Predictive Power of the Emissivity Angular Variation of Soils in the Thermal Infrared (8-14 μm) Region by Two Mie-Based Emissivity Theoretical Models

A confident knowledge of land surface emissivity at viewing zenith angles far from nadir is of prime interest to perform an accurate correction of the anisotropy effect in the measurements made by orbiting thermal infrared (TIR) sensors. It is also important for the correct treatment of angular measurements carried out by remote sensors such as AATSR/ENVISAT, MODIS/Terra-Aqua, or the recently launched SLSTR/Sentinel-3, which can also be used for the angular normalization of land surface temperature due to viewing geometry effect. In this letter, the anisotropy of TIR emissivity predicted by two analytical, Warren-Wiscombe-Dozier and Hapke, models based on Mie diffraction theory was compared with field-measured values under dry conditions. The results showed good agreement between models and measurements (RMSEs ranging from +/- 0.004 to +/- 0.030 depending on the sample, with an average value of +/- 0.016) if the compactness of the grains soil is taken into account, demonstrating the good performance of the cosine term of the zenith angle included in the expressions of both models. A significant disagreement between models and measurements is, however, obtained for some samples at high zenith angles, suggesting the necessity of a fudge factor to include in the compactness correction in that condition.