JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 97, NO. C7, PAGES 11,257-11,268, JULY 15, 1992

Isotopic Equilibration Between Dissolved and Suspended Particulate Lead in the Atlantic Ocean: Evidence From 210Pb and Stable Pb Isotopes

ROBERT M. $HERRELL AND EDWARD A. BOYLE

Department of Earth, Atmospheric and Planetary Sciences, Massachusetts Institute of Technology, Cambridge

BRUNO HAMELIN 1

Lamont-Doherty Geological Observatory, Columbia University, Palisades, New York

Decreased consumption of leaded gasoline in the United States over the past two decades has drastically altered the flux and isotopic composition of Pb entering the western North Atlantic from the atmosphere. Here we exploit the resulting temporal changes in the distribution and isotopic composition of oceanic Pb to investigate interactions between dissolved and particulate Pb in the oceanic water column. Measurements of dissolved Pb isotopic composition on samples collected in 1987 in the upper water column near Bermuda demonstrate that surface water 206pb/207pb decreased from -1.203 to -1.192 since 1983 and that a measurable change propagated to 300-500 m since the 1984 profile of Shen and Boyle (1988). The first accurate measurements of suspended particulate Pb in an open ocean profile show concentrations of 1-3 pmolFL, equal to 2-4% of total Pb. Vertical profiles of (1) the stable lead isotopic composition and (2) the ratio of total Pb to 210pb in suspended particles closely track contemporaneous depth variations in these ratios for dissolved Pb throughout the upper 2000 m of the Sargasso Sea near Bermuda. Thus suspended particles reach isotopic equilibrium with ambient seawater Pb on a time scale which is shorter than their residence time with respect to vertical removal, in agreement with equilibrium scavenging hypotheses based on interpretations of Th isotope distributions. A simple flux model suggests that the effect of deep ocean scavenging processes on the flux and isotopic composition of Pb sinking on large particles was minor throughout the preanthropogenic and most of the anthropogenic era but has become important as surface inputs decrease to pre-leaded gasoline levels and may exceed the contribution of surface-derived Pb flux in the next decade.

1. INTRODUCTION

How rapidly do metals exchange between solution and suspended particles? This central question in oceanic trace metal scavenging can be tested by analyzing suspended particles in the Sargasso Sea near Bermuda for total Pb, 210pb ' and stable Pb isotopic composition. If exchange is rapid, particles will exhibit similar depth variations in Pb isotopic composition to those observed for Pb dissolved in seawater.

The distribution of 230Th and 234Th, highly particle-reactive radionuclides with a uniform source throughout the ocean, has been shown to be consistent with a reversible uptake mechanism characterized by exchange times which are rapid (months) in relation to the residence time of suspended particles (years) [Bacon and Anderson, 1982; Nozaki et al., 1987]. The implied state of approximate exchange equilibrium has not been demonstrated for less reactive trace metals with sources at the

ocean's boundaries rather than in its interior. Lead, in particular, is less reactive than Th and has a deep ocean scavenging residence time of 100-200 years [Rama et al., 1961; Craig et al., 1973; Bacon et al., 1976; Nozaki et al., 1976, 1980] compared with a few decades for Th [Moore and Sackett, 1964; Anderson et al., 1983].

•Now at G•osciences de l'Environnement, Universit• d'Aix-Marseille, Marseille, France.

Copyright 1992 by the American Geophysical Union.

Paper number 92JC00759. 0148-0227/92/92JC-00759505.00

Anthropogenic input during the last century has completely altered the natural distribution of Pb in the Sargasso Sea [Schaule and Patterson, 1983; Boyle et al., 1986]. Downwind from North American industrial and automotive sources, the Pb distribution in

this region is no longer in steady state. The atmospheric flux of pollutant lead to the Sargasso Sea increased to reach a maximum in 1970, when the use of alkyl-leaded gasolines began to be curtailed in the United States, and has been decreasing ever since. The Pb concentration in the upper ocean near Bermuda, recorded in the skeletons of corals, has been decreasing in response since the early 1970s [Shen and Boyle, 1987].

Diminished lead levels have been accompanied by fi simultaneous change in stable Pb isotopic composition. The 206pb/207pb ratio of Bermuda surface water decreased from 1.195-1.210 in 1983-1984 [Shen and Boyle, 1988] to 1.187-1.197 in 1987 [Hamelin et al., 1988]. Recent work by A. Veron et al. (Response of lead cycling in the surface Sargasso Sea to changes in the tropospheric input, submitted to Journal of Geophysical Research, 1992, hereinafter referred to as Veron et al., 1992) demonstrates that the isotopic shift continued to 1989, when surface 206pb/207pb had fallen to 1.177-1.188. The rapid isotopic evolution is caused by a reduction of the contribution from North American gasoline Pb which was mined from unusually radiogenic ores, a relative increase in the fraction of Pb from less radiogenic industrial sources (U.S. iron/steel production, primary Pb production, coal combustion, and metal smelting) [Shen and Boyle, 1987], and a small but increasing fraction from European pollution sources (Veron et al., 1992). The future progression of the Pb isotopic composition of northwest Atlantic surface waters will depend largely on implementation of controls

11,257

11,258 SHERRELL ET AL.: ISOTOPIC EQUII.IltRATION OF LEAD IN SARGASSO SEA

on U.S. industrial mad European gasoline sources. As U.S. alkyl- Pb input becomes negligible, Pb concentrations may level off with little further alteration of the isotopic composition [Sheri and Boyle, 1988].

Pb-210 is produced in the atmosphere by decay of continentally derived gaseous 222Rn. It enters the surface of the western North Arianfie via atmospheric pathways similar to those followed by anthropogenic Pb [Settle et al., 1982; Talbot and Andten, 1983]. Thus in the Sargasso Sea, 210pb can be used to normalize short-term variations in surface water Pb concentrations

in order to reveal a secular trend in pollution Pb input [Boyle et al., 1986]. Pb-210 is also generated by decay of its dissolved grandparent 226Ra in the ocean interior. The oceanic distribution of 210pb ' in contrast to that of stable Pb, is in approximate steady state [Rama et al., 1961; Craig et al., 1973; Bacon et al., 1976; Nozaki et al., 1976]. As a result of the reduced anthropogenic Pb input, the dissolved pb/210pb prof'fie shows substantial changes in the upper water column of the northwest Arimatic over the last decade [Boyle et al., 1986; Shen and Boyle, 1988]. The pb/210pb ratio in Sargasso Sea surface water fell from -1250 pmol/dpm in 1979 to 690 pmol/dpm in 1984 [Boyle et al., 1986] mad to -450 pmol/dpm by 1986 [E. A. Boyle mad M.P. Bacon, unpublished data]

Thus the evolution of both 206pb/207pb and pb/210pb in surface waters provides the opportunity to use the ratios as complementary paired, semiindependent, time-variam tracers of Pb transport in the northwest Atlantic. The vertical propagation of these changing surface signals into the ocean interior has produced isotopic depth distributions which are evolving on a yearly to decadal time scale. Boyle et al. [1986] and Shen and Boyle [1988] used these variations as transient tracers to constrain the effect of therrnocline ventilation on chemical distributions in

the Sargasso Sea. In this study, we make use of this inadvertent tracer experiment, using measurements on suspended particulate matter to further investigate the role of particles in the transfer of Pb from surface waters to the deep ocean.

2. SAMPLING AND ANALYSIS

Suspended particulate matter was filtered in situ at depths of 10-4000 m on five separate occasions from September 1986 to April 1988 near station "S," -50 km southwest of Bermuda (32ø00'N, 64ø10'W, bottom depth -4200 m) [Sherrell and Boyle, 1992]. Sample contamination is a,-a acute concern in obtaining accurate Pb concentrations in natural waters [Schaule and Patterson, 1981], and so specially designed battery-powered pumps (the Rotating Automatic Pump for Particulate Inorganics Determination, RAPPID [Sherrell, 1991]) were used to collect the samples in situ. Details of the sampling mad analytical procedures are presented elsewhere [Sherrell, 1991; Sherrell and Boyle, 1992]. Briefly, 4-11 mg of suspended matter was collected from volumes of 200-1200 L of seawater on 142-mm-diameter 1.0-lxm pore size Nuclepore polycarbonate filters. Small suspended particles are expected to dominate exchange with dissolved Pb because of their abundance mad surface area. Therefore, a 53-lxm polyester prefilter was used on most deployments to exclude larger, fast-sinking particles (Table 1). The >53-lxm fraction contributed 5-25% of the total suspended mass, the fraction generally decreasing with depth through the upper thermocline. However, any compositional differences between the <53-lxm samples and total particulate matter samples, collected with no prefilter, could not be discerned for Pb or Pb isotopes given the

sample coverage mad measurement precision of this study. Therefore samples taken with mad without prefilters are treated as equivalent in this study. In addition, no temporal variability was resolvable between particulate profries obtained during this study; the results of five station occupations are presented as a composite profile.

Sample filters were rinsed with p H 8.3 distilled water, dried, weighed, mad subsampled. For determination of Pb and 210pb, three replicate subsamples (each 8% of total filter area) were dissolved separately in triple Vycor-distilled concentrated HNO3 with 40 !xL/mL subboiling distilled HF by heating in closed Teflon containers at ~100øC for 4 hours. Resulting solutions were then diluted by a factor of 5 with distilled, deionized water and analyzed for total Pb by graphite furnace atomic absorption spectroscopy, using the method of standard additions. Total uncertainty was limited not by a,-aalytical error (+3% based on replicate analyses of a single digest solution) but by subsampling uncertainty, which was generally +5-10%, determined as the standard deviation of 3-4 replicates. Blanks for both unused and "dipped" filters (full deployment on nonoperating pump) were generally <5% of sample signals, except in low-Pb near-surface samples, where blanks were <25%. Total uncertainty increased to about +35% in some surface samples. Known fractions of each subsample digest solution were pooled, representing ~20% of the total sample, and analyzed for 210pb by alpha spectrometric determination of the activity of ingrown daughter 210po (September 1987 mad March 1988 samples only). This was accomplished by adding 4.0 mL 70% perchloric acid, spiking with known quantities of 209pb mad carrier Pb, evaporating the solution to near dryness in a Teflon beaker, and thereafter following the method of Fleer and Bacon [1984]. Total procedural blanks for the 210pb a,-aalyses were 3-35% of sample values. Separate filter subsamples (4% of filter area, March 1987 and March 1988 samples only) were leached in 0.5 M HBr for stable Pb isotope measurements. The leachares were loaded directly onto 0.2-mL anion exchange resin columns (Dowex AG-1 X8) and eluted with 0.5 M HNO 3. The samples were loaded on rhenium filaments using standard H3PO4-SiO2 procedures mad were run on a VG Micromass 30 spectrometer, at temperatures of 1200ø-1300øC. Total Pb loaded varied from 2 to 22 ng. A series of about five blocks of six ratios were first acquired on a Daly detector; then five more blocks were taken at higher temperature using the Faraday cup. The two data sets were corrected for fractionafion using mean values of measurements of National Bureau of Standards Standard Reference Material 981 [aken

repeatedly with each detection system. The two data sets always agreed within error. Only Faraday cup data are presented here, with precision of the 206pb/207pb ratio equal to +0.001. Total Pb blanks for the purification steps were <10 pg. Blanks were somewhat lower than the atomic absorption blanks, probably because the milder leaching conditions resulted in a lower filter blank.

Lead concentration in the HBr leachates was determined using a 205pb spike and the method of isotope dilution. About 10 pg of the spike was added to the sample before purification. Thallium- 203 was measured to monitor 205T1 interference, which was always negligible in the temperature range 1150ø-1200øC. Uncertainties on Pb concentrations are estimated at 5%, mainly due to counting statistics on very small Pb peaks measured on the analog Daly detector. The particulate Pb concentrations determined by the mass spectrometric method using isotope dilution were equivalent to the atomic absorption analyses, with

SHERRELL ET AL.: ISOTOPIC EQUILIBRATION OF LEAD IN SARGASSO SEA 11,259

11,260 SIIERREI.I. ET AL.: ISOTOPIC EQUILIBRATION OF LEAD IN SARGA$$O SEA

Particulate

Pb (pmol/L)

o.o 1 .o 2.0

o

500

lOOO

1500

2000 ß

2500 ß

3000

3500 -

(a)

4000'

Particulate

Pb-210 (dpm/100kg)

3.0

o

1986-88

01 O.2 O.3 0.4 ß I -- ß I . ] •

(b)

1987-88

0.5

Fig. 1. (a) Composite vertical depth profile of suspended particulate Pb near Bermuda from five station occupations: September 1986 (open squares), March 1987 (open circles), September 1987 (open triangles), March 1988 (closed triangles), and April 1988 (closed squares). Replicate symbols connected by line segments represent replicate subsamples of the same f'dter. One point at 330 m (open circle with line) obtained with 0.4-1am filter (see text). (b) Suspended particulate 210pb near Bermuda in 1987-1988. Symbols as in Figure la.

subsampling uncertainty obscuring any clear depth dependence (mass spectrometry divided by atomic absorption = 0.91 + 0.16 [lo] for 14 samples). It is therefore reasonable to conclude that both leaching/dissolution procedures dissolved essentially all of the particulate Pb and that the isotopic ratios are representative of total particulate Pb. This agreement is consistent with an estimated refractory aluminosilicate Pb fraction of <5%, as discussed below. The only exception was the 3800-m sample, at which depth the aluminosilicate Pb fraction is estimated at 24% [Sherrell and Boyle, 1992] and which showed evidence of resuspension in Pb concentration and isotopic composition (see section 3.2 below).

Nine seawater samples were collected for dissolved Pb determination from the upper 600 m at the Bermuda station (31ø483q, 64ø07'W) in March 1987 using special "vane" samplers [Boyle et al., 1986]. These were analyzed by APDC/DDDC solvent extraction of 50-g aliquots following the method of Wallace et al. [1983]. Analysis of a few surface samples previously analyzed by APDC coprecipitation in our laboratory [Boyle et al., 1986] showed that the two methods agreed within

analytical uncertainty. Stable Pb isotopes were also measured on separate aliquots of the same samples [Hamelin et al., 1988]. The analytical technique for these measurements will be reported separately.

3. RESULTS

3.1. Pb and 210pb Concentrations

The 1986-1988 suspended particulate Pb concentrations increase from 0.8-1.3 pmol/L in near-surface waters (0-100 m) to a maximum of 2.9 pmol/L at about 500 m, then decrease to approximately constant deepwater values of 1.5-2.0 pmol/L (Table 1; Figure la). The profile shape resembles the 1987 dissolved Pb distribution but shows a steeper gradient at 0-500 m and a less pronounced decrease below 500 m (Figure 2a). This difference reflects the increasing fraction of Pb in particulate phases, from 1-2% of total Pb in surface waters to about 4% in deep water. The particulate Pb content (per gram dry weight of suspended material) increases from about 50 nmol/g in near-

SHEP. RE• ET AL.: ISOTOPIC EQUILJBRATION OF LEAD IN SARGASSO SEA 11,261

Total Pb (pmol/kg) Pb-210 (dpm/100kg)

0

0

5OO

IOO0

1500'

2OOO-

ß

2500-

5O IO0 150 5 10

1987 '%.

o

1984

o

o

o (a)

-I-

o

15 20 25

o -I- +

-I- -I-

1983-84

(b)

Fig. 2. (a) Dissolved (unfiltered) Pb at Bermuda. Data from 1984 for three station occupations from Boyle et al. [1986] (open s lv 210 circles, open triangles, crosses). Estimated discrete depth mean 1984 data u ed for disso ed Pb/ Pb calculation plotted in

Figure 3a are shown as line graph with small solid squares. Dissolved Pb data from 1987 shown as dotted line with solid circles. (b) Dissolved (unfiltered) 2IOpb at Bermuda for several station occupations in 1983-1984 from Boyle et al. [1986]. Estimated discrete depth means used for dissolved Pb/210pb calculation plotted in Figure 3a are shown as line graph with small solid squares.

surface waters to an approximately constant value of 350 nmol/g from 500 m to deep waters, and decreases again to -140 nmol/g in samples at 3800-4000 m because of dilution by low-Pb resuspended material in the nepheloid layer (Table 1).

A single comparison of 0.4- and 1.0-1•m Nuclepore filters (300 m, March 1987) showed no significant differences in Pb concentration, suggesting that the major portion of particulate Pb is retained on the larger pore size filter used throughout the study. More recent pore size comparison experiments carried out in the northeast Atlantic generally support this result, although significantly less Pb was collected on the 0.4-1•m filter in one comparison, suggesting a possible effect of flow rate on size fractionation or other operational effects (R. M. Sherrell and E. A. Boyle, manuscript in preparation). This same 300-m point coincidentally falls below the trend def'med by the rest of the profile (Figure la). Particulate Pb concentration for these samples is about half that expected from the profile defined by upper water column samples from other station occupations, probably a result of a deep late winter mixed layer, which forced particulate Pb values to levels typical of surface waters.

Based on average crustal Pb/A1 [Taylor, 1964] and total A1 measured in the particles [Sherrell and Boyle, 1992], less than 5% of particulate Pb is associated with refractory aluminosilicate particles, except in the nepheloid layer (25% at 3800-4000 m). In addition, it has been estimated that almost all anthropogenic Pb delivered from the atmosphere is already dissolved in rain

[Jickells et al., 1984; Mating and Duce, 1990; Veron et al., 1992]. Therefore most of the particulate Pb observed in this water column is is likely to have been acquired from solution, although not necessarily in the local environment. Similar arguments can be made for a minor fraction of nonexchangeable refractory 210pb.

We believe this is the first accurate measurement of fine

suspended particulate Pb in the subsurface open ocean. One published study of surface particle composition near Bermuda in 1974 gave particulate Pb concentrations -5 times higher than ours [Wallace et al., 1977], roughly consistent with the drop in surface water dissolved Pb concentrations over this period. Previous investigators of deep ocean particulate chemistry analyzed particles filtered from Niskin bottles (0.4-1xm filter) and obtained an average deepwater Pb concentration of 42 pmo!/kg for the North Atlantic in 1974-1975 [Buat-Menard and Chesselet, 1979]. Because dissolved Pb in surface waters has changed by only a factor of-2 during 1979-1987 [Schaule and Patterson, 1983; this study] and by much less in deep waters, such a high value is unlikely to be the result of historical input variations and is more likely due to sample contamination. A 1985 study of >10-[xm suspended particulate Pb in the northeast Atlantic found midwater and deepwater concentrations for this "large" particle fraction within a factor of 2 of our >l.0-1•m Sargasso Sea values, at stations with somewhat higher dissolved Pb [Lambert et al., 1991a,b].

11,262 SHERRELL ET AL.: ISOTOPIC EQUILIBRATION OF LEAD IN SARGASSO SEA

The adequacy of contamination control in our study is supported by three lines of evidence: (1) dipped procedural blanks are low in relation to sample values, (2) the profile is smoothly varying with depth, and (3) deepwater samples taken at similar depths several months apart give concentrations which agree within measurement uncertainty. Our recent results for particulate concentrations of several other trace metals similarly demonstrate that most previously published values were high by factors of 3 to 30 [Sherrell and Boyle, 1992].

An upper thermocline particulate 210pb maximum, similar to that found for particulate Pb, is also observed (Figure lb). Because no subsurface maximum is observed for dissolved 210pb' (Figure 2b), this result implies that the particulate Pb maximum is not simply a product of the temporal history of anthropogenic Pb input (which determines the shape of the dissolved profile) [Boyle et al., 1986], but that the maximum is caused by the higher affinity of suspended matter for Pb from 300-600 m. This may be due in part to association of Pb with Mn hydrous oxide phases, which also show a maximum at this depth [Sherrell and Boyle, 1992]. The maximum could alternatively be caused by a progessive enrichment of a single population of particles during transport from surface waters, without invoking the influence of a separate phase in this depth zone. Such a mechanism cannot be ruled out by the isotopic evidence, as will be explained below. Between 750 m and 2000 m, particulate 210pb increases by ~40%, quantitatively consistent with the increase in the particulate fraction of total Pb over this depth interval.

The new dissolved Pb data presented here show values of ~70 pmol/kg in the upper 300 m, with a subsurface maximum of ~100 pmo!/kg at 100-150 m (Table 2, Figure 2a). Below 300 m, values increase to converge with the 1984 profile at 600 m, as would be expected in view of the longer residence time of dissolved Pb in deeper waters. A mean profile of dissolved 210pb was estimated from previously published 1984 data (Figure 2b), and dissolved pb/210pb profiles were estimated for 1979, 1984, and 1987. The dissolved pb/210pb profile evolves from 1979 to 1987 because of the decrease in dissolved Pb in the upper ~500 m induced by decreased U.S. gasoline Pb usage (Figure 3a). By 1984, an upper thermocline pb/210pb maximum was generated by the interaction of decreased surface Pb input and thermocline ventilation [Boyle et al., 1986]. The new data demonstrate that the trend is continued to 1987, when the surface Pb/210pb fell to ~400 pmol/dpm.

TABLE 2. Dissolved Pb at Bermuda, March 15, 1987

Depth, Pb, m pmol/kg

47 71

98 103

148 97

198 71

299 75

349 109

399 115

500 179 *

601 126

*Contamination suspected.

The 1987-1988 profile of pb/210pb in suspended particles tracks the 1987 dissolved ratio from the surface down to 600 m

(Figure 3a). Measurement uncertainties may not allow resolution of temporal changes in pb/210pb on the time scale of 1 year, but the evolution of upper thermocline Pb/210pb in seawater and particles between 1984 and 1987 clearly exceeds analytical and estimation error (Figure 3a). Below 600 m, the 1987-1988 particulate pb/210pb profile tracks the 1984 dissolved profile.

3.2. Stable Pb Isotopes

Lead isotope determinations on 20 seawater samples collected in March 1987 in the northwest Atlantic gyre [Hamelin et al., 1988] demonstrate that surface water 206pb/207pb shifted from 1.203 in 1983 [Shen and Boyle, 1988] to about 1.192 in 1987. This decrease occurred because of the diminished contribution

from radiogenic gasoline Pb and may have been influenced as well by a change to less radiogenic ores for U.S. alkyl-Pb produced during this phase-out period. The profiles of suspended particulate and dissolved 206pb/207pb demonstrate that the change in the surface water Pb isotopic composition propagated to about 500 m under the influence of thermocline ventilation and

vertical particulate transport [Boyle et al., 1986] (Table 1; Figure 3b). At ~500 m, 1987-88 206pb/207pb values converge with the 1984 dissolved Pb isotope profile determined by Shen and Boyle [1988], again consistent with the longer residence time for Pb in deep waters.

The particulate Pb isotopic composition closely tracks contemporaneous dissolved ratios throughout the upper 2000 m of the water column (Table 1, Figure 3b; we report all measured ratios but plot only 206pb/207pb, since this ratio is measured with the greatest precision). As for Pb/210Pb ' the 300-m sample from March 1987 falls off the trend, more closely matching the 100-m sample. Below 2000 m, particulate 206pb/207pb increases again, following the trend of anthropogenic Pb input before 1930 [Shen and Boyle, 1988] (Figure 3b). The 3800m sample may be influenced by resuspended surface sediment Pb; its isotopic composition falls in the range of those observed for nonanthropogenic Atlantic abyssal sediments [Hamelin et al., 1990].

4. DISCUSSION

Results for both tracers, 206pb/207pb and Pb/210pb ' show a strong coherence between suspended particulate and dissolved Pb isotopic composition from surface waters to 2000 m. The preliminary implication is that particulate Pb undergoes complete isotopic exchange with dissolved Pb on a time scale which is short in relation to the residence time of suspended particles and to the time rate of change of the dissolved Pb isotope distribution.

The combined evidence from 206pb/207pb and pb/210pb makes a stronger argument for isotopic equilibration than would either tracer used alone. While stable Pb isotope ratios provide an absolute "fingerprint" of Pb source history, the Pb/210pb tracer currently has a larger surface/deep signal relative to measurement precision. In addition, the same particulate 206pb/207pb values occur at different depths in the 1987 profile (Figure 3b). On the basis of this tracer alone, for example, it cannot be determined whether particulate Pb at 1000 m is derived from ambient or surface waters. Pb/210Pb does not suffer the same ambiguity; the low surface value provides proof that deep particulate Pb content

SHERRELL ET AL.: ISOTOPIC EQUILIBRATION OF LEAD IN SARGASSO SEA 11,263

Pb/Pb-210

(pmol/dpm) Pb 206/207

500

lOOO

1500

2O00

2OO 400 600 800 1000 1200 , ...•1. , ! , ! , I I ]

,•, .: ! ! I

?, 1

ß ' I

/ / ' : I

1 984 i • 1 979 : I _.

• I : I ß

.: ... /

.: ... !

..' ! ! / !/

(a)

1400 118

o

500

lOOO

15oo

2000 -

3000 - .

4000 - ,

1.19 1.20 I • I •

:. <-dissolved 1984 ß

• .,.,"'

1.21

<-dissolved

1987

0

<-particulate 1987-88 ( b )

0

Fig. 3. (a) Dissolved and suspended particulate pb/210pb ratio. Dissolved profiles were estimated for 1979 (dashed line), 1984 (dotted line), and 1987 (solid line) from dissolved Pb data of Schaule and Patterson [1983] (profile 250 km northwest of Bermuda at 34'15'N, 66'17•N), Boyle et al. [1986], and this study, using mean 1984 210pb data (Figure 2b). Particulate values shown as individual points with error bars representing estimated precision (4-10%) on the ratio (sy•_bols as per Figure 1). Note 100-m point is two virtually superimposed particulate values. (b) Dissolved and particulate 206pb/207pb ratio. Dissolved profile for 1984 from Shen and Boyle [1988] (dotted line and open squares). Dissolved 1987 profile from this study (solid line and open circles). Particulate ratios for 1987-1988 shown as individual points (symbols as per Figure 1). Note depth scale change at 2200 m. Boxes below plot represent ranges for bulk sediment in 0- to 5-cm interval at two Sargasso Sea sites, the Hatteras Abyssal Plain 0t) and near Bermuda (B) [Hamelin et al., 1990].

must be governed by exchange with surrounding seawater, not by simple disaggregation of large surface-derived particles. Finally, surface 206pb/207pb may not fall below the pregasoline Pb value of-1.180 as long as non-gasoline anthropogenic sources are important [Sheri and Boyle, 1988], but pb/210pb will either level off or continue to decrease. The 1987 depth gradient seen in dissolved pb/210pb provides a signal that will continue to be a useful tracer of vertical Pb transport in the Sargasso Sea for the

next few decades. In contrast, the entire depth variation in stable Pb isotope ratios near Bermuda is very small in comparison with the global range observed in aerosols, sediments, and primary ore bodies [Hamelin et al., 1989]. The utility of 206pb/207pb might be enhanced in another oceanic water colunto influenced by a more variable isotopic history of Pb input or which is poised at a point in the pollutant source trend which results in larger isotopic depth gradients (e.g., eastern North Atlantic).

11,264 SHERRELL ET AL.: ISOTOPIC EQUILIBRATION OF LEAD IN SARGASSO SEA

4.1. Do Similar Isotopic Ratios Imply Chemical Exchange Equilibrium?

Use of isotopic ratios as tracers of chemical exchange among Pb reservoirs requires careful consideration of the implications of similarities observed in the isotopic signatures of two different Pb pools. While chemical equilibrium must result in isotopic equilibration, nearly identical isotopic signatures can be generated by irreversible processes.

For example, in surface waters, living or recently living forms dominate the suspended load; exogenous particles are a relatively minor fraction of the total mass [Sherrell and Boyle, 1992]. Because these particles are created de novo, they must have acquired their Pb directly from surface waters. Therefore, in the euphotic zone (upper 100 m), a close match between particulate and dissolved Pb composition is expected, and its existence cannot help distinguish irreversible uptake from exchange equilibrium as the process governing particulate Pb content.

Deeper in the water column, particle sources are more complex; some may be formed in situ, but most are likely to have settled vertically or to have been transported horizontally to their sampling location. Particles transported from other water masses with distinct Pb isotopic composition must exchange Pb with ambient seawater to acquire a matching isotopic signature. The agreement between the particulate and dissolved profiles below the euphotic zone thus suggests that suspended particles undergo complete isotopic exchange of Pb with the surrounding seawater before they are removed.

This conclusion would be thrown in doubt if the dominant

source of suspended particles at all depths were the disaggregation of surface-derived aggregates. An important consideration in this case is that the Pb content of surface particles (mol per gram dry weight) is about sevenfold lower than that of particles at 500 m (50 versus 350 nmol/g). From 500 m to 2000 m, Pb content is roughly constant. (The decrease in particulate Pb concentration per volume seawater thus reflects decreasing suspended mass, Figure l a.) Therefore assuming that Pb content is preserved during aggregation of surface suspended particles and during subsequent sinking and disaggregation at depth, the Pb-poor particles of recent surface origin could increase in Pb content by a factor of 7 and nearly acquire the deep dissolved isotopic signature by irreversible uptake, independent of chemical equilibrium.

This scenario seems unlikely, however, because such a simple model unrealistically assumes that suspended material at all depths is supplied only from the surface, i.e., any reaggregation of suspended material in subsurface layers results in direct removal to the sediments, with no subsequent disaggregation to contribute to the suspended pool in deeper waters. A more realistic model, supported by Th isotope distributions, suggests multiple aggregation/disaggregation exchanges occurring between fast- sinking large particles and the standing stock of suspended particles as material is transported to the seafloor [Bacon et al., 1985]. Thus the close match between particulate and dissolved Pb isotopic composition in the interval 500-2000 m, where Pb content is invariant, implies that particles transported vertically within this interval undergo rapid exchange of Pb with ambient seawater before being removed. Such a state of approximate exchange equilibrium has been suggested to control the water column distribution of Th isotopes, but the importance of reversible exchange in scavenging processes for other trace metals could not be evaluated [Bacon and Anderson, 1982; Nozaki et al., 1987]. We now see evidence from these isotopic tracers that suspended

particulate Pb, as well, is controlled by exchange with the much larger dissolved reservoir.

The evidence for exchange equilibrium suggests that dissolved/particulate fractionation for Pb might be predictably constant throughout the sub-surface ocean. Indeed, our comparison of "distribution coefficients" (mol/g particles divided by mol/g seawater) in midwaters of the Sargasso Sea and the northeast Pacific shows similar values over a fivefold difference

in dissolved Pb concentration, for particles with substantially different bulk composition (R. M. Sherrell and E. A. Boyle, manuscript in preparation). The applicability of an equilibrium exchange model for trace metals other than Th and Pb remains unknown, although recent evidence from the distribution of neodymium isotopes suggests that the behavior of this element is generally consistent with equilibrium exchange (C. Jeandel et al., 1992, Neodymium: a tool for tracing exchange processes within the oceanic water column? The Sargasso Sea example, submitted to Geochimica Cosmochimica Acta, 1992). In any case, we argue in the following discussion that the existence of dissolved/particulate exchange equilibrium in the deep ocean has little bearing on Pb removal fluxes.

4.2. A Simple Model of SuspendedlSinking Particle Interactions

Isotopic identification of anthropogenic Pb in surface sediments of the deep ocean has demonstrated that Pb can be transported rapidly from surface sources to the bottom [Veron et al., 1987; Hamelin et al., 1990], probably in the form of large, rapidly sinking aggregates [Fowler and Knauer, 1986; Lambert et al., 1991b]. This finding suggests that sediment Pb isotopic tracers might provide insight into chemical exchanges among Pb pools in the overlying water column. However, because of bioturbafion and slow sedimentation rates at deep ocean sims, the anthropogenic Pb isotope signature in surface sediments is an integrated mean of all historical input sources [Lambert et al., 1991b]. Thus water column processes which control the magnitude and isotopic composition of the Pb flux cannot be investigated using measurements in deep ocean sediments.

One might expect that analysis of material collected in deep sediment traps could shed light on the influence of exchanges between dissolved, suspended particulate, and sinking particle pools on the composition of the removal flux. However, few if any reliable stable Pb isotope data for sediment trap material have been published [Shen and Boyle, 1988]. Later we will show that pb/210pb ratios in deep time series traps near Bermuda may be too variable to provide unequivocal evidence of exchanges in the water column.

Faced with a lack of direct isotopic evidence from sinking particles, we turn to a simple model of particle interactions in the water column. Using the new suspended particulate Pb and 210pb data and sediment trap Pb data from a time series study in 1980- 1982 to constrain the model, we evaluate the relative influence of

surface inputs versus deep suspended particles on the time evolution of sinking Pb in the Sargasso Sea. We adopt a two-box model of particle interactions which arbitrarily divides the particle size continuum into two size classes: (1) small particles with negligible sinking rams which dominate exchange with the dissolved pool by virtue of their abundance and surface area and (2) large rapidly sinking aggregates, originating in surface water, which remove and regenerate the small-particle population through aggregation/disaggregation processes occurring at all depths but which have no direct chemical interaction with the dissolved pool [Bacon et al., 1985; Nozaki et al., 1987].

SttERRELL ET AL.: ISOTOPIC EQUILIBRATION OF LEAD IN SARGASSO SEA 11,265

Knowing the Pb content of small suspended particles in the deep water column (Pbp) and estimating suspended particle residence time (xp), it is possible to calculate the contribution of small particle removal (the "repackaging flux" FR) to the total Pb flux (FT) measured in deep sediment traps.

This model is described in full by Sherrell and Boyle [ 1992], where it is applied in a steady state form to evaluate the importance of deep water column scavenging processes for the removal of a suite of trace elements near Bermuda. Briefly, the model assumes a two-box one-dimensional water column where

the total deep flux is the sum of a primary flux originating in the surface box (Fs = 0-500 m) and the repackaging flux resulting from aggregation and sinking of suspended particles in the deep box (500-3200 m):

FT-- Fs + F g (1)

The repackaging flux is calculated from the suspended particulate Pb distribution and an estimate of the mean deep water column suspended particle residence time:

(2)

where Pbp equals the mean suspended concentration in the deep box, D is the depth of the deep box (2700 m), and xp = 6.5 + 1.5 years, as estimated independently from the distribution of particulate 230Th near Bermuda [Sherrell and Boyle, 1992]. Implicit is the assumption that Th and Pb are scavenged by the same population of particles.

The total Pb flux (FT) is calculated from the mean flux of eleven 2-month sediment trap deployments at 3200 m, taken within a few tens of kilometers of the pump stations over August 1980 to August 1982 [Jickells et al., 1984]. Because this was a period of anomalously high mass flux compared with the long- term (1978-1991) mean, we divide the mean measured Pb flux by 1.3 [Deuser et al., 1988; W. G. Deuser, personal communication, 1991], assuming that Pb flux is a linear function of mass flux on a

time scale of years to decades and that the 13-year mean is more representative of flux magnitude during the residence time of deep suspended particulate Pb measured in 1987-88.

The model results for Pb and for 210pb are presented in Table 3. The Pb calculation is keyed to 1981, the only year for which F T is measured. Assuming F R ls proportional to the deep box total Pb inventory, we estimate that F R for 1981 is ~10% greater than that calculated using the 1987-1988 suspended particle measurements, based on the difference between 1979 and 1984

dissolved Pb profiles [Boyle et al., 1986]. The surface-derived flux, FS, is calculated as a simple difference from equation (1).

For the year 1981, the model calculation indicates that small- particle "repackaging" contributed about 30% of the total deep Pb flux; the major flux _fraction is __therefore derived from the surface box. The steady state calculation for 210pb, using a mean deep box particulate concentration of 0.3 dpm/100 kg (this study and M.P. Bacon, unpublished data) and the mean 3200-m flux for 33 trap deployments during 1980-1986 (M.P. Bacon, unpublished data), demonstrates that in situ scavenging in the deep box contributed the same fraction of the total 210pb flux. The agreement in FR/F T for Pb and 210pb is a coincidental result of the progression of surface and deep Pb inventories, as explained in the next section. This calculation suggests that most of the Pb and 210pb in rapidly sinking particles must have originated in near- surface waters, and the isotopic composition of Pb reaching the sediments should not be greatly shifted from the surface value; certainly it will not equal the deep suspended particulate value. A simple mixing calculation using estimated mean 0- to 500-m and 500- to 2700-m isotopic composition yields pb/210pb of -900 pmol/dpm and 206pb/207pb of-1.203 for material sinking from deep water in 1981.

Measurements of Pb isotope ratios in sediment trap material should provide a test of the model results. However, recent attempts have yielded equivocal results because of suspected sample contamination and large compositional variations on short time scales. Using small traps in short-term deployments, Shen and Boyle [1988] found surface-type 206pb/207pb ratios in

TABLE. 3. Two-Box Flux Model Results for Pb and Pb-210 near Bermuda ^

Element Year

"Repackaging Surface Total Flux"* Flux Flux

FR F$ FT FR/FT FS/FR

1981 0.96 2.2 3.2II 0.30 2.3

1972 1.06 4.2 5.3 0.20 4.0

1979 0.98 2.8 3.8 0.26 2.9

1984 0.92 1.4 2.3 0.40 1.5

1988 0.87 0.7 1.6 0.55 0.8

1250 2930 4180 0.30 2.3

^Pb units in gmol/m2/yr, Pb-210 units in dpm/•/•. Italicized values are determined from measured quantities; others are estimates based on evolution of surface Pb concentrations and deepwater Pb inventories. See text for explanation.

*Calculated using mean deepwater suspended particulate Pb concentration, assuming suspended particle residence time is 6.5 years; 2700-m deep box.

#Mean flux from 2-year time series sediment trap deployment (August 1980 to August 1982) at 3200 m [Jickells et al., 19841.

@Lead-210 data from this study and M.P. Bacon (unpublished dam). FT(Pb-210) is mean of 1980-1986 deployments at 3200 m.

11,266 SItERREI J • ET AL.: ISOTOPIC EQUILIBRATION OF LEAD IN SARGASSO SEA

sinking particles at 860 m in August 1984, then found ratios more closely matching the dissolved profile in a subsequent deployment 6 months later. Their attempts to measure the isotopic composition of selected trap samples from Deuser's PARFLUX deployments gave ratios indicative of contamination. Similarly, pb/210pb in particles trapped at 3200 m (1980-1982 time series) varied over a range exceeding that exhibited by the 1981 dissolved pb/210pb profile, with no apparent dependence on mass flux (mean of 970 q- 560 [ lcq pmol/dpm [Jickells et al., 1984; Bacon et al., 1985; M.P. Bacon, unpublished data]). These results imply that the degree of interaction of the large-particle flux with the standing stock of suspended particles may be highly variable on seasonal or shorter time scales. Some of the variability may be procedural (e.g., Deuser's samples were rinsed and sieved before analysis, and Pb and 210pb were measured on separate splits), even if the mean values are accurate. A sediment-trapping study dedicated to this problem might provide more precise data, but these initial measurements suggest that frequent observations over a long time series will be necessary to test the model predictions.

Recently obtained 210pb flux data from Bermuda may add new constraints on this picture, requiring reconsideration of one of the model assumptions. Time series sediment trap measurements in midwater and deepwater indicate that 210pb flux at 1500 m averaged 63 q- 18% (lc0 of 3200-m flux over the period May 1984 to July 1986 (n = 9, two anomalously high ratios rejected), even though average mass flux was equal at the two depths (W. G. Deuser and M.P. Bacon, unpublished data). Assuming this flux difference represents a long-term mean flux gradient of constant slope in the deep water column, extrapolation of the 1500- to 3200-m gradient to the model's 2700-m deep box implies that the flux increase due to deep ocean scavenging processes accounts for -60% of the total 210pb flux at 3200 m. This is considerably more than the -1/3 fraction predicted by the model. If true, it violams the model assumption that large particles do not exchange Pb directly with the dissolved pool. Instead, large particles may scavenge about as much Pb from the dissolved pool as they scavenge by mediating the removal of suspended particles. The existence of a dissolved-large particle scavenging pathway was also proposed by Moore and Dymond [1988] and set forth as a possible resolution of unexpected discrepancies in a recent Th modeling effort [Murnane and Sarmiento, 1990].

In any case, because Pb on suspended material is in equilibrium with dissolved Pb, the pathway by which deep Pb is incorporated into sinking material does not affect the isotopic composition of the total deep Pb flux or the model predictions of surface-derived versus deep-derived Pb (FR/FT). Nevertheless, although the 210pb evidence must be considered preliminary because it depends on assumptions of similar surface source regions for particles caught at different depths (perhaps an unlikely scenario [Deuser et al., 1991' W. G. Deuser, personal communication, 1991]), it demonstrates that depth variations in chemical and mass fluxes need to be better understood to

formulate particle interaction and scavenging models which best describe processes occurring in the open-ocean water column.

43. Temporal Variation of Pb and Pb Isotope Fluxes

The degree to which the Pb isotopic composition of rapidly sinking particles is modified from the surface value depends on the magnitude of the surface-derived flux and the rate of small- particle removal in the deep water column. Because of the continuing reduction of anthropogenic Pb input to the northwest

Atlantic since the trap measurements were made, surface-derived Pb may have ceased to dominate the deep ocean Pb flux by 1988. The evolution of the model-derived sources of the deep Pb flux near Bermuda over the period 1972-1988 is outlined in Table 3.

The temporal trend in F S is estimated assuming that the yearly average surface-derived Pb flux is dependent on the dissolved Pb concentration in the mixed layer (normalized to 210pb to eliminate short-term source strength variability). Thus we expect that F S decreased by a factor of 1.5 from 1972 to 1979 (estimated from the coral record of Shen and Boyle, [ 1987]), by a factor of 2 from 1979 to 1984 [Boyle et al., 1986; Shen and Boyle, 1987], and by another factor of 2 by 1988 (this study; E. A. Boyle, unpublished data). Fixing F S for 1981 as determined above and assuming a linear decrease in the period 1979-1984, F S was estimated for the years 1972, 1979, 1984 and 1988. The repackaging flux, F R, has probably been much less variable. Assuming as above that FR is proportional to the total subsurface Pb inventory, values of FR for the four years listed above were estimated by allowing the 1988 suspended particle Pb inventory to reflect a linear decrease of-20% since 1972 [Boyle et al., 1986] (Table 3).

Thus the ratio of surface-derived to deep-derived Pb flux (FS/FR) is estimated to have diminished from -4 near the peak of the U.S. alkyl lead consumption in 1972 to <1.0 in 1988 (Table 3). During the buildup of alkyl Pb combustion from 1930 to 1970, the deepwater Pb inventory increased by about a factor of 3 while surface concentrations increased about fivefold over a 1930 value

which exceeded the 1988 surface levels [Shen and Boyle, 1988]. Therefore FS/FR was always >2 for this period, and the isotopic composition of deep sinking material has been dominated by contemporary surface input ratios for at least the last 50 years. Assuming amaospheric input of continental dust was the dominant source of Pb to central gyre open-ocean locations in the preanthropogenic era, this conclusion probably applies as well for Pb transport in the unperturbed condition. This assertion is supported by the model results for 210pb ' The steady state calculation gives Fs/F R = 2.3 for 210pb and probably provides a reasonable lower limit for preanthropogenic stable Pb.

If surface concentrations continue to decrease as leaded

gasoline use ceases in North America, the model predicts that Pb derived from surface waters may become a minor portion of the total deep ocean flux until the deepwater Pb inventory has diminished to a new steady state. During this period lasting many decades, the lead isotopic composition of deep sinking particles would be expected to resemble the deepwater values seen in the 1984 dissolved and 1988 particulate profiles. However, reconstruction of dissolved Pb concentrations from the Bermuda

coral record suggests that mixed layer Pb concentrations are now lower than the prealkyl levels of the 1920-1930s [Shen and Boyle, 1987]. Thus the future trend of Pb inputs to the Sargasso Sea may be determined largely by nongasoline North American industrial sources and European Pb inputs and may not decrease much below current values.

Considering this result along with our conclusions drawn from application of the same model to steady state distributions of several other trace elements [Sherrell and Boyle, 1992], we conclude generally that, apart from radionuclides with an in situ source and Pb in its current non-steady state condition, the trace element composition of suspended particles in the deep ocean may have only minor influence on the composition of material sinking to the sediments, with a greater role being assigned to surface ocean exchanges. Processes generating metal removal fluxes in

SHERRELL ET AL.' ISOTOPIC EQUILIBRATION OF LEAD IN SARGASSO SEA 11,267

the open oligotrophic ocean generally seem only partially coupled to deep ocean suspended particles. Therefore the observation of dissolved/particulate exchange equilibrium for Pb does not imply that whole ocean residence times for Pb or any trace element we have studied could be predicted solely from surface complexation models of deep ocean scavenging processes [e.g., Balistrieri et al., 1981; Erel and Morgan, 1991]; a model for upper water column fluxes would also be required.

5. CONCLUSIONS

1. Using in situ pump sampling at a station near Bermuda, we have measured the first open-ocean profile of suspended particulate Pb. Concentrations increase from-1 pmol/L in surface waters (0-i00 m) to a 2.9 pmoiœL maximum near 500 m, 'then decrease to 1.5-2.0 pmol/L in deep waters. Previous measurements using shipboard filtration from sampling bottles gave values that were >10 times higher, an apparent contamination artifact. The new data make possible a quantitative evaluation of the role of fine suspended particles in the transport of Pb through the oceanic water column.

2. New measurements of stable Pb isotopes dissolved in the upper water column near Bermuda indicate that surface water 206pb/207pb decreased from -1.203 to -1.192 over the period 1983 to 1987. This isotopic evolution has accompanied a drop in dissolved Pb concentration over this period as atmospheric input of unusually radiogenic gasoline-derived alkyl-Pb has continued to decrease. Over the same period, Pb/210Pb in surface water has decreased from 700 pmol/dpm to about 400 pmol/dpm. Water column profiles for 206pb/207pb and pb/210pb demonstrate that the surface signal has propagated into the upper thermocline, producing measurable differences from the 1984 profile to -500 m. This observation is consistent with the thermocline

ventilation/chemical transport model of Boyle et al. [1986] and Sheri and Boyle [1988].

3. Suspended particulate 206pb/207pb and pb/210pb profiles demonstrate that the isotopic composition of small particles matches that of ambient seawater throughout the upper 2000 m. The concordant evidence from two tracer ratios suggests that dissolved and particulate Pb equilibrate on a time scale which is short in relation to the particle residence time. This result implies that the Pb content of suspended particles is controlled by relatively rapid reversible exchange with dissolved Pb, consistent with equilibrium exchange models for the transport of Th.

4. A simple two-box flux model for the oceanic water column near Bermuda suggests that vertical removal of suspended particles below 500 m contributed about 30% of the total flux of Pb at 3200 m in 1981. We conclude that Pb flux out of the

Sargasso Sea water column may be dominated by processes in the surface water and upper thermocline, with a lesser influence of suspended particle "repackaging" in the deep water column. Therefore the whole ocean residence time for Pb may be largely independent of dissolved/suspended particulate exchange processes in the deep ocean.

5. Extension of the box model to temporal variations in surface and deepwater distribution of Pb suggests that removal of deep suspended Pb contributed only 20% of total deep Pb flux at the peak of U.S. Pb input in 1972, but this fraction increased to >50% in 1988 and may dominate the surface-derived flux if Pb inputs continue to diminish. Application of the flux model to the steady state distribution of 210pb indicates that deep scavenging of Pb on suspended particles contributed <30% of deep ocean Pb flux in its unperturbed preanthropogenic distribution.

Acknowledgments. The authors thank the officers and crew of the R/V Weatherbird, the R/V Endeavor, and the R/V Oceanus for superb support at sea. Gary Klinkhammer, John Trefry, Erik Brown, Scott Doney, Paula Rosenet, and Vernon Ross all gave valuable shipboard assistance. This manuscript benefited from discussions with Mike Bacon, Werner Deuser, Jim Bishop, Francois Morel, and Ed Sholkovitz. Mike Bacon is especially thanked for use of his laboratory, and we thank Becky Belastock and Alan Fleer for assistance with the 210pb analyses. We are grateful to Alan Zindler for giving access to the Lamont mass spectrometer and clean lab facilities, as well as for providing the 205pb spike required for this study. Gordon Wallace and Chris Krahforst were very generous in assisting with solvent extraction and analysis of dissolved Pb. This work was supported by NSF ornnt (Me•.R'/1 f}q9R nncl ONII• ornnt NlOlkqlA_Rfi_lT_f}q9• tn •. A •

and NASA grant NAGW-895 to B.H.

REFERENCES

Anderson, R. F., M.P. Bacon, an d P. G. Brewer, Removal of Th-230 and Pa-231 from the open ocean, Earth Planet. Sci. Lett., 62, 7-23, 1983.

Bacon, M.P., and R. F. Anderson, Distribution of thorium isotopes between dissolved and particulate forms in the deep sea, J. Geophys. Res., 87, 2045-2056, 1982.

Bacon, M.P., D. W. Spencer, and P. G. Brewer, Pb-210/Ra-226 and Po- 210/Pb-210 disequilibria in seawater and suspended particulate mauer, Earth Planet. Sci. Lett., 32,277-296, 1976.

Bacon, M.P., C.-A. Huh, A. P. Fleer, and W. G. Deuser, Seasonality in the flux of natural radionuclides and plutonium in the deep Sargasso Sea, Deep Sea Res. 32,273-286, 1985.

Balistrieri, L., P. G. Brewer, and J. W. Murray, Scavenging residence times of trace metals and surface chemistry of sinking particles in the deep ocean, Deep- Sea Res., 28A, 101-121, 1981.

Boyle, E. A., S. D. Chapnick, and G. T. Shen, Temporal variability of lead in the western North Atlantic, J. Geophys. Res., 91, 8573-8593, 1986.

Buat-Menard, P., and R. Chesselet, Variable influence of the atmospheric flux on the trace metal chemistry of oceanic suspended matter, Earth Planet. Sci. Lett., 42, 399411, 1979.

Craig, H., S. Krishnaswami, and B. L. K. Somayajulu, Pb-210 - Ra-226: Radioactive disequilibrium in the deep sea, Earth Planet. Sci. Lett., 17, 295-305, 1973.

Deuser, W. G., F. E. Muller-Karger, and C. Hemleben, Temporal variations of particle fluxes in the deep subtropical and tropical North Atlantic: Eulerian versus Lagrangian effects, J. Geophys. Res., 93, 685745862, 1988.

Deuser, W. G., D. A. Siegel, and L. Fricke, Sediment trapping in the Sargasso Sea: where do the particles come from? (abstract), Eos Trans. AGU, 72 (17), Suppl., 61, 1991.

Erel, Y., and J. J. Morgan, The effect of surface reactions on the relative abundances of trace metals in deep-ocean water, Geochim. Cosmochim. Acta, 55, 1807-1813, 1991.

Fleer, A. P., and M.P. Bacon, Determination of Pb-210 and Po-210 in seawater and marine particulate mauer, Nucl. Instrum. Methods Phys. Res., 223, 243-249, 1984.

Fowler, S. W., and G. A. Knauer, Role of large particles in the transport of elements and organic compounds through the oceanic water column, Prog. Oceanogr., 16, 147-194, 1986.

Hamelin, B., R. M. Sherrell, and E. A. Boyle, Pb isotopes in dissolved and particulate phases on depth profiles at Station "S" (Bermuda): Rapid change from 1983 to 1987 (abstract), Eos Trans. AGU, 69, 1255-1256, 1988.

Hamelin, B., F. E. Grousset, P. E. Biscaye, A. Zindler, and J. M. Prospero, Lead isotopes in trade wind aerosols at Barbados- The influence of European emissions over the North Atlantic, J. Geophys. Res., 94, 16,243-16,250, 1989.

Hamelin, B., F. E. Grousset, and E. R. Sholkovitz, Pb isotopes in surficial pelagic sediments from the North Atlantic, Geochim. Cosmochim. Acta, 54, 37-47, 1990.

11,268 SHERRELL ET AL.: ISOTOPIC EQUILIBRATION OF LEAD IN SARGASSO SEA

Jickells, T. D., W. G. Deuser, and A. H. Knap, The sedimentation rates of trace elements in the Sargasso Sea measured by sediment trap, Deep Sea Res., 31, 1169-1178, 1984.

Lambert, C. E., E. Nicolas, A. Veron, P. Buat-Menard, G. Klinkhammer, P. Le Corre, and P. Morin, Anthropogenic lead cycle in the northeastern Atlantic, Oceanol. Acta, 14, 59-66, 1991a.

Lambert, C. E., A. Veron, P. Buat-Menard, M. Heyraud, F. Grousset, and W. Simpson, The role of large biogenic particles in the transport of atmospheric Pb down to North Atlantic sediments, Oceanol. Acta, 14, 67-76, 1991b.

Maring, H. B., and R. A. Duce, The impact of atmospheric aerosols on trace metal chemistry in open ocean surface seawater, 3, Lead, J. Geophys. Res., 95, 5341-5347, 1990.

Moore, W. S., and J. Dymond, Correlation of Pb-210 removal with organic carbon fluxes in the Pacific Ocean, Nature, 331, 339-341, 1988.

Moore, W. S., and W. M. Sackett, Uranium and thorium series

inequilibrium in seawater, J. Geophys. Res., 69, 5401-5405, 1964. Mumane, R. J., and J. L. Sarmiento, Thorium isotopes, particle cycling

models, and inverse calculations of model rate constants, J. Geophys. Res., 95, 16,195-16,206, 1990.

Nozaki, Y., J. Thomson, and K. K. Turekian, The distribution of Pb-210 and Po-210 in the surface waters of the Pacific Ocean, Earth Planet. Sci. Lett.,32, 304-312, 1976.

Nozaki, Y., K. K. Turekian, and K. von Damm, Pb-210 in Geosecs profiles from the North Pacific, Earth Planet. Sci. Lett. 49, 393-400, 1980.

Nozaki, Y., H.-S. Yang, and M. Yamada, Scavenging of thorium in the ocean, J. Geophys. Res. 92,772-778, 1987.

Rama, M. Koide, and E. D. Goldberg, Pb-210 in natural waters, Science 134, 98, 1961.

Schaule, B. K., and C. C. Patterson, Lead concentrations in the northeast

Pacific: Evidence for global anthropogenic perturbations, Earth Planet. Sci. Lett., 54, 97-116, 1981.

Schaule, D. M., and C. C. Patterson, Perturbations of the natural Pb depth profile in the Sargasso Sea by industrial Pb, in Trace Elements in Seawater, edited by C. S. Wong, E. A. Boyle, K. Binland, D. Burton, and E. D. Goldberg, pp. 487-504, Plenum, New York, 1983.

Settle, D. M., C. C. Patterson, K. K. Turekian, and J. K. Cochran, Lead

precipitation fluxes at tropical oceanic sites determined from Pb-210 measurements, J. Geophys. Res., 87, 1239-1245, 1982.

Shen, G. T., and E. A. Boyle, Lead in corals: Reconstruction of historical industrial fluxesto the surface ocean, Earth Planet. Sci. Lett., 82,289- 304, 1987.

Shen, G. T., and E. A. Boyle, Thermocline ventilation of anthropogenic lead in the western North Atlantic, J. Geophys. Res., 93, 15,715- 15,732, 1988.

S[•errell, R. M., Collection of oceanic SPM for trace metal analysis using a new in situ pump, in Marine Particles: Analysis and Characterization, Geophys. Monogr. Ser., vol. 63, ed. D.C. Hurd and D. W. Spencer, pp. 285-294, AGU, Washington, D.C., 1991.

Sherrell, R. M., and E. A. Boyle, The trace metal composition of suspended particles in the oceanic water column near Bermuda, Earth Planet. Sci. Lett., 111,155-174, 1992.

Talbot, R. W. and A. W. Andren, Relationships between Pb and 210pb in aerosol and predpitafion at a semiremote site in northern Wisconsin, J. Geophys. Res., 88, 6752-6760, 1983.

Taylor, S. R., Abundance of chemical elements in the continental crest: A new table, Geochim. Cosmochim. Acta, 28, 1273-1285, 1964.

Veron, A., C. E. Lambert, A. Isley, P. Liner, and F. Grousset, Evidence of recent lead pollution in deep north-east Atlantic sediments, Nature, 326, 278-281, 1987.

Wallace, G. T., G. L. Hoffman, and R. A. Dace, The influence of organic matter and atmospheric deposition on the particulate trace metal concentrations of northwest Atlantic surface seawater, Mar. Chem., 5, 143-170, 1977.

Wallace, G. T., N. Dudek, R. Dulmage, and O. Mahoney, Trace element distributions in the Gulf Stream adjacent to the southeastern Atlantic continental shelf--influence of atmospheric and sheff water inputs, Can. J. Fish. Aquat. Sci., 40, Suppl. 2, 183-191, 1983.

R. M. Shenell and E. A. Boyle, Department of Earth, Atmospheric and Planetary $ciencea, E34-200, Massachusetts Institute of Technology, Cambridge, MA 02139.

B. Hamelin, G6oaciences de l'Environnement, Universit6 d'Aix- Marseille, F-13397 Marseille, France.

(Received September 5, 1991; revised February 24, 1992;

accepted February 25, 1992.)