an overturning cell, which transports large amounts
of heat northward across the equator. In the
Pacific and Indian Oceans, overturning cells are
confined to the thermocline with weaker deep cells
below. In the Pacific Ocean, significant northward
flow across the equator is required to supply the
Indonesian Throughflow.
The main mechanisms resulting in heat and
freshwater transport in the equatorial Indian
Ocean are completely different from those in the
Pacific and Atlantic, largely because the winds –
especially those along the equator – are so different. Thus the Somali Current, rather than the
equatorial Indian Ocean, seems to be the major
region of near-surface water mass generation.
Hence Indian Ocean equatorial flows must be
organized primarily to carry cold inflows and
warm outflows across the equator, rather than
upwelling them on the equator and distributing
them to both sides as in the other two oceans.
However, it is quite likely that the main transport mechanisms of each ocean are significant (if
minor) contributors in the other. Thus, winddriven upwelling in the eastern Indian Ocean plays a
role in SST cooling and water mass formation during interannual ‘Indian Ocean Dipole’ events (Saji
et al., 1999). Conversely, seasonal fluctuations of
Pacific near-surface equatorial currents, such as
the New Guinea Coastal Current, could provide a
mean cross-equatorial ‘flushing’ of water masses
like those that flush Somali Current water south.
In all three oceans, the Ekman transport across the
equator, generated by zonal winds that are antisymmetric about the equator, may be balanced (on
seasonal time scales) by flows at around 3000 m or
deeper. This is at least consistent with observations in the Atlantic and Indian Oceans, but less
is known about deep seasonal cross-equatorial
currents in the Pacific.
As noted in the Introduction, mechanisms of the
mixing processes that create new water masses are
highly non-linear, and mixing ‘hot spots’ are probably located quite tightly in space – and in time.
(For example, tidal mixing may be concentrated
near spring tide, while that due to intraseasonal
activity may be concentrated near westerly wind
bursts.) A variety of qualitatively different mechanisms have emerged, including heat release from
mid-ocean ridges as well as tidal mixing and
Kelvin–Helmholtz instability across shear zones.
The fact that (according to observed climatologies)
strong mixing occurs in Indonesia, within the very
high SSTs of the Warm Pool, suggests that locating
the source of the mixing is not just a matter of
locating low SSTs. So far, it has in fact not proved
possible to locate definitely a clear source of tidal
mixing in the Indonesian Seas. On the other hand,
progress is being made in understanding mixing at
the shear zones associated with the Deep Reversing
Jets. Refraction of internal waves may sustain these
jets in the presence of diffusion. Early process study
work suggested that mixing (probably by Kelvin–
Helmholtz instability) may extend well below the
EUC, in the Pacific; but such results could well be
revisited, using the much more extensive data now
available. Mixing processes in several coastal
regions also need direct observation.
The analysis of WOCE hydrographic section
data for purposes of estimating long-term means is
made more complicated by variability on interannual, seasonal and intraseasonal time scales,
which may cause aliasing in geostrophic current
estimates. The TAO/Triton mooring array can
assist greatly in disentangling such effects in the
upper few hundred metres of the Pacific, but model
studies suggest that variability on all three time
scales is important in the Indian Ocean. Intraseasonal variability is less prominent in the Atlantic,
slightly simplifying the analysis task there.
Acknowledgements
We wish to thank two anonymous reviewers, and
Dr J. McCreary, for their careful and constructive
comments on this manuscript, which (at least to
our eyes) have improved the coherence and readability of the result. Eric Firing and Roger Lukas
have offered many helpful items to this review.
SECTION 4 THE GLOBAL FLOW FIELD
246
of heat northward across the equator. In the
Pacific and Indian Oceans, overturning cells are
confined to the thermocline with weaker deep cells
below. In the Pacific Ocean, significant northward
flow across the equator is required to supply the
Indonesian Throughflow.
The main mechanisms resulting in heat and
freshwater transport in the equatorial Indian
Ocean are completely different from those in the
Pacific and Atlantic, largely because the winds –
especially those along the equator – are so different. Thus the Somali Current, rather than the
equatorial Indian Ocean, seems to be the major
region of near-surface water mass generation.
Hence Indian Ocean equatorial flows must be
organized primarily to carry cold inflows and
warm outflows across the equator, rather than
upwelling them on the equator and distributing
them to both sides as in the other two oceans.
However, it is quite likely that the main transport mechanisms of each ocean are significant (if
minor) contributors in the other. Thus, winddriven upwelling in the eastern Indian Ocean plays a
role in SST cooling and water mass formation during interannual ‘Indian Ocean Dipole’ events (Saji
et al., 1999). Conversely, seasonal fluctuations of
Pacific near-surface equatorial currents, such as
the New Guinea Coastal Current, could provide a
mean cross-equatorial ‘flushing’ of water masses
like those that flush Somali Current water south.
In all three oceans, the Ekman transport across the
equator, generated by zonal winds that are antisymmetric about the equator, may be balanced (on
seasonal time scales) by flows at around 3000 m or
deeper. This is at least consistent with observations in the Atlantic and Indian Oceans, but less
is known about deep seasonal cross-equatorial
currents in the Pacific.
As noted in the Introduction, mechanisms of the
mixing processes that create new water masses are
highly non-linear, and mixing ‘hot spots’ are probably located quite tightly in space – and in time.
(For example, tidal mixing may be concentrated
near spring tide, while that due to intraseasonal
activity may be concentrated near westerly wind
bursts.) A variety of qualitatively different mechanisms have emerged, including heat release from
mid-ocean ridges as well as tidal mixing and
Kelvin–Helmholtz instability across shear zones.
The fact that (according to observed climatologies)
strong mixing occurs in Indonesia, within the very
high SSTs of the Warm Pool, suggests that locating
the source of the mixing is not just a matter of
locating low SSTs. So far, it has in fact not proved
possible to locate definitely a clear source of tidal
mixing in the Indonesian Seas. On the other hand,
progress is being made in understanding mixing at
the shear zones associated with the Deep Reversing
Jets. Refraction of internal waves may sustain these
jets in the presence of diffusion. Early process study
work suggested that mixing (probably by Kelvin–
Helmholtz instability) may extend well below the
EUC, in the Pacific; but such results could well be
revisited, using the much more extensive data now
available. Mixing processes in several coastal
regions also need direct observation.
The analysis of WOCE hydrographic section
data for purposes of estimating long-term means is
made more complicated by variability on interannual, seasonal and intraseasonal time scales,
which may cause aliasing in geostrophic current
estimates. The TAO/Triton mooring array can
assist greatly in disentangling such effects in the
upper few hundred metres of the Pacific, but model
studies suggest that variability on all three time
scales is important in the Indian Ocean. Intraseasonal variability is less prominent in the Atlantic,
slightly simplifying the analysis task there.
Acknowledgements
We wish to thank two anonymous reviewers, and
Dr J. McCreary, for their careful and constructive
comments on this manuscript, which (at least to
our eyes) have improved the coherence and readability of the result. Eric Firing and Roger Lukas
have offered many helpful items to this review.
SECTION 4 THE GLOBAL FLOW FIELD
246