The Proposed Global Land Surface Temperature Product for ENVISAT’s AATSR: Scientific Basis, Algorithm Description and Validation Protocol

This paper provides an overview of, and rationale for, the definition of a multi-channel Land Surface Temperature (LST) data product from the AATSR. Land surface temperature (LST) is an extremely important parameter that controls the exchange of longwave radiation between the surface and atmosphere. Because of the extreme heterogeneity of most natural land surfaces, LST is a difficult parameter to estimate and to validate. Algorithms for deriving land surface temperature using split-window radiances are sufficiently advanced that accuracies of 1-3 K are possible. Better accuracies (about 1 K) are obtained at night, when differential surface heating is absent. Since surface spectral emissivities are not known to sufficient accuracy and not measured over large parts of the land surface, an LST algorithm that has no explicit dependence on surface emissivity is required. Angular effects on directional temperature and surface emissivity are also difficult to include, so the LST algorithm must rely on the nadir AATSR split-window radiances. Further research will investigate the incorporation of the forward view AATSR data into the LST algorithm.