Assessment of CMIP5 GCM Simulation Performance for Temperature Projection in the Tibetan Plateau

The performance of 33 Coupled Model Intercomparison Project 5 (CMIP5) general circulation models (GCMs) in temperature simulations in the Tibetan Plateau (TP) was comprehensively assessed by an improved score-based and multiple-criteria method using data collected from 1961 to 2005. Future temperatures were also simulated based on a multimodel ensemble coupled with the Delta downscaling method for near-term (2006-2050) and long-term (2051-2095) projections under Representative Concentration Pathways (RCP, scenarios RCP4.5 and RCP8.5). Our results demonstrated that all the GCMs evaluated in our study could capture the seasonal temperature patterns. However, most GCMs tended to underestimate temperatures by an average of -2.0 degrees C. All the GCMs could effectively simulate temporal distribution, with a mean correlation coefficient of 0.997. However, they did not perform well in reproducing spatial distribution. Different assessment criteria lead to inconsistent results; however, the improved rank score method with multiple criteria provided a robust assessment of GCMs performance. MPI-ESM-LR, CMCC-CMS, and GFDL-ESM 2M showed a better temperature simulation performance compared to the other GCMs that we have assessed. Topographic correction could effectively enhance spatial distribution simulation; however, this increased temperature underestimation. Future temperatures were projected to increase by 1.4 degrees C and 1.6 degrees C in near-term, and by 2.4 degrees C and 4.0 degrees C in long-term under RCP4.5 and RCP8.5 scenarios, respectively. High temperatures mainly occurring in the southeastern end of the Himalayas, as well as the northern and southeastern margins of the TP. The results are expected to provide valuable information on climate change and its impact on hydrology, ecology, and socioeconomics of the TP.