PHYSICAL-MECHANISMS IN HYPERSPECTRAL BRDF DATA OF GRASS AND WATERCRESS

A hyperspectral bidirectional reflectance data set of grass lawn and w atercress canopies acquired under controlled laboratory conditions at the JRC European Goniometric Facility, at Ispra, Italy was analyzed wi th respect to basic physical reflectance mechanisms and factors influe ncing reflectance anisotropy. By normalizing reflectance data with the nadir reflectance as a reference, a strong influence of reflectance i ntensity on observed reflectance an isotropy is demonstrated for both vegetation canopies. This is explained by multiple scattering effects inside vegetation canopies as a function of canopy spectral absorbance characteristics. Two quantities based on canopy reflectances are util ized for characterizing the spectrally variable dynamics of BRDF effec ts: the anisotropy factor (ANIF), which is simply a normalization with nadir reflectance, and the anisotropy index (ANIX) defined as the rat io between the maximum and minimum reflectance values In the principal plane (or defined azimuth plane) per spectral band. Using these two q uantities the study revealed that the basic physical reflectance mecha nisms described by Kimes (1983) for broad spectral bands are extendabl e to hyperspectral bidirectional reflectance data. In addition, the in fluences of canopy geometry and multiple scattering on BRDF are clearl y demonstrated for the erectophile grass lawn and the planophile water cress canopy. Since multiple scattering effects and canopy geometry ar e dominant factors regulating BRDF effects, vegetation canopy architec ture parameters such as the leaf area index might be derived from hype rspectral BRDF data. (C) Elsevier Science Inc., 1998.