Multi-channel and multi-angle algorithms for estimating sea and land surface temperature with ATSR data

Abstract A study has been carried out using LOWTRAN-7 simulations of the Along-Track Scanning Radiometer (ATSR) data at 11 and 12 μm wavelengths to compare the merits of the multi-angle technique with those of the currently used multi-channel technique (split-window method) to retrieve both sea surface temperature (SST) and land surface temperature (LST). To this end a simple single-channel double-angle viewing model is presented, which relates actual surface temperature to the two brightness temperatures measured from space in the two views of interest (ATSR nadir and forward). Subsequently, statistical retrieval coefficients for the double angle and split-window techniques are derived via a regression analysis of the synthetic measurements. The results show that the double angle technique is capable of producing SSTs with a standard deviation of 0.23 deg K if the satellite data are error free and, furthermore, confirm the advantage of the double-viewing angle technique in comparison with the split-window technique for LST determination in homogeneous surfaces if the emissivity's spectral variation and the emissivity's angular variation, are of the same order of magnitude. Finally we present the preliminary results obtained using the proposed model from ATSR data over a semi-arid region of New South Wales, Australia provided by Prata, and over the Pacific Ocean provided by Barton.