You must have a javascript-enabled browser and javacript and stylesheets must be enabled to use some of the functions on this site.


Causes of large-scale sea level variations in the Southern Ocean: Analyses of sea level and a finite element barotropic model

Frederic Vivier(1) , Kathryn A. Kelly(2) , and Migel Harismendy(3)

(1) LOCEAN (ex-LODYC), CNRS, Université Pierre et Marie Curie, T 45-55 Etage 4, 75005 Paris, France
(2) APL, University of Washington, 1013 NE 40th St, Seattle, WA 98195, United States
(3) LOCEAN (ex-LODYC), Université Pierre et Marie Curie, T 45-55 Etage 4, 75005 Paris, France


We analyze a decade of sea surface height (SSH) measurements in the Southern Ocean from the TOPEX/Poseidon and ERS altimeters, with a focus on the variability at timescales shorter than two years. Among the different processes contributing to large-scale SSH variations, the barotropic response to the winds dominates poleward of 50 S, while thermosteric processes dominate equatorward, except for resonant basins for the barotropic modes and regions of intense eddy activity. A finite element barotropic model has been developed to analyze the vorticity budget. The SSH from the model agrees well with observations. The leading barotropic mode, which is annular, is confined near Antarctica, and is responsible for most of the barotropic circumpolar transport. The barotropic transport associated with this mode is coherent with the zonally integrated eastward wind stress consistent with a free mode response. Although previously evidenced in bottom pressure data, this mode is only partially seen in altimeter data because of ice coverage. It nevertheless distinctly appears above the Pacific ridges where it expands meridionally up to midlatitudes. In the rest of the domain, several regions coherent with the local wind stress curl are found. These are regions isolated by f /H contours, mostly deep basins. An analysis of the vorticity budget shows that, generally, topographic Sverdrup balance is the leading process for periods larger than 50 days, but in some regions (resonant basins), diffusive and nonstationary terms are important. A model experiment shows that transients redistribute energy along f /H waveguides, contributing to drain resonant regions, as was hypothesized in previous works.


Workshop presentation

Full paper


                 Last modified: 07.10.03