Anisotropic reflectance of snow observed from space over the arctic and its effect on solar energy balance

Advanced Very High Resolution Radiometer (AVHRR) images of arctic snow reflectance, observed from different satellite passes with the closest pass times on the same day, show different patterns that are related to the sun-satellite geometry. The variation of snow reflectance with view angle can be large or small, depending on the view geometry. Thus, the snow reflectances of a common target, observed from different viewing angles, can be very different and hard to compare. Unfortunately, none represent the actual physical snow albedo, and the error in the converted broadband top of atmosphere (TOA) albedo could easily be larger than 10%, which implies a great effect on the solar radiation balance at TOA. For an accurate estimate of the solar energy budget from satellite observations, and anisotropic correction is required not only for those reflectance images with obvious angular variation, but also for seemingly "isotropic" scenes because this latter case might incorrectly estimate the albedo under general conditions. Two methods, on based on a radiative transfer model and the other on National Oceanic and Atmospheric Administration's (NOAA) Earth Radiation Budget (ERB) experimental data, have been applied to AVHRR data to implement and anisotropic correction. The results show that both methods can remove the systematic variation of snow reflectance with view angle and can significantly reduce the large observed differences in reflectances from different satellite observations of a common snow target. The radiative transfer model-based method, however, unlike the NOAA experimental method, corrects the reflectances in the two AVHRR channels separately, has higher angular resolution than the NOAA experimental method, and can account for variations in snow condition (e.g., grain size). Moreover, the model-based anisotropic correction shows superior self-consistency compared to that produced by NOAA's experimental-based method. Finally, the model-based method provides a sound theoretical basis for anisotropic correction of satellite images of arctic snow reflectance. (C) Elsevier Science Inc., 2001.