ARCTIC SEA-ICE VARIABILITY ON A TIMESCALE OF WEEKS AND ITS RELATION TO ATMOSPHERIC FORCING

Northern Hemisphere sea ice concentration, 500-hPa height, sea level pressure, and 1000-500-hPa thickness at 7-day intervals are examined for the period 1972-1989, with emphasis on the winter season. The temporal variability of sea ice concentration is largest along the climatological mean ice edge where its frequency distribution is strongly bimodal with ice-free and ice-covered conditions being observed much more frequently than partial ice cover. These results confirm impressions, based on visual inspection of satellite imagery, that most of the variability in these regions is associated with the advance and retreat of the ice edge. Relationships between large-scale patterns of atmospheric variability and sea ice variability are investigated, making use of singular value decomposition of the temporal covariance matrix. The analysis is conducted separately for the Atlantic and Pacific sectors. In agreement with earlier studies based upon monthly mean data on sea ice concentration, the strongest sea ice pattern is comprised of a dipole with opposing centers of action in the Davis Straits/Labrador Sea region and the Greenland and Barents seas. Its temporal variability is strongly coupled to the atmospheric North Atlantic oscillation (NAG). The relationship between the two patterns is strongest with the atmosphere leading the ocean by two weeks. An analogous dipole pattern is observed in the Pacific sector, with opposing centers of action in the Bening Sea and the Sea of Okhotsk, which is related to a distinctive pattern of atmospheric circulation anomalies in the Pacific sector. One polarity of the NAO and its Pacific counterpart is associated with blocking episodes, during which the influence of the atmosphere is strong enough to temporarily halt the climatological mean advance of the ice edge in some regions and substantially accelerate it in others. The relationships between the fields is indicative of local forcing of sea ice in most regions, with wind stress and thermodynamic fluxes at the air-sea interface both contributing. A possible exception is the Greenland Sea, where it may be necessary to invoke some form of remote forcing in order to explain the observed changes on the interannual time scale.