Abstract
A land-cover-based linear BRDF (bi-directional reflectance distribution function) unmixing (LLBU) algorithm based on the kernel-driven model is proposed to combine the compact airborne spectrographic imager (CASI) reflectance with the moderate resolution imaging spectroradiometer (MODIS) daily reflectance product to derive the BRDF/albedo of the two sensors simultaneously in the foci experimental area (FEA) of the Heihe Watershed Allied Telemetry Experimental Research (HiWATER), which was carried out in the Heihe River basin, China. For each land cover type, an archetypal BRDF, which characterizes the shape of its anisotropic reflectance, is extracted by linearly unmixing from the MODIS reflectance with the assistance of a high-resolution classification map. The isotropic coefficients accounting for the differences within a class are derived from the CASI reflectance. The BRDF is finally determined by the archetypal BRDF and the corresponding isotropic coefficients. Direct comparisons of the cropland archetypal BRDF and CASI albedo with in situ measurements show good agreement. An indirect validation which compares retrieved BRDF/albedo with that of the MCD43A1 standard product issued by NASA and aggregated CASI albedo also suggests reasonable reliability. LLBU has potential to retrieve the high spatial resolution BRDF/albedo product for airborne and spaceborne sensors which have inadequate angular samplings. In addition, it can shorten the timescale for coarse spatial resolution product like MODIS.
Highlights
Techniques to improve the understanding and quantifying of land surface reflectance anisotropy provide a critical foundation for deriving land surface parameters in earth system and in eco-hydro scientific research
The LLBU method presented in this paper, combines the spatial information from high-spatial-resolution data to characterize the within-class bidirectional reflectance distribution function (BRDF) variation, and the multi-angular information from coarse-spatial-resolution data to derivate the land-cover-specific archetypal BRDF
Its unmixing process allows anisotropic reflectance features to be extracted without identifying pure pixels on the coarse scale
Summary
Techniques to improve the understanding and quantifying of land surface reflectance anisotropy provide a critical foundation for deriving land surface parameters in earth system and in eco-hydro scientific research. One option for using airborne data for BRDF estimation is pixel-based (PB) BRDF fitting, requireing sufficient reflectance observation angles which can be acquired by sensors like airborne cloud absorption radiometer (CAR) [16,17,18] and airborne research scanning polarimeter (RSP) [19]. The CASI instrument was a push-broom imaging spectro-radiometer that measured the spectrum of each location on the ground in 48 spectral bands between 380 and 1055 nm It was flown at a relative altitude of 2 km above the surface and had a field of view (FOV) of 40° with 1500 across-track imaging pixels and a one-meter spatial resolution. The narrow range of NDVI indicates a similar structure (growing stage) of the cropland
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
