Le radar HF est actuellement le seul instrument courantométrique permettant d'obtenir une description synoptique à haute résolution spatiale et temporelle de la circulation côtière de surface. Un système radar déployé depuis 2010 en Méditerranée sur les côtes varoises offre pour la première fois une description exhaustive de la circulation, encore peu documentée dans cette zone.La cartographie des courants se fait classiquement en combinant les mesures d'au moins deux radars. Cependant des résultats significatifs ont été obtenus avec un seul radar concernant : l'identification de tourbillons méso-échelle ; la signature de phénomènes périodiques affectant la circulation superficielle dans les bandes diurne, inertielle et semi-diurne ; et les caractéristiques et les instabilités du Courant Nord Méditerranéen (CN).L'assimilation des mesures radar au moyen d'un lisseur de Kalman d'ensemble dans un modèle régional de la Méditerranée Nord Occidentale a été réalisée pour la première fois dans la zone d’étude. Cette méthode, qui contraint les courants de surface en optimisant le vent et les forçages aux frontières ouvertes, améliore la description de la veine du CN en vitesse et position / The HF radar is the sole instrument being able to monitor the surface coastal current at very high spatial and temporal resolution. A radar system deployed since 2010 on the Var coast (Western Mediterranean Sea) provides for the first time a comprehensive picture of the circulation, which remains poorly documented in this area.Surface current mapping is conventionally performed by combining measurements of at least two radars. However, significant results were obtained with a single radar concerning: the identification of mesoscale vortices; the signature of periodic phenomena affecting the surface circulation in the diurnal, inertial and semi-diurnal bands; and the features and instabilities of the North Mediterranean Current (NC).The assimilation of radar measurements using an ensemble Kalman smoother in a regional model of the North-Western Mediterranean Sea was performed for the first time in the study area. This method, which constraints the surface currents by wind and open boundary conditions optimisation, improves the simulation of the NC vein in terms of speed and position.

L'objectif général de cette thèse est de contribuer à l'avancement de la connaissance de l'impact de l'activité àméso-échelle du Courant Nord (CN) au large du Var sur sa circulation en aval et des interactions de ce courantde bord avec la dynamique côtière, particulièrement dans une baie semi-fermée peu profonde : la baie deHyères. Ces travaux se sont appuyés sur deux configurations numériques réalistes à haute résolution basées surle code NEMO et emboîtées avec AGRIF : une première de la façade méditerranéenne française à unerésolution spatiale de 1,2 km et une seconde le long des côtes varoises à 400 m de résolution.La comparaison des simulations obtenues avec des observations (radar HF, ADCP, glider, SST satellite) apermis de confirmer le réalisme des configurations, et de montrer l'apport d'une résolution de l'ordre de 400 msur la dynamique dans une baie mais également sur le CN et son écoulement en aval.Enfin, une étude de paramétrisation de l'advection horizontale et du mélange vertical a permis d'améliorerl'impact d'un downscaling dans la région d'étude, et particulièrement concernant la représentation de lacirculation au sein de la baie semi-fermée. / The main objective of this work is to improve our knowledge on the impact of the Northern Current (NC)mesoscale activity off the Var coast on its downstream flow and on the links between this boundary current andthe coastal dynamics, particularly in a semi-enclosed bay and shallow: the bay of Hyères. To do so, twonumerical realistic configurations at high-resolution were used. Based on the NEMO code and nested withAGRIF, the first one covers the French Mediterranean coasts at 1,2 km and the second one covers the Varcoasts with a spatial resolution of 400 m.Simulations comparisons with ocean observations (HF radar, ADCP, glider, satellite SST) confirm therealism of the configurations, and show the contribution of a 400 m spatial resolution on the simulateddynamics in the bay but also on the NC and its downstream flow.Finally, a parametrization study on the horizontal advection terms and vertical mixing provide an improvementof the impact of a downscaling in the studied area, and particularly for the simulated dynamics in the semienclosedbay.

The Northern Current flows cyclonically contouring the continental slope in the NW Mediterranean. At the entrance of the Gulf of Lions this current is about 20 -- 30 km wide and flows along the deepest half of the continental slope, i.e. over the 1000 to 2000 m isobaths approximately. Surface speeds are of 30 -- 50 cm s^{-1}. In the MATER HFF experiment (March -- May 1997) mesoscale variability of the Northern Current is observed from current meter records, SST images and hydrographic data. The HFF experimental box is 20 x 40 km, covering the upper half of the slope (i.e. covering from 250 m to 1250 m depth isobaths). Current meter and satellite data show that the site is embedded in a region of significant Northern Current meandering and eddy activity. From SST images, meander wavelengths are estimated larger than 60 km, embracing smaller structures. These flow patterns affect upper-layer waters down to at least 650 m depth. Current meter data distinguish two narrow energetic bands centred at 3.5 days and 7.5 days, respectively, in agreement with previous studies.Baroclinic instability is viewed as a possible mechanism to explain the generation of the Northern Current meanders. The analytical model of Tang (1975) predicts the development of unstable waves of wavelength (> 60 km) and periods compatible with the 7.5 day band recorded with current meter devices. The higher frequency band of 3.5 days is out of the frequency range predicted by the classical baroclinic instability theory and it is discussed as a restriction of quasi-geostrophic theory.Barotropic instability is studied using a laboratory model of a -westward' jet flowing over the lower half of the continental slope, which considers dynamic similarity with the Northern Current. The laboratory model is cross-validated with a corresponding numerical model. Jet instabilities of currents similar to the Northern Current (i.e. westward jets) occur at the edges of the jet, showing a clear meandering tendency over the mid-slope. Westward currents of Ro = 0.1 -- 0.2 develop instabilities of wavelengths (50 -- 75 km) similar to those observed from SST images, with periods (3.3 -- 3.8 days) compatible with the 3.5 days period band recorded with HFFE current meters.The laboratory and numerical experiments have reproduced westward jets (as the Northern Current), but also eastward jets, in order to have a full approach to better understand the role of the bottom topography on barotropic instabilities. The slope current instabilities are successfully explained by the Marcus and Lee theory (1998) of jets on a beta plane. This theory is valid for westward flows with Ro > 0.1 and for eastward flows with Ro > 0.2 (jets of the so-called Regime II flows in this thesis), and it states that the instabilities of each shear layer of the barotropic jet take the appearance of a Kelvin-Helmholtz-like pattern, associated with a Rossby wave (of topographic origin in our case). According to this theory, the differences between eastward and westward jets rely on the disposition of the Rossby waves --at the centre of the current in eastward flows and at the edges of the jet in westward currents. Jets over a sloping bottom with small Rossby numbers (Ro < 0.1 for westward jets; Ro < 0.2 for eastward jets) show a flow pattern (the so-called Regime I in this thesis) that has common characteristics for eastward and westward flows. In these -small'-Ro flows, Kelvin-Helmholtz-like instabilities dominate, whereas Rossby waves are too weak to produce any major difference between jets flowing in eastward or westward direction. This occurs when the topographic influence, assumed proportional to the Ro number of the jet, is small.The differences between eastward and westward slope currents observed in this work (and similar observations of jets on a beta-plane from previous works) are explained in this thesis by a simple scheme based on conservation of potential vorticity, considering there are two main components in balance: the shear-induced vorticity and the topographically induced vorticity. The signs of these two components are determined by the relative direction of the flow with respect to the inclination of the bottom topography. Once the critical Rossby number is overpassed so that the topographic effects are important (Ro > 0.1 for westward jets; Ro > 0.2 for eastward jets), conservation of potential vorticity tends to enhance vortices at the centre of eastward jets --eastward jets show meandering at the jet core. In westward jets, potential vorticity conservation is responsible of enhancing vortices at each edge of the jet. Thus, westward jets (as the Northern Current) are broad and meandering occurs at the jet edges.In Ro > 0.1 westward flows (i.e. Regime II westward jets) a topographic Rossby wave appears over the shelf break. This result is likely observed because of the specific topography used in this work --a continental slope and a continental shelf separated by a shelf break, producing a strong change in ambient potential vorticity. Numerical simulations reveal that this Rossby wave is triggered by the slope current. This topographic Rossby wave is a robust pattern, since it is independent of the position of the current over the slope, the shape of the velocity shear profile of the jet, and the jet width. Although this type of wave could not be inferred from the HFFE field data, it could be a focus of study in further field experiments. It also needs further analytical consideration. The general conclusion extracted from this thesis that tries to explain the mesoscale variability associated to the Northern Current is that both baroclinic and barotropic instability could explain part of the oceanic observations. As a consequence, mixed barotropic-baroclinic instability (which occurs at wavelengths which are between those corresponding to pure barotropic and pure baroclinic instability) is thought to play an important role on the observed mesoscale variability. The resulting wavelength would depend on the relative strength of both mechanisms.