A 450-kyr record of hydrological conditions on the western Agulhas Bank Slope, south of Africa

Marine Geology 180 (2002) 183^201 www.elsevier.com/locate/margeo

A 450-kyr record of hydrological conditions on the western Agulhas Bank Slope, south of Africa A.J. Rau a; *, J. Rogers a , J.R.E. Lutjeharms b , J. Giraudeau c , J.A. Lee-Thorp d , M.-T. Chen e , C. Waelbroeck f a

c

Department of Geological Sciences, University of Cape Town, Private Bag, Rondebosch 7701, South Africa b Department of Oceanography, University of Cape Town, Private Bag, Rondebosch 7701, South Africa De¨partement Ge¨ologie et Oce¨anographie, UMR 5805 CNRS, Universite¨ Bordeaux I, Avenue des Faculte¨s, 33405 Talence, France d Department of Archaeology, University of Cape Town, Private Bag, Rondebosch 7701, South Africa e Institute of Applied Geophysics, National Taiwan Ocean University, Keelung 20224, Taiwan f Laboratoire des Sciences du Climat et de l'Environnement, Domaine du CNRS, baªt. 12, 91198 Gif-sur-Yvette, France Received 1 July 2000; received in revised form 14 March 2001; accepted 25 May 2001

Abstract The Agulhas Bank region, south of Africa, is an oceanographically important and complex area. The leakage of warm saline Indian Ocean water into the South Atlantic around the southern tip of Africa is a crucial factor in the global thermohaline circulation. Foraminiferal assemblage, stable isotope and sedimentological data from the top 10 m of core MD962080, recovered from the western Agulhas Bank Slope, are used to indicate changes in water mass circulation in the southeastern South Atlantic for the last 450 kyr. Sedimentological and planktonic foraminiferal data give clear signals of cold water intrusions. The benthic stable isotope record provides the stratigraphic framework and indicates that the last four climatic cycles are represented (i.e. down to marine isotope stage (MIS) 12). The planktonic foraminiferal assemblages bear a clear transitional to subantarctic character with Globorotalia inflata and Neogloboquadrina pachyderma (dextral) being the dominant taxa. Input of cold, subantarctic waters into the region by means of leakage through the Subtropical Convergence, as part of Agulhas ring shedding, and a general cooling of surface waters is suggested by increased occurrence of the subantarctic assemblage during glacial periods. Variable input of Indian Ocean waters via the Agulhas Current is indicated by the presence of tropical/subtropical planktonic foraminiferal species Globoquadrina dutertrei, Globigerinoides ruber (alba) and Globorotalia menardii with maximum leakage occurring at glacial terminations. The continuous presence of G. menardii throughout the core suggests that the exchange of water from the South Indian Ocean to the South Atlantic Ocean was never entirely obstructed in the last 450 kyr. The benthic carbon isotope record and sediment textural data reflect a change in bottom water masses over the core location from North Atlantic Deep Water to Upper Southern Component Water. Planktonic foraminiferal assemblages and sediment composition indicate a profound change in surface water conditions over the core site approximately 200^250 kyr BP, during MIS 7, from mixed subantarctic and transitional water masses to overall warmer surface water conditions. ß 2002 Elsevier Science B.V. All rights reserved. Keywords: Indian^Atlantic exchange; Agulhas Current; Subtropical Convergence; subantarctic water; planktonic foraminifera; `Cape valve'

* Corresponding author. Fax: +27-21-650-3783.

E-mail address: [email protected] (A.J. Rau).

0025-3227 / 02 / $ ^ see front matter ß 2002 Elsevier Science B.V. All rights reserved. PII: S 0 0 2 5 - 3 2 2 7 ( 0 1 ) 0 0 2 1 3 - 4

MARGO 2974 25-2-02

184

A.J. Rau et al. / Marine Geology 180 (2002) 183^201

1. Introduction It has previously been demonstrated that the amount of water that escapes westward from the southern Indian Ocean has a direct e¡ect on the overturning of the whole Atlantic Ocean by the eventual increase in generation of North Atlantic Deep Water (NADW) (Weijer et al., 1999). Some researchers (e.g. Berger and Wefer, 1996) have suggested the possibility that the in£ux of warm saline Indian Ocean water into the South Atlantic

around the tip of Africa (referred to as the `Cape valve') was shut o¡ during glacial periods. Conceivably this inter-ocean exchange south of Africa could be shut o¡ by a substantial equatorward displacement of the Subtropical Convergence (De Ruijter, 1982) and/or changes in the inertia of the Agulhas Current (Matano et al., 1998). This possibility has recently been explored by Flores et al. (1999) with a core located south of Cape Town (35³49.2PS, 18³05.4PE), to the northwest of our core. Their low resolution study,

Fig. 1. A conceptual image of the currents that may a¡ect water masses over core MD962080 at present. A crossed circle shows the location of the core on the western slope of the Agulhas Bank. The region is dominated by the Agulhas Current (1) that follows the shelf edge closely, except when its trajectory is disturbed by a Natal Pulse (2). A well-developed pulse may cause an upstream leakage of water east of the Agulhas Plateau. South of the Agulhas Bank, shear edge plumes are formed on the landward border of the current. Some of these may be carried equatorward as Agulhas ¢laments (3), over the core site. At the Agulhas retro£ection (4) Agulhas rings are formed (5) that drift o¡ into the South Atlantic (6), partially carried by the Benguela Current. The rest of the water of the Agulhas Current £ows eastward along the STC as the Agulhas Return Current (7), shedding warm (8) and cold eddies (9). The former may become imbedded in the Antarctic Circumpolar Current. Wind-driven, coastal upwelling is shown in the Cape Town vicinity. BUS = Benguela Upwelling System, AFR = Agulhas Fracture Ridge.

MARGO 2974 25-2-02

A.J. Rau et al. / Marine Geology 180 (2002) 183^201

which focused on calcareous phytoplankton indicators (coccolithophorids), suggests the presence of a well-de¢ned glacial^interglacial cyclicity in the Agulhas Current retro£ection area in response to £uctuations in the Subtropical Convergence (STC). The location of the core discussed here (MD962080, 36³19PS, 19³28PE, Fig. 1) is particularly promising from a palaeoceanographic point of view for a number of reasons. First, it may be an appropriate site for monitoring the inter-ocean exchange of water from the Indian to the Atlantic Ocean (Lutjeharms and Cooper, 1996; De Ruijter et al., 1999), a crucial component of the climatically important global thermohaline circulation (Gordon, 1986). Second, this site may give indications on the temporal behaviour of water masses that move over the Agulhas Bank, the wide continental shelf south of Africa (Fig. 1). There is evidence that water derived from the Agulhas Current may move over this shelf region into the South Atlantic (Gordon et al., 1987), thus contributing to the inter-ocean leakage. There are also suggestions that the bottom water of the Agulhas Bank moves westward (Lutjeharms and Meyer, personal communication) and may cascade downward over the western slope. A core situated on this slope may therefore give a good indication of palaeoceanographic changes at this coastal margin. Recent studies (Giraudeau, 1993; Giraudeau et al., 2001; Niebler and Gersonde, 1998) document the reliability of planktonic foraminifera as tracers of present surface water masses in the Southeast Atlantic. The high abundance and good preservation of these calcareous microfossils in core MD962080 suggest that variability in species abundance allows reconstruction of the varying dominance of di¡erent surface water masses over the Agulhas Bank Slope. Hodell (1993) and Howard and Prell (1992) used foraminiferal data from the Atlantic and Indian sectors of the Southern Ocean, respectively, to interpret £uctuations in the Antarctic Polar Front Zone (APF) and the STC during the late Pleistocene. In this paper we examine marine sediments on the continental slope of the western Agulhas Bank to investigate the history of water column dynam-

185

ics in this critical region. We present data on planktonic foraminiferal assemblages, benthic foraminiferal stable isotope records and sediment composition and texture in order to examine the sensitivity of the system to global and regional climatic changes during the last 450 kyr. Our results show that there is a major hydrographic change around 200^250 kyr BP to overall warmer surface water conditions and clearly de¢ned glacial^interglacial cyclicity. A change in bottom water mass at the core location from NADW to Upper Southern Component Water (USCW) is re£ected in isotopic and textural records. We provide evidence that inter-ocean leakage was never entirely obstructed during the last 450 kyr. 2. Hydrographic setting Because it lies at the crossroads of the waters from a number of adjacent ocean basins, the oceanic region south of Africa is a veritable melange of water types. The major contributions of water come from the South Indian Ocean, mostly via the Agulhas Current ; from the South Atlantic Ocean, via the South Atlantic Current and the broader Benguela Current drift; and from occasional intrusions of subantarctic water from south of the STC (Fig. 1). For the purpose of this particular core, the water masses at the surface and at intermediate depths are probably of most importance. The Agulhas Current (Fig. 1) carries South Indian Subtropical Surface Water, Antarctic Intermediate Water and some remnants of Red Sea Water and Tropical Indian Surface Water into the Agulhas retro£ection region (Valentine et al., 1993). At the retro£ection, the Agulhas Current turns back on itself (Lutjeharms and van Ballegooyen, 1988a) and most of its waters subsequently £ow eastward as the Agulhas Return Current. During this process, the retro£ection loop can occlude to pinch o¡ an Agulhas ring that will subsequently drift o¡ into the South Atlantic Ocean, carrying its load of South Indian water masses with it (Fig. 1). This is the major mechanism for inter-ocean exchange south of Africa. The precise location of the retro£ection is variable

MARGO 2974 25-2-02

186

A.J. Rau et al. / Marine Geology 180 (2002) 183^201

and may, under extreme conditions, even lie far upstream (Lutjeharms and van Ballegooyen, 1988b), thus preventing this Indian Ocean water from even reaching the region south of Africa and contributing to inter-ocean exchange. The southern boundary of the Benguela Upwelling System can be considered as the Agulhas retro£ection area (Shannon and Nelson, 1996). This de¢nition is preferred to a de¢nition based on the extent of coastal upwelling, as Schumann et al. (1982) and others have shown that upwelling extends along much of the south coast of South Africa, certainly as far as 25³E. To the south, the large-scale circulation of the region is constrained by the STC. This front lies at about 42³S Lutjeharms, 1985) with a central temperature at the sea surface of 14³C. It forms the generic border to the Southern Ocean and its Antarctic Circumpolar Current (Fig. 1). It is characterised by extensive meridional meanders and a range of eddies (Lutjeharms and Valentine, 1988a), making it a most complex feature. Such complexity also distinguishes the location of our core for a number of other reasons. Based on investigations spanning a period of 10 years, it was shown that Agulhas rings may be present at the location of the core about 12% of the time (Lutjeharms and Valentine, 1988b). Agulhas ¢laments, tendrils of warm surface water from that current (Lutjeharms and Cooper, 1996), also move across this location since their preferred trajectories lie at the shelf edge (Fig. 1). Recent results have shown that such northward movement at this shelf edge may be interspersed by southward movement due to the presence of cyclonic (clockwise) eddies (Penven et al., 2001). These eddies (Fig. 1) are thought to be lee eddies, driven by the Agulhas Current moving past the eastern side of the Agulhas Bank. Agulhas rings, Agulhas ¢laments and lee eddies are not the only complicating factors hindering the understanding of water mass movements in the region. During the formation of an Agulhas ring, northward penetration of Subantarctic Surface Water from south of the STC has been observed between the Agulhas retro£ection and the newly formed ring (Lutjeharms and van Ballegooyen, 1988b). These intrusions (Fig. 1) may carry this

cold water over the core location (Shannon et al., 1989), particularly if a cyclonic lee eddy is present. With a reduced frequency of ring formation, the incidence of penetrations of subantarctic water into the region could also be decreased (Lutjeharms et al., 2001). What climatic changes could bring about reduced ring shedding is still largely a matter of speculation (Lutjeharms and de Ruijter, 1996). The core location (2488 m water depth) is presently bathed by oxygen-rich, nutrient-poor NADW that extends from 1700 m to 3600 m (Reid, 1989; Bainbridge, 1981) and is bounded (above and below) by southern component nutrient-rich waters (Upper and Lower Circumpolar Deep Water). Intermediate waters at the core location are derived from a complex combination of Indian, Atlantic and Southern Ocean sources (Shannon and Nelson, 1996). 3. Methods 3.1. Sediments Core MD962080 (Fig. 1) was retrieved from the western Agulhas Bank Slope (36³19PS, 19³28PE, 2488 m) during the IMAGES II-NAUSICAA cruise of the Marion Dufresne (Bertrand et al., 1997). The top 10 m of the core were sampled at 10-cm intervals. Each sample was subdivided into three sub-samples for foraminiferal census counts, isotope measurements and sedimentological analysis. Planktonic foraminifera were concentrated by de£occulating the foraminiferal sub-samples in a 3 M solution of sodium hexametaphosphate, prior to washing through a 125-Wm sieve. This size fraction was chosen to facilitate comparison with previously published planktonic foraminiferal assemblage data from surface sediments and sediment cores taken in the Southeast Atlantic Ocean (Giraudeau, 1993; Giraudeau et al., 2001; Little et al., 1997; Niebler and Gersonde, 1998). The s 125-Wm fractions were oven-dried overnight, then examined under a binocular microscope. On average, 250 specimens per sample were identi¢ed to species level following the classical taxon-

MARGO 2974 25-2-02

A.J. Rau et al. / Marine Geology 180 (2002) 183^201

omy of Be¨ (1967) and Parker (1962). Globorotalia menardii ss and Globorotalia tumida were combined into a single category Globorotalia menardii. The scarcity, but relative importance of this species as a potential tracer of warm tropical/subtropical Indian Ocean waters (e.g. Berger and Wefer, 1996) prompted a separate count on the whole s 125-Wm fraction. We subsequently expressed G. menardii abundance as accumulation rate (number of specimens/cm2 /kyr) by multiplying concentration by the bulk mass accumulation rate (BMAR). The raw counts for all other species were converted into percentages of the total number of specimens counted. Percent calcium carbonate was calculated by

187

measuring the mass di¡erence after the sub-samples were treated with 1 M HCl. Each sedimentological sub-sample was weighed to obtain the total mass, before the coarse fraction was separated by wet-sieving sub-samples through a 63-Wm sieve. The percentages of `total mud' and `total sand' were calculated by mass di¡erence. The acid-insoluble sand was washed on a 63-Wm sieve. Siliceous sponge spicules and radiolaria were £oated o¡ using a sodium polytungstate solution, with a speci¢c gravity of 2.5, in a separating funnel. The `heavy fraction' (quartz and glauconite) was separated, washed, oven-dried and weighed, the procedure being repeated for the `light fraction' (sponge spicules and

Fig. 2. (a) Oxygen isotope (N18 O) depth pro¢le for the benthic foraminifer Cibicidoides wuellerstor¢ which provided the stratigraphic framework. Age control points are shown to the right of the N18 O pro¢les. (b) Sedimentation rate showing depth in core versus calculated age.

MARGO 2974 25-2-02

188

A.J. Rau et al. / Marine Geology 180 (2002) 183^201

radiolaria). Fe-bearing glauconite was separated using a Franz isodynamic magnetic separator. The compositional data of the sand fraction are reported as percentages of the s 63-Wm fraction.

N 18 O ˆ ‰…18 O=16 O†sample =…18 O=16 O†standard 31Š per mil N 13 C ˆ ‰…13 C=12 C†sample =…13 C=12 C†standard 31Š per mil

3.2. Stable isotope analysis Specimens of the benthic foraminifer Cibicidoides wuellerstor¢ were picked from the whole s 125-Wm fraction. Carbon and oxygen isotopic measurements were performed at the Laboratoire des Sciences du Climat et de l'Environnement, Gif-sur-Yvette, France on a Finnigan MAT251 mass spectrometer. The mean external reproducibility of the measurements is 0.05x. Oxygen data are calibrated with respect to National Bureau of Standards NBS19 (Hut, 1987; Coplen, 1988). As per convention, isotope ratios are reported relative to the international standard, Vienna Pee Dee Belemnite (PDB), using the N notation:

3.3. Stratigraphy Stratigraphy for core MD962080 is based on the oxygen isotope record of the benthic foraminifer Cibicidoides wuellerstor¢. The isotope curve was compared to the SPECMAP stack (Imbrie et al., 1984) and a curve generated by Bassinot et al. (1994a). Signi¢cant peaks were correlated and provided the age control points for the time framework. Minima and maxima in N18 O were chosen rather than isotopic stage boundaries. Eighteen stable isotope events were identi¢ed within the top 10 m of the core (Fig. 2a). Sedimentation rates were calculated between those

Fig. 3. (a) Calcium carbonate content, (b) sediment texture and (c) BMAR. Horizontal grey-shaded bars represent inferred glacial MIS.

MARGO 2974 25-2-02

A.J. Rau et al. / Marine Geology 180 (2002) 183^201

time points and ages were then linearly interpolated to achieve a ¢nal age model. Sedimentation rates range between 1 and 5.5 cm/kyr with an average of 2.7 cm/kyr (Fig. 2b). We used the ¢nal age model to extrapolate the bulk mass accumulation rate (BMAR) according to the formula : BMAR ˆ SRUDBD where SR (cm/kyr) = sedimentation rate; DBD (g/ cm3 ) = dry bulk density = 2.65U(GRAPE31)/ (2.6531); GRAPE (g/cm3† = Q-ray attenuation.

4. Results 4.1. Sedimentology The core consists of interbedded grey to light grey foram-bearing ooze to foram nanno ooze and light yellowish-brown foram ooze (Bertrand et al., 1997; Chen et al., 1998; Rogers, 1999). Fig. 3 illustrates the carbonate content and textural characteristics of core MD962080. The total carbonate content of the upper 10 m of the core (Fig. 3a) varies between 62 and 85%. No clear glacial^interglacial trend is observed.

189

The texture is ¢nest (96% mud) in interglacial marine isotope stage (MIS) 11 (Fig. 3b). A clear glacial^interglacial cyclicity develops from interglacial MIS 7, with higher sand content during glacial intervals. In addition, a general trend towards coarser textures is evident up to Termination III (MIS 8/7). Coarser textures are best developed in glacial MIS 6, but are also evident in glacial MIS 12, 10, 8 and 2 (Fig. 3b). The generally good preservation of foraminiferal tests (the major component of the sand fraction) in intervals representative of glacial periods suggests that dissolution is unlikely to have had a major e¡ect on cyclical glacial^interglacial variations in texture. A potential driving mechanism for this textural variability is winnowing due to varying intensity of bottom currents. Wu and Berger (1991) showed that winnowing of ¢ne particles is important at least down to a depth of 2500 m (roughly the depth of this core). However, the BMAR record (Fig. 3c) shows that sediment accumulated at much higher rates during glacial intervals than during interglacials. The combination of high BMAR and coarser sediment instead suggests that enhanced productivity of planktonic foraminifera is the major driving force behind the observed glacial^interglacial textural variability. The sand fraction (Fig. 4) is dominated by

Fig. 4. Sedimentological composition of the sand fraction. Horizontal grey-shaded bars represent inferred glacial MIS.

MARGO 2974 25-2-02

190

A.J. Rau et al. / Marine Geology 180 (2002) 183^201

Table 1 Summary statistics for planktonic foraminiferal abundance counts Species

Minimum

Maximum

Mean

Neogloboquadrina pachyderma (s) Neogloboquadrina pachyderma (d) Turborotalia quinqueloba Globigerina bulloides Globorotalia in£ata Globorotalia truncatulinoides Globigerinita glutinata Globorotalia hirsuta Globorotalia crassaformis Globorotalia scitula Globigerina falconensis Globoquadrina dutertrei Globigerinoides sacculifer Globigerinella calida Globigerinoides ruber (alba) Globorotalia menardii

0 8.53 0 2.63 17.59 0 3.53 0 0 0.57 0.76 0 1.45 0 0 0

5.63 33.44 5.36 17.20 40.50 4.59 18.11 2.58 2.50 8.98 13.51 6.76 12.15 6.76 5.50 1.48

1.03 20.17 1.76 8.93 26.53 1.16 9.18 0.37 0.56 3.59 6.13 2.58 6.28 2.41 2.00 0.16

planktonic foraminifera (mean 86%) with subordinate amounts of terrigenous quartz (mean 11%) and minor amounts of authigenic glauconite (mean 0.6%). The last two components vary in sympathy with each other and £uctuate on glacial and interglacial time scales. Benthic sponge spicules and planktonic radiolaria are minor components (mean 1.6%). There is a change in mean trend of both the calcium carbonate and quartz contents of the sand fraction, occurring around 250 kyr BP. Calcium carbonate exhibits a general increase upcore and quartz content a general decrease upcore, until MIS 8/7. The amplitude of the £uctuations in both components is more constrained after this period. The two records have opposing glacial^interglacial cyclicity. 4.2. Foraminiferal assemblages A total of 16 taxa were identi¢ed in the s 125Wm fraction. Summary statistics are given in Table 1. Raw census counts are available on request from the ¢rst author. The foraminiferal population is dominated by two species: Globorotalia in£ata and Neogloboquadrina pachyderma (dextrally coiled). The subordinate taxa (mean s 1%), in order of decreasing average abundance are : Globigerinita glutinata, Globigerina bulloides, Globigerinoides sac-

culifer, Globigerina falconensis, Globorotalia scitula, Globoquadrina dutertrei, Globigerinella calida, Globigerinoides ruber (alba), Turborotalia quinqueloba, Globorotalia truncatulinoides and Neogloboquadrina pachyderma (sinistrally coiled). Individual species abundance variations exhibit high frequency erratic signals, probably re£ecting the complexity of the surface hydrological conditions over the Agulhas Bank Slope. We therefore performed Q-mode factor analysis (Imbrie and Kipp, 1971) in order to extract major groupings and trends not obvious from the individual species records. Factor analysis indicates that four faunal components account for more than 98% of the total variance in the census data. The factor scores which give the species composition of each assemblage are shown in Table 2. The varimax factor components, which give the composition of each sample in terms of the resultant assemblage, are plotted in Fig. 5. Each factor can be characterised by a dominant taxon and one or two subordinate taxa. Factor 1 (31.3% of the total variance) is de¢ned by Globorotalia in£ata. Globigerinoides sacculifer, Globigerina falconensis and Globigerinita glutinata characterise factor 2 (25.7% of the total variance). Factor 3 is characterised by Neogloboquadrina pachyderma (d) and Globigerina bulloides (25.6%). Factor 4 contributes only 16% to the total variance and is exclu-

MARGO 2974 25-2-02

A.J. Rau et al. / Marine Geology 180 (2002) 183^201

191

Table 2 Varimax factor scores derived from factor analysis of the planktonic foraminiferal census data Variable

Factor 1

Factor 2

Factor 3

Factor 4

Neogloboquadrina pachyderma (s) Neogloboquadrina pachyderma (d) Globigerina bulloides Globorotalia in£ata Globigerinoides ruber (alba) Globoquadrina dutertrei Globorotalia menardii Globorotalia scitula Globigerinita glutinata Globigerina falconensis Globorotalia hirsuta Globigerinoides sacculifer Globorotalia crassaformis Globigerinella calida Turborotalia quinqueloba Globorotalia truncatulinoides

0.018 30.064 0.069 0.954 0.035 0.026 30.001 30.001 30.187 30.169 0.001 30.107 0.023 30.057 30.025 0.016

30.066 30.116 0.256 0.229 0.116 0.09 0.01 0.159 0.621 0.473 0.025 0.434 0 0.093 0.064 0.08

0.074 0.725 0.638 30.011 0.033 30.01 0 30.069 30.085 0.12 30.009 30.172 0.005 0.058 0.024 30.006

0.03 0.66 30.647 0.155 30.1 0.019 0 0.106 0.222 30.057 30.003 0.2 30.004 0.042 0.055 30.043

Fig. 5. Planktonic foraminiferal assemblages as de¢ned by factor analysis. (a) The Transitional Assemblage is de¢ned by Globorotalia in£ata. (b) The Cosmopolitan Assemblage comprises Globigerinoides sacculifer, Globigerinita glutinata and Globigerina falconensis. (c) Neogloboquadrina pachyderma (dextral) and Globigerina bulloides de¢ne the Northern Subantarctic Assemblage. The near-monospeci¢c Subantarctic Assemblage is dominated by Neogloboquadrina pachyderma (dextral). (e) Summed percentages of the warm water species Globoquadrina dutertrei, Globigerinoides ruber (alba) and Globorotalia menardii make up the Tropical/Subtropical Assemblage. For details see text. Horizontal grey-shaded bars represent inferred glacial MIS.

MARGO 2974 25-2-02

192

A.J. Rau et al. / Marine Geology 180 (2002) 183^201

sively de¢ned by Neogloboquadrina pachyderma (d). For the following discussion, Be¨ and Tonderlund (1971), Be¨ and Hutson (1977) and Hutson (1980) are used as the general references for the environmental characterisation of the foraminiferal species. Globorotalia in£ata is the most abundant species of the transitional zone between subantarctic and subtropical waters. The present-day foraminiferal assemblage dominated by G. in£ata occurs north of the STC. The Transitional Assemblage (factor 1) £uctuates independently of global glacial^interglacial cycles (Fig. 5a). Variations in abundance occur on shorter time scales in response to local hydrographic changes. There is a change in mean contribution of this factor to the total foraminiferal population from relatively higher values prior to ca. 200 kyr to relatively lower values in the younger portion of the core. Factor 2 bears a cosmopolitan signature. Globigerinoides sacculifer is an abundant species in subtropical waters. Globigerinita glutinata is one of the most widespread species, occurring over a wide range of temperatures and salinity. Globigerina falconensis is the warm water variety of Globigerina bulloides. This assemblage shows the opposite trend to the Transitional Assemblage with lower relative contributions prior to 200 kyr and higher relative contributions thereafter (Fig. 5b). A glacial^interglacial cyclicity develops from MIS 7 upcore with higher values during warm periods. Factor 3 is de¢ned by the cold water foraminifera Neogloboquadrina pachyderma (d) and Globigerina bulloides. The two species are weighted equally (Table 2). Although these two species are present in o¡-shelf waters of the southern Benguela System, the location of the core makes it unlikely that their occurrence is a result of coastal upwelling, but is rather due to the presence of Southern Ocean waters. This speci¢c association is indicative of cool (10^18³C) northern subpolar waters in the vicinity of the STC. This assemblage has increased contributions during glacial periods over the last 200 kyr. Prior to this, contributions are constantly high (Fig. 5c). Factor 4 is a near-monospeci¢c assemblage, dominated by Neogloboquadrina pachyderma (d).

This is the cold end member of the foraminiferal population and re£ects cold subantarctic waters. This assemblage contributes the least to the overall variance, but bears a unique signal. Firstly, there is no change in overall contribution at 200 kyr BP as seen in the ¢rst three factors. Secondly, a clear glacial^interglacial cyclicity is expressed throughout the core, with higher values characterising cold periods (Fig. 5d). Although minor in the overall foraminiferal population in the core (thus not identi¢ed in the factor analysis), the subtropical and tropical species Globigerinoides ruber (alba), Globorotalia menardii and Globoquadrina dutertrei need to be considered as potential indicators of warm surface waters over the core site. We therefore summed their relative abundances to de¢ne a Tropical/ Subtropical Assemblage. Globoquadrina dutertrei is associated with boundary currents with highest percentages occurring in the sediments of equatorial regions and along continental margins. Globigerinoides ruber is a widespread species dominating in subtropical and tropical waters. Its dominance appears to increase in the oligotrophic Indian and Atlantic Central Waters. The pinkshelled variety is absent from the Indian Ocean today. The counts reported here refer to the white-shelled variety G. ruber (alba). Although only a minor component of the foraminiferal population, G. menardii is an important indicator species of tropical Indian Ocean waters (see Section 4.3). This Tropical/Subtropical Assemblage exhibits a clear glacial^interglacial cyclicity after 200 kyr BP (Fig. 5e) with minimum abundances occurring during glacial periods MIS 2, 4 and 6 and peak abundances in warm periods. No such cyclicity is present in the older portion of the core ( s 200 kyr BP), which displays, on average, lower abundance than in the younger interval. 4.3. Globorotalia menardii Berger and Wefer (1996) used the occurrence and disappearance of this species as an indicator of inter-ocean heat exchange around the southern tip of Africa. They proposed that Indian to Atlantic Ocean leakage of surface waters is a plausible mechanism for reseeding this taxon in the

MARGO 2974 25-2-02

A.J. Rau et al. / Marine Geology 180 (2002) 183^201

193

Fig. 6. (a) Absolute abundance and (b) accumulation rate of the tropical foraminifer Globorotalia menardii. Horizontal greyshaded bars represent inferred glacial MIS.

tropical Atlantic during warm periods (Ericson and Wollin, 1968). Although the relative contribution of Globorotalia menardii to the foraminiferal assemblage is negligible ( 6 2%), a separate count of absolute abundance reveals that this species is present throughout the length of core studied (Fig. 6). The pattern of accumulation rate of G. menardii (Fig. 6b) is dramatically di¡erent from the record of summed tropical/subtropical species (Fig. 5e). Accumulation peaks are centred on the ¢nal phase of glacial intervals, such as before Terminations II, III, and IV. The rather inconsistent pattern seen at MIS 2 (peak accumulation within MIS 2) might be a result of the poorly constrained stratigraphy toward the top of the core. 5. Discussion Our results suggest that a complex combination of hydrological changes occurred over the Agul-

has Bank slope at various intervals throughout the last 450 kyr: besides a long-term, linear evolution, as seen, for instance, in the trend of the sediment texture, we provide evidence for successive short-term events at glacial terminations, a glacial^interglacial cyclicity, and the occurrence of a major shift in surface hydrology at 200^250 kyr, which is mostly expressed in the foraminiferal record. This pattern is obviously related to the particular setting of the core studied, being at the crossroads of water masses from a number of adjacent ocean basins, which makes this location most sensitive to local (Agulhas water leakages), regional (dynamics of the STC and associated water masses) and global changes. 5.1. Sediment texture, productivity and bottom water chemistry As stated previously, the changes in productivity and burial rates of planktonic foraminiferal at our studied location are likely to be the main

MARGO 2974 25-2-02

194

A.J. Rau et al. / Marine Geology 180 (2002) 183^201

Fig. 7. (a) Carbon isotope (N13 C) record for the benthic foraminifer Cibicides wuellerstor¢, (b) sediment texture, and (c) relative abundance of Neogloboquadrina pachyderma (d). Horizontal grey-shaded bars represent inferred glacial MIS.

driving mechanism behind the recorded glacial^ interglacial textural variability. This assumption is essentially borne out by the combination of high BMAR and coarser sediment. Winnowing and carbonate dissolution, which are often critical in determining the texture of carbonate-rich sediments over continental margins, conventionally exert either negative feedbacks on both BMAR and sediment coarsening (in the case of dissolution) or an opposing e¡ect on those parameters (as in the case of winnowing). Interestingly, high amplitude changes in the benthic N13 C record are synchronous with major textural changes (Fig. 7a,b), low N13 C values being associated with high sand content. We interpret the excellent phase relationship of the stable isotope and textural record in glacial^interglacial cycles (as well as at shorter time scales) as an additional indication for enhanced productivity and associated burial rate during glacial intervals. Mackensen and Bickert (1999), in a recent review of the various factors a¡ecting the stable carbon isotope signature of epibenthic foramini-

fera in the southern Atlantic, identi¢ed a deviation of the isotope signature from ambient bottom water N13 CgCO2 in various domains of the Southern Ocean a¡ected by high seasonal productivity. This so-called `phytodetritus' e¡ect or `Mackensen' e¡ect (Mackensen et al., 1993) induces a lowering of the benthic N13 C values by 0.2^0.6x relative to the ambient bottom water N13 CgCO2 , a mechanism related to the rapid sedimentation and the degradation of phytodetritus, which produces a N13 C-depleted microenvironment at the sediment^water interface. Glacial to interglacial N13 C amplitude shifts in core MD962080 range from 0.6 to 1.2x, with maximum values at the 6/5, 8/7 and 10/9 transitions (Fig. 7a). The combination of (1) a change in bottom water mass at the site location (2500 m water depth) from NADW to USCW (N13 C of ca. 0.1x; Bickert, 1992) and (2) a global glacial^ interglacial change in the N13 CgCO2 of seawater is not su¤cient to explain the high glacial^interglacial amplitude shifts seen in our benthic N13 C record. An additional 0.2^0.4x lowering of the benthic N13 C during glacial intervals over the last

MARGO 2974 25-2-02

A.J. Rau et al. / Marine Geology 180 (2002) 183^201

450 kyr can be accounted for by an increase in primary productivity. Such highly seasonal productivity is typical of the subantarctic surface waters between the STC and the subantarctic front. We therefore surmise that, during glacial intervals for the time period studied, cold subantarctic water masses bearing a highly seasonal primary productivity reached at least the latitude of core MD962080 (36³S), and induced massive sedimentation of secondary carbonate producers (planktonic foraminifera). Increased abundance of the phytoplankton-grazing and subpolar waterthriving species Neogloboquadrina pachyderma (d) during glacial periods at the studied location (Fig. 7c) further corroborates this interpretation. The long-term trend in evolution of the sand content (coarsening of the sediment texture with decreasing age) seen in core MD962080 is consistent with earlier studies in the Atlantic, Paci¢c, and Indian Oceans (see review by Bickert and Wefer, 1996) and referred to as the Mid-Brunhes dissolution cycle (Bassinot et al., 1994b). As proposed by Peterson and Prell (1985), this dissolution signal might be driven by changes in deep water circulation. The overall trend displayed by the benthic N13 C record in core MD962080 (Fig. 7a) indeed con¢rms this hypothesis by documenting, through increasingly depleted carbon isotope values with age, an increased in£uence of corrosive Southern Component Water at the core location, which was probably associated with shoaling of the carbonate lysocline. The overall good preservation of planktonic foraminifera throughout the core is theoretically not consistent with the existence of such a dissolution cycle. A precise quanti¢cation of foraminiferal fragmentation in the near future might help solve this apparent inconsistency. 5.2. Sources of the terrigenous sand-fraction Quartz sand has been found in signi¢cant percentages in core MD962080, which is located well south of the southern tip of the African landmass at Cape Agulhas (just north of 35³S) in latitude 36³19PS. Four possible sources have been considered.

195

Firstly, under the highstand conditions prevailing during interglacials, a £uvial source of the quartz is unlikely, because the core site then lies far from the coastline. However, during Pleistocene glacials, lowstand conditions would prevail and the southern tip of the continent would shift close to the modern 130 m isobath (Dingle and Rogers, 1972). This would bring the southern tip of Africa close to latitude 36³S, although it would shift southeastwards to 21³E from the Holocene longitude of 20³E. In other words, the southern tip would move eastwards away from the site of core MD92080 on the western slope of the Agulhas Bank. A second possibility, an aeolian source, is hard to justify. No nearby sources exist. The nearest possible source is the Namib Desert, but this lies too far to the north, in both interglacial and glacial periods. A third possibility is ice-rafted detritus (IRD), which Ruddiman (1977), working mainly on cores from the North Atlantic, de¢ned as quartz and lithic sand coarser than 63 Wm, in other words terrigenous sand and gravel. Connolly and Ewing (1965), working on cores from the Southern Ocean, similarly chose 63 Wm as a lower size limit. A recent high resolution data set of IRD in the southern Atlantic (Kanfoush et al., 2000) revealed that ice-rafting events punctuated the last glacial periods and were recorded in bottom sediments up to 41³S, just south of the present location of the STC. Needham (1962) reported pebbles, cobbles and boulders, some faceted and striated, on the slope and rise due west of the Cape Peninsula and concluded that they were ice-rafted. These clasts were found equatorward of core MD962080 by 2 degrees of latitude. The acid-insoluble sand fraction contains trace amounts of basaltic rock fragments. There is no basalt on the continental shelf, nor in the hinterland of the Western Cape, South Africa. However, the main argument for the quartz sand in the cores being IRD is its increase in abundance with the subpolar planktonic foraminifer assemblage dominated by Neogloboquadrina pachyderma (d) during glacial periods (Fig. 7c). For icebergs to reach this environment su¤ciently intact to be carrying detritus, cooler surface water in the Southern

MARGO 2974 25-2-02

196

A.J. Rau et al. / Marine Geology 180 (2002) 183^201

Ocean would seem essential. In order to move rapidly enough northward to reach the coast of Africa ^ even with lower sea surface temperatures ^ increased wind stress would be required. Such consistently higher wind stresses over the Southern Ocean would lead to increased equatorward Ekman drift in the surface layers of this ocean and thus an equatorward shift of the STC as well as increased leakage of subantarctic water across the frontal zone. More cold water intrusions from this front would thus reach the core site. Three possibilities present themselves: the northward shift of the STC, a higher incidence of subantarctic intrusions as part of Agulhas ring shedding or a general cooling of surface water. We favour a combination of all three mechanisms. A fourth possibility for the bulk of the acidinsoluble sand is a shelf origin. Glauconite (Fig. 4c) varies in sympathy with quartz (Fig. 4b) in the sand fraction, albeit as a minor component. The outer shelf of the western Agulhas Bank is rich in glauconite (Birch, 1979) and concentrations of over 70% are found west of the Cape Peninsula, although the concentrations decrease rapidly southwards to the latitude of core MD962080. It is possible that the lowering of sea level during Pleistocene glacials, by simultaneously shifting the wave base closer to the slope, allowed the seaward transport of glauconite and quartz from the outer shelf to the slope during storms. The current data are presently insu¤cient to indisputably dictate whether the recorded lithogenic sand size fraction originates from the Antarctic ice sheet via iceberg transport, or from the nearby Agulhas Bank through downslope transfer of shelf material during low sea level stands. Although previous separate studies (see above) suggest that the most probable source is IRD, additional quantitative and mineralogical analyses of the terrigenous sand fraction in core MD962080 should be initiated, and compared to related investigations, such as those presently being conducted on material collected during ODP Leg 177 in the Atlantic part of the Southern Ocean (Gersonde et al., 1999).

5.3. Foraminiferal assemblages and dynamics of surface waters masses Most previous micropalaeontological studies undertaken on sediment material from the subantarctic Southern Ocean or from the southern border of subtropical gyres (Prell et al., 1979; Howard and Prell, 1992; Flores et al., 1999; among others) indicate that the palaeoceanographic history of this ocean realm is characterised by a series of uniform equatorward and poleward shifts of the belt of hydrological fronts (STC and APF) and associated water masses following a straightforward glacial^interglacial cyclicity. The planktonic foraminiferal data set obtained at core MD962080 suggests that the biotic response to surface hydrological changes near the southern tip of Africa is much more complex than supposed from these earlier studies and involves pulses additional to well-recognised glacial^interglacial cycles. Perhaps the most spectacular hydrological event recorded in the present investigation is a profound change in surface water conditions at 200^250 kyr, within MIS 7. Planktonic foraminiferal assemblages indicate that this shift represents a transition at the core location, from the presence of a mixed northern subantarctic and transitional surface water mass with limited variability on glacial^interglacial cycles during the MIS 12^7 interval, to overall warmer conditions as well as more pronounced glacial^interglacial cyclical £uctuations in the younger part of the record (Fig. 5). Sediment compositional data of the sand fraction also record a change at this time. Calcium carbonate and quartz contents show a distinct change in overall trend, with a cessation of general increasing and decreasing trends respectively (Fig. 4). Numerous locations in the Southern Hemisphere showed a transition to more `interglacial' conditions at about this time or earlier, in a change referred to as the `Mid-Brunhes Climatic Event' (Jansen et al., 1986). The shift in surface water mass dynamics above the slope of the Agulhas Bank as recorded in the studied core coincides with other changes observed in Southeast Atlantic sediments close to the MIS 8/7 boundary. For

MARGO 2974 25-2-02

A.J. Rau et al. / Marine Geology 180 (2002) 183^201

example, Giraudeau et al. (2001) report a change in planktonic foraminiferal distributional trend at 250 kyr BP in ODP Site 1087 located in the southern Benguela upwelling region. The abundance variations of the dominant foraminiferal species change from an ill-de¢ned glacial^interglacial relationship prior to MIS 7, to a clear glacial^interglacial pattern thereafter. Stuut et al. (this volume) show an overall decrease in the size of aeolian dust o¡ the Namibian coast from ca. 250 kyr BP to younger periods, re£ecting a relative decrease in trade-wind strengths. In addition, Schneider et al. (1996) provide evidence for a change in sediment deposits along the Namibian shelf from a carbonate-rich sedimentation between MIS 15 and 8, to a carbonate-poor sedimentation in younger intervals. This sedimentological pattern is probably indicative of a change in dominant primary producers (calcareous nannofossils and planktonic foraminifera versus diatoms) and might be interpreted as a response to a change in the nutrient content of upwelled water over the continental margin of SW Africa (Berger and Wefer, 1996). We argue that the above-mentioned shift in wind stress and nature of primary productivity along SW Africa is, to a certain degree, related to the displacement of subantarctic waters and associated hydrological fronts as identi¢ed in core MD962080. One might indeed consider that an overall northward position of polar and subpolar waters during the MIS 12^8 interval led, through an increase in pressure gradients, to increased trade-wind strength as well as enhanced subtropical gyral circulation. This enhanced Benguela circulation reduced the £ow of the poleward Angolan undercurrent along the SW African margin, which presently constitutes the main silicate source for upwelled waters in the Benguela region. Enhanced upwelling of silicate-poor subsurface waters o¡ SW Africa would consequently result in massive production and subsequent sedimentation of carbonate producers. As discussed previously (see Section 5.1), surface water changes following glacial^interglacial cyclicity are mostly expressed at the site location by advection of cold subpolar waters, bearing a unique signature of Neogloboquadrina pachyderma (d)-rich foraminiferal assemblage (Fig. 5d), during

197

glacial intervals. This probably re£ects equatorward shifts of the APF and associated subpolar waters during periods of Antarctic ice sheet growth. The presence of tropical and subtropical planktonic foraminifera throughout the length of the core studied (Fig. 5e), however, suggests that these advections of subpolar waters to the core location did not induce a drastic equatorward shift in the position of the STC. Conversely, we argue that any major displacements of the APF during glacial cold spells were not associated with uniform meridional shifts of the Southern Ocean hydrological belts south of Africa. Alternatively, our data indicate that any substantial equatorward displacement of the STC was limited to south of the core location (36³S). This is particularly signi¢cant since it implies that the exchange of water from the South Indian to the South Atlantic Ocean was never entirely obstructed by the STC during any of these periods. It is conceivable that a more equatorward location of this front could nonetheless force the Agulhas Current retro£ection eastward and limit the through£ow of warmer subtropical water, either as Agulhas rings or as Agulhas ¢laments. 5.4. Globorotalia menardii and Indian^Atlantic leakages The occurrence of Globorotalia menardii throughout the length of core studied implies that Indian Ocean waters passed over the Agulhas Bank (and very likely into the South Atlantic Ocean) throughout the last 450 kyr (Fig. 6). This supports the inference expressed above that the equatorward shift of the STC was never su¤cient during the last 450 kyr to shut o¡ interocean leakage south of Africa completely. Even if the Agulhas retro£ection had been constrained to remain far to the east, rings and ¢laments may still have passed through the gap remaining between the southern tip of the Africa (which would have shifted poleward during glacial period lowstands) and the STC. If this gap had been closed, water derived from the Agulhas Current could nonetheless have passed over the Agulhas Bank, but this leakage would have been of a di¡erent nature. It would have consisted of surface water

MARGO 2974 25-2-02

198

A.J. Rau et al. / Marine Geology 180 (2002) 183^201

only and would not have included deep-reaching Agulhas rings. The downcore pattern of accumulation rate of Globorotalia menardii bears a unique signature of maximum Indian water leakage at glacial Terminations II, III and IV. A comparison with the distributional trend of dominant foraminiferal species (Fig. 5) suggests that this speci¢c dynamic of Indian water in£ow acted independently of changes in the nature of surface water masses in the vicinity of the core location. Alternatively, this indicates that, at least for the last 450 kyr, the input of Indian water to the South Atlantic was not directly constrained by the dynamics of Southern Ocean waters and associated belt of hydrological fronts. Shannon et al. (1990), among others, suggest that the present zonal position of the Agulhas retro£ection regime may be controlled by the volume transport/inertia of the Agulhas Current. Understanding the Late Quaternary timing and causes of the observed pulsing of Indian water in£ow will therefore bene¢t from additional studies focused on circulation changes along Southeast Africa. 6. Conclusion The Agulhas Bank Slope is presently an oceanographically complex region with the interaction of numerous water masses, hydrological features and currents. Our data indicate that the hydrographic conditions were no less complex in the past. Inferences on the palaeoclimatic movements of the Agulhas system or the STC are therefore made with caution. Warm water at the site of the core may have been due to Agulhas rings and ¢laments, as well as water moving over the Agulhas Bank. Colder water may have come from a higher incidence of cold water intrusions across the STC, associated with non-uniform equatorward movements of the Southern Ocean hydrological fronts (STC and APF). However, the minor, but persistent occurrence of a tropical/subtropical foraminiferal assemblage indicates that the transfer of water from the South Indian Ocean to the South Atlantic Ocean was never completely obstructed throughout

the last 450 kyr. Thus, equatorward migration of the STC during this period was limited to south of the core site (36³S) and therefore not substantial enough to completely shut o¡ the `Cape valve'. The dominance of transitional and subantarctic foraminiferal assemblages suggests that the core site has lain predominantly under an area of cool-temperate mixed water masses. The site has, however, felt the increased impact of cold, subantarctic waters for at least short periods in the past, as seen in the increased abundance of the cold water foraminifer Neogloboquadrina pachyderma (d), particularly during glacial periods. This may have been a result of (1) general surface water cooling, (2) intrusions of subantarctic waters as part of Agulhas ring shedding and due to leakage of subantarctic water through the STC possibly as a result of equatorward migration of this frontal zone, (3) a decreased input of Agulhas Current waters because of a more easterly retro£ection of the Agulhas Current or (4) an unknown combination of these di¡erent mechanisms. Sediment textural data exhibit a glacial^interglacial cyclicity from MIS 7 with coarser textures dominating during glacial intervals. High amplitude variations in the N13 C signal are synchronous with textural changes, with low N13 C values associated with coarser sediments. This, in combination with high BMAR, suggests enhanced productivity and associated increased burial rate during glacial intervals. A change in the benthic N13 C record re£ects a change in dominating bottom water masses from NADW to USCW. A profound change in surface hydrological conditions occurs at 200^250 kyr BP, which is particularly evident in the planktonic foraminiferal record. The data indicate a transition from mixed subantarctic and transitional water masses to overall warmer conditions after 200 kyr BP. Our results indicate that the palaeoceanographic history of this ocean realm is characterised by a complex combination of hydrological changes over the Agulhas Bank Slope, which occurred at various intervals and is not simply driv-

MARGO 2974 25-2-02

A.J. Rau et al. / Marine Geology 180 (2002) 183^201

en by a series of equatorward and poleward migrations of the Southern Ocean hydrological fronts (STC and APF) on glacial and interglacial time scales. The core location lies at the crossroads of various water masses from adjacent ocean basins and records changes in local (Agulhas Current), regional (STC and associated subantarctic water masses) and global (glacial^interglacial cycles) ocean dynamics. Acknowledgements This study forms part of the IMAGES IINAUSICAA (Namibia Angola Upwelling System and Indian Connection to Austral Atlantic) project. The research was funded by the University of Cape Town, the National Research Foundation (South Africa), the Franco-South African Science and Technology Programme and the Centre National de la Recherche Scientifique (CNRS, France). R. Schneider, R. Thunell, and B. Christensen are thanked for their constructive comments and suggestions to improve the paper. Appendix A. Taxonomy of foraminifera Cibicidoides wuellerstor¢ (Schwager, 1866) Globorotalia in£ata (d'Orbigny, 1839) Neogloboquadrina pachyderma (Ehrenberg, 1861) Globigerinita glutinata (Egger, 1893) Globigerina bulloides (d'Orbigny, 1826) Globigerinoides sacculifer (Brady, 1877) Globigerina falconensis (Blow, 1959) Globorotalia scitula (Brady, 1882) Globoquadrina dutertrei (d'Orbigny, 1839) Globigerinoides ruber (d'Orbigny, 1839) Turborotalia quinqueloba (Natland, 1938) Globorotalia truncatulinoides (d'Orbigny, 1839) Globorotalia menardii (Parker, Jones and Brady, 1865) Globorotalia hirsuta (d'Orbigny, 1839) Globorotalia crassaformis (Galloway and Wissler, 1927) Globigerinella calida (Parker, 1962)

199

References Bainbridge, A.E., 1981. GEOSECS Atlantic Expedition, Hydrographic data, 1972^1973. National Science Foundation, US Government Printing O¤ce, Washington, DC, pp. 1^ 121. Bassinot, F.C., Labeyrie, L.D., Vincent, E., Quidelleur, X., Shackleton, N.J., Lancelot, Y., 1994a. The astronomical theory of climate and the age of the Brunhes^Matuyama magnetic reversal. Earth Planet. Sci. Lett. 126, 91^108. Bassinot, F.C., Beaufort, L., Vincent, E., Labeyrie, L.D., Rostek, F., Mu«ller, P.J., Quidelleur, X., Lancelot, Y., 1994b. Coarse fraction £uctuations in pelagic carbonate sediments from the tropical Indian Ocean: A 1500-kyr record of carbonate dissolution. Paleoceanography 9, 579^600. Be¨, A.W.H., 1967. Foraminiferal families: Globigerinidea et Globorotalidae. In: Fraser, J.H. (Ed.), Fiches d'Identi¢cation du Zooplankton. Cons. Permanent Int. Explor. Mar. Charlottenlund, Sheet 108, pp. 1^8. Be¨, A.W.H., Hutson, W.H., 1977. Ecology of planktonic foraminifera and biogeographic patterns of life and fossil assemblages in the Indian Ocean. Micropaleontology 23, 369^ 414. Be¨, A.W.H., Tonderlund, D.S., 1971. Distribution and ecology of living planktonic foraminifera in surface waters if the Atlantic and Indian Oceans. In: Funnel, B.M., Riedel, W.R. (Eds.), The Micropalaeontology of the Oceans. Cambridge University Press, Cambridge, pp. 105^149. Berger, W.H., Wefer, G., 1996. Expeditions into the past: Paleoceanographic studies in the South Atlantic. In: Wefer, G., Berger, W.H., Siedler, G., Webb, D.J. (Eds.), The South Atlantic: Past and Present Circulation. Springer-Verlag, Berlin, pp. 363^410. Bertrand, P., Balut, Y., Schneider, R., Chen, M.-T., Rogers, J., Shipboard participants, 1997. Scienti¢c report of the NAUSICAA-IMAGES coring cruise: La Reunion October 20, 1996 ^ La Reunion November 25, 1996 aboard the S./V. Marion Dufresne. Report l'Institut Francais pour la Recherche et la Technologie Polaires, 97-1, pp. 1^381. Bickert, T., 1992. Rekonstruktion der spa«tquarta«ren Bodenwasser Zirkulation im o«stlichen Su«datlantik u«ber stabile Isotope bentischer Foraminiferen. Ber. FB Geo Univ. Bremen 27, 205 pp. Bickert, T., Wefer, G., 1996. Late Quaternary deep water circulation in the South Atlantic: Reconstruction from carbonate dissolution and benthic stable isotopes. In: Wefer, G., Berger, W.H., Siedler, G., Webb, D.J. (Eds.), The South Atlantic: Past and Present Circulation. Springer-Verlag, Berlin, pp. 599^620. Birch, G.F., 1979. The nature and origin of mixed apatite/ glauconite pellets from the continental shelf o¡ South Africa. Mar. Geol. 29, 313^334. Chen, M.-T., Bertrand, P., Balut, Y., Schneider, R., Rogers, J., Taiwan IMAGES participants, 1998. IMAGES II Cruise (NAUSICAA) explores Quaternary climatic variability and linkage of Benguela and Agulhas Current Systems in the Southern Indian^Atlantic Ocean: Participation by consorti-

MARGO 2974 25-2-02

200

A.J. Rau et al. / Marine Geology 180 (2002) 183^201

um of Taiwan institutions (coordinated by National Taiwan University). J. Geol. Soc. China 41, 73^79. Connolly, J.R., Ewing, M., 1965. Ice-rafted detritus as a climatic indicator in Antarctic deep-sea cores. Science 150, 1822^1824. Coplen, T.B., 1988. Normalisation of oxygen and hydrogen isotope data. Chem. Geol. 72, 293^297. De Ruijter, W., 1982. Asymptotic analysis of the Agulhas and Brazil Current systems. J. Phys. Oceanogr. 12, 361^373. De Ruijter, W.P.M., Biastoch, A., Drijfhout, S.S., Lutjeharms, J.R.E., Matano, R.P., Pichevin, T., van Leeuwen, P.J., Weijer, W., 1999. Indian^Atlantic inter-ocean exchange: dynamics, estimation and impact. J. Geophys. Res. 104, 885^ 911. Dingle, R.V., Rogers, J., 1972. Pleistocene palaeogeography of the Agulhas Bank. Trans. R. Soc. S. Afr. 40, 155^165. Ericson, D.B., Wollin, G., 1968. Pleistocene climates and chronology in deep-sea sediments. Science 162, 1227^1234. Flores, J.A., Gersonde, R., Sierro, F.J., 1999. Pleistocene £uctuations in the Agulhas Current Retro£ection based on the calcareous plankton record. Mar. Micropaleontol. 37, 1^22. Gersonde, R., Hodell, D.A., Blum, P. et al., 1999. Proc. ODP Init. Rep., 177 [CD-ROM]. Available from: Ocean Drilling Program, Texas ApM University, College Station, TX. Giraudeau, J., 1993. Planktonic foraminiferal assemblages in surface sediments from the southwest African continental margin. Mar. Geol. 110, 47^62. Giraudeau, J., Pierre, C., Herve, L., 2001. A late Quaternary, high-resolution record of planktonic foraminiferal species distribution in the southern Benguela region: Site 1087. In: Wefer, G., Berger, W.H., Richter, C. (Eds.), Proc. ODP Sci. Results 175, 1^16. Gordon, A.L., 1986. Inter-ocean exchange of thermocline water. J. Geophys. Res. 91, 5037^5046. Gordon, A.L., Lutjeharms, J.R.E., Gru«ndlingh, M.L., 1987. Strati¢cation and circulation at the Agulhas Retro£ection. Deep-Sea Res. 34, 565^599. Hodell, D.A., 1993. Late Pleistocene paleoceanography of the South Atlantic Sector of the Southern Ocean: Ocean Drilling Programme Hole 704A. Paleoceanography 8, 47^67. Howard, W.R., Prell, W.L., 1992. Late Quaternary surface circulation of the Southern Ocean and its relationship to orbital variations. Paleoceanography 7, 79^117. Hut, G., 1987. Stable isotope reference samples for geochemical and hydrological investigations. Report to the Director General, International Atomic Energy Agency, Vienna. Hutson, W.H., 1980. The Agulhas Current during the Late Pleistocene: analysis of modern faunal analogs. Science 207, 64^66. Imbrie, J., Kipp, N., 1971. A new micropaleontological method for quantitative paleoclimatology: application to a Late Pleistocene Caribbean core. In: Turekian, K.K. (Ed.), Late Cenozoic Glacial Ages, Yale University Press, New Haven, CT, pp. 71^181. Imbrie, J., Hays, J.D., Martinson, D.G., McIntyre, A., Mix, A.C., Morely, J.J., Pisias, N.G., Prell, W.L., Shackleton

N.J., 1984. The orbital theory of Pleistocene climate: support for a revised chronology of the marine N18 O record. In: Berger, A., Imbrie, J., Hays, J., Kukla, G., Saltzman, B. (Eds.), Milankovitch and Climate, Part I, Reidel, Dordrecht, pp. 269^305. Jansen, J.H.F., Kuijpers, A., Troelstra, S.R., 1986. A MidBrunhes climatic event: long-term changes in global atmosphere and ocean circulation. Science 232, 619^622. Kanfoush, S.L., Hodell, D.A., Charles, C.D., Guilderson, T.P., Mortyn, P.G., Ninneman, U.S., 2000. Millenial-scale instability of the Antarctic ice sheet during the last glaciation. Science 288, 1815^1818. Little, M.G., Schneider, R.R., Kroon, D., Price, B., Bickert, T., Wefer, G., 1997. Rapid palaeoceanographical changes in the Benguela Upwelling System for the last 160 000 years as indicated by abundances of planktonic foraminifera. Palaeogeogr. Palaeoclimatol. Palaeoecol. 130, 135^161. Lutjeharms, J.R.E., 1985. Location of frontal systems between Africa and Antarctica: some preliminary results. Deep-Sea Res. 32, 1499^1509. Lutjeharms, J.R.E., Cooper, J., 1996. Interbasin leakage through Agulhas Current ¢laments. Deep-Sea Res. 43, 213^238. Lutjeharms, J.R.E., de Ruijter, W.P.M., 1996. The in£uence of the Agulhas Current on the adjacent coastal ocean: possible impacts of climate change. J. Mar. Syst. 7, 321^336. Lutjeharms, J.R.E., Valentine, H.R., 1988a. Eddies at the SubTropical Convergence south of Africa. J. Phys. Oceanogr. 18, 761^774. Lutjeharms, J.R.E., Valentine, H.R., 1988b. Evidence for persistent Agulhas rings southwest of Cape Town. S. Afr. J. Sci. 84, 781^783. Lutjeharms, J.R.E., van Ballegooyen, R.C., 1988a. The retro£ection of the Agulhas Current. J. Phys. Oceanogr. 18, 1570^1583. Lutjeharms, J.R.E., van Ballegooyen, R.C., 1988b. Anomalous upstream retro£ection in the Agulhas current. Science 240, 1770^1772. Lutjeharms, J.R.E., Monteiro, P.M.S., Tyson, P.D., 2001. The oceans around southern Africa and regional e¡ect of global change. S. Afr. J. Sci. (submitted). Mackensen, A., Hubberten, H.-W., Bickert, T., Fischer, G., Fu«tterer, D.K., 1993. The N13 C in benthic foraminiferal tests of Fontbotia wuellerstor¢ (Schwager) relative to the N13 C of dissolved inorganic carbon in Southern Oceandeep water: implications for glacial ocean circulation models. Paleoceanography 8, 587^610. Mackensen, A., Bickert, T., 1999. Stable carbon isotopes in benthic foraminifera: Proxies for deep and bottom water circulation and new production. In: Fischer, G., Wefer, G. (Eds.), Use of Proxies in Paleoceanography: Examples from the South Atlantic. Springer-Verlag, Berlin, pp. 229^254. Matano, R.P., Simionato, C.G., de Ruijter, W.P., van Leeuwen, P.J., Strub, P.T., Chelton, D.B., Schlax, M.G., 1998. Seasonal variability in the Agulhas retro£ection region. Geophys. Res. Lett. 25, 4361^4364. Needham, H.D., 1962. Ice-rafted rocks from the Atlantic

MARGO 2974 25-2-02

A.J. Rau et al. / Marine Geology 180 (2002) 183^201 Ocean o¡ the Cape of Good Hope. Deep-Sea Res. 9, 475^ 486. Niebler, H.-S., Gersonde, R., 1998. A planktonic foraminiferal transfer function for the southern South Atlantic Ocean. Mar. Micropaleontol. 34, 213^234. Parker, F.L., 1962. Planktonic foraminiferal species in Paci¢c sediments. Micropaleontology 8, 219^254. Penven, P., Lutjeharms, J.R.E., Marchesiello, P., Weeks, S.J., Roy, C., 2001. Generation of cyclonic eddies by the Agulhas Current in the lee of the Agulhas Bank. Geophys. Res. Lett. (submitted). Peterson, L.C., Prell, W.L., 1985. Carbonate dissolution in recent sediments of the eastern equatorial Indian Ocean: Preservation patterns and carbonate loss above the lysocline. Mar. Geol. 64, 259^290. Prell, W.L., Hutson, W.H., Williams, D.F., 1979. The Subtropical Convergence and Late Quaternary circulation in the Southern Ocean. Mar. Micropaleontol. 4, 225^234. Reid, J.L., 1989. On the total geostrophic circulation of the South Atlantic Ocean: Flow patterns, tracers and transports. Prog. Oceanogr. 23, 149^244. Rogers, J., 1999. Preliminary ¢ndings of the IMAGES-II Programme (International MArine Global changE Study) using giant piston cores from the continental slope and rise o¡ southern Africa. S. Afr. J. Geol. 102, 384^390. Ruddiman, W.F., 1977. Late Quaternary deposition of icerafted sand in the subpolar North Atlantic (lat. 40³ to 65³N). Geol. Soc. Am. Bull. 88, 1813^1827. Schneider, R.R., Mu«ller, P.J., Ruhland, G., Meinecke, G., Schmidt, H., Wefer, G., 1996. Late Quaternary surface temperatures and productivity in the east-equatorial Atlantic: Response to changes in trade/monsoon wind forcing and

201

surface water advection. In: Wefer, G., Berger, W.H., Siedler, G., Webb, D.J. (Eds.), The South Atlantic: Past and Present Circulation. Springer-Verlag, Berlin, pp. 527^ 551. Schumann, E.H., Perrins, L.-A., Hunter, I.T., 1982. Upwelling along the south coast of the Cape Province, South Africa. S. Afr. J. Sci. 78, 238^242. Shannon, L.V., Lutjeharms, J.R.E., Agenbag, J.J., 1989. Episodic input of Subantarctic water into the Benguela region. S. Afr. J. Sci. 85, 317^322. Shannon, L.V., Agenbag, J.J., Walker, N.D., Lutjeharms, J.R.E., 1990. A major perturbation in the Agulhas retro£ection area in 1986. Deep-Sea Res. 37, 493^512. Shannon, L.V., Nelson, G., 1996. The Benguela: Large scale features and processes and system variability. In: Wefer, G., Berger, W.H., Siedler, G., Webb, D.J. (Eds.), The South Atlantic: Past and Present Circulation. Springer-Verlag, Berlin, pp. 163^210. Stuut, J.-B.W., Prins, M.A., Schneider, R.R., Weltje, G.J., Jansen, J.H.F., Postma, G., this volume. A 300 kyr record of aridity and wind strength in southwestern Africa: inferences from grain-size distributions of sediments on Walvis Ridge, SE Atlantic. Valentine, H.R., Lutjeharms, J.R.E., Brundrit, G.B., 1993. The water masses and volumetry of the southern Agulhas Current region. Deep-Sea Res. 40, 1285^1305. Weijer, W., de Ruijter, W.P.M., Dijkstra, H.A., van Leeuwen, P.J., 1999. Impact of interbasin exchange on the Atlantic overturning circulation. J. Phys. Oceanogr. 29, 2266^2284. Wu, G., Berger, W., 1991. Pleistocene N18 O record from Ontong Java Plateau: e¡ects of winnowing and dissolution. Mar. Geol. 96, 193^209.

MARGO 2974 25-2-02