suggest a net divergence of fresh water much
smaller than previously thought: Holfort and
Siedler’s (1997) estimates of a 0.24 Sv convergence
between Bering Strait and 30°S is roughly half that
predicted by Baumgartner and Reichel (1975),
while the estimates at 40°S indicate almost no net
divergence over the Atlantic/Arctic. The Indian
Ocean direct estimates, however, are consonant
with a large net evaporation, as predicted previously (Fig. 6.2.6, see Plate 6.2.6, p. 492).
Hence, if the Pacific Ocean is not the source of
the excess oceanic fresh water required to supply
the Indian deficit, only one possibility remains:
excess precipitation and ice melt over the Southern
Ocean. Indeed Sloyan and Rintoul estimate that
0.54 Sv of excess fresh water is removed by the
Southern Ocean circulation south of 30–40°S,
highlighting the importance of the Southern Ocean
in the global oceanic freshwater balance. Based on
the independent estimates in Table 6.2.2, the
excess evaporation occurring north of Sloyan and
Rintoul’s lines add up to 0.39 Sv. The freshwater
global balance is thus 0.5490.39:0.15<0.29 Sv:
indistinguishable from zero.
6.2.6 Mechanisms of oceanic freshwater
transport
Stommel and Csanady (1980) pointed out that
ocean heat and freshwater transport is related to
the rates of conversion of water from one part of
temperature–salinity (T/S) space to another. They
went further and attempted to model this process
assuming salinity was a simple function of temperature. Recent analyses of freshwater fluxes across
ocean sections and in general circulation models
(Rahmstorf, 1996) reveal that this assumption is
incorrect, though the underlying idea remains a
powerful one, as it links forcing fluxes to water
mass volume and exchange in T/S space. Speer and
Tziperman (1992) recast this approach in terms of
density classes and express the competition between
density class conversion by surface fluxes and that
by interior diapycnal mixing, an idea employed in
Sloyan and Rintoul’s (2000b) analysis. The challenge in analysing ocean sections will be in distinguishing the water mass conversion occurring at the
surface and that due to internal mixing. Use of
freshwater fluxes will be a key element in this work.
Definition of a tracer transport mechanism
across an ocean section is somewhat ad hoc. Hall
and Bryden (1982) chose to form zonal averages
(and deviations) of velocity and properties on pressure surfaces and termed the resulting products the
‘overturning’ component of tracer transport, while
the residual, associated with the correlation of
velocity and tracer at a pressure level, was termed
the ‘horizontal’ or gyre component. A similar
decomposition can be carried out within density
layers (e.g. Wijffels et al., 1996a; Robbins and
Toole, 1997), and as density is largely determined
by temperature, more closely relates back to Stommel and Csanady’s (1980) suggestion. Freshwater
divergence is also achieved by the interbasin flows,
the ‘leakage’ term in equation (6.2.4). Unfortunately, few detailed decompositions of the oceanic
freshwater flux across hydrographic lines are
available, but those that are reveal interesting
mechanisms.
Wijffels et al. (1996a) find that at 10°N in the
Pacific, the small net freshwater divergence over
the Pacific relative to Bering Strait is achieved by a
balance between three major flux mechanisms: net
export of very fresh water through Bering Strait;
northward freshwater transport by the shallow
meridional circulation where the northward Ekman
flux is fresh and the compensating southward
thermocline flow more saline; and a southward
freshwater flux by the subthermocline horizontal
circulation where salty South Pacific waters flow
north in the eastern Pacific and fresh North Pacific
Intermediate water flows south in the western
Pacific. The deep and bottom water circulations in
the low-latitude Pacific achieve little net freshwater transport.
In the Indian Ocean, Robbins and Toole (1997)
diagnose a large net evaporation north of the section at 32°S of 0.31<0.09 Sv. They find three
mechanisms acting to import fresh water to supply
this net evaporation: a leakage term associated
with the inflow of fresh Indonesian Archipelago
waters that are distilled and leave across 32°S as
salty thermocline waters; upwelling of deep and
intermediate waters and their export as salty thermocline waters; and horizontal inflow of fresh
Antarctic Intermediate Water in the east that
leaves saltier in the west. Wijffels et al. (2000)
have similarly found the import of fresh water by
the recirculation and modification of Antarctic
intermediate waters to be an important transport
mechanism at 32°S in the Pacific. It is likely that
the production and northward export of fresh
SECTION 6 LARGE-SCALE OCEAN TRANSPORTS
486
smaller than previously thought: Holfort and
Siedler’s (1997) estimates of a 0.24 Sv convergence
between Bering Strait and 30°S is roughly half that
predicted by Baumgartner and Reichel (1975),
while the estimates at 40°S indicate almost no net
divergence over the Atlantic/Arctic. The Indian
Ocean direct estimates, however, are consonant
with a large net evaporation, as predicted previously (Fig. 6.2.6, see Plate 6.2.6, p. 492).
Hence, if the Pacific Ocean is not the source of
the excess oceanic fresh water required to supply
the Indian deficit, only one possibility remains:
excess precipitation and ice melt over the Southern
Ocean. Indeed Sloyan and Rintoul estimate that
0.54 Sv of excess fresh water is removed by the
Southern Ocean circulation south of 30–40°S,
highlighting the importance of the Southern Ocean
in the global oceanic freshwater balance. Based on
the independent estimates in Table 6.2.2, the
excess evaporation occurring north of Sloyan and
Rintoul’s lines add up to 0.39 Sv. The freshwater
global balance is thus 0.5490.39:0.15<0.29 Sv:
indistinguishable from zero.
6.2.6 Mechanisms of oceanic freshwater
transport
Stommel and Csanady (1980) pointed out that
ocean heat and freshwater transport is related to
the rates of conversion of water from one part of
temperature–salinity (T/S) space to another. They
went further and attempted to model this process
assuming salinity was a simple function of temperature. Recent analyses of freshwater fluxes across
ocean sections and in general circulation models
(Rahmstorf, 1996) reveal that this assumption is
incorrect, though the underlying idea remains a
powerful one, as it links forcing fluxes to water
mass volume and exchange in T/S space. Speer and
Tziperman (1992) recast this approach in terms of
density classes and express the competition between
density class conversion by surface fluxes and that
by interior diapycnal mixing, an idea employed in
Sloyan and Rintoul’s (2000b) analysis. The challenge in analysing ocean sections will be in distinguishing the water mass conversion occurring at the
surface and that due to internal mixing. Use of
freshwater fluxes will be a key element in this work.
Definition of a tracer transport mechanism
across an ocean section is somewhat ad hoc. Hall
and Bryden (1982) chose to form zonal averages
(and deviations) of velocity and properties on pressure surfaces and termed the resulting products the
‘overturning’ component of tracer transport, while
the residual, associated with the correlation of
velocity and tracer at a pressure level, was termed
the ‘horizontal’ or gyre component. A similar
decomposition can be carried out within density
layers (e.g. Wijffels et al., 1996a; Robbins and
Toole, 1997), and as density is largely determined
by temperature, more closely relates back to Stommel and Csanady’s (1980) suggestion. Freshwater
divergence is also achieved by the interbasin flows,
the ‘leakage’ term in equation (6.2.4). Unfortunately, few detailed decompositions of the oceanic
freshwater flux across hydrographic lines are
available, but those that are reveal interesting
mechanisms.
Wijffels et al. (1996a) find that at 10°N in the
Pacific, the small net freshwater divergence over
the Pacific relative to Bering Strait is achieved by a
balance between three major flux mechanisms: net
export of very fresh water through Bering Strait;
northward freshwater transport by the shallow
meridional circulation where the northward Ekman
flux is fresh and the compensating southward
thermocline flow more saline; and a southward
freshwater flux by the subthermocline horizontal
circulation where salty South Pacific waters flow
north in the eastern Pacific and fresh North Pacific
Intermediate water flows south in the western
Pacific. The deep and bottom water circulations in
the low-latitude Pacific achieve little net freshwater transport.
In the Indian Ocean, Robbins and Toole (1997)
diagnose a large net evaporation north of the section at 32°S of 0.31<0.09 Sv. They find three
mechanisms acting to import fresh water to supply
this net evaporation: a leakage term associated
with the inflow of fresh Indonesian Archipelago
waters that are distilled and leave across 32°S as
salty thermocline waters; upwelling of deep and
intermediate waters and their export as salty thermocline waters; and horizontal inflow of fresh
Antarctic Intermediate Water in the east that
leaves saltier in the west. Wijffels et al. (2000)
have similarly found the import of fresh water by
the recirculation and modification of Antarctic
intermediate waters to be an important transport
mechanism at 32°S in the Pacific. It is likely that
the production and northward export of fresh
SECTION 6 LARGE-SCALE OCEAN TRANSPORTS
486