A comparative study on three EOF analysis techniques using decades of Arctic sea-ice concentration data

Change in Arctic sea ice extent is one of the indicators of global climate changes. Spatio-temporal change and change patterns can be identified using various methods to facilitate human understanding global climate changes. Three empirical orthogonal function (EOF) techniques are discussed and applied to decades of sea-ice concentration (SIC) dataset in Arctic area for identifying independent patterns. It was found that: 1) discrepancies exist in magnitude and scope for each EOF pattern, however, the first two leading EOFs of variability possess high similarities in structure and shape; 2) Even though there are somewhat differences in amplitude of each PC mode, the first two leading PC modes maintain consistent in overall trend and periodicity; 3) There are significant discrepancies and inconsistencies in the third and fourth leading EOF and PC modes. The accuracies of three techniques are further validated in representing the physical phenomena of SIC anomaly patterns.