138
Baumann et ai.
trast, G. oceanica and relative abundances of G.
ericsonii show a slightly positive trend with increasing alkenone concentrations. It has, however,
not been demonstrated yet whether all of these
species actually do produce alkenones. Alkenones
have been identified in lipids of G. oceanica
(Volkman et al. 1995), as stated above, and therefore it seems reasonable to predict that other species of Gephyrocapsa might also contain these distinctive biomarkers. The relatively good correlation
of Gephyrocapsa spp. to the alkenone concentrations (Fig. 14) supports the assumption that temperature calculations based on the equation of
Volkman et al. (l99S) could yield more realistic
temperatures than those based on the calibration
of Prahl et al. (1988). However, as shown by MUller
et al. (1997), this is not the case (Fig. 13), and the
reason for this is still an open question.
Using the Uk'37-index as a proxy for SST, adetailed reconstruction of the temperature changes
over the last 200ka reveals a cyclic and large scale
fluctuation of about SoC between the maximum
interglacial and the minimum glacial temperature
estimates (Fig. 13). This confirms earlier findings
in the study area made by Summerhayes et al.
(l99S), although these authors found that minimum
temperatures occurred in the interstadial (stage 3).
In GeoB 1028, minimum temperatures occurred in
glacial stages 6, 4, and 2. The coccolith assemblage
generally tends to follow the Uk'37 quite closely.
Plots of both absolute numbers and relative abundance of the six key species versus SST (Uk'37)
for GeoBI028 demonstrate that only few show a
relatively close correlation with paleotemperature
estimates (Fig. IS). The abundances of G.
ericson ii, and U. sibogae clearly show a positive
correlation with increasing SSTs, whereas those of
G. muellerae, G. oceanica, and C. leptoporus
are correlated with cooler temperatures. Emiliania
huxleyi does not show any obvious SST -related
trend which may be related to the eurythermal
character of this species. Species indicating
warmer temperatures can be separated from those
indicative of warmer temperatures at about 20°C.
These findings independently confirm the above
mentioned trends in SST. The relatively cold-water adapted G. muellerae clearly predominates
exclusively in the cold stages, where SSTs are
continuously < 19-20°C. In contrast, the warm flora
(especially G. ericsonii and U. sibogae) is much
more abundant in interglacials where the temperature is constantl at >20-21 °C.
In general, the presented trends are very similar to those found by Jordan et al. (l99S) in the
upwelling region off Northwest Africa, except for
G. oceanica. This might rather be related to its
affinity to a relatively increased productivity than
to higher water temperatures. Also, the connections between the 'warmer' and the 'cooler' species are higher in the upwelling area off southwestern Africa than observed off Northwest Africa
(18°C). The high degree of scattering in all of the
plots, however, is not surprising because the distribution of coccolithophores is not necessarily determined by the surface-water temperature alone (see
above).
Conclusions and Perspective
Coccoliths are a major component in the sediments
of the South Atlantic, but surprisingly little quantitative information is available from this region.
However, as in many other parts of the world's
oceans, major oceanographic signals are preserved
in coccolith sediment assemblages. Thus, they reflect spatial and temporal changes in the circulation ofthe surface ocean very well. The following
conclusions and perspectives can be drawn from
the present study.
1) The distribution and abundances of both
coccolithophores in the plankton and coccoliths in
the underlying surface sediments of the study areas seem to be related to the combination of surface currents and upwelling. In addition, a correlation between productivity in the upwelling zone
and the production of carbonate by coccoliths
seems reasonable, although this could be masked
by any dilution ofterrigenous material and/or dissolution offragile species.
There is, however, an increasing need to determine present-day biogeographic distribution patterns and to compare them to external controls in
terms of biogeography, trophic level and seasonal
succession. Thus, more data on the variability of
coccolithophores in both species numbers and composition are useful for the identification of environ-
Baumann et ai.
trast, G. oceanica and relative abundances of G.
ericsonii show a slightly positive trend with increasing alkenone concentrations. It has, however,
not been demonstrated yet whether all of these
species actually do produce alkenones. Alkenones
have been identified in lipids of G. oceanica
(Volkman et al. 1995), as stated above, and therefore it seems reasonable to predict that other species of Gephyrocapsa might also contain these distinctive biomarkers. The relatively good correlation
of Gephyrocapsa spp. to the alkenone concentrations (Fig. 14) supports the assumption that temperature calculations based on the equation of
Volkman et al. (l99S) could yield more realistic
temperatures than those based on the calibration
of Prahl et al. (1988). However, as shown by MUller
et al. (1997), this is not the case (Fig. 13), and the
reason for this is still an open question.
Using the Uk'37-index as a proxy for SST, adetailed reconstruction of the temperature changes
over the last 200ka reveals a cyclic and large scale
fluctuation of about SoC between the maximum
interglacial and the minimum glacial temperature
estimates (Fig. 13). This confirms earlier findings
in the study area made by Summerhayes et al.
(l99S), although these authors found that minimum
temperatures occurred in the interstadial (stage 3).
In GeoB 1028, minimum temperatures occurred in
glacial stages 6, 4, and 2. The coccolith assemblage
generally tends to follow the Uk'37 quite closely.
Plots of both absolute numbers and relative abundance of the six key species versus SST (Uk'37)
for GeoBI028 demonstrate that only few show a
relatively close correlation with paleotemperature
estimates (Fig. IS). The abundances of G.
ericson ii, and U. sibogae clearly show a positive
correlation with increasing SSTs, whereas those of
G. muellerae, G. oceanica, and C. leptoporus
are correlated with cooler temperatures. Emiliania
huxleyi does not show any obvious SST -related
trend which may be related to the eurythermal
character of this species. Species indicating
warmer temperatures can be separated from those
indicative of warmer temperatures at about 20°C.
These findings independently confirm the above
mentioned trends in SST. The relatively cold-water adapted G. muellerae clearly predominates
exclusively in the cold stages, where SSTs are
continuously < 19-20°C. In contrast, the warm flora
(especially G. ericsonii and U. sibogae) is much
more abundant in interglacials where the temperature is constantl at >20-21 °C.
In general, the presented trends are very similar to those found by Jordan et al. (l99S) in the
upwelling region off Northwest Africa, except for
G. oceanica. This might rather be related to its
affinity to a relatively increased productivity than
to higher water temperatures. Also, the connections between the 'warmer' and the 'cooler' species are higher in the upwelling area off southwestern Africa than observed off Northwest Africa
(18°C). The high degree of scattering in all of the
plots, however, is not surprising because the distribution of coccolithophores is not necessarily determined by the surface-water temperature alone (see
above).
Conclusions and Perspective
Coccoliths are a major component in the sediments
of the South Atlantic, but surprisingly little quantitative information is available from this region.
However, as in many other parts of the world's
oceans, major oceanographic signals are preserved
in coccolith sediment assemblages. Thus, they reflect spatial and temporal changes in the circulation ofthe surface ocean very well. The following
conclusions and perspectives can be drawn from
the present study.
1) The distribution and abundances of both
coccolithophores in the plankton and coccoliths in
the underlying surface sediments of the study areas seem to be related to the combination of surface currents and upwelling. In addition, a correlation between productivity in the upwelling zone
and the production of carbonate by coccoliths
seems reasonable, although this could be masked
by any dilution ofterrigenous material and/or dissolution offragile species.
There is, however, an increasing need to determine present-day biogeographic distribution patterns and to compare them to external controls in
terms of biogeography, trophic level and seasonal
succession. Thus, more data on the variability of
coccolithophores in both species numbers and composition are useful for the identification of environ-