Understanding Model Diversity in Seasonal Prediction of the Winter Surface Air Temperature Variations Over China

Abstract Accurate prediction of winter surface air temperature (SAT) is crucial for proactive government responses to potential winter hazards. However, current state‐of‐the‐art operational seasonal forecast systems still face challenges in skillfully predicting winter SAT variations over China, with considerable model diversity in prediction skill. Here, we identified the critical physical processes that determine the prediction skill of winter SAT over China to better understand the sources of model diversity. The first empirical orthogonal function (EOF) mode of interannual variability of winter SAT over China is dominated by the East Asian winter monsoon (EAWM), whereas the second mode is dominated by the Arctic Oscillation (AO) along with a wave train over the Eurasian continent (EAWT). Although most models generally capture the spatial structure of the observed EOF patterns and their connections with critical atmospheric circulation drivers, substantial disparities exist in their ability to predict the principal components (PCs) associated with these modes. These disparities account for the differences in prediction skills for winter SAT over China across models. Among the prediction systems, those that more accurately predict the EAWM (the AO and the EAWT) show the highest skill in forecasting PC1 (PC2), respectively. Furthermore, we show that inter‐model differences in the prediction skill of driving factors primarily arise from differences in representing their teleconnections with the El Niño–Southern Oscillation (ENSO). While most models can skillfully predict ENSO itself, they exhibit large spread in simulating the ENSO–EAWM and ENSO–EAWT relationships, which ultimately leads to substantial differences in the prediction skill of winter SAT over China.