estimates calculated from observation-based Excess
CO 2 inventories (Gruber, 1998; see Fig. 6.3.1).
Two points should be made with respect to
Sarmiento et al.’s (1995) ‘observational’ budget
(Fig. 6.3.13b). First their contemporary export at
18°S was underestimated due to these numbers
having been calculated with an assumption of zero
net mass transport across the section (Section
6.3.4.3). This means that their preindustrial transport estimate at 18°S is also significantly too
low. Second, the observation-based estimates do
not appear to allow for a contribution of carbon
derived from riverine input within the North
Atlantic. The model-based estimate (Fig. 6.3.13a)
is surprising in that it implies that the Atlantic
Ocean north of 18°S was neither a source nor a
sink of CO 2 for the atmosphere during the preindustrial era. In other words, the model implies that
the present-day net CO 2 uptake by this region is
made up exclusively of the uptake of Excess CO 2 .
This is a counterintuitive result given that this
region as a whole is a region of signficant net heat
loss, which would tend to drive a natural net
uptake of CO 2 from the atmosphere (cf. Watson
et al., 1995). In this regard, it may be relevant that
Murnane et al. (1999) show that the net northward heat transport in the Atlantic in the Princeton Ocean Biogeochemical Model is a factor of
two to three times lower than heat transports estimated based on observations and transport calculations (e.g. MacDonald and Wunsch, 1996). It is
by no means clear whether such a model can adequately address the net uptake and meridional
transport of carbon in the Atlantic if the heat
transport is poorly represented.
The North Atlantic budget in Figure 6.3.13c, on
the other hand, based on within-ocean data only,
implies a very different picture. Note first that for
the contemporary Atlantic, the budget-derived estimate of the net air–sea CO 2 flux based on transport divergence and storage estimates, together
with an estimate for riverine carbon input, very
closely matches a completely independent estimate
based on sea-surface pCO 2 data (Takahashi et al.,
1995). Of course both estimates have very large
uncertainties, which implies the agreement could be
fortuitous. Second, note that in the preindustrial
era this region was a significant net sink for
CO 2 , in contrast with the model-based budget
(Fig. 6.3.13a). Finally, note that the implied Excess
CO 2 uptake estimated from the regional budget is
considerably smaller than the Excess CO 2 uptake
estimated by the GCM. Indeed, in the regional
budget (Fig. 6.3.13c), air–sea exchange supplies
only about 50% of the Excess CO 2 storage,
whereas in the model, Excess CO 2 uptake across
the air–sea interface supplies 975% of the storage.
This difference points to a fundamental difference
in the importance attached to the advective supply
of Excess CO 2 in the GCM as compared to the
observation-based transport estimates. Because
both Excess CO 2 and heat are carried into the
North Atlantic with the upper limb of the meridional overturning circulation, this difference may
be related to the underestimation of the meridional
heat transport in the GCM.
The regional budget based on transport calculations suggests that the net air–sea flux of CO 2
within the North Atlantic has not increased over
time by as much as one might expect based on the
observed storage of Excess CO 2 within the Basin
(e.g. the analysis by Gruber, 1998). The reason for
this is presumably that a large portion of the upper
layer waters that are presently entering the North
Atlantic have already had sufficient time to
approach equilibrium with the contemporary
(increased) atmospheric pCO 2 prior to their transit
across the equator. Hence the potential for these
waters to absorb further additional Excess CO 2
during their transit through the North Atlantic is
limited. Given that there is likely to be additional
Excess CO 2 uptake across the air–sea interface
between 20°S and the Equator, the transport estimates imply that most of the Excess CO 2 currently
stored within the North Atlantic water column
(e.g. see Fig. 6.3.1) will have actually crossed the
air–sea interface in the southern hemisphere ocean
prior to being carried northwards by the ocean
circulation. Present carbon cycle models do not
appear to replicate this situation and this requires
further investigation.
A counterintuitive consequence of excess CO 2 transport
This implied role for the transport of ‘pre-equilibrated’ surface waters into the North Atlantic
raises the previously unrecognized (and counterintuitive) possibility that certain ocean regions might
release Excess CO 2 back to the atmosphere (i.e.
Excess CO 2 uptake can be locally negative). In
other words, the local net air-to-sea flux of
CO 2 can actually be smaller now than it was in
1750, despite the fact that atmospheric pCO 2 has
SECTION 6 LARGE-SCALE OCEAN TRANSPORTS
518
CO 2 inventories (Gruber, 1998; see Fig. 6.3.1).
Two points should be made with respect to
Sarmiento et al.’s (1995) ‘observational’ budget
(Fig. 6.3.13b). First their contemporary export at
18°S was underestimated due to these numbers
having been calculated with an assumption of zero
net mass transport across the section (Section
6.3.4.3). This means that their preindustrial transport estimate at 18°S is also significantly too
low. Second, the observation-based estimates do
not appear to allow for a contribution of carbon
derived from riverine input within the North
Atlantic. The model-based estimate (Fig. 6.3.13a)
is surprising in that it implies that the Atlantic
Ocean north of 18°S was neither a source nor a
sink of CO 2 for the atmosphere during the preindustrial era. In other words, the model implies that
the present-day net CO 2 uptake by this region is
made up exclusively of the uptake of Excess CO 2 .
This is a counterintuitive result given that this
region as a whole is a region of signficant net heat
loss, which would tend to drive a natural net
uptake of CO 2 from the atmosphere (cf. Watson
et al., 1995). In this regard, it may be relevant that
Murnane et al. (1999) show that the net northward heat transport in the Atlantic in the Princeton Ocean Biogeochemical Model is a factor of
two to three times lower than heat transports estimated based on observations and transport calculations (e.g. MacDonald and Wunsch, 1996). It is
by no means clear whether such a model can adequately address the net uptake and meridional
transport of carbon in the Atlantic if the heat
transport is poorly represented.
The North Atlantic budget in Figure 6.3.13c, on
the other hand, based on within-ocean data only,
implies a very different picture. Note first that for
the contemporary Atlantic, the budget-derived estimate of the net air–sea CO 2 flux based on transport divergence and storage estimates, together
with an estimate for riverine carbon input, very
closely matches a completely independent estimate
based on sea-surface pCO 2 data (Takahashi et al.,
1995). Of course both estimates have very large
uncertainties, which implies the agreement could be
fortuitous. Second, note that in the preindustrial
era this region was a significant net sink for
CO 2 , in contrast with the model-based budget
(Fig. 6.3.13a). Finally, note that the implied Excess
CO 2 uptake estimated from the regional budget is
considerably smaller than the Excess CO 2 uptake
estimated by the GCM. Indeed, in the regional
budget (Fig. 6.3.13c), air–sea exchange supplies
only about 50% of the Excess CO 2 storage,
whereas in the model, Excess CO 2 uptake across
the air–sea interface supplies 975% of the storage.
This difference points to a fundamental difference
in the importance attached to the advective supply
of Excess CO 2 in the GCM as compared to the
observation-based transport estimates. Because
both Excess CO 2 and heat are carried into the
North Atlantic with the upper limb of the meridional overturning circulation, this difference may
be related to the underestimation of the meridional
heat transport in the GCM.
The regional budget based on transport calculations suggests that the net air–sea flux of CO 2
within the North Atlantic has not increased over
time by as much as one might expect based on the
observed storage of Excess CO 2 within the Basin
(e.g. the analysis by Gruber, 1998). The reason for
this is presumably that a large portion of the upper
layer waters that are presently entering the North
Atlantic have already had sufficient time to
approach equilibrium with the contemporary
(increased) atmospheric pCO 2 prior to their transit
across the equator. Hence the potential for these
waters to absorb further additional Excess CO 2
during their transit through the North Atlantic is
limited. Given that there is likely to be additional
Excess CO 2 uptake across the air–sea interface
between 20°S and the Equator, the transport estimates imply that most of the Excess CO 2 currently
stored within the North Atlantic water column
(e.g. see Fig. 6.3.1) will have actually crossed the
air–sea interface in the southern hemisphere ocean
prior to being carried northwards by the ocean
circulation. Present carbon cycle models do not
appear to replicate this situation and this requires
further investigation.
A counterintuitive consequence of excess CO 2 transport
This implied role for the transport of ‘pre-equilibrated’ surface waters into the North Atlantic
raises the previously unrecognized (and counterintuitive) possibility that certain ocean regions might
release Excess CO 2 back to the atmosphere (i.e.
Excess CO 2 uptake can be locally negative). In
other words, the local net air-to-sea flux of
CO 2 can actually be smaller now than it was in
1750, despite the fact that atmospheric pCO 2 has
SECTION 6 LARGE-SCALE OCEAN TRANSPORTS
518