AABW
It has been conventional to call the deepest water
mass in the North Atlantic ‘Antarctic Bottom
Water’ and, as its name implies, its ultimate source
is in the Antarctic region. Just as NADW represents
a mixture between various source waters, AABW is
a mixture of CDW, Weddell Sea Deep Water
(WSDW) and NADW. Its upper boundary within
the Atlantic is usually taken to be around 2.0°C,
potential temperature, or some related potential
density surface, rather than the 1.2°C surface as is
used in the Pacific.
Weddell Sea Deep Water enters the South
Atlantic through one or more gaps in the Falkland
Ridge, which separates the Georgia and Argentine
Basins. Along the northern flank of this topographic feature Whitworth et al. (1991) estimated
that 8.2 Sv of water colder than 0.2°C (their definition of WSDW) flows westward. The standard
deviation is 7 Sv and they identified a principal
period for variability of 70 days, apparently resulting from a meandering of the Circumpolar Current.
With a 14-month observation period this gives
about 12 degrees of freedom and a standard error
in the mean of 2 Sv. Whitworth et al. (1991) suggested that most of the 8.2 Sv is recirculated water
as the pure WSDW component has a westward
transport of just 2.5 Sv.
At the northern end of the Argentine Basin, the
Santos Plateau–Rio Grande Rise system is a major
impediment for the continued northward flow
of bottom water. Here Hogg et al. (1999; nos 7–9,
Fig. 4.5.2), through use of 2-year-long measurements, calculated that a total of 6.9 Sv of water
colder than 1.9°C flows into the Brazil Basin: this
is partitioned as 4 Sv through the Vema Channel
(no. 8) and 2.9<1.2 Sv through the Hunter Channel (no. 7) further to the east, with the relatively
large error resulting from poor spatial resolution
of the flow by the moored array (Zenk et al.,
1999a). Although Hogg et al. (1999) failed to
quantify the error in the earlier Vema Channel
transport measurements, Hogg et al. (1982)
reported a standard deviation for the transport of
1.2 Sv and Hogg and Owens (1999) computed an
integral time scale for the area of about 10 days.
This gives a standard error in the mean, based on
a 2-year observation programme, of <0.2 Sv.
Thus the flow into the Brazil Basin is 6.9<1.2 Sv,
with the error being dominated by that in the
Hunter Channel transport number.
The bottom water flowing into the Brazil Basin
has just two exits and both are at the equator: the
Romanche-Chain Fracture Zone complex in the
Mid-Atlantic Ridge (no. 3, Fig. 4.5.2) and a deep
equatorial passage that allows AABW entrance
into the western North Atlantic (no. 4). Both topographic constrictions were well planted with
moorings during WOCE as part of the Deep Basin
Experiment (see Section 4.5.3). With respect to the
first (Romanche-Chain FZ), Mercier and Speer
(1998) estimated that 1.22 Sv flows to the east.
This is a remarkably steady flow with a standard
deviation of just <0.25 Sv. Combined with a
reportedly short time scale of 10–20 days and a
2-year record, the standard error of the mean is
just <0.04 Sv. At the western exit Hall et al.
(1997) reported that 2.0<0.2 Sv of AABW flows
out of the Brazil Basin.
Further north there has been little direct measurement of AABW transports. Based on property
distributions it is apparent that the AABW follows
the western flank of the Mid-Atlantic Ridge, rather
than the western boundary of the basin where
Schmitz and McCartney (1993) suggested a rather
complicated set of recirculations. Some of the water
progressing northward along the Mid-Atlantic
Ridge does exit to the east through the Vema
Fracture Zone (not to be confused with the Vema
Channel) near 11°N. Fischer and Schott (1997)
computed 1.8–2.0 Sv based on synoptic data.
Atlantic summary
The Stommel (1958) scheme (Fig. 4.5.1) was for
the depth-integrated flow of the abyss below about
2000 m. The composite of the AABW and NADW
transports (Fig. 4.5.4) do conform to his picture in
a qualitative sense: there is, all along the western
boundary, a net transport toward the south. However, in a more quantitative sense there is disagreement as the transport should steadily decrease
while supplying fluid to the interior and it manifestly does not do so. The various investigators
who have produced the large transport numbers
in the tropics each ascribe them to possible recirculation further offshore but direct measurement, so
far, has failed to reveal it in a convincing manner.
4.5.2.3 The Indian Ocean
Although the moored arrays set in the Indian
Ocean had all been recovered by mid-1997 there
is, as of this writing, little or no presentation of the
4.5 Quantification of the Deep Circulation
265
Hogg
It has been conventional to call the deepest water
mass in the North Atlantic ‘Antarctic Bottom
Water’ and, as its name implies, its ultimate source
is in the Antarctic region. Just as NADW represents
a mixture between various source waters, AABW is
a mixture of CDW, Weddell Sea Deep Water
(WSDW) and NADW. Its upper boundary within
the Atlantic is usually taken to be around 2.0°C,
potential temperature, or some related potential
density surface, rather than the 1.2°C surface as is
used in the Pacific.
Weddell Sea Deep Water enters the South
Atlantic through one or more gaps in the Falkland
Ridge, which separates the Georgia and Argentine
Basins. Along the northern flank of this topographic feature Whitworth et al. (1991) estimated
that 8.2 Sv of water colder than 0.2°C (their definition of WSDW) flows westward. The standard
deviation is 7 Sv and they identified a principal
period for variability of 70 days, apparently resulting from a meandering of the Circumpolar Current.
With a 14-month observation period this gives
about 12 degrees of freedom and a standard error
in the mean of 2 Sv. Whitworth et al. (1991) suggested that most of the 8.2 Sv is recirculated water
as the pure WSDW component has a westward
transport of just 2.5 Sv.
At the northern end of the Argentine Basin, the
Santos Plateau–Rio Grande Rise system is a major
impediment for the continued northward flow
of bottom water. Here Hogg et al. (1999; nos 7–9,
Fig. 4.5.2), through use of 2-year-long measurements, calculated that a total of 6.9 Sv of water
colder than 1.9°C flows into the Brazil Basin: this
is partitioned as 4 Sv through the Vema Channel
(no. 8) and 2.9<1.2 Sv through the Hunter Channel (no. 7) further to the east, with the relatively
large error resulting from poor spatial resolution
of the flow by the moored array (Zenk et al.,
1999a). Although Hogg et al. (1999) failed to
quantify the error in the earlier Vema Channel
transport measurements, Hogg et al. (1982)
reported a standard deviation for the transport of
1.2 Sv and Hogg and Owens (1999) computed an
integral time scale for the area of about 10 days.
This gives a standard error in the mean, based on
a 2-year observation programme, of <0.2 Sv.
Thus the flow into the Brazil Basin is 6.9<1.2 Sv,
with the error being dominated by that in the
Hunter Channel transport number.
The bottom water flowing into the Brazil Basin
has just two exits and both are at the equator: the
Romanche-Chain Fracture Zone complex in the
Mid-Atlantic Ridge (no. 3, Fig. 4.5.2) and a deep
equatorial passage that allows AABW entrance
into the western North Atlantic (no. 4). Both topographic constrictions were well planted with
moorings during WOCE as part of the Deep Basin
Experiment (see Section 4.5.3). With respect to the
first (Romanche-Chain FZ), Mercier and Speer
(1998) estimated that 1.22 Sv flows to the east.
This is a remarkably steady flow with a standard
deviation of just <0.25 Sv. Combined with a
reportedly short time scale of 10–20 days and a
2-year record, the standard error of the mean is
just <0.04 Sv. At the western exit Hall et al.
(1997) reported that 2.0<0.2 Sv of AABW flows
out of the Brazil Basin.
Further north there has been little direct measurement of AABW transports. Based on property
distributions it is apparent that the AABW follows
the western flank of the Mid-Atlantic Ridge, rather
than the western boundary of the basin where
Schmitz and McCartney (1993) suggested a rather
complicated set of recirculations. Some of the water
progressing northward along the Mid-Atlantic
Ridge does exit to the east through the Vema
Fracture Zone (not to be confused with the Vema
Channel) near 11°N. Fischer and Schott (1997)
computed 1.8–2.0 Sv based on synoptic data.
Atlantic summary
The Stommel (1958) scheme (Fig. 4.5.1) was for
the depth-integrated flow of the abyss below about
2000 m. The composite of the AABW and NADW
transports (Fig. 4.5.4) do conform to his picture in
a qualitative sense: there is, all along the western
boundary, a net transport toward the south. However, in a more quantitative sense there is disagreement as the transport should steadily decrease
while supplying fluid to the interior and it manifestly does not do so. The various investigators
who have produced the large transport numbers
in the tropics each ascribe them to possible recirculation further offshore but direct measurement, so
far, has failed to reveal it in a convincing manner.
4.5.2.3 The Indian Ocean
Although the moored arrays set in the Indian
Ocean had all been recovered by mid-1997 there
is, as of this writing, little or no presentation of the
4.5 Quantification of the Deep Circulation
265
Hogg