National Oceanic and Atmospheric Administration (NOAA) is generating an operational multi-satellite daily global analysis sea surface temperature (SST), blending both operational polar and geostationary satellite SST products. The current SST inputs to the analysis are from the following platforms: GOES-11, -12 (4 km), NOAA-18, and MetOp-A (4 km). The generation of SST analysis field employs a recursive estimator which emulates the Kalman Filter and uses data-adaptive correlation length scales to provide a reasonable balance between noise reduction and detail preservation. The final product is a daily analysis SST field at 11 km spatial resolution, along with uncertainty estimates for each observational grid point. Comparison with Real Time Global (RTG) high resolution SST analysis shows a standard deviation (SD) of 0.45 K, and with Reynolds ¼° daily OI has a SD of 0.65 K. Evaluation against drifting buoy data results in a SD of ±0.47 K. This multi-satellite SST analysis meets the needs of the user community by clearly identifying the mesoscale oceanographic features, improving high seas forecasts, and providing information about the state of the coastal ecosystems.

The future operational global analysis will be generated at an improved spatial resolution of 5 km, which will include the following additional SST datasets: MetOp-A AVHRR (1 km), MTSAT-1R (4 km), Meteosat-9 SEVIRI (4 km), EOS AMSR-E (37 km), and Envisat AATSR (1 km). It will also take into account the estimates of diurnal warming. Inclusion of microwave data (AMSR-E) is deemed to improve the spatial coverage and the AATSR SST will provide lower-biased SSTs as a reference dataset in the generation of the SST analysis field.

This recursive estimation algorithm will be applied to update the current operational polar analysis and, owing to its generic nature, it will be used for generating analysis SST in the NPOESS and GOES-R era.