Cross-platform application of a sea ice classification method considering incidence angle dependency of backscatter intensity and its use in detecting deformed ice

Abstract:

Wide-swath C-band Synthetic-aperture Radar (SAR) has been used for sea ice classification and navigational ice charts ever since it first became widely available in 1990s. Still, automatic classifications only became available recently, with development of methods that take into account incidence angle dependency of SAR backscatter intensity. In this study we adapt one such novel sea ice classification method developed for Sentinel-1 data, and test its performance in classifying sea ice using Radarsat-2 data. Since the distribution of pressure ridges is of high climatological and navigational importance, we target the classification scheme to distinguish between level ice and deformed ice. We examine the classification results against ground truth data collected over the Norwegian Young sea ICE campaign (N-ICE2015) including Airborne Laser Scanner data, and also sea ice deformation products derived from Sentinel-1 data. Our initial validation results with manually drawn polygons of different ice types give similarly encouraging performance for both SAR platforms. We also demonstrate that wide-swath SAR with nominal resolution of 40-50 m has the capacity to distinguish between relatively smooth level ice and aggregations of deformed ice and large pressure ridges along shear zones and linear kinematic features. We expect that further developments of our method will provide reliable sea ice classifications of level and deformed ice based only on widely available SAR data and at least from the launch of Radarsat-2 onward (2000).