CANDIDATE HIGH-SPECTRAL-RESOLUTION INFRARED INDEXES FOR CROP COVER

The sensitivity of near-infrared / red ratio vegetation indices to soil reflectance and plant color result in ambiguous interpretation of plant condition and productivity. Measurements of the reflectance of crop canopies were made at high spectral resolution to investigate candidate vegetation indices, in the near- and middle-infrared (800-2500 nm), for their ability to unambiguously estimate foliage cover independently of the extraneous effects of variations in canopy color and soil background brightness. Reflectances were measured with an IRIS spectroradiometer over plots of sugar beet (Beta vulgaris L.) sown on different dates and at different densities to produce a wide range in canopy cover. Vegetation color was varied by infecting selected plots with sugar beet yellows virus. Soil brightness was varied by placing trays of peat between the plant rows. Selected narrow-band near- and middle-infrared reflectances were tested for their relationship with canopy cover and their sensitivity to variations in canopy color and soil brightness. The relationship between canopy cover and traditional near-infrared / red ratio indices was found to be sensitive to both canopy color and soil brightness. Most infrared indices were insensitive to the effects of canopy color, but those showing the highest correlations with cover tended to be significantly influenced by soil brightness. The most promising NIR reflectances were those beyond the range (760-900 nm) of near-infrared reflectance detected by current space-borne systems, such as Landsat TM and SPOT HRV.