Since 1980, the Canada Centre for Remote Sensing has undertaken work to determine the extent to which LANDSAT Multispectral Scanner (MSS) data can be used to assist managers of the rangelands of
Western Canada. Previous studies have focussed on the qualitative classification of range condition based on the tones of the enhanced satellite imagery. The aim of the present study is to evaluate the feasibility of monitoring the change in
rangeland condition using quantitative digital analysis techniques. 12 cloud-free fall images covering a 20,000 km2 area of southeastern Alberta were acquired between 1973 and 1988. A procedure previously developed at CCRS to convert
digital counts to reflectance values was applied to the individual scenes. This procedure corrects for sun angle and atmospheric effects, thus making possible between-date comparisons. The images were registered to the Universal Transverse Mercator
(UTM) projection and resampled to produce 50 m X 50 m pixels. MSS band 5 difference images were calculated for selected periods spanning the data coverage. The images showed year-to-year changes in range condition and land use. Land use
changes, particularly the conversion of rangeland into cropland or seeded pasture, were the dominant changes in the area. Significant differences in the level of range use between years were also observed in most difference images. A standard
deviation image product was then prepared in which the pixel value represents the standard deviation of band 5 for all the years. This product highlighted specific areas where land cover condition varied substantially over the 15 year period.
|