Blending Sea Surface Temperatures from Multiple Satellites and In Situ Observations for Coastal Oceans

Abstract

A two-dimensional variational data assimilation (2DVAR) method for blending sea surface temperature (SST) data from multiple observing platforms is presented. This method produces continuous fields and has the capability of blending multiple satellite and in situ observations. In addition, it allows specification of inhomogeneous and anisotropic background correlations, which are common features of coastal ocean flows. High-resolution (6 km in space and 6 h in time) blended SST fields for August 2003 are produced for a region off the California coast to demonstrate and evaluate the methodology. A comparison of these fields with independent observations showed root-mean-square errors of less than 1°C, comparable to the errors in conventional SST observations. The blended SST fields also clearly reveal the finescale spatial and temporal structures associated with coastal upwelling, demonstrating their utility in the analysis of finescale flows. With the high temporal resolution, the blended SST fields are also used to describe the diurnal cycle. Potential applications of this SST blending methodology in other coastal regions are discussed.