Trends and variability of rainfall, temperature and associated extremes, and their future projections in Rwanda using regional climate models

Abstract

This study examines observed trends, variability, and future projections of rainfall, temperature, and associated extremes in Rwanda using outputs from the Coordinated Regional Climate Downscaling Experiment-Coordinated Output for Regional Evaluation (CORDEX-CORE) regional climate models (RCMs). Observational trends and variability were analyzed for 19832021 using data from the Rwanda Meteorology Agency, while future projections were examined for 2026-2060 and 2066-2100 under two representative concentration pathways (RCP2.6 and RCP8.5). The performance of CORDEX-CORE RCMs in simulating observed rainfall was evaluated for 1983-2005. The study also assesses the influence of large-scale climate drivers, particularly the El Niño-Southern Oscillation (ENSO) and the Indian Ocean Dipole (IOD), on Rwanda’s rainfall patterns. For the observational period, trends and magnitudes were assessed using the Modified Mann-Kendall test and Theil-Sen estimator, while temporal variability was analyzed using standard deviation and coefficient of variation. Rainfall characteristics, namely onset day (OD) and cessation day (CD) of the rainy season, were determined using methods based on accumulation thresholds, rainy day counts, dry spell and potential evapotranspiration analysis. Climate extremes were evaluated following the methodology of the Expert Team on Climate Change Detection and Indices (ETCCDI). Results show that the multi-model ensemble (CORDEX-CORE-MME) outperforms individual models in simulating Rwanda’s present climate. During the observational period, maximum temperatures exhibited higher variability than minimum temperatures. However, minimum temperatures exhibited greater warming magnitudes than maximum temperatures. Rainfall showed high spatial and temporal variability, with mostly non-significant trends across regions. Rainfall onset and cessation dates revealed west-east and east-west progression patterns, respectively. Seasonal length (SL) and number of rainy days (RD) were generally longer and more frequent in the southern, western, and northern regions, while central and eastern areas had shorter seasons and fewer rainy days. Climate extreme indices for rainfall and temperature displayed spatially and seasonally heterogeneous trends, with significant increases or decreases depending on the index and emission scenario. Future projections suggest shifts in OD and CD, with earlier onsets and longer seasons more common under RCP2.6 than RCP8.5. Rainy days are projected to increase during the long rains, but exhibit more complex patterns in the short rains, including regional decreases under RCP8.5 by century’s end. Heavy vi (R10mm) and very heavy (R20mm) precipitation days are projected to decline during long rains, while consecutive wet days (CWD) are expected to rise, particularly under RCP2.6 in 2026-2060. For temperature extremes, cold days (Tx10p) are projected to increase during long rains for 20262060 but decline in 2066-2100, while warm days (Tx90p) show the opposite pattern. These findings provide crucial insights to inform climate-resilient planning and policy development across key socio-economic sectors, including agriculture, health, water resources, infrastructure, tourism, and disaster risk reduction, and support the formulation of effective adaptation strategies at local and national levels.