Seasonal prediction of the northern and southern temperature modes of the East Asian winter monsoon: the importance of the Arctic sea ice

Abstract

Previous research work found that two distinct surface air temperature (SAT) modes—the northern (N-) and the southern (S-) modes dominate the East Asian winter monsoon (EAWM) region. The inter-annual variations of these two modes were mainly attributed to preceding Siberian snow cover anomalies and El Nino–Southern Oscillation (ENSO). Observed evidence in this study shows that sea ice anomalies on Barents–Kara (BK) and Laptev Seas and those on Chukchi–Beaufort (CB) Sea in autumn have co-mingled effects on the two abnormal EAWM modes. If negative sea ice anomalies appear in the BK and northern CB Sea in September–October (SO), the following winter SAT anomaly over East Asia is featured by an obvious cooling north of 40° N, which highly resembles the SAT N-mode of EAWM. If negative sea ice anomalies occur in BK–Laptev Sea but positive anomalies in the southern CB Sea, the East Asia winter SAT anomaly exhibits a salient cold condition over the south of 40° N and corresponds to the EAWM S-mode. The above results indicate that autumn Arctic sea ice can provide another predictability source for the EAWM, besides Siberian snow cover and ENSO. To further verify its contribution to seasonal prediction of the EAWM, a series of physical–empirical models are established using the combinations of SO Arctic sea ice, autumn Siberian snow cover and ENSO. Hindcast experiment output shows when adding autumn Arctic sea ice into the predictors, the cross-validated prediction skill is significantly improved. The possible physical mechanisms on how preceding Arctic sea ice anomalies impact on the N- and S- modes are also discussed.