Determining rangeland brush canopy densities with Landsat MSS data

Abstract

Landsat multispectral scanner digital data was evaluated for use in determining brush canopy densities. A Landsat digital data simulation model was utilized to determine vegetation index sensitivity to changes in brush canopy density, soil background type and amount, and grass cover. The Landsat digital count simulation model was used to determine the sensitivity of the vegetation indices of S6, S7, ND6, ND7, Gr, Br, DD, GP, and GBa to changes in soil type. It was determined that of the vegetation indices simulated, the Kauth and Thomas Greeness index (Gr) was the least sensitive to change in soil type. Several combinations of bare soil amount, grass canopy cover, and brush canopy cover were simulated for three soil types. Here also the Gr index appears to be the most sensitive to scene greeness factors and least affected by soil type changes. A validation of this Landsat digital count simulation model for predicting rangeland Landsat vegetation index values was attempted. With the atmospheric and environmental data available to this investigator, the model did not accurately simulate rangeland Landsat scene data. A positive and statistically significant relationship was found to exist between aerial photographic estimates of brush canopy cover and several Landsat vegetation indices. The vegetation index which was the most highly correlated with brush canopy cover was GB (Kauth and Thomas Greeness/Kauth and Thomas Brightness, r('2) = .66). Any of several vegetation indices, including Greeness (Gr), could be used to predict brush canopy density on this site.Results of this study indicate that brush canopy over density can be quantified through the use of Landsat MSS data. The Greeness index (Gr) appears to be the least affected by changes in soil background and thus is the most likely candidate for use in a large area brush canopy monitoring program.