small in the LCDW layer, which lies below the
topography where mean geostrophic meridional
flow is possible. In the terminology of Section
4.6.3.2, these observations suggest the real ocean
is closer to case I.
Because the overturning circulation in numerical models depends on how the model is forced
(e.g. restoring or surface flux boundary conditions), how eddies are parameterized, and whether
the model is in steady state, they do not provide
conclusive evidence for or against deep versus
shallow compensation of the Ekman transport.
Nevertheless, a number of recent simulations, both
coarse and fine resolution, support the idea that a
significant fraction of the Ekman transport is balanced by a return flow shallower than the topography (case I). For example, Hirst and McDougall
(1998) show that a coarse-resolution level model
with the Gent and McWilliams (1990) parameterization of eddy-induced advection is consistent
with case I: in density coordinates, there is very little meridional overturning associated with the
Ekman transport between 40°S and 60°S. While
the degree of cancellation between the eddyinduced advection and the Ekman transport
(nearly complete in their model) depends on the
value chosen for the diffusivity, the model
improvements that result when such a parameterization is used support the notion that eddy-driven
transport likely plays an important role in the
overturning circulation of the Southern Ocean.
(The simulation is also improved by the decrease
in horizontal mixing permitted when the eddy
parameterization is used.) Killworth and Nanneh’s
(1994) analysis of the zonal momentum budget in
isopycnal layers in FRAM supports this conclusion: at the latitudes of Drake Passage, almost all
(at the southern side) to about half (at the northern side) of the northward transport of light water
returns south at densities that do not intersect
topography.
In summary, observations suggest significant
poleward flow in layers above topography, which
must be driven by divergence of the eddy (standing
and/or transient) interfacial form stress, in addition
to a poleward flow in density layers blocked by
topography. A variety of numerical simulations also
suggest that the presence of eddies permits a southward flux at densities above topography, which
compensates a large fraction of the northward
Ekman flux. In QG simulations this compensation
must be complete, since no diabatic transport is
permitted. In PE models that either resolve or
adequately parameterize the effect of eddies, there
is still a large degree of compensation. The time
and zonal mean flow at constant depth, in which
the Ekman mass transport returns at depths
blocked by topography, is thus decoupled from the
time and zonal mean flow at constant density,
in which most of the Ekman transport returns at
densities above topography.
4.6.6 Conclusions
Substantial progress has been made in understanding the circulation of the Southern Ocean during
the ‘WOCE decade’. This progress has relied on
advances in observations, theory and modelling.
Observations collected during WOCE represent a
significant achievement, given the challenges posed
by the remote and often hostile nature of the
Southern Ocean. Highlights include a circumpolar
survey of high-quality hydrographic, tracer and
ADCP data (including some of the first repeat sections obtained in the region), a number of mooring
arrays, float and drifter deployments, and satellite
measurements of sea surface height and temperature. Analytical models have provided insight into
the mechanisms responsible for setting the transport of the ACC. The last decade has also seen
rapid development of numerical models of the
Southern Ocean, including the first GCMs to
incorporate stratification, realistic topography,
and to resolve (or ‘permit’) eddies. The ability of
coarse-resolution models to simulate the Southern
Ocean has also improved significantly, in part due
to more effective parameterizations of the effect
of eddies. Fine-resolution models have achieved
sufficient realism that we can use them to estimate
the magnitude of individual terms in the momentum or vorticity budgets, to identify sites of strong
topographic influence, and to describe qualitatively the circumpolar structure of the complex,
filamented ACC, although it is still not clear
what resolution is necessary for a truly realistic
simulation.
The momentum, vorticity and buoyancy budgets – and as a consequence, the zonal and meridional circulations – are intimately linked. The
dynamics of the ACC differ in character from
those of strong currents in other ocean basins that
are zonally blocked. The now established fact that
SECTION 4 THE GLOBAL FLOW FIELD
300
topography where mean geostrophic meridional
flow is possible. In the terminology of Section
4.6.3.2, these observations suggest the real ocean
is closer to case I.
Because the overturning circulation in numerical models depends on how the model is forced
(e.g. restoring or surface flux boundary conditions), how eddies are parameterized, and whether
the model is in steady state, they do not provide
conclusive evidence for or against deep versus
shallow compensation of the Ekman transport.
Nevertheless, a number of recent simulations, both
coarse and fine resolution, support the idea that a
significant fraction of the Ekman transport is balanced by a return flow shallower than the topography (case I). For example, Hirst and McDougall
(1998) show that a coarse-resolution level model
with the Gent and McWilliams (1990) parameterization of eddy-induced advection is consistent
with case I: in density coordinates, there is very little meridional overturning associated with the
Ekman transport between 40°S and 60°S. While
the degree of cancellation between the eddyinduced advection and the Ekman transport
(nearly complete in their model) depends on the
value chosen for the diffusivity, the model
improvements that result when such a parameterization is used support the notion that eddy-driven
transport likely plays an important role in the
overturning circulation of the Southern Ocean.
(The simulation is also improved by the decrease
in horizontal mixing permitted when the eddy
parameterization is used.) Killworth and Nanneh’s
(1994) analysis of the zonal momentum budget in
isopycnal layers in FRAM supports this conclusion: at the latitudes of Drake Passage, almost all
(at the southern side) to about half (at the northern side) of the northward transport of light water
returns south at densities that do not intersect
topography.
In summary, observations suggest significant
poleward flow in layers above topography, which
must be driven by divergence of the eddy (standing
and/or transient) interfacial form stress, in addition
to a poleward flow in density layers blocked by
topography. A variety of numerical simulations also
suggest that the presence of eddies permits a southward flux at densities above topography, which
compensates a large fraction of the northward
Ekman flux. In QG simulations this compensation
must be complete, since no diabatic transport is
permitted. In PE models that either resolve or
adequately parameterize the effect of eddies, there
is still a large degree of compensation. The time
and zonal mean flow at constant depth, in which
the Ekman mass transport returns at depths
blocked by topography, is thus decoupled from the
time and zonal mean flow at constant density,
in which most of the Ekman transport returns at
densities above topography.
4.6.6 Conclusions
Substantial progress has been made in understanding the circulation of the Southern Ocean during
the ‘WOCE decade’. This progress has relied on
advances in observations, theory and modelling.
Observations collected during WOCE represent a
significant achievement, given the challenges posed
by the remote and often hostile nature of the
Southern Ocean. Highlights include a circumpolar
survey of high-quality hydrographic, tracer and
ADCP data (including some of the first repeat sections obtained in the region), a number of mooring
arrays, float and drifter deployments, and satellite
measurements of sea surface height and temperature. Analytical models have provided insight into
the mechanisms responsible for setting the transport of the ACC. The last decade has also seen
rapid development of numerical models of the
Southern Ocean, including the first GCMs to
incorporate stratification, realistic topography,
and to resolve (or ‘permit’) eddies. The ability of
coarse-resolution models to simulate the Southern
Ocean has also improved significantly, in part due
to more effective parameterizations of the effect
of eddies. Fine-resolution models have achieved
sufficient realism that we can use them to estimate
the magnitude of individual terms in the momentum or vorticity budgets, to identify sites of strong
topographic influence, and to describe qualitatively the circumpolar structure of the complex,
filamented ACC, although it is still not clear
what resolution is necessary for a truly realistic
simulation.
The momentum, vorticity and buoyancy budgets – and as a consequence, the zonal and meridional circulations – are intimately linked. The
dynamics of the ACC differ in character from
those of strong currents in other ocean basins that
are zonally blocked. The now established fact that
SECTION 4 THE GLOBAL FLOW FIELD
300