520
Lea et al.
glacial to interglacial change in seawater 8 1s O,
which might have been smaller than previously
thought (Schrag et al. 1996). Other marine SST
proxies (e.g. coral SrlCa) as well as terrestrial indicators (e.g. snow-line and ice core ( 18 0) suggest
tropical cooling of 4-6°C at the LGM (Guilderson
et al. 1994; Thompson et al. 1995). Experimental
data demonstrates that she1l8 1s O drops -0.002 and
-0.004%0 per Ilmollkg carbonate ion increase for
0. universa and G. bulloides, respectively (Spero
et al. 1997).
Anomalies in 8 1s O computed for an 80 ppmV
pC0 2 reduction range from -0.1 to -0.3%0, equivalent to an SST anomaly of approximately -0.5 to -
I.SoC (Fig. 2e, t). These anomalies in she1l8 1s O
are negative deviations from apparent equilibrium
which translate into overestimates of
paleotemperature during periods oflow pC0 2 • For
example, for an observed glacial-interglacial
planktic 8 1s O difference of 1.5%0 in tropical cores
(Broecker 1986; Shackleton et al. 1992; Linsley and
Dunbar 1994; Thunell et al. 1994), correcting for
the influence of higher carbonate ion on glacial shells
could increase the 8 1s O difference to 1.7%0. Assuming the more recent 1%0 estimate of the glacial- interglacial ice volume effect (Fairbanks 1989;
Schrag et al. 1996), the residual 0.7%0 is compatible with tropical cooling of up to 3.5°C. The influence of carbonate ion on shell 8 18 0 also influences
paleo-salinity estimates based on G. bulloides 8 18 0
(Duplessy et al. 1991). A 0.3%0 correction on8 18 0
would reduce calculated LGM salinity anomalies
by about 0.6 psu.
Conclusions
Our analysis ofthe influence of increased carbonate ion on glacial isotopic distributions suggests
potential solutions to some intractable paleoceanographic problems, but raises fundamental questions
as well. The carbonate ion effect can explain why
the amplitude of Southern Ocean G. bulloides 8 l3 C
records is larger than observed in other planktic
records. However, our analysis does not resolve
the extent to which planktic 8 l3 C records
reflect changes in the strength of the biological
pump (Shackleton et al. 1983a; Shackleton and
Pisias 1985; Curry and Crowley 1987; Shackleton
et al. 1992). If planktics alone require a correction
for higher glacial carbonate ion, the actual surface
- deep 8 l3 C difference during glacial episodes was
significantly larger (~0.4%0) than calculated from
the foraminiferal she1l8 l3 C differences. Ifbenthic
shell 8 l3 C is also affected by carbonate ion concentration, some or all of the negative carbon isotope shift during glacial episodes is likely to be due
to a whole-ocean carbonate ion increase. This attribution reduces the actual mean ocean 8 l3 C shift
and therefore diminishes the inferred amount of
terrestrial carbon transferred to the ocean during
the LGM. This change further enlarges the difference between pollen-based and ocean 8 l3 C-based
estimates of the amount of carbon transferred from
the terrestrial biosphere to the ocean during glacial
episodes (Shackleton 1977; Crowley 1995). Because the carbonate ion effect appears to be a
major factor influencing the foraminiferal isotope
record, calibration of the response in each ofthe
major paleoceanographic planktic and benthic species is a necessary next step towards fully understanding and utilizing the oceanic isotopic record.
Acknowledgments
The senior author would like to thank the Department of Geophysical Sciences, University of Chicago, for a Visiting Professorship, support, and facilities, while undertaking this research. Thanks go
to Alan Mix, Brad Linsley, Chris Charles, Tom
Crowley, Lloyd Keigwin and Michael Sarnthein for
sharing isotope data and to Jim Kennett and Nick
Shackleton for comments on the ms. Reviews by
W. Berger and an anonymous reviewer improved
the manuscript. This research was supported by the
US National Science Foundation (DWL and HJS)
and by the Deutsche Forschungsgemeinschaft
(Sonderforschungsbereich 261 at Bremen University, Contribution no. 204) and the Program for the
Advancement of Special Research Projects at the
Alfred Wegener Institute, Germany (JB).
Lea et al.
glacial to interglacial change in seawater 8 1s O,
which might have been smaller than previously
thought (Schrag et al. 1996). Other marine SST
proxies (e.g. coral SrlCa) as well as terrestrial indicators (e.g. snow-line and ice core ( 18 0) suggest
tropical cooling of 4-6°C at the LGM (Guilderson
et al. 1994; Thompson et al. 1995). Experimental
data demonstrates that she1l8 1s O drops -0.002 and
-0.004%0 per Ilmollkg carbonate ion increase for
0. universa and G. bulloides, respectively (Spero
et al. 1997).
Anomalies in 8 1s O computed for an 80 ppmV
pC0 2 reduction range from -0.1 to -0.3%0, equivalent to an SST anomaly of approximately -0.5 to -
I.SoC (Fig. 2e, t). These anomalies in she1l8 1s O
are negative deviations from apparent equilibrium
which translate into overestimates of
paleotemperature during periods oflow pC0 2 • For
example, for an observed glacial-interglacial
planktic 8 1s O difference of 1.5%0 in tropical cores
(Broecker 1986; Shackleton et al. 1992; Linsley and
Dunbar 1994; Thunell et al. 1994), correcting for
the influence of higher carbonate ion on glacial shells
could increase the 8 1s O difference to 1.7%0. Assuming the more recent 1%0 estimate of the glacial- interglacial ice volume effect (Fairbanks 1989;
Schrag et al. 1996), the residual 0.7%0 is compatible with tropical cooling of up to 3.5°C. The influence of carbonate ion on shell 8 18 0 also influences
paleo-salinity estimates based on G. bulloides 8 18 0
(Duplessy et al. 1991). A 0.3%0 correction on8 18 0
would reduce calculated LGM salinity anomalies
by about 0.6 psu.
Conclusions
Our analysis ofthe influence of increased carbonate ion on glacial isotopic distributions suggests
potential solutions to some intractable paleoceanographic problems, but raises fundamental questions
as well. The carbonate ion effect can explain why
the amplitude of Southern Ocean G. bulloides 8 l3 C
records is larger than observed in other planktic
records. However, our analysis does not resolve
the extent to which planktic 8 l3 C records
reflect changes in the strength of the biological
pump (Shackleton et al. 1983a; Shackleton and
Pisias 1985; Curry and Crowley 1987; Shackleton
et al. 1992). If planktics alone require a correction
for higher glacial carbonate ion, the actual surface
- deep 8 l3 C difference during glacial episodes was
significantly larger (~0.4%0) than calculated from
the foraminiferal she1l8 l3 C differences. Ifbenthic
shell 8 l3 C is also affected by carbonate ion concentration, some or all of the negative carbon isotope shift during glacial episodes is likely to be due
to a whole-ocean carbonate ion increase. This attribution reduces the actual mean ocean 8 l3 C shift
and therefore diminishes the inferred amount of
terrestrial carbon transferred to the ocean during
the LGM. This change further enlarges the difference between pollen-based and ocean 8 l3 C-based
estimates of the amount of carbon transferred from
the terrestrial biosphere to the ocean during glacial
episodes (Shackleton 1977; Crowley 1995). Because the carbonate ion effect appears to be a
major factor influencing the foraminiferal isotope
record, calibration of the response in each ofthe
major paleoceanographic planktic and benthic species is a necessary next step towards fully understanding and utilizing the oceanic isotopic record.
Acknowledgments
The senior author would like to thank the Department of Geophysical Sciences, University of Chicago, for a Visiting Professorship, support, and facilities, while undertaking this research. Thanks go
to Alan Mix, Brad Linsley, Chris Charles, Tom
Crowley, Lloyd Keigwin and Michael Sarnthein for
sharing isotope data and to Jim Kennett and Nick
Shackleton for comments on the ms. Reviews by
W. Berger and an anonymous reviewer improved
the manuscript. This research was supported by the
US National Science Foundation (DWL and HJS)
and by the Deutsche Forschungsgemeinschaft
(Sonderforschungsbereich 261 at Bremen University, Contribution no. 204) and the Program for the
Advancement of Special Research Projects at the
Alfred Wegener Institute, Germany (JB).