Assessing Sentinel-1 Imagery for Sea Ice Monitoring in the Denmark Strait

Abstract

Widely-available Synthetic Aperture Radar (SAR) imaging from the Sentinel-1 constellation offers opportunities for sea ice mapping and monitoring. In order to establish the viability of independent classification of new ice formation areas in the Marginal Ice Zone (MIZ) where sea ice presents maritime hazards, an independent classification was done to manually identify and delineate areas of new sea ice formation as conditions cool in autumn and sea ice pack extends into the Denmark Strait. A second analysis incorporated the use of climate data to revisit digitized new ice areas and determine analysis confidence. The analyses were performed for the three autumns that Sentinel-1 has now been available: 2014, 2015, and 2016. Obtained areas were used to run an optimized hot spot analysis to determine areas of significant clustering, perform a spatial analysis to identify patterns, and examine climate preconditions for optimal new ice formation. Autumn 2014 and 2015 showed a similar spatial pattern, while in 2016 ice did not appear in the study area until a month later than normal. 2014 was the year of greatest new formation in the MIZ in the Denmark Strait, followed by 2015, and distantly trailed by 2016. While imagery was very useful for interpretation, the interpretation was much stronger with the addition of corresponding climate conditions which should be considered when developing automated classifications based on Sentinel-1 images.Keywords: Arctic, sea ice, Denmark Strait, Marginal ice zone, Synthetic Aperture Radar (SAR), Sentinel-1, Mapping new ice, climate preconditioning, Optimized hot spot analysis.