LAUSR

The second empirical orthogonal function mode (EOF2) of winter surface air temperature (SAT) over 0°–180° E, 40°–90° N during 1979–2005 is defined as warm Arctic-cold Eurasia (WACE) pattern. The present study evaluates the performance of 25 Coupled Model Inter-comparison Project Phase 5 (CMIP5) models in simulating the WACE pattern based on historical runs. There exist large inter-model spreads in the simulation of the WACE pattern. Analyses show that the ability of a CMIP5 model in capturing the WACE pattern is connected with the model’s performance in representing the observed atmospheric circulation anomalies related to the winter sea ice concentration (SIC) variation over Barents–Kara Seas. Sea ice loss over Barents–Kara Seas can induce significant positive geopotential height anomalies over Arctic region and negative geopotential height anomalies around the Baikal Lake, resulting in warm anomalies over Barents–Kara Seas and cold anomalies over Eurasia. Further analysis shows that CMIP5 model’s performance in representing the SAT anomalies related to the WACE pattern is partly due to simulation of the amplitude of winter SIC variability over Barents–Kara Seas. Larger standard deviations of winter SIC over Barents–Kara Seas can instigate stationary wave-train more easily, which further induces the SAT anomalies.