Canopy directional emissivity: Comparison between models

Abstract

Land surface temperature plays an important role in many environmental studies, as for example the estimation of heat fluxes and evapotranspiration. In order to obtain accurate values of land surface temperature, atmospheric, emissivity and angular effects should be corrected. This paper focuses on the analysis of the angular variation of canopy emissivity, which is an important variable that has to be known to correct surface radiances and obtain surface temperatures. Emissivity is also involved in the atmospheric corrections since it appears in the reflected downwelling atmospheric term. For this purpose, five different methods for simulating directional canopy emissivity have been analyzed and compared. The five methods are composed of two geometrical models, developed by Sobrino et al. [J. A. Sobrino, V. Caselles, & F. Becker (1990). Significance of the remotely sensed thermal infrared measurements obtained over a citrus orchard. ISPRS Photogrammetric Engineering and Remote Sensing 44, 343–354] and Snyder and Wan [W. C. Snyder & Z. Wan, (1998). BRDF models to predict spectral reflectance and emissivity in the thermal infrared. IEEE Transactions on Geoscience and Remote Sensing 36, 214–225], in which the vegetation is considered as an opaque medium, and three are based on radiative transfer models, developed by François et al. [C. François, C. Ottlé, & L. Prévot (1997). Analytical parametrisation of canopy emissivity and directional radiance in the thermal infrared: Application on the retrieval of soil and foliage temperatures using two directional measurements. International Journal of Remote Sensing 12, 2587–2621], Snyder and Wan [W. C. Snyder & Z. Wan (1998). BRDF models to predict spectral reflectance and emissivity in the thermal infrared. IEEE Transactions on Geoscience and Remote Sensing 36, 214–225.] and Verhoef et al. [W. Verhoef, Q. Xiao, L. Jia, & Z. Su (submitted for publication). Extension of SAIL to a 4-component optical–thermal radiative transfer model simulating thermodynamically heterogenous canopies. IEEE Transactions on Geoscience and Remote Sensing], in which the vegetation is considered as a turbid medium. Over surfaces with sparse and low vegetation cover, high angular variations of canopy emissivity are obtained, with differences between at-nadir view and 80° of 0.03. Over fully vegetated surfaces angular effects on emissivity are negligible when radiative transfer models are applied, so in these situations the angular variations on emissivity are not critical on the retrieved land surface temperature from remote sensing data. Angular variations on emissivity are lower when the emissivity of the soil and the emissivity of the vegetation are closer. All the models considered assume Lambertian behaviour for the soil and the leaves. This assumption is also discussed, showing a different behaviour of directional canopy emissivity when a non-Lambertian soil is considered.