Evaluations of regional climate models for simulating precipitation and temperature over the Guder sub-basin of Upper Blue Nile Basin, Ethiopia

This study evaluates the performance of five Regional Climate Models (RCMs) (CCLM4-8-17, HIRHAM5, RACMO22T, RCA4 and REMO2009) driven by a Global Climate Model) (GCM) (ICHEC-EC-EARTH) for simulating rainfall and temperature in the Guder sub-basin of the Upper Blue Nile Basin in Ethiopia. The RCMs data were downloaded from the Earth System Grid Federation (ESGF) website and observational data were obtained from the National Meteorology Agency (NMA). The RCMs were evaluated against observed data on the basis of how they produce the monthly average, wet season (June-September), and annual average rainfall and temperature during the 1986-2005 periods. The result showed that in most stations RCA4, HIRHAM5, and CCLM4-8-17 models overestimate the rainfall with a minimum bias of 0.01 mm at Jeldu station and a maximum bias of 6.48 mm at Ambo station. On the other hand, all the models underestimate the maximum temperature between 0.46 and 10.04 degrees C and overestimate the minimum temperature in most climate stations within the ranges 0.14 degrees C-5.03 degrees C. In most of the statistical metrics, RACMO22T was superior while the RCA4 and HIRHAM5 RCMs models show the poorest performance in terms of capturing monthly and annual cycles of rainfall. But, for temperature simulation, HIRHAM5 simulation was relatively the best in simulating both the annual and wet season maximum and minimum temperature over the Guder sub-basin. This finding indicated that the best model for rainfall simulation had a poor performance for temperature. Therefore, this justifies the need for RCMs model evaluation in order to choose the most realistic model for a localized climate impact study.