Snow Mapping and Classification from Landsat Thematic Mapper Data

Abstract

Use of satellite multi-spectral remote-sensing data to map snow and estimate snow characteristics over remote and inaccessible areas requires that we distinguish snow from other surface cover and from clouds, and compensate for the effects of the atmosphere and rugged terrain. Because our space-borne radiometers typically measure reflectance in a few wavelength bands, for climate modeling we must use inferences of snow grain-size and contaminant amount to estimate snow albedo throughout the solar spectrum. Although digital elevation data may be used to simulate typical conditions for a satellite image, precise registration of an elevation data set with satellite data is usually impossible. Instead, an atmospheric model simulates combinations of Thematic Mapper (TM) band radiances for snow of various grain-sizes and contaminant amounts. These can be recognized in TM images and snow can automatically be distinguished from other surfaces and classified into clean new snow, older metamorphosed snow, or snow mixed with vegetation.