Conversion of Satellite-Measured Radiances to Fluxes: Accounting for Reflection Anisotropy

Abstract

The Earth is an anisotropic reflector of sunlight, and this anisotropy varies considerably from one surface to another. A knowledge of the anisotropy is needed to accurately determine surface physical properties from satellite remote sensing. It is also necessary for accurate conversion of satellite-observed radiances to fluxes, which are used in energy budget studies, and are especially important for observations and predictions of climate change. This thesis examines the empirical determination of anisotropic factors, and uses a seven year archive of Advanced Very High Resolution Radiometer (AVHRR) data to construct them. The mean rms error of all instantaneous fluxes found with the derived factors is 3.28%, dependent on surface type. This represents a significant improvement on previous results, due mainly to the reduction of errors in determining the viewed scene type. A theoretical model of reflection is used to find anisotropic functions from the factors, removing the error incurred by using angular bins. The functions are used in studies of both albedo and cloud-radiative forcing over the New Zealand region, investigating the dependence of the forcing on terrain, surface temperature, cloudiness and season. It is shown that the high values of cloudiness over New Zealand has a marked effect on the radiation budget, and it is inferred that clouds cool the region.