The effect of surface anisotropy and viewing geometry on the estimation of NDVI from AVHRR

Publisher Website
Google Scholar
Abstract
Terrestrial surfaces are observed routinely from space using the AVHRR instrument on the NOAA satellites. Observations in the red and near‐infrared spectral bands are used to compute vegetation indices (such as the NDVI), which are then interpreted in terms of various vegetation properties. Orbital and engineering constraints severely limit the range of illumination and viewing angles under which various ecosystems are observed. Since terrestrial surfaces are anisotropic, all spectral reflectance measurements obtained with a small instantaneous field of view instrument are specific to these angular conditions, and the value of the corresponding NDVI, computed from these bidirectional reflectances, is relative to the particular geometry of illumination and viewing at the time of the measurement. This paper documents the importance of these geometric effects through simulations of the AVHRR data acquisition process, and investigates the systematic biases that result from the combination of ecosystemspecific anisotropies with instrument‐specific sampling capabilities. Typical errors in the value of NDVI are estimated, and strategies to reduce these effects are explored.