Variability and trends in the Arctic Sea ice cover: Results from different techniques

Variability and trend studies of sea ice in the Arctic have been conducted using products derived from the same raw passive microwave data but by different groups using different algorithms. This study provides consistency assessment of four of the leading products, namely, Goddard Bootstrap (SB2), Goddard NASA Team (NT1), EUMETSAT Ocean and Sea Ice Satellite Application Facility (OSI-SAF 1.2), and Hadley HadISST 2.2 data in evaluating variability and trends in the Arctic sea ice cover. All four provide generally similar ice patterns but significant disagreements in ice concentration distributions especially in the marginal ice zone and adjacent regions in winter and meltponded areas in summer. The discrepancies are primarily due to different ways the four techniques account for occurrences of new ice and meltponding. However, results show that the different products generally provide consistent and similar representation of the state of the Arctic sea ice cover. Hadley and NT1 data usually provide the highest and lowest monthly ice extents, respectively. The Hadley data also show the lowest trends in ice extent and ice area at -3.88%/decade and -4.37%/decade, respectively, compared to an average of -4.36%/decade and -4.57%/decade for all four. Trend maps also show similar spatial distribution for all four with the largest negative trends occurring at the Kara/Barents Sea and Beaufort Sea regions, where sea ice has been retreating the fastest. The good agreement of the trends especially with updated data provides strong confidence in the quantification of the rate of decline in the Arctic sea ice cover. Plain Language Summary The declining Arctic sea ice cover, especially in the summer, has been the center of attention in recent years. Reports on the sea ice cover have been provided by different institutions using basically the same set of satellite data but different techniques for estimating key parameters such as ice concentration, ice extent, and ice area. In this study, a comparison of results from four different techniques that are frequently used shows significant disagreements in the characterization of the distribution of the sea ice cover primarily in areas that have a large fraction of new ice cover or significant amount of surface melt. However, the actual changes in the ice cover are consistently depicted and the trends in sea ice extent and ice area from the different data sets are practically the same providing strong confidence that satellite data are interpreted consistently by different scientists independently and confirming that the ice extent of the Arctic perennial ice is indeed declining at the rate of about 11% per decade. The results provide useful information for modelers, policy makers, and the general scientific public.