Projecting meteorological drought in Northern Malawi using SPEI and bias-corrected CMIP6 models

Northern Malawi’s hydroclimatic regime is highly sensitive to climate change, yet remains underrepresented in regional drought projection studies. This study provides a high-resolution, CMIP6-based characterization of future meteorological droughts across five synoptic stations (Bolero, Karonga, Mzimba, Mzuzu, Nkhata Bay) using the Standardized Precipitation Evapotranspiration Index (SPEI) at 6- and 12-month scales. Downscaled projections were bias-corrected using delta change and empirical quantile mapping (EQM) techniques applied to a five-model ensemble selected based on Taylor skill scores (R > 0.85) and distributional fidelity. Temporal drought dynamics were analyzed for mid-century (2021–2050) and mid–late century (2051–2080) under SSP2-4.5 and SSP5-8.5 scenarios. Results show statistically significant increases in drought severity and persistence under SSP5-8.5. Median SPEI-12 values at Mzimba declined from − 0.42 to − 0.63 between the two future periods (p < 0.001), with comparable changes at Mzuzu (–0.38 to − 0.57, p < 0.01) and Bolero (–0.33 to − 0.54, p < 0.05), indicating a strengthening of dryness. Seasonal droughts (SPEI-6) exhibited even steeper trends, with Mzimba’s median declining by 44% and Bolero’s by 40% from mid- to late century. Receiver Operating Characteristic (ROC) analysis confirmed model skill in detecting historical drought thresholds, with AUC values > 0.87 across all stations. These findings underscore the intensifying drought risk under warming scenarios and highlight the importance of integrating temperature-driven indices and ensemble diagnostics into regional adaptation planning.