Abstract:
This thesis consists of two loosely related problems, both of which analyze some
consequences of the failure of Sverdrup relation. In the first part, Chapters 2 and 3,
the Sverdrup relation is invalidated because substantial flow is obtained at the bottom
where topography exists. The eddies play an essential role in transfering momentum
vertically from the surface, where the forcing is applied, to the bottom, which is otherwise
unforced. If the topography has a structure in the longitudinal direction, then
the inviscid theory predicts the occurence of strong jets in the interior of the model
ocean. According to the structure of the topography these internal jets can occur in
both vertically homeogenous and baroclinic oceans. If the topographic slope changes
sign, then one kind of jets is observed both in stratified and in homogeneous oceans.
This phenomenon is robust to moderate amounts of dissipation and is not disturbed
by the occurrence of recirculating gyres within the basin.
If the topographic slope is constant, then another kind of internal jets is observed,
and it occurs in stratified models only. I was unable to observe this kind of jets in
the presence of weak dissipation. The reason for this failure is twofold: on one hand
friction, especially interfacial friction, tends to make the flow more barotropic (and
we believe that indeed this is one of the processes that the eddies accomplish in a
stratified fluid) and therefore the phenomena that rely strongly on baroclinicity are
discouraged. On the other hand, reduction of the dissipation leads to the onset of a
strong recirculating, inertial gyre which, although confined in space, affects the global
properties of the flow.
In the second part of the thesis (Chapters 4 and 5) I developed a simple model of
the recirculating, inertial gyre. Again the dynamics of this feature are far from being in
Sverdrup balance. In this case inertia is responsible for the failure of Sverdrup relation,
together with the eddy field which provides a mean for transfering momentum vertically
and laterally into regions away from where the forcing is applied. In this model there is
no direct forcing in the recirculation region, and the input of momentum is confined to
the boundary currents surrounding the gyre, for example the separated Gulf Stream.
One of the results of the recirculation model is the prediction of its transport. It
is shown that most of the transport is depth independent, i.e. it can be calculated
without detailed knowledge of the density structure of the ocean. It is also shown that
the "barotropic" part of the transport increases as the cube of the meridional extent
of the gyre.
