Meteorological drought and the most suitable probability distribution models for drought parameters in the Middle Euphrates region, Türkiye

The drought, which impacts nature, people, and every area with a relationship with water, is a natural, complex, and serious disaster that must be determined, monitored, and planned for the whole world. In this thesis, drought is analysed, and the best marginal distributions for drought parameters of the middle Euphrates basin are determined. Standard Precipitation Index (SPI), Standardized Precipitation Evapotranspiration Index (SPEI), and Reconnaissance Drought Index (RDI) drought indices were used to analyse the 1-, 3-, 6-, 9- and 12-month meteorological drought at 29 meteorological stations throughout the basin. The drought durations and severity computed using these drought indices were fitted with Lognormal, Weibull, Gamma, Normal, Logistic, and Exponential marginal distributions. The best distribution model for drought duration and drought severity of all time periods was determined by the Kolmogorov-Smirnov (K-S), Cramer-von Mises (CvM), Akaike's Information Criterion (AIC), Bayesian Information Criterion (BIC) and Maximum likelihood (ML) Goodness-of-fit tests. It has been found that the middle Euphrates basin generally has wet and mild dry seasons. Extreme drought was also noted in the area. In terms of drought severity, the best distributions were Gamma and Weibull during short time periods (1 and 3 months), whereas lognormal distribution was mostly seen over long time periods (6-, 9-, and 12-months). The lognormal marginal distribution was discovered to be the most suitable for the drought duration series of all time periods. Key Words: Meteorological Drought, SPI, SPEI, RDI, Marginal Distribution, Middle Euphrates Basin, Goodness-of-fit