Computed and downscaled Sentinel-2 level land surface albedo over the SAIL region in the East River watershed, Colorado, from September 2021 through June 2023. Using Geostationary Operational Environmental Satellite-R (GOES-R), we progressively downscale geostationary land-surface albedo to MODIS-level (500 m) and Sentinel-2-level (20 m) grids. The dataset contains three principal collections: 1. Computed 20-meter Sentinel-2-level land surface albedo: High-resolution reference products derived from Sentinel-2 observations. These include black-sky albedo, white-sky albedo, blue-sky albedo, broadband shortwave, snow-focused products, alternative hard and soft masking outputs, and associated quality-assurance layers. 2. Downscaled 500-meter MODIS-level land surface albedo: U-Net predictions produced by downscaling the approximately 2-kilometer GOES-R albedo observations to a MODIS-level 500-meter grid. 3. Downscaled 20-meter Sentinel-2-level land surface albedo: U-Net predictions produced by further downscaling the MODIS-level albedo products to a Sentinel-2-level 20-meter grid.Keywords: Land surface albedo, U-Net, deep learning, downscaling, MODIS, Sentinel-2, GOES-R, remote sensing, East River Watershed, SAIL.
Items connected by shared entities, co-authorship, citations, or semantic similarity.