Soil Loss Prediction in a Geographic Information System Format,
Abstract
Soil loss due to erosion from rainfall was accurately predicted for the Santa Paula 7.5 minute quadrangle, Ventura County, California, utilizing the VICAR/IBIS image processing and geographic information system to simulate the Universal Soil Loss Equation (USLE)). This work was part of a NASA funded research project investigating methods of incorporating collateral information in Landsat classification and modelling procedures (NSG-2377), performed at the University of California, Santa Barbara. Representing the rainfall, soil erodability, length of slope, slope gradient, crop management and soil loss tolerance coefficients of the USLE were data planes generated from digital Landsat data, USGS Digital Elevation Model topographic data, a digitized NOAA isopluvial map and digitized USDA soil conservation service soil maps. The Pearson product moment correlation coefficient, R, of soil loss predicted from the developed geobased model to a sample of manually derived soil losses was .91 after a log transform, a significant to the .0001 level. Estimates of accuracy for the intermediate data planes representing the rainfall, soil erodability, length of slope, slope gradient, crop management and soil loss tolerance ranged from a correlation coefficient, R, of .81 for the length of slope to 100 percent for the rainfall coefficient. The soil loss information system accurately targeted soil loss problem areas for subsequent analysis by Soil Conservation Service personnel. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- Jun 01, 1982
- Accession Number
- ADP001993
Entities
People
- A. H. Strahler
- J. E. Estes
- M. A. Spanner
Organizations
- Ames Research Center