Reducing Errors in Satellite-derived Sea Surface Temperature – Application to Aatsr Data Analyses

Operational multi-channel algorithms for deriving sea surface temperature (SST) from satellite data are derived using theoretical models of infrared radiation transfer through the atmosphere or by regression of satellite data with ground-based observations. In both cases the algorithm derivation implies a “first-guess” of the water vapour content along the atmospheric path between the satellite and the surface. Deviations from this first-guess water vapour profile may result in errors in the derived SST. The six channels of the AATSR (3 wavelengths x 2 views) account for much of this variation, but significant errors can still occur. These errors can be reduced by first using the AATSR data to derive some estimation of the water vapour content of the atmosphere, and then selecting SST algorithm coefficients from a pre-determined table. The derivation of water vapour content and distribution using AATSR data will be presented first followed by details of the two-step advanced approach to SST derivation.