components, the 3 Sv from the eastern basin and the
10 Sv coming southward along the Greenland coast,
gives very nearly the 13.3 Sv that Clarke (1984)
reported at the southern tip of Greenland (Cape
Farewell), again with unquantified errors.
The DWBC then traces the boundary of the
Labrador Sea before entering the subtropics at the
Grand Banks. At this point, just to the west of
the southern tip of the Banks, Pickart and Smethie
(1998) have calculated that 13.3 Sv of water with
927.8 kg m
93 flow westward. This estimate is
based on the average of three hydrographic sections
taken over a period of 4 years referenced with velocity profiling instruments. The reported variance is
4.2 Sv and, with just two degrees of freedom, the
error in the mean is 3 Sv: by this point along the
boundary there has been no significant change in
transport from that observed at Cape Farewell.
Further south the next well-measured site is at
27°N, where a current meter array monitored the
western boundary currents to the east of the island
of Abaco in the Bahamas for a period of almost
5 years from 1987 to 1992 (Lee et al., 1996). Here,
the DWBC was defined on the basis of the direction of the mean flow: above 800 m depth the flow
was primarily to the north (the Antilles Current)
while that below was southward (the DWBC).
After some judicious sampling was applied, as the
core of the deep flow was often not adequately
resolved by the array, the total deep transport was
found to be an astonishing 40<3 Sv! This enhancement over that seen further north and that inferred
necessary to balance the meridional overturning
(ϳ13 Sv) suggests that much of the 40 Sv is returned
to the north in an offshore recirculation zone that
was not adequately instrumented.
However, similar large southward transports
within the NADW further south have been reported.
At 8°N, Johns et al. (1993), on the basis of a single
mooring, several hydrographic sections, some Pegasus velocity profiles and a heuristic model of the current structure, computed a southward transport of
22 Sv for the flow below 2500 m – the Lower
NADW. Much of this is believed to be returned to
the north in another recirculation when the DWBC
encounters a topographic constriction near 6°N.
Closer to the equator at 44°W 2-year-long
moored array measurements gave 14–17 Sv for
the Upper NADW, depending on analysis method
(Schott et al., 1993a; Fischer and Schott, 1997;
no. 5, Fig. 4.5.2). Sparse sampling of the lower
NADW gave estimates varying between 4.5 Sv and
12 Sv so the total transport of NADW across the
equator has a large range: 18.5–29 Sv, consistent
with the work of Rhein et al. (1995), who found
that 26.8<7.0 Sv cross 35°W between 2°N and
5°S with 19.5<5.3 Sv reaching 5°S. The actual
pathways for flow crossing the equator appear
to be complex. Based on the trajectories of 14
SOFAR (SOund Fixing And Ranging) floats ballasted to be at 1800 m, Richardson and Fratantoni
(1999) suggested a combination of meridional
flow near the boundary and elongated zonal flows
in the interior: eastward jets at 2°N and 2°S and
a westward one on the equator, patterns that are
supported by tracer distributions (e.g. Weiss et al.,
1985; Andrié et al., 1998, 1999). Deep jets,
trapped to the equator and first observed in the
Atlantic by Ponte et al. (1990), have now been
shown to reverse with the seasons by Gouriou
et al. (1999) but to change little over a period of
5 years for the same season of the year.
At 18°S in the South Atlantic, Weatherly et al.
(2000; no. 6, Fig. 4.5.2) reported a total southward transport of NADW of 39 Sv. The rms variance is ϳ20 Sv, which, combined with a 20-day
integral time scale for the area (Hogg and Owens,
1999) and the 18-month record length, gives the
standard error of about 5 Sv shown on Fig. 4.5.4.
Again the authors appealed to a possible recirculation further offshore to return part of this large
flux, but had no measurements to support its existence. Indeed, direct measurements in the interior
by floats (Hogg and Owens, 1999, see below) have
shown only zonal flows. On the basis of dynamic
computations from hydrographic sections at 19°S
and 30°S, Zangenberg and Siedler (1998) computed a rapid decline in NADW transport from
27 Sv at 19°S to 6 Sv at 30°S and argued that this
divergence is taken up by a zonal flow to the east
in the latitude band 25–30°S. The float data, to be
described below, are in qualitative support.
It is curious that the DWBC transports of
NADW at subtropical and high latitudes are all of
the order needed to balance the meridional overturning (13 Sv), while those in the tropical band are
consistently higher by a factor of 2–3. Although
this could well point to the existence of repeated
and disconnected recirculation zones in this region
(e.g. Schmitz and McCartney, 1993), finding
the offshore northward return flow has been an
elusive quest.
SECTION 4 THE GLOBAL FLOW FIELD
264
10 Sv coming southward along the Greenland coast,
gives very nearly the 13.3 Sv that Clarke (1984)
reported at the southern tip of Greenland (Cape
Farewell), again with unquantified errors.
The DWBC then traces the boundary of the
Labrador Sea before entering the subtropics at the
Grand Banks. At this point, just to the west of
the southern tip of the Banks, Pickart and Smethie
(1998) have calculated that 13.3 Sv of water with
927.8 kg m
93 flow westward. This estimate is
based on the average of three hydrographic sections
taken over a period of 4 years referenced with velocity profiling instruments. The reported variance is
4.2 Sv and, with just two degrees of freedom, the
error in the mean is 3 Sv: by this point along the
boundary there has been no significant change in
transport from that observed at Cape Farewell.
Further south the next well-measured site is at
27°N, where a current meter array monitored the
western boundary currents to the east of the island
of Abaco in the Bahamas for a period of almost
5 years from 1987 to 1992 (Lee et al., 1996). Here,
the DWBC was defined on the basis of the direction of the mean flow: above 800 m depth the flow
was primarily to the north (the Antilles Current)
while that below was southward (the DWBC).
After some judicious sampling was applied, as the
core of the deep flow was often not adequately
resolved by the array, the total deep transport was
found to be an astonishing 40<3 Sv! This enhancement over that seen further north and that inferred
necessary to balance the meridional overturning
(ϳ13 Sv) suggests that much of the 40 Sv is returned
to the north in an offshore recirculation zone that
was not adequately instrumented.
However, similar large southward transports
within the NADW further south have been reported.
At 8°N, Johns et al. (1993), on the basis of a single
mooring, several hydrographic sections, some Pegasus velocity profiles and a heuristic model of the current structure, computed a southward transport of
22 Sv for the flow below 2500 m – the Lower
NADW. Much of this is believed to be returned to
the north in another recirculation when the DWBC
encounters a topographic constriction near 6°N.
Closer to the equator at 44°W 2-year-long
moored array measurements gave 14–17 Sv for
the Upper NADW, depending on analysis method
(Schott et al., 1993a; Fischer and Schott, 1997;
no. 5, Fig. 4.5.2). Sparse sampling of the lower
NADW gave estimates varying between 4.5 Sv and
12 Sv so the total transport of NADW across the
equator has a large range: 18.5–29 Sv, consistent
with the work of Rhein et al. (1995), who found
that 26.8<7.0 Sv cross 35°W between 2°N and
5°S with 19.5<5.3 Sv reaching 5°S. The actual
pathways for flow crossing the equator appear
to be complex. Based on the trajectories of 14
SOFAR (SOund Fixing And Ranging) floats ballasted to be at 1800 m, Richardson and Fratantoni
(1999) suggested a combination of meridional
flow near the boundary and elongated zonal flows
in the interior: eastward jets at 2°N and 2°S and
a westward one on the equator, patterns that are
supported by tracer distributions (e.g. Weiss et al.,
1985; Andrié et al., 1998, 1999). Deep jets,
trapped to the equator and first observed in the
Atlantic by Ponte et al. (1990), have now been
shown to reverse with the seasons by Gouriou
et al. (1999) but to change little over a period of
5 years for the same season of the year.
At 18°S in the South Atlantic, Weatherly et al.
(2000; no. 6, Fig. 4.5.2) reported a total southward transport of NADW of 39 Sv. The rms variance is ϳ20 Sv, which, combined with a 20-day
integral time scale for the area (Hogg and Owens,
1999) and the 18-month record length, gives the
standard error of about 5 Sv shown on Fig. 4.5.4.
Again the authors appealed to a possible recirculation further offshore to return part of this large
flux, but had no measurements to support its existence. Indeed, direct measurements in the interior
by floats (Hogg and Owens, 1999, see below) have
shown only zonal flows. On the basis of dynamic
computations from hydrographic sections at 19°S
and 30°S, Zangenberg and Siedler (1998) computed a rapid decline in NADW transport from
27 Sv at 19°S to 6 Sv at 30°S and argued that this
divergence is taken up by a zonal flow to the east
in the latitude band 25–30°S. The float data, to be
described below, are in qualitative support.
It is curious that the DWBC transports of
NADW at subtropical and high latitudes are all of
the order needed to balance the meridional overturning (13 Sv), while those in the tropical band are
consistently higher by a factor of 2–3. Although
this could well point to the existence of repeated
and disconnected recirculation zones in this region
(e.g. Schmitz and McCartney, 1993), finding
the offshore northward return flow has been an
elusive quest.
SECTION 4 THE GLOBAL FLOW FIELD
264