The variability and predictability of the East Asian (EA) winter climate has been studied, based on observed datasets and multi-model ensemble (MME) hindcast experiments archived at the APEC Climate Center (APCC). The focus is on the leading modes of wintertime variability over the eastern to northeastern part of Asia, which are identified based on multivariate EOF analysis of the monthly 850 hPa wind and temperature. The leading EA climate mode is characterized by continental-scale temperature anomalies covering a broad region from the northwestern flank of the Siberian high to northeast Asia. The second mode is associated with fluctuations of temperature and monsoon northerlies over the EA locations of Korea, Japan and eastern coastal China. Moreover, the first mode is found to be influenced by the Scandinavian (SCA) pattern, while the second mode is closely associated with the Eurasia (EU) pattern.
In general, the dominant circulation patterns of the EA wintertime variability from each MME member model compare well with their observational counterparts. However, the temporal variations of these modes are difficult to reproduce in the model simulations. The variation of the leading mode is found to be better predicted by most models, which leads to better predictions of the winter climate over continental northeast Asia, compared to the second mode. The MME performance is further assessed in the context of circulation changes during ENSO. It is found that most models have difficulty in capturing both the timing and strength of the observed second EA climate mode variations. Analysis based on observations shows that there is Rossby wave activity from Eurasia in early boreal winter during ENSO years. The Eurasian wave train, however, is either too weak or absent in the model simulations. Overall, these results highlight the difficulties in forecasting EA winter climate in the current framework of seasonal climate prediction.