Spatial-temporal trends and variability of rainfall and temperature in Northern Malawi

Understanding historical and projected climate variability is critical for developing effective adaptation strategies, particularly in regions highly dependent on rain-fed agriculture. This study analyzes the spatiotemporal trends of rainfall and temperature in Northern Malawi from 1991 to 2020 and projects future climate conditions up to 2080 using downscaled and bias-corrected Coupled Model Intercomparison Project Phase 6 (CMIP6) models. The analysis applies Pettitt's test for change point detection, followed by the Mann-Kendall trend test and Sen's slope estimator to assess segmented trends under stationary conditions. Results reveal a statistically significant warming trend across most locations, with mean annual maximum temperatures increasing by up to 0.029°C per year post-change, particularly at Mzimba, and minimum temperatures rising sharply at Mzuzu with a Sen's slope of 0.078°C/year (p < 0.01). Precipitation trends are more spatially variable: Bolero shows a significant pre-change increase of 7.25 mm/year (p = 0.032), while Mzuzu, despite a positive slope, displays no significant trend. Future projections under SSP2–4.5 and SSP5–8.5 scenarios indicate continued warming across the region, with maximum temperature increases reaching 2.51°C and minimum temperatures exceeding 3.1°C by 2080 under high-emission scenarios. Rainfall projections reveal both intra-annual and spatial heterogeneity—early- and late-season increases in rainfall contrast with mid-season drying, especially at Mzuzu, where a statistically significant annual wetting trend masks critical seasonal deficits. The 5th–95th percentile spread ranges further emphasize the uncertainty and potential intensification of hydroclimatic extremes. These findings underscore the need for spatially nuanced, seasonally aware adaptation policies and integrated climate risk assessments to enhance resilince.