Alkenone Ol3C as a Proxy for Past PC0 2 in Surface Waters
483
atmosphere, despite the fact that this is a highly
productive region (Sabine et al. 1997). The largest
natural source of CO, to the atmosphere nowadays
occurs in the equatorial Pacific, due to oceanic
upwelling ofCOz-rich waters in this high-productivity region (Feely et al. 1995; Murray et al. 1995).
The history of this region was investigated by
Jasper et al. (1994) who analysed alkenone 8 13 C
in a sediment core from the central equatorial Pacific. According to this work and the present study,
both regions, the central equatorial Pacific and
the Angola Basin, have been permanent oceanic
sources for CO, to the atmosphere during the last
200,000 years. In addition, the alkenone-derived
results suggest that both regions were even greater
oceanic sources of CO 2 to the atmosphere in glacial periods than in interglacial periods, by a factor
of at least 2 in the central equatorial Pacific and
by a factor of 2 - 3 in the Angola Basin. Based on
bulk sediment 813C, Pedersen et al. (1991) obtained
similar results for the Panama Basin. These results
all suggest that the lower pC0 2 level of the glacial
atmosphere was not the result of increased productivity in low-latitude upwelling regions. Since spatial heterogeneity is an important feature of upwelling systems (Bertrand et al. 1996), additional
studies are needed to investigate the evolution
of paleo-PC0 2 and productivity in low-latitude
upwelling regions.
Conclusions
There is evidence that the isotopic fractionation of
organic carbon by phytoplankton is a function of
both intracellular and extracellular CO 2 (aq) concentrations (Laws et al. 1995; Bidigare et al. 1997).
Estimates of ancient CO,( aq) concentrations from
sedimentary 8 13 C records must therefore consider
the effect of varying growth rates which determine
the CO 2 (aq) difference between the water and
the phytoplankton cell. In order to calibrate the
alkenone 8 13 C signal for paleo-PC0 2 reconstructions, we investigated core-top samples from the
equatorial and the South Atlantic, and found that
the sedimentary alkenone Ep value is closely related
to the phosphate concentration in overlying surface
waters (0-50 m). This is assumed to reflect differences in growth rates because our core-top sampies cover a wide range of oligotrophic to highly
eutrophic conditions. Based on theoretical grounds
(Jasperet al. 1994; Rau et al. 1996), for algae with
exclusive CO 2 (aq) utilization and diffusive CO,
transport, a simple hyperbolic equation can be used
to relate Ep to ambient [CO,(aq)] and intracellular
carbon demand: Ep = 25 %0 - b/[C0 2 (aq)]. In this
model, different growth rates are reflected by different b-values. Such a relationship is probably also
valid for our study area because our core-top measurements reveal a strong linear correlation between
the b-values and the phosphate concentration in surface waters. To reconstruct ancient changes in the
b-value, we used the bulk sediment81~ which also
shows a close inverse relationship to the phosphate
concentration in the surface waters.
On this basis and in accordance with a previous study (Muller et al. 1994), our results suggest
that the eastern Angola Basin was a permanent
oceanic source for CO 2 to the atmosphere during
the last 200,000 years. In contrast to the previous
study, however, our alkenone based reconstruction
considering growth rate effects suggests that this
region was an even greater CO 2 source during
glacials compared to interglacials. Evidently, the
delivery of CO 2 to surface waters via upwelling
was greater than the extraction of carbon by photosynthesis and export production. This indicates
that the low pC0 2 levels of the glacial atmosphere
were not the result of increased productivity in lowlatitude upwelling regions.
Until now, no well established standard method
exists for using alkenone 8 13 C as a proxy para
meter for ancientPC0 2 • However, there are several reasons why alkenone 8 13 C seems to be especially suitable for PC0 2 reconstruction. Knowledge about alkenones and the controls on their isotopic composition is much better than for other individual biomarkers. Both culture experiments and
natural studies show that alkenone Ep is dependent on [C0 2 (aq)] (Bidigare et al. 1997; Popp et al.
1998a). In contrast to other algae (Burns and
Beardalll987; Raven and Johnston 1991, Pancost
et al. 1997), there are also good indications that
alkenone Ep is not affected by active CO 2 (aq)
transport or bicarbonate usage during photosynthesis (e.g. Paasche 1964; Bidigare et al. 1997),
by the cell geometry (Popp et al. 1998b), or by
483
atmosphere, despite the fact that this is a highly
productive region (Sabine et al. 1997). The largest
natural source of CO, to the atmosphere nowadays
occurs in the equatorial Pacific, due to oceanic
upwelling ofCOz-rich waters in this high-productivity region (Feely et al. 1995; Murray et al. 1995).
The history of this region was investigated by
Jasper et al. (1994) who analysed alkenone 8 13 C
in a sediment core from the central equatorial Pacific. According to this work and the present study,
both regions, the central equatorial Pacific and
the Angola Basin, have been permanent oceanic
sources for CO, to the atmosphere during the last
200,000 years. In addition, the alkenone-derived
results suggest that both regions were even greater
oceanic sources of CO 2 to the atmosphere in glacial periods than in interglacial periods, by a factor
of at least 2 in the central equatorial Pacific and
by a factor of 2 - 3 in the Angola Basin. Based on
bulk sediment 813C, Pedersen et al. (1991) obtained
similar results for the Panama Basin. These results
all suggest that the lower pC0 2 level of the glacial
atmosphere was not the result of increased productivity in low-latitude upwelling regions. Since spatial heterogeneity is an important feature of upwelling systems (Bertrand et al. 1996), additional
studies are needed to investigate the evolution
of paleo-PC0 2 and productivity in low-latitude
upwelling regions.
Conclusions
There is evidence that the isotopic fractionation of
organic carbon by phytoplankton is a function of
both intracellular and extracellular CO 2 (aq) concentrations (Laws et al. 1995; Bidigare et al. 1997).
Estimates of ancient CO,( aq) concentrations from
sedimentary 8 13 C records must therefore consider
the effect of varying growth rates which determine
the CO 2 (aq) difference between the water and
the phytoplankton cell. In order to calibrate the
alkenone 8 13 C signal for paleo-PC0 2 reconstructions, we investigated core-top samples from the
equatorial and the South Atlantic, and found that
the sedimentary alkenone Ep value is closely related
to the phosphate concentration in overlying surface
waters (0-50 m). This is assumed to reflect differences in growth rates because our core-top sampies cover a wide range of oligotrophic to highly
eutrophic conditions. Based on theoretical grounds
(Jasperet al. 1994; Rau et al. 1996), for algae with
exclusive CO 2 (aq) utilization and diffusive CO,
transport, a simple hyperbolic equation can be used
to relate Ep to ambient [CO,(aq)] and intracellular
carbon demand: Ep = 25 %0 - b/[C0 2 (aq)]. In this
model, different growth rates are reflected by different b-values. Such a relationship is probably also
valid for our study area because our core-top measurements reveal a strong linear correlation between
the b-values and the phosphate concentration in surface waters. To reconstruct ancient changes in the
b-value, we used the bulk sediment81~ which also
shows a close inverse relationship to the phosphate
concentration in the surface waters.
On this basis and in accordance with a previous study (Muller et al. 1994), our results suggest
that the eastern Angola Basin was a permanent
oceanic source for CO 2 to the atmosphere during
the last 200,000 years. In contrast to the previous
study, however, our alkenone based reconstruction
considering growth rate effects suggests that this
region was an even greater CO 2 source during
glacials compared to interglacials. Evidently, the
delivery of CO 2 to surface waters via upwelling
was greater than the extraction of carbon by photosynthesis and export production. This indicates
that the low pC0 2 levels of the glacial atmosphere
were not the result of increased productivity in lowlatitude upwelling regions.
Until now, no well established standard method
exists for using alkenone 8 13 C as a proxy para
meter for ancientPC0 2 • However, there are several reasons why alkenone 8 13 C seems to be especially suitable for PC0 2 reconstruction. Knowledge about alkenones and the controls on their isotopic composition is much better than for other individual biomarkers. Both culture experiments and
natural studies show that alkenone Ep is dependent on [C0 2 (aq)] (Bidigare et al. 1997; Popp et al.
1998a). In contrast to other algae (Burns and
Beardalll987; Raven and Johnston 1991, Pancost
et al. 1997), there are also good indications that
alkenone Ep is not affected by active CO 2 (aq)
transport or bicarbonate usage during photosynthesis (e.g. Paasche 1964; Bidigare et al. 1997),
by the cell geometry (Popp et al. 1998b), or by