SOLAR ZENITH ANGLE EFFECTS ON VEGETATION INDEXES IN TALLGRASS PRAIRIE

Nadir spectral reflectances for a range of grass-land conditions on the Konza prairie in Kansas were acquired in 1987 from June to October during the First International Satellite Land Surface Climatology Project (ISLSCP) Field Experiment, or FIFE. Measurements in the red (0.668 +/- 0.005-mu-m) and near-infrared (0.821 +/- 0.006-mu-m) were obtained at numerous solar zenith angles (SZAs) during the diurnal cycle. Reflectance factors were utilized for computation of two spectral vegetation indices (VIs) - the simple ratio (SR) and the normalized difference vegetation index (NDVI). The diurnal change in the narrow red and near-infrared bands and their VIs as a function of SZA, and the relationship of these spectral data to biophysical measurements, were examined throughout the growing season. The assumption of independence of VIs from illumination geometry was specifically addressed. Both the SR and the NDVI were significantly affected by SZA for many prairie conditions. SR was significantly related to more canopy variables over a wider SZA range (27-degrees-61-degrees) than was NDVI. For intermediate green LAI (0.5 < LAI < 2.0), both indices were at a minimum value at solar noon, increasing with SZA. Conversely, the higher biomass sites (LAI > 2.0) with especially tall grass exhibited either a maximum value at solar noon which decreased as SZA increased or a moderate to high constant VI. Low density canopies (LAI < 0.5) exhibited no change in their low VI values over the diurnal course. These qualitatively different responses precluded the use of a general "correction" to remove the observed SZA effects on the measured VI values. The strength of the relationships between VI and LAI also varied, depending on SZA, with lower predictive capability at the traditionally used "high sun" condition than at oblique sun angles (SZA > 40-degrees). The restriction of surface measurements to a standard acquisition SZA of 45-degrees produced a good correspondence of both the SR and NDVI to LAI (r2 = 0.69). The best predictive linear model for the SR was obtained at SZA = 45-degrees (r2 = 0.82) and was based on two canopy variables, the green LAI and the canopy standing green to standing dead dry weight ratio. Therefore, a standard SZA for data acquisitions at 45-degrees is recommended over the traditional "high sun" practice. The results suggest that the VIs measure the instantaneous, projected, and illuminated LAI canopy fraction relative to the illuminated substrate.