LAUSR

The land surface thermal-infrared broadband emissivity (BBE) is a vital variable for estimating land surface radiation budgets (SRBs). We develop a framework for retrieving the 30 m BBE from Landsat surface reflectance data to estimate SRBs at finer scales and validate coarse resolution data. In the developed framework, the land surface is classified as bare soils and vegetated surfaces to allow different algorithms to be used for the BBE estimation. We propose a downscaling algorithm that uses the empirical relationship between the Advanced Spaceborne Thermal Emission and Reflectance Radiometer (ASTER) BBE and Landsat surface reflectance at 90 m to retrieve the 30 m BBE over bare soils. A look-up table (LUT)-based algorithm is proposed for vegetated surfaces. The BBE is interpolated from a LUT that is constructed from the 4SAIL radiative transfer model with inputs of the leaf BBE, the soil background BBE, and the leaf area index (LAI). Ground measurements that were collected at 11 relatively homogeneous sandy sites during three independent field campaigns are used to validate the proposed algorithm over bare soils. The average difference between the retrieved and field-measured BBEs is 0.012. We produce the land surface BBE of China in 2008 by using the developed framework and composited winter and summer seasonal BBE maps. The composited seasonal BBE maps are compared to the seasonal BBE maps derived from the ASTER emissivity product. The bias is within ±0.005 over bare soils and ranges from 0.012 to 0.019 over vegetated surfaces. Combined with the validated results in this study and published references, the comparison results demonstrate the good performance of the developed framework. This study provides a new perspective on estimating BBEs from sensors with only a thermal-infrared channel.