Synergistic use of DAIS bands to retrieve land surface emissivity and temperature

Land surface temperature is an important key for environmental studies like energy balances and climate models. As is well-known, to retrieve land surface temperature from remotely sensed data, land surface emissivities are needed due to the nondeterministic nature of the temperature/emissivity separation from thermal infrared measurements: if thermal radiation is measured in N spectral bands, there will be N+1 unknowns (N emissivities and a single surface temperature). In this paper, two methods to estimate land surface emissivity from the Digital Airborne Imaging Spectrometer (DAIS) sensor are considered: the temperature/emissivity separation method (TES) developed by Gillespie et al. (1998), in which thermal infrared data are needed, and the NDVI thresholds method (NDVI^THM) developed by Sobrino et al. (2001), in which visible and near-infrared data are also needed. Once the LSE has been estimated, LST can be retrieved using a single-channel or two-channel method using only one thermal channel or a combination of two thermal channels respectively. These methods usually use the atmospheric water vapor content as input data, sp a method based on the ratios between absorbent and transparent bands in the red and near-infrared region to estimate atmospheric water vapor is also presented. The final results obtained for the validation carried out over the Barrax test site (Albacete, Spain) in the framework of the DAISEX (DAIS experiment) campaigns supported by ESA show deviations of around 0.01 for LSE and deviations between 1 K and 1.5 K for LST.