Type of Report : Semi - Annual , July - December 1993 Alfredo

Several soiland atmospheric correcting variants of the normalized difference vegetation index (NDVI) have been proposed to improve its accuracy in estimating biophysical plant parameters. In this study, a sensitivity analyses, utilizing simulated and observational data, was conducted on the NDVI and improved variants by analyzing the atmospheric-and soilperturbed responses of each VI as a continuous function of two vegetation parameters (LAI and %green cover). Absolute and percent relative errors, and vegetation equivalent noises (VEN) were calculated for soil and atmospheric influences, separately and combined. The NDVI variants included the soil-adjusted vegetation index (SAVI), the atmospherically resistant vegetation index (ARVI), the soil and atmospherically resistant vegetation index (SARVI), the modified SAVI (MSAVI), and modified SARVI (MSARVI). =20 Soil and atmospheric error were of similar magnitudes, but varied with the vegetation index. All new variants outperformed the NDVI. It was found that the atmospherically resistant versions minimized atmospheric noise, but enhanced soil noise while the soil adjusted variants minimized soil noise, but remained sensitive to the atmosphere. The SARVI, which had both a soil and atmosphere calibration term, performed the best with an absolute error of +/0.06, relative error of 10%, and VEN of +/-0.33 LAI. By contrast, the NDVI had absolute and relative errors of 0.12, 20%, and 0.82 LAI, respectively. =20 ____________________________________________________________ The structure of the SARVI equation may be written as: SARVI = [(ρ*nir ρ*rb) / (ρ*nir + ρ*rb + L)] (1 + L), ρ*rb = ρ*r γ (ρ*b ρ*r). The 'L' term provides feedback for soil background noise removal and the g term in conjunction with the blue band minimizes atmospheric sources of noise attributed to varying aerosol contents. The SARVI thus requires correction (ρ*) for atmospheric gases, namely Rayleigh scattering and ozone absorption. Although the SARVI greatly reduced soil and atmospheric influences, relative to the NDVI, they were not both reduced is a systematic, predictable manner. This is because the SARVI minimizes atmospheric and soil noise through independent means without considering combined or 'coupled', interactive effects of soil and atmosphere (see Figs. 1 & 2).