Angular effect on surface temperature estimation from AATSR data

The estimation of sea and land surface temperature using multi-angular algorithms with Advanced Along Track Scanning Radiometer (AATSR) data require an accurate knowledge of the angular variation of surface emissivity in the thermal infrared. A range of operative and accurate multiangle algorithms for estimating surface temperature has been developed using MODTRAN 3.5 simulations of the AATSR data at 11 and 12 μm wavelengths, making use of the dual-angle viewing capability of the AATSR. To build the simulated database, emissivity values have been used. An experimental investigation has been made in order to determine the angular variation of the infrared emissivity as a well-known parameter to estimate ST in multiangular methods. The results show a general decrease of the emissivity with increasing viewing angles. Finally, the proposed algorithms have been applied to a series of satellite images acquired by the CSIRO in Australia, retrieving LST with a standard deviation less than 0.6 K if the satellite data are error free; in this case, the emissivity and temperature values necessary to retrieve the ST have been acquired from CSIRO in situ data.