such ocean heat transport estimates, however, is
entirely dependent on the validity of the circulation
scheme. Is there enough cold deep water flowing
equatorward and enough warm surface water
flowing poleward? Are the boundary currents and
gyre-scale circulations represented reasonably? Are
the wind-driven Ekman transports accurate? Are
the eddy fluxes realistically included? All such
questions (and more) must be answered affirmatively before the estimated ocean heat transport
can be considered a reliable measure of the ocean’s
role in maintaining the global heat balance. Ultimately, estimating ocean heat transport is a matter
of quantifying the ocean circulation. Ocean heat
transport can be a useful summary of the circulation; but heat transport values without a firm
understanding of the circulation are vacuous.
Direct estimates of ocean heat transport are
usually made using zonal coast-to-coast hydrographic sections. Prior to WOCE, successful heat
transport estimates had been made primarily
where the magnitude of the western boundary current transport was known: at 24°N in the Atlantic
(Hall and Bryden, 1982), where the Gulf Stream
transport through Florida Straits is 30 Sv
(1 Sv:10
6 m
3 s
91
) (Niiler and Richardson, 1973;
Larsen, 1992); and at 24°N in the Pacific (Bryden
et al., 1991), where the Kuroshio transport over
the Okinawa Trough is 27 Sv (Bingham and Talley,
1991). The reason is that, in combination with the
requirement for total mass conservation for the
section, the boundary current transport puts a
strong constraint on the interior circulation of the
ocean determined by geostrophic calculations
across the breadth of the interior ocean. It is well
known that geostrophic velocity estimates depend
on a reference level velocity or, alternatively, the
choice of a zero velocity surface. Without constraints imposed by the western boundary current
transports, very small changes in reference velocity
can create enormous changes in the interior circulation. For example, across the 6000-km wide
24°N Atlantic section, a modification in reference
velocity of 0.1 cm s
91 changes the interior ocean
transport by 28 Sv. Thus, a small change in the reference level velocity (or zero velocity surface) can
dramatically alter the strength of the basin-scale
circulation and resultant heat transport estimates.
Knowledge of the boundary current transports
constrains the interior circulation and effectively
determines the overall reference level velocity.
During WOCE, a large number of zonal hydrographic sections were carried out: across six latitudes in the Pacific, ten in the Atlantic and four in
the Indian Ocean (Fig. 3.1.4a, see Plate 3.1.4a,
p. 172). Current meter arrays at five of these latitudes (24°N and 30°S in the Atlantic, 20–32°S in
the Indian, and 30°N and 30–32°S in the Pacific)
were deployed to measure the strength of the deep
and/or shallow boundary currents so that estimates of meridional ocean heat transport could be
made more accurately. Analysis of the circulation
and meridional heat transport across these five
critical latitudes has not yet been completed so, in
a sense, we await the best direct estimates of circulation and heat transport that will be made from
WOCE. Nevertheless, notable progress has
recently been made on the Atlantic circulation and
heat transport, using WOCE sections as well as
hydrographic sections from the 1980s, and a
global inverse analysis of high-quality pre-WOCE
sections has derived a consistent pattern of circulation and heat transport. While there has been
progress in the Pacific and Indian Ocean analyses,
much remains to be done. We will review these
results here before discussing the challenges in
developing circulation and heat transport estimates for all of the WOCE sections.
6.1.4.1 Atlantic Ocean heat transport
Atlantic transoceanic sections across 55°N, 24°N,
30°S and 45°S early in WOCE have already been
analysed for meridional ocean heat transport
(Saunders and King, 1995b; Bacon, 1997; Lavín
et al., 1998; Holfort and Siedler, 2001). In addition,
three transoceanic sections across 14°N, 11°N and
11°S during the 1980s have been recently analysed
for heat transport (Friedrichs and Hall, 1993; Klein
et al., 1995; Speer et al., 1996). Each of these
studies includes a consideration of the flow
through Bering Strait, whose effect is small for
Atlantic heat transport; and each considers the
uncertainty in the resulting estimate of ocean heat
transport. There is general agreement that the
uncertainty in careful estimates of ocean heat
transport is about 0.3 PW.
Each individual study of Atlantic heat transport
shows the vertical distribution of the zonally averaged meridional circulation, which is responsible
for nearly all of the heat transport. For each
section, except 55°N, which is too far north for
Antarctic water masses to penetrate, the deep
SECTION 6 LARGE-SCALE OCEAN TRANSPORTS
460
entirely dependent on the validity of the circulation
scheme. Is there enough cold deep water flowing
equatorward and enough warm surface water
flowing poleward? Are the boundary currents and
gyre-scale circulations represented reasonably? Are
the wind-driven Ekman transports accurate? Are
the eddy fluxes realistically included? All such
questions (and more) must be answered affirmatively before the estimated ocean heat transport
can be considered a reliable measure of the ocean’s
role in maintaining the global heat balance. Ultimately, estimating ocean heat transport is a matter
of quantifying the ocean circulation. Ocean heat
transport can be a useful summary of the circulation; but heat transport values without a firm
understanding of the circulation are vacuous.
Direct estimates of ocean heat transport are
usually made using zonal coast-to-coast hydrographic sections. Prior to WOCE, successful heat
transport estimates had been made primarily
where the magnitude of the western boundary current transport was known: at 24°N in the Atlantic
(Hall and Bryden, 1982), where the Gulf Stream
transport through Florida Straits is 30 Sv
(1 Sv:10
6 m
3 s
91
) (Niiler and Richardson, 1973;
Larsen, 1992); and at 24°N in the Pacific (Bryden
et al., 1991), where the Kuroshio transport over
the Okinawa Trough is 27 Sv (Bingham and Talley,
1991). The reason is that, in combination with the
requirement for total mass conservation for the
section, the boundary current transport puts a
strong constraint on the interior circulation of the
ocean determined by geostrophic calculations
across the breadth of the interior ocean. It is well
known that geostrophic velocity estimates depend
on a reference level velocity or, alternatively, the
choice of a zero velocity surface. Without constraints imposed by the western boundary current
transports, very small changes in reference velocity
can create enormous changes in the interior circulation. For example, across the 6000-km wide
24°N Atlantic section, a modification in reference
velocity of 0.1 cm s
91 changes the interior ocean
transport by 28 Sv. Thus, a small change in the reference level velocity (or zero velocity surface) can
dramatically alter the strength of the basin-scale
circulation and resultant heat transport estimates.
Knowledge of the boundary current transports
constrains the interior circulation and effectively
determines the overall reference level velocity.
During WOCE, a large number of zonal hydrographic sections were carried out: across six latitudes in the Pacific, ten in the Atlantic and four in
the Indian Ocean (Fig. 3.1.4a, see Plate 3.1.4a,
p. 172). Current meter arrays at five of these latitudes (24°N and 30°S in the Atlantic, 20–32°S in
the Indian, and 30°N and 30–32°S in the Pacific)
were deployed to measure the strength of the deep
and/or shallow boundary currents so that estimates of meridional ocean heat transport could be
made more accurately. Analysis of the circulation
and meridional heat transport across these five
critical latitudes has not yet been completed so, in
a sense, we await the best direct estimates of circulation and heat transport that will be made from
WOCE. Nevertheless, notable progress has
recently been made on the Atlantic circulation and
heat transport, using WOCE sections as well as
hydrographic sections from the 1980s, and a
global inverse analysis of high-quality pre-WOCE
sections has derived a consistent pattern of circulation and heat transport. While there has been
progress in the Pacific and Indian Ocean analyses,
much remains to be done. We will review these
results here before discussing the challenges in
developing circulation and heat transport estimates for all of the WOCE sections.
6.1.4.1 Atlantic Ocean heat transport
Atlantic transoceanic sections across 55°N, 24°N,
30°S and 45°S early in WOCE have already been
analysed for meridional ocean heat transport
(Saunders and King, 1995b; Bacon, 1997; Lavín
et al., 1998; Holfort and Siedler, 2001). In addition,
three transoceanic sections across 14°N, 11°N and
11°S during the 1980s have been recently analysed
for heat transport (Friedrichs and Hall, 1993; Klein
et al., 1995; Speer et al., 1996). Each of these
studies includes a consideration of the flow
through Bering Strait, whose effect is small for
Atlantic heat transport; and each considers the
uncertainty in the resulting estimate of ocean heat
transport. There is general agreement that the
uncertainty in careful estimates of ocean heat
transport is about 0.3 PW.
Each individual study of Atlantic heat transport
shows the vertical distribution of the zonally averaged meridional circulation, which is responsible
for nearly all of the heat transport. For each
section, except 55°N, which is too far north for
Antarctic water masses to penetrate, the deep
SECTION 6 LARGE-SCALE OCEAN TRANSPORTS
460