Fluctuations and seasonality in the Arctic sea ice area: A sudden regime shift in 2007?

Abstract: Since the beginning of satellite observations, the Arctic sea ice extent has shown a downward trend. The decline has been weaker in the March maximum than in the September minimum and masked by inter-annual fluctuations. One of the less understood aspects of the sea ice response is the persistence times for fluctuations, which could indicate the dominant physical processes behind the sea ice decline. To determine the fluctuation persistence times, however, it is necessary to first filter out the dominant effect of the seasonal cycle. In the current study, we thus develop a statistical model, which accurately decomposes the ice area changes into: (1) a variable seasonal cycle component with a constant shape and (2) a residual (short term) fluctuation. We find the persistence time of fluctuations to be only about three weeks, independently from season, which is substantially shorter than previously reported. Such short time scale points to the dominance of atmospheric forcing. The shape of the seasonal cycle is surprisingly constant for the whole observational record despite the rapid decline. This is in agreement with the suggestion that the asymmetry of the seasonal cycle is an effect of Arctic land-sea geography, which has not changed with climate change. The analysis suggest a jump in the annual sea ice area amplitude occurring in 2007, from which it has not yet recovered, possibly revealing a permanent amplitude shift. In physical sense, this could imply a shift towards the younger, thinner and more susceptible ice cover commencing after the immense 2007 multi-year ice loss.