of anthropogenic carbon from “natural”, leading to the conclusion that the physical
transfer of CO 2 in the Southern Ocean is now into the ocean, whereas without fossil-
fuel carbon it would be out. Elsewhere, the directions of transfers are unchanged, but
the amounts have changed greatly, for example the reduction of export by almost half
in the tropics. The global carbon budget diagram (Fig. 16.9) shows that net changes
are small relative to overall transfers. Thus, both input and output of CO 2 to the ocean
are about 90 GtC yr−1, with a difference of about 2.2 GtC yr−1, the anthropogenic
input. Again, virtually all primary production is respired on very short time scales,
with very little organic matter stored in water or sediment. Thus, the small net
changes are widely dispersed across the oceans, making them extremely difficult to
measure. It is remarkable that completely different estimates of ocean uptake agree so
closely.
(^) The climatic impact of the industrial era increase of atmospheric carbon is estimated
by complex models of the circulation of the global atmosphere, the biosphere–
atmosphere interaction with respect to CO 2 (also methane, etc.), air–sea energy
exchange, and the incoming–outgoing energy budget. Making and running these
models is now a regular industry with dozens of practitioners. Apart from numerical
problems such as necessarily coarse grid scales, the main difficulties with these
models are the large uncertainties deriving from the complexity of the system and
from critical processes that are not well quantified. Examples include: (i) the poor
characterization of the effect of increased CO 2 on primary productivity, probably
small for phytoplankton but modestly important on land; (ii) the impact of warming
on cloud formation, altitude and coverage, and thus global albedo; and (iii) the impact
of warming on deep-water formation and thus deep-sea ventilation rates.
CO 2 and the Glacial–Interglacial Cycle
(^) At least a part of global climate variation on glacial–interglacial time scales is coupled
to this same transfer of CO 2 into (and out of) the deep-sea that accounts for fossil fuel
carbon not in the air. This topic, too, is freighted with complexities. Ruddiman
(2007a) has sorted some of them out in intuitive fashion and should be consulted for
fuller treatment. The ocean below its permanent pycnoclines has the largest reservoir
of labile carbon on Earth, some 38,000 GtC, compared to 830 GtC in the atmosphere
(Fig. 16.9). If deep-water formation were to slow or stop, as it apparently does at least
intermittently in glacial eras, the driving force for both horizontal deep-sea circulation
and ventilation of accumulated CO 2 at evasion sites would decrease markedly. During
a long re-equilibration interval, there would be a net transfer of CO 2 to the deep sea
by the biological pump, and atmospheric CO 2 levels would drop. According to gas