Trivariate Uncertainty Analysis of Hydro-Climatic Drought Risk Using Bootstrap Method (Case Study: Esteghlal Dam Basin, Minab)

Pakdaman, Zohreh; Bazrasshan, Omolbanin

doi:10.22059/ije.2025.395400.1871

Trivariate Uncertainty Analysis of Hydro-Climatic Drought Risk Using Bootstrap Method (Case Study: Esteghlal Dam Basin, Minab)

Articles in Press

Document Type : Research Article

Authors

¹ Department of Statistics, University of Hormozgan, Bandar Abbas, Iran

² - Professor, Department of Natural Resources Engineering, Faculty of Agriculture , University of Hormozgan,Bandar Abass, Iran

10.22059/ije.2025.395400.1871

Abstract

In this study, the maximum likelihood estimation (MLE) method was employed for multivariate analysis of hydro-meteorological drought. To this end, monthly precipitation and runoff data from 1991 to 2021 (1370 to 1400 in the Iranian calendar) were collected from the Esteghlal Dam watershed in Minab and normalized using the Gamma distribution. The Joint Drought Index (JDI) was constructed by integrating the SPI and SRI indices, and dependence analyses were conducted using copula functions. Using the standardized data, the hydro-meteorological drought index was calculated, and drought characteristics including severity, duration, and magnitude were extracted.

Subsequently, trivariate conditional probabilities of drought severity were predicted using copula functions, and the uncertainty associated with these predictions was quantified for different values of drought magnitude and conditional probability. The findings showed that after confirming significant correlations among drought variables and selecting the best-fitting marginal distributions, copula functions from the Archimedean and elliptical families were used for bivariate frequency analysis. Based on the Sn goodness-of-fit statistic, the optimal copula function was identified, and its parameters were estimated via MLE. Among the fitted copulas for severity, duration, and magnitude, the Gumbel copula showed the best performance. For example, at a conditional probability of 0.1

Keywords