Ludwig, Valentin; Spreen, Gunnar; Pedersen, Leif Toudal (2021): 1 km sea-ice concentration from Sentinel-2 reflectances in the Arctic marginal seas between February and May 2019 [dataset]. PANGAEA, https://doi.org/10.1594/PANGAEA.933913

This dataset comprises quality-checked sea-ice concentration at 1 km spatial resolution from Sentinel-2 imagery. 83 Sentinel-2 scenes distributed over the East Siberian, Laptev, Kara, Barents and Beaufort Seas as well as the Fram Strait between February and May 2019 are analysed. For each scene, we investigated histograms of the Level 1C (L1C) reflectance of band 4 (665 nm) at 10 m resolution and manually identified two thresholds for each scene to separate it into water, thin ice and thick ice. We use the classified water/thin-ice/thick-ice images to create two sea-ice concentration datasets at 1 km spatial resolution: One which contains the thin ice (sic_thin in the netCDF files) and one which contains the thick ice (sic_thick in the netCDF files). To this end, we average the images over 100x100 pixels and interpret the ratio of sea-ice pixels within these 100x100 pixel windows as sea-ice concentration. The total sea-ice concentration can be obtained as the sum of sic_thin and sic_thick. The ice thickness of thin-ice pixels is likely below 50 cm. The data are gridded using a Transverse Mercator projection. Details about the projection can be found in the netCDF metadata. Details about the dataset are presented in https://doi.org/10.3390/rs12193183.