The application of Landsat-1 imagery for monitoring strip mines in the New River Watershed in northeast Tennessee
The objectives of this study were to determine the accuracy of Landsat-1 imagery for measuring strip mines of the size common in east Tennessee and to examine the possibility of detecting and mapping strip mine subcategories. The New River Drainage Basin in northeast Tennessee was chosen as the general study area because of the significant impact of surface mining activity in the region. Seven sample disturbed-land areas were selected for detailed analysis.
Landsat imagery and supplementary aircraft photography of the test areas were obtained and submitted to a multilevel analysis using conventional photo interpretation methods, densitometric techniques, multispectral analysis, and statistical testing. The zoom transfer scope was essential for the mapping and classification of the data. Using the analog/digital system, the color and density characteristics of the Landsat imagery were tested. Imagery enhancement was accomplished by use of the multispectral viewer.
The seven sample areas were mapped and tested for accuracy. The Landsat area measurements were compared with low altitude area measurements because State-published measurements do not include total disturbed areas. Contour strips with widths as little as 67 meters (220 feet) were mapped. The average accuracy over all the mined-land sample areas mapped from the Landsat-1 image was 90 percent. Some of the differences in area are real since some of the mines were operative during this period and others revegetating.
The discrimination of strip mine subcategories is somewhat limited on Landsat-1 imagery. A mine site, whether active or inactive, can be inferred by lack of vegetation, shape, and image texture. Mine ponds are difficult or impossible to detect due to their small size and turbidity. Unless bordered and contrasted with vegetation, haulage roads are impossible to delineate. And preparation plants and refuse areas are not detectable. Density slicing of Landsat band 7 proved most useful in the detection of reclamation progress within the mined areas.
At present, it appears that for most state requirements for around-the-year, surface mined-land monitoring, Landsat is of limited value. However, for periodic, e.g., yearly updating of regional surface mine maps, Landsat may provide sufficient accuracies for some users. with appropriate baseline data, Landsat can provide information rapidly and economically for updating progress of new mining and reclamation of mined land.
Thesis77.S437.pdf
6.53 MB
Adobe PDF
044fe53554590d17c091ad03f0b55ddf