to suggest that wave–wave interactions were weaker
at low relative to mid-latitudes.)
In contrast, the Gulf Stream, one of the most
significant ocean currents on the globe, does not
appear to support significant turbulent diapycnal
mixing. Gregg and Sanford (1980) and Osborn
(1980) inferred diffusivity values from microstructure profiles in the Stream that average much less
than 10
94 m
2 s
91
. And Winkel et al. (1992) report
an average diffusivity value of 4.110
95 m
2 s
91
across the Florida Current outside the 10- to 50m-thick surface and bottom boundary layers.
One attribute of these large-scale currents that is
thought important to mixing is their horizontal
shear. Ocean areas characterized by negative
relative vorticity are often sites of enhanced internal wave energy density. As discussed by Kunze
(1985) and Kunze and Boss (1998), low-frequency
internal waves sense an effective inertial period
that depends on the relative vorticity of the largescale flow in which they are imbedded. Flows with
negative relative vorticity are capable of forming
wave-trapping zones in the horizontal and critical
layers in the vertical. At the latter, wave energy
accumulates and eventually dissipates. Kunze et al.
(1995) and Kunze and Toole (1997) offer examples of this process within a warm core ring of the
Gulf Stream and within a rectified anticyclonic
flow atop a mid-latitude seamount, respectively.
In the latter case, the diapycnal buoyancy flux
above the seamount was some 200 times that
of the background interior value (on a per unit
area basis). But this enhanced mixing above the
seamount may be due as much to the abrupt
bathymetry being a strong source of internal wave
energy (see below) as to the over-lying vortex that
traps that energy.
5.2.5.3 Near-bottom mixing
Isopycnals characteristically found in the main
thermocline of the ocean interior approach, and
commonly intersect, the bottom over the continental slopes and shelves, mid-ocean ridges and seamounts. In those regions where the bottom doesn’t
slope too steeply, a (virtually) homogeneous layer
some 10 or (sometimes many) more metres thick
is often observed: the bottom boundary layer
(e.g. Caldwell, 1976; Lentz and Trowbridge,
1991). These layers are sometimes believed to be
the manifestation of turbulent mixing driven by
bottom stresses associated with the overlying flow,
but exceptional levels of turbulent kinetic energy
dissipation don’t always span the depth of the
homogeneous layers. This is, consequently, a
rather inefficient mixing mechanism as the bulk of
the turbulence lies in previously mixed water (see
discussions by Armi, 1978, 1979 and Garrett,
1979, 1991). Indeed, thick bottom-layer formation
can also be the result of lateral convergence in the
boundary layer flow and involve little or no mixing.
Breaking of boundary-reflected internal waves
in the stratified water column above the bottom
layer is believed to be a far more efficient mixing
mechanism. Linear bottom-reflection kinematics
for sloping bathymetry predicts a range of incident
waves that reflect with smaller vertical scale and
greater energy density (Phillips, 1977; Eriksen,
1982). The effect is greatest about the critical
wave frequency: c :[N
2 sin
2 (␣);f
2 cos
2 (␣)]
1/2 ,
where ␣ is the angle of the bottom from horizontal. Incident waves at this frequency reflect with a
group velocity vector parallel to the bottom (and
so are not able to carry energy away from the
slope. Linear theory in fact breaks down at this
point.) It is surmised that subsequent breaking of
the reflected waves may in turn support enhanced
turbulent mixing above slopes (Eriksen, 1985;
Garrett and Gilbert, 1988; Slinn and Riley, 1999),
although destructive interference of bottomreflected waves above concave bathymetry and
generation of along-isobath mean flows may limit
or distort the effect (Gilbert and Garrett, 1989;
Slinn, 1999, respectively). A key point with this
mixing mechanism relates to the waves’ ability to
propagate vertically; wave breaking and associated
turbulent dissipation can occur remote from the
bottom (and any homogenized bottom layer) and
thus support mixing of water properties. But as is
the case for bottom-boundary-layer mixing, the
ability of secondary and tertiary circulations to
remove the products of wave breaking and thus
sustain the stratification is central to the effectiveness of these near-boundary mixing processes
(Garrett, 1991; Garrett et al., 1993a).
In support of these ideas, enhanced internal
wave energy density has been reported about the
estimated critical wave frequencies in moored
current meter records obtained above sloping
bathymetry (Eriksen, 1985, 1998), though the evidence is not universal (e.g. Gilbert, 1993). And
Toole et al. (1997) found enhanced dissipation
supporting diffusivities of around 510
94 m
2 s
91
SECTION 5 FORMATION AND TRANSPORT OF WATER MASSES
350
at low relative to mid-latitudes.)
In contrast, the Gulf Stream, one of the most
significant ocean currents on the globe, does not
appear to support significant turbulent diapycnal
mixing. Gregg and Sanford (1980) and Osborn
(1980) inferred diffusivity values from microstructure profiles in the Stream that average much less
than 10
94 m
2 s
91
. And Winkel et al. (1992) report
an average diffusivity value of 4.110
95 m
2 s
91
across the Florida Current outside the 10- to 50m-thick surface and bottom boundary layers.
One attribute of these large-scale currents that is
thought important to mixing is their horizontal
shear. Ocean areas characterized by negative
relative vorticity are often sites of enhanced internal wave energy density. As discussed by Kunze
(1985) and Kunze and Boss (1998), low-frequency
internal waves sense an effective inertial period
that depends on the relative vorticity of the largescale flow in which they are imbedded. Flows with
negative relative vorticity are capable of forming
wave-trapping zones in the horizontal and critical
layers in the vertical. At the latter, wave energy
accumulates and eventually dissipates. Kunze et al.
(1995) and Kunze and Toole (1997) offer examples of this process within a warm core ring of the
Gulf Stream and within a rectified anticyclonic
flow atop a mid-latitude seamount, respectively.
In the latter case, the diapycnal buoyancy flux
above the seamount was some 200 times that
of the background interior value (on a per unit
area basis). But this enhanced mixing above the
seamount may be due as much to the abrupt
bathymetry being a strong source of internal wave
energy (see below) as to the over-lying vortex that
traps that energy.
5.2.5.3 Near-bottom mixing
Isopycnals characteristically found in the main
thermocline of the ocean interior approach, and
commonly intersect, the bottom over the continental slopes and shelves, mid-ocean ridges and seamounts. In those regions where the bottom doesn’t
slope too steeply, a (virtually) homogeneous layer
some 10 or (sometimes many) more metres thick
is often observed: the bottom boundary layer
(e.g. Caldwell, 1976; Lentz and Trowbridge,
1991). These layers are sometimes believed to be
the manifestation of turbulent mixing driven by
bottom stresses associated with the overlying flow,
but exceptional levels of turbulent kinetic energy
dissipation don’t always span the depth of the
homogeneous layers. This is, consequently, a
rather inefficient mixing mechanism as the bulk of
the turbulence lies in previously mixed water (see
discussions by Armi, 1978, 1979 and Garrett,
1979, 1991). Indeed, thick bottom-layer formation
can also be the result of lateral convergence in the
boundary layer flow and involve little or no mixing.
Breaking of boundary-reflected internal waves
in the stratified water column above the bottom
layer is believed to be a far more efficient mixing
mechanism. Linear bottom-reflection kinematics
for sloping bathymetry predicts a range of incident
waves that reflect with smaller vertical scale and
greater energy density (Phillips, 1977; Eriksen,
1982). The effect is greatest about the critical
wave frequency: c :[N
2 sin
2 (␣);f
2 cos
2 (␣)]
1/2 ,
where ␣ is the angle of the bottom from horizontal. Incident waves at this frequency reflect with a
group velocity vector parallel to the bottom (and
so are not able to carry energy away from the
slope. Linear theory in fact breaks down at this
point.) It is surmised that subsequent breaking of
the reflected waves may in turn support enhanced
turbulent mixing above slopes (Eriksen, 1985;
Garrett and Gilbert, 1988; Slinn and Riley, 1999),
although destructive interference of bottomreflected waves above concave bathymetry and
generation of along-isobath mean flows may limit
or distort the effect (Gilbert and Garrett, 1989;
Slinn, 1999, respectively). A key point with this
mixing mechanism relates to the waves’ ability to
propagate vertically; wave breaking and associated
turbulent dissipation can occur remote from the
bottom (and any homogenized bottom layer) and
thus support mixing of water properties. But as is
the case for bottom-boundary-layer mixing, the
ability of secondary and tertiary circulations to
remove the products of wave breaking and thus
sustain the stratification is central to the effectiveness of these near-boundary mixing processes
(Garrett, 1991; Garrett et al., 1993a).
In support of these ideas, enhanced internal
wave energy density has been reported about the
estimated critical wave frequencies in moored
current meter records obtained above sloping
bathymetry (Eriksen, 1985, 1998), though the evidence is not universal (e.g. Gilbert, 1993). And
Toole et al. (1997) found enhanced dissipation
supporting diffusivities of around 510
94 m
2 s
91
SECTION 5 FORMATION AND TRANSPORT OF WATER MASSES
350