Past changes in ocean carbonate chemistry - Ocean Acidification - page 29

2.4.2 Glacial-interglacial changes

The glacial-interglacial changes in surface-ocean
CO 2 chemistry have been invoked to explain changes
in shell weights of surface-dwelling foraminifera.
For example, calcite shells of different planktonic
foraminiferal species recovered from deep-sea sedi-
ment cores in the North Atlantic and Indian Ocean
show higher shell weights during the last glacial
period compared with the Holocene (e.g. Barker and
Elderi eld 2002 ; de Moel et al. 2009). These authors
suggest that lower p CO 2 and elevated surface [CO 3 2- ]
caused higher initial shell weights during the last
glacial stage. On the other hand, shell weights of
planktonic foraminifera have been used as an indi-
cator of carbonate sediment dissolution and thus as
a proxy for [CO 3 2- ] in the deep sea , rather than the sur-
face (e.g. Broecker and Clark 2003). Clearly, the issue
is complicated due to various factors, including pos-
sible effects of growth temperature, [CO 3 2- ], nutri-
ents, and other environmental parameters on initial
shell weight, as well as dissolution in sediments
and/or the water column (e.g. Bijma et al. 2002 ).
Interrelations between coccolithophore species, coc-
colith weight/chemistry, primary production, and
the carbon cycle appear to be even more complex
(for discussions see, e.g., Zondervan et al. 2001 ;
Beaufort et al. 2007 ; Rickaby et al. 2007 ).

Over the past 800 000 years or so, atmospheric CO 2
has varied periodically between ~200 ppmv and
~280 ppmv (Siegenthaler et al. 2005 ; Lüthi et al.
2008). These glacial-interglacial cycles were accom-
panied by periodic changes in surface-ocean CO 2
chemistry, while deep-sea pH and carbonate ion
concentration are believed to have been relatively
stable (Zeebe and Marchitto 2010). Compared with
interglacials, glacial surface-ocean conditions were
characterized by lower temperatures, higher pH,
and higher carbonate ion concentration (e.g. Sanyal
et al. 1995 ; Hönisch and Hemming 2005 ; Foster
2008). For example, an interglacial surface-seawater
sample at T = 15°C, S = 35, C T = 2000 μmol kg -1 , and
A T = 2284 μmol kg -1 , has a p CO 2 of 280 μatm, pH T =
8.17, and [CO 3 2- ] = 198 μmol kg -1 . Corresponding

glacial conditions may have been T = 12°C, S = 36,
C T = 2006 μmol kg -1 , and A T = 2353 μmol kg -1 , which
yields a p CO 2 of 200 μatm, pH T = 8.30 and [CO 3 2- ] =
238 μmol kg -1 . This indicates a difference in the gla-
cial-interglacial saturation state of about 20%. Note
that while this scenario assumes 3% higher glacial
A T , various other scenarios are possible, which
would also modify the calculated pH change (e.g.
Archer et al. 2000). Nevertheless, it illustrates the
order of magnitude of glacial-interglacial changes
in surface-ocean carbonate chemistry.
When considering the time evolution of the sys-
tem over glacial-interglacial cycles, it is clear that
surface-ocean pH and saturation state decline dur-
ing the course of a deglaciation. One might thus
think of a deglaciation as an 'acidii cation event',
albeit a very slow and moderate one. In terms of
rate and magnitude, it is important to realize that a
deglaciation is not a past analogue for the current
anthropogenic perturbation. For example, the rate
of surface-ocean pH change during the most recent
deglaciation may be estimated as 0.1 to 0.2 units per
10 000 years, or 0.001 to 0.002 units per century on
average. In contrast, under business-as-usual CO 2
emissions, humans may cause a surface-ocean pH
change of 0.7 units per 500 years, or 0.14 units per
century on average. Thus, changes in surface-ocean
chemistry during the Anthropocene are expected to
be about three to seven times larger and 70 times
faster than during a deglaciation.

2.4.3 Pleistocene and Pliocene (~5 Myr to
~12 000 yr ago)

Ice-core records of atmospheric CO 2 are limited by
the oldest samples available in Antarctic ice cores,
which go back at most about 1 Myr. Beyond that,
estimates of palaeo- p CO 2 levels and ocean chemistry
have to rely on other proxies. Based on stable boron
isotopes in foraminifera, glacial p CO 2 levels before
the mid-Pleistocene Transition (MPT; ~1 Myr) were
estimated to have been about 30 μatm higher than
after the transition. Estimates of pre-MPT intergla-
cial values appear similar to those obtained from ice
cores during the late Pleistocene (Hönisch et al.
2009). Note that stable boron isotopes are actually a
proxy for seawater pH and that one other CO 2 sys-
tem parameter is required to reconstruct atmos-
pheric CO 2 . Regardless, the boron isotope data
indicate that surface-ocean pH SWS over the past 2
Myr has varied periodically between ~8.1 and ~8.3
( Hönisch et al. 2009). So far, no major excursions or