Soil background effects on the spectral response of a three-component rangeland scene Academic Article uri icon

abstract

  • Landsat data were simulated for a three-component rangeland scene consisting of grass, brush, and soil to evaluate the effect of soil background reflectance on the sensitivity of ratio and orthogonal spectral indices to vegetation density. The simulations were done using soils with low, intermediate, and high reflectances for conditions representing periods of active growth in which brush and grass are green, and summer dormancy in which brush is green and grass is dormant. When soil reflectance was low, the sensitivities of the ratio vegetation index and the normalized difference decreased as the density of green vegetation increased. When soil reflectance was high, the sensitivity of these same indices increased as vegetation density increased. The sensitivity of greenness, calculated from published coefficients, was greater when soil reflectance was low, and differences due to soil reflectance decreased as canopy cover increased. The sensitivity of soil-specific greenness, calculated from coefficients derived from the simulated data, was greater when soil reflectance was high, and differences due to soil reflectance increased as canopy cover increased. Soil reflectance also affected the sensitivity of the indices to the density of dormant grass. The results suggest that discrimination of brush and vegetation may not be possible during active growth when both are green, but may succeed during summer dormancy when the grass is dormant. The results further suggest that even if soil-specific calibrations and soil lines are used, the accuracy of any estimates of rangeland vegetation density derived from the indices may vary with the soil background. © 1986.

published proceedings

  • Remote Sensing of Environment

author list (cited authors)

  • Heilman, J. L., & Boyd, W. E

citation count

  • 21

complete list of authors

  • Heilman, JL||Boyd, WE

publication date

  • April 1986