Operational sea surface temperature bias adjustment using AATSR data

Abstract

One objective of the sea surface temperature (SST) component of the Global Monitoring for the Environment and Security (GMES) MERSEA project was to define the combined use of all SST products in an optimal interpolation scheme. Most SST products are now available in a common format but show regional differences between datasets that must be removed before use. We have defined a simple method using the SST derived from the Advanced Along Track Scanning Radiometer (AATSR) onboard the European Space Agency (ESA) ENVISAT platform as a common reference for all SST products. This method is based on an analysis of the differences between a given dataset and the AATSR SST over the 10 preceding days and over a 5° regular grid. These space–time characteristics have been defined empirically in a preliminary experiment. The efficiency of the method has been evaluated on a daily basis for a two year period (2007–2008) based on two SST products from the European Organisation for the Exploitation of Meteorological Satellites (EUMETSAT): the Spinning Enhanced Visible and Infra-Red Imager (SEVIRI) derived SST over the Atlantic and the Advanced Very High Resolution Radiometer (AVHRR) SST derived at fine scale over the North-East Atlantic. They have been compared to the AATSR and buoy SSTs and to the Naval Oceanographic Office (NAVOCEANO) AVHRR global SST product. The SEVIRI SST bias, which was reduced from −0.40K to −0.02K for the 0–20N latitude band by this approach provides a good example of the performance of the method. Reductions in the bias are also demonstrated for the North Atlantic, where cloudiness related problems were expected.