Abstract: Based on monthly precipitation and temperature dataset (1986-2005) of 142 stations in the Yangtze River basin, the capability of the COSMO model in Climate Mode (CCLM) to reproduce the spatial and temporal variability of the observed precipitation and temperature was examined. The bias between simulation and observation was corrected by the equidistant cumulative distribution functions (EDCDF). The analysis shows that the spatial changes of simulated annual precipitation and mean temperature are broadly comparable with the observations. CCLM tends to overestimate temperature at both seasonal and annual time scales. CCLM fails to simulate the monsoon precipitation, thereby to underestimate precipitation in summer and autumn, and overestimate it in winter and spring. On the whole, temperature is better simulated than precipitation. Based on the bias-corrected projection, annual mean temperature over the whole basin will increase 0.66℃, while annual precipitation will reduce 2.2% in 2016-2035 under the RCP4.5 scenario, relative to the observations of 1986-2005.

Key words: regional climate model, Yangtze River basin, temperature, precipitation, RCP4.5 scenario