Overview:
SOSE (pronounced saucy) is a Southern Ocean State Estimate. A modern general circulation model, the MITgcm, is least squares fit to all available ocean observations. This is accomplished iteratively through the adjoint method. The result is a physically realistic estimate of the ocean state. The estimation period for SOSE is 2005 through 2006. Analysis of this state estimate can be used to help understand the biology, chemistry, climate, and dynamics of the Southern Ocean. SOSE is being produced by Matthew Mazloff under the primary guidance of Carl Wunsch and Patrick Heimbach as part of the ECCO-GODAE project and the ECCO2 project. Computational support for the work comes from SDSC.
SOSE specifications:
The model domain is from latitude 78 South to 24 South. The horizontal resolution is 1/6 of a degree. There are 42 levels in the vertical. First guess initial conditions and open boundary conditions are derived from a coarse resolution global state estimate produced by Gael Forget. SOSE uses an atmospheric boundary layer scheme and is coupled to a sea-ice model. It employs the KPP and GM-Redi parameterizations. The first guess atmospheric state is from NCEP reanalysis. The adjoint method systematically perturbs the atmospheric state and initial conditions, within their uncertainty, to find a model ocean state consistent with the observations. SOSE is being carried out on SDSC's DataStar supercomputer.
Observations:
Satellite altimeters: Jason and GEOSAT follow on (GFO) from PODAAC. ENVISAT from AVISO
Mean dynamic ocean topography from the GRACE Project (Tapley et al. 2005)
Sea-ice concentration from NSIDC
In situ temperature and salinity profiles: ARGO from IFREMER. CTD from various sources, including WOCE HP. SEaOS from M. Meredith. XBT from D. Behringer (NCEP)
SST: Reynold's and TMI
Hydrographic climatology from Gouretski and Koltermann (2004) and NOAA World Ocean Atlas 2001. Conkright et al., 2002
For 2005 and 2006 there are about 15 million altimetry observations and about 5 million in-situ observations. When we include the other observations we have on the order of 1 billion terms in our optimization.
Access:
Anyone may use the data. Contact Matt Mazloff (mmazloff at mit.edu) to get what you need. This may include requests for SOSE state (U,V,W,S,T,P,eta...), or requests for corrected atmospheric state, or requests for sea-ice cover, or requests about how well the model fits the observations, or requests for the observations used, or requests for information on how we do it....or whatever...we are not trying to hide anything...but we would like to know what you intend to use the information for.
More info:
SOSE press:
SOSE links:
Ryan Abernathey is studying eddy stirring across the ACC using SOSE: (effective diffusivity)