Geochimico er Cosmochimica Acta Vol. 55. pp. 3321-3331 Copyright 8 1991 Pergamon Press plc. Printed in U.S.A.

oOl6-7037/91/$3.00 + .LM

Barium in planktonic foraminifera

DAVID W. LEA * ,+ and EDWARD A. BOYLE

Department of Earth, Atmospheric, and Planetary Sciences, Massachusetts Institute of Technology, Cambridge, MA 02 139, USA

(Received May 10, 1990; accepted in revised form August 15, 199 1)

Abstract-Reconstructions of Ba distributions in ancient oceanic surface waters could provide new insight into paleoceanographic change. Calcite shells of planktonic foraminifera potentially provide a means of reconstructing such paleo-Ba distributions if lattice-bound Ba can be determined on shells recovered from deep-sea cores. Planktonic foraminifera shells from a series of cores were purified of non-lattice- bound Ba associated with organic or sedimentary phases by a combination of physical agitation, oxidative- reductive steps, acid leaches, and a novel alkaline-DTPA step to dissolve bar&e. A sequential dissolution of a large sample of cleaned shells of the planktonic foraminifer Globigerinoides conglobatus indicates homogeneous distribution of Ba in the shell material. Comparison of shells from sediments, sediment traps, and plankton tows indicates no significant differences in the Ba content of the purified shells. Cleaned samples of the planktonic foraminifera Globigerinoides sacculifea, G. ruber, G. conglobatus, Orbulina spp., and Globoquudrina dutertrei from the equatorial Pacific, North Atlantic, and Mediterranean Sea have Ba/Ca ratios between 0.6 and 1.0 rmol/mol(O.8 to 1.4 ppm). Variation in foraminiferal Ba contents between the three basins is consistent with the trend in surface seawater Ba. The calculated distribution coefficient for Ba incorporation in these five species based on these data is 0.19 * 0.05. Several species of the non-spinose planktonic foraminifera Globorotalia have Ba/Ca ratios ranging from 2 to 13 pmol/mol; these high Ba contents might be explained by differences in the way these foraminifera precipitate their shells. A temporal record of Ba/Ca in samples of Globigerinoides and Orbulina from a core in the northwest Atlantic suggests that the Ba concentration of surface waters at this site has not changed by more than 20% over the last 14 kyr.

INTRODUCIION

THE INCORPORATION OF trace elements into marine carbon- ates has proved to be a useful method for probing paleo- chemical changes in ancient oceans. Since these tracers are characterized by unique marine geochemical cycles, each element can provide novel insight into oceanic change. The observation that water masses are often characterized by contrasting trace element signatures is one of the keys to the utility of these paleo-metal probes. Reconstructions of past trace metal distributions accomplished by using an array of sediment cores yield direct insight into changes in oceanic circulation parameters.

Foraminiferal Cd has received the most attention in this regard (BOYLE, 1981, 1988; BOYLE and KEIGWIN, 1982, 1985, 1987); the effectiveness of the Cd method rests on the similarity of Cd distributions to that of the limiting nutrient phosphorous. A recent study demonstrated that a new pa- leochemical tracer, Ba in benthic foraminifera, can be used to reconstruct Ba in past bottom waters (LEA and BOYLE,

1989). Like Cd, Ba is also a nutrient-Like tracer, but its oceanic distribution contrasts with that of Cd due to differences in uptake and regeneration of Ba in the water column.

The distribution of barium in the oceans reflects Ba’s in- volvement in the biological cycle of uptake in surface waters and regeneration in deep waters (BISHOP, 1988; CHAN et al., 1977; CHOW and GOLDBERG, 1960). Barium is depleted in

* Present address: Department of Geological Sciences, University of California, Santa Barbara, CA 93 106, USA.

’ Author to whom correspondence should be addressed.

tropical and sub-tropical surface waters; removal of dissolved Ba from surface waters appears to be controlled by precipi- tation of barite (BaSO,) in decaying marine particulate matter (BISHOP, 1988; CHOW and GOLDBERG, 1960; DEHAIRS et al., 1980). The barite in the sinking particulates dissolves deep in the water column and/or in the sediments, creating deep water maxima throughout the world’s oceans (CHAN et al., 1977).

The concentration of Ba in oceanic surface waters (ex- cluding high latitudes) is quite uniform, about 34 nmol/kg in the Pacific and 41 nmol/ kg in the Atlantic ( CHAN et al., 1977; OSTLUND et al., 1987). Higher Ba values in Atlantic surface waters are probably due to the large river inputs that drain into the Atlantic. Because the residence time of Ba in surface waters is about 10 times longer than that typical for more bio-active elements like Cd or P, Ba anomalies due to high Ba river inputs persist within ocean basins.

The magnitude of depletion of Ba in surface waters (as low as 20% of the highest deep water values) is much smaller than is observed for elements like P, NOs , Si, and Cd, which are highly depleted in surface waters. Unlike these bio-limiting elements, the concentration of Ba in surface waters must in part reflect the mean Ba concentration of the ocean; simple models suggest that surface Ba rises with increasing mean ocean Ba (LEA, 1990). Oceanic Ba is primarily supplied by riverine input of Ba, although Ba from deep ocean hydro- thermal vents might account for up to 20% of the Ba input to the oceans ( VON DAMM et al., 1985).

Although Ba is a large ion relative to Ca ( 1.35 vs. 1 .OO A in 6-fold coordination; SHANNON, 1976), it can substitute for Ca in calcite (KITANO et al., 1971). The Ba content of

3321

3322 D. W. Lea and E. A. Boyle

calcitic planktonic foraminifera could be used as a means of

reconstructing surface water Ba concentrations in past oceans if Ba is substituted into the lattice of the calcite shell in pro- portion to its dissolved concentration in seawater. One of the immediate difficulties in verifying proportional substitution is that there is a very limited range of Ba variability in surface

waters of the low- to mid-latitude oceans: 34 nmol/ kg in oli- gotrophic Pacific surface waters to 53 nmol/kg in the Med-

iterranean (LEA, 1990; OSTLUND et al., 1987 ). This, coupled with the inherent 5-10% variability that apparently charac- terizes planktonic Ba/Ca (see later and LEA and BOYLE, 1990b) limits the quality of the empirical calibration curve

that can be constructed from core-top foraminifera.

foraminiferal species. Ba can reside in a number ofextraneous phases associated with shells: detrital grains, fine-grained CaCO,, organic matter, Mn and Fe oxides, and barite ( BaSO,) . Initial cleaning steps follow methods developed for foraminiferal Cd (BOYLE, 198 1; BoYLE and KEIGWIN, 1985): detrital grains and fine grained material are removed by physical agitation in distilled water and methanol, organic matter by oxidation in hydrogen peroxide-sodium hydroxide solution, ferro-manganese oxide coatings by reduction in hydrazine-ammo- nium citrate, and adsorbed metals by a 0.001 N nitric acid leach. (For details of these steps, see BOYLE, 1981: BOYLE and KEIGWIN, 1985.)

A study undertaken recently to directly culture Orbulina

universa in seawater solutions of known Ba content unam- biguously confirms incorporation of Ba in shells in proportion to seawater Ba concentration (LEA and SPERO, 199 1). The study also demonstrated that the distribution coefficient that

characterizes shell incorporation is about 0.2, the same value found for the samples in this study by empirical calibration

( see later).

A unique cleaning problem for foraminiferal Ba is the presence of barite ( BaS04) in marine sediments. Barite is present in marine sed- iments over a range of about 10 to 10.000 ppm (CHURCH, 1979). with highest concentrations occurring in sediments underlying oceanic high productivity belts. SEM examination of sediment samples from the Panama Basin (core WH482-483) indicates that spheres of barite approximately one micron in diameter occur in conjunction with diatom debris coating foraminifera shells. These barite crystals are a potential source of contamination to determinations of lattice-bound Ba in foraminifera. Barite is found in samples from both sediment cores and sediment traps, so at least some portion of the barite must be formed in the water column. The amount of barite present in unpurified samples of separated shells is up to tens of ppm. while Ba in cleaned planktonic shells is generally a ppm or less.

BOYLE ( 198 1) found high and variable Ba contents for

planktonic foraminifera from an equatorial Atlantic sediment core, suggesting that a careful evaluation of cleaning proce- dures is required to ensure that lattice-bound foraminiferal Ba is not obscured by extraneous Ba-rich sedimentary phases. This study details an assessment of cleaning methods for de- termination of Ba/Ca in various species of planktonic fo-

raminifera from a number of sediment cores, plankton tows, and sediment traps. Results indicate that shells of certain species of Globigerinoides and Orbulina can be cleaned of non-lattice-bound Ba and therefore can be used as a means

of reconstructing past surface water Ba concentrations.

ANALYTICAL METHODS

Samples were analyzed for Ba by a modification of the graphite furnace atomic absorption technique. Graphite tubes were modified by insertion of a 0.127 mm thick rolled tantalum liner ( SEN GUF-TA, 1984). These Ta liners increase sensitivity for Ba by a factor of I2 and reduce the carryover blank characteristic of carbide-forming elements like Ba. [Typical sensitivities were 0.01 absorbance units (peak height) for a 20 ILL injection of Ba = 10 nmol/L.] Ta-lined graphite tubes require atomization temperatures of only 2500°C compared to >27OO”C required for conventional graphite tubes. Al- though the Ta-lined tubes often performed well for -50 injections, it was impossible to predict when the liners would fail, at which point the precision would become intolerably poor. In addition, it was very difficult to produce consistent lined tubes; typically only a quarter of the batches prepared worked effectively.

Barite is removed by cleaning the separated shells in a solution of alkaline diethylenetriamine-pentaacetic acid (DTPA). DTPA forms strong complexes with Ba under alkaline conditions ( RINGBOM, 1963), so DTPA concentrations many orders of magnitudes in excess of the barite present causes rapid dissolution of the barite (SILL and WILLIS, 1964, 1966). Since the DTPA also has a strong affinity for Ca, an unavoidable side effect of the alkaline-DTPA treatment is that some foraminiferal calcite is lost. For this reason the concentration of DTPA and length of cleaning time must be carefully balanced to ensure dissolution of the barite without excessive loss of foraminifera. Upon completion of the cleaning steps used for Cd purification. 0.25 mL of alkaline-DTPA solution (0.2 mol/L DTPA in 2.5 N NaOH) is added to centrifuge vials containing 1 to 3 mg of shell material. The vials are then placed in a boiling water bath for 3 to 15 min; the length of the cleaning time that can be used depends on the initial sample size. Samples are repeatedly ultrasonicated throughout the treatment to encourage dissolution of barite. Upon completion of the cleaning. three to five rinses of ammonium hydroxide are used to rinse out the alkaline-DTPA solution, followed by three to five rinses of distilled water to rinse out the ammonium hydroxide.

Ba was quantified by comparison to standards prepared in a CaNOj matrix matched to the average sample Ca content. Absorption peaks for digests of foraminifera shells with less than 1 ppm Ba, typical for samples of Globigerinoides, Orbulina. and Globoquadrinia. were about 0.0 1 absorbance units, with signal to noise ratios of about 10. Digests of Globorotalia samples, which have higher Ba contents (see later), gave absorbances of about 0.05. Ca was analyzed by flame atomic absorption with inter-run reproducibility of 2%. Inter-run precision of the Ba/Ca ratio was about 6% for consistency standards with optimal Ba concentrations: however, samples with low Ba, typical of Globigerinoides, Orbulina. and Globoquadrinia samples, have precisions of about 10%.

A series of measurements has been made to assess Ba in various sedimentary phases (Fig. 1,2; core locations in Table 1; data compiled in Table 2). Planktonic foraminifera treated with ultrasonication to remove loosely adhering sedimentary particles have between 2-8 pmol/mol Ba/Ca ( I rmol/mol Ba/Ca = 1.37 ppm). Those subject to oxidative cleaning for organically bound Ba. reductive cleaning to remove oxide coatings and alkaline DTPA to remove barite have Ba/Ca ratios of only 0.8 f 0.2 pmol/mol (an exception are the Globorotalia-see below). Comparison of samples cleaned with and without both the reductive step and the DTPA step indicates that the proportion of Ba these cleaning steps remove depends strongly on the sediment type, with the cleaning steps becoming increasingly more important in sediments from more productive regions of the ocean where barite abundances (and possibly the proportion of oxide coatings) are greater. Samples of both sediment and sediment trap material from the Panama Basin indicate the importance of the DTPA step for removing adhering barite (Fig. 2). The final strategy has been to use the most rigorous cleaning method on all samples under the assumption that we cannot predict how sediment properties might change throughout the length of a core.

ADDITIONAL NOTES ON CLEANING

CLEANING METHODS

During the course of this study two unique cases particular to two cores arose:

A number of cleaning methods were evaluated in this study to 1) Foraminifera from a sediment core treated with sodium hexa- determine which would be effective over a large range of cores and metaphosphate in another laboratory (often used to enhance dis-

Barium content of foram shells 3323

EWCa (pmol/mol)

•l

q

distilled waterlultrasonfc

+RED/+DTPA full cleaning

Orbulina ruber conglobatus dutertrei

Planktonic Species

FIG. 1. Comparison of Ba/Ca determinations for Holocene planktonic foraminifera cleaned with ultrasonication in distilled water vs. the full reductive/alkaline-DTPA cleaning described in text. Values represent averages of many analyses (Table 2).

aggregation of sediments) yielded Ba/Ca ratios two to twelve times minations were made on cores that were disaggregated in deionized higher than normal, even after prolonged cleaning (see data for water. core RC 13-228, Table 2 ) This is presumably due to precipitation 2) Planktonic foraminifera from 56 and 96 cm in Core V22-174 of an insoluble Ba-metaphosphate phase. All other Ba/Ca deter- yielded Ba/Ca ratios 10 to 100 times higher than those from 5

71

4

Ba/Ca (pmol/mol)

3

6

5

2

1

0 I

k1.5 (rl=3)

-RED/-DTPA

. 1;1’g(n=3Ji, pG--, 4

+RED/-DTPA

Cleaning Method

+REDI+DTPA

FIG. 2. Comparative cleaning of G. dutertrei shells from a sediment trap sample collected in the Panama Basin (WH4 11). Key: -RED/ -DTPA = all cleaning steps outlined in text (see Cleaning Method) with omission of the hydrazine reductive step and alkaline-DTPA step. +RED/-DTPA = all cleaning steps outlined in text with omission of the alkaline-DTPA step only. +RED/ +DTPA = All cleaning steps included.

3324 D. W. Lea and E. A. Boyle

Table 1: Locations of sediment cores, sediment traps and plankton tow used in this study

Latitude Longitude Water Depth

Sediment Cores m Atlantic Ocean: AI1107 65GGC 32’ 02’ S 36” 11’ W 2795 AI1107 67GGC 31055’S 36’ 12’ W 2587 AI1107 71GGC 31° 31’ s 35056’ w 1887 CHN82 4PC 41’43’N 32’ 51-W 3427 tXN82 11PC 42’23’ N 31”48’ W 3209 EN120GGCl 33” 4O’N 57O 37’W 4450 EN66 1OGGC 06’39’ N 21’54 W 3527 OC173-4 Core G 31”54’ N 64’ 18’ W 4469 RC13-228 220 20’ s 110 12’w 3204 v22- 174 10004’S 12’49’W 2630 Pacific Ocean: KNR73 3PC 00022’ s 106” 11’W 3606 TR163-27 02O 15’ s 86035’ w 3180 TR163-28 02O’19’ s 86’14’ W 3200 TRl63-31B 03O57’ s 85”58’ W 3210 TR163-32 02’29’ S 82’59’ W 2890 WH482-483 5’22’N 82’ 6’ W 3885

Sediment Traps SCIFF 31°N 64OW 3400 WH411 5’N 82’W 3354

Plankton Tow OC52-2 between 40“ N, 9’ W and O-10

70 s, 330 w

cm, even after protracted cleaning with alkaline-DTPA (Table 2). Similar results were found for foraminifera from this core in an earlier effort to quantify Ba in planktonic foraminifera (BOYLE,

I98 1). Examination of specimens of these individuals by SEM

indicates that the sediment in the high-Ba intervals was subject to a chemical event (possibly hydrothermal) which caused occlu- sion of pore spaces and precipitation of hydrothermal bar&es. Although the cleaning method employed did not remove the barite

Table 2: Ba/Ca in Planktonic Fotaminifera

Species Core Depth BalCa SD n r Cleaning Notes cm pmol/mol

Globigerinoiaks conglobahcs AII107.65GGC 6 0.82 0.11 2 reductive + DTPA AI1107.67GGC

f 0.82 0.05 2 reductive + DTPA

AI1107.71GGC 0.82 1 reductive + DTPA CHN82.4PC 5 0.60 reductive + DTPA CHN82.1 IPC

:“7 1.35 ; reductive + DTPA

ENlZO.GGCl 0.73 0.11 5 reductive + DTPA GCl73-4.G : 0.72 0.08 11 reductive + DTPA GCl73-4.G 0.77 0.08 2 std reductive GCl73-4.G : 1.53 0.89 4 cleaning

no reductive steu Globoquadrina dutertrei EN66.1 OGGC

;::6 1.0 3 reductive + DTPA

K73.4.3PC 12 0.03 3 reductive + DTPA K73.4.3PC 12 0.7 1 0.17 2 std reductive TR163.27 15 0.61 0.05 3 cleaning

reductive + DTPA TR163.28 ; 0.55 0.05 2 reductive + DTPA TR163.28 0.77 std reductive TRl63.31B 2 0.64 0.14 t8 3 reductive cleaning

+DTPA TR163.31B : 0.87 1 std reductive TR163.32 0.57 cleaning

reductive + DTPA WH411

:

WI-I411 !T 0.86 0.11 reductive + DTPA 2.0 0.9 3 std reductive WH411 ST 4.7 1.5 3 no reductive cleaning

WH482-483 1-6 1.29 0.05 2 step reductive + DTPA

WH482-483 1-6 7.8 2.2 3 no reductive step (note: SEM examination of shells from WH482-483 indicated ubiquitous ) batite occurencej

Globorotalia hirsuta OCl73-4.G : 4.8 0.9 7 1 std reductive GCl73-4.G 4.5 2.6 4 reductive + DTPA cleaning

Globorotalia menardii CKY73-4.G 13.1 6.2 3 std reductive (X52-2 L 6.5 2.2 2 cleaning

Orbulina universa std reductive cleaninr!

CHN82.4PC 5 0.63 0.06 2 reductive + DTPA CHN82.1 IPC 13 0.96 0.04 2 no reductive ENIZO.GGCl 5-55 0.66 0.24 15 1 reductive +DTPA step

EN66.1 OGGC : 0.98 0.26 2 1 reductive +DTPA GCl73-4.G 0.73 reductive + DTPA GC173-4.G z 0.86 0.12

: std reductive GC173-4.G 0.87 0.05 2 acid leach cleaning

GCl73-4.G 5 6.3 0.9 only

2 OC52-2 PT 0.50

water rinse only 1

RC13.228 :

2.2 1.6 3 std reductive cleaning

1 reductive + DTPA RC13.228 3.4 1.5 8 SCEF ST 0.83 0.02 2

std reductive cleaning std reductive TR163.3lB 2 1.40 1 cleaning no reductive step

Barium content of foram shells 3325

Table 2: Ba/Ca in Planktonic Foraminifera (cont.)

Species Core Depth BaKa SD n r Cleaning Notes Cm ~molhnol

Globigerinoides ruber EN12OGGCl 45-51 0.85 0.18 3 reductive + DTPA EN66.1OGCC 2 0.95 0.01 2 reductive + DTPA clC173-4.G

: 0.73

: reductive + DTPA

GC173-4.G 1.18 0.13 std reductive cleaning (X173-4.G :. 1.14 0.03 2 acid leach only OC173-4.G 3.52 0.16 2 water rinse only GC52-2 PT 1.01 0.12 5 std reductive cleaning wH411 ST 1.96 0.42 2 no reductive step WH482-483 1-6 6.10 no reductive step

Globigerinoides sacculifera EN 120.GGC 1 21-65 0.64 0.01 f!~ 1 reductive + DTPA EN66.lOGGC 2 0.94 0.21 4 reductive + DTPA KNR134-7.13GGC 0 - 1 1.04 0.19 4 reductive + DTPA CX!173-4.G & 0.85 0.06 2 std reductive cleaning OC52-2 0.66 0.02 2 std reductive cleaning TR163.28 2 0.80 1 std reductive cleaning TR163.31B 2 1.76 no reductive step v22- 174 5 0.55

; reductive + DTPA

V22-174 2:

22.5 :

reductive + DTPA V22-174 46.5 9.7 reductive + DTPA wH411 ST 1.59 1 WH482-483

std reductive cleaning 1-6 0.58

WH482-483 1-6 2.33 ; std reductive cleaning no reductive steu

Globorotalia truncatulinoides ENlZO.GGCl :

4.2 0.3 3 reductive + DTPA GC173-4.G 2.7 1.4 24 reductive + DTPA (X173-4.G 5 6.1 1.2 19 1 SCIFF

stdrcductivecleaning :; 3.3 2.5 3 reductive + DTPA

SCIFF 2.4 0.4 2 std reductive cleaning KEY: SD = standard deviation; n = number of samples included in mean; r = number of samples excluded from mean: PT = plankton tow; ST = sediment trap: cleaning notes: reductive + DTPA = full Cd/Ca cleaning procedure from Boyle and Keigwin (1985) and alkaline-DTPA step; std reductive cleaning = full Cd/Ca cleaning procedure; no reductive step = Cd/Ca cleaning procedure without Mn,Fe-oxide reduction step; acid leach only = sample was leached with O.lN nitric acid; water rinse only = sample was ultrasonicated and rinsed in distilled water

from the foraminifera in this core, this was the only such instance encountered in this study, as well as in determination of Ba con- tents of henthic foraminifera in over 40 Quatemary cores (LEA and BOYLE, 1989) and in two DSDP cores of Pliocene age (LEA, unpub. data). The hydrothermal barites that crystallized in fora- minifem shells from core V22- 174 are about 2 microns in diameter and are morphologically distinct from barites precipitated in sur- face waters. Inability to remove these barks during cleaning might be due to their precipitation in pore spaces in the shells.

Ba/Ca IN 3 GLOBIGERINOIDES SPECIES, ORBULINA SPP., AND GLOBOQUADRINA DUTERTREI

The range of Ba in the low latitude surface waters of the

world’s oceans is relatively small. Thirteen GEOSECS stations in the western Atlantic between 7Y’N and 33”s had dissolved Ba = 41 + 2 nmol/kg (CHAN et al., 1977). Six GEOSECS stations in the Pacific between 30”s and 30’N had dissolved Ba = 34 + 1 nmol/kg ( OSTLUND et al., 1987). The surface waters of the Southern Ocean and North Pacitic have enriched Ba due to out-cropping of cold, nutrient-enriched waters.

Ba/Ca in cleaned samples of the spinose planktonic fo- raminifera Globigerinoides sacculifera, G. ruber, G. conglo- batus, and Orbulina universa recovered from Holocene sed- iments primarily in the sub-tropical and Equatorial Atlantic have ratios of 0.8 f 0.1 rcmol/mol (Table 3). Ba/Ca in the non-spinose planktonic foraminifera Globoquadrina dutertrei from the equatorial Pacific have ratios of 0.6 f 0.1 rmol/ mol. The uniformity of values in the Atlantic and lower values for G. dutertrei from the Pacific both agree with the distri- bution of Ba in the surface oceans, although the difference is not clearly resolved above the standard deviation of the pooled data.

Measurements of Ba/Ca in foraminifera samples from sediment traps and plankton tows can be used to ascertain that Ba present in the foramimfera shells recovered from cores is not an artifact of sediment contamination. Three sets of samples were used (Fig. 3): 0. universa from a North Atlantic sediment trap (SCIFF) placed at 3400 m (3 1 “N, 64”W), G. dutertrei from an equatorial Pacific sediment trap (WH4 11) placed at 3354 m (5”N, 82”W); and G. sacculifera and G. ruber from a plankton tow in the northeast Atlantic (cruise Oceanus 52-2). These samples were cleaned with slightly dif- ferent procedures, as indicated in Table 2. Figure 3 indicates that there are no obvious differences between the Ba/Ca ratios of planktonic foraminifera samples from plankton tows, sed- iment traps, and sediment core tops. Therefore, the Ba con- tent of the shells does not appear to be changed by contact with sediments.

A further test for lattice-bound metals is sequential dis- solution of shells accompanied by determination of the metal to calcium ratio in each partial dissolution fraction.

Table 3. Mean Ba/Ca for certain planktonic foraminifera

Species No. of cores BalCa (pmol/mol)

G. conglobatw 6 (t=l) 0.75 f 0.09 Atlantic Orbulina spp. 4 G. ruber 3

G. sacculifera 2

ALL 15

0.75 f 0.16 0.84f0.11

0.79f0.15 0.77 f 0.12

Pacific G. dutertrei

r = number of rejected cores 5 (r=l) 0.57 f 0.07

3326 D. W. Lea and E. A. Boyle

Ba/Ca (~mol/mol)

1.4,

1 Pacific !

1.2 -

1.0 -

0.8 -

0.6 -

G. cfutertrei Orbulina spp.

Atlantic

G. her G. sacmlifera

Species

FIG. 3. Average Ba/Ca for 4 species of planktonic foraminifera from Atlantic and Pacific Ocean sediment core tops, a plankton tow from the Atlantic (OC52-2), and an eastern equatorial Pacific sediment trap (WH41 I). Standard deviations of the averages are indicated by error bars. Not all samples were cleaned with the alkaline-DTPA treatment (see Table 2). Data are tabulated in Table 2; sediment core top averages are listed in Table 3.

Reproducible values that do not change over the course of the dissolution suggest that Ba is in the calcite lattice, while a lack of reproducibility or systematically decreasing values might indicate that Ba is associated with another phase (BOYLE, 1981). A 4.5 mg sample of G. conglobatus (about 75 individuals) from box core OC173-4.G taken at 4469 m on the Bermuda Rise were cleaned as described above and sequentially dissolved. Ba/Ca was measured on nine disso- lution fractions (Fig. 4). Within the error of the measurement there was no change in Ba/Ca of the leachates. The results of this experiment indicate no evidence for contribution of Ba from external phases. Additionally, the relative constancy of the values suggests a homogeneous distribution of Ba in the calcite lattice of the foraminifera.

Measurements of Ba/Ca in G. sacculifea from the core top of an Eastern Mediterranean core by the ICP-MS method described elsewhere (LEA and BOYLE, 1989, 1990b; LEA. 1990) yield Ba/Ca = 1.0 t 0.2. Measurement of surface water Ba near this site indicates Ba of ‘about 53 nmol/ kg (LEA, 1990). The averages for planktonic foraminiferal Ba/ Ca for Holocene sediments from the Atlantic, Pacific, and Mediterranean are used to calculate a distribution coefficient D = { Ba/Ca}r,,,,,/ { Ba/Ca},,,,,. Excluded from this cal- culation are determinations on Globorotalia shells and one set of unexplained high values for G. dutertrei from core EN66- 1OGGC in the eastern equatorial Atlantic (see Table

2). For the five species of foraminifera from these cores D =

0.19 -t 0.05 (see Table 4). Figure 5 indicates the averaged Ba/Ca values plotted as a function of surface water Ba con- centration for each basin; although the one SD error bars are large enough to limit confidence in the relationship, the trend is consistent with proportional incorporation.

LEA and BOYLE ( 1989) report a mean distribution coef- ficient for Ba of 0.37 + 0.06 for three species of benthic fo- raminifera. The difference between the planktonic D reported here and the benthic D found previously might be attributed to the influence of temperature, pressure or other factors on Ba incorporation. Kitano and co-authors (KITANO et al., 197 1) found that under conditions of slow precipitation the distribution coefficient for Ba in inorganic calcite is about 0.1. However, during rapid precipitation achieved by “vig- orous stirring,” Ba distribution coefficients were as high as 3 during the early stages of precipitation and never dropped as low as 0.5 during the final stages of precipitation. However, distribution coefficients obtained via inorganic precipitation experiments are often not directly comparable to those de- termined by empirical calibration of a natural system (MORSE and BENDER, 1990).

BARIUM IN FOUR GLOBOROTALZA SPECIES

Analysis of Ba/Ca in the non-spinose foraminifera Glo- borotalia truncatulinoides and G. hirsuta from several cores

Barium content of foram shells 3327

&,/Ca ‘I ? (umol/mol) 2’o

1.5-

l.O-

- ~*3----++***~*-~* -....... +...._. .,.................... 0.5 -

H . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . g..” .*..- *p

. . . . . . . . . +..... G. conglobatus

_3t_ G. truncatulinoides

0.0 11 0 20 40 60 80 100

% Sample Dissolved

FIG. 4. Partial dissolutions of -5 mg samples of cleaned G. conglobatus shells and cleaned G. truncatulinoides shells. Ba/Ca is plotted vs. the % sample dissolved at each stage (see text). Ba/Ca of the first fraction of the shell dissolved is plotted on the far left of the figure: Ba/Ca of the last fraction of the shell dissolved is plotted on the far right of the figure. Dotted line. G. conglobatw solid line, G. truncatulinoides.

and a northwest Atlantic sediment trap indicate ratios three to nine times higher than those measured in the five species discussed in the preceding section (Table 2). A few mea- surements of Ba/Ca in Globorotalia menardii and G. inflata indicate high values for these species as well (Table 2). The only non-spinose species that yields Ba/Ca ratios consistently below 1 rmol/mol is Globoquadrina dutertrei taken from eastern equatorial Pacific sediment traps and core tops (Table 2). However, G. dutertrei shells from a single eastern equa- torial Atlantic core (EN66-IOGGC) have high Ba contents comparable to the values found for the Globorotalia shells (Table 2).

Partial dissolution of a G. truncatulinoides sample did not result in a reduction of Ba/Ca (Fig. 4). Attempts to use more extensive cleaning with alkaline-DTPA did eventually yield lower values, although Ba/Ca is never as low as in Globi- gerinoides (Fig. 6). Examination of G. truncatulinoides shells by SEM did not reveal the presence of barite particles, al- though this does not rule out the existence of barite particles smaller than 0.1 pm. The high Ba contents of sediment trap

samples of G. truncatulinoides suggests that Ba enrichment of the Globorotalia specimens is not simply related to sedi- ment contamination occurring when shells are deposited on the ocean floor.

The origin of high Ba/Ca in Globorotalia shells might be related to the unique lifestyle of this genera. For one, Glo- borotalia are known to calcify in the thermocline at depths of up to 1000 m (DEUSER et al., 1981; LOHMANN and SCHWEITZER, 1990). Since Ba increases with depth, Globor- otalia would be expected to precipitate some part of their shells in waters with higher Ba concentrations. In addition. the higher distribution coefficients which characterize benthic foraminiferal uptake of Ba (LEA and BOYLE, 1989) suggest that Ba incorporation increases with increasing depth. How- ever. even a combination of the distribution coefficient effect and the increase in Ba at depth would not fully account for the observed Ba/Ca ratios. since most Globorotalia samples have Ba/Ca values higher than what is found for benthic foraminifera at the same sites (LEA and BOYLE, 1989). Ex- planation of the high values might also be related to the

Table 4. Calculation of foraminiferal distribution co&Cents. (SW = surface water; foram = planktonic foraminifera)

Basin

Atlantic Pacific MeditClTaneaJl

@&;hlyam SW Ba nmoVkg

0.77M.12 419 0.57M.07 34fl 1.04M.19 53+1

SW Ca (Ba/Ca),, Droram mmovkg pmol/mol

10.3 4.0 0.19zLo.03 10.0 3.4 0.17zkO.02 11.3 4.7 0.22M.04

3328 D. W. Lea and E. A. Boyle

1.4

D=0.19+0.05

Foram BalCa (pmol/mol) 0.6-

0.4-

0.2-

0.0 I. I - I ’ I ’ I - I ’ 0 10 20 30 40 50 60 70

Surface Water Ba (nmol/kg)

PIG. 5. Mean Ba/Ca determinations for G. dutertrei from the equatorial Pacific (surface water Ba = 34 nmol/kg), 4 planktonic foraminifera species from the Atlantic (surface water Ba = 41 nmol/kg), and G. sacculifera from the Mediterranean Sea (surface water Ba = 53 nmol/kg) plotted vs. the respective surface water Ba concentrations for each basin (see Tables 3 and 4).

Table 5: Ba/Ca in EN120-GGCl Planktonic Foraminifera

Depth SW= BalCa SD n r cm

5 5 7

pmol/mol conglob 0.73 0.13

Orb [1.25,2.11] conglob 0.65 0.06

9 -tib 0.73 15 Orb 0.55 0.14 21 25 31

z: 39 39 4.5 45 45 51

::

:: 73

I:

:: 95

Orb sacc

% sacc Orb

rubcr sacc Orb

rubcr sacc Orb sacc

Z sacc Orb Orb [2.07] Orb

::z 0.62 0.65 0.58 0.58 0.65 0.58 0.19 0.68 0.64 0.01 0.46 0.82 0.64 0.57 0.62 0.77

%!I 0.59

0.74

4 2

2 1 1 3 1 1 1 1 1 1

: 1 2 1 1 1 1 1 1 1 1 1

1 1 1

1 1 1

!

: 1 1 1 1

;; conglob 0.64 sacc [I.751

115 0.67 121 2: 0.54 129 Fr: 0.63 135 0.59 147 Orb 0.73 153 conglob 0.86 153 Ek 0.66 167 0.83 179 0.82 205 :: 0.72 217 Orb 0.87 1

Notes: SD = standard deviation; n = number of samples included in mean; r = number of samples rejected from mean

precipitation of a late calcite crust (LOHMANN and

SCHWEITZER, 1990); large differences between the shell and the crust have been observed for Sr and Mg (P. LOHMANN,

pet-s. comm.).

RECONSTRUCTION OF SURFACE Ba IN THE ATLANTIC OVER THE LAST 14 KYR

Core EN 120~GGC 1 from the Bermuda Rise in the north- west Atlantic was raised from 4450 m water depth at 33”4O’N, 57’37’W (BOYLE and KEIGWIN, 1987). It contains a high sedimentation record (averaging 18 cm/ kyr) of the last 15,000 y; the oxygen isotope record (cibicidoides) indicates oxygen- 18 enriched glacial values (>3.7%) from 16 1 cm to the bot- tom of the core at 273 cm (Table 5). Age models based on radiocarbon dates on another Bermuda Rise core (GPCJ ) indicate higher sedimentation rates in the glacial intervals, so that the bottom 100 cm ofthe core covers less than a 1000 y ( KEIGWIN and JONES, 1989; L. KEIGWIN, pers. comm.).

Ba/Ca was analyzed in cleaned samples of Orbulina spp. over the top 2 17 cm, corresponding to the last 14 kyr (Fig. 7 ) . Orbulina spp. abundances fall off markedly in the glacial intervals, especially below 200 cm. In addition to the Orbulina spp. analyses, a few samples of G. conglobatus, G. ruber, and G. sacculifera from various depths in the core were analyzed for Ba. These are also shown on Fig. 7. No clear differences in Ba/Ca were noted from species to species, although it should be noted that the scatter between species at a given depth is equivalent to the full range of Ba/Ca values found for the Orbulina spp.-FIa/Ca record.

The data in Fig. 7 indicate that there might be a tendency to higher values during the late glacial period. Orbulina shells from the depth interval O-147 cm (O-12 kyr) have Ba/Ca

Barium content of foram shells 3329

6

BaEa (umol/mol) 4

q 0 0

l + 0

++ . ++ + l + +

2 ++++ +

.** + alkaline-DTPA 0 reductive cleaning only

01 . I 1 1 1 . 1 0 10 20 30 40 50 (

Recovery (%)

FIG. 6. Effect of alkaline-DTPA treatment on samples of G. truncutulinoides. Ba/Ca of samples is plotted vs. percent recovery of calcite (i.e., final CaCOs after cleaning divided by initial CaCOj before cleaning). Percent recovery is a measure of the DTPA cleaning time, with samples with the lowest recoveries having experienced the greatest degree of cleaning.

ratios of 0.62 + 0.08 rmol/mol (n = 19); Orbulina shells what factors might cause surface waters of the northwest At- from the depth interval 153-217 cm (12-14 kyr) have Ba/ lantic to have higher Ba concentrations. Evidence from a Ca ratios of 0.78 f 0.09 rmol/mol (n = 5), which is about survey of benthic Ba/Ca at the last glacial maximum indicates 25% higher. Although the scatter in the data reduces confi- that mean ocean Ba might have been 10 to 15% lower (LEA dence in the significance of this trend, it is worth considering and BOYLE, 1990b). Lower mean ocean Ba would lead to

100

Depth

(cm)

150

25

Ba/Ca (pmollmol)

0.0 0.2 0.4 0.6 0.8 t

4

6

8

10

Age W)

12

PIG. 7. Ba/Ca of cleaned planktonic foraminiferal shells in core ENl20-GCCI from the Bermuda rise (4450 m water depth) plotted as a function of sediment depth. Key to species code: squares = G. congiobatus; open circles = G. sacculifera; crosses = G. ruber; dotted squares = Orbulina spp. A line is drawn through the Orb&a data. Ages from L. KJXGWIN, pers. comm. and KEIGWIN and JONES (1989).

3330 D. W. Lea and E. A. Boyle

lower Ba in surface waters (LEA, 1990), the opposite trend to what is observed in Core EN1 20-GGC.

Increased riverine discharge or a meltwater spike draining into the Atlantic could cause an increase in surface Ba con- centrations. The residence time of Ba in warm surface waters (O-200 m) is about twenty-six years (calculated from a seven-

box ocean model in LEA and BOYLE, 1990b). Therefore, a meltwater spike discharging into the Gulf of Mexico via the Mississippi River system would probably retain its Ba sig- nature as it was advected into the North Atlantic via the Gulf Stream. If such a meltwater spike retained Ba concentrations comparable to the effective present-day river average of 400

nmol/kg (CHAN et al., 1976). had a flux of 1080 km3/yr (TELLER, 1990), and was integrated over a surface water residence time for Ba of twenty-six years, the total Ba anomaly would be:

26 y* 1080 km3/y* 10” L/km3*400 nmol/L

= 1.1 X 10” mol Ba. (1)

If this flux were spread out over 10% of the Atlantic surface

area and mixed to a depth of 200 m, the Ba concentration

anomaly would be

I. 1 X 10 lo mol Ba/

(0.10*82* lo6 km2* lo6 m/km’*200 m* 1000 L/m3)

= 7 nmol/L. (2)

This anomaly would lead to an enrichment in Atlantic surface waters of about 17%, within the bounds of the enrichment observed in shells from the 12- 14 kyr interval in core EN 120- GGCl.

The magnitude of a surface water Ba anomaly can also be evaluated by conversion of salinity anomalies found in an-

other Bermuda Rise core, KNR31 GPC-5 (mEWIN and JONES, 1989 ). The planktonic foraminiferal oxygen isotope

record from this core shows two clear 6 “0 minima dated at 13.5 kyr and I2 kyr B .P.; these events are interpreted as melt- water spikes caused by melting of the Laurentide ice sheet.

The 6’*0 anomalies of about -0.75%0 suggest negative sa- linity anomalies of about 0.5%0; this would correspond to a I .4% meltwater component, which would give rise to a Ba anomaly of about 10%. again using mean river Ba. However, this should be taken as a minimum anomaly for Ba since, as the meltwater mixes into the Atlantic, the salinity anomaly will be altered by evaporation on a much more rapid time scale than the Ba anomaly, given the long residence time of Ba in surface waters.

The scatter in the Ba/ Ca data from core EN 120-GGC 1 is in part attributable to analytical scatter. A consistency stan- dard of comparable concentration to EN 120-GGC 1 fora- minifera solutions was reproducible to about f 10% over the course of these analyses. The ability to do precise measure- ments of foraminiferal Ba by inductively coupled plasma mass

spectrometry suggests that improved planktonic Ba/Ca rec- ords can be obtained (LEA. 1990; LEA and BOYLE, 1989, 1990a. 1990b).

CONCLUSIONS

The Ba content of purified shells of three species of the planktonic foraminifera Globigerinoides conglobactus. G.

ruben, G. sacculijera, Orbulina spp., and Globoquadrinia du- tertrei from plankton tows, sediment traps, and core tops appear to reflect the Ba concentration of the seawater from which they precipitated. However, the small range of variation in surface water Ba, coupled with variability in the planktonic Ba/Ca data, limits confidence in this relationship. Barium contents of four species of Globorotalia are enriched over other planktonic foraminifera. possibly reflecting different depths, temperatures, and modes of shell precipitation; these

species appear to be poor candidates for reconstructing surface water Ba. Recovery of meaningful Ba/Ca values for plank- tonic foraminifera requires extensive purification of the shells to remove residual sedimentary phases which can contribute

Ba to the analysis. A record of planktonic foraminiferal Ba/Ca for a northwest

Atlantic core covering the last 14 kyr suggests that Ba in surface waters of the Atlantic might have been up to 25% higher during the interval 12-14 kyr, although the trend is

not clearly resolvable above the scatter in the data. Higher surface Ba might be of an indication of increase in riverine or melt-water fluxes to the Atlantic.

Acknowledgments-We thank Werner Deuser, Ellen Druffel, and Sus Honjo for providing sediment trap samples and Lloyd Keigwin and Ed Sholkovitz for providing sediment samples for this work. Com- ments and suggestions from Pat Lohmann, Lex van Geen, Paula Rosener, Kristin Orians, Debra Colodner, Kelly Kenison Falkner, Gary Km&hammer, and Robbie Toggweiler aided in the development of this study. We thank John Edmond for access to the MIT plas- maquad and use of the GEOSECS barium spike. Help and advice from members of the MIT plasmaquad squad was greatly appreciated. Core curation at WHOI, LDGO, and URI is supported by NSF. This research was supported by NSF grant 0CE87 IO 168 (to Edward A. Boyle) and a Joint Oceanographic Institutions/Ocean Drilling Pro- gram Fellowship to the senior author for 1987-1988.

Editorial handling: E. R. Sholkovitz

REFERENCES

BISHOP J. K. B. ( 1988) The barite-opal-organic carbon association in oceanic particulate matter. Nature 332, 341-343.

BOYLE E. A. ( 1981) Cadmium, zinc, copper, and barium in fora- minifera tests. Earth Planet. Sci. Lett. 53, 11-35.

BOYLE E. A. ( 1988) Cadmium: Chemical tracer ofdeepwater paleo- ceanography. Paleoceanography 3.47 l-489.

BOYLE E. A. and KEICWIN L. D. ( 1982) Deep circulation of the North Atlantic over the last 200,000 years: Geochemical evidence. Science 218,784-787.

BOYLE E. A. and KEIGWIN L. D. ( 1985) Comparison of Atlantic and Pacific paleochemical records for the last 215,000 years: Changes in deep ocean circulation and chemical inventories. Earth Planet. Sci. Lett. 16, 135-150.

BOYLE E. A. and KEIGWIN L. D. ( 1987) North Atlantic thermohahne circulation during the past 20,000 years linked to high-latitude surface temperature. Nature 330, 35-40.

CHAN L. H.. EDMOND J. M., STALLARD R. F., BROECKER W. S., CHUNG Y. C., WEISS R. F., and Ku T.-L. (1976) Radium and barium at GEOSECS stations in the Atlantic and Pacific. Earth Planet. Sci. Lett. 32, 258-267.

CHAN L. H.. DRUMMOND D., EDMOND J. M., and GRANT B. ( 1977 ) On the barium data from the Atlantic GEOSECS Expedition. Deep- Sea Rex 24,6 13-649.

CHOW T. J. and GOLDBERG E. D. ( 1960) On the marine geochemistry of barium. Geochim. Cosmochim. Acta 20, 192-198.

CHURCH T. M. ( 1979) Marine Barite. In Marine Minerals (ed. R. G. BURNS): Reviews in Mineralogy 6. pp. 175-209.

DEHAIRS F., CHESSELET R.. and JEDWAR J. ( 1980) Discrete sus-

Barium content of foram shells 3331

pended particles of barite and the barium cycle in the open ocean. Earth Planet. Sci. Lett. 49, 528-550.

DEUSER W. G., Ross E. H., HEMLEBEN C., and SPINDLER M. ( 198 I ) Seasonal changes in species composition, numbers, mass, size, and isotopic composition of planktonic foraminifera settling into the deep Sargasso Sea. Paleogeogr. Paleoclimatol. Paleoecol. 33, 103- 127.

KEIGWIN L. D. and JONES G. A. ( 1989) Glacial-Holocene stratig- raphy, chronology, and paleoceanographic observations on some North Atlantic sediment drifts. Deep-Sea Res. 36,845-867.

KITANO Y., KANAMOR~ N., and O~MORI T. ( 1971) Measurements of distribution coefficients of strontium and barium between car- bonate precipitate and solution-Abnormally high values of dis- tribution coefficients at early stages of carbonate formation. J. Geochem. 4, 183-206.

LEA D. W. ( 1990) Foraminiferal and coralline barium as paleocean- ographic tracers. Ph. D. thesis, Massachusetts Institute of Tech- nology/Woods Hole Oceanographic Institution.

LEA D. W. and BOYLE E. A. ( 1989) Barium content of benthic fo- raminifera controlled by bottom water composition. Nature 338, 751-753.

LEA D. W. and EWYLE E. A. ( 1990a) A 2 lO,OOO-year record ofbarium variability in the deep northwest Atlantic Ocean. Nature 347, 269- 272.

LEA D. W. and BOYLE E. A. ( 1990b) Foraminiferal reconstruction of barium distributions in water masses of the glacial oceans. Pa- leoceanography 5, 7 19-742.

LEA D. W. and SPERO H. J. (1991) Experimental verification ofbar- ium uptake in individual shells of the planktonic foraminifera Or- bulina Universa. Geochim. Cosmochim. Acta (in review).

LDHMANN G. P. and SCHWEITZER P. N. ( 1990) Globorotalia trun-

catulinoides’ growth and chemistry as probes of the past thermo- cline: 1. Shell size. Paleoceanography 5, 55-75.

MORSE J. W. and BENDER M. L. ( 1990) Partition coefficients in calcite-examination of factors influencing the validity of exper- imental results and their application to natural systems. Chem. Geol. 82,265-277.

OSTLUND H. G., CRAIG H., BROECKER W. S., and SPENCER D. W. ( 1987 j GEOSECS Atlantic. Pacitic. and Indian Ocean Expeditions. vol. i, Shorebased Data and &-aphics. National Science Foun- dation.

RINGBOM A. ( 1963) Complexation in Analytical Chemistry. Inter- science.

SEN GUFTA J. G. ( 1984) Determination of cerium in silicate rocks by electrothermal atomization in a furnace lined with tantalum foil. Talanta 31, 1053-1056.

SHANNON R. D. ( 1976) Revised effective ionic radii and systematic studies of interatomic distances in halides and chalcogenides. Acta Cryst. A32, 751-767.

SILL C. W. and WILLIS C. P. ( 1964) Precipitation of submicrogram quantities of thorium by barium sulfate and application to flou- rometric determination of thorium in mineralogical and biological samples. Anal. Chem. 36, 622-630.

SILL C. W. and WILIX C. P. ( 1966) Determination of radioisotopes of cerium, barium, lanthanum and neptunium after separation by barium sulfate. Anal. Chem. 38, 97-102.

TELLER J. T. ( 1990) Meltwater and precipitation runoff to the North Atlantic, Arctic and Gulf of Mexico from the Laurentide Ice Sheet and adjacent refions during the Younger Dryas. Paleoceunography 5,897-905.

VON DAMM K. L., EDMOND J. M., GRANT B., MEASURES C. I., WALDEN B., and WEISS R. F. ( 1985) Chemistry of submarine hydrothermal solutions at 2 I “N, East Pacific Rise. Geochim. Cos- mochim. Acta 49, 2 197-2220.