Coccolithophores as Indicators of Ocean Water Masses
133
as recorded by alkenone data (Fig. 13, also
Summerhayes et al. 1995; MUller et al. 1997).
At first glance this does not seem to be in good
agreement with recent coccolith data provided from
a sediment trap located at the northern crest of the
Walvis Ridge. As noted by Cepek & Wefer (in
press), maximum values of coccolith fluxes occur
during austral spring and autumn, when temperatures, recorded by alkenone data, are lowest
(Fig. 12). On the other hand this observation corroborates the relatively high numbers of coccoliths,
and especially of G. oceanica and E. huxleyi,
during the glacial stage 6. Emiliania huxleyi even
reached high coccolith numbers during glacial
stage 2. In addition, long-term evolutionary trends
in coccolithophore development are possibly superimposed on short-term ecological trends, especially
if one considers the very rapid evolutionary development in the genus Gephyrocapsa, which also
include E. huxleyi (e.g. Samtleben 1980). Thus, at
least the change in dominance from G. muellerae
/G. ericsonii to E. huxleyi may represent a
phylogenetic development during which changing
abundances may have been caused by variations
in ecological conditions.
In addition, E. huxleyi is a eurythermal species
with a worldwide geographic distribution and is not
useful in the reconstruction of surface water temperatures. This species is mainly influenced by the
productivity of the surface-water. Counts of E.
huxleyi often were removed from the quantitative
analysis (e.g. Giraudeau 1992; Winter and Martin
1990) in order to resolve difficulties by underestimating abundances of other species. The rapid increase in abundance of E. huxleyi after isotope
stage 5 (see Fig. 11) is an ubiquitous phenomenon
(Thierstein et al. 1977) whereas the dominance of
G. ericsonii in isotope stage 5e has also been reported from other areas (e.g. Winter 1982; Winter
and Martin 1990). The maxima ofthe latter species during interglacials and especially in substages
5e, 5c, and 5a indicate that G. ericsonii may be
considered as a relatively warm-water species.
This species is progressively replaced by G.
muellerae toward higher latitudes in stage 5
(Baumann 1990; Su 1996; Flores et al. 1997). Thus,
the ratio of both species probably reflects changes
in surface-water conditions with higher values of
G. muellerae indicating lower sea-surface temperatures. In addition to the above mentioned species, highest numbers of total coccoliths in GeoB
1028 are also due to few other species, such as C.
leptoporus and G. oceanica (Fig. 11). A change
in dominance as an indication for relative temperature changes has also been reported for the ratio
of G. muellerae to G. oceanica (Weaver and
Pujol 1988). However, G. oceanica is important
only during isotope stage 6 and seems to be more
indicative for relatively increased productivity
than for higher water temperatures. In contrast,
maxima in the abundance of taxa such as
Umbilicosphaera sibogae and C. leptoporus
probably indicate higher sea-surface temperatures
during most of the interglacial as well as interstadial
periods as they are known to prefer relatively warm
surface waters.
Geochemical Approach
In recent years, past SSTs have increasingly been
determined on the basis of relative abundances of
C 37 alkenones (Uk'37-index) in marine sediments
(Brassell et al. 1986). Studies by Volkman et al.
(1980a,b, 1995) revealed that these lipids are produced by the coccolithophores E. huxleyi and G.
oceanica inter alia. Several culture experiments
(e.g. Brassell et al. 1986; Prahl et al. 1988) have
led to the conclusion that the unsaturation ratio Uk'37
is closely correlated with water temperature and
seems to be relatively unaffected during sedimentation. Since the calibration derived by Prahl et al.
(1988) has yielded reasonable SSTs in various regions ofthe modern ocean (e.g. McCaffrey et al.
1990; Kennedy and Brassell 1992; Rosell-Mele et
al. 1995; Schneider et al. 1995) it has become a
standard calibration method for water temperature
estimates in paleoceanographic studies (Eglington
et al. 1992; Lyle et al. 1992; Rostek et al. 1993;
Zhao eta!' 1993; Schneider eta!' 1995, 1996). The
application of this method to older sediments assumes that E. huxleyi is the dominant species
within these sediments. However, this is only true
for the last 58 kyrs in core GeoB 1028. Prior to the
dominance of E. huxleyi , and even before its first
appearance 268 kyrs ago (Thierstein et al. 1977),
species ofthe genus Gephyrocapsa dominated the
133
as recorded by alkenone data (Fig. 13, also
Summerhayes et al. 1995; MUller et al. 1997).
At first glance this does not seem to be in good
agreement with recent coccolith data provided from
a sediment trap located at the northern crest of the
Walvis Ridge. As noted by Cepek & Wefer (in
press), maximum values of coccolith fluxes occur
during austral spring and autumn, when temperatures, recorded by alkenone data, are lowest
(Fig. 12). On the other hand this observation corroborates the relatively high numbers of coccoliths,
and especially of G. oceanica and E. huxleyi,
during the glacial stage 6. Emiliania huxleyi even
reached high coccolith numbers during glacial
stage 2. In addition, long-term evolutionary trends
in coccolithophore development are possibly superimposed on short-term ecological trends, especially
if one considers the very rapid evolutionary development in the genus Gephyrocapsa, which also
include E. huxleyi (e.g. Samtleben 1980). Thus, at
least the change in dominance from G. muellerae
/G. ericsonii to E. huxleyi may represent a
phylogenetic development during which changing
abundances may have been caused by variations
in ecological conditions.
In addition, E. huxleyi is a eurythermal species
with a worldwide geographic distribution and is not
useful in the reconstruction of surface water temperatures. This species is mainly influenced by the
productivity of the surface-water. Counts of E.
huxleyi often were removed from the quantitative
analysis (e.g. Giraudeau 1992; Winter and Martin
1990) in order to resolve difficulties by underestimating abundances of other species. The rapid increase in abundance of E. huxleyi after isotope
stage 5 (see Fig. 11) is an ubiquitous phenomenon
(Thierstein et al. 1977) whereas the dominance of
G. ericsonii in isotope stage 5e has also been reported from other areas (e.g. Winter 1982; Winter
and Martin 1990). The maxima ofthe latter species during interglacials and especially in substages
5e, 5c, and 5a indicate that G. ericsonii may be
considered as a relatively warm-water species.
This species is progressively replaced by G.
muellerae toward higher latitudes in stage 5
(Baumann 1990; Su 1996; Flores et al. 1997). Thus,
the ratio of both species probably reflects changes
in surface-water conditions with higher values of
G. muellerae indicating lower sea-surface temperatures. In addition to the above mentioned species, highest numbers of total coccoliths in GeoB
1028 are also due to few other species, such as C.
leptoporus and G. oceanica (Fig. 11). A change
in dominance as an indication for relative temperature changes has also been reported for the ratio
of G. muellerae to G. oceanica (Weaver and
Pujol 1988). However, G. oceanica is important
only during isotope stage 6 and seems to be more
indicative for relatively increased productivity
than for higher water temperatures. In contrast,
maxima in the abundance of taxa such as
Umbilicosphaera sibogae and C. leptoporus
probably indicate higher sea-surface temperatures
during most of the interglacial as well as interstadial
periods as they are known to prefer relatively warm
surface waters.
Geochemical Approach
In recent years, past SSTs have increasingly been
determined on the basis of relative abundances of
C 37 alkenones (Uk'37-index) in marine sediments
(Brassell et al. 1986). Studies by Volkman et al.
(1980a,b, 1995) revealed that these lipids are produced by the coccolithophores E. huxleyi and G.
oceanica inter alia. Several culture experiments
(e.g. Brassell et al. 1986; Prahl et al. 1988) have
led to the conclusion that the unsaturation ratio Uk'37
is closely correlated with water temperature and
seems to be relatively unaffected during sedimentation. Since the calibration derived by Prahl et al.
(1988) has yielded reasonable SSTs in various regions ofthe modern ocean (e.g. McCaffrey et al.
1990; Kennedy and Brassell 1992; Rosell-Mele et
al. 1995; Schneider et al. 1995) it has become a
standard calibration method for water temperature
estimates in paleoceanographic studies (Eglington
et al. 1992; Lyle et al. 1992; Rostek et al. 1993;
Zhao eta!' 1993; Schneider eta!' 1995, 1996). The
application of this method to older sediments assumes that E. huxleyi is the dominant species
within these sediments. However, this is only true
for the last 58 kyrs in core GeoB 1028. Prior to the
dominance of E. huxleyi , and even before its first
appearance 268 kyrs ago (Thierstein et al. 1977),
species ofthe genus Gephyrocapsa dominated the