Assessment of future climate change scenario over Rwanda using statistical downscaling approach

Abstract

Rwanda has been facing adverse consequences of climate change affecting its socio-economic sectors. This study adopts a statistical downscaling approach to refine CMIP5 GCM outputs (CanESM2) under three Representative Concentration Pathways (RCP2.6, RCP4.5, and RCP8.5) for 2010-2039 (2020s), 2040-2069 (2050s) and 2070-2099 (2080s) at the meteorological station level, aligning them with local requirements for impact assessment. The screen process was performed to select the predictors for rainfall, minimum temperature and maximum temperature. The performance of CanESM2 model statistically downscaled using statistical downscaling model (SDSM) results was based on evaluation metrics namely correlation (R), Root Mean Square error (RMSE), Mean Square Error (MSE), and Index of Agreement (IOA) during both the calibration and validation phases. Projected changes were performed based on to 1976 – 2005 reference period. The results show that the model is able to capture the annual cycle for both rainfall and temperatures (minimum and maximum). For calibration period (1983-1995), the correlation measure indicated a correlation ranging between 0.99 and 1 for maximum temperature, 0.86 and 0.99 for minimum temperature and 0.83 and 0.99 for rainfall during calibration. The RMSE ranges between 0.01 and 0.03 for maximum temperature, 0.01 and 0.72 for minimum temperature and 0.01 and 1.35 for rainfall. The MSE ranges between 0 and 0.01 for maximum temperature, 0 and 0.51 for minimum temperature and 0 and 1.83 for rainfall while IOA ranges between 0.99 and1 for maximum temperature, 0.23 and 0.99 for minimum temperature and 0.89 and 0.99 for rainfall. For evaluation (1996 -2005), the correlation measure indicated a correlation ranging between 0.36 and 0.99 for maximum temperature, 0.61 and 0.98 for minimum temperature and 0.70 and 0.94 for rainfall during evaluation. The RMSE ranges between 0.11 and 0.51 for maximum temperature, 0.27 and 0.99 for minimum temperature and 0.81 and 2.4 for rainfall. The MSE ranges between 0.01 and 0.26 for maximum temperature, 0.07 and 0.98 for minimum temperature and 0.58 and 3.94 for rainfall while IOA ranges between 0.59 and 0.99 for maximum temperature, 0.63 and 0.98 for minimum temperature and 0.57 and 0.95 for rainfall. It expected that the majority of meteorological stations will experience an increase ranging between 0.05°C and 11.8°C in minimum temperature across all emission scenarios and for all future periods, except for a few stations which are expected to experience a reduction ranging between -1°C and -2.5°C. For v maximum temperature, a projected increase ranging between 0.3°C and 8.5°C is expected in the station located in Kigali city and Eastern region, the remaining stations are expected to experience a reduction ranging between -0.07°C and -17°C across all emission scenarios and for all future periods. It expected that the most of meteorological stations will experience an increase ranging between 11.5mm to 49.97 mm in rainfall across all emission scenarios and for all future periods, while a reduction of -4.11mm to -4.39 mm is expected over a few stations. The obtained results are essential for developing suitable mitigation and adaptation measures to mitigate and adapt to the effect of climate change on socio-economic sectors in Rwanda.