296
Diekmann et al.
Temporal Fluctuations of the Kaolinite/
Chlorite-Ratio
Downcore fluctuations of the kaolinite/chlorite-ratio
in the investigated sediment cores correlate
with glacial-interglacial cycles, which in most
investigated sediment cores are displayed by
foraminiferal 8 18 0 records (Figs. 5, 6, 7, 9, 10, 12,
13). Generally interglacial periods are represented
by increased kaolinite/chlorite-ratio compared to
glacial periods. In PS2495-3, beside the variability
of the kaolinite/chlorite-ratio, percentages of
kaolinite and chlorite are negatively correlated and
show opposite curves (Fig. 7). Maximal kaolinite/
chlorite-ratio correlates with maxima of %kaolinite and minima of%-chlorite. Depending on
the temporal resolution of the sediment cores the
signatures of kaolinite/chlorite ratio curves display
even main and sub stages of glacial and interglacial periods in varying quality. Interglacial substages
are generally more pronounced than glacial
substages. In sediment cores PS2495-3, PS24981, and PS 1768-8, for instance, peak values ofthe
kaolinite/chlorite-ratio are related to isotope stage
5.5, the climate optimum of the last interglacial period (Figs. 5,6,9). In PS 1768-8 even the
Holocene sequence reveals differences in
kaolinite/ratios with maximum values around 9-10
ka BP (Fig. 10), which corresponds to the southern hemispheric climate optimum (Salinger 1981;
Gersonde et al. 1996). Isotope stages 2 and 3 in
PS 1768-8 are moreover characterized by shortterm variations of the kaolinite/chlorite-ratio on a
millennial time scale (stadials and interstadials). A
maximum of the kaolinite/chlorite-ratio around
9-10 ka BP is also indicated in sediment core
PS 1786-1 from the South Sandwich Trench, which
comprises the Holocene and Last Glacial Maximum (Fig. 13).
Apart from these temporal variations of
the kaolinite/chlorite-ratio, a latitudinal zonation
through time is evident. The stacked plot of
kaolinite/chlorite ratio time-series deduced from the
Mid-Atlantic Ridge sediment cores reflects the
decrease of the kaolinite/chlorite-ratio in a southern direction from PS2495-3 over PS2498-1 to
PS2499-5 over the presented time interval from isotope stage 5 to the Holocene (Fig. 6).
In the study area, the highest values of the
kaolinite/chlorite-ratio (0.7 to > 1.5) occur in the
northernmost sediment core GeoB211 0-4 from the
Vema Channel, lowest values in the southernmost
sediment cores PS1768-8 (0.25-0.75) from the
western SW Indian Ridge and PS 1786-1 «0.5)
from the South Sandwich Trench. PS 1768-8 and
PS1786-1 are from similar latitudes (52 0 and 54 0
S) and hence reflect additional longitudinal gradients ofthe kaolinite/chlorite-ratio within the ACC
region.
In order to determine the correlations between
global climatic cyclicity and the clay mineral parameters kaolinite/chlorite ratio, %-kaolinite and %chlorite by statistical means, we have conducted
cross-spectral analyses (method after Blackman
and Tukey 1958) between these three clay mineral
parameters in sediment core PS2495-3 and the
8 18 0 record of the reversed SPECMAP stack
(Imbrie et al. 1984), using the AnalySeries software
(Paillard et al. 1996). The technique of cross spectral analysis for paleoceanographic applications was
described in detail by Imbrie et al. (1989). The reversed SPECMAP stack is regarded as a proxy of
global ice volume variability in response to climate
changes that are triggered by orbital forcing.
Sediment core PS2495-3 from the Mid-Atlantic Ridge offers a suitable age frame to deduce time
frequencies in the frequency range of orbital parameters. Average sedimentation rates of3.0 cm/
ka and a maximum age of300 ka provide an appropriate temporal resolution. The applied age
model depends on isotope stratigraphy deduced
from the benthic foraminiferal 8 18 0 record of
Cibicidoides wuellerstorfi (Mackensen, unpublished), which reveals high coherencies and
little phase differences to frequencies of the 8 18 0
SPECMAP stack (Table 2).
The analyses of cross spectra requires the recalculation of the basic data sets to equidistantly
stepped time series through interpolation. Therefore, before conducting cross-spectral analyses, the
significance of periodic spectral variances of clay
mineral parameters in the basic data sets were
estimated according to the 'Siegel' test (Siegel
1980) with the' Spectrum' software (Schulz 1996).
The latter permits harmonic time-series analyses
on non-equidistantly stepped data sets. The
Diekmann et al.
Temporal Fluctuations of the Kaolinite/
Chlorite-Ratio
Downcore fluctuations of the kaolinite/chlorite-ratio
in the investigated sediment cores correlate
with glacial-interglacial cycles, which in most
investigated sediment cores are displayed by
foraminiferal 8 18 0 records (Figs. 5, 6, 7, 9, 10, 12,
13). Generally interglacial periods are represented
by increased kaolinite/chlorite-ratio compared to
glacial periods. In PS2495-3, beside the variability
of the kaolinite/chlorite-ratio, percentages of
kaolinite and chlorite are negatively correlated and
show opposite curves (Fig. 7). Maximal kaolinite/
chlorite-ratio correlates with maxima of %kaolinite and minima of%-chlorite. Depending on
the temporal resolution of the sediment cores the
signatures of kaolinite/chlorite ratio curves display
even main and sub stages of glacial and interglacial periods in varying quality. Interglacial substages
are generally more pronounced than glacial
substages. In sediment cores PS2495-3, PS24981, and PS 1768-8, for instance, peak values ofthe
kaolinite/chlorite-ratio are related to isotope stage
5.5, the climate optimum of the last interglacial period (Figs. 5,6,9). In PS 1768-8 even the
Holocene sequence reveals differences in
kaolinite/ratios with maximum values around 9-10
ka BP (Fig. 10), which corresponds to the southern hemispheric climate optimum (Salinger 1981;
Gersonde et al. 1996). Isotope stages 2 and 3 in
PS 1768-8 are moreover characterized by shortterm variations of the kaolinite/chlorite-ratio on a
millennial time scale (stadials and interstadials). A
maximum of the kaolinite/chlorite-ratio around
9-10 ka BP is also indicated in sediment core
PS 1786-1 from the South Sandwich Trench, which
comprises the Holocene and Last Glacial Maximum (Fig. 13).
Apart from these temporal variations of
the kaolinite/chlorite-ratio, a latitudinal zonation
through time is evident. The stacked plot of
kaolinite/chlorite ratio time-series deduced from the
Mid-Atlantic Ridge sediment cores reflects the
decrease of the kaolinite/chlorite-ratio in a southern direction from PS2495-3 over PS2498-1 to
PS2499-5 over the presented time interval from isotope stage 5 to the Holocene (Fig. 6).
In the study area, the highest values of the
kaolinite/chlorite-ratio (0.7 to > 1.5) occur in the
northernmost sediment core GeoB211 0-4 from the
Vema Channel, lowest values in the southernmost
sediment cores PS1768-8 (0.25-0.75) from the
western SW Indian Ridge and PS 1786-1 «0.5)
from the South Sandwich Trench. PS 1768-8 and
PS1786-1 are from similar latitudes (52 0 and 54 0
S) and hence reflect additional longitudinal gradients ofthe kaolinite/chlorite-ratio within the ACC
region.
In order to determine the correlations between
global climatic cyclicity and the clay mineral parameters kaolinite/chlorite ratio, %-kaolinite and %chlorite by statistical means, we have conducted
cross-spectral analyses (method after Blackman
and Tukey 1958) between these three clay mineral
parameters in sediment core PS2495-3 and the
8 18 0 record of the reversed SPECMAP stack
(Imbrie et al. 1984), using the AnalySeries software
(Paillard et al. 1996). The technique of cross spectral analysis for paleoceanographic applications was
described in detail by Imbrie et al. (1989). The reversed SPECMAP stack is regarded as a proxy of
global ice volume variability in response to climate
changes that are triggered by orbital forcing.
Sediment core PS2495-3 from the Mid-Atlantic Ridge offers a suitable age frame to deduce time
frequencies in the frequency range of orbital parameters. Average sedimentation rates of3.0 cm/
ka and a maximum age of300 ka provide an appropriate temporal resolution. The applied age
model depends on isotope stratigraphy deduced
from the benthic foraminiferal 8 18 0 record of
Cibicidoides wuellerstorfi (Mackensen, unpublished), which reveals high coherencies and
little phase differences to frequencies of the 8 18 0
SPECMAP stack (Table 2).
The analyses of cross spectra requires the recalculation of the basic data sets to equidistantly
stepped time series through interpolation. Therefore, before conducting cross-spectral analyses, the
significance of periodic spectral variances of clay
mineral parameters in the basic data sets were
estimated according to the 'Siegel' test (Siegel
1980) with the' Spectrum' software (Schulz 1996).
The latter permits harmonic time-series analyses
on non-equidistantly stepped data sets. The