Spectral characteristics of selected soils and vegetation in Northern Nevada and their discrimination using band ratio techniques

Abstract

Characterizing arid region soils and vegetation conditions from remotely sensed imagery is limited by low interband and intraspectral reflectance contrast between soil and vegetation. This study has evaluated the spectral response of semiarid soils and vegetation and the utility of four calculated Landsat Thematic Mapper (TM) band ratios and band transformations for discriminating soil and vegetation. Ground-level reflectance spectra were taken of 62 soil and 236 vegetation surfaces. Mean reflectances were calculated for the equivalent TM Bands 1, 2, 3 and 4. All two-band ratios, the Normalized Difference Vegetation Index (NDVI), and the brightness, greenness, and yellowness transformation were calculated. Soil reflectance spectra are highly variable, yet predictable: They increased directly with wavelength over the visible-near infrared (NIR) region and have low interband contrast. Vegetation spectra are less predictable because various plant structures and phenology affect the spectral response and the visible-to-NIR reflectance contrast. Ratio techniques can separate most pure vegetation samples from pure soils, but the degree of separation varies with the technique. Ratioing is effective for surfaces with high interband spectral contrast but is not effective for surfaces with low contrast. The NIR/red and the NDVI ratios indicate an association with plant-available water gradients and the drought tolerance or drought-avoidance mechanisms.