EFFECTS OF STANDING LITTER ON THE BIOPHYSICAL INTERPRETATION OF PLANTCANOPIES WITH SPECTRAL INDEXES

Litter is frequently present within vegetation canopies and thus contr ibutes to the overall spectral response of a canopy. Consequently, lit ter will affect spectral indices designed to be sensitive to green veg etation, soil brightness or other features. The main objectives of the current research were to 1) evaluate the spectral properties of green vegetation and litter and 2) quantify the effect of standing litter o n the performance of spectral indices. The SAIL (scattering by arbitra rily inclined leaves) model teas used to generate canopy reflectance ' 'mixtures'' and to estimate fractions of absorbed photosynthetically a ctive radiation (fAPAR) with varying leaf area index (LAI), soil backg round, combinations of vegetation, component spectral properties, and one or two horizontal vegetation layers. Spectral measurements of diff erent bare soils and mature green and senescent leaves of representati ve plant species at the HAPEX-Sahel (Hydrological Atmospheric Pilot Ex periment) study sites were used as input. The normalized difference ve getation index (NDVI), the soil adjusted vegetation index (SAVI), and the modified NDVI (MNDVI) and mixture model spectral indices were sele cted to evaluate their performance with respect to standing litter and green vegetation mixtures. Spectral reflectance signatures of leaf li tter varied significantly, but strongly resembled soil spectral charac teristics. The biophysical parameters (LAI, fAPAR), derived from spect ral vegetation indices, tended to be overestimated for randomly distri buted, sparse green and fitter vegetation cover mixtures, and underest imated for randomly distributed dense green and litter vegetation cove r mixtures. All spectral indices and their biophysical interpretation were significantly altered by variability in 1) green leaf, leaf litte r, and bark optical properties, 2) the amount and position of standing leaf litter, 3) leaf angle distribution, and 4) soil background. The NDVI response to these variables was inconsistent, and was the most af fected by litter. The spectral mixture model indices, designed to be s ensitive to litter were shown to be promising for the identification o f litter present among different ecosystems.