The USGS Landsat Snow Covered Area Science Data Products

Abstract:

The Landsat Thematic Mapper (TM), Enhanced Thematic Mapper + (ETM+), and Operational Land Imager (OLI) together provide more than 30 years of 30 m resolution multi-spectral image data well suited for snow cover mapping. The Landsat data record includes regular 16-day acquisitions for the conterminous US and a few other regions since 1984 and regular 16-day acquisitions for most other land areas since 2013. These data can provide unique information regarding the spatial and temporal distribution of snow cover, particularly in mountainous environments where substantial snow cover heterogeneity is common at scales finer than 1 km. The USGS Landsat snow covered area science data product maps per-pixel snow cover fraction, often referred to as fractional snow covered area (fSCA), and will be available on-demand for nearly all Landsat scenes acquired since 1984. Fractional snow covered area is computed from Landsat surface reflectance data using TMSCAG, a spectral unmixing approach that also provides vegetation and rock/soil fractions for each pixel. In addition to the standard, sensor-viewable fSCA product, vegetation fraction is also used to produce a canopy-adjusted fSCA product. Early validation results indicate good agreement between Landsat-derived fSCA and fSCA computed from high spatial resolution imagery as well as from in situ sensor arrays. Data from sensor arrays also indicate the canopy adjustment approach substantially improves agreement between Landsat-derived fSCA and in situ fSCA in areas with forest cover. A comprehensive validation effort currently underway will provide a more complete assessment of product accuracy across a wide range of snow conditions and land cover types. An alternative cloud mask provided also facilitates the use of Landsat snow covered area data in mountainous regions with mixtures of snow, rock, and vegetation where existing cloud masking approaches often result in false identification of cloud cover. The availability of the Landsat-derived snow covered area dataset presents the opportunity for new analysis of spatial and temporal snow cover variability at higher spatial resolutions and over a longer period of record than has previously been possible.