Land Surface Temperature (LST) is a good indicator of the energy balance at the Earth/Atmosphere interface. It is therefore an important parameter for many applications of Numerical Weather Prediction.

There are several techniques in the literature for retrieval of LST from thermal infrared remote sensing platforms. These techniques must account for atmospheric absorption, emission and scattering and for reflection of atmospheric radiation from the surface. This problem is ill-conditioned as there are more unknown parameters than data. The atmospheric effects may be approximated by use of radiative transfer models on standard atmospheres. The surface reflection requires knowledge of surface emissivity. There exists in the literature a range of different land surface type atlases with associated surface emissivity classification which are used to make the problem determined. Techniques to determine emissivity using empirical relations between radiances in different bands are also common.

The Met Office operates the ARIES Fourier transform interferometer. This instrument makes measurements in the spectral range 3.3 - 16 microns and has been flown at low level over a range of surface types: sea ice; snow covered forest and crops; senescent broadleaf forest; arid soils and agricultural crops. Coincident measurements with other radiometers and with video cameras allows high resolution land classification. We use these data to evaluate some of the LST retrieval schemes and to validate some of the land surface atlases available.