Using wavelet analysis to examine four-dimensional changes in sea ice and relating these to atmospheric patterns

Issues of climatic change in the Polar Regions have prompted concern regarding the melting of sea ice. To examine the processes that occur between the atmosphere-ice-ocean interface, one is often forced to involve the interactions between the atmosphere/ocean and a region of ice cover or development. Therefore a climatic or meteorological signal must be analysed for trends and compared with a temporally changing area of sea ice – a task that can be both daunting and time consuming.

The following study outlines one method for qualitatively examining the physical processes between ice and atmosphere. Temporally changing sea ice, on a large spatial scale, can be investigated for overall changing patterns using passive microwave satellite imagery. Satellite imagery can provide excellent spatial and temporal coverage of the Polar Regions if spatial resolution is not a concern. Sea ice changes on various scales can be determined using principal components analysis (PCA) with derivation of leading modes. Finally a smoothing of the loading pattern of sea ice can be achieved by using a two-dimensional wavelet transform. PCA allows the user to become familiar with both temporal and spatial variability related to decreasing variance, whilst the wavelet transform filters the higher frequency signal from the lower frequency pattern. The same wavelet technique can be performed on atmospheric data, in this case the leading modes of an empirical orthogonal function analysis of an atmospheric teleconnection pattern. Therefore interactions between sea ice and atmosphere and potential feedbacks can be qualitatively observed through a correlation of filtered patterns.