Copula-based risk evaluation of global meteorological drought in the 21st century based on CMIP5 multi-model ensemble projections

Abstract Climate warming is expected to have significant impacts on the global hydrologic cycle, including changes in precipitation-induced extremes such as droughts. Using 28 CMIP5 global climate models (GCMs), this study presents a global-scale analysis of the joint return periods (T) of meteorological drought characteristics (duration D, severity S, and peak P) at the 6- and 12-month scales under the Representative Concentration Pathways scenarios RCP2.6 and RCP4.5. The D, S, and P estimated based on the Runs theory are used to calculated from the Standardized Precipitation Index (SPI) at the global 1o× 1o grids. Six marginal probability distributions are used to fit S, P and D, whereas three Archimedean copula functions (Clayton, GH and Frank) are used to estimate T for the paired drought characteristics (S-D, P-D and S-P). Large spatial variability is found globally in the best-fitted copulas for the paired drought characteristics, with the Frank (Frank and GH) of the largest global percentage for SPI6 (SPI12). Relative to the baseline (1971–2000), the T of the paired drought characteristics above the moderate drought (S>1, P>1 and D>3) is projected to decrease mostly in North America and Asia (