PERFORMANCE EVALUATION OF THE CORDEX AFRICA REGIONAL CLIMATE MODELS IN SIMULATING RAINFALL AND AIR TEMPERATURE THE CASE OF WABI SHEBELE BASIN, ETHIOPIA

Abstract

The impact of the climate change such as an increase in the temperature, and rainfall seasonal shift are observable in the world and particularly in Ethiopia. These changes may affect the socio economic life of the people if it is not assessed and predicted well. Regional Climate Models (RCMs) are mostly used in local scale assessment of climate change. But due to reflection of the inherent and methodological uncertainties in climate modeling, the reliability of individual models needs to be assessed before using their output for impact and climate change assessment. The objective of this study is to evaluate the performance of RCMs under the project of Coordinated Regional Climate Down-Scaling Experiment (CORDEX) African domain under Representative Concentration Pathways (RCP) of 4.5 and 8.5 scenarios. The RACMO22T, CCLM4, RCA4 and, HIRHAM5 models which were driven by ICHEC-EC-EARTH are evaluated against observed rainfall and air temperature data for a period of (1986 to 2005) over the Wabi Shebele basin, Ethiopia. The statistical measures like Bias, RMSE and, r were used for model performance evaluation. The evaluation is mainly on how best RCMs performed in simulating annual rainfall, maximum, and minimum air temperature over the study area. All models best simulated in dry months. In air temperature simulations, all models under simulated the maximum temperature and overestimated the minimum temperature in most months. Based on statistical result for performance evaluation, monthly cycle, annual and seasonal variability of the RCMs, RCA4 was well performed in both simulating rainfall and air temperature. After the models performance was checked, the best performed model, RCA4 was selected and bias corrected. With the use of Rstudio software wich is high programming language trend was analysised for time series of annual rainfall, maximum and minimum air temperature for observed (1986-2005). Further,the best performed model bias corrected data under both RCP scenarios for a period of 2031-2050 and 2051-2070 years was done. The results of trend analysis indicated that a decrease in annual rainfall and an increasing trend in average maximum and minimum air temperature for selected scenarios over the Wabi Shebele basin. In this regard, RCA4 regional climate models have to be used for any climate change impact study and for local-scale climate projections over the Wabi Shebele basin