Bivariate homogenous regions and projections based on copula function using RDI and SPI indices for drought risk assessment in Pakistan

This study aims to construct bivariate homogenous regions and projections for drought and water resources planning in Pakistan. The Reconnaissance Drought Index and Standardized Precipitation Index are applied to extract variables of drought duration and severity for the selected sites which have high correlation and skewed shape. Five homogenous regions are constructed WRT drought duration and severity, simultaneously. Best-fit regional probability distribution out of five 3-parameter distributions is selected and estimated using L-moment statistical technique. For joint distribution, the best-fit Gumbel-Hougaard copula is selected based on graphical and numerical statistical tools. The selected copula has the property to measure upper-tail dependence which is highly important in drought risk assessment. Potential drought risks are estimated and compared through univariate and bivariate joint return periods at selected years. The drought duration and severity increase with an increase in return periods. The contour lines for combinations of drought duration and severity at selected return periods are graphed. Most of the combinations in the regions are near the lines of lower return periods and have smaller interarrival times while some with larger interarrival times. Both the indices have no significant differences in results. Pakistan is facing frequent droughts and water scarcity. These results might help water resources planners, disaster management units, drought mitigation, and irrigation departments for better climate change planning in the country.