5.2.6 Mixing in the abyss
5.2.6.1 The Brazil Basin tracer release
experiment
Understanding mixing in the abyssal oceans,
where isopycnals do not outcrop locally, is perhaps a simpler problem as the issues relating to
time-variable air–sea exchanges are absent. However, progress here has been slowed by the logistical difficulties of working at great depth, but that
is starting to change. Motivated by the large
diapycnal diffusivities inferred from abyssal heat
budgets for semi-enclosed basins, WHOI (Woods
Hole Oceanographic Institution) investigators
recently conducted a joint finestructure/microstructure and tracer study in the Brazil Basin in conjunction with the WOCE Deep Basin Experiment.
Full water column profiles of temperature, salinity
and horizontal velocity along with turbulent dissipation rate information were obtained on two
cruises spanning the zonal interval between the
3800 m isobath on the continental slope east of
Brazil and the Mid-Atlantic Ridge crest (Polzin
et al., 1997; Fig. 5.2.5, see Plate 5.2.5, p. 428).
Those stations occupied in the western part of the
Basin, including those taken over the smoothly
sloping South American continental rise where
a deep western boundary current is located (see
Durrieu De Madron and Weatherly, 1994), were
characterized by background-intensity internal
waves and weak turbulent mixing. Kunze and
Sanford (1996) also found background-intensity
internal wave shears in the deep Sargasso Sea
(a region characterized by smooth bathymetry).
In contrast, above the rough flanks of the MidAtlantic Ridge in the Brazil Basin, internal wave
energy was much enhanced and accompanied by
energetic microstructure that implied diffusivity values of order 10
93 m
2 s
91
. Rough areal averaging of
these microstructure data suggest there may be sufficient mixing in the abyssal Brazil Basin to close
the heat budget reported by Hogg et al. (1982).
Polzin et al. (1997) postulated that the enhanced
internal wave energy (and ultimately the energy
source of the strong turbulence) is the product of
internal wave generation by flow over the rough
bathymetry of the mid-ocean ridge. In particular,
barotropic tidal currents were implicated, mean
and mesoscale flows being thought too weak in
the eastern Brazil Basin. Critical bottom reflection
of waves radiating down from the thermocline
might have also contributed, but it is unlikely they
caused the observed fortnightly modulation in
depth-integrated kinetic energy dissipation that
motivated a tidal explanation (Ledwell et al.,
2000). Building on the idea of bottom wave generation, Polzin (2000) went on to take a locally generated wave field as a bottom boundary condition
in a model that predicts the dissipation profile in
terms of vertical wave propagation and wave–
wave interaction. The model predictions for the
turbulent dissipation rate magnitude and vertical
structure are in good agreement with the observations. Interestingly, Munk and Wunsch (1998)
independently explored the ideas of internal tide
generation and mixing from an astronomical viewpoint. The decay rate of the moon’s orbit sets the
global energy dissipation rate; the difficult part is
quantifying what (possibly small) fraction of this
energy supports mixing in the deep ocean. Indeed,
a large fraction of the barotropic tidal dissipation
may happen in shallow seas within bottom boundary layers. Though very uncertain, Munk and
Wunsch indicate that perhaps as much as 10
12 W
of energy is supplied to open-ocean tides that, in
turn, dissipates and supports diapycnal mixing in
the ocean interior. This amount of energy could
support an order 10
94 m
2 s
91 globally averaged
deep ocean diffusivity. Recent work by Egbert and
Ray (2000) gives supporting evidence for barotropic
tidal energy scattering into baroclinic motions at
ocean ridge systems (that in turn presumably dissipates and supports diapycnal mixing). They estimate an M 2 barotropic-to-baroclinic tidal energy
conversion rate in the deep ocean of 0.7<
0.15 TW, and by extrapolation for all tidal constituents, a deep-ocean tidal conversion rate of
1<0.25 TW.
5.2.6.2 Other abyssal mixing mechanisms
Enhanced deep-ocean internal wave shears have
also been observed in the Southern Ocean south of
Australia by Polzin and Firing (1997). Rather than
the tides, as were implicated in the Brazil Basin,
Polzin and Firing suggest internal wave generation
in the deep Southern Ocean is associated with
mean and mesoscale eddy currents responding to
bottom roughness. Based on the extended Henyey
et al. (1986) internal-wave-energy parameterization, they inferred K ϳ510
94 m
2 s
91 below
1000 m within the Antarctic Circumpolar Current
(see also Polzin, 1999). In addition to supporting
SECTION 5 FORMATION AND TRANSPORT OF WATER MASSES
352
5.2.6.1 The Brazil Basin tracer release
experiment
Understanding mixing in the abyssal oceans,
where isopycnals do not outcrop locally, is perhaps a simpler problem as the issues relating to
time-variable air–sea exchanges are absent. However, progress here has been slowed by the logistical difficulties of working at great depth, but that
is starting to change. Motivated by the large
diapycnal diffusivities inferred from abyssal heat
budgets for semi-enclosed basins, WHOI (Woods
Hole Oceanographic Institution) investigators
recently conducted a joint finestructure/microstructure and tracer study in the Brazil Basin in conjunction with the WOCE Deep Basin Experiment.
Full water column profiles of temperature, salinity
and horizontal velocity along with turbulent dissipation rate information were obtained on two
cruises spanning the zonal interval between the
3800 m isobath on the continental slope east of
Brazil and the Mid-Atlantic Ridge crest (Polzin
et al., 1997; Fig. 5.2.5, see Plate 5.2.5, p. 428).
Those stations occupied in the western part of the
Basin, including those taken over the smoothly
sloping South American continental rise where
a deep western boundary current is located (see
Durrieu De Madron and Weatherly, 1994), were
characterized by background-intensity internal
waves and weak turbulent mixing. Kunze and
Sanford (1996) also found background-intensity
internal wave shears in the deep Sargasso Sea
(a region characterized by smooth bathymetry).
In contrast, above the rough flanks of the MidAtlantic Ridge in the Brazil Basin, internal wave
energy was much enhanced and accompanied by
energetic microstructure that implied diffusivity values of order 10
93 m
2 s
91
. Rough areal averaging of
these microstructure data suggest there may be sufficient mixing in the abyssal Brazil Basin to close
the heat budget reported by Hogg et al. (1982).
Polzin et al. (1997) postulated that the enhanced
internal wave energy (and ultimately the energy
source of the strong turbulence) is the product of
internal wave generation by flow over the rough
bathymetry of the mid-ocean ridge. In particular,
barotropic tidal currents were implicated, mean
and mesoscale flows being thought too weak in
the eastern Brazil Basin. Critical bottom reflection
of waves radiating down from the thermocline
might have also contributed, but it is unlikely they
caused the observed fortnightly modulation in
depth-integrated kinetic energy dissipation that
motivated a tidal explanation (Ledwell et al.,
2000). Building on the idea of bottom wave generation, Polzin (2000) went on to take a locally generated wave field as a bottom boundary condition
in a model that predicts the dissipation profile in
terms of vertical wave propagation and wave–
wave interaction. The model predictions for the
turbulent dissipation rate magnitude and vertical
structure are in good agreement with the observations. Interestingly, Munk and Wunsch (1998)
independently explored the ideas of internal tide
generation and mixing from an astronomical viewpoint. The decay rate of the moon’s orbit sets the
global energy dissipation rate; the difficult part is
quantifying what (possibly small) fraction of this
energy supports mixing in the deep ocean. Indeed,
a large fraction of the barotropic tidal dissipation
may happen in shallow seas within bottom boundary layers. Though very uncertain, Munk and
Wunsch indicate that perhaps as much as 10
12 W
of energy is supplied to open-ocean tides that, in
turn, dissipates and supports diapycnal mixing in
the ocean interior. This amount of energy could
support an order 10
94 m
2 s
91 globally averaged
deep ocean diffusivity. Recent work by Egbert and
Ray (2000) gives supporting evidence for barotropic
tidal energy scattering into baroclinic motions at
ocean ridge systems (that in turn presumably dissipates and supports diapycnal mixing). They estimate an M 2 barotropic-to-baroclinic tidal energy
conversion rate in the deep ocean of 0.7<
0.15 TW, and by extrapolation for all tidal constituents, a deep-ocean tidal conversion rate of
1<0.25 TW.
5.2.6.2 Other abyssal mixing mechanisms
Enhanced deep-ocean internal wave shears have
also been observed in the Southern Ocean south of
Australia by Polzin and Firing (1997). Rather than
the tides, as were implicated in the Brazil Basin,
Polzin and Firing suggest internal wave generation
in the deep Southern Ocean is associated with
mean and mesoscale eddy currents responding to
bottom roughness. Based on the extended Henyey
et al. (1986) internal-wave-energy parameterization, they inferred K ϳ510
94 m
2 s
91 below
1000 m within the Antarctic Circumpolar Current
(see also Polzin, 1999). In addition to supporting
SECTION 5 FORMATION AND TRANSPORT OF WATER MASSES
352