Present day bias and future change signal of temperature over China in a series of multi-GCM driven RCM simulations
Abstract Simulation of surface air temperature over China from a set of regional climate model (RCM) climate change experiments are analyzed with the focus on bias and change signal of the RCM and driving general circulation models (GCMs). The set consists of 4 simulations by the RCM of RegCM4 driven by 4 different GCMs for the period of 1979–2099 under the mid-range RCP4.5 (representative concentration pathway) scenario. Results show that for present day conditions, the RCM provides with more spatial details of the distribution and in general reduces the biases of GCM, in particular in DJF (December–January–February) and over areas with complex topography. Bias patterns show some correlation between the RCM and driving GCM in DJF but not in JJA (June–July–August). In JJA, the biases in RCM simulations show similar pattern and low sensitivity to the driving GCM, which can be attributed to the large effect of internal model physics in the season. For change signals, dominant forcings from the driving GCM are evident in the RCM simulations as shown by the magnitude, large scale spatial distribution, as well as interannual variation of the changes. The added value of RCM projection is characterized by the finer spatial detail in sub-regional (river basins) and local scale. In DJF, profound warming over the Tibetan Plateau is simulated by RCM but not GCMs. In general no clear relationships are found between the model bias and change signal, either for the driving GCMs or nested RCM.