Kaolinite and Chlorite as Tracers of Modern and Late Quaternary Deep Water Circulation
289
et al. 1996). Using the weighting factors ofBiscaye
(1965), percentages of each clay mineral group
were computed: smectite (Ix), kaolinite (2x),
chlorite (2x), illite (4x).
Age models for the sediment cores are based
on the correlations of benthic or planktic
foraminiferal 8 18 0 records with the 8 18 0 curve of
the SPECMAP stack (Imbrie et al. 1984). Age
models for sediment cores, which are totally or
partly barren of biogenic carbonate (PS 1786-1,
PS2082-1, PS2499-5), were inferred from diatom
and radiolarian biofluctuation stratigraphies and
biostratigraphies according to Hays et al. (1976);
Burckle (1982); Abelmann and Gersonde (1988)
and Gersonde et al. (1990). Additional chronostratigraphic information was deduced from the
relative abundances of planktic foraminifera ofthe
Globorotalia menardii complex according to
Damuth (1977) (GeoB 2110-4) and 14C AMS
and 23°Thex measurements (PS 1768-8, PS2499-5)
(Frank et al. this volume). Most of the applied age
models have been published elsewhere (Table 1).
Accumulation rates of terrigenous sediment
(AR te ,,) were calculated from sediment dry bulk
density multiplied with the relative proportion ofthe
terrigenous fraction and the linear sedimentation
rate (SR) that depends on the applied age model
(Van Andel et al. 1975). For AR calculation and
better comparison we used averaged SR for whole
isotope stages, regardless of the actual temporal
resolution estimated in a sediment core.
AR of the individual clay minerals kaolinite and
chlorite were estimated from AR ofthe terrigenous
clay fraction and the relative clay mineral proportions within the clay fraction. In the studied sediment cores, the non-carbonate clay fraction also
includes biogenic opal and minor amounts of quartz
and feldspar with subordinate amphibole. Ratios of
XRD peak areas of the clay minerals to the nonclay minerals show no distinct downcore variations.
So we assume a constant clay mineral proportion
in the terrigenous clay fraction throughout the
entire sediment cores. Since the oriented preparation of clay mounts favors the diffraction of
phyllosilicates, an accurate determination of nonclay mineral proportions is impossible. For the calculation of clay mineral AR we assume that the
four main clay mineral groups constitute 80% of
the entire terrigenous clay fraction. Due to high
errors in quantitative opal determination in the clay
fraction of diatomaceous muds and oozes, AR for
clay minerals were only calculated for the opal-poor
sediment cores GeoB 2110-4 and PS2495-3 that
also reveal marked variations of relative downcore
%-kaolinite and %-chlorite distributions. Keeping
in mind such restrictions, mass accumulation rates
of individual clay minerals give a first-order estimation (Fagel et al. 1997).
Time series analyses of clay mineralogical
data were conducted for the well-dated sediment
core PS2495-3 using the 'Spectrum' software
(Schulz 1996) and the 'AnalySeries 1.0a7' software (Paillard et al. 1996).
Modern Deep Water Advection and
Kaolinite/Chlorite Distribution in Surface
Sediments
The South Atlantic plays a key role in global thermohaline circulation as it forms the flow paths for the
major abyssal water masses, the North Atlantic
Deep Water (NADW), the Circumpolar Deep
Water (CPDW), and the Antarctic Bottom Water
(AABW) (Whitworth and Nowlin 1987; Peterson
and Whitworth 1989; Reid 1989; Gordon 1996;
Schmitz 1995). These water masses are fed into
the Southern Ocean and distributed to the other
world oceans via the Antarctic Circumpolar Current (ACC) (Figs. 1,3). Very cold (0 to -0.7 0C)
AABW originates from the marginal seas
of Antarctica, particularly from the Weddell Sea
(Carmack and Foster 1975; Foldvik and
Gammelsrod 1988). It occupies the basal water
column of the eastward flowing ACC and spreads
to the north through the western South Atlantic.
CPDW is the main water mass of the ACC, where
it occurs in water depths below 500 m. North of
about 45° S CPDW is split off due to injection of
southward advecting saline, nutrient-poor NADW
at intermediate depths (2000-3500 m) (Fig. 3).
Lower CPDW overlies the AABW in the western South Atlantic basins and bathes the deepest
parts ofthe eastern South Atlantic basins below approximately 4000 m water depth, respectively (Fig.
3). Basins in the eastern South Atlantic north ofthe
Walvis Ridge are completely filled up with NADW.
289
et al. 1996). Using the weighting factors ofBiscaye
(1965), percentages of each clay mineral group
were computed: smectite (Ix), kaolinite (2x),
chlorite (2x), illite (4x).
Age models for the sediment cores are based
on the correlations of benthic or planktic
foraminiferal 8 18 0 records with the 8 18 0 curve of
the SPECMAP stack (Imbrie et al. 1984). Age
models for sediment cores, which are totally or
partly barren of biogenic carbonate (PS 1786-1,
PS2082-1, PS2499-5), were inferred from diatom
and radiolarian biofluctuation stratigraphies and
biostratigraphies according to Hays et al. (1976);
Burckle (1982); Abelmann and Gersonde (1988)
and Gersonde et al. (1990). Additional chronostratigraphic information was deduced from the
relative abundances of planktic foraminifera ofthe
Globorotalia menardii complex according to
Damuth (1977) (GeoB 2110-4) and 14C AMS
and 23°Thex measurements (PS 1768-8, PS2499-5)
(Frank et al. this volume). Most of the applied age
models have been published elsewhere (Table 1).
Accumulation rates of terrigenous sediment
(AR te ,,) were calculated from sediment dry bulk
density multiplied with the relative proportion ofthe
terrigenous fraction and the linear sedimentation
rate (SR) that depends on the applied age model
(Van Andel et al. 1975). For AR calculation and
better comparison we used averaged SR for whole
isotope stages, regardless of the actual temporal
resolution estimated in a sediment core.
AR of the individual clay minerals kaolinite and
chlorite were estimated from AR ofthe terrigenous
clay fraction and the relative clay mineral proportions within the clay fraction. In the studied sediment cores, the non-carbonate clay fraction also
includes biogenic opal and minor amounts of quartz
and feldspar with subordinate amphibole. Ratios of
XRD peak areas of the clay minerals to the nonclay minerals show no distinct downcore variations.
So we assume a constant clay mineral proportion
in the terrigenous clay fraction throughout the
entire sediment cores. Since the oriented preparation of clay mounts favors the diffraction of
phyllosilicates, an accurate determination of nonclay mineral proportions is impossible. For the calculation of clay mineral AR we assume that the
four main clay mineral groups constitute 80% of
the entire terrigenous clay fraction. Due to high
errors in quantitative opal determination in the clay
fraction of diatomaceous muds and oozes, AR for
clay minerals were only calculated for the opal-poor
sediment cores GeoB 2110-4 and PS2495-3 that
also reveal marked variations of relative downcore
%-kaolinite and %-chlorite distributions. Keeping
in mind such restrictions, mass accumulation rates
of individual clay minerals give a first-order estimation (Fagel et al. 1997).
Time series analyses of clay mineralogical
data were conducted for the well-dated sediment
core PS2495-3 using the 'Spectrum' software
(Schulz 1996) and the 'AnalySeries 1.0a7' software (Paillard et al. 1996).
Modern Deep Water Advection and
Kaolinite/Chlorite Distribution in Surface
Sediments
The South Atlantic plays a key role in global thermohaline circulation as it forms the flow paths for the
major abyssal water masses, the North Atlantic
Deep Water (NADW), the Circumpolar Deep
Water (CPDW), and the Antarctic Bottom Water
(AABW) (Whitworth and Nowlin 1987; Peterson
and Whitworth 1989; Reid 1989; Gordon 1996;
Schmitz 1995). These water masses are fed into
the Southern Ocean and distributed to the other
world oceans via the Antarctic Circumpolar Current (ACC) (Figs. 1,3). Very cold (0 to -0.7 0C)
AABW originates from the marginal seas
of Antarctica, particularly from the Weddell Sea
(Carmack and Foster 1975; Foldvik and
Gammelsrod 1988). It occupies the basal water
column of the eastward flowing ACC and spreads
to the north through the western South Atlantic.
CPDW is the main water mass of the ACC, where
it occurs in water depths below 500 m. North of
about 45° S CPDW is split off due to injection of
southward advecting saline, nutrient-poor NADW
at intermediate depths (2000-3500 m) (Fig. 3).
Lower CPDW overlies the AABW in the western South Atlantic basins and bathes the deepest
parts ofthe eastern South Atlantic basins below approximately 4000 m water depth, respectively (Fig.
3). Basins in the eastern South Atlantic north ofthe
Walvis Ridge are completely filled up with NADW.