Refinements in the remote sensing of snow-covered area

Snow-covered area can be mapped in the visible bands by Landsat MSS, SPOT, and GOES systems. But the white snow is difficult to distinquish from white clouds and saline soils. Conifers and other vegetation can obscure the satellite view of a snowpack. There can also be discontinuous snow within a pixel for any of the available resolutions. Refinements in the operational techniques have been investigated. Adjusting image brightnesses to approximate vertical sun and flat terrain and then subtracting a summer reference image increases the detection of snow in mountainside conifers. Such change detection also seems appropriate for saline soils. Fractional snow coverage, as seen by Landsat, was compared with the brightness of snow in the coarse AVHRR pixels. Vegetative cover and snowpack aging make it inappropriate to adopt a simple transformation from snow brightness into percent snow-covered area with a pixel.