ABSTRACT In the era of global warming, drought remains one of the most pressing, insidious and complex climatic phenomena. Although Global Climate Models (GCMs) are indispensable for simulating past and future climate patterns, they continue to face persistent challenges in accurately characterising drought under diverse climate conditions due to inherent model uncertainties. Multi‐model ensembles (MMEs) are widely adopted to reduce these uncertainties; however, conventional MME approaches often overlook three critical aspects: (i) the independent relative contribution of individual GCMs, (ii) the influence of point‐wise value deviations and (iii) the sequential order of predictor inclusion in the analysis. Neglecting these factors can result in biassed weighting, reduced interpretability and overestimation of drought severity in subsequent assessments. This study introduces an integrated two‐stage framework, Integrated Ensemble Based on Sequential Predictor Contribution Analysis for Standardised Climate Indices (IEBSPCASCI), for future drought assessment using GCM projections, combining a novel weighting strategy with an optimised drought index formulation. In Stage 1, we propose the sequential relative impact weighting ensemble (SRIWE), a new GCM ensemble weighting scheme that accounts for point‐wise discrepancies and sequential predictor contributions with addressing multicollinearity among models; in Stage 2, we develop the Standardised Partitioned Impact Severity Drought Index (SPISDI), an optimal‐standardisation‐based index enhancing drought severity quantification under future climate scenarios. The proposed IEBSPCASCI framework is applied to 94 locations in Pakistan using 22 GCM from the Coupled Model Intercomparison Project Phase 6 (CMIP6) for the period 1950–2014. The Stage 1 weighting scheme, SRIWE, outperforms conventional MMEs, achieving the highest median and mean Kling–Gupta Efficiency (KGE) (0.2650, 0.2435) and Nash–Sutcliffe efficiency (NSE) (0.2074, 0.1075), and the lowest median and mean values of Mean Absolute Error (MAE) (12.5572, 18.027), indicating superior predictive skill and smaller errors. In stage 2, SPISDI‐based drought assessments reveal a consistent increase in drought severity and frequency in all SSP scenarios. Under SSP5‐8.5, the intensity of drought peaks markedly towards 2080–2100, signalling substantial climate‐induced risks to water resources and agriculture. Steady‐state probability analysis indicates that normal drought conditions dominate; however, extreme drought (ED) and severe drought (SD) events remain significant, with ED peaking at 2.37% (SSP2–4.5, 48 months) and SD reaching 8.6% (SSP1–2.6, 48 months). These findings underscore the urgent need for adaptive water management policies, climate‐resilient agricultural strategies and proactive drought risk governance to mitigate future socio‐economic impacts.
Abbas et al. (2026) studied this question.