Mapping short-wave albedo of agricultural surfaces using airborne PolDER data

Abstract

This study focuses on albedo mapping over agricultural surfaces using multidirectional and multispectral remote sensing data. These data were acquired using the airborne PolDER sensor during the Remote Sensing Data Assimilation (ReSeDA) experiment. The data set allowed to perform a validation over the growth cycles of several crops. Problems induced by mixed pixels were reduced since the ground spatial resolution was 20 m. First, linear kernel-driven bidirectional reflectance distribution function (BRDF) models were used to retrieve the BRDF and then to compute the hemispherical reflectance in the PolDER channels. We tested the four most classical models: Li-Ross, MRPV, Roujean, and Walthall. They presented similar interpolation performances, whereas the quality of the hemispherical reflectance estimates was also driven by the extrapolation performances. Second, the albedo was computed as a linear combination of the waveband hemispherical reflectances. We used several sets of coefficients proposed in the literature for different sensors. The validation of the albedo maps against field measurements showed that it was possible to achieve a relative accuracy about 9% when using an appropriate coefficient set.