Dominant patterns of winter Arctic surface wind variability

Abstract

Dominant statistical patterns of winter Arctic surface wind (WASW) variability and their impacts on Arctic sea icemotion are investigated using the complex vector empirical orthogonal function (CVEOF) method. The results indicate that theleading CVEOF of Arctic surface wind variability, which accounts for 33% of the covariance, is characterized by two differentand alternating spatial patterns (WASWP1 and WASWP2). Both WASWP1 and WASWP2 show strong interannual and decadalvariations, superposed on their declining trends over past decades. Atmospheric circulation anomalies associated with WASWP1and WASWP2 exhibit, respectively, equivalent barotropic and some baroclinic characteristics, differing from the Arctic dipoleanomaly and the seesaw structure anomaly between the Barents Sea and the Beaufort Sea. On decadal time scales, the declinetrend of WASWP2 can be attributed to persistent warming of sea surface temperature in the Greenland—Barents—Kara seas fromautumn to winter, reflecting the effect of the Arctic warming. The second CVEOF, which accounts for 18% of the covariance, alsocontains two different spatial patterns (WASWP3 and WASWP4). Their time evolutions are significantly correlated with the NorthAtlantic Oscillation (NAO) index and the central Arctic Pattern, respectively, measured by the leading EOF of winter sea levelpressure (SLP) north of 70°N. Thus, winter anomalous surface wind pattern associated with the NAO is not the most importantsurface wind pattern. WASWP3 and WASWP4 primarily reflect natural variability of winter surface wind and neither exhibits anapparent trend that differs from WASWP1 or WASWP2. These dominant surface wind patterns strongly influence Arctic sea icemotion and sea ice exchange between the western and eastern Arctic. Furthermore, the Fram Strait sea ice volume flux is onlysignificantly correlated with WASWP3. The results demonstrate that surface and geostrophic winds are not interchangeable interms of describing wind field variability over the Arctic Ocean. The results have important implications for understanding andinvestigating Arctic sea ice variations: Dominant patterns of Arctic surface wind variability, rather than simply whether thereare the Arctic dipole anomaly and the Arctic Oscillation (or NAO), effectively affect the spatial distribution of Arctic sea iceanomalies.