Inversion of the asymmetry factor for desert areas of China

Abstract

Deserts, which have high surface albedo and wide area, are important components of the earth system. It is very important for the research of surface radiation and energy balance to understand the anisotropic scattering of desert areas. The emergence and development of multi-angle remote sensing made possi]e the inversion of the anisotropic scattering of desert areas at the regional or global scale. Firstly, this paper explored the accuracy of the inversion of asymmetry factor using the Hapke model and the simulated single- and multi-phase MISR data. Based on the results, the asymmetry factor of representative surface of desert areas in northwestern China was retrieved. The values of the asymmetry factor retrieved from MISR data were compared with the values retrieved from the laboratory data. The results showed that the single-phase MISR data could be used for the inversion of asymmetry parameter of desert areas. The sign of the asymmetry parameter for the laboratory measurements was positive, which suggests that the surface of laboratory samples is forward scattering. The sign of the asymmetry parameter for MISR data was negative; that is, it is backscattering. The values of the asymmetry parameters retrieved from MISR data were related to the character of the land surface. At Loulan, where the surface was smoother than other sites, retrieved values exhibited the largest negative values of asymmetry factor, suggesting the strongest backscattering. The sand dune area of the Kumtag Desert, which has the greatest roughness, had only slightly negative asymmetry factor values. These findings indicated that at the sensor scale, a rough surface (e.g., dunes) does not necessarily mean more backscattering than a smooth surface. This finding has significant implications for empirical methods (e.g., using the normalized index of backward-scattered radiance minus forward-scattered radiance as an indicator of surface roughness), which should be used carefully for analyzing surface roughness from the remote sensing data.