全球变化知识资源中心

The present study evaluates the performance of 19 climate models that participated in the coupled model intercomparison project phase 5 (CMIP5) in reproducing climatology, the standard deviation, and the dominant mode of spring surface air temperature (SAT) variations over the mid-high latitudes of Eurasia based on historical runs. Future change of the Eurasian spring SAT under the anthropogenic global warming is also examined by comparing the historical and RCP4.5 run. All the 19 CMIP5 models capture well the observed spatial structure of climatological spring SAT, with the pattern correlation coefficients all larger than 0.94. However, most of the models tend to underestimate the SAT over north Europe and north Siberia and overestimate the SAT south of 50 degrees N. There exists large inter-model spreads in the standard deviation of the spring SAT. Most of the models capture realistically the observed dominant mode of interannual variations of spring SAT. Analyses show that the ability of a CMIP5 model in capturing the dominant mode of Eurasian spring SAT variations is connected with the model's performance in representing the observed atmospheric circulation anomalies related to the Arctic Oscillation and the dominant mode of the atmospheric variations over Eurasia. Six best models are selected based on the ability in simulating the dominant mode of the spring SAT variations in the historical runs. These six models project an increase in the SAT climatology but a decrease in the standard deviation over most of Eurasia. These six models project a decrease in the explained variance as well as in the amplitude of the spring SAT and atmospheric anomalies related to the dominant mode.