Estimation of sea ice drift and concentration during melt season using C-band dual-polarimetric Sentinel-1 data

Abstract

Characterizing sea ice changes over the Arctic can give proper insight into its extent in the current context of climate change. Sea ice motion's effect on ice area is influenced by seasonal conditions and the prevailing climate. However, the motion of thinner ice during the melt season can result in polynyas and cause significant ice loss that lasts the entire year. Sea ice drift and concentration during melt season are rarely discussed. The present study attempts to estimate sea ice drift during the melt season (July–August) using Differential Synthetic Aperture Radar Interferometry (DInSAR) with the usability of C-band Sentinel-1 data. The displacements/drifts were estimated using the phase shifts between two complex images with a temporal extent of 12 days. The study also proposed estimating sea ice concentration using polarimetric descriptors (power, purity, and randomness). The results revealed drifts of about 14–15 cm in the northeastern and southern parts followed by 45–50 cm in the middle parts and 20–25 cm along the coasts and open areas. Based on three polarimetric descriptors, thin ice and other ice fragments over open areas have shown less sea ice concentration as compared to the more concentrated pancake and pack ice with more compactness over the ice crystals. The study can be useful for computing drift in the melt season and carrying out spatiotemporal time-series analysis of seasonal sea ice drift/deformation studies. The algorithm and results can be used as a base study for the upcoming L- and S-band carrying NASA-ISRO Synthetic Aperture Radar (NISAR) mission.