thermodynamic feedback (Carton et al., 1996a;
Chang et al., 1997). However, the investigation of
the modes of variability associated with air–sea
coupling in the equatorial Atlantic has not been
sought after experimentally until recently. It is only
in the last couple of years, and in particular with
the PIRATA (PIlot Research Array in the Tropical
Atlantic) array of upper ocean and meteorological
moorings (Servain et al., 1998) and other recent
programmes that new experimental investigation
is been carried out.
Comparisons of Pacific and Atlantic Equatorial
Undercurrents were made in the previous section.
The section on the tropical Atlantic will therefore
be focused on what we learnt during the WOCE
decade about the circulation and its variability,
and about the associated heat and fresh water
transports, based on repeated hydrographic sections, mooring arrays and floats. Forward modelling efforts that are most relevant are the ones at
high resolution, which incorporate the average
large-scale thermohaline circulation. A large part
of the investigation up to now has focused on the
western boundary and subsurface equatorial circulations, in order to understand how these circulations are linked and how they participate in the
large-scale meridional thermohaline cells.
At the beginning of WOCE, the tropical
Atlantic circulation was well known only near the
surface and away from the western boundary. The
upper thermocline circulation had been outlined;
the presence of rings shed by the North Brazil Current retroflection had been identified, while deeper
currents could only be sketched very roughly. The
pathways of southern water to the subtropical
Northern Hemisphere, evidenced from water mass
analysis, were not clearly known. However, it was
already estimated that 13 Sv of the transport of the
Florida Current originated from the Southern
Hemisphere, primarily in the warm surface layer
and below the thermocline (Schmitz and Richardson,
1991). The ocean circulation was known to
carry heat northward and fresh water southward
(Roemmich, 1983; from the inversion of 8°N and
8°S hydrographic sections). This was also known
to involve the thermohaline circulation with a net
flow to the north in the upper 1000 m (surface,
thermocline and Antarctic Intermediate Water)
and near the ocean bottom (AntArctic Bottom
Water, AABW). A compensating southward flow
occurred in-between with waters of different
northern Atlantic origins labelled ‘upper, middle
and lower North Atlantic Deep Water’, (uNADW,
mNADW and lNADW). There was some suggestion, based on transient tracer observations in the
uNADW, that the southward path of the North
Atlantic water across the equatorial ocean might
not be a direct one. These data revealed a tongue
of higher CFC (chlorofluorocarbon) concentration
in 1982 along the equator in the western equatorial Atlantic Ocean (Weiss et al., 1985).
One main way in which our knowledge of the
subsurface circulation and of its variability has
been acquired since the mid-1980s has been
through repeat hydrographic sections, particularly
in the western Atlantic Ocean, both across the
equator and near the western boundary. In addition, a few mooring arrays have been implemented.
These have been supported by float observations,
mostly near 1000 m depth, and by transient tracer
observations, which have been the most instrumental for understanding the deeper circulation. Highresolution modelling efforts have contributed to
the investigation of the circulation in the upper
ocean and its variability. The WOCE effort has led
to more precise estimates of water mass transformation, heat and freshwater transport. However,
there is still considerable uncertainty, and the synthesis of these different sets of observations is far
from being accomplished. Many papers describing
the state of the analysis were published in DeepSea Research 46, numbers 1–2 (1999), or in Journal of Geophysical Research 104, number C9
(1999).
4.3.3.2 Equatorial circulation in the upper
1000 m
Meridional repeat sections have been carried out
along 35°W, which can be combined to provide an
average picture of the equatorial circulation and
some idea of its seasonal variability (Stramma and
Schott, 1996; Schott et al., 1998; Bourles et al.,
1999b). The average circulation bears some similarities with the Pacific Ocean. An Equatorial
UnderCurrent (EUC) (Fig. 4.3.5) in the equatorial
thermocline (core near isopycnal 24.5) is surrounded by westward currents. This is the SEC. A
seasonal surface trapped North Equatorial Countercurrent occurs north of 5°N (Fig. 4.3.6). At the
equator, the EUC usually overlies a westward EIC
bounded by eastward currents at 4°N (NEUC) and
3–4°S (SEUC). These have their largest velocities
4.3 The Tropical Ocean Circulation
227
Godfrey, Johnson, McPhaden, Reverdin and Wijffels
Chang et al., 1997). However, the investigation of
the modes of variability associated with air–sea
coupling in the equatorial Atlantic has not been
sought after experimentally until recently. It is only
in the last couple of years, and in particular with
the PIRATA (PIlot Research Array in the Tropical
Atlantic) array of upper ocean and meteorological
moorings (Servain et al., 1998) and other recent
programmes that new experimental investigation
is been carried out.
Comparisons of Pacific and Atlantic Equatorial
Undercurrents were made in the previous section.
The section on the tropical Atlantic will therefore
be focused on what we learnt during the WOCE
decade about the circulation and its variability,
and about the associated heat and fresh water
transports, based on repeated hydrographic sections, mooring arrays and floats. Forward modelling efforts that are most relevant are the ones at
high resolution, which incorporate the average
large-scale thermohaline circulation. A large part
of the investigation up to now has focused on the
western boundary and subsurface equatorial circulations, in order to understand how these circulations are linked and how they participate in the
large-scale meridional thermohaline cells.
At the beginning of WOCE, the tropical
Atlantic circulation was well known only near the
surface and away from the western boundary. The
upper thermocline circulation had been outlined;
the presence of rings shed by the North Brazil Current retroflection had been identified, while deeper
currents could only be sketched very roughly. The
pathways of southern water to the subtropical
Northern Hemisphere, evidenced from water mass
analysis, were not clearly known. However, it was
already estimated that 13 Sv of the transport of the
Florida Current originated from the Southern
Hemisphere, primarily in the warm surface layer
and below the thermocline (Schmitz and Richardson,
1991). The ocean circulation was known to
carry heat northward and fresh water southward
(Roemmich, 1983; from the inversion of 8°N and
8°S hydrographic sections). This was also known
to involve the thermohaline circulation with a net
flow to the north in the upper 1000 m (surface,
thermocline and Antarctic Intermediate Water)
and near the ocean bottom (AntArctic Bottom
Water, AABW). A compensating southward flow
occurred in-between with waters of different
northern Atlantic origins labelled ‘upper, middle
and lower North Atlantic Deep Water’, (uNADW,
mNADW and lNADW). There was some suggestion, based on transient tracer observations in the
uNADW, that the southward path of the North
Atlantic water across the equatorial ocean might
not be a direct one. These data revealed a tongue
of higher CFC (chlorofluorocarbon) concentration
in 1982 along the equator in the western equatorial Atlantic Ocean (Weiss et al., 1985).
One main way in which our knowledge of the
subsurface circulation and of its variability has
been acquired since the mid-1980s has been
through repeat hydrographic sections, particularly
in the western Atlantic Ocean, both across the
equator and near the western boundary. In addition, a few mooring arrays have been implemented.
These have been supported by float observations,
mostly near 1000 m depth, and by transient tracer
observations, which have been the most instrumental for understanding the deeper circulation. Highresolution modelling efforts have contributed to
the investigation of the circulation in the upper
ocean and its variability. The WOCE effort has led
to more precise estimates of water mass transformation, heat and freshwater transport. However,
there is still considerable uncertainty, and the synthesis of these different sets of observations is far
from being accomplished. Many papers describing
the state of the analysis were published in DeepSea Research 46, numbers 1–2 (1999), or in Journal of Geophysical Research 104, number C9
(1999).
4.3.3.2 Equatorial circulation in the upper
1000 m
Meridional repeat sections have been carried out
along 35°W, which can be combined to provide an
average picture of the equatorial circulation and
some idea of its seasonal variability (Stramma and
Schott, 1996; Schott et al., 1998; Bourles et al.,
1999b). The average circulation bears some similarities with the Pacific Ocean. An Equatorial
UnderCurrent (EUC) (Fig. 4.3.5) in the equatorial
thermocline (core near isopycnal 24.5) is surrounded by westward currents. This is the SEC. A
seasonal surface trapped North Equatorial Countercurrent occurs north of 5°N (Fig. 4.3.6). At the
equator, the EUC usually overlies a westward EIC
bounded by eastward currents at 4°N (NEUC) and
3–4°S (SEUC). These have their largest velocities
4.3 The Tropical Ocean Circulation
227
Godfrey, Johnson, McPhaden, Reverdin and Wijffels