of deep water in the North Atlantic. As it becomes ice covered far to the south, the
Gulf Stream does not carry high-salinity water beneath very cold arctic winds, so
cooling does not raise its density sufficiently to sink it to the seafloor. There can also
be intermittent spreading of glacial meltwater across the Atlantic, enhancing
stratification. It has been suggested that formation of North Atlantic Deep Water
(NADW) may be replaced by subduction of shallower, less-voluminous “Glacial
North Atlantic Intermediate Water”. There is also likely to be inhibition of gas
exchange by year-around presence of sea ice around Antarctica to much lower
latitudes than today. When deep water is not formed or re-exposed at the surface,
ventilation is diminished, and organic matter exported to depth and converted to CO 2
remains in the sea. The relatively rapid returns (a few kyr) occur in periods when
heating allows resumption of deep-water formation, thus, ventilation. Retreat of
antarctic sea ice, allowing ventilation via upwelling in the Southern Ocean, is
certainly also significant – perhaps the leading event in warming cycles. Carbonate in
antarctic foraminifera showed a strong drop in ^14 C age during the last glacial
termination (Skinner et al. 2010), in agreement with return to the surface, and
exchange with the atmosphere, of carbon long sequestered at depth. The drops in ^14 C
age actually occurred in several pulses associated with short-term variations in the
deglaciation rate known from various stratigraphically documented events long
hypothesized to be associated with variations in deep-water formation rates: Heinrich
Stadial 1, Bølling–Allerod, and Pre-Boreal–Younger Dryas.
(^) The mechanism for sequestering carbon when glacial-era processes slow ventilation
must primarily be the biological pump. We would like to be able to project back to
glacial conditions, say via models, to determine the global rate of sequestration.
However, that probably cannot be done convincingly, because the annual amounts
involved are extremely small. Bringing CO 2 down by ∼70 ppmv in ∼30 kyr (290 to
220 ppmv, the “rapid” initial drawdown of the Wisconsin glaciation) only required
sequestration of ∼145 GtC from the atmosphere’s interglacial content of ∼600 GtC,
which is ≈3 years’ global marine primary productivity. Thus, the rate was less than
0.005 GtC yr−1, a miniscule change readily produced by even small changes in
ventilation or export. It would have been immeasurable by any direct means, had we
been about with our ships, sediment traps, and thorium methods. In fact, a modest
change in the amount of vertical mixing of the oceans by swimming animals (Dewar
et al. 2006; Katija & Dabiri 2009) could have changed exchange that much.
Comparing 0.005 GtC yr−1 to the 2.2 GtC yr−1 currently moving into the ocean from
fossil fuel gives a strong sense of the magnitude of current human involvement in
climate control and ocean chemistry.
(^) It was for a time popular to suppose that something about the glacial eras enhanced
primary production in order to generate the organic matter to increase vertical export