uncertainty in Excess CO 2 concentrations based on
their method is of the order of 10 mol kg
91 (i.e.
about 20% of the current surface ocean Excess CO 2
concentration). For the calculation of inventories, as
suggested by the discussion below, an important
and difficult-to-assess issue is whether these errors
are randomly or systematically distributed.
Atlantic Ocean
Gruber (1998) presented estimates of Excess CO 2
concentrations in the Atlantic Ocean based mainly
on pre-WOCE data (TTO and SAVE), but also
including data from two WOCE cruises in the
South Atlantic (Meteor 15/3 cruise along A9 (19°S)
and the Meteor 11/5 cruise along A12 (~ 58°S) and
A21 (Drake Passage)). Qualitatively, the distributions closely match those of other transient tracers,
such as the CFCs, with large accumulations and
deep penetration found within the thermocline
waters of the subtropical gyres of both hemispheres.
This accumulation in the mid-latitude water column
reflects convergence of Excess CO 2 driven by the
convergence of surface waters. Relatively shallow
penetration and low inventories were found in the
tropical and equatorial zone (20°S to 20°N),
whereas very deep penetration was observed in high
northern latitudes associated with deep-water formation. Relatively low deep-water concentrations
and low inventories were detected in high southern
latitudes (950°S) apparently because of the relatively short time available for newly formed deep
waters in this region to pick up Excess CO 2 from
the atmosphere, in part because ice cover restricts
gas exchange (Poisson and Chen, 1987). Dilution
by rapid mixing with old waters also makes the
Excess signal hard to detect in this region, and
Caldeira and Duffy (2000) suggest that any
absorbed Excess CO 2 is rapidly transported out of
the region along isopycnals.
According to Gruber (1998) Excess CO 2 at levels95 mol kg
91 were found all the way to the
bottom throughout much of the North Atlantic
north of 40–50°N in 1982. South of this, zonally
averaged Excess CO 2 levels dropped to levels
approaching zero below about 2000 m, although a
weak Excess CO 2 signal (:5 mol kg
91 ) associated with Antarctic Bottom Water was detected
spreading from the south below about 4000 m.
In contrast to the surface distribution of CFCs,
Excess CO 2 levels are highest in warm tropical
waters (approx. 40–50 mol kg
91 in 1989) and
are lower (30–40 mol kg
91
) in colder, high-latitude waters. This is a consequence of the strong
temperature dependence of the buffer capacity of
seawater with respect to CO 2 addition (Revelle
factor), which more than compensates for the fact
that both CO 2 and the CFCs are both more soluble in colder water. For a given pCO 2 increase,
warm waters are able to absorb significantly more
Excess CO 2 from the atmosphere than cold
waters.
Holfort et al. (1998) analysed data from three
WOCE sections in the South Atlantic between 10
and 30°S (A8, A9, A10) as well as several preWOCE sections from the same region. Their
primary aim was to examine carbon transports
(Section 6.3.4.3). However, in order that the transport of Excess CO 2 could be estimated, it was necessary to estimate the Excess CO 2 distribution
across all of the sections. The Excess CO 2 calculation method used was not described in their paper,
and is described schematically in Figure 6.3.7. A
significant difference with the method used by
Gruber (1998) was that it was nowhere assumed
that deep water was uncontaminated with Excess
CO 2 in order to calculate air–sea disequilibria.
Rather a factor analagous to Gruber’s air–sea disequilibrium term was extrapolated from regions
where transient tracers were detectable into
regions of the water column where transient tracers (including CCl 4 ) were undetectable.
Figure 6.3.8 shows a comparison between the
zonal-average profiles resulting from Holfort
et al.’s (1998) estimates and Gruber’s (1998) estimates for the South Atlantic between 10 and 30°S.
Overall the profiles are very similar: noticeable in
all the profiles are relatively high and uniform concentrations in the Antarctic Intermediate Water
(AAIW) between 500 and 1000 m, with Excess
CO 2 levels in this water mass being higher towards
the south, closer to the AAIW formation region
(cf. Brewer, 1978). The main difference between
the two analyses is the presence in the Holfort
et al. (1998) profiles of a background, non-zero
level of Excess CO 2 (3–5 mol kg
91
) throughout
the deep-water column. In contrast, Gruber’s
(1998) values drop to zero between 2000 and
4000 m. Whereas this 3–5 mol kg
91 background
level is well within the acknowledged uncertainty
of both separation techniques, it represents a systematic difference that can contribute significantly
to total water column inventories. This possibility
6.3 Storage and Transport of Excess CO 2 in the Oceans
505
Wallace
their method is of the order of 10 mol kg
91 (i.e.
about 20% of the current surface ocean Excess CO 2
concentration). For the calculation of inventories, as
suggested by the discussion below, an important
and difficult-to-assess issue is whether these errors
are randomly or systematically distributed.
Atlantic Ocean
Gruber (1998) presented estimates of Excess CO 2
concentrations in the Atlantic Ocean based mainly
on pre-WOCE data (TTO and SAVE), but also
including data from two WOCE cruises in the
South Atlantic (Meteor 15/3 cruise along A9 (19°S)
and the Meteor 11/5 cruise along A12 (~ 58°S) and
A21 (Drake Passage)). Qualitatively, the distributions closely match those of other transient tracers,
such as the CFCs, with large accumulations and
deep penetration found within the thermocline
waters of the subtropical gyres of both hemispheres.
This accumulation in the mid-latitude water column
reflects convergence of Excess CO 2 driven by the
convergence of surface waters. Relatively shallow
penetration and low inventories were found in the
tropical and equatorial zone (20°S to 20°N),
whereas very deep penetration was observed in high
northern latitudes associated with deep-water formation. Relatively low deep-water concentrations
and low inventories were detected in high southern
latitudes (950°S) apparently because of the relatively short time available for newly formed deep
waters in this region to pick up Excess CO 2 from
the atmosphere, in part because ice cover restricts
gas exchange (Poisson and Chen, 1987). Dilution
by rapid mixing with old waters also makes the
Excess signal hard to detect in this region, and
Caldeira and Duffy (2000) suggest that any
absorbed Excess CO 2 is rapidly transported out of
the region along isopycnals.
According to Gruber (1998) Excess CO 2 at levels95 mol kg
91 were found all the way to the
bottom throughout much of the North Atlantic
north of 40–50°N in 1982. South of this, zonally
averaged Excess CO 2 levels dropped to levels
approaching zero below about 2000 m, although a
weak Excess CO 2 signal (:5 mol kg
91 ) associated with Antarctic Bottom Water was detected
spreading from the south below about 4000 m.
In contrast to the surface distribution of CFCs,
Excess CO 2 levels are highest in warm tropical
waters (approx. 40–50 mol kg
91 in 1989) and
are lower (30–40 mol kg
91
) in colder, high-latitude waters. This is a consequence of the strong
temperature dependence of the buffer capacity of
seawater with respect to CO 2 addition (Revelle
factor), which more than compensates for the fact
that both CO 2 and the CFCs are both more soluble in colder water. For a given pCO 2 increase,
warm waters are able to absorb significantly more
Excess CO 2 from the atmosphere than cold
waters.
Holfort et al. (1998) analysed data from three
WOCE sections in the South Atlantic between 10
and 30°S (A8, A9, A10) as well as several preWOCE sections from the same region. Their
primary aim was to examine carbon transports
(Section 6.3.4.3). However, in order that the transport of Excess CO 2 could be estimated, it was necessary to estimate the Excess CO 2 distribution
across all of the sections. The Excess CO 2 calculation method used was not described in their paper,
and is described schematically in Figure 6.3.7. A
significant difference with the method used by
Gruber (1998) was that it was nowhere assumed
that deep water was uncontaminated with Excess
CO 2 in order to calculate air–sea disequilibria.
Rather a factor analagous to Gruber’s air–sea disequilibrium term was extrapolated from regions
where transient tracers were detectable into
regions of the water column where transient tracers (including CCl 4 ) were undetectable.
Figure 6.3.8 shows a comparison between the
zonal-average profiles resulting from Holfort
et al.’s (1998) estimates and Gruber’s (1998) estimates for the South Atlantic between 10 and 30°S.
Overall the profiles are very similar: noticeable in
all the profiles are relatively high and uniform concentrations in the Antarctic Intermediate Water
(AAIW) between 500 and 1000 m, with Excess
CO 2 levels in this water mass being higher towards
the south, closer to the AAIW formation region
(cf. Brewer, 1978). The main difference between
the two analyses is the presence in the Holfort
et al. (1998) profiles of a background, non-zero
level of Excess CO 2 (3–5 mol kg
91
) throughout
the deep-water column. In contrast, Gruber’s
(1998) values drop to zero between 2000 and
4000 m. Whereas this 3–5 mol kg
91 background
level is well within the acknowledged uncertainty
of both separation techniques, it represents a systematic difference that can contribute significantly
to total water column inventories. This possibility
6.3 Storage and Transport of Excess CO 2 in the Oceans
505
Wallace