Evaluation of a regional climate model for maximum and minimum temperature in Rwanda under Climate change

Abstract

The aim of this research is to evaluate the ability of a Regional Climate Model (RegCM 4.7) used in the Coordinated Regional Climate Downscaling Experiment (CORDEX) to reproduce the seasonal mean maximum temperature (TX) and mean minimum temperature (TN) patterns over Rwanda during March, April, May (MAM) and October, November, December (OND) seasons. The Regional Climate Model (RCM) outputs were evaluated using observational gridded datasets collected from the Rwanda Meteorology Agency between 1983 and 2005 period. All simulations performed are at a resolution of 25 km (0.22o x 0.22o) over the project domain. The analysis is based on determining how effectively the Regional Climate Model reproduces climatological trends and inter-annual variability in minimum and maximum temperatures. We compared forecasts of Minimum and Maximum temperature changes by 2021 to 2050 and 2051 to 2080 for one Regional Climate Model using two Representative Concentration Pathways, RCP 2.6 and RCP 8.5 to assess uncertainty in these variables. The Bias, Root Mean Square Error, and trend analysis were used as statistical metrics of model performance. The findings from this study show that a Regional Climate Model (RegCM 4.7) reproduced interannual variability and trends in both minimum and maximum temperatures. In general, for both March to May (MAM) and October to December (OND) seasons, the Biases and Root Mean Square Error (RMSE) resulting from the Regional Climate Model and driving General Climate Model in simulating Minimum temperature and maximum temperature are relatively small. Therefore the model can be used to simulate the future climate of Rwanda. The results from projected minimum temperature (TN) and maximum temperature (TX) for March to May and October to December seasons using both Representative Concentration Pathways, RCP 2.6 and RCP 8.5 showed that minimum and maximum temperatures will generally increase for both seasons and Representative Concentration Pathways (RCPs), with an emphasis on the eastern region of the country. The results from projected change for minimum temperature (TN) and maximum temperature (TX) during the March to May and October to December wet seasons using both Representative v Concentration Pathways, RCP 2.6 and RCP 8.5 showed a positive change (warming) for minimum temperature and maximum temperature almost everywhere in the country, however the change in March to May is higher compared to the change in October to December. For most locations, average changes in minimum temperature were greater than associated changes in maximum temperature for both MAM and OND across the two RCPs. Results from trends show the highest trend of 0.04757oC/year for maximum temperature during March to May (MAM) season with RCP 8.5 scenario and the lowest trend of -0.0006oC/year for maximum temperature is observed during MAM with RCP 2.6. Similarly, Results show the highest trend of 0.0429oC/year for minimum temperature during MAM with RCP 8.5 scenario and the lowest trend of 0.0005oC/year for maximum temperature is observed during October to December with RCP 2.6.