Is there a dominant timescale of natural climate variability in the Arctic?

A frequency-domain singular value decomposition performed jointly on centur y-long (1903-94) records of North Atlantic sector sea ice concentration and sea level pressure poleward of 40 degreesN reveals that fluctuations on th e interdecadal and quasi-decadal timescales account for a large fraction of the natural climate variability in the Arctic. Four dominant signals, with periods of about 6-7, 9-10, 16-20, and 30-50 yr, are isolated and analyzed . These signals account for about 60%-70% of the variance in their respecti ve frequency bands. All of them appear in the monthly (year-round) data. Ho wever, the 9-10-yr oscillation especially stands out as a winter phenomenon . Ice variability in the Greenland, Barents, and Labrador Seas is then linked to coherent atmospheric variations and certain oceanic processes. The Gree nland Sea ice variability is largely due to fluctuations in ice export thro ugh Fram Strait and to the local wind forcing during winter. It is proposed that variability in the Fram Strait ice export depends on three different mechanisms, which are associated with different timescales: 1) wind-driven motion of anomalous volumes of ice from the East Siberian Sea out of the Ar ctic (6-7-yr timescale); 2) enhanced ice motion forced by winter wind anoma lies when they align parallel to the Transpolar Drift Stream (9-10-yr times cale); 3) wind-driven motion of old, thick, and very low salinity ice from offshore northern Canada into the outflow region (16-20-yr timescale). Also , a marked decreasing trend in ice extent since around 1970 (30-50-yr times cale) is linked to a recently reported warming in the Arctic. The Barents Sea ice variability is associated with the nature of the penetr ation of Atlantic waters into the Arctic Basin, which is affected by two di stinct mechanisms: 1) changes in the intensity of the northward-flowing Nor wegian Current, which is linked to variability in the North Atlantic oscill ation (NAO) pattern (9-10-yr timescale); and 2) changes in the upper-ocean temperature of the Norwegian Current waters, which is likely related to the advection of temperature anomalies by the ocean gyres (16-20-yr timescale) . Ice variability in the Labrador Sea, on the other hand, appears to be mai nly determined by thermodynamical effects produced by the local wind forcin g., which is closely related to the NAO pattern (9-10-yr timescale), and by oceanic advection of ice anomalies into this sea from the Greenland-Irming er Sea by the East Greenland Current (6-7-yr timescale).