Potential for Semiarid Community Type Differentiation

Abstract Community-level vegetation type differentiation is regarded as an important application of remote sensing in monitoring the health of semiarid environments and yet it has proved difficult to discriminate more than two or three community types with confidence when using coarse spatial resolution data from off-nadir scanning sensors such as the AVHRR, even when multi-temporal datasets are used. Attempts to relate vegetation phenology to community type using AVHRR data have produced mixed results and large classification errors have been attributed to subpixel scale topographc and soil background variations, lack of in-flight calibration, inaccurate reflectance retrieval and poor registration. While these are important factors, anisotropy in surface reflectance as a result of the bidirectional reflectance distribution function (BRDF) also causes severe perturbations in the signal and yet this is rarely accounted for adequately. Recent years have seen a wider acknowledgement of this fundamental aspect of optical remote sensing, although it is usually approached as a problem. The extent of the improvement in community type differentiation may be considerable when a BRDF model is used to separate isotropic and anisotropic components. Here this potential is tested using data from the AVHRR over semi-arid regions.