The Arctic Oscillation (AO) is the dominant mode of Northern Hemisphere (NH) winter climate variability. An accurate prediction of the AO could lead to improved seasonal temperature forecasts over the major population centers in eastern North America and Europe as it explains the largest fraction of temperature variance of any other climate mode over these regions. In previous studies, a relationship between mean October snow cover extent over Eurasia and the phase of the wintertime AO has been documented. Enhanced snow cover over Eurasia during October stimulates the development and westward expansion of the Siberian high, which subsequently produces anomalous upward fluxes of wave energy into the stratosphere that lead to warming and a weakening of the polar vortex. Circulation anomalies of the same sign in the troposphere tend to follow this weakening of the stratospheric vortex leading to a negative phase of the AO in the troposphere.

Snow indices that consider only monthly mean snow cover extent over Eurasia have a moderate, but statistically significant correlation with the DJF AO. We have developed a snow advance index (SAI) that measures the mean daily rate of increase in snow cover over interior Eurasia (south of 60ºN) during the month of October. This new index is strongly correlated (r = 0.86) with the DJF AO. Anomaly correlations of DJF surface temperature and geopotential heights with the AO and SAI have remarkably similar patterns. Although this index is based on the limited number of years (1997-2010) for which daily snow cover data exists, the high correlation between the SAI and the DJF AO suggests that the wintertime AO is indeed predictable, rather than a by-product of the stochastic behavior of internal atmospheric dynamics.