Resumen de Análisis de remolinos mesoescalares oceánicos mediante trayectorias de boyas de deriva y un modelo de velocidad angular
This thesis examines the temporal evolution of mesoscalar eddies. ln the first part of the thesis we carry out an exhaustive analysis of field and numerical buoys tracking four eddies with different characteristics.
We initially determine the time that each buoy remains within an eddy and then we carry out an exhaustive analysis of the orbitaltrajectories and the tendency of the orbitalradius, from which we may infer the radial velocity and get an estlmate of the angular velocity. Generally the buoys remain within the during several weeks, although one buoy gets to stay during 200 days. The orbital periods range between 2 and 10 days, the radial velocities are very variable (typically of the order of 1 km in 10 days), and the angular velocity ranges between 1O5 y lfas-1. The second part of the thesis develops a simple radial advection--diffusíon model of angular velocity in an axisymmetric and vertically homogeneous eddy.
We explore several versions of the diffusive model, with the final version obtained from the analysis of the momentum equations in isopycnal coordinates. From these equations we are also able to include the radial advective term. ln order to apply the model we require the horizontal diffusion coefficient and the radial velocity. The diffusive coefficient is estimated from the buoy located within the eddies. verifying that the inertial stability condition is satisfied. The radial veloc¡ty is assumed to be proportional to the sign and magnitude of the horizontal convergence, so il can be expressed as a function of the angular velocity itself. The model equations are solved by applying the Mathematica sofh¡,¡are and their behaviour is calibrated with the data from all buoys.
The analysis of the buoys' trajectories is cornbined with the model's results. This suggests that radial advection is of minor importance as compared with diffusion, except forthe anticyclonic eddy from the Loop Current where its effect appears to be significant. The parameterization of the radial velocity u, is given as a fundion of the radius r and the angular velocity rr;, by the relation ur = c o) r, using c = 0,001.
For the eddies South of the Canary lslands, of barotropic character, the initial radius lays between 25 km (cyclonic) and 30 km (anticyclonic), the angular velocity is 1,8 (cycfonic) and - 2,5 (anticyclonic) 106sl, and the diffusion coeffícient is 30 (cyclonic) and 10 {anticyclonic) nr? s1 . For the anticyclonic eddy of the Algerian Cunent, of baroclinic charac{er, the initial radius is 45 km, the angular velocity is also - 2,5 1Os s^(-1), but the diffusion coefficient is one order of m4nitude laryer, '100 rn2 s' . For the anticyclonic eddy of the Loop Cunent, the initial angular velocity is much larger, - 2,21O4 s1, and the diffusion coefficient remains large, 100 rn2 s-t . ln this last case the best adjustment corresponds to an initial band with solid-body type rotation between 25 and 45 km from the eddy center. The modeldescribes reasonably well the data at short times after the eddy's formation but it is not adequate for long times, probably because eddies are not isolated structures, rather they interact with other mesoscalar slructures and with the bottom topography. Finally, the angular velocities predided by the model are analyzed under the assumption of a two layer ocean, with the top layer active and the bottom one at rest, This allows us to reconstruct the temporal evolution of the interface between both layers, which provides an overall view of the likely temporal evolution of the eddy's structure.
