Laboratoire des Écoulements Géophysiques et Industriels

Nos tutelles


Nos partenaires


Accueil > Équipes > Equipe MEIGE : Modélisation, Expériences et Instrumentation pour la Géophysique et l’Environnement > Diffusion scientifique > Séminaires internes

Turbulent oceanic western-boundary layers at low latitude

Cataria Quam Cyrille AKUETEVI, 02/05/2013

Low latitude oceanic western-boundary layers range within the most turbulent regions in the worlds ocean. The
Somali current system with the Great Whirl and the Brazilian current system with its eddy shedding are the most
prominent examples. Results from analytical calculations and integration of a one layer reduced-gravity fine reso-
lution shallow water model is used to entangle this turbulent dynamics. Two types of wind-forcing are applied : a
remote Trade wind forcing with maximum shear along the equator and a local Monsoon wind forcing with maxi-
mum shear in the vicinity of the boundary. For high values of the viscosity (> 1000m2 s−1 ) the stationary solutions
compare well to analytical predictions using Munk and inertial layer theory. When lowering the friction parameter
time dependence results. The onset of instability is strongly influenced by inertial effects. The unstable boundary
current proceeds as a succession of anti-cyclonic coherent eddies performing a chaotic dynamics in a turbulent
flow. The dynamics is governed by the turbulent fluxes of mass and momentum. We determine these fluxes by
analyzing the (potential) vorticity dynamics. We demonstrate that the boundary-layer can be separated in four sub-
layers, which are (starting from the boundary) : (1) the viscous sub-layer (2) the turbulent buffer-layer (3) the layer
containing the coherent structures and (4) the extended boundary layer. The characteristics of each sub-layer and
the corresponding turbulent fluxes are determined, as are the dependence on latitude and the type of forcing. A
new pragmatic method of determining the eddy viscosity, based on Munk-layer theory, is proposed. Results are
compared to observations and solutions of the multi-level primitive equation model (DRAKKAR).