Sun and view angle corrections on reflectances derived fromNOAA/AVHRR data

A new method is proposed for the reduction of the noise-like fluctuations associated with variations of Sun-target-sensor geometry in multitemporal AVHRR data in the visible and near-infrared bands. Its principle is to adjust, over a monthly period, a three-parameter model of surface bidirectional reflectance on a time series of cloud-free atmospherically corrected AVHRR data. One parameter of the model represents the surface reflectance corrected for angular effects. Time profiles of corrected reflectances are obtained by making the monthly period slide over the annual vegetation cycle. This procedure is applied to an annual cycle of AVHRR data on seven test sites in France representative of bare soils, agricultural crops, and forested thematic areas. The method is evaluated according to two criteria, which are the ability of the bidirectional reflectance model to reduce the amount of noise-like temporal fluctuations of AVHRR data, and the stability of the retrieved parameters when random noise is artificially added to the original AVHRR data set. The perturbations induced by the coupling between diffuse sky radiance effects and the non-Lambertian behavior of ground reflectance are also discussed. The method is proved to give satisfactory results, and can potentially be used to compare satellite-derived reflectances obtained not only at different times, but also at different places and with different sensors