Dominant patterns of sea ice variability around Antarctica

. A frequency-domain decomposition is performed on three different sets of sea ice concentration data, spatially gridded around Antarctica from 50S to the South Pole. The method used for the analysis is the Multi Taper Method - Singular Value Decomposition (MTM-SVD), a multivariate frequency-domain decomposition technique recently developed by M. Mann and J. Park. This technique seeks to identify statistically significant narrow-band oscillations that are correlated among a large number of independent time series (grid points). During the last years, the MTM-SVD technique has been successfully applied by an increasing number of investigators on different datasets and regions of the world. This includes a very recent analysis of sea ice variability in the Arctic Ocean and marginal seas.

The data analyzed in this work are obtained from three different sources (satellite, boats and buoys, model outputs) and the records span the last 22, 30 and 42 years respectively. The decomposition in the frequency domain allows for the detection of the dominant timescales of sea ice variability. The comparison between datasets yields a fairly robust picture of the spatial and temporal behavior of the sea ice extent around Antarctica during the last decades. The analyses reveal a generally clockwise propagation of sea ice anomalies around Antarctica with dominant variability on the interannual timescale. A seasonal breakdown of the data is also performed to test the sensibility of the results to the warm and cold seaso