514
Leaet al.
terrestrial biosphere to the oceans, it is consistent
with an addition to the glacial oceans of 450 Gt
terrestrial carbon with a 8 l3 C of -25%0, or about
21 % of the present biospheric carbon reservoir (including dead soil organic matter) (1 Gt = 10 15 g C).
Pollen databases from the LGM appear to be consistent with a substantially larger reduction of the
terrestrial biosphere (750 - 1350 Gt), although there
is considerable uncertainty associated with estimating terrestrial carbon mass and in predicting how
efficiently carbon is transferred to the ocean during glaciations (Crowley 1995). An additional complication is that some ofthe glacial carbon addition
might have come from the erosion of continental
shelves exposed by lower sea level (Broecker
1982b).
Planktic foraminiferal8 l3 C records are complicated by isotopic offsets between species, variations in 8 J3 C with shell size, and disequilibrium effects attributable to the presence of symbiotic algae (Berger et al. 1978; Spero and Williams 1988;
Oppo and Fairbanks 1989; Spero and Lea 1996).
Records of Globigerinoides sacculifer, the principal species used for tropical reconstructions, reveal that shells deposited in LGM sections average
only about 0.2%0 more negative in 8 J3 C than their
Holocene counterparts (Curry and Crowley 1987;
Oppo and Fairbanks 1989). In addition, low latitude
planktic 8 J3 C records are characterized by the most
negative 8 J3 C values occurring on the glacial terminations rather than during the LGM. Despite the
complexity ofthese records, an increase in plankticbenthic 8 J3 C differences in the glacial sections of
tropical cores has been cited as evidence of a
stronger biological pump during glacial periods
(Shackleton et al. 1983a; Shackleton and Pisias
1985; Curry and Crowley 1987; Shackleton et al.
1992). The largest shift in planktic8 J3 C is observed
in cores from the Southern Ocean, where shells
taken from the glacial sections are typically 1 %0
more negative than interglacials (Labeyrie and
Duplessy 1985; Charles and Fairbanks 1990;
Charles et al. 1996). Neither the magnitude nor
direction of the Southern Ocean 8 J3 C shifts have
been adequately explained because the large nutrient changes implied by the shifts are not supported by coincident changes in other proxies
(Boyle 1988; Lea 1995), are opposite to those expected from a strengthened biological pump
(Broecker 1982a), and are greater than could be
reasonably inferred from changes in gas exchange
(Charles et al. 1993).
Influence of Seawater Carbonate Ion
Concentration on Foraminiferal Isotopes
Culturing experiments demonstrate that the 8 J3 C
and 8180 composition of planktic foraminifera shells
responds to changes in seawater carbonate ion
concentration (Spero et al. 1997) (Bijma et al. this
volume). The average response of she1l8 J3 C in the
growth experiments is -0.006%0 for Orbulina
universa and -0.012%0 for Globigerina bulloides
per ~moI/kg increase in seawater carbonate ion.
The response for shell 8 18 0 is -0.002 and -0.004%0
per ~mollkg carbonate ion for 0. universa and G.
bulloides, respectively. Carbonate ion concentration is utilized as the independent variable because,
unlike pH, under different experimental conditions
isotopic deviations are a linear function of carbonate ion. However, because pH and carbonate ion
vary together in seawater, we do not have proof that
carbonate ion concentration is the controlling variable. The isotopic response of the shells to changes
in carbonate ion concentration presumably reflects
a kinetic response during calcite precipitation
(Spero et al. 1997). Newer results reveal a similar
affect for two tropical, surface-dwelling species,
Globigerinoides ruber and Gs. sacculifer (Bijma
et al. 1998).
Ice core records indicate that atmospheric pCO z
levels dropped to about 200 ppm V during the LGM,
although the cause of the reduction remains unexplained (Barnola et al. 1987; 10uzel et al. 1993).
Glacial reconstructions demonstrate that temperature and salinity changes alone cannot account for
more than a small fraction of this pCO z reduction
(Broecker 1982a; Broecker and Peng 1993). Because the surface ocean must on average have the
same pCO z as the atmosphere and because sea
surface temperatures did not drop sufficiently to
cause the observed pC0 2 reduction (Broecker and
Peng 1993), lower glacial pC0 2 1eveis must reflect
an increase in average sea surface carbonate ion,
or more specifically an increase in the ratio of alkalinity (excess base) to COz' Based on the results
Leaet al.
terrestrial biosphere to the oceans, it is consistent
with an addition to the glacial oceans of 450 Gt
terrestrial carbon with a 8 l3 C of -25%0, or about
21 % of the present biospheric carbon reservoir (including dead soil organic matter) (1 Gt = 10 15 g C).
Pollen databases from the LGM appear to be consistent with a substantially larger reduction of the
terrestrial biosphere (750 - 1350 Gt), although there
is considerable uncertainty associated with estimating terrestrial carbon mass and in predicting how
efficiently carbon is transferred to the ocean during glaciations (Crowley 1995). An additional complication is that some ofthe glacial carbon addition
might have come from the erosion of continental
shelves exposed by lower sea level (Broecker
1982b).
Planktic foraminiferal8 l3 C records are complicated by isotopic offsets between species, variations in 8 J3 C with shell size, and disequilibrium effects attributable to the presence of symbiotic algae (Berger et al. 1978; Spero and Williams 1988;
Oppo and Fairbanks 1989; Spero and Lea 1996).
Records of Globigerinoides sacculifer, the principal species used for tropical reconstructions, reveal that shells deposited in LGM sections average
only about 0.2%0 more negative in 8 J3 C than their
Holocene counterparts (Curry and Crowley 1987;
Oppo and Fairbanks 1989). In addition, low latitude
planktic 8 J3 C records are characterized by the most
negative 8 J3 C values occurring on the glacial terminations rather than during the LGM. Despite the
complexity ofthese records, an increase in plankticbenthic 8 J3 C differences in the glacial sections of
tropical cores has been cited as evidence of a
stronger biological pump during glacial periods
(Shackleton et al. 1983a; Shackleton and Pisias
1985; Curry and Crowley 1987; Shackleton et al.
1992). The largest shift in planktic8 J3 C is observed
in cores from the Southern Ocean, where shells
taken from the glacial sections are typically 1 %0
more negative than interglacials (Labeyrie and
Duplessy 1985; Charles and Fairbanks 1990;
Charles et al. 1996). Neither the magnitude nor
direction of the Southern Ocean 8 J3 C shifts have
been adequately explained because the large nutrient changes implied by the shifts are not supported by coincident changes in other proxies
(Boyle 1988; Lea 1995), are opposite to those expected from a strengthened biological pump
(Broecker 1982a), and are greater than could be
reasonably inferred from changes in gas exchange
(Charles et al. 1993).
Influence of Seawater Carbonate Ion
Concentration on Foraminiferal Isotopes
Culturing experiments demonstrate that the 8 J3 C
and 8180 composition of planktic foraminifera shells
responds to changes in seawater carbonate ion
concentration (Spero et al. 1997) (Bijma et al. this
volume). The average response of she1l8 J3 C in the
growth experiments is -0.006%0 for Orbulina
universa and -0.012%0 for Globigerina bulloides
per ~moI/kg increase in seawater carbonate ion.
The response for shell 8 18 0 is -0.002 and -0.004%0
per ~mollkg carbonate ion for 0. universa and G.
bulloides, respectively. Carbonate ion concentration is utilized as the independent variable because,
unlike pH, under different experimental conditions
isotopic deviations are a linear function of carbonate ion. However, because pH and carbonate ion
vary together in seawater, we do not have proof that
carbonate ion concentration is the controlling variable. The isotopic response of the shells to changes
in carbonate ion concentration presumably reflects
a kinetic response during calcite precipitation
(Spero et al. 1997). Newer results reveal a similar
affect for two tropical, surface-dwelling species,
Globigerinoides ruber and Gs. sacculifer (Bijma
et al. 1998).
Ice core records indicate that atmospheric pCO z
levels dropped to about 200 ppm V during the LGM,
although the cause of the reduction remains unexplained (Barnola et al. 1987; 10uzel et al. 1993).
Glacial reconstructions demonstrate that temperature and salinity changes alone cannot account for
more than a small fraction of this pCO z reduction
(Broecker 1982a; Broecker and Peng 1993). Because the surface ocean must on average have the
same pCO z as the atmosphere and because sea
surface temperatures did not drop sufficiently to
cause the observed pC0 2 reduction (Broecker and
Peng 1993), lower glacial pC0 2 1eveis must reflect
an increase in average sea surface carbonate ion,
or more specifically an increase in the ratio of alkalinity (excess base) to COz' Based on the results