Variability in Leaf and Litter Optical Properties: Implications for BRDF Model Inversions Using AVHRR, MODIS, and MISR

Canopy radiative transfer models simulate the bidirecmodel inversions by decreasing the number of observations required to retrieve canopy structural and biophysitional reflectance distribution function (BRDF) of vegetation covers with differing leaf and soil spectral and cancal information from multiangle remotely sensed data. Elsevier Science Inc., 1998 opy structural characteristics. Numerical inversion of these models has provided estimates of vegetation structural and biophysical characteristics from multiangle, remotely INTRODUCTION sensed optical data. The number of angularly unique observations compared to BRDF model parameters largely The spectral and angular dependence of photons reflected off a surface is governed by the bidirectional reflectance determines the accuracy of retrievals. To increase this ratio, additional observations of a target must be acquired distribution function (BRDF). In the specific case of vegetation canopies, this reflectance distribution is anisotropic and the BRDF models and inversions must be simplified. The former will occur when the EOS instruments become and primarily a function of leaf optical properties, canopy architecture, soil surface attributes, illumination conavailable. Previous studies suggest that simplification of BRDF model inversions may best be accomplished by ditions, and viewing geometry (Ross, 1981; Goel, 1988; Shultis and Myneni, 1988; Myneni et al., 1988; Jacqueconstraining the leaf optical parameters. This study focused on full-range (400–2500 nm) leaf and litter spectral moud et al., 1992). The spectral attributes of a plant canproperties convolved to AVHRR, MODIS, and MISR opopy are linked to scattering processes at the leaf level. tical channels. Using a diverse array of woody plant and Leaf-level scattering varies with leaf structure, water grass species, we found robust and readily usable interrecontent, and the concentration of carbon constituents lationships among spectra through correlation, regres(e.g., lignin, cellulose, starch), chlorophyll, and nitrogen sion, and principal components analyses. Significant dif(Gates et al., 1965; Thomas et al., 1971; Wooley, 1971; ferences between green leaf and litter optical properties Walter-Shea and Norman, 1991; Fourty et al., 1996; Jacand their sensor-specific interrelationships indicate that quemoud et al., 1996). Although the angular variation of green leaf optical constraints may be useful with BRDF scattered photons at the canopy level is tied to leaf and retrievals to detect the onset of canopy senescence. These soil spectral properties, it is highly dependent on canopy findings will provide increased efficiency in canopy BRDF structural characteristics such as leaf area index and leaf angle distribution (Goel, 1988; Myneni and Asrar, 1993). Plant canopy radiative transfer (RT) models have ad*Center for the Study of Earth from Space/CIRES and Departvanced from simple light extinction algorithms to those ment of Environmental, Population and Organismic Biology, University based explicitly on scattering theory using turbid medium of Colorado, Boulder and turbid-medium/geometric-optical methods (e.g., Goel, †National Center for Atmospheric Research, Boulder ‡Department of Rangeland Ecology and Management, Texas A&M 1988; Myneni et al., 1989; Li et al., 1995). Since these University, College Station models are based on scattering processes, they provide a Address correspondence to G. P. Asner CIRES/CSES, Campus valuable tool for quantifying the light regime within canBox 216, University of Colorado, Boulder, CO 80309-0216. Received 19 January 1997; revised 22 August 1997. opies and for simulating the scattering of photons from