llpadding=0 cellspacing=0 width=100%>
Yahoo! is not affiliated with the authors of this page or responsible for its content.
Oceanic Fluxes in the South Atlantic
Oceanic Fluxes in the South Atlantic
E
LAINE
L. M
C
D
ONAGH AND
B
RIAN
A. K
ING
Southampton Oceanography Centre, Empress Dock, Southampton, United Kingdom
(Manuscript received 30 December 2002, in final form 23 June 2004)
ABSTRACT
A box inverse of the World Ocean Circulation Experiment A10 (30°S) and A11 (nominally 45°S) sections
in the South Atlantic Ocean was undertaken. The authors find a heat flux across A10 of 0.22
0.08 PW,
consistent with previous studies, and a heat flux of 0.43
0.08 PW across A11. The A11 heat flux is lower
than some previous analyses of this section but implies a plausible oceanic heat convergence (heat loss to
the atmosphere) of 0.21
0.10 PW. The difference is principally due to adding a cyclonic component to the
circulation in the Cape Basin. As compared with the solution of other studies, the anticyclonic circulation
in the surface and intermediate water of the subtropical gyre is weakened. The circulation of the deep water
is cyclonic rather than anticyclonic; this is in better agreement with previously published circulation schemes
based on examination of water properties. A southward freshwater flux of 0.7 Sv (1 Sv
10
6
m
3
s
1
) at A11,
consistent with previous inverse studies, is still inconsistent with the net Atlantic evaporation inferred from
integrated surface climatologies. Results suggest a small gain of freshwater (0.2
0.1 Sv) between the
sections.
1. Introduction
The overturning circulation in the South Atlantic
consists of surface and intermediate waters traveling
northward above North Atlantic Deep Water (NADW)
flowing southward; below that is found bottom water
that formed in the Antarctic, also flowing northward.
The surface and intermediate waters ultimately feed
the production of NADW in the northern limb of the
overturning. Determining the strength of this overturn-
ing circulation gives us insight into the elements that
contribute to the thermohaline circulation, as well as
allowing diagnosis of the net contribution of the ocean
to heat, freshwater, and nutrient fluxes in the South
Atlantic.
During the World Ocean Circulation Experiment
(WOCE), 10 coast-to-coast quasi-zonal sections were
occupied in the Atlantic. These were designed to mea-
sure the meridional fluxes of properties such as heat.
The two southernmost of these, designated A10 (30°S)
and A11 (nominally 45°S, Fig. 1), were occupied nearly
simultaneously between 27 December 1992 and 31
January 1993.
We use box inverse methods (Wunsch 1996) to pro-
duce a circulation with consistent fluxes of mass, heat,
salt, and silicate through the A10 and A11 sections.
Although the sections were occupied nearly simulta-
neously, we do not intend to produce a circulation
scheme that represents the ocean state in that month.
Rather, we intend our solution to represent the mean
circulation. To this end we invoke the arguments made
by Bryden and Imawaki (2001) that the density field is
slowly varying and represents the mean circulation over
some time scale if average boundary currents and Ek-
man fluxes are used. An advantage of using such a
regional inverse with only two sections is that it is fea-
sible to make a careful study of detailed aspects of the
flow across each section. As well as the careful consid-
eration of boundary currents crucial to any inverse, we
also consider the flow in deep channels, effects of indi-
vidual eddies in the flux of properties, and the differ-
ences in circulation that cause the bulk diagnostics of
this solution to differ from those in other solutions, as
described below.
2. Heat and freshwater fluxes in the
South Atlantic
In a study of the fluxes on the WOCE A11 section,
Saunders and King (1995b, hereinafter SKb) diagnosed
a northward heat flux of 0.53
0.1 PW (Table 1). This
heat flux, calculated to represent a climatological mean,
was somewhat higher than previous South Atlantic heat
flux estimates (typically
0.3 PW: e.g., Rintoul 1991
and Macdonald 1993). However, these estimates were
Corresponding author address: Elaine L. McDonagh, Southamp-
ton Oceanography Centre, Empress Dock, Southampton, Hamp-
shire SO14 3ZH, United Kingdom.
E-mail: elm@soc.soton.ac.uk
J
ANUARY
2005
M C D O N A G H A N D K I N G
109
© 2005 American Meteorological Society
JPO2666 made from sections that were up to 15° of latitude far-
ther north than the A11 section.
Integrated surface flux climatologies suggest uncer-
tainty about the oceanic heat convergence between
A11 and A10 in the South Atlantic. The Bunker fields
suggest convergence (heat loss to the atmosphere) of
0.05 PW, while the Southampton Oceanography Cen-
tre (SOC) climatology (Josey et al. 1999) suggests a
divergence (heat gain by the ocean) of 0.2 PW. How-
ever, this divergence becomes a 0.1-PW convergence if
the SOC climatology is uniformly adjusted to account
for its global imbalance of 30 Wm
2
. This is almost
exactly equal to the 29 W m
2
adjustment to the heat
flux that Grist and Josey (2003) calculated for the re-
gion between A10 and A11 when they adjusted the
SOC climatology using ocean heat transport constraints
including those from this study. Any of these results
falls short of the convergence of
0.2 PW implied by
the SKb result.
The global box inverse study of Ganachaud (1999)
derived heat fluxes of 0.66
0.12 PW and 0.35
0.15
PW across A11 and A10, respectively (Table 1). This
gives an implied oceanic convergence of 0.31
0.19
PW. The magnitude of this convergence is high when
compared with the integrated surface flux climatolo-
gies. In an inverse analysis of South Atlantic hydro-
graphic sections Holfort and Siedler (2001) calculated
northward heat fluxes of 0.37
0.02 PW and 0.29
T
ABLE
1. Summary of previous estimates of mass (Tg s
1
or Sv), heat (PW), and salt (Gg s
1
or Sv psu) fluxes across A10 and A11
sections (positive fluxes are northward).
A10
A11
Source
Mass
(Tg s
1
or Sv)
Heat (PW)
Salt
(Gg s
1
or
Sv psu)
Mass
(Tg s
1
or Sv)
Heat (PW)
Salt
(Gg s
1
or
Sv psu)
Saunders and King (1995b)
0
0.53
0.1
0.8
5.0
Ganachaud (1999)
1
3
0.35
0.15
1
3
0.66
0.12
Ganachaud and Wunsch (2003) and
A. Ganachaud (2003, personal
communication)
0
9
4*
0
9
5*
Holfort and Siedler (2001)
0.53
0.03
0.29
0.05
26.75
0.77
0.56
0.03
0.37
0.02
26.37
0.73
This study
0
0.22
0.08
9.7
2.9
0
0.43
0.08
2.9
2.9
* These numbers are reported as fluxes of salinity anomaly relative to the section average salinity and as such are equivalent to the zero
net mass flux solution.
F
IG
. 1. Positions of A10 and A11 CTD stations. The positions of the two rings sampled on A11 are marked. Surface circulation is
based on that of Stramma and England (1999).
110
J O U R N A L O F P H Y S I C A L O C E A N O G R A P H Y
V
OLUME
35 0.05 PW across A11 and A10, respectively (Table 1).
The magnitude of the oceanic heat convergence that
they find is consistent with the integrated surface flux
climatologies. However, they describe difficulties when
they include the A11 section in their inversion. They
ascribe this inconsistency to intense mixing between the
A10 and A11 sections at the BrazilFalklands (Malvi-
nas) confluence and water mass formation at the sub-
tropical and polar fronts, and the failure of their model
to accommodate these processes. We allow mixing be-
tween layers in this study.
In this study the net northward heat flux on A11 is
0.43 PW, smaller than SKbs number of 0.53 PW and
Ganachauds (1999) 0.66 PW. The value of 0.22 PW at
A10 is consistent with previous published estimates
(Table 1). The solution we present has a convergence of
heat across the perimeter of the box, implying a net loss
to the atmosphere of 0.21 PW, with a formal uncer-
tainty (one standard error) of 0.1 PW. The statistics of
the determination are therefore marginal, but a net
heat loss to the atmosphere is likely.
In section 5 we will look at the elements of the cir-
culation that change, resulting in the different heat
fluxes in ours and other solutions. We will focus on A11
where the different solutions give the most widely vary-
ing results (Table 1).
SKb estimated a freshwater flux that was inconsistent
with the results of Wijffels et al. (1992). In a global
study of freshwater divergence based on the Baumgart-
ner and Reichel (1975) climatology, Wijffels et al. cal-
culated a climatological freshwater loss to the atmo-
sphere of 0.7 Sv (1 Sv
10
6
m
3
s
1
) over the Atlantic
between the Arctic and 35°S. This implies that the vol-
ume flux across A11 should be
0.7 Sv less than that
through Bering Strait (approximate because A11 is on
average south of 35°S) while the salt flux is the same (as
salt is conserved). For a Bering Strait volume and salt
flux of 0.8 Sv and 26.0 Sv psu (Coachman and Aagaard
1988), then according to SKb, the freshwater flux across
A11 would need to be 0.75 Sv southward to conserve
salt (Table 2). This implies a near-zero freshwater di-
vergence over the Atlantic. In this study we intend to
combine the A10 and A11 sections to test the conclu-
sions of SKb, as suggested in the last paragraph of that
manuscript.
In a comprehensive review of ocean freshwater trans-
port, Wijffels (2001) found that integrated surface flux
products gave A