Available online at www.sciencedirect.

com
R

Marine Geology 204 (2004) 43^57

www.elsevier.com/locate/margeo

Calcium carbonate corrosiveness in the South Atlantic

during the Last Glacial Maximum as inferred from changes
in the preservation of Globigerina bulloides:
A proxy to determine deep-water circulation patterns?
A.N.A. Volbers
, R. Henrich
University of Bremen, Faculty of Geosciences, Department of Paleoceanography and Sedimentology, P.O.Box 330440, D-28334 Bremen,
Germany
Received 11 April 2003; received in revised form 12 November 2003; accepted 9 December 2003

Abstract

The modern Atlantic Ocean, dominated by the interactions of North Atlantic Deep Water (NADW) and
Antarctic Bottom Water (AABW), plays a key role in redistributing heat from the Southern to the Northern
Hemisphere. In order to reconstruct the evolution of the relative importance of these two water masses, the NADW/
AABW transition, reflected by the calcite lysocline, was investigated by the Globigerina bulloides dissolution index
(BDXP). The depth level of the Late Glacial Maximum (LGM) calcite lysocline was elevated by several hundred
metres, indicating a more corrosive water mass present at modern NADW level. Overall, the small range of BDXP
data and the gradual decrease in preservation below the calcite lysocline point to a less stratified Atlantic Ocean
during the LGM. Similar preservation patterns in the West and East Atlantic demonstrate that the modern west^east
asymmetry did not exist due to an expansion of southern deep waters compensating for the decrease in NADW
formation.
9 2004 Elsevier B.V. All rights reserved.

Keywords: calcium carbonate dissolution; calcite lysocline; planktic foraminifera; Globigerina bulloides; Last Glacial Maximum;
South Atlantic

1. Introduction and stability are critical issues in understanding

the present state of climate and possibilities of
The pathways of large amounts of heat, fresh- future change (e.g. Gaunachaud and Wunsch,
water and nutrients, their transport mechanisms 2000). Deep-sea sediments reveal sudden spikes
in oceanic conditions, which correspond to cli-
mate shifts on land as recorded in the Greenland
* Corresponding author. Present address: Bundesanstalt ice cap (Bond et al., 1993; Marchitto et al., 1998;
fu«r Geowissenschaften und Rohsto¡e, Stilleweg 2,
30655 Hannover, Germany. Tel.: +49-511-6432785;
Severinghaus et al., 1998). There is evidence that
Fax: +49-511-6433663. climatic shifts are strongly related to deep-water
E-mail address: a.volbers@bgr.de (A.N.A. Volbers). reorganisations of a few-hundred-year time scales

0025-3227 / 04 / $ ^ see front matter 9 2004 Elsevier B.V. All rights reserved.
doi:10.1016/S0025-3227(03)00372-4

MARGO 3444 4-2-04

44 A.N.A. Volbers, R. Henrich / Marine Geology 204 (2004) 43^57

or less (e.g. Adkins et al., 1997). In particular, 1992; Labeyrie et al., 1996; Rutberg et al.,
deep-water formation in high northern latitudes 2000). Marchal et al. (2000) suggest a reduction
does not only seem to be the key when recon- of the thermohaline circulation by not more than
structing past climate changes but also when sim- 30%, whereas a 50% reduction of NADW was
ulating future climate conditions. suggested by Scha«fer-Neth and Paul (2001). As
Simulating the climate of the Last Glacial Max- the density of the sinking waters was reduced,
imum (LGM) is important because estimates of they would have penetrated less deeply, resulting
future change must be consistent with the sensi- in the formation of Glacial North Atlantic Inter-
tivity of the climate system to altered forcing pa- mediate Water (GNAIW, Sarnthein et al., 1994;
rameters. In addition, climate models must be Oppo and Lehmann, 1993; Broecker, 1997).
able to simulate the full range of dynamical be- Changes in the oceanic surface freshwater £uxes
haviour of the climate system, and extreme cli- are considered to have been the dominant forcing
matic periods, such as the LGM, may be the mechanism for the reduced Atlantic overturning
most critical tests (Stocker, 1998). During the (Schmittner et al., 2002). The recent work of Gro«-
LGM around 50 million km3 of water were ger et al. (2003) on grain-size analysis and carbon-
locked in huge ice sheets, lowering sea level by ate preservation in the western equatorial Atlantic
more than 120 m. Allen et al. (1999) provide evi- reports a reduced £ow velocity of Lower NADW
dence for a closely coupled ocean^atmosphere at a core position about 15 000 km down-current
system of the Northern Hemisphere during the from the production area, which generally sup-
LGM. The North Atlantic region and Europe re- ports the view of reduced glacial Lower NADW
ceived a strong cooling of more than 10‡C (CLI- £ux to the Atlantic Ocean. Although there is sub-
MAP, 1981; Guiot et al., 1989) and, according to stantial support for changes in water mass distri-
Lynch-Stieglitz et al. (1999), the speed of the Gulf bution in the deep Atlantic by Cd/Ca and 231 Pa/
230
Stream decreased by about 35% during this time Th data, neodymium isotope ratios, calcium
period. The Gulf Stream is an important part of carbonate and other proxy parameters (e.g. Mar-
today’s global ocean circulation as it carries warm tin and Lea, 1998; Venz et al., 1999; Marchal et
surface water northwards where it cools and sinks al., 2000; Rutberg et al., 2000; Gro«ger et al.,
to the sea£oor to form North Atlantic Deep 2003; Frenz and Henrich, 2004), Cd/Ca, Ba/Ca
Water (NADW, Dickson and Brown, 1994). As and 231 Pa/230 Th data of other studies suggest
long as salty water from the south is supplied and only little change from the glacial to the intergla-
vertical mixing continually removes surface water, cial periods (Lea and Boyle, 1990; Boyle, 1992;
the system seems to be stable regarding the for- Yu et al., 1996).
mation of NADW (Rahmstorf, 1997), and deep As emphasised by Rutberg et al. (2000), the
water moves across the deep ocean basins to the importance of ocean circulation in modulating
Indian Ocean and the Paci¢c Ocean, where it is the global climate warrants the use of additional
reheated and ¢nally turns back northward. proxies to resolve this con£ict. Here we use lyso-
Extensive reorganisation in the intermediate cline shifts to reveal speci¢c oceanic reorganisa-
and deep-water circulation has been hypothesised tions during the LGM. We present calcium car-
to have occurred during the LGM. There is evi- bonate preservation data determined by the
dence that around 21 kyr ago ocean circulation Globigerina bulloides dissolution index (BDXP,
was sluggish (e.g. Duplessy, 1999). Annual mean Volbers and Henrich, 2002a) to locate the posi-
meridional heat transport seems to have been re- tion of the LGM calcite lysocline. Early studies
duced by about 30% at low latitudes according to proposed that the position of the modern calcite
model data (Lynch-Stieglitz et al., 1999). Accord- lysocline is linked to the transition of NADW and
ingly, NADW production appears to have been Antarctic Bottom Water (AABW ; Berger, 1968)
decreased (Curry and Lohmann, 1983; Boyle and its position during the LGM could provide
and Keigwin, 1987; Oppo and Fairbanks, 1987; information on the distribution pattern of
Duplessy et al., 1988; Charles and Fairbanks, NADW and AABW during this time period

MARGO 3444 4-2-04

 A.N.A. Volbers, R. Henrich / Marine Geology 204 (2004) 43^57 45

(Berger and Wefer, 1996). Via core-top calibra- by the submarine barriers of the Mid-Atlantic
tions we are able to infer changes in its position Ridge in the west, and by the Walvis Ridge in
in the western and eastern South Atlantic basins. the south. Since only small quantities of AABW
can enter the Guinea and Angola Basin via the
Romanche Fracture Zone and the Walvis Pas-
2. Modern and past South Atlantic deep-water sage, they are predominately ¢lled by NADW
circulation (Van Bennekom and Berger, 1984; Shannon and
Chapman, 1991; Warren and Speer, 1991; Mer-
The world’s oceans circulation was ¢rst de- cier et al., 1994).
scribed by Broecker as the ‘Global Ocean Con- When simulating past climate changes, the for-
veyor Belt’, in which the Atlantic Ocean plays a mation of deep water at high northern latitudes
key role (Broecker et al., 1985; Gordon, 1986; seems to be a crucial point (e.g. Ganopolski et al.,
Broecker, 1991). Not only South Atlantic surface 1998; Ganopolski and Rahmstorf, 2001). Accord-
water masses cross the equator but also intermedi- ing to the computer model of Ganopolski et al.
ate and deep waters from the south move north- (1998) simulating the atmospheric and oceanic
wards. According to Gordon et al. (1992), most of circulations of the LGM, oceanic circulation
the net meridional £ow within the South Atlantic changes might have ampli¢ed Northern Hemi-
heading towards the north occurs in the AAIW sphere cooling during the LGM by about 50%.
layer rather than in the thermocline layer. This In response to sea-ice advance from around
£ow is compensated by the supply of NADW 75‡N to 55‡N, a southward shift of NADW for-
characterised by its low nutrient concentration mation occurred (Ganopolski et al., 1998; Stock-
and relatively high N13 C values of dissolved inor- er, 1998). As the density of the sinking waters was
ganic carbon (Dickson and Brown, 1994). On its reduced, they might have penetrated less deeply,
way to the southern ocean, it subdivides the Cir- forming Glacial North Atlantic Intermediate
cumpolar Deep Waters (CDW) from the south at Water (GNAIW ; e.g. Boyle and Keigwin, 1987;
mid-depth. The upper branch of CDW, the Upper Zahn, 1992; Weaver et al., 1994; Broecker, 1997).
Circumpolar Deep Water (UCDW), can be traced According to Oppo and Lehmann (1993), deep
through the Atlantic Ocean due to its silica max- waters from the south may have replaced
imum (Talley, 1996). UCDW and AAIW have NADW below around 2000 m water depth while
low [CO233 ] compared to the surrounding water GNAIW occupied the shallower depths.
mass bodies and can be traced by their low ara- Accordingly, the glacial counterpart of AABW
gonite preservation potential (Gerhardt and Hen- could have penetrated further into the northern
rich, 2001). The Lower Circumpolar Deep Water North Atlantic as inferred from palaeoceano-
(LCDW) from the Drake Passage overrides the graphic data (e.g. Duplessy et al., 1988; Sarnthein
denser waters from the Weddell Sea and ¢lls the et al., 1994). Recently, Hagen and Keigwin (2002)
deep western basin and the Cape Basin. LCDW assumed that during stadial coolings, Southern
and the Weddell Sea Deep Water (WSDW) are Component Water (SCW) dominated the deep
commonly referred to as AABW, which is a water in the North Atlantic, probably to as shal-
cold water mass of low salinity (Siedler et al., low as approximately 3000 m at the Blake Outer
1996). AABW and NADW di¡er signi¢cantly in Ridge (around 33‡N). Sarnthein et al. (2001) in-
their physical properties, nutrient concentration, ferred the presence of a glacial derivate of south-
and [CO23 3 ]. Whereas NADW is a relatively ern-source bottom water in the deep eastern At-
warm and well-ventilated water mass formed in lantic continental margin up to approximately
the Greenland, Iceland, and Labrador Seas, 2800^2600 m depth along the Rockall Plateau
AABW originates near Antarctica and contains (around 56‡N).
less oxygen and higher nutrients (e.g. Reid, The modern asymmetry between the western
1996). The in£ow of South Atlantic deep waters and eastern South Atlantic regarding its deep
into the eastern South Atlantic basin is controlled water mass distribution may not have existed dur-

MARGO 3444 4-2-04

46 A.N.A. Volbers, R. Henrich / Marine Geology 204 (2004) 43^57

ing times of decreased NADW production. Bick- However, as summarised by Sigman and Boyle
ert (1992) inferred from N13 C data that SCW was (2000), continued studies of sea£oor dissolution
present in the Angola Basin during glacial peri- have indicated that pore water undersaturation
ods. These late Pleistocene shifts in the NADW/ due to oxidation of organic matter within the
AABW transition as determined by proxies such sediments is also a signi¢cant driver of calcite
as N13 C or Ca/Cd are also re£ected by changes in dissolution. The importance of this e¡ect was al-
the calcium carbonate preservation pattern since ready introduced by Emerson and Bender (1981)
they are linked with a redistribution of the car- twenty years ago but so far treated as an e¡ect of
bonate system, including a reduction in the atmo- minor importance. But nevertheless, when this ad-
spheric CO2 content and a shallowing of the cal- ditional ‘respiratory dissolution’, is taken into ac-
cite lysocline in the North Atlantic (e.g. Hagen count, the depth of the lysocline does not neces-
and Keigwin, 2002) and South Atlantic deep ba- sarily correspond to the calcite saturation depth
sins (e.g. Bickert and Wefer, 1996, Anderson and (Volbers, 2001; Volbers and Henrich, 2002b). For
Archer, 2002, this study). the same reason, calci¢cation of benthic forami-
nifera in 13 C-depleted pore water should be con-
sidered when interpreting N13 C data (e.g. Hagen
3. The imprint of South Atlantic deep-water mass and Keigwin, 2002). For this reason, we focus on
distribution on the BDXPP record obtained from several di¡erent environments of the South Atlan-
surface sediments tic Ocean to determine the calcium carbonate
preservation state. In order to determine the mod-
Around 55% of the ocean’s £oor is covered by ern position of the calcite lysocline, all South At-
carbonate-rich sediments (Milliman, 1993). In lantic basins were investigated and compared to
particular, within the South Atlantic, most bio- an extensive LGM data set.
genic sediments are calcareous and originate
from various micro- and nannofossil groups. Re- 3.1. Eastern South Atlantic
cent investigations con¢rm that planktic forami-
nifera are the main contributors to the South At- In the Cape Basin the modern position of the
lantic deep-sea sediments besides coccolithoporids calcite lysocline is encountered at 4000 m and also
and pteropods, particular in the equatorial and re£ected by a decrease in CaCO3 values and an
Mid-Atlantic Ridge region (Volbers, 2001). increase in Total Organic Carbon (TOC) content
Deep-sea calcium carbonate burial is mediated (Fig. 1, Volbers and Henrich, 2002b). The
by the depth of the calcite lysocline which is NADW/AABW transition correlates with the po-
marked by the ¢rst signs of signi¢cant dissolution sition of the calcite lysocline in the northern Cape
in deep-sea sediments (Milliman et al., 1999). One Basin. AABW ¢lls the deep Cape Basin below
of the major controls on the depth of the lysocline 4000 m re£ected by the onset of decreasing Glo-
is the calcite saturation depth. At this depth level, bigerina bulloides preservation at around 4000 m.
the carbonate ion concentration of the bottom Whereas the calcite lysocline within the Cape Ba-
water [CO23 3 ]bw equals the carbonate ion concen- sin appears to be predominantly a water mass
tration at calcite saturation [CO23
3 ]sat . Above this signal, the situation is di¡erent for the Guinea
level, sea waters bathing sediments are supersatu- and Angola Basin. Here, the calcite lysocline is
rated with respect to calcite and calcareous sedi- encountered at 4000 m water depth in the deep
ments should be well preserved, whereas dissolu- Angola Basin (Volbers and Henrich, 2002b). As
tion proceeds in the undersaturated waters below due to the absence of AABW the [CO23 3 ] content
this depth level. As calcite solubility increases in the water column stays nearly constant with
with depth (pressure e¡ect), water depth and the increasing water depth (Broecker and Takahashi,
[CO233 ] of the surrounding water mass should be 1978), the observed (bad) preservation pattern
the dominant controls on calcium carbonate pres- was attributed to additional dissolution due to
ervation. microbial degradation of organic matter in the

MARGO 3444 4-2-04

 A.N.A. Volbers, R. Henrich / Marine Geology 204 (2004) 43^57 47

Fig. 1. (a) Determination of the sedimentary calcite lysocline within the northern Cape Basin. The progressive dissolution of Glo-
bigerina bulloides ultrastructure, expressed as BDXP, is compared with CaCO3 data, TOC, and the burial ratio which could be
used as calcium carbonate dissolution proxies in this region (taken from Volbers and Henrich, 2002b). Below the calcite lysocline,
determined from a BDXP value of 3 (Volbers and Henrich, 2002a,b), BDXP values increase rapidly as well as TOC and the burial
ratio, whereas the CaCO3 ratio decreases. (b) the BDXP preservation curve pattern is compared to the [CO23 3 ] of the di¡erent
water masses that ¢ll the Cape Basin today. Water masses with higher [CO23 3 ], e.g. NADW, are marked in grey whereas water
masses of lower [CO23 3 ], e.g. the more corrosive AABW, are marked in dark grey. The NADW/AABW transition coincides with
the calcite lysocline determined via the BDXP in the modern Cape Basin.

uppermost sediment layer. By this, the decrease in 3.2. Western South Atlantic
[CO23
3 ] must be in the order of 30 Wmol/kg to
cross the calcite saturation curve at 4000 m (Volb- In order to depict the modern position of the
ers and Henrich, 2002b). The very low sedimenta- calcite lysocline in the western South Atlantic and
tion rates within the Mid-Atlantic Ridge transect compare it with its position during the LGM, 75
(Milliman, 1993) may account for enhanced surface sediment samples were investigated (for
supralysoclinal dissolution within the surface sedi- detailed description of the BDXP, see Section
ments of this region. 4.2). Fig. 2 shows the present water mass con¢g-

MARGO 3444 4-2-04

48 A.N.A. Volbers, R. Henrich / Marine Geology 204 (2004) 43^57

Fig. 2. BDXP results from surface samples introduced in this study plotted into the western Atlantic water mass distribution pat-
tern determined by Gerhardt and Henrich (2001). Samples within upper NADW show best preservation of Globigerina bulloides
tests, whereas samples within AAIW and UCDW have slightly increased BDXP values, indicating some (local) calcite dissolution
at the intermediate depth level today. Although deeper in the water column (increased pressure e¡ect), BDXP values report better
calcium carbonate preservation in upper NADW (a water mass with higher [CO23 3 ]) compared to preservation of samples bathed
in AAIW (shallower in the water column but with a lower [CO23 3 ]). Globigerina bulloides preservation declines at the NADW/
AABW transition zone, where the deep calcite lysocline is encountered. For samples indicated by ‘x’, the BDXP failed because
the minimum number of G. bulloides, which is ten tests, was not available.

uration inferred from GEOSECS data (modi¢ed with the AABW/NADW transition but seems to
from Gerhardt and Henrich, 2001) in which re- be shallower in the equatorial region because of
sults from Globigerina bulloides ultrastructural in- additional dissolution by organic matter degrada-
vestigation were plotted. We observe moderate tion.
preservation of G. bulloides test features (2.9^
3.1) from samples in the intermediate depth level,
where AAIW and UCDW are located. Samples 4. Materials and methods
that are bathed by NADW reveal better preserva-
tion even below 3000 m depth than samples lo- 4.1. LGM core material
cated within AAIW and UCDW (Fig. 2). A sig-
ni¢cant decrease in G. bulloides preservation is The LGM was de¢ned as the 23^19 ka calendar
observed at around 3500 m water depth at 35‡S age period (or 19.5^16.1 ka corrected 14 C age) on
latitude. This level descends to 4000 m depth at the ¢rst EPILOG workshop in May 1999 in Del-
20‡S and reaches ¢nally 3900 m in the equatorial menhorst (compare Mix et al., 2001). In this
region. Up to 20‡S, the calcite lysocline as deter- study, 48 cores from all over the South Atlantic
mined by the BDXP seems to be closely linked (Fig. 3) were investigated in order to determine

MARGO 3444 4-2-04

 A.N.A. Volbers, R. Henrich / Marine Geology 204 (2004) 43^57 49

Fig. 3. Map of core locations from which LGM samples were determined. BDXP results are shown in brackets.

their calcium carbonate preservation state. The 10 kN accelerating voltage) and carried out at a
majority of these cores were recovered on various magni¢cation of 4000U. In order to do this, at
RV METEOR cruises, whereas the rest was taken least ten specimens per sample were glued to a
from RV Sonne and Victor Hensen cruises. The SEM stub and gold coated.
stratigraphic control in all investigated cores is
based on a combination of N18 O and 14 C 4.2. The Globigerina bulloides dissolution index
(AMS). If 14 C dates have not been available, oxy- (BDXP)
gen isotope stage 2.22 was identi¢ed, which cor-
responds to a 14 C age of about 16 ka and a cal- Globigerina bulloides is among the most abun-
endar age of 18.9 ka (Mix and Ruddiman, 1985; dant planktic foraminifera in the eastern South
Bard, 1998). Further details on stratigraphy are Atlantic today (Kemle-von Mu«cke and Oberha«ns-
given in Niebler et al. (2003). LGM material li, 1999). It is a reliable indicator of foraminiferal
was wet-sieved to separate the ¢ne from the carbonate preservation via the identi¢cation of six
coarse fraction which was then dried and dissolution stages based on four distinct ultra-
weighted. Whenever available, 30 Globigerina bul- structural features (spines, pores, ridges, interpore
loides from the LGM coarse fraction were inves- area). In this study, the (sedimentary) ‘calcite ly-
tigated to determine the position of the sedimen- socline’ as de¢ned by Milliman et al. (1999) is set
tary calcite lysocline in the South Atlantic. to the depth level where the ¢rst observation of
Ultrastructural investigations were performed on signi¢cant carbonate dissolution in calcareous
a ZEISS DSM 940 A (10 mm working distance, sediments was observed, which is re£ected by

MARGO 3444 4-2-04

50 A.N.A. Volbers, R. Henrich / Marine Geology 204 (2004) 43^57

the BDXP preservation stage of 3 (Volbers and four. This method assures that every criterion is
Henrich, 2002a). The following is a brief summary judged equally when calculating the BDXP:
of the preservation features and preservation
BDX 0 ¼ 4 P=nT ð2Þ
stages used in this study:
Stage 0^1 (excellent preservation): recognised With 4P as the sum of all individual preserva-
by round pores, smooth interpore areas, intact tion states of Globigerina bulloides divided by the
ridges, and high spines. It is only observed on number of investigated tests (nT).
Globigerina bulloides tests taken from mooring
samples (after storage in mercury-saturated sea-
water which, of course, may have led to some 5. Results and discussion
additional dissolution features).
Stage 1^2 (very good to good preservation): According to the present west^east asymmetry
identi¢ed by slightly widened pores, partly etched of South Atlantic deep-water mass distribution
interpore areas, slightly denuded ridges and pattern, we provide two graphs, one covering
spines. Only observed at Globigerina bulloides the western Atlantic basins (Fig. 4a), and the oth-
tests taken from mooring samples, hardly ob- er the eastern Atlantic basins (Fig. 4b).
served in surface sediment samples.
Stage 2^3 (good to moderate preservation): rec- 5.1. Western South Atlantic
ognised by funnellike pores, etched and rough-
ened interpore areas, denuded ridges and spines. As determined via the BDXP, LGM sediment
This preservation stage is normally found in samples reveal good to moderate preservation
surface sediments under slightly corrosive condi- throughout the water column until the calcite ly-
tions. socline is encountered. In the western Atlantic
Stage 3^4 (bad preservation) : revealed by Ocean, the calcite lysocline is located at around
partly interconnected pores, partial removal of 3500 m water depth in the northern and southern
the surface layer, strongly reduced ridges and parts of the study area, constantly rising towards
spines. It is an indicator of moderate corrosive the south. This is several hundred metres shal-
conditions much as occur during the respiration lower than today. Broecker and Clark (2001) re-
of organic matter in the eastern South Atlantic ported that at the 2800-m depth level in the vicin-
surface sediments. ity of the Ceara Rise, the calculated [CO23 3 ] was
Stage 4^5 (very bad preservation): indicated by higher than today whereas it was lower at 4.0-km
interconnected pores, removal of the surface depth during the LGM. The depth separating the
layer, missing ridges and spines. This preservation waters with higher concentrations from those with
stage is found in surface sediments under highly lower concentrations was located at around 3.5
corrosive conditions. km, in accordance with our ¢ndings in this area.
Stage 5: indicated by the absence of intact tests Our data imply that the position of the LGM
due to severe dissolution. calcite seemed to have been generally elevated in
The BDXP value was calculated according to the Atlantic Ocean during the LGM besides the
the following equations: deepening around the equatorial region (Fig.
P ¼ ðPðporesÞ þ Pðinterpore areaÞ þ PðspinesÞ
4a,b). In the modern South Atlantic Ocean, the
calcite lysocline is shallowest in the equatorial re-
þPðridgesÞ Þ=4 ð1Þ gion compared to the open ocean but at its deep-
est position of all eastern South Atlantic basins
The preservation state (P, between 0.5 and 4.5) during LGM. We attribute this deepening to an
of each single Globigerina bulloides is calculated increase in coccolith production because of stron-
from the preservation of each criterion investi- ger LGM upwelling intensity (Kinkel et al., 2004).
gated (pores, interpore area, spines, and ridges) The resultant higher supply of carbonate to the
divided by the number of criteria, which are sea£oor and the rapid embedding process seems

MARGO 3444 4-2-04

 A.N.A. Volbers, R. Henrich / Marine Geology 204 (2004) 43^57 51

to account for a better preservation of the calcar- ably lower during the LGM than during the Ho-
eous material. locene.
As indicated by Fig. 2, BDXP data obtained
from surface sediment samples re£ect modern 5.2. Eastern South Atlantic basins
South Atlantic oceanography well. In contrast
to the strati¢ed modern South Atlantic Ocean, As in the western South Atlantic, BDXP data
BDXP values determined from LGM samples point to a signi¢cantly less strati¢ed water mass in
are rather evenly distributed throughout the water the eastern South Atlantic basins. When com-
column. Compared to the modern situation, pared to the core-top data set of the eastern South
we do not see any signi¢cant changes in BDXP Atlantic (Volbers and Henrich, 2002b), BDXP val-
preservation ^ could this account for an ocean ues are within a narrower range. BDXP values
that was signi¢cantly less strati¢ed in the inter- vary only slightly within the 0^3000-m depth in-
mediate and upper deep water layer during the terval and do not decrease as rapidly below the
LGM? calcite lysocline as observed in the modern ocean.
Similar results were reported by Gerhardt and The position of the calcite lysocline during the
Henrich (2004) who investigated the preservation LGM was encountered at around 3200 m water
state of Limacina in£ata, a pteropod which secrets depth in the southernmost part of the investigated
an aragonitic shell. Limacina in£ata index (LDX) area dropping towards the equator. As in the
data from surface sediments re£ect the modern western Atlantic Ocean a deep calcite lysocline
water mass distribution pattern. Good preserva- was observed in the equatorial region.
tion of L. in£ata tests was found within the sur- Our ¢ndings of a shallower calcite lysocline
face waters and the upper NADW whereas rather during the LGM are in accordance with the cal-
poor preservation was reported from samples that culations of Anderson and Archer (2002) who de-
are bathed by aragonite-corrosive intermediate termined a decrease in Atlantic deep water [CO23 3 ]
water masses of southern origin, such as AAIW across a broad depth range. Following our data
and UCDW (Gerhardt and Henrich, 2001). In and the variations in size composition of LGM
contrast, they report a rather regular increase in and core-top sediments which was determined
their LDX data with water depth in western by Frenz and Henrich (2004), the glacial calcite
South Atlantic sediments of LGM age. lysocline rose several hundred metres compared to
The modern contrast in both Limacina in£ata its modern position. In addition, we observe a
and Globigerina bulloides test preservation pat- rather gradual increase in dissolution below the
terns was obviously diminished during the LGM lysocline compared to today, indicating
LGM. These results seem to point to an inter- that the dissolution gradient was less steep. Our
mediate water mass that is only slightly corrosive data are consistent with those of Anderson and
to aragonite and calcite, most probably re£ecting Archer (2002) who found some indication that the
the presence of GNAIW in the western South v[CO233 ] in the deep water s 4000 m had in-
Atlantic. There is evidence from the N13 C and creased.
Cd/Ca data of Oppo and Horowitz (2000) that In addition, whereas there is a contrast in the
GNAIW extended at least as far south as 28‡S modern deep-water mass distribution between the
during the LGM. According to Gerhardt and Guinea and Angola Basins (both dominated by
Henrich (2004), the aragonite saturation state at NADW to great depths) and the Cape Basin
the sediment^water interface within GNAIW (AABW below 4000 m), this north^south asym-
might have been similar to that within modern metry seems to have been diminished during the
UNADW conditions. This would also explain LGM (this study). N13 C data of benthic forami-
rather constant BDXP data within the 0^2000 m nifera point towards the presence of SCW in the
interval. In addition, Bickert and Mackensen eastern South Atlantic basin (Bickert, 1992). The
(2004) report that vertical N13 C gradients espe- upper part of the glacial SCW, characterised by
cially in the deep western Atlantic are consider- N13 C values between 0.18x and 0.44x, could

MARGO 3444 4-2-04

52 A.N.A. Volbers, R. Henrich / Marine Geology 204 (2004) 43^57

MARGO 3444 4-2-04

 A.N.A. Volbers, R. Henrich / Marine Geology 204 (2004) 43^57 53

£ow across the Walvis Ridge ¢lling the deepest mass with respect to aragonite and calcite and
part of the eastern Angola and Guinea Basins, the local undersaturation in the sediments due
spreading from there in the western and northern to the production of metabolic CO2 . To date
basins (Bickert and Mackensen, 2004). The simi- these individual factors can hardly be separated
lar calcium carbonate preservation patterns of the and the usage of the calcite lysocline or the ara-
western and eastern South Atlantic deep-sea ba- gonite compensation depth to determine the
sins also refer to a rather uniform water mass South Atlantic water mass con¢guration during
distribution pattern in the western and eastern the LGM might therefore be biased to some ex-
South Atlantic Ocean in contrast to the modern tent. The usage of the calcite lysocline for the
situation. reconstruction of past water mass distribution
patterns and the calculation of past [CO23 3 ] distri-
5.3. Deep-water mass distribution during the LGM bution has to be carried out with caution in as
many di¡erent regions as possible to obtain an
The recent work of Broecker and Clark (2001) overview. For this reason we use an extensive
reported a broad range in the [CO23 ] of the west- data set of surface sediment samples and LGM
ern equatorial Atlantic waters in the vicinity of material from all the di¡erent hydrographical set-
the Ceara Rise during the LGM as calculated tings of the South Atlantic.
from planktic foraminiferal shell loss. Compared What we ¢nally see from Globigerina bulloides
with the investigated section between 2800 and ultrastructural investigation of LGM sediment
4000 m in their study, the modern CO2 3 distribu- material is that: (1) in general, BDXP data are
tion looks rather homogenous (Broecker and within a smaller range throughout the whole
Clark, 2001). If we convert their ¢ndings to water column compared to the modern ocean;
BDXP preservation stages, we can expect good (2) the G. bulloides state of preservation does
preservation around 2800 m and very bad preser- not decrease as rapidly below the calcite lysocline
vation around 4000 m. Our data, taken from a as observed in the modern ocean; (3) the position
slightly di¡erent location between around 2^6‡N of the calcite lysocline seems to have been at the
in the western Atlantic, show good to moderate same depth level in both the western and eastern
preservation (BDXP of 2.8^3.0) between 2200 and South Atlantic Ocean during the LGM ; (4) it was
3200 m and bad preservation (BDXP of 3.2^3.3) several hundred metres shallower in the northern
below 3500 m. According to the [CO23 3 ] calculated and southern parts of the investigated area; and
by Broecker and Clark (2001), we would have (5) the course of the calcite lysocline is opposite to
expected to ¢nd a much wider range in BDXP its modern mode in the equatorial region with a
values. This is obviously not the case. As noted deep lysocline below 4000 m.
already, the reconstruction of past CO23 3 su¡ers Along the western continental margin no indi-
from the common drawback that much of the cation for aragonite/calcite-corrosive glacial
dissolution occurs in the sediment pore waters AAIW was found. Although Haddad and Droxler
and DeVilliers (2003) recently questioned the pro- (1996) assume that AAIW was formed during the
posed use of shell weight as a reliable proxy of LGM, it must have been restricted to the South-
carbonate saturation. Calcium carbonate dissolu- ern Hemisphere since they did not ¢nd any in£u-
tion in deep-sea sediments is controlled by several ence of corrosive, southern source water in the
important factors of capital importance, the gen- Caribbean at this time. Our data indicate that, if
eral undersaturation of the surrounding water glacial AAIW was formed it did not spread as far

Fig. 4. BDXP data determined in this study are separated into western (a) and eastern Atlantic transects (b). BDXP results were
plotted in the LGM water mass distribution model presented by Bickert (1992). BDXP data reveal constant good to moderate
preservation throughout the water column until the calcite lysocline is encountered. The position of the calcite lysocline during
the LGM was shallower compared to its modern position with exception of the equatorial region.

MARGO 3444 4-2-04

54 A.N.A. Volbers, R. Henrich / Marine Geology 204 (2004) 43^57

as 10‡S. A rather uniform water mass, which is calcite dissolved several hundred metres above the
supersaturated with respect to calcite, is indicated present level in consequence of a more calcite-cor-
by good to moderate preservation of Globigerina rosive water mass present at NADW level during
bulloides to 3500 m depth, indicating a strength- the LGM. The LGM calcite lysocline was found
ened GNAIW £ow into the South Atlantic Ocean. at a similar depth level in both the western and
Our ¢ndings were con¢rmed by the work of Ger- eastern South Atlantic basins demonstrating the
hardt and Henrich (2004) who report slightly bet- presence of SCW, the glacial counterpart of
ter preservation of Limacina in£ata shells (arago- AABW, in the Angola and Guinea Basin at
nite) within the 2000^3000-m level during the mid-depth. In order to penetrate into the eastern
LGM compared to the modern environment. basins, SCW must have expanded in the water
Their ¢ndings were attributed to the presence of column to ¢nally override the critical sill depths
GNAIW to around 3000 m, whereas the arago- of the Walvis Ridge and the Mid-Atlantic Ridge
nite compensation depth (ACD) was suggested to that currently prevent AABW in£ow (Bickert and
re£ect the GNAIW/SCW boundary. According to Wefer, 1999; this study). BDXP data suggest that
Duplessy et al. (1988), the GNAIW core was re- SCW rose up to 3500 m and ascended towards
constructed by calculated N13 C values of total dis- 3000 m in the south, providing support for
solved CO2 as determined from benthic foraminif- the theory of reduced glacial dominance of
eral N13 C values down to almost 3000 m between NADW in the Atlantic Ocean. Overall, BDXP
15‡S and 15‡N, including a mixing zone down to preservation data point to a less strati¢ed ocean
3500 m. Since the modern west^east asymmetry during the LGM, re£ected by the narrow range of
was not observed by Duplessy et al. (1988), Bick- BDXP values and the gradually declining G. bul-
ert (1992) and in this study, a rise of SCW in the loides preservation pattern below the calcite lyso-
South Atlantic water column to depth levels cline.
around 3500 m seemed to have occurred during
the LGM. The position of the calcite lysocline
during the LGM as determined by the BDXP Acknowledgements
could re£ect the upper edge of the SCW. Bickert
(1992) suggested to even subdivide this glacial We are grateful to M. Frenz who made the
water mass into an upper (USCW) and a lower sand fraction of LGM core material available to
part (LSCW, Fig. 4a,b) by N13 C values. It cannot us and we thank S. Gerhardt for discussion. For
be clearly determined whether USCW might be a thoughtful reviews, we thank D. Archer and B.
separate water mass as suggested by Bickert Ruddiman. BDXP values presented in this paper
(1992) or might re£ect the mixing between are archived in the PANGAEA database at the
NCW/GNAIW and SCW. When referred to Alfred Wegener Institute for Polar and Marine
BDXP values, which are in a narrow range and Research (http://www.pangaea.de/home/avolb-
re£ect continuous worsening below the calcite ly- ers). This research was funded by the German
socline, the suggested USCW water mass rather Research Foundation, DFG (Collaborative Re-
seems to re£ect the mixing between NCW/ search Center SFB 261 at the University of Bre-
GNAIW and SCW. men).

6. Conclusions References

Ultrastructural investigations of the planktic Adkins, J.F., Boyle, E.A., Keigwin, L., Cortijo, E., 1997. Var-
foraminifera Globigerina bulloides taken from iability of the North Atlantic thermohaline circulation dur-
ing the last interglacial period. Nature 390, 154^156.
LGM core materials reveal large-scale changes Allen, J.R.M., Brandt, U., Brauer, A., Hubberten, H.-W.,
in the position of the calcite lysocline in the At- Huntley, B., Keller, J., Kraml, M., Mackensen, A., Mini-
lantic Ocean. In contrast to the modern settings, gram, J., Negendank, J.F.W., Nowaczyk, N.R., Oberha«nsli,

MARGO 3444 4-2-04

 A.N.A. Volbers, R. Henrich / Marine Geology 204 (2004) 43^57 55

H., Watts, W.A., Wulf, S., Zolitschka, B., 1999. Rapid en- lysocline depth and in situ carbonate ion concentration.
vironmental changes in southern Europe during the last gla- Deep-Sea Res. 25, 65^95.
cial period. Nature 400, 740^743. Charles, C.D., Fairbanks, R.G., 1992. Evidence from Southern
Anderson, D.M., Archer, D., 2002. Glacial^interglacial stabil- Ocean sediments for the e¡ect of North Atlantic deep-water
ity of ocean pH inferred from foraminifer dissolution rates. £ux on climate. Nature 315, 21^26.
Nature 416, 70^73. CLIMAP Project Members, 1981. Seasonal reconstructions of
Bard, E., 1998. Geochemical and geophysical implications of the Earth’s surface at the Last Glacial Maximum. Geol. Soc.
the radiocarbon calibration. Geochim. Cosmochim. Acta 62, Am., GSA Map and Chart Ser. MC-36, Boulder, CO, 18 pp.
2025^2038. Curry, W.B., Lohmann, G., 1983. Reduced advection into At-
Berger, W.H., 1968. Planktonic foraminifera: Selective solu- lantic Ocean deep eastern basins during last glaciation max-
tion and paleoclimatic interpretation. Deep-Sea Res. 15, imum. Nature 308, 317^342.
31^43. DeVilliers, 2003. A 425 kyr record of foraminiferal shell weight
Berger, W.H., Wefer, G., 1996. Central themes of South At- variability in the western equatorial Paci¢c. Paleoceanogra-
lantic circulation. In: Wefer, G., Berger, W.H., Siedler, G., phy 18, 1080.
Webb, D.J (Eds.), The South Atlantic: Present and Past Dickson, R.R., Brown, J., 1994. The production of North
Circulation. Springer, Berlin, pp. 1^11. Atlantic Deep Waters: Sources, rates and pathways. J. Geo-
Bickert, T., 1992. Rekonstruktion der spa«tquarta«ren Boden- phys. Res. 99, 12319^12341.
wasserzirkulation im o«stlichen Su«datlantik u«ber stabile Iso- Duplessy, J.C., Shackleton, N.J., Fairbanks, R.G., Labeyrie,
tope benthischer Foraminiferen. Ber. Fachbereich Geowiss. L., Oppo, D., Kallel, N., 1988. Deep water source variation
Univ. Bremen 27, 205 pp. during the last climatic cycle and their impact on global
Bickert, T., Mackensen, A., 2004. Last Glacial to Holocene deepwater circulation. Paleoceanography 3, 343^360.
changes in South Atlantic deep water circulation. In: Wefer, Duplessy, J.-C., 1999. Climate and the Gulf Stream. Nature
G., Mulitza, S., Ratmeyer, V. (Eds.), The South Atlantic in 402, 593^594.
the Late Quaternary: Reconstruction of Material Budget Emerson, S., Bender, M., 1981. Carbon £uxes at the sediment^
and Current Systems. Springer, Berlin, in press. water interface of the deep-sea: Calcium carbonate preser-
Bickert, T., Wefer, G., 1996. Late Quaternary deep water cir- vation. J. Mar. Res. 39, 139^162.
culation in the South Atlantic: Reconstruction from carbon- Frenz, M., Henrich, R., 2004. Carbonate dissolution revealed
ate dissolution and benthic stable isotopes. In: Wefer, G., by silt grain-size distributions: Comparison of Holocene and
Berger, W.H., Siedler, G., Webb, D.J. (Eds.), The South last glacial maximum sediments from the pelagic South At-
Atlantic: Present and Past Circulation. Springer, Berlin, lantic. Sedimentology, submitted.
pp. 599^620. Ganopolski, A., Rahmstorf, S., Petoukhow, V., Claussen, M.,
Bickert, T., Wefer, G., 1999. South Atlantic and benthic for- 1998. Simulation of modern and glacial climates with a
aminifer N13 C deviations: Implications for reconstructing the coupled global model of intermediate complexity. Nature
late Quaternary deep-water circulation. Deep-Sea Res. II 46, 391, 351^356.
437^452. Ganopolski, A., Rahmstorf, S., 2001. Rapid changes of glacial
Bond, G., Broecker, W., Johnsen, S., McManus, J., Labeyrie, climate simulated in a coupled climate model. Nature 409,
L., Jouzel, J., Bonani, G., 1993. Correlations between cli- 153^158.
mate records from North Altnaitc sediments and Greenland Gaunachaud, A., Wunsch, C., 2000. Improved estimates of
ice. Nature 365, 143^147. global ocean circulation, heat transport and mixing from
Boyle, E.A., 1992. Cadmium and N13 C paleochemical ocean hydrographic data. Nature 408, 453^457.
distributions during the Stage 2 glacial maximum. Annu. Gerhardt, S., Henrich, R., 2001. Shell preservation of Limacina
Rev. Earth Planet. Sci. 20, 245^287. in£ata (Pteropoda) in surface sediments from the Central
Boyle, E.A., Keigwin, L., 1987. North Atlantic thermohaline and South Atlantic Ocean: A new proxy to determine the
circulation during the past 20,000 years linked to high-lat- aragonite saturation state of water masses. Deep-Sea Res. I
itude surface temperature. Nature 350, 35^40. 48, 2051^2071.
Broecker, W.S., 1991. The great ocean conveyor. Oceanogra- Gerhardt, S., Henrich, R., 2004. Intermediate water circulation
phy 4, 79^89. during the Last Glacial Maximum in the South Atlantic
Broecker, W.S., 1997. Thermohaline circulation, the Achilles Ocean inferred from changes in the shell preservation of
heel of our climate system: Will man made CO2 upset the Limacina in£ata (Pteropoda). Deep-Sea Res. I, submitted.
current balance? Science 278, 1582^1588. Gordon, A.L., 1986. Interocean exchange of thermocline
Broecker, W.S., Clark, E., 2001. Glacial-to-Holocene redistri- water. J. Geophys. Res. 91, 5037^5046.
bution of carbonate ion in the deep sea. Nature 294, 2152^ Gordon, A.L., Weiss, R.F., Smethie, W.M., Jr., Warner, M.J.,
2155. 1992. Thermocline and intermediate water communication
Broecker, W.S., Peteet, D.M., Rind, D., 1985. Does the ocean- between the South Atlantic and Indian Oceans. J. Geophys.
atmosphere system have more than one stable mode of op- Res. 97, 7223^7240.
eration? Nature 315, 21^26. Gro«ger, M., Henrich, R., Bickert, T., 2003. Glacial^interglacial
Broecker, W.S., Takahashi, T., 1978. The relationship between variability in lower North Atlantic deep water: Inference

MARGO 3444 4-2-04

56 A.N.A. Volbers, R. Henrich / Marine Geology 204 (2004) 43^57

from silt grain-size analysis and carbonate preservation in cesses of the Ice Age: Land, oceans, glaciers (EPILOG).
the western equatorial Atlantic. Mar. Geol. 201, 321^332. Quat. Sci. Rev. 20, 627^657.
Guiot, J., Pons, A., de Beaulieu, J.L., Reille, M., 1989. A Mix, A.C., Ruddiman, W.F., 1985. Structure and timing of the
140,000-year continental climate reconstruction from two last deglaciation: Oxygen isotope evidence. Quat. Sci. Rev.
European pollen records. Nature 338, 309^313. 4, 59^108.
Haddad, G.A., Droxler, A.W., 1996. Metastable CaCO3 dis- Niebler, H.-S., Arz, H.W., Donner, B., Mulitza, S., Pa«tzold, J.,
solution at intermediate water depths of the Caribbean and Wefer, G., 2003. Sea-surface temperature in the equatorial
western North Atlantic: Implications for intermediate water and South Atlantic Ocean during the Last Glacial Maxi-
circulation during the past 200,000 years. Paleoceanography mum (23^19 ka). Paleoceanography 18, 1069.
11, 701^716. Oppo, D.W., Fairbanks, R.G., 1987. Variability in the deep
Hagen, S., Keigwin, L.D., 2002. Sea-surface temperature var- and intermediate water circulation of Atlantic Ocean during
iability and deep water reorganisation in the subtropical the past 25,000 years: Northern Hemisphere modulation of
North Atlantic during Isotope Stage 2^4. Mar. Geol. 189, the Southern Ocean. Earth Planet. Sci. Lett. 86, 1^15.
145^162. Oppo, D.W., Horowitz, M., 2000. Glacial deep water geome-
Kemle-von Mu«cke, S. and Oberha«nsli, H., 1999. The distribu- try: South Atlantic benthic forminiferal Cd/Ca and N13 C
tion of living planktic foraminifera in relation to Southeast evidence. Paleoceanography 15, 147^160.
Atlantic Oceanography. In: Fischer, G., Wefer, G. (Eds.), Oppo, D.W., Lehmann, S.J., 1993. Mid-depth circulation of
Use of Proxies in Paleoceanography: Examples from the the subpolar North Atlantic during the Last Glacial Max-
South Atlantic. Springer, Berlin, pp. 91^115. imum. Science 259, 1148^1152.
Kinkel, H., Dittert, N., R. Henrich, 2004. Calcareous plankton Rahmstorf, S., 1997. Risk of sea-change in the Atlantic. Na-
record of the Equatorial Atlantic: A 300 kyr record of cli- ture 388, 825^826.
mate feedback, productivity and dissolution. Palaeogeogr. Reid, J.R., 1996. On the circulation of the South Atlantic
Palaeoclimatol. Palaeoecol., submitted. Ocean. In: Wefer, G., Berger, W.H., Siedler, G., Webb,
Labeyrie, L., Labracherie, M., Gorfti, N., Pichon, J.J., Vautra- D.J. (Eds.), The South Atlantic: Present and Past Circula-
vers, M., Arnold, M., Duplessy, J.C., Paterne, M., Michel, tion. Springer, Berlin, pp. 13^44.
E., Duprat, J., Caralp, M., Turon, J.L., 1996. Hydrographic Rutberg, R.L., Hemming, S.R., Goldstein, S.L., 2000. Re-
changes of the Southern Ocean (southeast Indian sector) duced North Atlantic Deep Water £ux to the glacial South-
over the last 230 kyr. Paleoceanography 11, 57^76. ern Ocean inferred from neodymium isotope ratios. Nature
Lea, D.W., Boyle, E.A., 1990. Foraminiferal reconstruction of 2000, 935^938.
barium distributions in water masses of the glacial ocean. Sarnthein, M., Winn, K., Jung, S.J.A., Duplessy, J.C., Labeyr-
Paleoceanography 5, 719^742. ie, L.D., Erlenkeuser, H., Ganssen, G., 1994. Changes in
Lynch-Stieglitz, J., Curry, W.B., Slowey, N., 1999. Weaker East Atlantic deepwater circulation over the last 30,000
Gulf Stream in the Florida Straits during the Last Glacial years: Eight time slice reconstructions. Paleoceanography
Maximum. Nature 402, 644^647. 9, 209^267.
Marchal, O., Francois, R., Stocker, T., Joos, F., 2000. Ocean Sarnthein, M., Stattegger, K., Dreger, D., Erlenkeuser, H.,
thermohaline circulation and sedimentary 231 Pa/230 Th ratio. Grootes, P., Haupt, B.J., Jung, S., Kiefer, T., Kuhnt, W.,
Paleoceanography 15, 625^641. P£aumann, U., Scha«rfer-Neth, C., Schulz, H., Schulz, M.,
Marchitto, T.M., Curry, W.B., Oppo, D.W., 1998. Millennial- Seidov, D., Simstich, J., van Kreveld, S., Vogelsang, E.,
scale changes in North Atlantic circulation since the last Vo«ker, A. and Weinelt, M., 2001. Fundamental modes and
glaciation. Nature 393, 557^561. abrupt changes in North Atlantic circulation and climate
Martin, P.A., Lea, D.W., 1998. Comparison of water mass over the last 60 ky ^ Concepts, reconstruction and numerical
changes in the deep tropical Atlantic derived from Cd/Ca modelling. In: Scha«fer, P., Ritzrau, W., Schlu«ter, M.,
nd carbon isotope records: Implications for changing Ba Thiede, J. (Eds.), The Northern North Atlantic: A Changing
composition of deep Atlantic water masses. Paleoceanogra- Environment. Springer, Berlin, pp. 365^410.
phy 13, 572^585. Scha«fer-Neth, C., Paul, A., 2001. Circulation of the glacial
Mercier, H., Speer, K.G., Honnorez, J., 1994. Flow pathways Atlantic: A synthesis of global and regional modelling. In:
of bottom water through the Romanche and Chain Fracture Schafer, P., Ritzrau, W., Schlu«ter, M., Thiede, J. (Eds.), The
Zones. Deep-Sea Res. 41, 1457^1477. Northern North Atlantic. Springer, Berlin, pp. 441^462.
Milliman, J.D., 1993. Production and accumulation of calcium Schmittner, A., Meissner, K.J., Eby, M., Weaver, A.J., 2002.
carbonate in the ocean: Budget of a nonsteady state. Glob. Forcing of the deep ocean circulation in simulations of the
Biogeochem. Cycles 7, 927^957. Last Glacial Maximum. Paleoceanography 17, 1029.
Milliman, J.D., Troy, P.J., Balch, W.M., Adams, A.K., Li, Severinghaus, J.P., Sowers, T., Brook, E.J., Alles, R.B., Bend-
Y.-H., Mackenzie, F.T., 1999. Biologically mediated disso- er, M.L., 1998. Timing of abrupt climate change at the end
lution of calcium carbonate above the chemical lysocline? of the Younger Dryas interval from thermally fractionated
Deep-Sea Res. I 46, 1653^1669. gases in polar ice. Nature 391, 141^146.
Mix, A.C., Bard, E., Schneider, R., 2001. Environmental pro- Shannon, L.V., Chapman, P., 1991. Evidence of Antarctic

MARGO 3444 4-2-04

 A.N.A. Volbers, R. Henrich / Marine Geology 204 (2004) 43^57 57

Bottom Water in the Angola Basin at 32‡S. Deep-Sea Res. Indicators: Production, Preservation and Reconstruction of
38, 1299^1304. Upwelling Intensity. Implications from Late Quaternary
Siedler, G., Mu«ller, T.J., Onken, R., Arhan, M., Mercier, H., South Atlantic Sediments. Ph.D. Thesis, Ber. Fachbereich
King, B.A., Saunders, P.M., 1996. The Zonal WOCE Sec- Geowiss. Univ. Bremen 184, 114 pp.
tions in the South Atlantic. In: Wefer, G., Berger, W.H., Volbers, A.N.A., Henrich, R., 2002a. Late Quaternary varia-
Siedler, Webb, D.J. (Eds.), The South Atlantic: Present and tions in calcium carbonate preservation of deep-sea sedi-
Past Circulation. Springer, Berlin, pp. 83^104. ments in the northern Cape Basin: Results from a multi-
Sigman, D.M., Boyle, E.A., 2000. Glacial/interglacial varia- proxy approach. Mar. Geol. 180, 203^220.
tions in atmospheric carbon dioxide. Nature 407, 859^869. Volbers, A.N.A., Henrich, R., 2002b. Present water mass cal-
Stocker, T.F., 1998. A glimpse of the glacial. Nature 391, 338^ cium carbonate corrosiveness in the eastern South Atlantic
339. inferred from ultrastructural breakdown of Globigerina bul-
Talley, L.D., 1996. Antarctic Intermediate Water in the South loides in surface sediments. Mar. Geol. 186, 471^486.
Atlantic. In: Wefer, G., Berger, W.H., Siedler, G., Webb, Warren, B., Speer, K.G., 1991. Deep circulation in the eastern
D.J. (Eds.) The South Atlantic: Present and Past Circula- South Atlantic ocean. Deep-Sea Res. 38, 281^322.
tion. Springer, Berlin, pp. 219^238. Weaver, A.J., Aura, S.M., Myers, P.G., 1994. Interdecadal
Van Bennekom, A.J., Berger, G.W., 1984. Hydrography and variability in an idealised model of the North Atlantic.
silica budget of the Angola Basin. Neth. J. Sea Res. 17, 149^ J. Geophys. Res. 99, 12423^12442.
200. Yu, E.-J., Francois, R., Bacon, M.P., 1996. Similar rates of
Venz, K.A., Hodell, D.A., Stanton, C., Warnke, D.A., 1999. A modern and last glacial ocean thermohaline ciculation in-
1.0 Myr record of Glacial North Atlantic Intermediate ferred from radiochemical data. Nature 379, 689^694.
Water variability from ODP site 982 in the northeast Atlan- Zahn, R., 1992. Deep ocean circulation puzzle. Nature 356,
tic. Paleoceanograpy 14, 42^52. 744^746.
Volbers, A., 2001. Planktic Foraminifera as Paleoceanographic

MARGO 3444 4-2-04