Operational relationships between NOAA-advanced very high resolution radiometer vegetation indices and daily fraction of absorbed photosynthetically active radiation, established for Sahelian vegetation canopies

摘要

abstract_textpTo improve the estimation of primary production ata-regional scale, an assessment of the utility of fraction of absorbed photosynthetically active radiation (fAPAR) estimated from spectral vegetation indices (VI) for the case of the Sahelian vegetation was studied. Simulations using a three-dimensional radiative transfer model were conducted for two types of structurally distinct vegetation canopies: millet crop (regularly distributed clumps of vegetation over bare soil) and savanna (mixture of a grass layer and a sparse woody layer). A realistic range of values is extracted for each vegetation input variable (leaf area index, ground cover, height and spatial distribution) from published literature. Bidirectional reflectance factors were calculated in the NOAA-advanced very high resolution radiometer (AVHRR) spectral bands for a geometric configuration representative of the NOAA satellite series. Two vegetation indices were tested : normalized difference vegetation index (NDVI) and modified soil adjusted vegetation index (MSAVI). The simulations indicate that the fAPAR-VI relationship is sensitive to the geometry of measurement and soil optical properties, especially in the case of a millet crop. Simple linear models that include this variability and are directly applicable to atmospherically corrected AVHRR data-are proposed, and the error of estimation of fAPAR is evaluated. MSAVI gives better results, especially when the brightness of soil is known. For natural vegetation the relationship between fAPAR and MSAVI is then reduced to a coefficient that is a constant (similar to 2.2). For millet canopies this coefficient is a function of Sun and view angles. Finally, the utility of our approach is illustrated with two examples of fAPAR-VI relationships applied to a NOAA time series./p/abstract_text