and sequestration during ice-sheet growth. The primary candidate was enhanced dust
delivery to the oceans. It was likely that the mean global wind velocity did increase as
the thermal gradients from the equator to the poles intensified, and that would have
carried dust farther seaward from sources on land, where more soil open to wind
erosion would have resulted from greater aridity. Dust delivery might well have been
most significant in the subantarctic sector of the Atlantic and western Indian Oceans,
with the Argentinian pampas upwind in the westerlies belt to supply the dust.
Something like the present-day iron-limitation of subantarctic production might have
been alleviated to a degree by that increase in dust. However, look again at Fig. 16.10.
The initial drawdown of the Wisconsin (and the prior two glaciations at least)
occurred before much dust appeared. Dust does spike at the glacial maxima, but not
during the main CO 2 drawdown. That argument is not conclusive, since the ice
records available are far inland in Antarctica, isolated by several opposing wind belts
from the likely main paths of over-ocean dust transport. Dust flux data from an ice
core representing 800 kyr at a site called Dome C (75°S) (Lambert et al. 2008) shows
that the pattern of low flux during interglacials and in initial glaciation (∼0.03 mg m−2
yr−1) and high dust later (to ∼15 mg m−2 yr−1) recurred through eight glacial cycles.
(^) There is, however, some confirmation of low dust over the Southern Ocean during
initial phases of reglaciation. Study of a single sediment core from the Atlantic
subantarctic (Martinez-Garcia et al. 2009) shows that deposited dust and iron were
low during the present and previous interglacials and for periods of ∼30 Myr
afterward (Fig. 16.14), only rising as full glacial development was approached.
Eventually, possibly due to iron enrichment, both diatom deposition (not measured by
Martinez-Garcia et al., but evident from gamma-ray attenuation data at the associated
ODP site 1090) and alkenone deposition (a signature of prymnesiophyte abundance)
peaked, but only at the glacial maxima. As stated by Martinez-Garcia et al., the initial
phase of CO 2 drawdown must have been mediated by physics, not dust or its
contained iron. The obvious candidate is reduced ventilation. That could have been
reduced North Atlantic deep-water formation or expanded, annually prolonged
Antarctic sea-ice coverage or both. In contrast to these subantarctic data, results from
four late Pleistocene cores inside the Antarctic polar front at both Pacific and Atlantic
longitudes (Anderson et al. 2009) show enhanced diatom sedimentation just during
the last deglaciation, then ceasing. Anderson et al. suggest this uptick in diatom
deposition was caused by increased silicic acid availability during the resumption of
ventilation. It is not clear why it would stop after full onset of the interglacial. Data
from ODP site 1094, also inside the polar front, show low diatom deposition rates (at
least based on approximate opal proxies) during the CO 2 drawdowns following the
peaks of earlier interglacials.
Fig. 16.14 Comparison of dust deposition rates (black lines) with interpretive