The Impact of Assimilating Sea Surface Salinity Observations in a High Resolution Ocean Prediction System

Remotely sensed sea surface temperature (SST) observations collected from a myriad of different satellites make up the backbone of many of our operational ocean prediction systems. Depending on the region and cloud cover, we routinely will have analyses that have complete coverage of SST observations. What if in addition to SSTs, we also had global coverage of sea surface salinity (SSS) observations? How would that affect our ocean prediction systems? For this presentation, a series of observing system simulation experiments (OSSEs) were performed in an attempt to answer these questions.

Over the past several years there has been a lot of interest in remotely sensed SSS observations; i.e. SMOS, SMAP, and Aquarius. Observations from these satellites, however, have relatively high errors and therefore are not typically used in operational systems. There is an ongoing effort to improve the accuracy of remotely sensed SSS, and to have satellites that can provide global SSS coverage. The OSSEs that were set up to test this possible scenario assumed a SSS observing satellite in the same orbit as VIIRS. These experiments were performed: 1) on a very high resolution grid (1 km) to see if SSS obs help improve the resolution of submesoscale features, 2) in the North Arabian Sea which can exhibit strong salinity gradients, and 3) with both 3DVAR and 4DVAR assimilation solvers to see if 4DVAR is better able to propagate SSS observation information into the interior of the domain.