FY-3E WindRAD Arctic Sea Ice-Water Classification Products (2021–2024)

This dataset provides daily Arctic sea ice–water classification products derived from the FengYun-3E (FY-3E) WindRAD dual-frequency (C- and Ku-band) scatterometer using a Random Forest classifier trained with high-resolution Sentinel-1 SAR labels. Three model configurations are included:- WRADC: C-band only- WRADKu: Ku-band only- WRADDual: Dual-frequency fusion (recommended primary product) Each daily file contains binary classification maps (sea_ice_water: 0=open water, 1=sea ice, 255=fill/land) at 10 km spatial resolution on the NSIDC Sea Ice Polar Stereographic North grid (EPSG:3413), covering the Arctic region (>60°N) from August 2021 to April 2024. Geographic coordinates (latitude/longitude) and the land mask are provided in a separate static file (FY3E_WindRAD_GridInfo_10km_PSG_V1.h5) to avoid redundant storage. The land mask was extracted from the original ice_edge variable where value=3 indicates land. For full methodological details, validation statistics, and usage examples, please refer to the README.txt file included in this repository and the associated paper (Zhai et al., 2026, GIScience & Remote Sensing). Related identifiers:- Dataset DOI: 10.5281/zenodo.21872089- Input FY-3E WindRAD L1 data: National Satellite Meteorological Center (NSMC), CMA- SAR training labels: Wang & Li (2021), ESSD, https://doi.org/10.5194/essd-13-2723-2021