In this study, the interannual variations of winter precipitation over central Asia (CA) are investigated over the period 1948–2018 using both observational analysis and a linear baroclinic model (LBM). The focus is on the characteristics and factors of the leading empirical orthogonal function mode (EOF1) of winter precipitation over CA. The results show that the key circulation anomalies associated with the positive phase of EOF1 feature a tripole pattern that is responsible for transporting moisture from the subtropical North Atlantic region to CA. An examination of the lower boundary conditions indicates that anomalous North Atlantic sea surface temperature (SST), Kara Sea ice in the preceding autumn, and simultaneous snow cover in central Asia can promote large-scale atmospheric waves that contribute to the EOF1-related anomalous tripole pattern. This result is verified by an analysis of the apparent heat source (Q1) in the atmosphere and the LBM experiments. Linear regression (LR) models were constructed using the precursors revealed by the above observational analyses to perform hindcasts for EOF1 over the period 1950–2018. The North Atlantic SST and the Kara Sea ice in the preceding autumn are shown to be effective predictors in the LR model that can capture the variation in EOF1 during this period. The seasonal forecasts of winter precipitation over CA for the 2011–18 period based on the LR models outperform those of the Climate Forecast System version 2 (CFSv2) over central and northern CA.

seasonnal forecasting,forecast skill,climate variability,interannual variability