theorem, and the absence of trends was tested using the modified Mann-Kendall trend test. Appropriate marginal distribution functions for duration and severity drought were derived by using the linear momentchr('39')s method. In addition, copula’s parameters for Frank, Clyton and Gumbel families were calculated by both direct and indirect methods. The best copulas were selected by the goodness of fit tests. Finally, the joint and conditional return periods for duration and severity drought were derived for each station. The results showed that drought severity and duration for all hydrometery stations in Karkhe basin followed both the generalized extreme value marginal distribution function and Gumbel copulas family, which could be used for regional copulas modeling.

Keywords

• archimedean copulas
• goodness of fitness
• joint return period
• conditional return period.