A Universal BRDF Archetype Database for Normalizing Directional Reflectance to Nadir BRDF Adjusted Reflectance

Abstract

Surface reflectance varies with changes in solar and viewing angles, exhibiting distinct anisotropic reflectance characteristics, which can be characterized using the bidirectional reflectance distribution function (BRDF). Normalizing reflectance from various directions to a common solar-viewing geometry can eliminate the effects of reflectance anisotropy, thereby enhancing the accuracy of surface parameter retrieval from reflectance data. In previous studies, the reflectance angular normalization typically relies on prior knowledge extracted from coarse-resolution BRDF products. In this study, we developed a universal BRDF archetype database and explored the applicability of each BRDF archetype across various solar and viewing angles. The results indicate that the following: 1) The 3 × 3 universal BRDF archetype database is sufficient to represent the surface anisotropy of different spectral bands. 2) The optimal BRDF archetype in the lookup table (LUT) undergoes notable variations under different solar and viewing geometries, and specific prior BRDF knowledge is applicable only to certain solar and viewing geometries. 3) Validations based on simulated MODerate resolution imaging spectroradiometer reflectance suggests that in the red and near-infrared (NIR) bands, when the observation zenith angle is 60°, the percentages of directions with the root mean square error (RMSE) values below 0.05 and 0.06 constitute 82.74% and 75.70% of the entire viewing hemisphere, respectively. And validations based on polarization and directionality of the Earth's reflectance observations indicate that after normalization with the BRDF archetype LUT, the RMSE decreases from 0.0291 to 0.0249. The BRDF archetype LUT is applicable to the vast majority of solar and viewing geometries, even for the large solar zenith angles. Additionally, the LUT demonstrates good noise resistance and high applicability in observations at different scales, particularly in the NIR band. This study provides a new approach for reflectance angular normalization, offering the potential to provide a more reliable data foundation for future scientific research and practical applications.

Keywords

• Kernel-driven bidirectional reflectance distribution function (BRDF) model
• reflectance angular normalization
• reflectance anisotropy
• universal BRDF archetype