Abstract
Data from the Tourbillon Experiment Intensive Period in the northeast Atlantic presented evidence of a vertical phase shift with time of the main eddy, interpreted as an occurrence of internal barotropic instability. In order to justify this: (i) the idealized case of an isolated eddy immersed in a stratified environment, whose characteristics correspond to the Tourbillon site (see Part I) and (ii) the realistic case using the full three-dimensional data from the experiment as initial conditions (Part II) are modeled. Both studies use a quasi-geostrophic periodic spectral model with six vertical normal modes and a horizontal 128×128 grid. It is demonstrated that a some what realistic “forecast” can be obtained for an integration time of up to one month.
While a linear instability analysis revealed that the Tourbillon eddy is very slowly unstable (Part I), its encounter with a Mediterranean Water tongue caused a large-amplitude baroclinic perturbation, triggering a nonlinear destabilization of the eddy, and hence the observed tilting of its vertical axis with time. One failure of the model concerns the final fate of the eddy: at the end of the intensive measurement period, the eddy is observed to remain a single entity, while the quasi-geostrophic modeling predicts its fragmentation into two vertically smaller structures by the strong instability.
