Retrieval of Surface Albedo Based on BRDF Model

Abstract

Land Surface Albedo is an important surface parameter and is widely applied to the surface energy balance, mid-term and long-term weather forecast and atmospheric general circulation model. GF-4 satellite is the first geostationary orbit satellite which combines high spatial resolution and high temporal resolution in China. In order to explore the feasibility of semi-empirical kernel-driven BRDF model applied on GF-4 satellite data, the earth's surface classification is joined to give kernel factors initial value, and Powell iteration algorithm is used to optimize the result of model. Then the land surface narrow band albedos of each band could be gained through angle integration on BRDF model. On this basis, combining spectral library with spectral response function of GF-4 satellite, the conversions of narrow to broadband albedo for GF-4 satellite data is built for the first time. And albedo inversion in short wave band (0.3-3µm) are acquired. Finally, cross validation used by MODIS albedo product indicates that an accurate land surface albedo could be acquired by this method.