The Adjusted Normalized Emissivity Method (ANEM) for land surface temperature and emissivity recovery

In this work we propose a methodology for the retrieval of land surface temperatures and emissivities from thermal infrared data. It is based on the combination of the Vegetation Cover Method (VCM) and the Normalized Emissivity Method (NEM). NEM uses a first guess for emissivity, ε_NEM, which is constant for all channels and pixels. The method provides accurate relative emissivity variations, but the accuracy of absolute emissivity depends on the closeness of ε_NEM to the actual maximum emissivity value. The concept in the new methodology is to adjust ε_NEM initially by estimating channel emissivities in a pixel-by-pixel basis accounting for the spatial variation of emissivity with vegetation cover. This estimation can be done with the VCM, using coefficients adapted to each band. ε_NEM is taken as the maximum emissivity value obtained for each pixel, which is then introduced into the NEM algorithm, ANEM was applied to a series of scenes recorded by the Digital Airborne Imaging Spectrometer (DAIS), which was flown over the test site of Barrax, Spain, during the DAISEX campaigns in 1998-2000. Retrieved temperatures and emissivities were validated with ground measurements carried out coincidentally to the DAIS flights. The results showed good agreement between retrieved and measured values, and demonstrated the capability of the method to achieve nearly unbiased temperatures with uncertainties within ±0.8 K, and emissivities with errors mostly within ±0.01.