- Email: [email protected]
Carbonate dissolution in the South Atlantic Ocean: evidence from ultrastructure breakdown in Globigerina bulloides
Deep-Sea Research I 47 (2000) 603}620
Carbonate dissolution in the South Atlantic Ocean: evidence from ultrastructure breakdown in Globigerina bulloides Nicolas Dittert*, RuK diger Henrich Fachbereich 5 } Geowissenschaften, Universita( t Bremen, 28334 Bremen, Germany Received 16 June 1998; received in revised form 24 February 1999; accepted 23 June 1999
Abstract Ultrastructure dissolution susceptibility of the planktic foraminifer Globigerina bulloides, carbonate ion content of the water column, calcium carbonate content of the sediment surface, and carbonate/carbon weight percentage ratio derived from sediment surface samples were investigated in order to reconstruct the position of the calcite saturation horizon, the sedimentary calcite lysocline, and the calcium carbonate compensation depth (CCD) in the modern South Atlantic Ocean. Carbonate ion data from the water column refer to the GEOSECS locations 48, 103, and 109 and calcium carbonate data come from 19 GeoB sediment surface samples of 4 transects into the Brazil, the Guinea, and the Cape Basins. We present a new (paleo-) oceanographic tool, namely the Globigerina bulloides dissolution index (BDX). Further, we give evidence (a) for progressive G. bulloides ultrastructural breakdown with increasing carbonate dissolution even above the lysocline; (b) for a sharp BDX increase at the sedimentary lysocline; and (c) for the total absence of this species at the CCD. BDX puts us in the position to distinguish the upper open ocean and the upwelling in#uenced continental margin above from the deep ocean below the sedimentary lysocline. Carbonate ion data from water column samples, calcite weight percentage data from surface sediment samples, and carbonate/carbon weight percentage ratio appear to be good proxies to con"rm BDX. As shown by BDX both the calcite saturation horizon (in the water column) and the sedimentary lysocline (at the sediment}water interface) mark the boundary between the carbonate ion undersaturated and highly corrosive Antarctic Bottom Water and the carbonate ion saturated North Atlantic Deep Water (NADW) of the modern South Atlantic. ( 2000 Elsevier Science Ltd. All rights reserved.
* Corresponding author. Present address: Institut Universitaire EuropeH en de la Mer, UMR 6539, Technopo( le Brest-Iroise, F-29280 PlouzaneH , France. Tel.: 0033-2-9849-8673; fax: 0033-2-9849-8645. E-mail address: [email protected] (N. Dittert) 0967-0637/00/$ - see front matter ( 2000 Elsevier Science Ltd. All rights reserved. PII: S 0 9 6 7 - 0 6 3 7 ( 9 9 ) 0 0 0 6 9 - 2
604
N. Dittert, R. Henrich / Deep-Sea Research I 47 (2000) 603}620
Keywords: Calcite dissolution; Globigerina bulloides; Ultrastructural breakdown; Lysocline; Carbonate compensation depth; Deep-water masses; Antarctic Bottom Water; North Atlantic Deep Water; South Atlantic Ocean; Brazil Basin; Guinea Basin; Cape Basin
1. Introduction The thermohaline circulation of the global oceans (Broecker and Denton, 1989; Broecker, 1997) is an essential factor determining the climate on the earth and is crucial for understanding climate change. It contributes substantially to the global transport of heat and water and is important for the global carbon cycle. Since the oceans cover over 70% of the earth's surface, and since deep and intermediate waters of the oceans contain about 60 times more CO than the atmosphere (Broecker and 2 Peng, 1982), they represent `sleeping giantsa in the control of increased CO in the 2 atmosphere. Regarding the reconstruction of the deep-water circulation in the South Atlantic, judgement of both position of the calcite lysocline and thickness of the calcite transition zone is of major climatic importance (Archer, 1996). They are determined by characteristic extent and corrosivity of Antartic Bottom Water (AABW) and North Atlantic Deep Water (NADW), whose di!erent impact is re#ected by distinct preservation patterns of calcareous sediments. Our main objective is to determine the modern sedimentary calcite lysocline and the Calcium carbonate Compensation Depth (CCD) as well as their inter-basin variations by means of the ultrastructure dissolution susceptibility of the planktic foraminifer Globigerina bulloides d'Orbigny 1826. We investigated 19 sediment surface samples of four transects into the Brazil, the Guinea, and the Cape Basins. This led us to the Globigerina bulloides dissolution index (BDX), which is based on the recognition of preservation features in the ultrastructure of planktic foraminiferal tests at the SEM (BeH et al., 1975) and has been evaluated for Neogloboquadrina pachyderma (Henrich, 1989; Baumann and Meggers, 1996) and for G. bulloides (Van Kreveld, 1996). 1.1. Oceanographic sketch Today the circulation in the deep South Atlantic Ocean is dominated by interactions between the AABW and the NADW, which have contrasting features. NADW is indicated by oxygen-enriched, nutrient-depleted water masses of high CO2~ and low 3 CO contents. AABW can be distinguished as an extremely cold, oxygen-depleted and 2 nutrient-enriched water mass of low CO2~ and high CO contents (Kroopnick, 1985; 3 2 Boyle, 1988). Today's mixing zone between AABW and NADW in the South Atlantic is close to the 90 lmol/kg CO2~ isoline (Bainbridge, 1981). The relatively warm 3 and saline NADW occupies the depth interval between 2000 m and 4000 m, while below 4000 m AABW is encountered. With respect to the modern South Atlantic Ocean, the border of AABW with NADW and the calcite saturation horizon coincide (Bickert and Wefer, 1996). Other deep-water reconstructions (e.g., Bainbridge, 1976;
N. Dittert, R. Henrich / Deep-Sea Research I 47 (2000) 603}620
605
Takahashi et al., 1980; Reid, 1996) elucidate the contrasting in#uence of the carbonate ion depleted, CO ameliorated, and corrosive AABW. That is, below the lysocline 2 destruction and dissolution of calcareous sediments become progressively more intense due to the predominance of the AABW. Dissolution subsequently increases in proportion to the fourth power of D CO2~ (Keir, 1980) until the rate of calcite 3 sedimentation is totally compensated for by the rate of calcite dissolution (Bramlette, 1961; Archer, 1996).
2. Material and methods Sediment surface samples (giant box core) were collected during R/V Meteor cruises M9/4, M12/1, M20/2 (Wefer et al., 1989,1990; Schulz et al., 1992) at water depths from 1007 m down to 5213 m (Fig. 1; Table 1). Total carbon (TC) and total organic carbon (TOC) content of bulk sediments were measured with an LECO-CS 244 infrared analyzer. Calcium carbonate content was calculated in weight percentage of the bulk
Fig. 1. General map of the investigated areas; for data consult Table 1.
606
N. Dittert, R. Henrich / Deep-Sea Research I 47 (2000) 603}620
Table 1 Locations and water depths of the investigated surface sediment samples (0}1 cm) and GEOSECS stations 48, 103, 109 Sample
Latitude
Longitude
Depth [m]
Transect 1: MOR } Brazil Basin GeoB 1115-4 03.5583S GeoB 1116-1 03.6233S GeoB 1117-3 03.8173S GeoB 1118-2 03.5603S GeoB 1119-2 02.9983S GEOSECS 48 04.0003S
12.5803W 13.1873W 14.9033W 16.4323W 18.3783W 29.0003W
2921 3471 3977 4675 5213 11}5075
Transect 2: MOR } Guinea Basin GeoB 1106-5 01.7603S GeoB 1105-3 01.6653S GeoB 1104-5 01.0163S GeoB 1103-3 00.6023N GeoB 1102-3 00.5153N GEOSECS 109 02.0003S
12.5523W 12.4283W 10.7053W 09.2633W 08.5883W 04.5003W
2471 3231 3724 4321 4779 2}5132
Transect 3: Walvis Ridge } Western Cape Basin GeoB 1217-1 24.9453S GeoB 1207-2 24.5983S GeoB 1208-1 24.4923S GeoB 1209-1 24.5123S GeoB 1211-1 24.4733S GeoB 1212-2 24.3323S GEOSECS 103 23.9953S
06.7253E 06.8553E 07.1133E 07.2833E 07.5373E 08.2503E 08.5033E
2007 2593 2971 3303 4089 4669 5}4572
Transect 4: Cape Basin } Namibia Continental Margin GeoB 1709-3 23.5883S GeoB 1710-2 23.4303S GeoB 1711-5 23.3173S GeoB 1712-2 23.2553S
10.7583E 11.7033E 12.4623E 12.8033E
3837 2987 1964 1007
sample by CaCO (wt%)"(TC (wt%)!TOC (wt%))8.33. (1) 3 Within the particle #ux (or `raina) crossing the thermocline the ratio between carbonate and organic carbon varies substantially as a function of fertility (Berger and Keir, 1984). Assuming persistent productivity within each transect (Berger et al., 1987; Siedler et al., 1996) the change in the carbonate/carbon weight percentage ratio thus may indicate changing sea#oor carbonate dissolution. The use of the word `lysoclinea is rather fuzzy: sometimes it is used in the traditional sense to mean the shallowest depth at which dissolution is evident and other times it is used to mean the transition zone between the onset of dissolution and the CCD. The latter makes more sense linguistically, and the best solution is
N. Dittert, R. Henrich / Deep-Sea Research I 47 (2000) 603}620
607
to use the phrase `transition zonea when the entire zone is meant (pers. comm. David Archer). For dissolution investigations, the samples were washed through a 63 lm sieve under a weak spray of water to protect them from additional fragmentation. The whole sample '63 lm was sieved on a 150 lm sieve (Imbrie and Kipp, 1971) according to paleoceanographic and paleoclimatic investigation conventions (CLIMAP, 1976). The taxonomy used follows that of Hemleben et al. (1989). The number of G. bulloides specimens to be investigated is settled on the con"dence limit of 95%. Consequently, for determination of dissolution stages regarding ultrastructural breakdown, at least 40 G. bulloides specimens per sample were hand-picked and mounted on a carbon tape, glued on SEM stub, and then gold coated. On the apertural side of each test the ultimate and the penultimate chambers were investigated using ZEISS DSM 940A at 9 mm working distance and 9 kV accelerating voltage. Overview investigations were carried out online at 200] magni"cation; ultrastructure investigations were executed online at 3000] magni"cation on two randomly chosen transects covering the whole chamber. In order to show the feature of `high spinesa, `peeled calcite layersa, and `partial removal of the surface layera the pictures R1-1, R3-1, and R4-1 are shown at a somewhat larger scale (cf. Plate 1). At the end of the experiment every third sample was re-measured `blinda, i.e., without knowing the water depth from which they originated, in order to avoid a subjective component to the greatest possible extent. As based on SEM evidence, G. bulloides ultrastructure distinguishes two principal surface textures, namely the reticulate and the crystalline morphotype. We di!erentiate "ve dissolution stages by the distinct preservation features (A } ) of spines, spine 15 bases, ridges, interpore areas, and pores, each of which gets worse shaped with progressing dissolution: Dissolution stage R : High spines with well-preserved spine bases (Plate 1 R1-1); round 1 pores (Plate 1 R1-2); smooth interpore area (Plate 1 R1-3); intact ridges (Plate 1 R1-4). Dissolution stage R : Slightly denuded spine bases or ridges (Plate 1 R2-1); 2 broadened or funnel-like pores (Plate 1 R2-2); etched interpore areas (Plate 1 R2-3); reduced spine bases (Plate 1 R2-4). Dissolution stage R : Peeled calcite layers (Plate 1 R3-1); formation of haircracks 3 (Plate 1 R3-2); strongly undermined spine bases (Plate 1 R3-3); strongly reduced spine bases or ridges (Plate 1 R3-4). Dissolution stage R : Partial removal of the surface layer (Plate 1 R4-1); forced 4 formation of cracks and holes on ridges or spine bases (Plate 1 R4-2); pores becoming interconnected (Plate 1 R4-3); missing spine bases or ridges (Plate 1 R4-4). Dissolution stage R : Specimens are not preserved even as fragments. 5 Specimens of the crystalline morphotype appeared subordinately, and progressing dissolution hence can be de"ned almost exclusively by worsening of the calcite crystals: Dissolution stage C : Sharp crystal edges (Plate 1 C-1). 1 Dissolution stage C : Weakly rounded edges; pitted surface (Plate 1 C-2). 2
608
N. Dittert, R. Henrich / Deep-Sea Research I 47 (2000) 603}620
Plate 1. Progressive dissolution stages of Globigerina bulloides d'Orbigny; the reticulate tests (R }R ) as 1 4 well as the crystalline tests (C }C ) worsen in preservation features due to increasing dissolution. Pores 1 4 become frayed, spine bases and ridges denude until they are totally dissolved, hair-cracks make the tests disintegrade into fragments, and smooth interpore areas convert to rugged appearance.
Dissolution stage C : Stronger rounded edges; slightly distorted crystallites (Plate 1 C-3). 3 Dissolution stage C : Rounded edges and angles; buried pores; completely distorted 4 crystallites (Plate 1 C-4). Dissolution stage C : Specimens are not preserved even as fragments. 5
N. Dittert, R. Henrich / Deep-Sea Research I 47 (2000) 603}620
609
Dissolution stages R and C were not observed in surface sediment samples. Each 0 0 BDX value between 0 and 5 was calculated weighted on the frequency of appearance per sample: BDX"+ (BDX@)/number of investigated tests,
(2a)
BDX@"+ (A
(2b)
1}5
)/number of features per test obtained,
where A } is the preservation feature. 15 The calcite saturation horizon was estimated by the intercept of seawater carbonate ion content and the concentration of carbonate ions in equilibrium with seawater (saturation state: D CO2~"0 lmol/kg) for calcite as proposed by Takahashi et al. 3 (1980) and Broecker and Takahashi (1978). The depth of the saturation horizon determined from sedimentary evidence is broadly consistent with that determined from carbonate ion measurements (Broecker and Peng, 1982). The thickness of the calcite transition zone covers the range from the saturation horizon to the D CO2~ at 3 the CCD as de"ned by Archer (1996). The determination of carbonate dissolution strongly depends on the ability to measure the solubility product K@ "(Ca2`)(CO2~), 3 41
(3)
where (Ca2`) and (CO2~) are the total concentrations of calcium and carbonate ions 3 in mol/kg of seawater. The values of K@ found in the literature (see compilation by 41 Morse et al., 1980) are di$cult to evaluate because of the diverse experimental techniques employed. Particularly important to every technique is the value of the second dissociation constant of carbonic acid, which is needed to evaluate the distribution of the carbonate alkalinity as bicarbonate and carbonate ions (cf. Mehrbach et al., 1973). The use of the apparent (stoichiometric) constant K@ is more 41 convenient than that of the thermodynamic constant (e.g., Millero, 1979), because applications of the latter require a knowledge of the single ion activity coe$cients of calcium and carbonate ions at the temperatures, salinities, and pressures studied (Ingle et al., 1973). In fact, there is a serious di!erence between the various approaches by 10 lmol and more; thus, depth estimates of the saturation horizon (water column) and the lysocline (sediment) are subject to uncertainties of several tens to hundreds of meters. All data presented are archived in the PANGAEA database at the Alfred Wegener Institute for Polar and Marine Research (http://www.pangaea.de).
3. Results 3.1. Globigerina bulloides dissolution index Each sample is represented by at least 40 BDX values; standard deviation ranges from 0.0 to 0.8 (Fig. 2). Transect 1 (Brazil Basin) BDX mean values increase from 1.1
610
N. Dittert, R. Henrich / Deep-Sea Research I 47 (2000) 603}620
Fig. 2. BDX values of the investigated test surfaces regarding calcite dissolution of Globigerina bulloides.
to 5 with a distinct shift from 1.5 to 2.9 below 3977 m. Transect 2 (Guinea Basin) BDX values intensify from 1.1 to 2; transect 3 (western Cape Basin) BDX values increase from 0.7 to 2.6; transect 4 (Namibia Continental Margin) BDX values rise from 0.9 to 2.4. BDX values versus water depth (Fig. 3A) show a signi"cant correlation on the Guinea Basin transect (r2"0.98), the western Cape Basin transect (r2"0.81), and the Namibia Continental Margin transect (r2"0.99). Brazil Basin transect data show a meaningfully exponential correlation (r2"0.93). BDX values versus * CO2~ (Fig. 3B) also show a relevant correlation on the 3 Guinea Basin transect (r2"0.99), the western Cape Basin transect (r2"0.77), and the Namibia Continental Margin transect (r2"0.28). Brazil Basin transect data show an important exponential correlation (r2"0.96). 3.2. Geochemical parameters The calcium carbonate content (Fig. 4) of the sediment surface samples decreases on all investigated transects with increasing water depth. Transect 1 (Brazil Basin)
N. Dittert, R. Henrich / Deep-Sea Research I 47 (2000) 603}620
611
Fig. 3. (A) Globigerina bulloides dissolution index (BDX) versus water depth. BDX rises with increasing water depth due to increasing carbonate dissolution. Clusters of BDX values distinguish the upper open ocean and the upwelling in#uenced continental margin above the lysocline from the deep ocean below the lysocline. BDX shows strong increase at the lysocline and total absence of G. bulloides at the CCD. (B) Globigerina bulloides dissolution index (BDX) versus * CO2~. BDX rises with decreasing * CO2~ due to 3 3 increasing carbonate dissolution.
CaCO -values drop from 98.5 to 11.5 wt%; transect 2 (Guinea Basin) CaCO -values 3 3 decrease from 98.2 to 79.6 wt%; transect 3 (western Cape Basin) CaCO -values 3 diminish from 96.4 to 50.3 wt%; transect 4 (Namibia Continental Margin) CaCO 3 values decrease from 97.5 to 4.2 wt%. In comparison to inorganic carbon, the C content (Fig. 4) of the sediment 03' surface samples increases on the Brazil Basin transect (0.16 to 0.8 wt%), the Guinea Basin transect (0.14 to 0.48 wt%), and the western Cape Basin transect (0.1 to 0.44 wt%). Continental margin C values decrease from 5.69 to 03' 0.24 wt%. The carbonate/carbon weight percentage ratio (Fig. 4) rises on the Brazil Basin transect from 0.002 to 0.069; on the Guinea Basin transect from 0.001 to 0.006; on the western Cape Basin transect from 0.001 to 0.009. Namibia continental margin carbonate/carbon weight percentage ratio "rst drops from 0.067 to 0.003 and subsequently increases to 0.121.
612
N. Dittert, R. Henrich / Deep-Sea Research I 47 (2000) 603}620
Fig. 4. Modi"cation of CaCO wt%, C wt%, and carbonate/carbon weight percentage ratio versus 3 03' water depth. High carbonate/carbon ratio may be put down to the facts that (a) production of organic carbon is increased substantially, which may lead to supralysoclinal dissolution, and (b) CaCO values 3 decrease due to sublysoclinal dissolution. Where the rain rate of calcitic and non-calcitic material is constant and neither productivity nor dilution is enhanced, carbonate/carbon ratio of organic to inorganic carbon remains constant and low. Data refer to GeoB surface sediment samples.
4. Discussion 4.1. Globigerina bulloides dissolution index As can be visualized immediately by looking at the photographs, the ultrastructure of G. bulloides worsens in regular patterns due to increasing carbonate dissolution (Plate 1). Regarding the reticulate morphotype, spines and spine bases become reduced, spine bases and ridges become denuded, pores become funnel-like, the surface becomes chapped until complete parts will crack o!. Regarding the crystalline morphotype, edges and angles grow round, pores become buried, and at last crystals become completely distorted. However, in order to attain a speci"c dissolution stage not all preservation features of one preservation category have to occur. Applying Eq. (2a) BDX values result in non-integer numbers, which allow us to document even minor di!erences in the dissolution extent. If regarded as data clusters BDX values can be assigned to three di!erent oceanic realms (Fig. 3A). The "rst cluster covers dissolution stages 0.7 and 2.49 and is found in
Fig. 5. Above: * CO2~ of the water column versus CaCO wt% of the sediment. The thickness of the calcite transition zone covers the range from the saturation 3 3 horizon to the * CO2~ at the CCD. The calcite saturation horizon was estimated by the intercept of seawater carbonate ion content and the concentration of 3 carbonate ions in equilibrium with seawater (saturation state: * CO2~"0 lmol/kg) for calcite mineral phase. Below: * CO2~ of the water column versus water 3 3 depth below 2000 m. Applying several methods, the positions of the calcite saturation horizon and the sedimentary lysocline as well as the extent of the calcite transition zone were calculated. * CO2~ values refer to the GEOSECS locations 48, 103, and 109, whereas the calcium carbonate data come from GeoB sediment 3 surface samples.
N. Dittert, R. Henrich / Deep-Sea Research I 47 (2000) 603}620 613
614
N. Dittert, R. Henrich / Deep-Sea Research I 47 (2000) 603}620
surface sediments of the upper open ocean from 2000 to 4300 m, which corresponds to a carbonate ion undersaturation state D CO2~*!5 lmol/kg. The second cluster 3 contains dissolution stages 2.5}5 and indicates the deep open ocean from 4600 to 5300 m water depth, which corresponds to a carbonate ion undersaturation state D CO2~(!5 lmol/kg. Between these two clusters, where dissolution becomes rap3 idly stronger until even fragments are completely dissolved, we locate the border between upper and deep open ocean, namely the sedimentary lysocline. This can also be shown by the decrease in the total number of planktic foraminifera per gram sediment (Dittert et al., 1999). The third cluster includes dissolution stages 0.9}2.4 and is found in sediments of the upwelling in#uenced continental margin from 1000 to 3900 m water depth, which corresponds to a carbonate ion undersaturation state D CO2~*8 lmol/kg. This area is re#ected by high productivity within the Benguela 3 system, where signi"cantly higher amounts of organic carbon than in the open ocean are delivered to the depth. Metabolic activities of benthic organisms lead to respiration and remineralization and subsequently to CO enrichment at the sedi2 ment}water interface. This metabolically driven carbonate dissolution, which was earlier described as supralysoclinal dissolution (Emerson and Bender, 1981; Jahnke et al., 1994; Freiwald, 1995; Martin and Sayles, 1996), results in a higher basic dissolution potential. The dissolution stages R }R (reticulate morphotype) and C }C (crystalline 1 5 1 5 morphotype) based on our G. bulloides examinations are comparable to those of G. bulloides (`no dissolutiona to `severe dissolutiona) investigated by Van Kreveld (1996) and to those of N. pachyderma (D `no dissolutiona to D `severe 0 4 dissolutiona) introduced by Henrich et al. (1989) and Henrich (1989). Similar to our examinations on G. bulloides, sediment investigations on G. bulloides (Van Kreveld, 1996) and laboratory experiments on Globigerinoides trilobus (Hecht et al., 1975) show that the initial dissolution "rst reduces spines and a!ects spine bases (Plate 1). Other micropaleontological approaches undertaken to reconstruct the sedimentary lysocline depth include (1) percent fragmented planktic foraminiferal tests (e.g., Keigwin, 1976; Le and Shackleton, 1992); (2) proportions of solution-susceptible and solutionresistant planktic foraminifera species (e.g., Schott, 1935; Berger, 1979; Boltovskoy and Totah, 1992); (3) the ratio of one organism group to another (e.g., Kennett, 1966; Hay, 1970; Parker, 1971; Berner, 1977; Peterson and Prell, 1985). Actually, each of these parameters is a potential dissolution index, but their variations are controlled by ecological rather than by geochemical factors (see compilation by Dittert et al., 1999). Berger et al. (1982) developed the `elbowa model, which illustrates the rapidly increasing dissolution at the lysocline based on the relationship between foraminiferal preservation states and amount of calcite dissolution as a function of water depth. With our BDX tool we are able to expand Berger's `elbowa model and may distinguish the upper part into the upper open ocean and the upwelling in#uenced continental margin realm. At the bend we position the sedimentary lysocline or the top of the calcite transition zone, respectively, and dissolution stage 5 corresponds to the CCD (Figs. 3A and 5).
N. Dittert, R. Henrich / Deep-Sea Research I 47 (2000) 603}620
615
4.2. Geochemical parameters In order to assess the BDX tool we sought methods, to corroborate our results. By means of the carbonate ion content of the water column (Takahashi et al., 1980) and the calcium carbonate content of the sediment surface samples our conclusions were con"rmed. The carbonate ion content at GEOSECS station 48 (Brazil Basin) ranges from 235 lmol/kg at the surface to 77 lmol/kg at 5075 m water depth. At GEOSECS station 109 (Guinea Basin) the values decrease from 247 lmol/kg at the surface to 90 lmol/kg at 5132 m. At GEOSECS station 103 (western Cape Basin) the carbonate ion content drops from 219 lmol/kg at the surface to 84 lmol/kg at 4572 m. According to the equation introduced by Broecker and Takahashi (1978) the calcite saturation in the water column rises from 48 lmol/kg at the surface to 114 lmol/kg at 5500 m water depth. Using several geochemical approaches we were able to determine the position of the calcite saturation horizon, the sedimentary lysocline, and the CCD (Fig. 5). The calcite saturation horizon (Broecker and Takahashi (1978)) is situated highest in the Cape Basin (transects 3 and 4) at 3950 m. In the Brazil Basin it is located at 4250 m and in the Guinea Basin at 4350 m water depth (Fig. 5). The reconstruction from sedimentary calcite (Figs. 4 and 5) locates the top of the calcite transition zone 100 to 450 m below the calcite saturation horizon. The 80% CaCO isopleth (Figs. 4 and 5; Farrell and 3 Prell, 1989) varies from 4150 m (Namibia continental margin) to 4350 m (Brazil Basin) to 4400 m (western Cape Basin) to 4550 m (Guinea Basin). Since at the calcite saturation horizon D CO2~ becomes zero, above the calcite saturation horizon the 3 seawater is by de"nition supersaturated with respect to carbonate ion content. Nevertheless, an undersaturation of 10 lmol/kg is enough to dissolve almost all the calcite falling to the sea#oor (Broecker and Peng, 1982). The bottom of the calcite transition zone (Archer (1996)) varies from 4950 m (Cape Basin) to 5300 m (Brazil and Guinea Basins). The 10% CaCO isopleth (Figs. 4 and 3 5), which is a marker of the CCD (Bramlette (1961)) varies from 4950 m (Namibia continental margin) to 5300 m (Brazil Basin). That is, in the Guinea Basin and the Cape Basin the CCD is not attained. Where the carbonate content drops below 10 wt%, we can ascertain that both the geochemical and the sedimentary reconstruction lead to a corresponding depth of the bottom of the calcite transition zone. Presumably, this accordance is due to the fact that the 10% CaCO isopleth and the 3 zero intercept of the %CaCO versus D CO2~ relation coincide approximately with 3 3 the progressive dissolution increase of D CO2~ (Keir, 1980) below the top of the 3 calcite transition zone. Another clue for the assessment of the lysocline depth is the in#uence of organic carbon. Our results show that at the open ocean realm of all three South Atlantic basins the amount of sedimentary C is below 1 wt% (Fig. 4). However, the carbonate/carbon 03' weight percentage ratio increases rapidly at distinct water depths. Where the values exceed 0.004, the lysocline (Farrell and Prell (1989)) is attained exactly (Fig. 4). At the continental margin it will be noticed immediately that even in lower water depths high amounts of organic carbon (up to 5.69 wt%) are present, which is explained by the enormous productivity of the Benguela upwelling system. The high carbonate/carbon
616
N. Dittert, R. Henrich / Deep-Sea Research I 47 (2000) 603}620
weight percentage ratio (up to 0.067) and the smooth break of carbonate content (down to 71 wt%) at water depths of 350 m down to 1400 m lead us to assume a certain amount of supralysoclinal dissolution (Emerson and Bender, 1981; Jahnke et al., 1994). 4.3. Comparison of Globigerina bulloides dissolution index and geochemical parameters As we can show by the increase of the BDX values above the lysocline (Fig. 3), carbonate dissolution obviously a!ects the G. bulloides ultrastructure even at supersaturation, as stated by Broecker and Peng (1982). According to BDX values the lysocline is attained only by transect 1 (Brazil Basin) and transect 3 (western Cape Basin), which is documented by a rapid increase of BDX values. In the Brazil Basin and the western Cape Basin, our lysocline reconstruction through BDX matches the approaches by Farrell and Prell (1989). In the Brazil Basin, BDX values increase between 3977 m (GeoB 1117-3) and 4675 m (GeoB 1118-2) by leaps and bounds, which covers the calculated sedimentary lysocline depth of 4350 m (according to Farrell and Prell, 1989). In the western Cape Basin, BDX values increase between 4089 m (GeoB 1211-1) and 4669 m (GeoB 1212-2), which covers the calculated depth of 4400 m (according to Farrell and Prell, 1989). However, the exact determination of the lysocline depth through BDX is not possible due to the low sample density. Nevertheless, dissolution even becomes evident at supersaturation (Fig. 3). We attribute the o!set between the calcite saturation horizon and the sedimentary lysocline to the fact that dissolution occurs mostly at the sediment}water interface rather than in the water column. That is, at this interface [CO2~] is very likely to be 3 higher than at the equivalent water depth in the water column (Le and Shackleton, 1992). However, that is not necessarily the case, since oxic metabolism can make the carbonate ion content of pore waters drop below that of the overlaying bottom water. Of course, variations in the percentage of CaCO in a single sample cannot be simply 3 interpreted as an index of preservation, because the relative abundance of carbonate is controlled by the balance of productivity over dissolution and by dilution due to the in#ux of non-carbonate sedimentary components. Only in the ideal situation, where the rain rate of calcitic and non-calcitic material is constant in space and time could the amount of calcite lost to dissolution be calculated from the percentage of CaCO 3 in the sediment. Otherwise, the calcite content of the sediment from above the lysocline has to be used as the reference for the amount of dissolution, that has occurred in samples from the transition zone (the `depth-transect approacha; e.g. Farrell and Prell, 1989; Curry and Lohmann, 1990; Bickert et al., 1997; Dittert et al., 1999). However, the fact that CaCO contents of supralysoclinal sediment average about 3 90% in the world ocean (Archer, 1996) raises the problem that quite large amounts of dissolution create only very small changes in the carbonate content. According to BDX values the CCD is attained only by transect 1 (Brazil Basin), which is shown by dissolution stage 5 at about 5213 m (Fig. 3). Our reconstruction "ts the approaches by Archer (1996) and Bramlette (1961), which locate the CCD at about 5300 m (Fig. 5). Regarding the top and the bottom of the sedimentary calcite transition zone, all reconstructions show a consistent picture, which we attribute to the bottom water
N. Dittert, R. Henrich / Deep-Sea Research I 47 (2000) 603}620
617
distribution. While AABW } as a composite of Weddell Sea Deep Water and Lower Circumantarctic Deep Water } is moving more or less unhampered northwards within the Brazil Basin, it is detained within the Cape and the Guinea Basin by the Mid Atlantic Ridge and the Walvis Ridge (Fig. 1). The in#ow of AABW into the Guinea Basin is restricted to only small quantities passing the sills through the Romanche Fracture Zone, the Chain Fracture Zone, and the Walvis Passage (Van Bennekom and Berger, 1984; Shannon and Chapman, 1991; Warren and Speer, 1991; Mercier et al., 1994). Consequently, the deepest parts of the Guinea Basin are "lled almost exclusively by O ameliorated, CO depleted, and carbonate ion saturated 2 2 NADW (Kroopnick, 1985). The Cape Basin, although located east of the Mid Atlantic Ridge, is dominated by AABW below 4000 m due to a bottom water passage, which allows AABW to enter the basin from the south (Reid, 1989). Therefore, lysocline depth and sublysoclinal carbonate dissolution pattern in the open ocean seem to depend on both the in#uence of the corrosive AABW and the asymmetry of its oceanographic distribution.
Acknowledgements We thank P.J. MuK ller for providing sedimentary carbon and carbonate data. M. Matthies and R. Schlotte assisted the processing of the GEOSECS data set. Discussions with colleagues (H. Meggers, S. Mulitza, H.-S. Niebler) were helpful. Comments from two anonymous reviewers improved the manuscript. We appreciate technical collaboration by H. Heilmann, R. Henning, and C. Wienberg. This investigation was funded by the German Research Foundation (SFB No. 261). This is SFB 261 contribution no. 275 at Bremen University.
References Archer, D.E., 1996. An Atlas of the distribution of calcium carbonate in sediments of the deep sea. Global Biogeochemical Cycles 10 (1), 159}174. Bainbridge, A.E., 1976. GEOSECS Atlantic Expedition, Vol. 2. Sections and Pro"les. National Science Foundation, Superintendent of Documents, US Government Printing O$ce, Washington, DC, 198pp. Bainbridge, A.E., 1981. GEOSECS Atlantic Expedition, Hydrographic Data, 1972}1973. National Science Foundation, Superintendent of Documents, US Government Printing O$ce, Washington, DC, 121pp. Baumann, K.-H., Meggers, H., 1996. Paleoceanographical change in the Labrador Sea during the last 3.1 MY: evidence from calcareous plankton records. In: Moguilevsky, A., Whatley, R. (Eds.), Microfossils and Oceanic Environments. Aberystwyth-Press, Aberystwyth, pp. 131}154. BeH , A.W.H., Morse, J.W., Harrison, S.M., 1975. Progressive dissolution and ultrastructural breakdown of planktonic foraminifera. Cushman Foundation For Foraminiferal Research Special Publication Vol. 13, 27}55. Berger, W.H., 1979. Preservation of foraminifera. In: Lipps, J.H., Berger, W.H., Buzas, M.A., Douglas, R.G., Ross, C.A. (Eds.), Foraminiferal Ecology and Palecology. Society of Economic Paleontologists and Mineralogists, Houston, pp. 105}155. Berger, W.H., Bonneau, M.-C., Parker, F.L., 1982. Foraminifera on the deep-sea #oor: lysocline and dissolution rate. Oceanologica Acta 5 (2), 249}258.
618
N. Dittert, R. Henrich / Deep-Sea Research I 47 (2000) 603}620
Berger, W.H., Fischer, K., Lai, C., Wu, G., 1987. Ocean productivity and organic carbon #ux (Part I. Overview and maps of primary production and export production). University of California, San Diego, SIO Report, Vols. 87}30, 67pp. Berger, W.H., Keir, R.S., 1984. Glacial-Holocene changes in atmospheric CO and the deep-sea record. In: 2 Hansen, J.E., Takahashi, T. (Eds.), Climate Processes and Climate Sensitivity. Geophysical Monographs, Vol. 29. Maurice Ewing Series. Washington, DC, pp. 337}351 Berner, R.A., 1977. Sedimentation and dissolution of pteropods in the ocean. In: Andersen, N.R., Malaho!, A. (Eds.), The Fate of Fossil Fuel CO in the Oceans. Plenum Press, New York, pp. 243}260. 2 Bickert, T., Cordes, R., Wefer, G., 1997. Late Pliocene to middle Pleistocene (2.6 to 1.0 Ma) carbonate dissolution in the western equatorial Atlantic : results of Leg 154, Ceara Rise. Proceedings of the Ocean Drilling Program. Scienti"c Results Vol. 154, 229}237. 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: Present and Past Circulation. Springer, Berlin, pp. 599}620. Boltovskoy, E., Totah, V.I., 1992. Preservation index and preservation potential of some foraminiferal species. Journal of Foraminiferal Research 22 (3), 267}273. Boyle, E.A., 1988. Vertical oceanic nutrient fractionation and glacial/interglacial CO cycles. Nature 331, 2 55}56. Bramlette, M.N., 1961. Pelagic sediments. In: Sears, M. (Ed.), Oceanography. AAAS Publication, New York, pp. 345}366. Broecker, W.S., 1997. Thermohaline circulation, the Achilles heel of our climate system: will man made CO upset the current balance?. Science 278, 1582}1588. 2 Broecker, W.S., Denton, G.H., 1989. The role of ocean}atmosphere reorganizations in glacial cycles. Geochimica et Cosmochimica Acta 53, 2465}2501. Broecker, W.S., Peng, T.H., 1982. Tracers in the Sea. Eldigio Press, New York, pp. 689. Broecker, W.S., Takahashi, T., 1978. The relationship between lysocline depth and in situ carbonate ion concentration. Deep-Sea Research 25F (1), 65}95. CLIMAP Project members, 1976. The surface of the ice-age earth. Science 191, 1131}1137. Curry, W.B., Lohmann, G.P., 1990. Reconstructing past particle #uxes in the tropical Atlantic Ocean. Paleoceanography 5, 487}506. d'Orbigny, A., 1826. Tableau meH thodique de la classe des CeH phalopodes. Annales des Sciences Naturelles 7, 245}314. Dittert, N., Baumann, K.-H., Bickert, T., Henrich, R., Huber, R., Kinkel, H., Meggers, H., 1999. Carbonate dissolution in the deep-sea: methods, quanti"cation and paleoceanographic application. In: Fischer, G., Wefer, G. (Eds.), Use of Proxies in Paleoceanography * Examples from the South Atlantic. Springer, Berlin, pp. 255}284. Emerson, S., Bender, M., 1981. Carbon #uxes at the sediment-water interface of the deep-sea: calcium carbonate preservation. Journal of Marine Research 39, 139}162. Farrell, J.W., Prell, W.L., 1989. Climatic change and CaCO preservation: an 800,000 year bathymetric 3 reconstruction from the central equatorial Paci"c Ocean. Paleoceanography 4 (4), 447}466. Freiwald, A., 1995. Bacteria-induced carbonate degradation: a taphonomic case study of Cibicides lobolatus from a high-boreal carbonate setting. Palaios 10, 337}346. Hay, W.W., 1970. Calcareous nannofossils from cores recovered on Leg 4. Initial Report Deep Sea Drilling Project Vol. 4, 455}501. Hecht, A.D., Eslinger, E.V., Garmon, L.B., 1975. Experimental studies on the dissolution of planktonic foraminifera. Cushman Foundation For Foraminiferal Research Special Publication Vol. 13, 56}96. Hemleben, C., Spindler, M., Anderson, O. R., 1989. Modern Planktonic Foraminifera. Springer, New York, 363pp. Henrich, R., 1989. Glacial/interglacial cycles in the Norwegian Sea : sedimentology, paleoceanography, and evolution of late Pliocene to Quaternary northern hemisphere climate. Proceedings of the Ocean Drilling Program. Scienti"c Results Vol. 104, 189}232. Henrich, R., Kassens, H., Vogelsang, E., Thiede, J., 1989. Sedimentary facies of glacial}interglacial cycles in the Norwegian Sea during the last 350 ka. Marine Geology 86, 283}319.
N. Dittert, R. Henrich / Deep-Sea Research I 47 (2000) 603}620
619
Imbrie, J., Kipp, N.G., 1971. A new micropaleontological method for quantitative paleoclimatology: application to a Late Pleistocene Carribean core. In: Turekian, K.K. (Ed.), Late Cenocoic Glacial Ages. Yale University Press, New Haven, pp. 71}181. Ingle, S.E., Culberson, C.H., Hawley, J.E., Pytkowicz, R.M., 1973. The solubility of calcite in seawater at atmospheric pressure and 35% salinity. Marine Chemistry 1, 295}307. Jahnke, R.A., Craven, D.B., Gaillard, J.F., 1994. The in#uence of organic matter diagenesis on CaCO 3 dissolution at the deep-sea #oor. Geochimica et Cosmochimica Acta 58, 2799}2809. Keigwin, L.D., 1976. Late Cenozoic planktonic foraminiferal biostratigraphy and paleoceanography of the Panama Basin. Micropaleontology 22, 419}422. Keir, R.S., 1980. The dissolution kinetics of biogenic calcium carbonates in seawater. Geochimica et Cosmochimica Acta 44, 241}252. Kennett, J.P., 1966. Foraminiferal evidence of a shallow calcium carbonate solution boundary, Ross Sea, Antarctica. Science 153, 191}193. Kroopnick, P.M., 1985. The distribution of 13C of z CO in the World Oceans. Deep-Sea Research 32 (1), 2 57}84. Le, J., Shackleton, N.J., 1992. Carbonate dissolution #uctuations in the western equatorial Paci"c during the Late Quaternary. Paleoceanography 7, 21}42. Martin, W.R., Sayles, F.L., 1996. CaCO dissolution in sediments of the Ceara Rise, western equatorial 3 Atlantic. Geochimica et Cosmochimica Acta 60 (2), 243}263. Mehrbach, C., Culberson, C.H., Hawley, J.E., Pytkowicz, R.M., 1973. Measurement of the apparent dissociation constants of carbonic acid in seawater at atmospheric pressure. Limnological Oceanography 18, 892}907. Mercier, H., Speer, K.G., Honnorez, J., 1994. Tracing the Antarctic Bottom Water through the Romanche and Chain Fracture Zones. Deep-Sea Research I 41, 1457}1477. Millero, F.J., 1979. The thermodynamics of the carbonate system in seawater. Geochimica et Cosmochimica Acta 43, 1651}1661. Morse, J.W., Mucci, A., Millero, F.J., 1980. The solubility of calcite and aragonite in seawater of 35% salinity at 253C and atmospheric pressure. Geochimica et Cosmochimica Acta 44, 85}94. Parker, F.L., 1971. Distribution of planktonic foraminifera in recent deep-sea sediments. In: Funnell, B.M., Riedel, W.R. (Eds.), The Micropaleontology of oceans. Cambridge University Press, Cambridge, pp. 289}307. 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. Marine Geology 64, 259}290. Reid, J.L., 1989. On the total geostrophic circulation of the South Atlantic Ocean: #ow patterns, tracers and transports. Progress in Oceanography 23, 149}244. Reid, J.R., 1996. On the circulation of the South Atlantic. In: Wefer, G., Berger, W.H., Siedler, G., Webb, D.J. (Eds.), The South Atlantic: present and Past Circulation. Springer, Berlin, pp. 13}44. Schott, W., 1935. Die Foraminiferen in dem aK quatorialen Teil des Atlantischen Ozeans. Wissenschaftliche Ergebnisse der Deutschen Atlantik Expedition Meteor 1925}1927, Vol. 3, 135pp. Schulz, H.D., cruise participants, 1992. Bericht und erste Ergebnisse uK ber die Meteor-Fahrt M20/2, Abidjan-Dakar, 27.12.1991}3.2.1992. Berichte, Fachbereich Geowissenschaften, UniversitaK t Bremen, Vol. 25, 173pp. Shannon, L.V., Chapman, P., 1991. Evidence of Antarctic Bottom Water in the Angola Basin at 323S. Deep-Sea Research 38, 1299}1304. Siedler, G., MuK ller, T.J., Onken, R., Arhan, M., Mercier, H., King, B.A., Saunders, P.M., 1996. The zonal WOCE sections in the South Atlantic. In: Wefer, G., Berger, W.H., Siedler, G., Webb, D.J. (Eds.), The South Atlantic: Present and Past Circulation. Springer, Berlin, pp. 83}104. Takahashi, T., Broecker, W. S., Bainbridge, A. E., Weiss, R. F., 1980. Carbonate chemistry of the Atlantic, Paci"c and Indian Oceans: the results of the GEOSECS expeditions 1972}1978. Lamont-Doherty Geological Observatory, Palisades, NY, 211pp. Van Bennekom, A.J., Berger, G.W., 1984. Hydrography and silica budget of the Angola Basin. Netherlands Journal of Sea Research 17, 149}200.
620
N. Dittert, R. Henrich / Deep-Sea Research I 47 (2000) 603}620
Van Kreveld, S.A., 1996. Calcium carbonate dissolution indices on the northeast Atlantic sea #oor. In: Van Kreveld-Alfane, S.A. (Ed.), Late Quaternary Sediment Records of Mid-Latitude Northeast Atlantic Calcium Carbonate Production and Dissolution. FEBO B.V., Enschede, pp. 135}184. Warren, B.A., Speer, K.G., 1991. Deep circulation in the eastern South Atlantic Ocean. Deep-Sea Research 38, 281}322. Wefer, G., Bleil, U., cruise participants, 1990. Bericht uK ber die Meteor-Fahrt M12/1, Kapstadt-Funchal, 13.3.1990}14.4.1990. Berichte, Fachbereich Geowissenschaften, UniversitaK t Bremen, Vol. 11, 66pp. Wefer, G., Bleil, U., Schulz, H. D., cruise participants, 1989. Bericht uK ber die Meteor-Fahrt M9-4, Dakar-Santa Cruz, 19.2.}16.3.1989. Berichte, Fachbereich Geowissenschaften, UniversitaK t Bremen, Vol. 7, 103pp.
Carbonate dissolution in the South Atlantic Ocean: evidence from ultrastructure breakdown in Globigerina bulloides Nicolas Dittert*, RuK diger Henrich Fachbereich 5 } Geowissenschaften, Universita( t Bremen, 28334 Bremen, Germany Received 16 June 1998; received in revised form 24 February 1999; accepted 23 June 1999
Abstract Ultrastructure dissolution susceptibility of the planktic foraminifer Globigerina bulloides, carbonate ion content of the water column, calcium carbonate content of the sediment surface, and carbonate/carbon weight percentage ratio derived from sediment surface samples were investigated in order to reconstruct the position of the calcite saturation horizon, the sedimentary calcite lysocline, and the calcium carbonate compensation depth (CCD) in the modern South Atlantic Ocean. Carbonate ion data from the water column refer to the GEOSECS locations 48, 103, and 109 and calcium carbonate data come from 19 GeoB sediment surface samples of 4 transects into the Brazil, the Guinea, and the Cape Basins. We present a new (paleo-) oceanographic tool, namely the Globigerina bulloides dissolution index (BDX). Further, we give evidence (a) for progressive G. bulloides ultrastructural breakdown with increasing carbonate dissolution even above the lysocline; (b) for a sharp BDX increase at the sedimentary lysocline; and (c) for the total absence of this species at the CCD. BDX puts us in the position to distinguish the upper open ocean and the upwelling in#uenced continental margin above from the deep ocean below the sedimentary lysocline. Carbonate ion data from water column samples, calcite weight percentage data from surface sediment samples, and carbonate/carbon weight percentage ratio appear to be good proxies to con"rm BDX. As shown by BDX both the calcite saturation horizon (in the water column) and the sedimentary lysocline (at the sediment}water interface) mark the boundary between the carbonate ion undersaturated and highly corrosive Antarctic Bottom Water and the carbonate ion saturated North Atlantic Deep Water (NADW) of the modern South Atlantic. ( 2000 Elsevier Science Ltd. All rights reserved.
* Corresponding author. Present address: Institut Universitaire EuropeH en de la Mer, UMR 6539, Technopo( le Brest-Iroise, F-29280 PlouzaneH , France. Tel.: 0033-2-9849-8673; fax: 0033-2-9849-8645. E-mail address: [email protected] (N. Dittert) 0967-0637/00/$ - see front matter ( 2000 Elsevier Science Ltd. All rights reserved. PII: S 0 9 6 7 - 0 6 3 7 ( 9 9 ) 0 0 0 6 9 - 2
604
N. Dittert, R. Henrich / Deep-Sea Research I 47 (2000) 603}620
Keywords: Calcite dissolution; Globigerina bulloides; Ultrastructural breakdown; Lysocline; Carbonate compensation depth; Deep-water masses; Antarctic Bottom Water; North Atlantic Deep Water; South Atlantic Ocean; Brazil Basin; Guinea Basin; Cape Basin
1. Introduction The thermohaline circulation of the global oceans (Broecker and Denton, 1989; Broecker, 1997) is an essential factor determining the climate on the earth and is crucial for understanding climate change. It contributes substantially to the global transport of heat and water and is important for the global carbon cycle. Since the oceans cover over 70% of the earth's surface, and since deep and intermediate waters of the oceans contain about 60 times more CO than the atmosphere (Broecker and 2 Peng, 1982), they represent `sleeping giantsa in the control of increased CO in the 2 atmosphere. Regarding the reconstruction of the deep-water circulation in the South Atlantic, judgement of both position of the calcite lysocline and thickness of the calcite transition zone is of major climatic importance (Archer, 1996). They are determined by characteristic extent and corrosivity of Antartic Bottom Water (AABW) and North Atlantic Deep Water (NADW), whose di!erent impact is re#ected by distinct preservation patterns of calcareous sediments. Our main objective is to determine the modern sedimentary calcite lysocline and the Calcium carbonate Compensation Depth (CCD) as well as their inter-basin variations by means of the ultrastructure dissolution susceptibility of the planktic foraminifer Globigerina bulloides d'Orbigny 1826. We investigated 19 sediment surface samples of four transects into the Brazil, the Guinea, and the Cape Basins. This led us to the Globigerina bulloides dissolution index (BDX), which is based on the recognition of preservation features in the ultrastructure of planktic foraminiferal tests at the SEM (BeH et al., 1975) and has been evaluated for Neogloboquadrina pachyderma (Henrich, 1989; Baumann and Meggers, 1996) and for G. bulloides (Van Kreveld, 1996). 1.1. Oceanographic sketch Today the circulation in the deep South Atlantic Ocean is dominated by interactions between the AABW and the NADW, which have contrasting features. NADW is indicated by oxygen-enriched, nutrient-depleted water masses of high CO2~ and low 3 CO contents. AABW can be distinguished as an extremely cold, oxygen-depleted and 2 nutrient-enriched water mass of low CO2~ and high CO contents (Kroopnick, 1985; 3 2 Boyle, 1988). Today's mixing zone between AABW and NADW in the South Atlantic is close to the 90 lmol/kg CO2~ isoline (Bainbridge, 1981). The relatively warm 3 and saline NADW occupies the depth interval between 2000 m and 4000 m, while below 4000 m AABW is encountered. With respect to the modern South Atlantic Ocean, the border of AABW with NADW and the calcite saturation horizon coincide (Bickert and Wefer, 1996). Other deep-water reconstructions (e.g., Bainbridge, 1976;
N. Dittert, R. Henrich / Deep-Sea Research I 47 (2000) 603}620
605
Takahashi et al., 1980; Reid, 1996) elucidate the contrasting in#uence of the carbonate ion depleted, CO ameliorated, and corrosive AABW. That is, below the lysocline 2 destruction and dissolution of calcareous sediments become progressively more intense due to the predominance of the AABW. Dissolution subsequently increases in proportion to the fourth power of D CO2~ (Keir, 1980) until the rate of calcite 3 sedimentation is totally compensated for by the rate of calcite dissolution (Bramlette, 1961; Archer, 1996).
2. Material and methods Sediment surface samples (giant box core) were collected during R/V Meteor cruises M9/4, M12/1, M20/2 (Wefer et al., 1989,1990; Schulz et al., 1992) at water depths from 1007 m down to 5213 m (Fig. 1; Table 1). Total carbon (TC) and total organic carbon (TOC) content of bulk sediments were measured with an LECO-CS 244 infrared analyzer. Calcium carbonate content was calculated in weight percentage of the bulk
Fig. 1. General map of the investigated areas; for data consult Table 1.
606
N. Dittert, R. Henrich / Deep-Sea Research I 47 (2000) 603}620
Table 1 Locations and water depths of the investigated surface sediment samples (0}1 cm) and GEOSECS stations 48, 103, 109 Sample
Latitude
Longitude
Depth [m]
Transect 1: MOR } Brazil Basin GeoB 1115-4 03.5583S GeoB 1116-1 03.6233S GeoB 1117-3 03.8173S GeoB 1118-2 03.5603S GeoB 1119-2 02.9983S GEOSECS 48 04.0003S
12.5803W 13.1873W 14.9033W 16.4323W 18.3783W 29.0003W
2921 3471 3977 4675 5213 11}5075
Transect 2: MOR } Guinea Basin GeoB 1106-5 01.7603S GeoB 1105-3 01.6653S GeoB 1104-5 01.0163S GeoB 1103-3 00.6023N GeoB 1102-3 00.5153N GEOSECS 109 02.0003S
12.5523W 12.4283W 10.7053W 09.2633W 08.5883W 04.5003W
2471 3231 3724 4321 4779 2}5132
Transect 3: Walvis Ridge } Western Cape Basin GeoB 1217-1 24.9453S GeoB 1207-2 24.5983S GeoB 1208-1 24.4923S GeoB 1209-1 24.5123S GeoB 1211-1 24.4733S GeoB 1212-2 24.3323S GEOSECS 103 23.9953S
06.7253E 06.8553E 07.1133E 07.2833E 07.5373E 08.2503E 08.5033E
2007 2593 2971 3303 4089 4669 5}4572
Transect 4: Cape Basin } Namibia Continental Margin GeoB 1709-3 23.5883S GeoB 1710-2 23.4303S GeoB 1711-5 23.3173S GeoB 1712-2 23.2553S
10.7583E 11.7033E 12.4623E 12.8033E
3837 2987 1964 1007
sample by CaCO (wt%)"(TC (wt%)!TOC (wt%))8.33. (1) 3 Within the particle #ux (or `raina) crossing the thermocline the ratio between carbonate and organic carbon varies substantially as a function of fertility (Berger and Keir, 1984). Assuming persistent productivity within each transect (Berger et al., 1987; Siedler et al., 1996) the change in the carbonate/carbon weight percentage ratio thus may indicate changing sea#oor carbonate dissolution. The use of the word `lysoclinea is rather fuzzy: sometimes it is used in the traditional sense to mean the shallowest depth at which dissolution is evident and other times it is used to mean the transition zone between the onset of dissolution and the CCD. The latter makes more sense linguistically, and the best solution is
N. Dittert, R. Henrich / Deep-Sea Research I 47 (2000) 603}620
607
to use the phrase `transition zonea when the entire zone is meant (pers. comm. David Archer). For dissolution investigations, the samples were washed through a 63 lm sieve under a weak spray of water to protect them from additional fragmentation. The whole sample '63 lm was sieved on a 150 lm sieve (Imbrie and Kipp, 1971) according to paleoceanographic and paleoclimatic investigation conventions (CLIMAP, 1976). The taxonomy used follows that of Hemleben et al. (1989). The number of G. bulloides specimens to be investigated is settled on the con"dence limit of 95%. Consequently, for determination of dissolution stages regarding ultrastructural breakdown, at least 40 G. bulloides specimens per sample were hand-picked and mounted on a carbon tape, glued on SEM stub, and then gold coated. On the apertural side of each test the ultimate and the penultimate chambers were investigated using ZEISS DSM 940A at 9 mm working distance and 9 kV accelerating voltage. Overview investigations were carried out online at 200] magni"cation; ultrastructure investigations were executed online at 3000] magni"cation on two randomly chosen transects covering the whole chamber. In order to show the feature of `high spinesa, `peeled calcite layersa, and `partial removal of the surface layera the pictures R1-1, R3-1, and R4-1 are shown at a somewhat larger scale (cf. Plate 1). At the end of the experiment every third sample was re-measured `blinda, i.e., without knowing the water depth from which they originated, in order to avoid a subjective component to the greatest possible extent. As based on SEM evidence, G. bulloides ultrastructure distinguishes two principal surface textures, namely the reticulate and the crystalline morphotype. We di!erentiate "ve dissolution stages by the distinct preservation features (A } ) of spines, spine 15 bases, ridges, interpore areas, and pores, each of which gets worse shaped with progressing dissolution: Dissolution stage R : High spines with well-preserved spine bases (Plate 1 R1-1); round 1 pores (Plate 1 R1-2); smooth interpore area (Plate 1 R1-3); intact ridges (Plate 1 R1-4). Dissolution stage R : Slightly denuded spine bases or ridges (Plate 1 R2-1); 2 broadened or funnel-like pores (Plate 1 R2-2); etched interpore areas (Plate 1 R2-3); reduced spine bases (Plate 1 R2-4). Dissolution stage R : Peeled calcite layers (Plate 1 R3-1); formation of haircracks 3 (Plate 1 R3-2); strongly undermined spine bases (Plate 1 R3-3); strongly reduced spine bases or ridges (Plate 1 R3-4). Dissolution stage R : Partial removal of the surface layer (Plate 1 R4-1); forced 4 formation of cracks and holes on ridges or spine bases (Plate 1 R4-2); pores becoming interconnected (Plate 1 R4-3); missing spine bases or ridges (Plate 1 R4-4). Dissolution stage R : Specimens are not preserved even as fragments. 5 Specimens of the crystalline morphotype appeared subordinately, and progressing dissolution hence can be de"ned almost exclusively by worsening of the calcite crystals: Dissolution stage C : Sharp crystal edges (Plate 1 C-1). 1 Dissolution stage C : Weakly rounded edges; pitted surface (Plate 1 C-2). 2
608
N. Dittert, R. Henrich / Deep-Sea Research I 47 (2000) 603}620
Plate 1. Progressive dissolution stages of Globigerina bulloides d'Orbigny; the reticulate tests (R }R ) as 1 4 well as the crystalline tests (C }C ) worsen in preservation features due to increasing dissolution. Pores 1 4 become frayed, spine bases and ridges denude until they are totally dissolved, hair-cracks make the tests disintegrade into fragments, and smooth interpore areas convert to rugged appearance.
Dissolution stage C : Stronger rounded edges; slightly distorted crystallites (Plate 1 C-3). 3 Dissolution stage C : Rounded edges and angles; buried pores; completely distorted 4 crystallites (Plate 1 C-4). Dissolution stage C : Specimens are not preserved even as fragments. 5
N. Dittert, R. Henrich / Deep-Sea Research I 47 (2000) 603}620
609
Dissolution stages R and C were not observed in surface sediment samples. Each 0 0 BDX value between 0 and 5 was calculated weighted on the frequency of appearance per sample: BDX"+ (BDX@)/number of investigated tests,
(2a)
BDX@"+ (A
(2b)
1}5
)/number of features per test obtained,
where A } is the preservation feature. 15 The calcite saturation horizon was estimated by the intercept of seawater carbonate ion content and the concentration of carbonate ions in equilibrium with seawater (saturation state: D CO2~"0 lmol/kg) for calcite as proposed by Takahashi et al. 3 (1980) and Broecker and Takahashi (1978). The depth of the saturation horizon determined from sedimentary evidence is broadly consistent with that determined from carbonate ion measurements (Broecker and Peng, 1982). The thickness of the calcite transition zone covers the range from the saturation horizon to the D CO2~ at 3 the CCD as de"ned by Archer (1996). The determination of carbonate dissolution strongly depends on the ability to measure the solubility product K@ "(Ca2`)(CO2~), 3 41
(3)
where (Ca2`) and (CO2~) are the total concentrations of calcium and carbonate ions 3 in mol/kg of seawater. The values of K@ found in the literature (see compilation by 41 Morse et al., 1980) are di$cult to evaluate because of the diverse experimental techniques employed. Particularly important to every technique is the value of the second dissociation constant of carbonic acid, which is needed to evaluate the distribution of the carbonate alkalinity as bicarbonate and carbonate ions (cf. Mehrbach et al., 1973). The use of the apparent (stoichiometric) constant K@ is more 41 convenient than that of the thermodynamic constant (e.g., Millero, 1979), because applications of the latter require a knowledge of the single ion activity coe$cients of calcium and carbonate ions at the temperatures, salinities, and pressures studied (Ingle et al., 1973). In fact, there is a serious di!erence between the various approaches by 10 lmol and more; thus, depth estimates of the saturation horizon (water column) and the lysocline (sediment) are subject to uncertainties of several tens to hundreds of meters. All data presented are archived in the PANGAEA database at the Alfred Wegener Institute for Polar and Marine Research (http://www.pangaea.de).
3. Results 3.1. Globigerina bulloides dissolution index Each sample is represented by at least 40 BDX values; standard deviation ranges from 0.0 to 0.8 (Fig. 2). Transect 1 (Brazil Basin) BDX mean values increase from 1.1
610
N. Dittert, R. Henrich / Deep-Sea Research I 47 (2000) 603}620
Fig. 2. BDX values of the investigated test surfaces regarding calcite dissolution of Globigerina bulloides.
to 5 with a distinct shift from 1.5 to 2.9 below 3977 m. Transect 2 (Guinea Basin) BDX values intensify from 1.1 to 2; transect 3 (western Cape Basin) BDX values increase from 0.7 to 2.6; transect 4 (Namibia Continental Margin) BDX values rise from 0.9 to 2.4. BDX values versus water depth (Fig. 3A) show a signi"cant correlation on the Guinea Basin transect (r2"0.98), the western Cape Basin transect (r2"0.81), and the Namibia Continental Margin transect (r2"0.99). Brazil Basin transect data show a meaningfully exponential correlation (r2"0.93). BDX values versus * CO2~ (Fig. 3B) also show a relevant correlation on the 3 Guinea Basin transect (r2"0.99), the western Cape Basin transect (r2"0.77), and the Namibia Continental Margin transect (r2"0.28). Brazil Basin transect data show an important exponential correlation (r2"0.96). 3.2. Geochemical parameters The calcium carbonate content (Fig. 4) of the sediment surface samples decreases on all investigated transects with increasing water depth. Transect 1 (Brazil Basin)
N. Dittert, R. Henrich / Deep-Sea Research I 47 (2000) 603}620
611
Fig. 3. (A) Globigerina bulloides dissolution index (BDX) versus water depth. BDX rises with increasing water depth due to increasing carbonate dissolution. Clusters of BDX values distinguish the upper open ocean and the upwelling in#uenced continental margin above the lysocline from the deep ocean below the lysocline. BDX shows strong increase at the lysocline and total absence of G. bulloides at the CCD. (B) Globigerina bulloides dissolution index (BDX) versus * CO2~. BDX rises with decreasing * CO2~ due to 3 3 increasing carbonate dissolution.
CaCO -values drop from 98.5 to 11.5 wt%; transect 2 (Guinea Basin) CaCO -values 3 3 decrease from 98.2 to 79.6 wt%; transect 3 (western Cape Basin) CaCO -values 3 diminish from 96.4 to 50.3 wt%; transect 4 (Namibia Continental Margin) CaCO 3 values decrease from 97.5 to 4.2 wt%. In comparison to inorganic carbon, the C content (Fig. 4) of the sediment 03' surface samples increases on the Brazil Basin transect (0.16 to 0.8 wt%), the Guinea Basin transect (0.14 to 0.48 wt%), and the western Cape Basin transect (0.1 to 0.44 wt%). Continental margin C values decrease from 5.69 to 03' 0.24 wt%. The carbonate/carbon weight percentage ratio (Fig. 4) rises on the Brazil Basin transect from 0.002 to 0.069; on the Guinea Basin transect from 0.001 to 0.006; on the western Cape Basin transect from 0.001 to 0.009. Namibia continental margin carbonate/carbon weight percentage ratio "rst drops from 0.067 to 0.003 and subsequently increases to 0.121.
612
N. Dittert, R. Henrich / Deep-Sea Research I 47 (2000) 603}620
Fig. 4. Modi"cation of CaCO wt%, C wt%, and carbonate/carbon weight percentage ratio versus 3 03' water depth. High carbonate/carbon ratio may be put down to the facts that (a) production of organic carbon is increased substantially, which may lead to supralysoclinal dissolution, and (b) CaCO values 3 decrease due to sublysoclinal dissolution. Where the rain rate of calcitic and non-calcitic material is constant and neither productivity nor dilution is enhanced, carbonate/carbon ratio of organic to inorganic carbon remains constant and low. Data refer to GeoB surface sediment samples.
4. Discussion 4.1. Globigerina bulloides dissolution index As can be visualized immediately by looking at the photographs, the ultrastructure of G. bulloides worsens in regular patterns due to increasing carbonate dissolution (Plate 1). Regarding the reticulate morphotype, spines and spine bases become reduced, spine bases and ridges become denuded, pores become funnel-like, the surface becomes chapped until complete parts will crack o!. Regarding the crystalline morphotype, edges and angles grow round, pores become buried, and at last crystals become completely distorted. However, in order to attain a speci"c dissolution stage not all preservation features of one preservation category have to occur. Applying Eq. (2a) BDX values result in non-integer numbers, which allow us to document even minor di!erences in the dissolution extent. If regarded as data clusters BDX values can be assigned to three di!erent oceanic realms (Fig. 3A). The "rst cluster covers dissolution stages 0.7 and 2.49 and is found in
Fig. 5. Above: * CO2~ of the water column versus CaCO wt% of the sediment. The thickness of the calcite transition zone covers the range from the saturation 3 3 horizon to the * CO2~ at the CCD. The calcite saturation horizon was estimated by the intercept of seawater carbonate ion content and the concentration of 3 carbonate ions in equilibrium with seawater (saturation state: * CO2~"0 lmol/kg) for calcite mineral phase. Below: * CO2~ of the water column versus water 3 3 depth below 2000 m. Applying several methods, the positions of the calcite saturation horizon and the sedimentary lysocline as well as the extent of the calcite transition zone were calculated. * CO2~ values refer to the GEOSECS locations 48, 103, and 109, whereas the calcium carbonate data come from GeoB sediment 3 surface samples.
N. Dittert, R. Henrich / Deep-Sea Research I 47 (2000) 603}620 613
614
N. Dittert, R. Henrich / Deep-Sea Research I 47 (2000) 603}620
surface sediments of the upper open ocean from 2000 to 4300 m, which corresponds to a carbonate ion undersaturation state D CO2~*!5 lmol/kg. The second cluster 3 contains dissolution stages 2.5}5 and indicates the deep open ocean from 4600 to 5300 m water depth, which corresponds to a carbonate ion undersaturation state D CO2~(!5 lmol/kg. Between these two clusters, where dissolution becomes rap3 idly stronger until even fragments are completely dissolved, we locate the border between upper and deep open ocean, namely the sedimentary lysocline. This can also be shown by the decrease in the total number of planktic foraminifera per gram sediment (Dittert et al., 1999). The third cluster includes dissolution stages 0.9}2.4 and is found in sediments of the upwelling in#uenced continental margin from 1000 to 3900 m water depth, which corresponds to a carbonate ion undersaturation state D CO2~*8 lmol/kg. This area is re#ected by high productivity within the Benguela 3 system, where signi"cantly higher amounts of organic carbon than in the open ocean are delivered to the depth. Metabolic activities of benthic organisms lead to respiration and remineralization and subsequently to CO enrichment at the sedi2 ment}water interface. This metabolically driven carbonate dissolution, which was earlier described as supralysoclinal dissolution (Emerson and Bender, 1981; Jahnke et al., 1994; Freiwald, 1995; Martin and Sayles, 1996), results in a higher basic dissolution potential. The dissolution stages R }R (reticulate morphotype) and C }C (crystalline 1 5 1 5 morphotype) based on our G. bulloides examinations are comparable to those of G. bulloides (`no dissolutiona to `severe dissolutiona) investigated by Van Kreveld (1996) and to those of N. pachyderma (D `no dissolutiona to D `severe 0 4 dissolutiona) introduced by Henrich et al. (1989) and Henrich (1989). Similar to our examinations on G. bulloides, sediment investigations on G. bulloides (Van Kreveld, 1996) and laboratory experiments on Globigerinoides trilobus (Hecht et al., 1975) show that the initial dissolution "rst reduces spines and a!ects spine bases (Plate 1). Other micropaleontological approaches undertaken to reconstruct the sedimentary lysocline depth include (1) percent fragmented planktic foraminiferal tests (e.g., Keigwin, 1976; Le and Shackleton, 1992); (2) proportions of solution-susceptible and solutionresistant planktic foraminifera species (e.g., Schott, 1935; Berger, 1979; Boltovskoy and Totah, 1992); (3) the ratio of one organism group to another (e.g., Kennett, 1966; Hay, 1970; Parker, 1971; Berner, 1977; Peterson and Prell, 1985). Actually, each of these parameters is a potential dissolution index, but their variations are controlled by ecological rather than by geochemical factors (see compilation by Dittert et al., 1999). Berger et al. (1982) developed the `elbowa model, which illustrates the rapidly increasing dissolution at the lysocline based on the relationship between foraminiferal preservation states and amount of calcite dissolution as a function of water depth. With our BDX tool we are able to expand Berger's `elbowa model and may distinguish the upper part into the upper open ocean and the upwelling in#uenced continental margin realm. At the bend we position the sedimentary lysocline or the top of the calcite transition zone, respectively, and dissolution stage 5 corresponds to the CCD (Figs. 3A and 5).
N. Dittert, R. Henrich / Deep-Sea Research I 47 (2000) 603}620
615
4.2. Geochemical parameters In order to assess the BDX tool we sought methods, to corroborate our results. By means of the carbonate ion content of the water column (Takahashi et al., 1980) and the calcium carbonate content of the sediment surface samples our conclusions were con"rmed. The carbonate ion content at GEOSECS station 48 (Brazil Basin) ranges from 235 lmol/kg at the surface to 77 lmol/kg at 5075 m water depth. At GEOSECS station 109 (Guinea Basin) the values decrease from 247 lmol/kg at the surface to 90 lmol/kg at 5132 m. At GEOSECS station 103 (western Cape Basin) the carbonate ion content drops from 219 lmol/kg at the surface to 84 lmol/kg at 4572 m. According to the equation introduced by Broecker and Takahashi (1978) the calcite saturation in the water column rises from 48 lmol/kg at the surface to 114 lmol/kg at 5500 m water depth. Using several geochemical approaches we were able to determine the position of the calcite saturation horizon, the sedimentary lysocline, and the CCD (Fig. 5). The calcite saturation horizon (Broecker and Takahashi (1978)) is situated highest in the Cape Basin (transects 3 and 4) at 3950 m. In the Brazil Basin it is located at 4250 m and in the Guinea Basin at 4350 m water depth (Fig. 5). The reconstruction from sedimentary calcite (Figs. 4 and 5) locates the top of the calcite transition zone 100 to 450 m below the calcite saturation horizon. The 80% CaCO isopleth (Figs. 4 and 5; Farrell and 3 Prell, 1989) varies from 4150 m (Namibia continental margin) to 4350 m (Brazil Basin) to 4400 m (western Cape Basin) to 4550 m (Guinea Basin). Since at the calcite saturation horizon D CO2~ becomes zero, above the calcite saturation horizon the 3 seawater is by de"nition supersaturated with respect to carbonate ion content. Nevertheless, an undersaturation of 10 lmol/kg is enough to dissolve almost all the calcite falling to the sea#oor (Broecker and Peng, 1982). The bottom of the calcite transition zone (Archer (1996)) varies from 4950 m (Cape Basin) to 5300 m (Brazil and Guinea Basins). The 10% CaCO isopleth (Figs. 4 and 3 5), which is a marker of the CCD (Bramlette (1961)) varies from 4950 m (Namibia continental margin) to 5300 m (Brazil Basin). That is, in the Guinea Basin and the Cape Basin the CCD is not attained. Where the carbonate content drops below 10 wt%, we can ascertain that both the geochemical and the sedimentary reconstruction lead to a corresponding depth of the bottom of the calcite transition zone. Presumably, this accordance is due to the fact that the 10% CaCO isopleth and the 3 zero intercept of the %CaCO versus D CO2~ relation coincide approximately with 3 3 the progressive dissolution increase of D CO2~ (Keir, 1980) below the top of the 3 calcite transition zone. Another clue for the assessment of the lysocline depth is the in#uence of organic carbon. Our results show that at the open ocean realm of all three South Atlantic basins the amount of sedimentary C is below 1 wt% (Fig. 4). However, the carbonate/carbon 03' weight percentage ratio increases rapidly at distinct water depths. Where the values exceed 0.004, the lysocline (Farrell and Prell (1989)) is attained exactly (Fig. 4). At the continental margin it will be noticed immediately that even in lower water depths high amounts of organic carbon (up to 5.69 wt%) are present, which is explained by the enormous productivity of the Benguela upwelling system. The high carbonate/carbon
616
N. Dittert, R. Henrich / Deep-Sea Research I 47 (2000) 603}620
weight percentage ratio (up to 0.067) and the smooth break of carbonate content (down to 71 wt%) at water depths of 350 m down to 1400 m lead us to assume a certain amount of supralysoclinal dissolution (Emerson and Bender, 1981; Jahnke et al., 1994). 4.3. Comparison of Globigerina bulloides dissolution index and geochemical parameters As we can show by the increase of the BDX values above the lysocline (Fig. 3), carbonate dissolution obviously a!ects the G. bulloides ultrastructure even at supersaturation, as stated by Broecker and Peng (1982). According to BDX values the lysocline is attained only by transect 1 (Brazil Basin) and transect 3 (western Cape Basin), which is documented by a rapid increase of BDX values. In the Brazil Basin and the western Cape Basin, our lysocline reconstruction through BDX matches the approaches by Farrell and Prell (1989). In the Brazil Basin, BDX values increase between 3977 m (GeoB 1117-3) and 4675 m (GeoB 1118-2) by leaps and bounds, which covers the calculated sedimentary lysocline depth of 4350 m (according to Farrell and Prell, 1989). In the western Cape Basin, BDX values increase between 4089 m (GeoB 1211-1) and 4669 m (GeoB 1212-2), which covers the calculated depth of 4400 m (according to Farrell and Prell, 1989). However, the exact determination of the lysocline depth through BDX is not possible due to the low sample density. Nevertheless, dissolution even becomes evident at supersaturation (Fig. 3). We attribute the o!set between the calcite saturation horizon and the sedimentary lysocline to the fact that dissolution occurs mostly at the sediment}water interface rather than in the water column. That is, at this interface [CO2~] is very likely to be 3 higher than at the equivalent water depth in the water column (Le and Shackleton, 1992). However, that is not necessarily the case, since oxic metabolism can make the carbonate ion content of pore waters drop below that of the overlaying bottom water. Of course, variations in the percentage of CaCO in a single sample cannot be simply 3 interpreted as an index of preservation, because the relative abundance of carbonate is controlled by the balance of productivity over dissolution and by dilution due to the in#ux of non-carbonate sedimentary components. Only in the ideal situation, where the rain rate of calcitic and non-calcitic material is constant in space and time could the amount of calcite lost to dissolution be calculated from the percentage of CaCO 3 in the sediment. Otherwise, the calcite content of the sediment from above the lysocline has to be used as the reference for the amount of dissolution, that has occurred in samples from the transition zone (the `depth-transect approacha; e.g. Farrell and Prell, 1989; Curry and Lohmann, 1990; Bickert et al., 1997; Dittert et al., 1999). However, the fact that CaCO contents of supralysoclinal sediment average about 3 90% in the world ocean (Archer, 1996) raises the problem that quite large amounts of dissolution create only very small changes in the carbonate content. According to BDX values the CCD is attained only by transect 1 (Brazil Basin), which is shown by dissolution stage 5 at about 5213 m (Fig. 3). Our reconstruction "ts the approaches by Archer (1996) and Bramlette (1961), which locate the CCD at about 5300 m (Fig. 5). Regarding the top and the bottom of the sedimentary calcite transition zone, all reconstructions show a consistent picture, which we attribute to the bottom water
N. Dittert, R. Henrich / Deep-Sea Research I 47 (2000) 603}620
617
distribution. While AABW } as a composite of Weddell Sea Deep Water and Lower Circumantarctic Deep Water } is moving more or less unhampered northwards within the Brazil Basin, it is detained within the Cape and the Guinea Basin by the Mid Atlantic Ridge and the Walvis Ridge (Fig. 1). The in#ow of AABW into the Guinea Basin is restricted to only small quantities passing the sills through the Romanche Fracture Zone, the Chain Fracture Zone, and the Walvis Passage (Van Bennekom and Berger, 1984; Shannon and Chapman, 1991; Warren and Speer, 1991; Mercier et al., 1994). Consequently, the deepest parts of the Guinea Basin are "lled almost exclusively by O ameliorated, CO depleted, and carbonate ion saturated 2 2 NADW (Kroopnick, 1985). The Cape Basin, although located east of the Mid Atlantic Ridge, is dominated by AABW below 4000 m due to a bottom water passage, which allows AABW to enter the basin from the south (Reid, 1989). Therefore, lysocline depth and sublysoclinal carbonate dissolution pattern in the open ocean seem to depend on both the in#uence of the corrosive AABW and the asymmetry of its oceanographic distribution.
Acknowledgements We thank P.J. MuK ller for providing sedimentary carbon and carbonate data. M. Matthies and R. Schlotte assisted the processing of the GEOSECS data set. Discussions with colleagues (H. Meggers, S. Mulitza, H.-S. Niebler) were helpful. Comments from two anonymous reviewers improved the manuscript. We appreciate technical collaboration by H. Heilmann, R. Henning, and C. Wienberg. This investigation was funded by the German Research Foundation (SFB No. 261). This is SFB 261 contribution no. 275 at Bremen University.
References Archer, D.E., 1996. An Atlas of the distribution of calcium carbonate in sediments of the deep sea. Global Biogeochemical Cycles 10 (1), 159}174. Bainbridge, A.E., 1976. GEOSECS Atlantic Expedition, Vol. 2. Sections and Pro"les. National Science Foundation, Superintendent of Documents, US Government Printing O$ce, Washington, DC, 198pp. Bainbridge, A.E., 1981. GEOSECS Atlantic Expedition, Hydrographic Data, 1972}1973. National Science Foundation, Superintendent of Documents, US Government Printing O$ce, Washington, DC, 121pp. Baumann, K.-H., Meggers, H., 1996. Paleoceanographical change in the Labrador Sea during the last 3.1 MY: evidence from calcareous plankton records. In: Moguilevsky, A., Whatley, R. (Eds.), Microfossils and Oceanic Environments. Aberystwyth-Press, Aberystwyth, pp. 131}154. BeH , A.W.H., Morse, J.W., Harrison, S.M., 1975. Progressive dissolution and ultrastructural breakdown of planktonic foraminifera. Cushman Foundation For Foraminiferal Research Special Publication Vol. 13, 27}55. Berger, W.H., 1979. Preservation of foraminifera. In: Lipps, J.H., Berger, W.H., Buzas, M.A., Douglas, R.G., Ross, C.A. (Eds.), Foraminiferal Ecology and Palecology. Society of Economic Paleontologists and Mineralogists, Houston, pp. 105}155. Berger, W.H., Bonneau, M.-C., Parker, F.L., 1982. Foraminifera on the deep-sea #oor: lysocline and dissolution rate. Oceanologica Acta 5 (2), 249}258.
618
N. Dittert, R. Henrich / Deep-Sea Research I 47 (2000) 603}620
Berger, W.H., Fischer, K., Lai, C., Wu, G., 1987. Ocean productivity and organic carbon #ux (Part I. Overview and maps of primary production and export production). University of California, San Diego, SIO Report, Vols. 87}30, 67pp. Berger, W.H., Keir, R.S., 1984. Glacial-Holocene changes in atmospheric CO and the deep-sea record. In: 2 Hansen, J.E., Takahashi, T. (Eds.), Climate Processes and Climate Sensitivity. Geophysical Monographs, Vol. 29. Maurice Ewing Series. Washington, DC, pp. 337}351 Berner, R.A., 1977. Sedimentation and dissolution of pteropods in the ocean. In: Andersen, N.R., Malaho!, A. (Eds.), The Fate of Fossil Fuel CO in the Oceans. Plenum Press, New York, pp. 243}260. 2 Bickert, T., Cordes, R., Wefer, G., 1997. Late Pliocene to middle Pleistocene (2.6 to 1.0 Ma) carbonate dissolution in the western equatorial Atlantic : results of Leg 154, Ceara Rise. Proceedings of the Ocean Drilling Program. Scienti"c Results Vol. 154, 229}237. 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: Present and Past Circulation. Springer, Berlin, pp. 599}620. Boltovskoy, E., Totah, V.I., 1992. Preservation index and preservation potential of some foraminiferal species. Journal of Foraminiferal Research 22 (3), 267}273. Boyle, E.A., 1988. Vertical oceanic nutrient fractionation and glacial/interglacial CO cycles. Nature 331, 2 55}56. Bramlette, M.N., 1961. Pelagic sediments. In: Sears, M. (Ed.), Oceanography. AAAS Publication, New York, pp. 345}366. Broecker, W.S., 1997. Thermohaline circulation, the Achilles heel of our climate system: will man made CO upset the current balance?. Science 278, 1582}1588. 2 Broecker, W.S., Denton, G.H., 1989. The role of ocean}atmosphere reorganizations in glacial cycles. Geochimica et Cosmochimica Acta 53, 2465}2501. Broecker, W.S., Peng, T.H., 1982. Tracers in the Sea. Eldigio Press, New York, pp. 689. Broecker, W.S., Takahashi, T., 1978. The relationship between lysocline depth and in situ carbonate ion concentration. Deep-Sea Research 25F (1), 65}95. CLIMAP Project members, 1976. The surface of the ice-age earth. Science 191, 1131}1137. Curry, W.B., Lohmann, G.P., 1990. Reconstructing past particle #uxes in the tropical Atlantic Ocean. Paleoceanography 5, 487}506. d'Orbigny, A., 1826. Tableau meH thodique de la classe des CeH phalopodes. Annales des Sciences Naturelles 7, 245}314. Dittert, N., Baumann, K.-H., Bickert, T., Henrich, R., Huber, R., Kinkel, H., Meggers, H., 1999. Carbonate dissolution in the deep-sea: methods, quanti"cation and paleoceanographic application. In: Fischer, G., Wefer, G. (Eds.), Use of Proxies in Paleoceanography * Examples from the South Atlantic. Springer, Berlin, pp. 255}284. Emerson, S., Bender, M., 1981. Carbon #uxes at the sediment-water interface of the deep-sea: calcium carbonate preservation. Journal of Marine Research 39, 139}162. Farrell, J.W., Prell, W.L., 1989. Climatic change and CaCO preservation: an 800,000 year bathymetric 3 reconstruction from the central equatorial Paci"c Ocean. Paleoceanography 4 (4), 447}466. Freiwald, A., 1995. Bacteria-induced carbonate degradation: a taphonomic case study of Cibicides lobolatus from a high-boreal carbonate setting. Palaios 10, 337}346. Hay, W.W., 1970. Calcareous nannofossils from cores recovered on Leg 4. Initial Report Deep Sea Drilling Project Vol. 4, 455}501. Hecht, A.D., Eslinger, E.V., Garmon, L.B., 1975. Experimental studies on the dissolution of planktonic foraminifera. Cushman Foundation For Foraminiferal Research Special Publication Vol. 13, 56}96. Hemleben, C., Spindler, M., Anderson, O. R., 1989. Modern Planktonic Foraminifera. Springer, New York, 363pp. Henrich, R., 1989. Glacial/interglacial cycles in the Norwegian Sea : sedimentology, paleoceanography, and evolution of late Pliocene to Quaternary northern hemisphere climate. Proceedings of the Ocean Drilling Program. Scienti"c Results Vol. 104, 189}232. Henrich, R., Kassens, H., Vogelsang, E., Thiede, J., 1989. Sedimentary facies of glacial}interglacial cycles in the Norwegian Sea during the last 350 ka. Marine Geology 86, 283}319.
N. Dittert, R. Henrich / Deep-Sea Research I 47 (2000) 603}620
619
Imbrie, J., Kipp, N.G., 1971. A new micropaleontological method for quantitative paleoclimatology: application to a Late Pleistocene Carribean core. In: Turekian, K.K. (Ed.), Late Cenocoic Glacial Ages. Yale University Press, New Haven, pp. 71}181. Ingle, S.E., Culberson, C.H., Hawley, J.E., Pytkowicz, R.M., 1973. The solubility of calcite in seawater at atmospheric pressure and 35% salinity. Marine Chemistry 1, 295}307. Jahnke, R.A., Craven, D.B., Gaillard, J.F., 1994. The in#uence of organic matter diagenesis on CaCO 3 dissolution at the deep-sea #oor. Geochimica et Cosmochimica Acta 58, 2799}2809. Keigwin, L.D., 1976. Late Cenozoic planktonic foraminiferal biostratigraphy and paleoceanography of the Panama Basin. Micropaleontology 22, 419}422. Keir, R.S., 1980. The dissolution kinetics of biogenic calcium carbonates in seawater. Geochimica et Cosmochimica Acta 44, 241}252. Kennett, J.P., 1966. Foraminiferal evidence of a shallow calcium carbonate solution boundary, Ross Sea, Antarctica. Science 153, 191}193. Kroopnick, P.M., 1985. The distribution of 13C of z CO in the World Oceans. Deep-Sea Research 32 (1), 2 57}84. Le, J., Shackleton, N.J., 1992. Carbonate dissolution #uctuations in the western equatorial Paci"c during the Late Quaternary. Paleoceanography 7, 21}42. Martin, W.R., Sayles, F.L., 1996. CaCO dissolution in sediments of the Ceara Rise, western equatorial 3 Atlantic. Geochimica et Cosmochimica Acta 60 (2), 243}263. Mehrbach, C., Culberson, C.H., Hawley, J.E., Pytkowicz, R.M., 1973. Measurement of the apparent dissociation constants of carbonic acid in seawater at atmospheric pressure. Limnological Oceanography 18, 892}907. Mercier, H., Speer, K.G., Honnorez, J., 1994. Tracing the Antarctic Bottom Water through the Romanche and Chain Fracture Zones. Deep-Sea Research I 41, 1457}1477. Millero, F.J., 1979. The thermodynamics of the carbonate system in seawater. Geochimica et Cosmochimica Acta 43, 1651}1661. Morse, J.W., Mucci, A., Millero, F.J., 1980. The solubility of calcite and aragonite in seawater of 35% salinity at 253C and atmospheric pressure. Geochimica et Cosmochimica Acta 44, 85}94. Parker, F.L., 1971. Distribution of planktonic foraminifera in recent deep-sea sediments. In: Funnell, B.M., Riedel, W.R. (Eds.), The Micropaleontology of oceans. Cambridge University Press, Cambridge, pp. 289}307. 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. Marine Geology 64, 259}290. Reid, J.L., 1989. On the total geostrophic circulation of the South Atlantic Ocean: #ow patterns, tracers and transports. Progress in Oceanography 23, 149}244. Reid, J.R., 1996. On the circulation of the South Atlantic. In: Wefer, G., Berger, W.H., Siedler, G., Webb, D.J. (Eds.), The South Atlantic: present and Past Circulation. Springer, Berlin, pp. 13}44. Schott, W., 1935. Die Foraminiferen in dem aK quatorialen Teil des Atlantischen Ozeans. Wissenschaftliche Ergebnisse der Deutschen Atlantik Expedition Meteor 1925}1927, Vol. 3, 135pp. Schulz, H.D., cruise participants, 1992. Bericht und erste Ergebnisse uK ber die Meteor-Fahrt M20/2, Abidjan-Dakar, 27.12.1991}3.2.1992. Berichte, Fachbereich Geowissenschaften, UniversitaK t Bremen, Vol. 25, 173pp. Shannon, L.V., Chapman, P., 1991. Evidence of Antarctic Bottom Water in the Angola Basin at 323S. Deep-Sea Research 38, 1299}1304. Siedler, G., MuK ller, T.J., Onken, R., Arhan, M., Mercier, H., King, B.A., Saunders, P.M., 1996. The zonal WOCE sections in the South Atlantic. In: Wefer, G., Berger, W.H., Siedler, G., Webb, D.J. (Eds.), The South Atlantic: Present and Past Circulation. Springer, Berlin, pp. 83}104. Takahashi, T., Broecker, W. S., Bainbridge, A. E., Weiss, R. F., 1980. Carbonate chemistry of the Atlantic, Paci"c and Indian Oceans: the results of the GEOSECS expeditions 1972}1978. Lamont-Doherty Geological Observatory, Palisades, NY, 211pp. Van Bennekom, A.J., Berger, G.W., 1984. Hydrography and silica budget of the Angola Basin. Netherlands Journal of Sea Research 17, 149}200.
620
N. Dittert, R. Henrich / Deep-Sea Research I 47 (2000) 603}620
Van Kreveld, S.A., 1996. Calcium carbonate dissolution indices on the northeast Atlantic sea #oor. In: Van Kreveld-Alfane, S.A. (Ed.), Late Quaternary Sediment Records of Mid-Latitude Northeast Atlantic Calcium Carbonate Production and Dissolution. FEBO B.V., Enschede, pp. 135}184. Warren, B.A., Speer, K.G., 1991. Deep circulation in the eastern South Atlantic Ocean. Deep-Sea Research 38, 281}322. Wefer, G., Bleil, U., cruise participants, 1990. Bericht uK ber die Meteor-Fahrt M12/1, Kapstadt-Funchal, 13.3.1990}14.4.1990. Berichte, Fachbereich Geowissenschaften, UniversitaK t Bremen, Vol. 11, 66pp. Wefer, G., Bleil, U., Schulz, H. D., cruise participants, 1989. Bericht uK ber die Meteor-Fahrt M9-4, Dakar-Santa Cruz, 19.2.}16.3.1989. Berichte, Fachbereich Geowissenschaften, UniversitaK t Bremen, Vol. 7, 103pp.