6.16 Tracers of Past Ocean Circulation J. Lynch-Stieglitz Columbia University, Palisades, NY, USA

6.16.1 INTRODUCTION 433 6.16.2 NUTRIENT WATER MASS TRACERS 434 6.16.2.1 Carbon Isotopes 434 6.16.2.1.1 Controls on d13C of oceanic carbon 434 6.16.2.1.2 Carbon isotope ratios in benthic foraminifera 435 6.16.2.1.3 Cadmium in benthic foraminifera 437 6.16.2.1.4 Cadmium in seawater 437 6.16.2.1.5 Cd/Ca in foraminifera 438 6.16.2.2 Ba/Ca 439 6.16.2.3 Zn/Ca 439 6.16.3 CONSERVATIVE WATER MASS TRACERS 439 6.16.3.1 Mg/Ca in Benthic Foraminifera 439 6.16.3.2 Pore-water Chemistry 439 6.16.3.3 Oxygen Isotopes in Benthic Foraminifera 440 6.16.3.4 Carbon Isotope Air–Sea Exchange Signature 441 6.16.4 NEODYMIUM ISOTOPE RATIOS 441 6.16.5 CIRCULATION RATE TRACERS 441 6.16.5.1 Radiocarbon 441 6.16.5.2 231Pa/ 230Th Ratio 442 6.16.5.3 Geostrophic Shear Estimates from d18O in Benthic Foraminifera 442 6.16.6 NON-GEOCHEMICAL TRACERS OF PAST OCEAN CIRCULATION 443 6.16.7 OCEAN CIRCULATION DURING THE LAST GLACIAL MAXIMUM 443 6.16.8 CONCLUSIONS 447 ACKNOWLEDGMENTS 448 REFERENCES 448

6.16.1 INTRODUCTION flow patterns and mixing of subsurface water masses. Information about how the ocean circulated Several decades ago it was realized that during the past is useful in understanding changes chemistry of the shells of benthic foraminifera in ocean and atmospheric chemistry, changes in (carbon isotope and Cd/Ca ratios) carried an the fluxes of heat and freshwater between the imprint of the nutrient content of deep-water ocean and atmosphere, and changes in global wind masses (Shackleton, 1977; Broecker, 1982; Boyle, patterns. The circulation of surface waters in the 1981). This led rapidly to the recognition that the ocean leaves an imprint on sea surface tempera- water masses in the Atlantic Ocean were arrayed ture, and is also inextricably linked to the patterns differently during the last glacial maximum than of oceanic productivity. Much valuable infor- they are today, and the hypothesis that the glacial mation about past ocean circulation has been arrangement reflected a diminished contribution inferred from reconstructions of surface ocean of low-nutrient North Atlantic deep water temperature and productivity, which are covered (NADW) (Curry and Lohmann, 1982; Boyle and in separate chapters. Here the focus is on the Keigwin, 1982). More detailed spatial reconstruc- geochemical tracers that are used to infer the tions indicated a shallow nutrient-depleted water

433 434 Tracers of Past Ocean Circulation mass overlying a more nutrient-rich water mass in nutrient tracer reflects the increased contribution the glacial Atlantic. These findings spurred from the regeneration of organic material as the advances not only in geochemistry but in water mass ages, it also carries a signature of the oceanography and climatology, as workers in initial nutrient content of the source regions. In these fields attempted to simulate the inferred today’s Atlantic, the contrast between high- glacial circulation patterns and assess the nutrient Southern Ocean source water and low- vulnerability of the modern ocean circulation to nutrient NADW is strong, and the residence time of changes such as observed for the last ice age. deep water in the Atlantic Ocean is short, allowing While the nutrient distributions in the glacial little additional accumulation of nutrients from Atlantic Ocean were consistent with a diminished sinking organic material. In this case the nutrient flow of NADW, they also could have reflected an tracers can be used to document the relative increase in inflow from the South Atlantic and/or a mixtures of Northern and Southern water masses. shallower yet undiminished deep-water mass.

Clearly, tracers capable of giving information on 13 deep-water flow rate, rather than nutrient content 6.16.2.1.1 Controls on d C of oceanic alone, were needed. Differences between surface carbon water (measured on planktonic foraminifera) and During photosynthesis, organisms preferen- deep-water (measured on coexisting benthic tially take up the lighter isotope of carbon (12C) foraminifera) radiocarbon concentrations pro- increasing surface ocean d13C. When the 13C- vided the first rate constraint (Broecker et al., depleted organic matter is decomposed, this CO2 1988; Shackleton et al., 1988). Reduced amounts release decreases the d13C of seawater DIC at of protactinium relative to the more particle- depth (Deuser and Hunt, 1969; Craig, 1970; reactive thorium in the glacial Atlantic suggested Kroopnick, 1985). If this process were the only that deep water was exported from the Atlantic one responsible for the distribution of d13Cin during glacial times (Yu et al., 1996). More the ocean, d13C of oceanic carbon would recently, density gradients in upper waters have decrease by ,1.1‰ for every 1 mmol kg21 been used to infer changes in the upper ocean increase in oceanic PO4 (Broecker and Maier- return flow that compensates the deep-water Reimer, 1992). Reflecting this biological cycling, export (Lynch-Stieglitz et al., 1999b). as a deep water mass with a homogeneous source Many of these tracers of paleo-ocean flow have ages (such as in the deep Indian and Pacific been applied to all of the ocean basins, and have 13 oceans) the d CandPO4 increase in this been extended in time throughout the Neogene. biologically expected ratio (Figure 1). However, Despite this progress, a consistent picture of the the discrimination against 13C during photosyn- circulation of the ocean during even the last ice thesis increases from about 19‰ in the warm age has yet to emerge. While circulation tracers surface ocean to about 30‰ in the Antarctic, suggest a rearrangement of water masses in the 13 which means that the PO4 and d C changes Atlantic, there is still considerable disagreement associated with the uptake and regeneration of about the water masses and circulation in the rest organic material can vary regionally. of the World Ocean. Some of this arises from still Carbon isotope fractionation during air–sea insufficient data coverage, but some is the result of exchange is also an important factor in determin- conflicting information from the various deep ing the isotopic composition of seawater. If the circulation tracers. In this chapter, we examine in CO2 in the atmosphere were in isotopic equili- more detail the methods used to reconstruct past brium with the dissolved inorganic carbon (DIC) ocean circulation, which will illuminate the source in the ocean, the DIC would be enriched in 13C of some of this conflicting information. relative to the atmospheric CO2 by ,8‰ at 20 8C (Inoue and Sugimura, 1985). This equilibrium fractionation depends on the temperature of 6.16.2 NUTRIENT WATER MASS TRACERS equilibration, with the oceanic carbon more 6.16.2.1 Carbon Isotopes enriched relative to the atmospheric value by about 0.1‰ per degree of cooling (Mook et al., Measuring the carbon isotope ratio in the calcite 1974). If the surface ocean were everywhere in tests of bottom dwelling (benthic) foraminifera is isotopic equilibrium with atmospheric CO2, perhaps the most widespread method for recon- the 30 8C range in ocean temperatures would structing the distribution and properties of deep- cause a 3‰ range in oceanic d13C with higher water masses. The carbon isotope ratios in the values where surface temperatures are cold, and foraminifera reflect that ratio in seawater. In turn, lower values in warm surface waters. This range is the carbon isotope ratio in the deep sea is primarily similar to the magnitude of d13C change induced controlled by the regeneration of 13C-poor organic by biological processes. However, isotopic equili- material and has a distribution much like a nutrient brium is rarely reached, because the timescale for such as phosphate in the modern ocean. While a carbon isotope equilibration between the ocean Nutrient Water Mass Tracers 435

1.5 d 13 C d 13 as =0 C as NADW =0.5

1.0 WSBW

CDCDWW 0.5 C (‰ PDB) 13 d Biology

0.0

d 13 C as =–0.5 Air–sea exchange –0.5 1.0 1.5 2.0 2.5 3.0 µ –1 PO4 ( mol kg )

13 Figure 1 Deep water (.2,000 m) d C and PO4 data from the modern ocean after Lynch-Stieglitz et al. (1995). The 13 13 lines indicate a “Redfield slope” and d Cas indicates the deviation from the mean ocean Redfield slope for d C and PO4 (Broecker and Maier-Reimer, 1992). While the deep Indian and Pacific values follow the slope expected for biologic processes, deep waters which form in the North Atlantic and Southern Ocean have, respectively, lower and higher d13C due to air–sea exchange.

13 and atmosphere (,10 years for a 50 m deep mixed By definition, water with d Cas of 0‰ has the layer) is about 10 times longer than chemical same air–sea exchange signature as mean ocean 13 equilibration between atmospheric CO2 and DIC deep water. A positive value of d Cas means that 13 (Broecker and Peng, 1974; Tans, 1980). This the water has a higher d Cas than mean ocean deep isotopic equilibration time is significantly longer water (more of an influence of air–sea exchange at than surface waters stay in one place in the ocean. cold temperatures/less at warm temperatures), 13 The exchange of atmospheric CO2 and surface where as a negative d Cas value implies less d13 ocean aqueous CO2 has the potential to leave C enrichment due to air–sea exchange than for 13 mean ocean deep water. oceanic carbon depleted in C in areas of CO2 13 d Cas in surface waters is highest in the invasion (e.g., high northern latitudes, low-pCO2 surface waters), and enriched in 13C in areas of the Antarctic and the North Pacific due to exchange with the atmosphere at low temperatures. The ocean where CO2 is outgassed (e.g., equatorial 13 lowest values of d Cas are found in the centers of upwelling regions, high-pCO2 surface waters) (Lynch-Stieglitz et al., 1995). This is because the subtropical gyres. Surface waters circulate in the CO gas that is dissolved in the ocean and these anticyclonic gyres long enough to allow the 2 oceanic dissolved inorganic carbon (DIC) to exchanges between the ocean and atmosphere is equilibrate isotopically with the atmosphere at isotopically lighter than the much larger pool of high temperatures despite the relatively low rates the DIC that consists mainly of bicarbonate and of gas exchange. The situation in the North Atlantic carbonate ions. This effect is observed in the seems slightly more complex. Air–sea exchange at ocean, because the timescale for carbon equili- 13 cold temperatures will tend to raise the d Cas of brium between the ocean and atmosphere (which the northward flowing upper waters, but does not drives the air–sea exchange of the CO2) is shorter have time to overcome the effects of CO2 invasion than the timescale for isotopic equilibrium. 13 13 and the remnant low-d Cas of the northward In general, surface-ocean d C lies somewhere flowing surface water. Deep waters leaving the in between the value predicted if biological cycling surface in the North Atlantic and the Antarctic alone controlled the distribution of d13C in the 13 reflect the low and high d Cas, respectively. ocean and the value which would be at equilibrium with the atmosphere (Broecker and Peng, 1982; 6.16.2.1.2 Carbon isotope ratios in benthic Broecker and Maier-Reimer, 1992). The departure foraminifera from the biologically expected value can be expressed as an air–sea exchange signature Belanger et al. (1981) demonstrated that the 13 13 13 (d Cas ¼d C2(2.721.1£ PO4)) (Broecker and d C in the benthic foraminifera Planulina Maier-Reimer, 1992; Lynch-Stieglitz et al., 1995). wuellerstorfi decreased from highest values in 436 Tracers of Past Ocean Circulation the Norwegian-Greenland Sea, to lower values in time of this study, there was little quality water the South Atlantic and lowest values on the East column d13C data with which to compare the core Pacific Rise. They argued that P. wuellerstorfi top values. However, in subsequent years the more were reliable recorders of bottom-water condi- extensive and improved data sets have confirmed tions and not overly influenced by the incorpor- that, on average, Planulina and Cibicidoides d13C ation of metabolic carbon. (While wuellerstorfi is reflect water column d13C with very little or no variously assigned to the genera Planulina, offset. Some of the scatter in Figure 2 may be Cibicides, Cibicidoides, and Fontbotia by various attributed to errors in estimated oceanic d13Cor workers, there is no disagreement on the assign- core tops which are not Holocene in age. However, ment of the species.) They also found that d13Cin the more recent data sets also show that there can the genera Oridorsalis and Pyrgo did not follow be significant offsets between the foraminifera and the expected deep-water progression in d13C and water column values at some locations even when argued that the shell chemistry reflected the lower living foraminifera are compared with water d13C of their pore-water habitat. The same year, column d13C collected at the same location. Graham et al. (1981) showed, with a larger set Perhaps it is not unexpected that benthic of core top data, that the d13C from the tests foraminifera do not always cleanly record the P. wuellerstorfi and Cibicidoides kullenbergi also isotopic composition of bottom water. The fact followed the expected relationship with seawater that all of the benthic foraminifera have d13C that chemistry. Graham et al. (1981) also found that is significantly lighter than what would be these species show d13C values that are closest to expected for equilibrium with bottom water what would be expected for precipitation provides the first clue that additional processes in seawater. While they found that the d13C may be at work. As postulated by Belanger et al. values in the P. wuellerstorfi and C. kullenbergi (1981), some foraminifera may live in the pore reflect the d13C of the ambient water, they are a waters where d13C of the DIC is lower than in the full 1‰ lower than expected for calcite formed in overlying water column. Even so, the carbon equilibrium with this water (Woodruff et al., isotopic composition of the calcite skeleton of 1980). However, carbon isotope equilibrium most calcitic organisms does not reflect the between calcite and dissolved carbon is not well composition of the ambient waters (Wefer and known at the temperatures of the deep ocean Berger, 1991) but is dominated by physiological (Grossman, 1984a; Romanek et al., 1992). These or “vital” effects which usually involve the studies paved the way for the widespread incorporation of isotopically light metabolic 13 reconstruction of deep-water d C that followed carbon or kinetic fractionation during the transport soon afterward (Curry and Lohmann, 1982; Boyle of carbon to the calcification site or during the and Keigwin, 1982; Duplessy et al.,1984, 1988; calcification itself. The isotopic ratios of Curry et al., 1988). planktonic foraminifera are also dominated by Shown in Figure 2 is a comparison of Planulina these vital effects and are, in general, not in 13 and Cibicidoides d C with estimated water equilibrium with the seawater (e.g., Spero and 13 column d C from Duplessy et al. (1984). At the Lea, 1996; Spero et al., 1997). A more detailed picture at how benthic 1.60 foraminifera record bottom-water d13C has been PDB)

gained through careful analysis of stained (live, or

‰ recently living) benthic foraminifera collected 1.20 from surface sediments combined with measure- ments of both water column and pore-water d13C 0.80 (Grossman, 1984a,b; Mackensen and Douglas, 1989; McCorkle et al., 1990, 1997; Mackensen et al., 1993; Rathburn et al., 1996; Tachikawa and 0.40 Elderfield, 2002). These studies confirm earlier notions that different species show different 0.00 microhabitat preferences, with some species (including most Cibicidoides and Planulina)

C reconstructed from foraminifera ( living on or near the sediment surface and others 13

d –0.40 (including Uvigerina) living within the sediments. –0.40 0.00 0.40 0.80 1.20 1.60 The deeper-dwelling species show depleted d13C Estimated seawater d13C (‰ PDB) values relative to bottom waters that, in general, Figure 2 Carbon isotopic composition of core top reflect the magnitude of the pore-water depletions. benthic foraminifera in the genera Cibicidoides and These studies suggest that this microhabitat effect Planulina versus estimated carbon isotopic composition is the dominant control on benthic foraminiferal 13 of DIC in the water column above core location (after d C, and the variations in the vital effects Duplessy et al., 1984). between species are small. While the number of Nutrient Water Mass Tracers 437 Cibicidoides analyzed has been small, the Cibici- the most enriched concentrations in deep waters doides species tend to have the smallest isotopic (Boyle et al., 1976). Many high-quality cadmium depletion relative to bottom waters. However, the measurements in seawater since the early 1980s isotopic depletion even for Cibicidoides can be have confirmed this nutrient-like behavior, with significant, especially in areas where productivity, cadmium showing an almost linear relationship and thus the supply of organic material to the with phosphorus in seawater. While it is still not sediments, is high. Mackensen et al. (1993) clear whether cadmium is taken up as an essential showed that living and dead specimens of nutrient or inadvertently, the cycling within the P. wuellerstorfi showed substantial depletions ocean is dominated by the incorporation into relative to bottom waters, particularly beneath organic tissue in the surface ocean, and the high-productivity fronts in the Southern Ocean. regeneration of elemental cadmium as this While P. wuellerstorfi has been reported to live material breaks down below the surface. The above the sediment surface (Lutze and Thiel, relationship with phosphorus is not quite linear, 1989), Mackensen et al. (1993) postulate that the with a lower slope at lower phosphorus values foraminifera are calcifying within the organic fluff (Figure 3). This structure may be due to the fact layer that can be found in high-productivity 13 that as Cd/P in seawater decreases so will Cd/P in environments, with their low-d Cvalue organic matter (Elderfield and Rickaby, 2000). reflecting either the low-d13C environment within That is, a fixed (“Redfield”) ratio of Cd/P in the fluff layer or high growth rates of the organic matter should not be expected. Instead, if foraminifera within these environments. the cadmium is incorporated actively by organ- These relationships have been generally con- 13 isms, the ecosystem finds a way to manage with firmed with down-core d C studies. These studies show that the d13CfromUvigerina less of this “nutrient,” or less cadmium is (infaunal) is always more negative than for incorporated inadvertently into particles, simply Cibicidoides, and that this offset is variable in because there is less. Intermediate nutrient values space and time, consistent with the microhabitat can result from either mixing of a very low and model. While many early isotopic studies were high nutrient water mass (which would produce a performed using Uvigerina because it was felt linear Cd–P curve that passes through the origin that it best reflected oxygen isotope equilibrium, and the Cd/P of deep water), or through the most deep-water d13C reconstructions now use progressive drawdown of nutrients in surface Cibicidoides or Planulina species. It is broadly water (which would produce a curved relation- acknowledged that these records, particularly in ship). As a consequence, there are real differences high-productivity areas, might be influenced in the Cd–P relationships in different ocean by a microhabitat effect. However, by avoiding regions (Elderfield and Rickaby, 2000), which high-productivity areas and using supplemental may of course change through time. Because information about changes in bottom-water characteristics and overlying productivity, this method has continued to yield regionally coher- 1.4 ent and sensible patterns of deep-water d13C. 1.2 6.16.2.1.3 Cadmium in benthic foraminifera 1 ) –1 The use of Cd/Ca measurements in the tests of 0.8 benthic foraminifera for deep-water nutrient

reconstructions was developed in parallel with (nmol kg 0.6 the carbon isotope method. Cadmium concen- SW trations in seawater follow a nutrient-like distri- Cd 0.4 bution, while calcium concentrations simply reflect variations in salinity. The benthic forami- 0.2 nifera incorporate the cadmium and calcium into their shells in proportion to their presence in 0 seawater, which allows for the reconstruction of 0.0 0.8 1.6 2.4 3.2 µ –1 deep-water cadmium (and thus macronutrient) PSW ( mol kg ) distributions. Figure 3 Seawater cadmium measurements from the modern ocean plotted versus PO4 (after Elderfield and 6.16.2.1.4 Cadmium in seawater Rickaby, 2000). The curvature in this relationship is modeled by Elderfield and Rickaby (2000) (solid line) as Cadmium in seawater has a nutrient-like profile, resulting from the preferential uptake of cadmium in with depleted values in warm surface waters, and marine organic matter. 438 Tracers of Past Ocean Circulation cadmium is removed from seawater via CdS that for aragonitic foraminifera Hoeglundina precipitation in suboxic sediments, the global elegans D ¼ 1; which shows no depth dependence. seawater cadmium inventory, and thus the global McCorkle et al. (1995) observed an unexpected Cd–P relationship can change as the extent of decrease in the apparent distribution coefficient of suboxic sediments changes (Rosenthal et al., cadmium and other trace metals with depth 1995; van Geen et al., 1995). The residence time between 2.5 km and 4 km in the deep Pacific, for cadmium and phosphorous in seawater is which suggested dissolution-driven loss of cad- sufficiently long that the cadmium concentration mium from the foraminiferal calcite after burial. in the last glacial is unlikely to be lower than today Elderfield et al. (1996) favor the idea that the lower by more than 5% (Boyle and Rosenthal, 1996), D in deep Pacific waters may be a calcification allowing us to make meaningful inferences about response to the low carbonate saturation state the cycling of the major nutrients via reconstruc- rather than representing post-depositional dis- tions of the cadmium concentration in seawater. solution, a possibility also acknowledged by McCorkle et al. (1995). McCorkle et al. (1995) also observe a small decrease in the d13C content of foraminifera with depth, but this decrease is well 6.16.2.1.5 Cd/Ca in foraminifera 13 within the other uncertainties related to the d C Boyle (1981) showed that with careful attention tracer. There has been no systematic attempt to to cleaning, benthic foraminifera shells reflect the correct paleoceanographic cadmium reconstruc- concentration of some trace metals (including tions for this apparent dissolution (or low-carbon- cadmium) in seawater, with a preferential incor- ate saturation state) effect. 13 poration of cadmium into foraminifera. The To the extent that the d C values of infaunal 13 empirical distribution coefficient (D,ratioof benthic foraminifera reflect the low d C of pore Cd/Ca in foraminifera to Cd/Ca in seawater) was waters, it would be expected that the Cd/Ca will shown to be 2.9 (Boyle, 1988). More core top also reflect the interstitial, not bottom-water measurements at all water depths have sub- concentrations. However, while interstitial d13C sequently shown that this distribution coefficient values decrease with increasing depth in the increases with depth, starting at 1.0 at ,1 km and sediment due to the regeneration of isotopically increasing until stabilizing at ,2.9 at a water depth light organic matter, the distribution of cadmium of 3 km (Boyle, 1992). Boyle (1992) argues that the in pore water is more complex. In shallow pore depth dependence is unlikely to be a temperature waters cadmium increases due to the regeneration effect. However, once this depth-dependent distri- of organic material, but in suboxic sediments, it bution coefficient is applied, the reconstructed then decreases due to the precipitation of CdS cadmium from foraminifera tests shows a linear (McCorkle and Klinkhammer, 1991). As a con- relationship to cadmium estimated for the over- sequence, unlike for d13C there is no systematic lying water (Figure 4). Boyle et al. (1995) showed relationship between Cd/Ca of foraminifera that dwell on the surface and those below the sediment surface. Tachikawa and Elderfield (2002) showed 1.2 that, for three sites in the North Atlantic pore ) 1

– waters were 2–4 times enriched in cadmium kg

relative to bottom waters. They found that 1.0 P. wuellerstorfi, whose d13C values suggested that they calcified in the bottom waters at the 0.8 sediment surface, had a D of 3–4, consistent with culture studies (Havach et al., 2001). The species which calcify within the pore water (infaunal 0.6 species) showed lower D (,1) with respect to the pore-water cadmium concentrations. The net 0.4 effect is that the Cd/Ca in P. wuellerstorfi and the infaunal species is quite similar. Further, Tachikawa and Elderfield argue that the depth- 0.2 dependent D for foraminifera (which compared Cd/Ca ratios from primarily Uvigerina and Cd reconstructed from foraminifera (nmol 0.0 Cibicidioides/Planulina) actually reflects a trend 0.0 0.2 0.4 0.6 0.8 1.0 1.2 in the enrichment of pore-water cadmium relative –1 Estimated seawater Cd (nmol kg ) to bottom-water cadmium. So, as with the d13C Figure 4 Seawater cadmium reconstructed from measurements, the Cd/Ca on P. wuellerstorfi or foraminifera using a depth-dependent empirical distri- other surface dwelling foraminifera will give a bution coefficient versus estimated bottom-water cad- more direct measure of bottom-water cadmium. mium (sources Boyle, 1988, 1992). The D for this species is not well constrained but is Conservative Water Mass Tracers 439

between 3 and 4. If infaunal foraminifera are used, ) 12 1 the depth-dependent D of Boyle, which takes into – kg account both the enrichment of pore-water 10 cadmium relative to bottom water and the uptake of cadmium into the foraminifera, is more appropriate. As with d13C, careful interpretation 8 aided by considering supplemental information about changes in bottom-water characteristics has 6 continued to yield regionally coherent and sen- sible patterns of deep-water cadmium (e.g., Boyle, 4 1992; Rickaby et al., 2000). 2 6.16.2.2 Ba/Ca 0 Lea and Boyle (1989) showed that the barium Zn reconstructed from foraminifera (nmol 0 246810 12 content in foraminifera shells is, like cadmium, Estimated seawater Zn (nmol kg–1) controlled by the barium content in bottom waters. Figure 5 Seawater zinc reconstructed from foramini- Barium, in a broad sense, also cycles like a fera using an empirical distribution coefficient which nutrient (depleted in surface waters, and higher in depends on the degree of carbonate unsaturation versus deep waters), but its regeneration occurs deeper estimatedbottom-waterzinc(afterMarchittoetal.,2000). than the organic matter. This results in a close correlation between barium and alkalinity in masses in the past despite the complication that today’s ocean (Lea, 1993). Like Cd/Ca, Ba/Ca in the nutrient content of deep water depends both on foraminifera has also been used, as a paleo-tracer the nutrient content at the source of the deep water of water masses (e.g., Lea and Boyle (1990), but and the addition of nutrients via the regeneration suffers the same carbonate-saturation-state-linked of organic matter as the water mass ages. This is effect as Cd/Ca (McCorkle et al., 1995). primarily due to the large and relatively readily measured contrasts in the nutrient-related tracers. 6.16.2.3 Zn/Ca Headway is now being made in the use of conservative tracers that reflect the mixing of Benthic foraminiferal Zn/Ca is also controlled source waters alone (such as temperature and by bottom-water-dissolved zinc concentration salinity in the modern ocean). (Marchitto et al., 2000), which also follows a nutrient-type profile (Bruland et al., 1978). Zinc is necessary nutrient for phytoplankton and, 6.16.3.1 Mg/Ca in Benthic Foraminifera especially, diatom growth. This results in a very strong correlation between oceanic zinc and Mg/Ca in benthic foraminifera reflect the silicon concentrations. Like Cd/Ca, the apparent calcification temperature (Rosenthal et al., distribution coefficient for zinc decreases in waters 1997b; Rathburn and DeDeckker, 1997; Martin which are more undersaturated with respect to et al., 2002). In principle, temperature can be used calcium carbonate. This change in D is linearly as a conservative tracer for subsurface water related to the degree of undersaturation and, once masses, but at this point the uncertainties in this effect is taken into account, the Zn/Ca in the converting the Mg/Ca data into temperature data foraminifera reflect this well in the overlying water overwhelm the deep-water temperature gradients mass (Figure 5). Marchitto et al. (2000) suggest that distinguish water masses of different origins. that because they respond to saturation state Also, due to the nonlinear nature of the relation- differently, Cd/Ca and Zn/Ca can be used together ship between Mg/Ca and temperature, small errors to assess both the deep-water mass properties and in Mg/Ca can lead to large inferred temperature carbonate ion concentrations. Because of the close changes, especially at cold temperatures. How- relationship between silicon and zinc, Zn/Ca in ever, development in this field is rapid, and there foraminifera is particularly useful for tracing the is promise that Mg/Ca could be used to assess silicon-rich bottom water that originates in the the contributions to intermediate-water masses Southern Ocean (Marchitto et al., 2002). where the temperature signals are more easily distinguished.

6.16.3 CONSERVATIVE WATER MASS TRACERS 6.16.3.2 Pore-water Chemistry In general, nutrient-type tracers have been used The d18O of seawater is set by processes to infer the sources and mixing of deep-water that occur at the ocean surface (evaporation, 440 Tracers of Past Ocean Circulation precipitation, sea ice formation, and melt) (Craig masses in the deep ocean. The use of d18Oasa and Gordon, 1965), and can be used as con- conservative tracer does not require the separation servative tracers in the subsurface ocean. The of the temperature and d18O/salinity components d18O of glacial-age water has been reconstructed of the foraminiferal d18O signal. Despite the fact by measuring the d18O of pore water in deep-sea that oxygen isotope data are routinely gathered sediments (e.g., Schrag and DePaolo, 1993; simultaneously with the carbon isotope data, the Schrag et al., 2002). The maximum contribu- use of d18O as a tracer for ocean circulation has tion to d18O from the LGM appears at a depth of lagged behind the use of d13C. Zahn and Mix 20–60 m within the sediments. The actual d18Oof (1991) showed that the gradients in deep-water the glacial-age deep water must be reconstructed foraminiferal d18O were small, due to the fact by modeling the diffusion of water within the that while NADW is warmer than Antarctic sediments. LGM estimates for the bottom-water Bottom Water (which would lead to lower temperature (also a conservative tracer) can then foraminiferal d18O), the d18O of NADW is higher 18 be made by using the benthic foraminiferal d O (counteracting the temperature effect). They also with the information from pore fluids about the 18 demonstrated that there were significant issues d O of the water in which it calcified. Similarly, with interlaboratory and interspecies calibration salinity can be reconstructed using pore fluid [Cl] that interfered with the effective use of this tracer. (Adkins and Schrag, 2001). While the number of However, recent studies have shown that Cibici- locations with such data are limited, Adkins et al. 18 doides and Planulina do a good job of tracking (2002) show the promise of using pore-water d O calcification temperature (Lynch-Stieglitz et al., and [Cl] to reconstruct the properties of deep- 1999a; Duplessy et al., 2002)(Figure 6). While water masses. As more data are collected, these early studies showed that Cibicidoides d18O did tracers will also be very useful for reconstructing not reflect equilibrium calcification (O’Neil et al., the sources and mixing of deep-water masses. 1969), later determinations of equilibrium calci- fication are consistent with the observed values for core-top Cibicidoides (Kim and O’Neil, 1997; 6.16.3.3 Oxygen Isotopes in Benthic Lynch-Stieglitz et al., 1999a). Some of the inter- Foraminifera laboratory calibration problems arise, because The d18O of foraminiferal calcite reflects the most mass spectrometry laboratories routinely d18O of the water in which it calcified and the calibrate using a carbonate standard (NBS-19) that temperature of calcification. Since both tempera- can be up to 6‰ lower than benthic foraminifera ture and d18O of seawater are imprinted on a water from the deep ocean during glacial periods. Care mass as at the surface, the d18O of the foraminifera must be taken to calibrate regularly with standards is a conservative tracer that can be used to spanning a large range of isotopic ratios, and determine the sources and mixing of water appropriate corrections must be made for many

4

3

(“PDB”) 2 water O

18 1 d

0 (PDB) Ð –1 calcite O 18

d –2

–3 0 51015 20 25 30 Temperature (ûC)

Figure 6 The isotopic fractionation, expressed as the difference between the d18O of foraminiferal calcite and the d18O of seawater (converted from SMOW to a “Peedee Belemnite (PDB)” like scale by subtracting 0.27 in accordance with paleoceanographic tradition) versus temperature of calcification. Cibicidoides and Planulina from Lynch-Stieglitz et al. (1999a) and (— þ) and Duplessy et al. (2002) (— †) are shown along with the inorganic precipitation experiments of Kim and O’Neil (1997) (K). Circulation Rate Tracers 441 mass spectrometers (e.g., Ostermann and Curry, active mixing between these water masses (the 2000). Duplessy et al. (2002) showed coherent Atlantic Ocean and the Southern Ocean) the patterns of deep-ocean foraminiferal d18O, which seawater (and manganese–iron precipitates) could be used to make inferences about deep- have intermediate values, as is the case for the water sources. While the range of foraminiferal nutrient and conservative water mass tracers d18O in the deep ocean is small, the foraminifera (Rutberg et al., 2000). seem to be able to record these differences well. Mackensen et al. (2001) and Matsumoto et al. (2001) also showed regionally coherent patterns of benthic foraminiferal d18O in the Southern 6.16.5 CIRCULATION RATE TRACERS Ocean. With careful attention to laboratory calibration, deep- and intermediate-water forami- The water mass tracers discussed above can be niferal d18O reconstructions show great promise used to assess the sources of subsurface water as a conservative tracer for ocean circulation. masses and the mixture of waters from different sources. However, this information can be used only to assess the relative rates of renewal of these 6.16.3.4 Carbon Isotope Air–Sea Exchange deep-water masses from the various source Signature regions (e.g., LeGrand and Wunsch, 1995). There are several geochemical techniques that Because the d13C in seawater is influenced by can be used to assess the rates of ocean circulation. both biological cycling and air–sea exchange, one can use co-occurring nutrient data (influenced only by biological cycling) to isolate d13C 6.16.5.1 Radiocarbon differences that arise from air–sea exchange (Lynch-Stieglitz and Fairbanks, 1994). Because Carbon-14 (radiocarbon) is produced in the it is reset only at the surface, the air–sea exchange atmosphere and decays with a half-life of signature is a conservative tracer for ocean 5,730 yr. Natural radiocarbon levels in the ocean circulation. Today, newly formed NADW and are highest in the surface ocean, where they 13 Weddell Sea bottom water (WSBW) have d C partially equilibrate with atmosphere. Deep-water air–sea exchange signatures that differ by 0.8‰ concentrations are highest in newly formed (Lynch-Stieglitz et al.,1995). However, the NADW and decrease to the lowest values in the patterns in the air–sea exchange signature are deep Pacific, where the deep waters have longest 13 only as good as the contributing d C and Cd/Ca been out of contact with the atmosphere (Stuiver data, and areas where either dissolution or et al., 1983). Benthic foraminifera record the productivity overprints may be a problem must radiocarbon concentrations of the deep water in be avoided, thus limiting the use of this tracer. which they grow, but the 14C in the foraminifera will decay over time. In order to reconstruct the radiocarbon content at the time the foraminifera 6.16.4 NEODYMIUM ISOTOPE RATIOS were living, radiocarbon measurements can be made on benthic and planktonic foraminifera from The deep-water masses of the ocean have the same interval in the core (Broecker et al., distinctive neodymium isotope ratios that are 1988, 1990; Shackleton et al., 1988) in order to recorded both in manganese nodules (Albare`de determine the age of the benthic foraminifera. and Goldstein, 1992; Albare`de et al.,1997) and However, the radiocarbon content of the plank- dispersed manganese–iron oxides in deep-sea tonic foraminifera will reflect not only the age of sediments (Rutberg et al., 2000). Precisely how the planktonic foraminifera, but the initial radio- the water masses acquire their neodymium isotope carbon content of the near-surface water in which ratios is still unresolved (Albare`de et al., 1997). they grew. The initial radiocarbon in surface This ratio is neither a conservative nor a nutrient waters and in near-surface waters can be signifi- water mass tracer. Neodymium is unaffected by cantly less than expected for equilibrium with the biological cycling in the ocean, but there are both atmosphere due to the relatively long time sources and sinks of neodymium beneath the sea required for isotopic equilibrium between carbon surface, particularly at the sediment–water inter- in the surface ocean and atmosphere. One must face where it is mobilized from detrital material, also account for the fact that the radiocarbon and is precipitated in metallic crusts. The content in the atmosphere changes with time, and neodymium isotope ratios in NADW are high, will also impact the initial radiocarbon content of reflecting the addition of neodymium from the surface waters (Adkins and Boyle, 1997). predominantly old continentally derived detrital Also complicating the use of concurrent benthic material, and the neodymium isotope ratios in the and planktonic radiocarbon concentrations to North Pacific are low, reflecting the influence of assess water mass age is the requirement that the recent volcanic activity. In regions where there is benthic and planktonic foraminifera measured at 442 Tracers of Past Ocean Circulation the same interval actually be the same age. If the that this pattern of protactinium export persisted in peak abundance of planktonic and benthic for- glacial times, and a low deep-water residence time aminifera occurs at different depths in the core, for the LGM Atlantic has been inferred (Yu et al., bioturbation can introduce a spread in the 1996). However, the 231Pa/230Th ratio in sedi- planktonic–benthic ages in cases with low ments depends on the complex interplay between accumulation rate, which is unrelated to changes deep-water flow and the differential scavenging of in bottom-water age (Peng et al., 1977; Broecker the two elements, which depends both on the et al., 1999). While these problems can be composition of the particles and the particle flux overcome by using cores with high sedimentation (Marchal et al., 2000). Biogenic opal fractionates rate (e.g., Keigwin and Schlegel, 2002), the thorium from protactinium much less effectively widespread use of this techniques has been limited than other particle types and the ratios in Southern primarily by the identification of suitable cores Ocean sediments will be very sensitive to the with high enough foraminiferal abundances for particle composition (Luo and Ku, 1999; Yu et al., accurate radiocarbon determinations. Sediments 2001). While the Atlantic 231Pa/230Th fluxes are with high accumulation rates tend to either have low and today predominantly reflect the short low foraminiferal abundances (due to dilution by residence time of deep waters in this basin, in the terrigenous material), or be in high-productivity Pacific Ocean 231Pa/230Th ratios are most sensitive regions of the ocean, where surface waters are to particle flux, with higher ratios where the expected to have variable and relatively low initial scavenging is most intense at the boundaries (Yu radiocarbon content. et al., 2001). Reconstructing circulation patterns Another way to measure the deep-water radio- from 231Pa/230Th ratios will, in general, require a carbon ventilation age in the past is by using deep- detailed knowledge of particle fluxes and dwelling benthic corals (Adkins et al., 1998; composition. Mangini et al., 1998; Goldstein et al., 2001). Here, the age of the coral can be independently determined using uranium-series dating allowing 6.16.5.3 Geostrophic Shear Estimates from the radiocarbon content of the coral to be used to 18 determine the radiocarbon content of deep water d O in Benthic Foraminifera at the time the coral grew. This method is Physical oceanographers routinely use the unaffected by bioturbation, but is currently limited distribution of density in the ocean to help by the availability of deep-sea benthic corals. quantify ocean circulation. A balance between Changes in the ventilation of the ocean can 14 the pressure gradient force and Coriolis term also be assessed by looking at how the C (geostrophic balance) is assumed. Velocity can content of the atmosphere has changed through then be calculated from density information using time (e.g., Hughen et al., 1998). The atmos- 14 either a measured or assumed velocity at a pheric C content is controlled both by the rate reference level, or by combining the geostrophic of 14C production in the atmosphere and the rate 14 balance with other constraints (such as conserva- at which C is transferred into the ocean. While tion of tracer flow across a section). This approach a decrease in the ventilation of the ocean causes has also been applied to paleoceanography by a buildup of radiocarbon in the atmosphere, one using vertical density profiles reconstructed from must also account for changes in production the oxygen isotopic composition of foraminiferal before interpreting the atmospheric radio- calcite (Lynch-Stieglitz et al., 1999a,b). carbon changes in terms of ocean circulation The d18O from the calcite tests of benthic (Marchal et al., 2001). foraminifera preserved in ocean sediments can be used to estimate upper-ocean density because both 18 18 the d O of calcite (d Ocalcite) and density 6.16.5.2 231Pa/230Th Ratio increase as a result of increasing salinity or decreasing temperature (Lynch-Stieglitz et al., Both protactinium and thorium are produced in 1999a). The fractionation between calcite pre- the water column and scavenged by particles onto cipitated inorganically and the water in which it the seafloor. Because protactinium is scavenged forms increases by ,0.2‰ for every 1 8C decrease less efficiently by marine particles (Anderson in temperature (Kim and O’Neil, 1997). The 18 et al., 1983), much of the protactinium produced relationship between d Ocalcite and salinity is 18 18 in the Atlantic Ocean is transported to the more complex. The d Ocalcite reflects the d Oof Southern Ocean by deep waters before it is the water in which the foraminifera grew. The 18 18 scavenged onto the seafloor. This produces a low d O of seawater (d Owater) primarily reflects (below production) 231Pa/230Th ratio in Atlantic patterns of evaporation and freshwater influx to Ocean and a high 231Pa/230Th ratio in Atlantic the surface of the ocean. Because salinity also Ocean in the Southern Ocean (Yu et al., 1996). reflects these same processes, salinity and 231 230 18 Reconstruction of the Pa/ Th ratio suggests d Owater are often well correlated in the ocean. Ocean Circulation during the Last Glacial Maximum 443 Although the exact relationship varies in different the flux of organic material from overlying surface areas of the surface ocean (Craig and Gordon, productivity (e.g., Miller and Lohmann, 1982; 1965), the vast majority of surface and warn Loubere, 1991), limiting its use as a water mass subsurface waters in the ocean have salinity and tracer to low-productivity regions (Schnitker, 18 d Owater values that scatter around a linear trend. 1994). For times in the geologic past, the ability to 18 reconstruct density from the d Ocalcite is most limited by the knowledge of the relationships 18 6.16.7 OCEAN CIRCULATION DURING between d Owater and salinity, as well as the relationships between temperature and salinity. THE LAST GLACIAL MAXIMUM However, it is the gradients in density that are Some of these methods have been applied to used for the geostrophic calculations, and the examine ocean circulation throughout the Cen- reconstructed density gradients are less affected ozoic era, and others to examine century scale by these limitations than the density itself. changes in the very recent past. However, all of This technique has been used to estimate and the methods have been used to look at the ocean reconstruct the flow of the Gulf Stream in the circulation during the last glacial maximum Florida Straits (Lynch-Stieglitz et al., 1999b)by (LGM, ca. (1.9–2.3) £ 104 yr ago) and have reconstructing vertical density profiles on either their maximum geographic coverage during this side of this current. Perhaps more useful for the time. Here we will briefly examine the picture of reconstruction of the large-scale ocean circulation ocean circulation that these techniques have is the fact that the difference in density between produced for the LGM. Each of the paleoceano- the eastern and western margins of the ocean graphic methods outlined above has its advantages basins reflects the strength of the geostrophic and pitfalls, times and places where it should work overturning circulation (Lynch-Stieglitz, 2001). well, and environments where variables other than At this point, the vertical density structure for ocean circulation will complicate the interpreta- the past ocean can be reconstructed using benthic tions. There is no single ocean circulation scenario foraminifera only where the seafloor intersects the for the LGM that is consistent with all of the upper water column (ocean margins, islands, available data from all of the tracers. In the shallow seamounts). Although planktonic forami- following description, regionally coherent signals nifera calcify at various depth within the upper are weighted more heavily than data from isolated water column, we have no way to quantitatively sediment cores. Data in regions where the reconstruct the depth at which they calcified. complications may overwhelm the signal (e.g., However, deep dwelling planktonic foramini- high-productivity regions for d13C data, the very fera can be used to reconstruct the spatial deep ocean for cadmium data, etc.) are weighted pattern of upper ocean flows (Matsumoto and less heavily. A quantitative description of past Lynch-Stieglitz, 2003). ocean circulation will require many more data, better understanding of the tracers, combined with better techniques for dealing with the kinds of 6.16.6 NON-GEOCHEMICAL TRACERS uncertainties and multiple processes that affect the OF PAST OCEAN CIRCULATION paleoceanographic proxies. Both Cd/Ca and d13C distributions suggest a A review of tracers of past ocean circulation strong stratification in the North Atlantic Ocean, would not be complete without mentioning some with a low nutrient, high d13C water mass (often of the sedimentological and paleontological called glacial North Atlantic intermediate water, approaches to reconstructing past ocean flow. GNAIW) occupying depths down to about Sediment grain size has been used to reconstruct 2,000 m, and a high nutrient, low d13C water current intensity since the 1970s (Ledbetter and mass of southern origin underneath (e.g., Boyle Johnson, 1976). More recent grain-size studies and Keigwin, 1987; Oppo and Lehman, 1993; emphasize the mean size of the “sortable silt” Duplessy et al., 1988)(Figure 7). These data are fraction (.10 mm), which is least likely to display generally interpreted as supporting a shallower a cohesive behavior (McCave et al., 1995). overturning circulation in the glacial Atlantic, Stronger currents show larger sortable silt size similar to that observed in some models of LGM due to changes in both deposition and winnowing ocean circulation. with current speed. However, it is clear from many tracers that Benthic foraminifera assemblages have been there was continued ventilation of deep waters related to distinct deep-water masses and have in the North Atlantic as well. Boyle and been used for tracking these water masses during Keigwin (1982) argued that the fact that recon- the past (Streeter, 1973; Schnitker, 1974). In structed cadmium concentrations in the deep subsequent years, it has been determined that the North Atlantic were still lower than average deep-sea benthic assemblage responds strongly to ocean values suggests that there was not a 444 Tracers of Past Ocean Circulation complete shutdown in NADW production. More in the North Atlantic below 2,500 m water depth, recently, Rickaby et al. (2000) find that recon- supporting a North Atlantic source of deep waters structed glacial cadmium concentrations were (Figure 8). The reconstructed d13C are also higher lower in the western basin than the eastern basin in the North Atlantic than in the South Atlantic during the LGM, suggesting continued ventilation of deep waters from the north (e.g., Matsumoto 0 and Lynch-Stieglitz, 1999; Broecker, 2002). d13C Atlantic Ocean Modern Barium (Lea and Boyle, 1990) and zinc +0.5 (Marchitto et al., 2002) data demand continued +0.7 ventilation from the north as well. +0.9 NADW Yu et al. (1996) find that the 231Pa/230Th ratio in 2,500 +1.1 the LGM Atlantic is lower than the production ratio, implying relatively short residence time for AABW waters in the LGM Atlantic. However, this could be accomplished by rapid flushing of deep or intermediate waters either from the north or from 5,000 the south, so does not demand strong ventilation of deep waters in the North Atlantic. Lynch-Stieglitz 0 d13C Atlantic Ocean et al. (1999b) find a reduced vertical shear in the LGM geostrophic velocity in the Florida Straits, con- sistent with a weaker overturning circulation. This +0.9 does not contradict the idea of continued 2,500 +0.6 ventilation of deep- and intermediate-water masses in the North Atlantic during the LGM, Ð0.2 +0.2 but does suggest that this ventilation was either Ð0.4 slower or accomplished by mechanisms that did not draw large quantities of surface water northward. 5,000 Radiocarbon reconstructions for the deep 40ûS 20ûS 0 20û N 40û N North Atlantic show older ventilation ages than Figure 7 Modern and LGM distributions of d13Cin today (Broecker et al.,1990; Keigwin and the Atlantic Ocean inferred from benthic foraminifera Schlegel, 2002), but still younger than waters in (Cibicidoides)(Labeyrie, 1992) (reproduced by per- the South Atlantic (Goldstein et al., 2001), mission of Elsevier from Quat. Sci. Rev., 1992, 11, suggesting that while this deep water was 401–413). ventilated from the North during the LGM, it

Ð1 Ð1 Cdsw (nmol kg )Cdsw (nmol kg ) 0 0.2 0.4 0.6 0 0.20.4 0.6 0 Holocene Glacial

1,000

2,000

3,000 Water depth (km) Water

4,000

5,000 (a) (b)

Figure 8 Cadmium reconstructions for the North Atlantic Ocean, showing the presence of a shallow, nutrient- depleted water mass overlying deeper, nutrient-rich waters (Rickaby et al., 2000). Cores west of the Mid-Atlantic Ridge are plotted in open symbols, with lower nutrient values suggesting a source of new deep waters in the North Atlantic (reproduced by permission of Elsevier from Geochim. Cosmochim. Acta., 2000, 64, 1229–1236). Ocean Circulation during the Last Glacial Maximum 445

Pacific holocene d13C NPIW

1 -0.2 0.1 0.1±0.2 Ð0.1 (n=3) x xx 2 0.4 0.1 0.3±0.2xx 0.3 xxx Ð0.2 0.1 ± (n=6) x xÐ0.4 0.1 x (n=4) Ð0.2 x 0.4 Ð0.1 0.5 0.5 0.3 x 3 0.5 0.2±0.1 x x 0 x 0.3 0.3 0.3±0.1 (n=8) x xx xx x (n=3) x x 0.3±0.1 x x (n=5) x x x xx

Water depth (km) Water 0.3 0.2±0.1x 0.3 0.1±0.0 x (n=8) x 4 (n=3) xx x 0.4 0.4 0.4

5

Pacific LGM d13C 0.4 ?

1 0.1 0.4 0.1 0.2±0.0 (n=3) Ð0.2 Ð0.1 x Ð0.4 ? x x Ð0.1 0.0 2 Ð0.1 Ð0.4 ± xx Ð0.1 -0.4±0.1 0.0 0.1 x x (n=5) x 0.0 xxx (n=4) x x Ð0.1 x Ð0.1 Ð0.2 Ð0.4±0.1x Ð0.4 Ð0.1 Ð0.2 Ð0.2 (n=6) x 3 Ð0.1 Ð0.1 ± x Ð0.1 Ð0.5 Ð0.1 0.0 xx ± x (n=3) x Ð0.3 0.1x Ð0.1±0.1 x (n=5) xx x x x (n=5) x x Water depth (km) Water x Ð0.1 Ð0.3±0.1 Ð0.2 0.0 (n=4) x Ð0.2 4 xx 0.1 0.0 0.0

5

80 60 40 200 20 40 60 South Latitude North Figure 9 Modern and LGM distributions of benthic foraminiferal (Cibicidoides) d13C in the Pacific Ocean (after Matsumoto et al., 2002). 446 Tracers of Past Ocean Circulation was either ventilated more slowly than today or 3.7 that the glacial NADW formed from surface waters which had radiocarbon concentrations that ) 3.9 were less well equilibrated with the atmosphere ‰

O ( 4.1 Pacific than today. 18 d The processes involved in the ventilation of the 4.3 upper and lower deep waters in the North Atlantic were probably different for each water mass and 4.5 South Atlantic North Atlantic are not well understood at this time. However, a 4.7 LGM benthic simple shoaling (and possible southward shift) of South Ocean the NADW overturning cell is probably a very 4.9 simplistic view of the circulation of the LGM Ð1.5 Ð1.0 Ð0.5 0.0 0.5 1.0 1.5 d 13 Atlantic. LGM benthic C (‰) Less is known about the state of the LGM Figure 10 LGM distribution of benthic foraminiferal Pacific Ocean, largely because the poor carbon- d13C and d18O in sediment cores from the deep ocean ate preservation in the Pacific Ocean limits the (mostly .2,000 m water depth) (Duplessy et al., 2002). availability of cores with benthic foraminifera. The very high d18O and low d13C in the Southern Ocean The undersaturation with respect to calcium suggest a separate water mass from those filling the carbonate in the Pacific also complicates the deep Atlantic and Pacific basins (Duplessey et al., 2002) interpretation of the trace-metal-based methods (reproduced by permission of Elsevier from Quat. Sci. that suffer from dissolution or incorporation Rev., 2002, 21, 327). effects. However, like in the Atlantic, the carbon and oxygen isotope data suggest a deep hydro- graphic boundary at ,2,000 m water depth (Mix values in the North Atlantic, to lowest in the et al., 1991; Herguera et al., 1992; Keigwin, North Pacific. This indicates either a distinct 1998; Matsumoto et al.,2002), with a low-d13C, source of deep water to the Pacific Ocean, or high-d18O water mass below this divide and a progressive mixing with the overlying low-d18O high-d13C, low-d18O water mass above this water mass as the deep water ages. divide (Figure 9). The LGM Indian Ocean is Finally, there appears to be an even denser also characterized by a benthic divide at water mass that occupies the deepest portion of ,2,000 m, with higher d13C and lower cadmium the Southern Ocean. This deep water has high- reconstructed for the upper water mass (Kallel d18O values and very low d13C (e.g., Figure 10). et al., 1988; Boyle et al., 1995; McCorkle et al, This water mass is characterized by extremely low 1998). This deep hydrographic boundary is also (about 21‰) d13C values and is observed in the seen off Southern Australia (Lynch-Stieglitz Atlantic sector of the Southern Ocean (Figure 11). et al., 1996). While there was some thought that the low-d13C For both the upper and lower water masses, values may reflect a productivity-related overprint there is a general progression from highest d13C (Mackensen et al., 1993), a convincing case has and Cd/Ca in the North Atlantic, to intermedi- been made that the d13C values are regionally ate values in the South Atlantic and South coherent and unrelated to other measures of Australia, to progressively lower d13C values in overlying productivity (Mackensen et al., 2001; the North Pacific. Goldstein et al. (2001) also Ninnemann and Charles, 2002). One core within argue, based on limited glacial-age radiocarbon this deep-water mass shows LGM neodymium data, that the modern progression of youngest isotope values that have less of an Atlantic deep waters in the North Atlantic, intermediate signature than today (Rutberg et al.,2000). values in the Southern Ocean, and oldest values Although isolated low-d13C values have been in the deep Pacific was preserved during glacial found, there is currently insufficient spatial cover- times. The d18O values in the upper water mass age, especially in deep waters, to document this are similar from North Atlantic to North deep-water mass in the Indian and Pacific sectors Pacific, within the considerable scatter in the of the Southern Ocean. The Cd/Ca data are ever data. Whether or not the GNAIW reached the more limited than the d13C data in the Southern Pacific (after modification in the Southern Ocean (due largely to low foraminiferal abun- Ocean) or there was another source nutrient dance, and potential saturation-state-related over- poor upper deep water in the North Pacific prints), but suggest that the low-d13C values are remains a source of debate (Lynch-Stieglitz not accompanied by significant increases in et al.,1996; Keigwin, 1998; Oppo and nutrients (Rosenthal et al., 1997a,b). This Horowitz, 2000; Matsumoto et al., 2002). suggests that air–sea exchange may have been Duplessy et al. (2002) (Figure 10) show a responsible, at least in part for the lower d13Cin progression of oxygen isotope values in the the deep southern Ocean relative to the rest of deep (.2,000 m) water mass from highest the deep LGM ocean (Toggweiler, 1999). Conclusions 447

Latitude (ûS) Ð50 Ð45 Ð40 Ð35 Ð30 Ð25 Ð20 Ð15 2 PF SAF STF

40 52 90 80 45

49 74 3 NADW 71 .59 .70 87 40 .70

.59 Water depth (km) Water .60 77 4 .40 .39 67

.80 d13C .44

Holocene (4Ð0 ka) 5 Latitude (ûS) WR

Ð50 Ð45 Ð40 Ð35 Ð30 Ð25 Ð20 Ð15 2 PF SAF STF

.03 Ð27 .12 Ð40 0 3 Ð.20 .22 Ð52

Ð.10 .10

Ð67 .07

Water depth (km) Water Ð.20

4 Ð.30 Ð35 Ð41 Ð.10 .10 d13C Ð33

Ð.30 LGM (22Ð16 ka) .50 5 WR Figure 11 Modern and LGM distributions of benthic foraminiferal (Cibicidoides) d13C in the South Atlantic (Mackensen et al., 2001). Data have been corrected for a photodetritus effect, but still show very low d13C(,20.5 per mil) in the deepest portions of the Antarctic (reproduced by permission of Elsevier from Global Planet. Change, 2001, 30, 216–217).

Reconstructions of salinity from pore water [Cl] 6.16.8 CONCLUSIONS (Adkins et al., 2002) suggest that this deep Southern Ocean water mass was considerably Considerable strides have been made since the saltier than deep waters in the Atlantic and Pacific early 1980s in developing methods to reconstruct basins. Adkins et al. (2002) also argue that the past ocean circulation. There are now a number of pore-water d18O suggests that this dense, salty different geochemical methods that can yield water is produced as a by-product of sea ice information on paleo-ocean flow. However, as formation. the use of each method develops, it becomes clear 448 Tracers of Past Ocean Circulation that there are multiple processes affecting the Boyle E. A. and Keigwin L. D. (1987) North Atlantic tracers themselves, and the recording of the thermohaline circulation during the last 20,000 years linked to high latitude surface temperature. Nature 330, geochemical tracers into the sediments. We are 35–40. still just beginning to understand most of these Boyle E. A. and Rosenthal Y. (1996) Chemical hydrography of methods, and have limited amounts of data the South Atlantic during the last glacial maximum: Cd vs available. Learning how to fully appreciate and d13C. In The South Atlantic: Present and Past Circulation (eds. G. Wefer, W. H. Berger, G. Siedler, and D. Webb). separate out these factors, combined with multiple Springer, Berlin, pp. 423–443. measurements using methods affected by different Boyle E. A., Sclater F. R., and Edmond J. M. (1976) On the processes, can ultimately lead to well-constrained marine geochemistry of cadmium. Nature 263, 42–44. reconstructions of past ocean circulation. How- Boyle E. A., Labeyrie L., and Duplessy J.-C. (1995) Calcitic ever, even now these methods are capable of foraminiferal data confirmed by cadmium in aragonitic Hoeglundina: application to the last glacial maximum giving us a tantalizing glimpse of LGM ocean in in the northern Indian Ocean. Paleoceanography 10, which the water masses were arranged in a 881–900. distinctly different pattern than today. Broecker W. S. (1982) Glacial to interglacial changes in ocean chemistry. Prog. Oceanogr. 11, 151–197. Broecker W. S. (2002) Constraints on the glacial operation of the Atlantic Ocean’s conveyor circulation. Israel J. Chem. 42, 1–14. ACKNOWLEDGMENTS Broecker W. S. and Peng T.-H. (1974) Gas exchange rates between air and sea. Tellus 26, 21–35. I am grateful to Tom Marchitto for his help Broecker W. S. and Peng T.-H. (1982) Tracers in the Sea. in preparing this chapter and Dan McCorkle for Eldigio Press, Palisades, NY. his thorough review. This work was supported Broecker W. S. and Maier-Reimer E. (1992) The influence of air and sea exchange on the carbon isotope distribution in the by an NSF CAREER award to the author sea. Global Biogeochem. Cycles 6, 315–320. (OCE-9984989). Broecker W. S., Andree M., Bonani G., Wolfli W., Oeschger H., Klas M., Mix A., and Curry W. (1988) Preliminary estimates for the radiocarbon age of deep water in the glacial ocean. Paleoceanography 3, 659–669. Broecker W. S., Peng T.-H., Trumbore S., Bonani G., and REFERENCES Wolfli W. (1990) The distribution of radiocarbon in the glacial ocean. Global Biogeochem. Cycles 4, 103–107. Adkins J. F. and Boyle E. A. (1997) Changing atmospheric 14 Broecker W., Matsumoto K., Clark E., Hajdas I., and Bonani G. D C and the record of deep water paleoventilation ages. (1999) Radiocarbon are differences between coexisting Paleoceanography 12, 337–344. foraminiferal species. Paleoceanography 14, 431–436. Adkins J. F. and Schrag D. P. (2001) Pore fluid constraints on Bruland K. W., Knauer G. A., and Martin J. H. (1978) Zinc in deep ocean temperature and salinity during the Last Glacial north-east Pacific water. Nature 271, 741–743. Maximum. Geophys. Res. Lett. 28, 771–775. Craig H. (1970) Abyssal carbon 13 in the south Pacific. Adkins J. F., Cheng H., Boyle E. A., Druffel E. R. M., and J. Geophys. Res. 75, 691–695. Edwards R. L. (1998) Deep-sea coral evidence for rapid Craig H. and Gordon L. I. (1965) Deuterium and oxygen 18 change in ventilation of the deep North Atlantic 15,400 years variations in the ocean and the marine atmosphere: stable ago. Science 280, 725–728. isotopes in oceanographic studies and paleotemperatures. In Adkins J. F., McIntyre K., and Schrag D. P. (2002) The salinity, 18 Proceedings of the 3rd Spoleto Conference, Spoleto, Italy temperature, and d O of the glacial deep ocean. Science (ed. E. Tongiori). Sischi and Figli, Pisa, pp. 9–130. 298, 1769–1773. Curry W. B. and Lohmann G. P. (1982) Carbon isotopic Albare`de F. and Goldstein S. L. (1992) World map of Nd changes in benthic foraminifera from the western South isotopes in sea-floor ferromanganese deposits. Geology 20, Atlantic: reconstructions of glacial abyssal circulation 761–761. patterns. Quat. Res. 18, 218–235. Albare`de F., Goldstein S. L., and Dautel D. (1997) The Curry W. B., Duplessy J.-C., Labeyrie L. D., and Shackleton 13 neodymium isotopic composition of manganese nodules N. J. (1988)P Changes in the distribution of d C of deep from the Southern and Indian oceans, the global oceanic water CO2 between the last glaciation and the Holocene. neodymium budget, and their bearing on deep ocean Paleoceanography 3, 317–341. circulation. Geochim. Cosmochim. Acta 61, 1277–1291. Deuser W. G. and Hunt J. M. (1969) Stable isotope ratios of Anderson R. F., Bacon M. P., and Brewer P. G. (1983) dissolved inorganic carbon in the Atlantic. Deep-Sea Res. Removal of 231Pa and 230Th from the open ocean. Earth 16, 221–225. Planet. Sci. Lett. 62, 7–23. Duplessy J.-C., Shackleton N. J., Matthews R. K., Prell W., Belanger P. E., Curry W. B., and Matthews R. K. (1981) Core- Ruddiman W. F., Caralp M., and Hendy C. H. (1984) 13C top evaluation of benthic foraminiferal isotopic ratios for record of benthic foraminifera in the last interglacial ocean: paleo-oceanographic interpretations. Paleogeogr. Paleocli- implications for the carbon cycle and the global deep water matol. Paleoecol. 33, 205–220. circulation. Quat. Res. 21, 225–243. Boyle E. A. (1981) Cadmium, zinc, copper, and barium in Duplessy J.-C., Shackleton N. J., Fairbanks R. G., Labeyrie L., foraminifera tests. Earth Planet. Sci. Lett. 53, 11–35. Oppo D., and Kallel N. (1988) Deepwater source variations Boyle E. A. (1988) Cadmium: chemical tracer of deepwater during the last climatic cycle and their impact on the paleoceanography. Paleoceanography 3, 471–489. global deepwater circulation. Paleoceanography 3, Boyle E. A. (1992) Cadmium and d13C paleochemical ocean 343–360. distributions during the stage 2 glacial maximum. Ann. Rev. Duplessy J.-C., Labeyrie L., and Waelbroeck C. (2002) Earth Planet. Sci. 20, 245–287. Constraints of the ocean oxygen isotopic enrichment Boyle E. A. and Keigwin L. D. (1982) Deep circulation of the between the Last Glacial Maximum and the Holocene: North Atlantic over the last 200,000 years: geochemical paleoceanographic implications. Quat. Sci. Rev. 21, evidence. Science 218, 784–787. 315–330. References 449

Elderfield H. and Rickaby R. E. M. (2000) Oceanic Cd/P ratio Loubere P. (1991) Deep-sea benthic foraminiferal assemblage and nutrient utilization in the glacial Southern Ocean. Nature response to a surface ocean productivity gradient: a test. 405, 305–310. Paleoceanography 6, 193–204. Elderfield H., Bertram C. J., and Erez J. (1996) A biominer- Luo S. and Ku T.-L. (1999) Oceanic 231Pa/230Th ratio alization model for the incorporation of trace elements into influenced by particle composition and remineralization. foraminiferal calcium carbonate. Earth Planet. Sci. Lett. Earth Planet. Sci. Lett. 167, 183–195. 142, 409–423. Lutze G. F. and Thiel H. (1989) Epibenthic foraminifera Goldstein S. J., Lea D. W., Chakraborty S., Kashgarian M., and from elevated microhabitats: Cibicidoides wuellerstorfi Murrell M. T. (2001) Uranium-series and radiocarbon and Planulina ariminensis. J. Foraminiferal Res. 19, geochronology of deep-sea corals: implications for Southern 153–158. Ocean ventilation rates and the oceanic carbon cycle. Earth Lynch-Stieglitz J. (2001) Using ocean margin density to Planet. Sci. Lett. 193, 167–182. constrain ocean circulation and surface wind strength in the Graham D. W., Corliss B. H., Bender M. L., and Keigwin L. D. past. Geochem. Geophys. Geosys. 2, 2001GC000208. (1981) Carbon and oxygen isotopic disequilibria of recent Lynch-Stieglitz J. and Fairbanks R. G. (1994) A conservative deep-sea benthic foraminifera. Mar. Micropaleontol. 6, tracer for glacial ocean circulation from carbon isotope and 483–497. paleonutrient measurements in benthic foraminifera. Nature Grossman E. L. (1984a) Carbon isotopic fractionation in live 369, 308–310. benthic foraminifera-comparison with inorganic precipitate Lynch-Stieglitz J., Stocker T. F., Broecker W. S., and studies. Geochim. Cosmochim. Acta 48, 1505–1512. Fairbanks R. G. (1995) The influence of air–sea exchange Grossman E. L. (1948b) Stable isotope fractionation in live on the isotopic composition of oceanic carbon: observations benthic foraminifera from the southern California border- and modeling. Global Biogeochem. Cycles 9, 653–665. land. Palaeogeogr. Paleoclimatol. Palaeoecol. 47, Lynch-Stieglitz J., van Geen A., and Fairbanks R. G. (1996) 301–327. Interocean exchange of Glacial North Atlantic Intermediate Havach S. M., Chandler G. T., Wilson-Finelli A., and Shaw Water: evidence from sub-Antarctic Cd/Ca and carbon T. J. (2001) Experimental determination of trace element isotope measurements. Paleoceanography 11, 191–201. partition coefficients in cultured benthic foraminifera. Lynch-Stieglitz J., Curry W. B., and Slowey N. (1999a) A Geochim. Cosmochim. Acta 65, 1277–1283. geostrophic transport estimate for the Florida Current from Herguera J. C., Jasen E., and Berger W. H. (1992) Evidence for the oxygen isotope composition of benthic foraminifera. a bathayal front at 2,000 m depth in the glacial Pacific, based Paleoceanography 14, 360–373. on a depth transect on Ontong-Java Plateau. Paleoceano- Lynch-Stieglitz J., Curry W. B., and Slowey N. (1999b) graphy 7, 273–288. Weaker gulf stream in the Florida Straits during the last Hughen K. A., Overpeck J. T., Lehman S. J., Kashgarian M., glacial maximum. Nature 402, 644–648. Southon J., Peterson L. C., Alley R., and Sigman D. M. (1998) Mackensen A. and Douglas R. G. (1989) Down-core distribution of live and dead deep-water benthic foraminifera Deglacial changes in ocean circulation from an extended in box cores from the Weddell Sea and the California radiocarbon calibration. Nature 391, 65–68. continental borderland. Deep-Sea Res. 36, 879–900. Inoue H. and Sugimura Y. (1985) Carbon isotopic fractionation Mackensen A., Hubberten H.-W., Bickert T., Fischer G., and during CO exchange process between air and sea water 2 Fu¨tterer D. K. (1993) The d13C in benthic foraminiferal tests under equilibrium and kinetic conditions. Geochim. Cosmo- of Fontbotia wuellerstorfi (Schwager) relative to the d13Cof chim. Acta 49, 2453–2460. dissolved inorganic carbon in southern ocean deep water: Kallel N., Labeyrie L. D., Juillet-Leclerc A., and Duplessy implications for glacial ocean circulation models. J.-C. (1988) A deep hydrological front between intermediate Paleoceanography 8, 587–610. Nature and deep water masses in the glacial Indian Ocean. Mackensen A., Rudolph M., and Kuhn G. (2001) Late 333, 651–655. Pleistocene deep-water circulation in the sub-Antarctic Keigwin L. D. (1998) Glacial-age hydrography of the far eastern Atlantic. Global Planet. Change 30, 197–229. northwest Pacific Ocean. Paleoceanography 13, 323–339. Mangini A., Lomitschka M., Eichstadter R., Frank N., Vogler Keigwin L. D. and Schlegel M. A. (2002) Ocean ventilation S., Bonani G., Hajdas I., and Patzold J. (1998) Coral and sedimentation since the glacial maximum at 3 km in the provides way to age deep water. Nature 392, 347–348. western North Atlantic. Geochem. Geophys. Geosys. 3, Marchal O., Francois R., Stocker T. F., and Joos F. (2000) 10.1029/2001GC000283. Ocean thermohaline circulation and sedimentary 231Pa/230Th Kim S.-T. and O’Neil J. (1997) Equilibrium and none- ratio. Paleoceanography 15(6), 625–641. quilibrium oxygen isotope effects in synthetic carbonates. Marchal O., Stocker T. F., and Muscheler R. (2001) Geochim. Cosmochim. Acta 61, 3461–3475. P 13 Atmospheric radiocarbon during the Younger Dryas: Kroopnick P. M. (1985) The distribution of Cof CO2 in the Production, ventilation, or both? Earth Planet. Sci. Lett. world oceans. Deep-Sea Res. 32, 57–84. 185, 383–395. Labeyrie L. D. (1992) Changes in the vertical structure of the Marchitto T. M., Curry W. B., and Oppo D. W. (2000) Zinc North Atlantic Ocean between glacial and modern times. concentrations in benthic foraminifera reflect seawater Quat. Sci. Rev. 11, 401–413. chemistry. Paleoceanography 15, 299–306. Lea D. W. (1993) Constraints on the alkalinity and circulation Marchitto T. M., Oppo D. W., and Curry W. B. (2002) Paired of glacial circumpolar deep water from benthic foraminiferal benthic foraminiferal Cd/Ca and Zn/Ca evidence for barium. Global Biogechem. Cycles 7, 695–710. a greatly increased presence of Southern Ocean water in Lea D. W. and Boyle E. A. (1989) Barium content of benthic the glacial North Atlantic. Paleoceanography 17(3), foraminifera controlled by bottom-water composition. art. no. 1038. Nature 338, 751–753. Martin P. A., Lea D. W., Rosenthal Y., Shackleton N. J., Lea D. W. and Boyle E. A. (1990) Foraminiferal reconstruc- Sarnthien M., and Papenfuss T. (2002) Quaternary deep sea tions of barium distributions in water masses of the glacial temperature histories derived from benthic foraminiferal oceans. Paleoceanography 5, 719–742. Mg/Ca. Earth Planet. Sci. Lett. 198, 193–209. Ledbetter M. T. and Johnson D. A. (1976) Increased transport Matsumoto K. and Lynch-Stieglitz J. (1999) Similar glacial of Antarctic bottom water in the Vema Channel during the and Holocene deep water circulation inferred from southeast Last Ice Age. Science 194, 837–839. Pacific benthic foraminiferal carbon isotope composition. LeGrand P. and Wunsch C. (1995) Constraints from paleo- Paleoceanography 14, 149–163. tracer data on the North Atlantic circulation during the last Matsumoto K. and Lynch-Stieglitz J. (2003) Persistence of glacial maximum. Paleoceanography 10, 1011–1045. Gulf Stream separation during the Last Glacial Period: 450 Tracers of Past Ocean Circulation

implications for current separation theories. J. Geophys. Res. compositions of living (stained) benthic foraminifera from Oceans 108(C6), art. no. 3174. the Sulu and South China seas. Deep-Sea Res. I 43, Matsumoto K., Lynch-Stieglitz J., and Anderson R. F. (2001) 1617–1646. Similar glacial and Holocene Southern Ocean hydrography. Rickaby R. E. M., Greaves M. J., and Elderfield H. (2000) Cd Paleoceanography 16, 445–454. in planktonic and benthic foraminiferal shells determined by Matsumoto K., Oba T., Lynch-Stieglitz J., and Yamamoto H. thermal ionisation mass spectrometry. Geochim. Cosmo- (2002) Interior hydrography and circulation of the glacial chim. Acta 64, 1229–1236. Pacific Ocean. Quat. Sci. Rev. 1693–1704. Romanek C. S., Grossman E. L., and Morse J. W. (1992) McCave I. N., Manighetti B., and Robinson S. G. (1995) Carbon isotopic fractionation in synthetic aragonite and Sortable silt and fine sediment size/composition slicing: calcite: effect of temperature and precipitation rate. parameters for palaeocurrent speed and palaeoceanography. Geochim. Cosmochim. Acta 56, 419–430. Paleoceanography 10, 593–610. Rosenthal Y., Boyle E. A., Labeyrie L., and Oppo D. (1995) McCorkle D. C. and Klinkhammer G. P. (1991) Porewater Glacial enrichments of authigenic Cd and U in sub-Antarctic cadmium geochemistry and the porewater cadmium–delta- sediments—a climatic control on the elements oceanic C-13 relationship. Geochim. Cosmochim. Acta 55, 161–168. budget. Paleoceanography 10, 395–413. McCorkle D. C., Keigwin L. D., Corliss B. H., and Emerson Rosenthal Y., Boyle E. A., and Labeyrie L. (1997a) Last glacial S. R. (1990) The influence of microhabitats on the carbon maximum paleochemistry and deepwater circulation in the isotopic composition of deep-sea benthic foraminifera. Southern Ocean: evidence from foraminiferal cadmium. Paleoceanography 5, 161–185. Paleoceanography 12, 787–796. McCorkle D. C., Martin P. A., Lea D. W., and Klinkhammer Rosenthal Y., Boyle E. A., and Slowey N. (1997b) Tempera- G. P. (1995) Evidence of a dissolution effect on benthic ture control on the incorporation of magnesium, strontium, foraminiferal shell chemistry: d13C, Cd/Ca, Ba/Ca, and Sr/ fluorine, and cadmium into benthic foraminiferal shells from Ca results from the Ontong Java Plateau. Paleoceanography Little Bahama Bank: prospects for thermocline paleoceano- 10, 699–714. graphy. Geochim. Cosmochim. Acta 61, 3633–3643. McCorkle D. C., Corliss B. H., and Farnham C. A. (1997) Rutberg R. L., Hemming S. R., and Goldstein S. L. (2000) Vertical distributions and stable isotopic compositions of Reduced North Atlantic deep water flux to the glacial live (stained) benthic foraminifera from the North Carolina Southern Ocean inferred from neodymium isotope ratios. and California continental margins. Deep-Sea Res. I 44, Nature 405, 935–938. 983–1024. Schnitker D. (1974) West Atlantic abyssal circulation during McCorkle D. C., Heggie D. T., and Veeh H. H. (1998) Glacial the past 120,000 years. Nature 248, 385–387. and Holocene stable isotope distributions in the southeastern Schnitker D. (1994) Deep-sea benthic foraminifers: food and Indian Ocean. Paleoceanography 13, 20–34. bottom water masses. In Carbon Cycling in the Glacial Miller K. G. and Lohmann G. P. (1982) Environmental Ocean: Constraints on the Ocean’s Role in Global Change, distribution of recent benthic foraminifera on the northeast NATO ASI Ser. I. 17. (eds. R. Zahn, T. F. Pedersen, United States continental slope. Geol. Soc. Am. Bull. 93, M. A. Kaminiski, and L. Labeyrie). Springer, Berlin, 200–206. pp. 539–554. Mix A. C., Pisias N. G., Zahn R., Rugh W., Lopez C., and Schrag D. P. and DePaolo D. J. (1993) Determination of d18O Nelson K. (1991) Carbon 13 in Pacific deep and intermediate of seawater in the deep ocean during the Last Glacial waters, 0–370 ka: implications for ocean circulation and Maximum. Paleoceanography 8, 1–6. Pleistocene CO2. Paleoceanography 6, 205–226. Schrag D. P., Adkins J. F., McIntyre K., Alexander J. L., Mook W. G., Bommerson J. C., and Staverman W. H. (1974) Hodell D. A., Charles C. D., and McManus J. F. (2002) Carbon isotope fractionation between dissolved bicarbonate The oxygen isotopic composition of seawater during the and gaseous carbon dioxide. Earth Planet. Sci. Lett. 22, Last Glacial Maximum. Quat. Sci. Rev. 21, 331–342. 169–176. Shackleton N. J. (1977) Tropical rainforest history and the Ninnemann U. S. and Charles C. D. (2002) Changes in the equatorial Pacific carbonate dissolution cycles. In Fate of mode of Southern Ocean circulation over the last glacial Fossil Fuel CO2 in the Oceans (eds. N. R. Anderson and cycle revealed by foraminiferal stable isotopic variability. A. Malahoff). Plenum, NY, pp. 401–427. Earth Planet. Sci. Lett. 201, 383–396. Shackleton N. J., Duplessy J. C., Arnold M., Maurice P., Hall O’Neil J. R., Clayton R. N., and Mayeda T. K. (1969) Oxygen M. A., and Cartlidge J. (1988) Radiocarbon age of last isotope fractionation in divalent metal carbonates. J. Chem. glacial Pacific deep water. Nature 335, 708–711. Phys. 5, 5547–5558. Spero H. J. and Lea D. W. (1996) Experimental determination Oppo D. W. and Horowitz M. (2000) Glacial deep water of stable isotope variability in Globigerina bulloides: geometry: South Atlantic benthic foraminiferal Cd/Ca and implications for paleoceanographic reconstructions. Mar. d13C evidence. Paleoceanography 15, 147–160. Micropaleontol. 28, 231–246. Oppo D. W. and Lehman S. J. (1993) Mid-depth circulation of Spero H. J., Bijma J., Lea D. W., and Bemis B. E. (1997) Effect the subpolar North Atlantic during the last glacial maximum. of seawater carbonate chemistry on planktonic foraminiferal Science 259, 1148–1152. carbon and oxygen isotope values. Nature 390, 497–500. Ostermann D. R. and Curry W. B. (2000) Calibration of stable Streeter S. S. (1974) Bottom water and benthonic foraminifera isotopic data: an enriched delta O-18 standard used for in the North Atlantic glacial-interglacial contrasts. Quat. source gas mixing detection and correction. Paleoceano- Res. 3, 131–141. graphy 15, 353–360. Stuiver M., Quay P. D., and Ostlund H. G. (1983) Abyssal Peng T.-H., Broecker W. S., Kipphut G., and Shackleton N. water C-14 distribution and the age of the world oceans. (1977) Benthic mixing in deep-sea cores as determined by Science 219, 849. 14C dating and its implications regarding climate stratigra- Tachikawa K. and Elderfield H. (2002) Microhabitat effects on 13 phy and the fate of fossil fuel CO2.InFate of Fossil Fuel Cd/Ca and d C of benthic foraminifera. Earth Planet. Sci. CO2 in the Oceans (eds. N. R. Anderson and A. Malahoff). Lett. 202, 607–624. Plenum, NY, pp. 355–373. Tans P. P. (1980) On calculating the transfer of carbon-13 in Rathburn A. E. and DeDeckker P. (1997) Magnesium and reservoir models of the carbon cycle. Tellus 32, 464–469. strontium compositions of recent benthic foraminifera from Toggweiler J. R. (1999) Variation of atmospheric CO2 by the Coral Sea, Australia and Prydz Bay, Antarctica. Mar. ventilation of the ocean’s deepest water. Paleoceanography Micropaleo. 32, 231–248. 14, 571–588. Rathburn A. E., Corliss B. H., Tappa K. D., and Lohmann K. C. van Geen A., McCorkle D. C., and Klinkhammer G. P. (1995) (1996) Comparisons of the ecology and stable isotopic Sensitivity of the phosphate-cadmium–carbon isotope References 451

relation in the ocean to cadmium removal by suboxic circulation inferred from radiochemical data. Nature 379, sediments. Paleoceanography 10, 159–169. 689–694. Wefer G. and Berger W. H. (1991) Isotope paleontology: Yu E. F., Francois R., Bacon M. P., et al. (2001) Fluxes of growth and composition of extant calcareous species. Mar. Th-230 and Pa-231 to the deep sea: implications for the Geol. 100, 207–248. interpretation of excess Th-230 and Pa-231/Th-230 profiles Woodruff F., Savin S. M., and Douglas R. G. (1980) Biological in sediments. Earth Planet Sci. Lett. 191(3–4), 219–230. fractionation of oxygen and carbon isotopes by recent benthic Zahn R. and Mix A. C. (1991) Benthic foraminiferal d18Cin foraminifera. Mar. Micropaleo. 5, 3–11. the ocean’s temperature–salinity–density field: constraints Yu E.-F., Francois R., and Bacon M. P. (1996) Similar on Ice Age thermohaline circulation. Paleoceanography 6, rates of modern and last-glacial ocean thermohaline 1–20.

q 2003, Elsevier Ltd. All rights reserved Treatise on Geochemistry No part of this publication may be reproduced, stored in a retrieval system or ISBN (set): 0-08-043751-6 transmitted in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, without prior written permission of the Publisher. Volume 6; (ISBN: 0-08-044341-9); pp. 433–451