The importance of soil conservation and the devastating effects of soil erosion have been noted throughout recorded human history. In the United States it has been estimated that up to two-thirds of the nation's cropland is losing soil at rates up to 5 tons/A (11 mt/ha) while nearly 15% is losing more than 10 ton/A (22 mt/ha) per year. The Soil Conservation Service of the USDA and the Environmental Protection Agency are actively engaged in programs to reduce soil loss and subsequent stream pollution. Rapid methods of pinpointing areas undergoing accelerated erosion are needed to assist in these efforts.

The use of remote sensing techniques such as conventional black and white photography has been utilized in soil survey for many years. In recent years Landsat (formerly ERTS) satellite data have been used in soil mapping for both large scale and more detailed mapping. Digital analysis techniques developed at LARS have been used to produce spectral maps with a scale as large as 1:15840 which are used to assist in mapping soils in the field.

The objective of this study was to produce a ground cover classification map using computer-implemented analysis of Landsat multispectral scanner data and to evaluate the usefulness of this map in delineating severely eroded soil areas under cultivated conditions.

Date of this Version