on the distribution of copper, nickel, and cadmium in the surface...

JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 86, NO. C9, PAGES 8048-8066, SEPTEMBER 20, 1981

On the Distribution of Copper, Nickel, and Cadmium in the Surface Waters of the North Atlantic

and North Pacific Ocean ß

EDWARD A. BOYLE, SARAH $. HUESTED AND SUSAN P. JONES

Department of Earth and Planetary Sciences, Massachusetts Institute of Technology, Cambridge, Massachusetts 02139

Concentrations of copper, nickel, and cadmium have been determined for about 250 surface water samples. Nonupwelling open-ocean concentrations of these metals are Cu, 0.5-1.4 nmol/kg; Ni, 1-2 nmol/kg; and Cd, less than 10 pmol/kg. In the equatorial Pacific upwelling zone, concentrations of Ni (3 nmol/kg) and Cd (80 pmol/kg) are higher than in the open ocean, but Cu (0.9 nmol/kg) is not significantly enriched. Metal concentrations are higher in cool, nutrient-rich eastern boundary currents: Cu, 1.5 nmol/kg; Ni, 3.5 nmol/kg, and Cd, 30--50 pmol/kg. Copper is distinctly higher in the coastal waters of the Gulf of Panama (3-4 nmol/kg) and also higher in the shelf waters north of the Gulf Stream (2.5 nmol/kg); these copper enrichments may be caused by copper remobilized from mildly reducing sheff sediments and maintained by a coastal nutrient trap. In the open ocean, events of high-Cu water (1.5-3.5 nmol/kg) are seen on scales up to 60 km; presumably, these are due to the advection of coastal water into the oceanic interior. The lowest copper concentrations in the North Pacific central gyre (0.5 nmol/kg; (Bruland, 1980)) are lower than in the Sargasso Sea (1.3 nmol/kg), while for nickel the lowest concentrations are 2 nmol/kg in both the North Pacific and the North Atlantic. Nickel and cadmium, while generally correlated with the nutrients in surface waters, show distinct regional changes in their element-nutrient correlations. The residual concentrations of trace metals in the surface waters of the ocean can be

explained if biological discrimination against trace metals relative to phosphorus increases as productivity decreases.

INTRODUCTION anti-fouling coating of the ship. However, the surface maxi-

Recent work on copper, nickel, and cadmium based on mum observed at GEOSECS station 340 [Boyle et al., 1977] about a dozen vertical profiles in the ocean elucidated the ma- (Figure la) consisted of six data points, four of which were be- jot processes controlling their distributions [Boyle and Ed- low the base of the mixed layer. It is difficult to contaminate

the upper thermocline by a research' vessel because the verti- mond, 1975b; $clater et al., 1976; Boyle et al., 1976; Martin et al., 1976; Bender and Gagher, 1976; Boyle et al., 1977; Bruland cal diffusivity is low [Jenkins, 1980]. It is also worth noting et al., 1978; Moore, 1978; Bruland, 1980]. From a comparison that these samples were taken with a rosette sampler with bot- of the profile data of these elements with those of the nutri- ties tripped as the rosette was returning to the surface, so the ents, it is clear that all of these elements are involved in the surface samples are no more prone to sampler contamination

than the deeper samples. Boyle et al. [1977] considered that biogeochemical cycle of removal from surface and near-surface waters by organisms and subsequent regeneration into the surface maximum might be due to a continental shelf

source analogous to "-SRa [Kaufman et al., 1973], but the solution at depth from falling biological debris. The regeneration depth differs between elements: cadmium, and some of radium data were not available until later, when Knauss et al. the nickel being regenerated at shallow depths, copper and the [1978] showed that there indeed was high '-'-aRa near the sta- remaining nickel being regenerated at or near the bottom. tion where the surface copper maximum was observed (Figure Most of the variation in the deep distributions of these ele- lb). Therefore, it is at least plausible that continental shelves ments can be explained by the interaction of these processes may serve as a source of dissolved copper that can be ad- with the general circulation of the ocean. vected seaward and create surface copper maxima.

This investigation was undertaken to clarify these questions These investigations have raised two problems concerning the distribution of trace elements in near-surface ocean wa- concerning the mechanisms controlling the distribution of Cu, ters: (1) In nutrient-depleted waters, cadmium and zinc [Bru- Ni, and Cd in the surface waters of the ocean by providing a land, 1980] are depleted to very low levels (less than 1/100 of more extensive and precise data base on which to base hy- their deep water concentrations) but there are substantial re- potheses and models. sidual quantities of copper, nickel, barium [Chan et al., 1977], Se [Measures et al., 1980], and chromium [Cranston and Mur- ray, 1978]. (2) Boyle et. al. [1977] reported evidence for a surface copper maximum and suggested an aeolean source similar to that for '-•øPb [Nozaki et al., 1976], but Bruland [1980]

SAMPLE COLLECTION AND ANALYSIS

The elements studied are ubiquitous on research vessels and in oceanographic laboratories (e.g., copper antifouling paint, brass fittings, nickel-plate, and alloys; cadmium/copper hi-

presented data from the North Pacific ruling out such a 2løpb: trate columns, etc. ad infinitum) so contamination is a serious Cu correlation. Bruland suggested that surface extrema may problem requiring suitable precautions. To avoid possible arise from contamination in the vicinity of a research vessel. contamination of waters by the hull of the ship, samples in In two of the three GEOSECS stations, the surface maxima this study were taken while the ship was steaming at 2 kn, us- consisted of only two data points within the mixed layer and ing a 15-20 foot pole extended over the side of the ship. At the plausibly could be attributed to local contamination by the end of the pole a plexiglas holder was fitted to which the bot-

tles were attached by tygon straps. The bottles were immersed Copyright ¸ 1981 by the American Geophysical Union. upside down (to minimize surface slicks), turned over, and al-

Paper number IC0755. 0148-0227/81/001 C-0755501.00

8048

BOYLE ET AL.: Cu, Ni, AND Cd IN N. HEMISPHERE SURFACE WATERS 8049

Copper. nmol/kg

200

0 2

6O0 ß ' ' OEOSECS Stn.

Fig. la. Shallow profile for copper at GEOSECS Station 340 [Boyle et al., 1977] (10ø28'N; 123ø38'W); two-sigma error bars are in- dicated and replicates are connected by a line.

when available, in a laminar flow clean bench (DeSteiguer 1979, Gillis 1979).

Thirty-five gram portions of the samples were analyzed by a modification of the cobalt-APDC coprecipitation method de- scribed by Boyle and Edmond [1975a, 1977]. The precipitation procedure was carried out in $0 ml teflon centrifuge tubes. Af- ter being allowed to stand overnight, the precipitate was centrifuged, the seawater siphoned off, the precipitate rinsed with 10 ml distilled water, and then centrifuged and siphoned again. The precipitate is then digested with 0.$ ml 6N HNO3, evaporated to dryness, and then redissolved in 1 ml 0. IN HNO3. These concentrates are then analyzed by flameless atomic absorption spectrometry, using a Perkin-Elmer model $000 spectrophotometer, model $00 heated graphite atomizer, and AS-1 autosampler. Argon was used as the tube purge gas. Maximum power heating was utilized for Cu and Ni, and a 1- s ramp atomization was employed for Cd. Regular graphite tubes were used for Cd, while pyro-coated tubes were used for Cu and Ni. To compensate for small losses of precipitate during siphoning, cobalt was determined on diluted portions of the solution by flame atomic absorption spectrometry, using a

lowed to fill. They were then emptied and rerinsed by the Perkin-Elmer model 403 spectrophotometer (using a narrower same procedure prior to being filled and immediately capped. slit setting than recommended to improve linearity). The trace Clean polyethylene gloves were always worn when handling metal values then were normalized for cobalt recovery. Blanks the sample bottles, and the bottles were kept covered when were run with each batch and generally were below 0.2 nmol/ not being handled. The Thomas Washington 1976 samples kg for Cu, 0.05 for Ni, and 0.005 for Cd. The Cd blank on the were taken by lowering a 1-1 bottle from the bow of the ship Pierce 1979, Oceanus 1978, and Gillis 1979 data sets were as it stopped for station work. The DeSteiguer 1979 profile was higher, thereby decreasing Cd precision and increasing detec- obtained from rosette-mounted 'Go-Flo' bottles. tion limits. Blank corrections were applied in the calculation

Two independent unfiltered samples were collected at each of the final data. site. Occasionally, additional samples were taken for filtration through 0.4/z Nucleopore filters, using a vacuum filtration ap- partus. The filter holder was rinsed with distilled water and a portion of the sample; the tiltrate was then collected directly in a clean sample bottle placed within the vacuum chamber. Table la compares several filtered versus unfiltered samples; it is clear that filtration is not significant for Cu and Ni but may be significant for Cd. The results reported in Tables 2-9 are for unfiltered samples. Within a few days of collection samples were acidified with 1 ml of double vycor-distilled 6N HCI, as were distilled water blanks as well. The distilled water blanks were of different known volumes so that blank contri- butions from the distilled water and the acid could be re-

It was observed that the precision for replicate samples within a single batch was superior to precision for replicates between batches. This difference may arise from errors in the preparation of standards or from inexact matrix matching. To maintain precision within large data sets, two control samples with different concentrations were run within each batch. The

data were then adjusted to be consistent with the mean values determined for these control samples. The standard additions precipitation recovery efficiencies were calculated after correc- tion to the mean quality controls: Cu, 87 + 5%; Ni, 91 + 4%; Cd, 80 +_ 7%. On the basis of our independent determinations of the chemical recovery [Boyle and Edmond, 1975a] these apparent recoveries may be UP to 10% low from a systematic un-

solved. This process was carried out in a sheltered area, or derestimate of the mean for the control samples. These uncer-

228Ra Copper

o S 20 0 o N 20 Lotitude

Fig. lb. Surface 228Ra data from GEOSECS Pacific section along 123øW (units dpm/100 kg) [Knauss et al., 1978]; surface copper concentrations reported from GEOSECS stations 340 and 345 are also plotted (units: nmol/kg).

8050 BOYLE ET AL.: Cu, Ni, AND Cd IN N. HEMISPHERE SURFACE WATERS

TABLE la. Comparison of Filtered and Unfiltered Samples

Copper Nickel Cadmium

Unfiltered Filtered Unfiltered Filtered Unfiltered Filtered

Number 1 2 3 4 1 2 3 4 1 2 3 4 Sample

I 1.33 1.25

34 0.97, 1.05 0.92, 1.43 0.90, 0.89 1.02, 1.14 92 0.74, 0.82 0.62, 0.69 0.63, 0.62 0.70, 0.82

Oceanus 1978

DeSteiguer 19 79

0.075 0.024

0.024, 0.028 0.013, 0.018 0.010, 0.005 0.016, 0.013

1 12.137 3.80 3.70 3.917 2.56 2.81 0.055 0.044 0.066 2 3.62 3.95 3.52 3.52 2.63 2.65 2.99 2.68 0.039 0.049 0.029 6 2.07 2.14 1.69 2.86 2.48 2.30 2.19 2.32 0.025 0.025 0.010

60 1.17 0.95 1.21 2.30 2.06 1.98 0.010 0.013 0.018

0.033 0.013

Two independent unfiltered samples (1 and 2) were collected at the same time as a larger sample to be filtered. Two samples (3 and 4) were successively filtered through the same filter into separate bottles.

tainties do not affect the final data sets, and comparison of our data with other investigators (see below) confirms a lack of a systematic bias in excess of our stated precision. After adjust- ment in this way, the best one sigma precisions are estimated as Cu, 0.13 nmol/kg or 5%, whichever is greater; Ni, 0.18 or 5%; Cd, 0.01 or 10%. For the Oceanus 1978 and Gillis 1979 data, the Cd and Ni precisions are somewhat poorer; for Cd this is because of a higher reagent blank in the cobalt solution. The poorer Cd precision in general is 'probably caused by in- complete salt removal and its effect on the flameless AA sensi- tivity; a second distilled water wash seems worthwhile if more precise Cd data were desired. At most stations two samples were analyzed and replicates run if the two samples disagreed. High points contradicted by low duplicate samples were discarded as contaminated; out of over 400 total analyses for each element, less than 5% were discarded in this fashion. No low points were excluded. In the data tables, average values have been given, although in some of the figures the individual analyses have been plotted to serve as a guide to the precision.

Temperature measurements were made with a bucket ther- mometer, and salinity was determined by using various con- Station* ductivity salinometers. Silicate, phosphate, and nitrate were 9130 determined by using standard methods [Strickland and Par- 9131 9132

sons, 1968]. In some cases, the nutrients are based on ship- Sample, board data (Thomas Washington 1976, Gillis 1979)but other- 32 wise silicate and nitrate were determined in the shore lab from 33

the acidified trace element samples. 34 35

RESULTS 36 37

This data set (Tables 2-8, Figures 2-8) is comparable in 38 magnitude to the entire published literature for reliable oce- 39 anic data for these elements; of this literature only about 10% 40 are for surface samples. There is clearly a high level of inter-

nal consistency within the data sets as well as general agree- Stations ment with expectations based on the reliable literature. In 23 Table lb our data is compared directly with that of Moore 21 [1978] and Bruland [1980] for sample locations overlapping 20 within about 100 kin. Our copper data averages about 0.5 Stationô

5

nmol?kg lower than that of Moore [1978]. Comparison of our 9 data with Bruland [1980] is complicated by an obvious eddy 11 enriched in Cu, Ni, and Cd (our stations 11 and 17); excluding 17 those points our copper data is 0.3 nmol?kg higher than Bru- land's data and the nickel data agrees exactly. Since the surface Cd data at this area is near our detection limit, we cannot compare these data; however, using Bruland's nitrate--phos-

phate and cadmium--phosphate regression equations we cal- culate 1.09 nmol?kg Cd at the nutrient maximum in the De- Steiguer 1979 profile; we observe 1.11, 1.12, and 1.11. In general, we believe that our data agree with that of Bruland [1980] within 20 of our stated precision.

The Thomas Washington 1976 data deserves some comment on the apparent 'spikiness' in all parameters observed in the western North Pacific. The hydrographic data, obtained by the Scripps GOG group, is of high. quality; the noise is real and results from the pervasive eddy structure of the western North Pacific [Kenyon, 1978] (Figure 3b). Similar variability in the trace element data is likely to be a true representation of the western North Pacific.

Copper

In the copper data sets, several features stand out: (1) High surface values (2-4 nmol/kg) are observed in the Gulf of Pan-

TABLE lb. Comparison of This Data Set With Overlapping Data Sets From the Literature

Latitude, øN Longitude, øW Cu

21 o29' 24o29 ' 1.7 20 ø 18' 21 o 49' 1.4 20o50 ' 18o55 ' 1.3

22ø57 ' 17ø39 ' 0.77 21ø0' 18ø0 ' 0.88 20ø28 ' 18 ø 1' 0.95 20o28 ' 18 ø 1' 0.68 20o28 ' 18ø1 ' 0.70 20o28 ' 18ø1 ' 0.99 20o28 ' 18 ø 1' 0.84 20o28 ' 18ø1 ' 0.73 20o16 ' 17o60 ' 1.10

Latitude, øN Longitude, øW Cu Ni Cd

35ø32 ' 128ø32 ' 1.14 3.44 0.020 34ø54 ' 132ø39 ' 0.86 3.12 0.003 34 ø13' 136o39 ' 0.85 3.02 0.002

35ø0 ' 128ø0 ' 1.40 3.22 35ø0 ' 130o0 ' 1.10 3.18 35ø1 ' 133o57 ' 1.86 4.30 35o0 ' 138ø3.1 ' 1.69 3.15

* Moore [1978]. , This work (Oceanus 1978). $ Bruland [1980]. ô This work (Thomas Washington 1976).


TABLE 2. Data From the Thomas Washington, March 23 to April 30, 1976

Station Latitude, Longitude, Number øN øE T S O2 POn Si NO3 NO2 Cu Ni Cd

I 3503.8 ' -12203.4 ' 5 3500.3 ' -127059.6 ' 9 35 ø0.0' - 13000.0 '

11 35ø1.2 ' -133056.9 '

17 34059.8 ' -138ø3.1 ' 23 34059.9 ' -14400.7 ' 31 3500.3 ' -153056.5 ' 38 34059.7 ' -15900.8 ' 41 3500.7 ' -16201.0 ' 43 34ø59.1 ' -16401.0 ' 45 34055.8 ' -165059.7 ' 47 35ø2.1 ' -169059.9 '

51 34056.2 ' -172ø2.1 ' 53 3500.4 ' -175054.3 ' 55 34057.9 ' -177059.4 ' 59 34059.0 ' 179059.0 ' 65 34059.8 ' 174ø-1.4 ' 67 34059.6 ' 171058.7 ' 69 34058.4 ' 17004.3 ' 75 3503.9 ' 164013.5 '

77 34058.9 ' 16206.0 ' 79 3503.4 ' 16009.3 ' 81 34ø59.1 ' 157057.3 ' 87 35ø1.9 ' 151ø55.1 ' 89 34059.9 ' 149054.4 ' 90 34059.2 ' 148ø59.1 ' 91 34054.5 ' 14800.6 ' 92 35 ø 11.0' 14700.4 '

93 34ø58.1 ' 146ø1.1 ' 94 34ø56.1 ' 144058.4 ' 95 34058.4 ' 14401.0 ' 96 3500.4 ' 142059.9 ' 96 3500.4 ' 142059.9 ' 97 3502.8 ' 14201.2 '

12.11 33.418 6.25 0.33 13.19 33.201 6.17 0.33 14.17 33.341 6.03 0.28 15.23 33.400 5.92 0.25 15.93 33.996 5.79 0.08 16.66 34.121 5.74 0.06 17.29 34.333 5.60 0.03 15.59 34.334 5.83 14.75 34.201 6.02 0.12 15.47 34.336 5.96 0.06 14.62 34.328 6.14 0.08 15.28 34.391 6.00 0.09 14.69 34.376 6.10 0.23 15.02 34.438 5.98 0.17 15.39 34.485 5.85 0.14

15.06 34.435 5.94 0.18 15.40 34.607 5.78 0.18

7.2 2.65 0.13 2.31 4.16 0.169 4.9 0.15 0.03 1.40 3.22 0.045 3.2 0.00 0.00 1.10 3.18 0.028 3.4 0.03 0.00 1.86 4.30 0.189 4.3 0.08 0.00 1.69 3.15 0.090 5.7 0.00 0.00 0.90 2.77 0.026 4.8 0.00 0.00 0.99 2.91 0.033

0.90 3.10 0.041 6.7 0.00 0.00 0.97 3.22 0.063 7.5 0.01 0.00 1.32 2.45 0.050 6.2 0.00 0.00 0.90 2.85 0.038 6.8 0.00 0.00 1.13 2.75 0.063

8.3 1.24 0.00 1.03 2.88 0.155 4.0 1.11 1.51 2.71 0.088

1.54 2.91 0.099 7.4 0.77 1.10 3.04 0.075 6.1 1.99 1.75 2.67 0.153

15.60 34.630 5.69 0.41 10.4 5.47 0.31 1.06 4.25 0.129 34.582 4.99 0.14 5.7 1.33 0.23 0.87 7.73 0.071

15.00 34.658 5.83 0.39 8.2 2.88 1.66 2.95 0.130 13.80 34.501 6.12 0.40 12.0 4.58 0.33 1.17 3.34 0.174 15.80 34.702 5.68 0.27 8.0 1.51 0.29 1.31 3.46 0.150 16.10 34.720 5.58 0.11 5.6 1.99 0.38 1.25 2.58 0.094 15.66 34.715 5.77 0.21 5.6 2.75 0.12 1.15 2.91 0.136 15.96 34.616 5.96 0.39 10.4 2.75 0.39 1.15 2.87 0.125 15.95 34.646 5.95 1.94 2.80 0.126 18.87 34.732 5.58 0.16 4.1 0.00 0.00 1.10 2.69 0.085 18.68 34.712 5.57 1.41 3.13 0.153 15.25 34.511 6.09 0.13 11.8 4.52 0.14 1.26 3.23 0.196 19.34 34.724 5.44 1.29 2.53 0.074 17.61 34.733 5.86 0.00 0.6 0.00 0.00 1.46 3.69 0.200 17.66 34.742 5.78 1.28 2.62 0.110 17.66 34.742 5.78 1.20 2.45 0.085 18.30 34.767 5.55 0.06 5.4 1.68 0.55 0.94 2.48 0.058

Trace element samples were analyzed only once for this cruise, and a different collection method was used. Nutrient data from SIO data report (INDOPAC). Units used here and throughout are T, øC; S, %0; 02, ml/l; POn, Si, NO3, NO2,/•rnol/kg; Cu, Ni, Cd, nmol/kg.

ama on two cruises (DeSteiguer 1979 and Gillis 1979) (Figures base levels are somewhat lower (2.0 nmol/kg) but there is still 5 and 8). (2) Copper values exceeding 2 nmol/kg are also ob- a tendency to be slightly higher in cooler, nutrient-rich waters. served at 80-9 ø north in the Guatemala Basin (Figure 6). (3) The northern boundary of the Gulf Stream (shelf water) has Cadmium copper concentrations greater than 2.5 nmol/kg (Figure 7). (4) Cadmium covaries with the labile nutrients. The lowest val- Throughout the eastern boundary regions of the North Pacific ues observed are less than 10 pmol/kg (our detection limit), and the North Atlantic, 'events' of copper greater than 1.5 and the highest values (75-160 pmol/kg) are found in high- nmol/kg are observed occasionally. (5) In general, the con- nutrient waters. centration of copper in the central ocean is less than 1 nmol/

kg, but there is an increase to about 1.5 nmol/kg going into DISCUSSION the cool eastern boundary currents. (6) The Sargasso Sea has a higher copper concentration (• 1.3 nmol/kg) than the cen- Copper tral North Pacific (0.5 nraol/kg) [Bruland, 1980], although The high copper levels found in the Gulf of Panama lower copper concentrations are observed elsewhere in the unambiguously confirm the existence of a significant source of North Atlantic (Oceanus 1978 data). (7) There is no general copper for surface waters of the ocean. A plausible ex- covariance between copper and silicate or other nutrients in planation of this phenomenon is remobilization of copper the surface waters sampled here. from continental shelf sediments and subsequent diffusion

into the overlying water. Some alternative mechanisms that Nickel could produce the enrichment can be ruled out: (1) An aeo-

In the eastern Pacific, nickel shows a general correlation lean source similar to that for 21øpb is precluded by the lack of with the nutrients, being higher in the zone influenced by a central gyre copper maximum [Bruland, 1980]. (2) Up- coastal upwelling (Gulf of Panama and California current) welling of high-copper deep water as a source of high surface and the equatorial upwelling zone (within 2 ø of the equator). concentrations can be dismissed by two lines of evidence. AI- The lowest levels are about 2 nmol/kg (this work and Bruland though there are nutrient enrichments in the Gulf of Panama [1980]) increasing to 3-3.5 nmol/kg in the California current. due to upwelling processes, a comparable degree of nutrient There is a strong correlation between nickel and surface tem- enrichment in surface waters at the equator is accompanied perature in this region (Figure 9a). In the North Atlantic, the by no significant copper enhancement. Also, vertical profiles

TABLE 3. Data From the R/V Oceanus, October 29 to November 23, 1978

Sample Number Latitude, øN Longitude, øW T S Si Cu Ni Cd

1 38o28.8 ' 9o40.5 ' 0.47 1.33 2.20 0.075 2 37o28.2 ' 11ø20.4' 20.3 36.43 0.38 1.14 2.22 0.024 3 36o0.5 ' 13o13.4 ' 20.0 36.22 0.38 1.05 0.022 4 35045.3 ' 14ø6.5 ' 20.7 36.51 0.44 0.95 _<0.000 6 34040.5 ' 15056.0 ' 21.0 36.50 0.42 0.84 _<0.009 7 3300.0 ' 16ø0.0 ' 21.1 36.75 0.46 1.00 1.91 _<0.007 8 3209.7 ' 16027.0 ' 22.1 36.75 0.55 0.99 _<0.009 9 31016.0 ' 15057.5 ' 0.49 0.80 _<0.000

10 31o16.0 ' 15o57.5 ' 0.75 1.11 0.051 11 31016.0 ' 15057.5 ' 21.7 36.62 0.50 0.86 1.98 _<0.010 12 31016.0 ' 15057.5 ' 0.66 0.87 _<0.016 13 31016.0 ' 15057.5 ' 0.49 0.91 0.041 14 31o16.0 ' 15o57.5 ' 0.49 1.36 0.028 15 2900.0 ' 15ø0.0 ' 22.3 36.70 0.94 0.84 1.86 _<0.004 16 29030.0 ' 14048.6 ' 22.9 36.73 0.50 0.93 _<0.007 17 2900.0 ' 15ø0.0 ' 22.7 36.71 0.39 1.08 1.91 _<0.006 18 28034.6 ' 17013.0 ' 22.8 36.90 0.52 1.30 1.90 _<0.018 19 2700.0 ' 16ø0.0 ' 22.7 36.67 0.38 1.02 2.08 _<0.001 20 27024.3 ' 17038.0 ' 23.4 36.86 0.46 0.91 2.47 _<0.010 21 26033.0 ' 19013.0 ' 23.5 36.69 0.35 1.11 2.09 _<0.007 23 2500.0 ' 19ø0.0 ' 23.7 36.86 0.42 1.00 2.19 _<0.003 24 24017.2 ' 17053.7 ' 23.6 36.75 0.42 0.89 _<0.002 25 2300.0 ' 17ø0.0 ' 22.4 36.49 0.83 0.97 2.01 _<0.012 26 2304.3 ' 17ø35.2 ' 21.7 36.32 0.83 0.93 1.63 _<0.016 27 22057.0 ' 17039.0 ' 21.5 36.34 1.45 0.86 2.04 _<0.013 28 22057.0 ' 17039.0 ' 0.59 0.72 2.16 _<0.011 29 22o57.0 ' 17o39.0 ' 0.66 0.77 2.25 0.026 30 22057.0 ' 17039.0 ' 0.89 0.76 2.19 _<0.014 31 22057.0 ' 17039.0 ' 1.40 0.77 2.24 _<0.018 32 22057.0 ' 17039.0 ' 0.61 0.77 2.26 _<0.011 33 21ø0.0 ' 18ø0.0 ' 23.2 36.28 1.41 0.88 2.19 0.046 34 20028.4 ' 18ø0.9 ' 0.69 0.95 _<0.016 35 20028.4 ' 18ø0.9 ' 0.77 0.68 2.24 _<0.017 36 20028.4 ' 18ø0.9 ' 0.74 0.70 2.51 _<0.015 37 20o28.4 ' 18ø0.9 ' 0.64 0.99 0.028 38 20028.4 ' 18ø0.9 ' 0.61 0.84 _<0.014 39 20028.4 ' 18ø0.9 ' 1.50 0.73 1.91 _<0.003 40 20o16.2 ' 17o59.9 ' 0.77 1.10 0.026 41 18ø0.0 ' 18ø0.0 ' 25.5 36.06 0.82 0.89 2.25 _<0.008 42 17o41.5 ' 18o10.0 ' 28.2 34.08 2.13 0.82 0.028 46 13o30.6 ' 19ø0.4 ' 28.4 35.05 1.10 1.22 1.92 0.137 48 11 ø52.7' 20044.2 ' 28.8 35.09 0.64 0.69 1.64 _<0.012 50 10ø0.0 ' 22030.0 ' 28.3 35.37 0.77 0.75 1.92 _<0.013 52 8o59.0 ' 21o40.6 ' 28.4 35.61 0.69 1.26 0.030 54 7ø21.8 ' 2008.5 ' 29.7 34.72 1.22 0.77 _<0.000 55 6034.2 ' 19ø27.1 ' 28.6 34.61 1.07 0.77 1.63 _<0.008 56 6034.2 ' 19ø27.1 ' 1.25 0.80 1.73 _<0.008 57 6034.2 ' 19ø27.1 ' 1.30 0.76 1.82 _<0.002 58 6034.2 ' 19ø27.1 ' 1.27 0.83 1.67 _<0.000 59 6034.2 ' 19ø27.1 ' 1.25 0.70 1.82 _<0.000 60 6034.2 ' 19ø27.1 ' 27.3 36.37 0.90 0.81 _<0.000 61 6034.2 ' 19ø27.1 ' 26.5 0.72 0.83 1.58 _<0.009 62 6034.2 ' 19ø27.1 ' 27.0 35.70 0.65 0.69 1.71 _<0.000 63 6034.2 ' 19ø27.1 ' 26.5 36.09 0.80 0.78 1.62 _<0.007 64 6034.2 ' 19ø27.1 ' 1.28 0.78 1.49 _<0.000 65 6034.2 ' 19ø27.1 ' 26.7 36.15 1.38 0.78 1.53 _<0.001 66 6034.2 ' 19ø27.1 ' 1.26 1.09 1.44 _<0.013 67 6034.2 ' 19ø27.1 ' 1.24 0.78 1.68 0.041 71 2035.0 ' 18ø9.0 ' 27.2 36.79 0.78 1.07 _<0.012 73 2ø14.0 ' 20055.0 ' 27.0 0.70 0.55 _<0.001 75 200.5 ' 22042.7 ' 26.5 35.71 0.56 0.84 1.92 _<0.006 77 1ø31.8 ' 25046.7 ' 26.8 36.15 0.80 0.55 _<0.010

78 1ø28.1 ' 26042.6 ' 26.5 36.00 0.78 0.56 2.21 0.021 83 -1ø40.5 ' 30047.7 ' 26.3 36.22 0.86 0.53 2.18 _<0.006 84 --1ø40.5 ' 30047.7 ' 0.93 0.45 2.13 _<0.001 85 -1ø40.5' 30o47.7 ' 0.95 0.63 1.99 0.055 86 - 1 ø40.5' 30047.7 ' 0.94 0.53 2.05 _<0.004 87 --1ø40.5 ' 30047.7 ' 0.98 0.71 2.85 _<0.017 88 - 1 ø40.5' 30047.7 ' 0.99 0.54 2.25 _<0.007 89 --3038.7 ' 29023.0 ' 26.5 36.17 0.74 0.54 _<0.009 91 --5ø57.0' 2906.0 ' 26.5 36.26 0.70 0.82 2.22 _<0.005 92 --5029.5 ' 29051.0 ' 26.3 36.25 0.71 0.70 _<0.000 95 -7020.0 ' 32ø51.1 ' 26.5 36.20 0.66 0.49 _<0.007

Groups of samples listed at the same position were taken at 5 min intervals while the ship was steaming at 2 kn. Note that Cd blanks were higher than in other data sets.

BOYLE ET AL..' Cu, Ni, AND Cd IN N. HEMISPHERE SURFACE WATERS 8053

Sample Number

1

2

3 4

5 6 7

8

9

10

11 12

13 14

15

16 17

TABLE 4. Data From R/V Gillis, March 1-3, 1979

Latitude, øN Longitude, øW T S NO3 PO,• Si Cu Ni Cd

7025.5 ' 79055.7 ' 26.2 34.81 0.4 0.30 1.1 2.34 2.52 0.047 6058.0 ' 80022.0 ' 26.4 34.68 0.2 0.26 0.2 2.09 2.27 0.038 6037.0 ' 80045.0 ' 26.1 34.61 0.4 0.30 0.2 1.92 2.21 0.057 6ø10.4 ' 81ø14.1 ' 28.6 32.91 0.1 0.20 0.0 1.41 1.69 --<0.023 5046.3 ' 81035.3 ' 28.1 33.00 0.1 0.26 0.0 1.31 2.31 -<0.026 5ø21.1 ' 81051.2 ' 29.3 33.80 0.1 0.22 0.0 1.49 1.98 0.043 4057.0 ' 82022.0 ' 29.0 32.86 0.1 0.22 0.0 1.51 1.92 -<0.021 4030.0 ' 82ø51.1 ' 29.0 32.79 0.1 0.22 0.0 1.17 1.86 -<0.026 406.6 ' 83016.5 ' 27.6 33.42 1.5 0.37 2.3 1.39 2.49 0.040 3ø51.0' 83038.0 ' 27.4 33.34 0.2 0.32 1.0 1.24 2.26 0.074 3026.5 ' 84ø0.1 ' 27.5 33.33 0.1 0.30 1.1 1.39 2.57 -<0.011 303.4 ' 84ø23.1 ' 28.4 32.31 0.1 0.24 0.5 1.29 1.95 --<0.028 2042.3 ' 84039.5 ' 28.9 32.15 0.0 0.22 0.5 1.18 1.84 -<0.025 2ø17.0 ' 84057.0 ' 29.2 31.65 0.0 0.20 0.2 1.44 1.99 -<0.014 1ø51.4 ' 85014.0 ' 31.75 0.0 0.20 0.2 1.43 1.91 -<0.028 1ø32.7 ' 85031.5 ' 31.68 0.5 0.2 1.45 1.69 -<0.018 1ø14.6 ' 85059.0 ' 28.8 31.92 1.55 1.92 0.045

Note that NO3- and Si data may be offset from the true value due to an overcompensated blank. High Cd blank.

of copper in the ocean, such as the profile in Figure 10, show that values as high as 3 nmol/kg are not usually reached until 1-2 km; mid-latitude upwelling never draws on source waters that deep [Smith, 1968]. (3) River water contains significantly more dissolved Cu than seawater (20-30 nmol/kg) [Boyle, 1979]; ocean waters whose salinity was appreciably lowered by rivers would have higher copper concentrations. But in the Panama basin there are no significant rivers (the low salinity observed during DeSteiguer 1979 results from the excess of precipitation over evaporation [Sverdrup et al., 1942]). Also, no significant lowering of salinity is observed in the Gillis 1979 data, while a comparable copper excess is seen.

Although other mechanisms calling upon special aeolean inputs to near-coastal waters or radical differences in trace element uptake and recycling in coastal waters could be pro- posed to explain the coastal copper enrichment, there is no substantive evidence for such processes. Since the best docu- mented copper surface maximum is that in the Gulf of Pan- ama, consideration has to be given to the possibility that ship traffic (coated with copper-based antifouling paint) through the Panama Canal might be responsible for the excess copper. Calculations based on U.S. (extrapolated to world) production of antifouling paints (as reported in Chemical Week, July 25, 1979) suggest that if the residence time of water in the

Gulf of Panama is about I year then about 1% of the antifouling copper production would have to be released in the Gulf of Panama. This seems much too high to be realistic. It appears that the most plausible souce of excess copper in coastal waters is diffusive flux from near-shore sediments. Such a source has been definitively established for 228Ra, which originates in sediments from the decay of insoluble 2•Th [e.g., Kaufman et aL 1973]. Copper has no radiogenic source and must originate from diagenetic reactions in sediments. A similar coastal enrichment mechanism has been pro- posed for Mn [e.g., Yeats et al., 1979; Landing and Bruland, 1980]. In the deep ocean there is a strong source of copper to the bottom water as a result of the oxidative decomposition of organic material in surface sediments that has been enriched in copper by scavenging during descent through the water column. This process contributes a flux of about 2 nmol/cm•/ year [Boyle et al., 1977; Klinkhammer, 1980]. A plausible source of copper from continental shelf sediments is diagenetic reactions under mildly reducing conditions. Copper is strongly bound by iron and manganese oxides under oxic conditions; terrigenous sediments contain significant quantities of copper in this form [Gibbs, 1977]. When oxygen is depleted by biological processes using organic matter as an energy source, manganese and iron oxides can be employed as oxidant, re-

Sample Number

9

10

11

13 14 15

16

17 18

19 20 21

22

23

TABLE 5. Data From R/V Knorr, July 7-14, 1979

Latitude, øN Longitude, øW T S Cu Ni

11 ø29.0' 87 ø 12.0' 33.928 0.89 1.68 11ø28.0 ' 87016.0 ' 28.2 33.979 1.13 1.96 11 ø24.0' 87022.0 ' 28.6 33.901 1.42 1.96 11 ø 16.0' 87024.0 ' 28.4 33.822 1.03 1.97 11 ø9.0' 87035.0 ' 28.8 33.837 1.42 1.85 10043.0 ' 87059.0 ' 33.676 0.90 1.95 10042.0 ' 8804.0 ' 28.8 33.597 1.08 2.04 10015.0 ' 88028.0 ' 28.9 33.527 0.94 2.12 9036.0 ' 8909.0 ' 28.0 33.481 1.00 1.86 8059.0 ' 89046.0 ' 27.7 33.599 3.08 2.38 800.0 ' 90043.0 ' 28.1 33.357 2.49 1.70 704.0 ' 91041.0 ' 28.3 33.583 0.92 1.83 5028.0 ' 93017.0 ' 27.3 33.413 1.24 1.65 408.0 ' 94022.0 ' 26.5 33.921 1.40 1.75

Cd

-<0.007 -<0.031

0.O69 -<0.033 -<0.046

-<0.019 -<0.036 -<0.018 -<0.043

0.059 0.120

-<0.013

-<0.049

-<0.033

High Cd blank.


TABLE 6. Data From G. W. Pierce, July 22-29, 1979

Sample Number Latitude, øN Longitude, øW T Cu Ni Cd 22.0 33022.0 ' 62 ø 10.0' 24.5 1.22 2.06 <--0.013 22.5 33041.0 ' 61 ø39.0' 25.0 1.77 2.12 <-0.023 23.0 34ø31.0' 6007.0 ' 24.5 1.53 1.86 <-0.009 23.5 34ø31.0' 6007.0 ' 24.5 1.20 1.92 <-0.014 24.0 3603.0 ' 61 ø48.0' 25.2 1.41 1.86 <-0.011 25.0 37034.0 ' 62045.0 ' 26.0 1.52 1.85 <-0.019 27.0 39ø1.0 ' 64015.0 ' 25.2 1.43 1.91 <-0.012 27.5 38057.0 ' 64035.0 ' 1.28 1.91 <-0.029 28.0 39038.0 ' 65055.0 ' 25.6 1.39 2.15 <-0.012 29.0 40o38.0 ' 68o0.0 ' 17.4 2.63 4.12 0.164

No nutrient data were obtained on this cruise.

leasing Fe 2+ and Mn 2+ into the pore waters of the sediment biological activity is considerably higher than in the deep sea. [Froelich et al., 1979]. The trace metals bound in the oxides It should be noted however that many continental shelf also are released into solution and are free to migrate by diffusion. However, deeper in the sediment column, surfate reduction can lead to precipitation of copper sulfides. Copper remobilized in the upper portion of the sediment column can be permanently trapped in the deeper sediments by this process so that not all of the remobilized copper can escape. If the

sediments are relic gravels and sandstones [e.g., C, uilcher, 1968] and contain little copper [e.g., Bothner et al., 1980], so these sediments will not serve as good sources of copper (or Mn or 2:8Ra). The flux out of some shelf sediments must be higher than calculated above. Also, the above calculation neglects horizontal mixing, so a lower average residence time is

conditions in the upper sediments are only slightly reducing, a likely and thus a higher average flux would then be required. significant fraction of the copper can be trapped when up- It seems likely that the enhancement of copper in coastal wa- ward-diffusing Mn :+ encounters molecular oxygen and pre- ter is maintained by having a primary sediment flux recycled cipitates as MnO:. If the sediments were strongly reducing, by a nutrient-trap set via coastal upwelling. The flux of copper sulfide would be encountered within a short distance of the from the bottom would be derived then from two sources, a sediment-water interface, so that much of the copper again primary addition from the reduction of primary terrigenous would be trapped in the sediments. To derive a significant flux metal oxides and a cyclic component from the biological up- of copper, the sediments must be at an intermediate redox take and decomposition of the resulting organic debris. The state: The sediments must be sufficiently reducing so that sedimentary sources of copper, :28Ra, and manganese or other Mn :+ does not encounter dissolved oxygen until close enough elements remobilized in shelf sediments may differ sub- to the interface that its oxidation kinetics [Stumm and Mor- stantially depending on the redox chemistry of the elements gan, 1970] minimize its precipitation within the sediments, but not so reducing as to form sulfides near the sediment surface. The flux of copper is highest when there is a terrigenous source containing abundant metal-containing oxides and when there is just enough readily utilizable carbon to con- sume the diffusive flux of oxygen (and nitrate) within a few

and of the sediments. Mn :+ can be stopped at the surface by oxidation but will not be trapped deeper in the sediments as a sulfide, while ::8Ra will not be affected by either precipitation reaction. It appears likely that distinctive 'dye plumes' of these elements may originate from different sediment regimes.

Copper-enriched coastal water may leave the continental centimeters of the sediment-water interface. If this organic shelf either as part of a steady current or as an isolated eddy. carbon were just sufficient to reduce all of the labile metal The surface copper enrichments observed in the open ocean, oxides but less than enough to promote rapid sulfate reduc- such as Geosecs station. 340 (Figure 1), the two high points tion, there would be a strong flux of copper from the sedi- from the Knorr 1979 cruise (Figure 6), and the Atlantic Shelf ments into the overlying water column. Biological stirring of Water (Figure 7) are probably due to such movements of the upper sediment will tend to enhance the net flux, both by coastal waters into the open ocean. This signature of its providing a greater effective surface area and by moving solid coastal origins will be maintained until obliterated by mixing material into the remobilization zone. or removed by biological activity and general scavenging. The

The magnitude of the flux required to produce the Gulf of observation of a surface copper maximum at GEOSECS sta- Panama surface maximum can be shown to be within reason. tion 340 implies that this identity can be maintained for dis- Consider a 50 m deep water column impinging upon a conti- tances of the order of at least 1000 km. Noise in the copper nental shelf and traveling longshore for 1000 km at a speed of versus location plots can be interpreted as encounters with 5 cm/s [Mayer et al., 1979; Butman et al., 1979; Halliwell and waters having remnants of this coastal signature. This process Mooers, 1979]. This water will be above the shelf for 2/3 year will be variable, depending on seasonal changes in the cur- (neglecting exchange perpendicular to the flow, which will be rents and w, inds and perhaps even on temporal variations in discussed later). If the copper concentration of this water in- the chemistry and biology of the shelf sediments. There was creases from 1 to 3 nmol/kg, it will have to have been exposed no surface copper maximum observed in the DeSteiguer 1979 to a flux of 5 x 10 -7 nmol/cm2/s during that time. By using a transect as it crossed 9øN at 100øW (November) (Figure 8) al- plausible effective diffusion coefficient of 2 x 10 -6 cm:/s [Ber- though the maximum was observed in stations to the east ner, 1980], this flux would require a gradient of 250 nmol/kg/ (Knorr, July 1979, Figure 6) and west (GEOSECS 340, June cm in the pore waters at the sediment-water interface. Gradi- 1974, Figure 1) for samples taken in the summer. The large ents of copper in deep sea sediments are lower than this by seasonal variability of the currents in this region of the ocean about an order of magnitude. The mechanism of enrichment [Wyrtki, 1965] must be responsible for these temporal in coastal waters is different, however, and productivity and changes.

TABLE 7a. Data From USNS DeSteiguer, November 6 to December 1, 1979, Leg 1

Sample Number Latitude, øN Longitude, øW T S Si NO3 Cu Ni Cd

1 8025.5 ' 79034.2 ' 29.35 5.40 0.07 3.76 2.69 2 809.9 ' 79035.2 ' 29.19 5.13 0.16 3.65 2.74 3 7022.6 ' 79036.4 ' 29.24 5.44 0.19 3.91 2.68 4 709.2 ' 79037.2 ' 29.06 5.45 0.13 3.86 2.64 5 6026.2 ' 80011.3 ' 30.63 1.91 0.10 2.37 2.01 6 6023.0 ' 80026.2 ' 30.62 1.50 0.10 2.11 2.32 7 6ø17.1 ' 80051.6 ' 30.64 2.96 0.18 2.35 2.21 8 6ø13.3 ' 81ø5.9 ' 30.67 2.85 0.00 2.54 2.30 9 5059.3 ' 81034.0 ' 27.0 31.08 1.39 0.24 1.83 2.20

10 5ø38.1 ' 8300.8 ' 26.9 31.50 0.94 0.12 1.65 2.22 11 5ø24.1 ' 84ø1.5 ' 27.1 32.66 1.05 0.16 1.15 2.17 12 5ø16.4 ' 84038.7 ' 27.0 33.22 0.86 0.10 0.82 2.08 13 5ø13.2 ' 84ø54.1 ' 27.0 1.02 0.73 2.05 14 4044.9 ' 86023.8 ' 26.1 0.92 0.81 2.10 15 4o38.5 ' 86o55.2 ' 1.07 0.15 0.77 2.18 16 4o29.3 ' 87o31.4 ' 33.33 1.11 0.73 2.07 17 4ø15.4 ' 88o42.0 ' 33.34 1.16 1.22 1.97 18 4ø12.5 ' 88055.9 ' 27.0 33.32 1.12 1.04 2.11 19 4o4.7 ' 89o31.2 ' 26.8 33.29 1.12 0.75 2.16 20 3o59.8 ' 89o51.3 ' 26.8 33.32 1.07 0.33 0.86 2.21 21 3o50.8 ' 90o18.8 ' 26.8 33.24 1.12 0.67 2.21 22 3ø28.1 ' 91048.7 ' 26.6 33.16 1.04 0.90 2.09 23 3020.2 ' 92ø26.1 ' 26.7 33.21 1.06 0.82 2.23 24 3ø9.1 ' 92o54.8 ' 26.7 33.26 1.06 0.90 2.32 25 2058.4 ' 93015.4 ' 26.6 33.35 1.11 0.18 0.78 1.94 26 2053.5 ' 93023.5 ' 26.4 33.29 1.01 0.83 2.04 27 2o35.9 ' 93o55.8 ' 26.3 33.32 1.11 0.83 1.97 28 1ø48.1 ' 95o7.4 ' 25.8 33.56 1.14 0.16 0.85 2.02 29 lø26.8 ' 95o8.0 ' 25.4 33.81 1.26 0.10 1.59 2.05 30 1 ø5.5' 95 ø 12.9' 25.2 33.88 1.38 0.28 0.88 2.09 31 0o43.5 ' 95o15.7 ' 25.1 33.99 1.59 0.15 0.94 2.41 32 0 ø 15.5' 9507.3 ' 25.0 33.96 1.72 0.76 0.82 2.00 34 -1ø59.7 ' 9502.0 ' 25.0 34.00 1.51 0.73 0.74 2.29 35 -lø12.3 ' 95o2.0 ' 22.9 34.47 2.43 5.43 0.98 2.73 36 -2o59.6 ' 94o59.5 ' 22.9 34.55 3.25 6.18 0.86 2.66 37 -3045.8 ' 94058.0 ' 23.2 34.52 0.31 4.04 0.83 2.31 38 -3ø44.1 ' 95046.2 ' 22.9 34.54 4.08 0.71 2.00 39 -3ø44.1 ' 96032.8 ' 22.5 34.79 1.62 7.15 0.78 2.47 40 -3045.3 ' 9703.8 ' 23.0 35.00 5.13 8.46 0.78 2.92 41 -3o43.7 ' 97o28.2 ' 23.0 35.02 5.10 8.35 0.80 2.91 42 -3 o58.6' 99o22.5 ' 23.0 34.93 4.64 8.72 0.98 2.94 43 -2ø1.8 ' 99o45.5 ' 23.1 34.98 4.91 8.34 0.69 2.97 44 -2ø1.0 ' 10003.3 ' 23.0 35.02 5.06 8.67 0.96 2.85 45 -2024.2 ' 100010.4 ' 22.8 34.99 5.24 9.41 0.80 2.61 46 -2o48.3 ' 100 ø 15.0' 22.5 34.96 5.35 8.80 1.06 2.89 47 -lø19.0 ' 99o52.2 ' 22.5 34.83 5.17 8.70 0.81 2.92 48 -1ø44.9 ' 99055.8 ' 22.4 34.61 4.69 6.78 1.17 2.63 49 004.8 ' 100 ø 1.5' 22.8 34.57 4.17 6.36 0.81 2.78 50 0033.0 ' 10003.8 ' 22.6 34.55 3.85 6.58 0.92 2.71 51 0050.4 ' 10002.5 ' 23.0 34.60 1.90 5.24 1.00 2.59 52 1 ø22.5' 10006.5 ' 23.8 34.39 2.62 3.82 0.78 2.45

,

53 2ø11.8 ' 100o7.8 ' 26.0 34.47 0.93 0.10 0.95 2.15 54 2043.9 ' 100ø0.1 ' 26.0 33.75 1.15 0.23 0.82 1.75 55 3 ø 14.5' 10002.5 ' 26.2 33.93 1.27 0.68 1.96 56 3040.7 ' 100ø0.1 ' 26.5 33.89 1.11 0.88 1.'86 57 4ø17.3 ' 99o59.0 ' 27.0 33.93 1.07 0.58 2.19 58 5o3.9 ' 99o55.7 ' 27.0 33.87 1.14 0.68 1.95 59 5029.4 ' 99054.2 ' 27.2 33.74 1.12 0.75 2.04 60 6ø1.1 ' 99o58.6 ' 27.8 33.09 1.09 1.11 2.11 61 6o29.5 ' 99ø57.1 ' 28.0 32.82 1.10 0.77 1.82 62 7o2.8 ' 99o59.6 ' 28.0 33.08 1.21 0.69 1.70 63 7o34.6 ' 99o58.5 ' 27.7 33.14 1.22 0.88 2.02 64 8o30.7 ' 99o59.6 ' 28.1 33.23 1.28 0.88 1.99 65 9o0.8 ' 99o60.0 ' 28.2 33.48 1.19 0.12 0.90 2.15 66 9o26.2 ' 99o58.2 ' 28.0 33.75 1.28 0.33 0.86 2.17 67 9053.5 ' 99055.8 ' 27.9 33.87 1.57 0.73 1.04 2.21 68 10o23.9 ' 99o57.3 ' 27.1 33.08 1.99 1.35 1.11 2.22 69 10051.3 ' 99053.3 ' 27.0 33.93 2.04 2.09 1.07 2.24 70 11 o 19.6' 99o57.4 ' 26.9 33.74 2.47 2.52 1.05 2.24 71 11037.6 ' 10000.6 ' 26.0 33.14 4.71 6.93 1.01 2.85 72 12035.4 ' 100ø5.1 ' 26.7 33.89 3.67 5.61 1.06 2.37 73 12054.4 ' 10003.9 ' 27.5 33j09 2.14 2.00 0.87 2.25 74 13ø5.4 ' 100ø12.1 ' 27.6 33.23 1.36 0.50 1.20 1.98 75 14010.0 ' 100010.4 ' 28.1 33.93 1.35 0.16 1.00 2.04 76 14o43.0 ' 100o8.5 ' 28.0 32.82 1.89 0.25 1.21 2.21

0.049

0.034 0.042 0.049 0.014 0.017

0.023 0.030 0.030 0.025 0.018 0.018

_<0.014

0.019

_<0.011 _<0.010 _<0.008

0.016 0.021 0.022

_<0.008

_<0.013 0.025 0.016

_<0.009

_<0.010

_<0.013 0.019

_<0.012 _<0.012

0.021

0.016

0.017 0.061 0.055 0.073 0.049

0.065

0.073 0.080 0.064 0.071 0.071 0.078

0.078 0.062 0.080 0.064 0.091 0.082 0.046 0.019

_<0.008 _<0.006

_<0.010 0.017

_<0.008 _<0.011

_<0.013 _<0.013 _<0.009

_<0.011

0.020

_<0.011 0.021

_<0.014 0.039 0.033 0.045

0.085 0.083 0.041 0.036 0.024 0.046


TABLE 7b. Same as Table 7a, Leg 2.

Sample Number Latitude, øN Longitude, øW T S Si Cu Ni Cd

1 16049.0 ' 103043.5 ' 28.0 34.37 1.51 1.97 2.17 0.028 2 16044.3 ' 10405.2 ' 28.6 33.85 1.54 1.35 1.48 0.022 3 16047.5 ' 104047.5 ' 28.2 33.94 1.45 1.60 1.76 0.023 4 16053.0 ' 105026.7 ' 27.5 33.34 1.64 1.90 2.08 --<0.015 5 16049.3 ' 106010.9 ' 26.0 34.49 1.79 0.89 0.98 -<0.013 6 16050.5 ' 106053.3 ' 27.0 34.38 1.61 0.94 1.03 0.019 7 16053.4 ' 107042.2 ' 26.8 34.43 1.87 0.99 1.09 0.016 8 17ø1.4 ' 10803.6 ' 26.9 34.53 1.76 0.97 1.07 0.024 9 17015.5 ' 108042.0 ' 24.9 1.22 1.72 1.89 0.033

10 17030.9 ' 10908.2 ' 24.2 34.37 1.07 0.83 0.91 0.059 11 17037.5 ' 109055.9 ' 26.5 34.15 1.25 1.03 1.13 0.028 12 17045.0 ' 110018.5 ' 25.9 34.61 1.93 1.04 1.15 0.019 13 17ø54.1 ' 110045.7 ' 25.8 34.52 1.89 1.03 1.14 0.023 14 18ø7.0 ' 111ø24.0' 25.9 34.52 1.82 0.94 1.03 0.048 15 18023.9 ' 112ø7.9 ' 25.4 34.52 1.90 0.94 1.04 -<0.006 16 18ø46.1 ' 112027.8 ' 24.3 34.42 2.08 1.03 1.13 0.016 17 19ø17.0' 112052.0 ' 24.8 34.51 2.08 0.99 1.09 -<0.008 18 19050.5 ' 113014.9 ' 24.2 34.51 2.24 1.05 1.16 0.020 19 20ø21.0' 113035.0 ' 24.2 34.55 1.89 1.48 1.63 -<0.008 20 20059.2 ' 113ø56.0 ' 23.3 34.61 1.93 1.15 1.26 -<0.014 21 21ø32.0' 114014.8 ' 22.0 34.43 2.93 0.95 1.04 0.024 22 22ø0.1 ' 114ø43.0 ' 20.9 34.00 2.19 1.24 1.36 -<0.014 23 22019.6 ' 11507.5 ' 22.0 34.20 2.30 1.07 1.17 -<0.012 24 22037.0 ' 115017.0 ' 20.7 34.17 2.17 1.18 1.30 0.038 25 2307.2 ' 115040.5 ' 20.2 34.22 2.76 1.07 1.18 0.027 26 23ø32.1 ' 116011.4 ' 19.4 33.89 2.38 1.12 1.23 -<0.014 27 24ø2.0' 116033.0 ' 19.4 34.10 2.43 1.41 1.55 <0.015 28 24040.0 ' 117ø0.6 ' 20.1 33.93 2.47 1.56 1.72 <0.013 29 25ø1.7 ' 117019.2 ' 20.1 33.99 2.30 1.07 1.17 0.025 30 25027.3 ' 117ø44.1 ' 19.9 34.00 2.57 2.40 2.64 0.028 31 25043.7 ' 117049.2 ' 20.0 34.03 2.59 2.02 2.22 0.018 32 26017.0 ' 117ø43.5 ' 20.1 34.05 2.64 1.35 1.49 <0.014 33 26056.6 ' 117ø45.7 ' 20.3 34.12 2.68 1.16 1.28 <0.007 34 27033.6 ' 117ø43.7 ' 20.5 34.14 2.66 1.33 1.46 <0.012 35 28013.0 ' 117046.0 ' 19.0 33.61 2.34 1.32 1.45 0.024 36 28053.7 ' 117ø50.2 ' 19.0 33.65 2.54 1.12 1.23 0.022 37 29029.0 ' 117ø55.4 ' 19.0 33.70 2.74 1.33 1.46 0.018 38 30012.4 ' 117054.5 ' 19.0 33.71 2.93 1.36 1.49 <0.015 39 30048.8 ' 117059.4 ' 19.1 33.63 2.77 1.12 1.23 <0.015 40 31029.0 ' 11802.6 ' 17.9 33.52 2.25 1.15 1.27 0.023 41 3200.0 ' 118ø6.9 ' 17.1 33.65 2.40 1.87 2.06 0.089 42 32031.4 ' 11809.2 ' 16.7 33.60 2.30 1.02 1.12 0.039

The general increase of copper proceeding from the tropical variance is consistent with Ni-P correlations reported for pro- ocean into the cool eastern boundary currents could be due in file data. part to this coastal enrichment. But the magnitude of the in- In profile, nickel shows a dual covariance with phosphate crease and the general similarity of the increase to that of and silicate arising from common regeneration depths. As nickel and silicate in the eastern North Pacific suggests that the outcropping of cooler waters in the eastern boundary currents is probably sufficient to account for the increase.

An ironic consequence of the copper enrichment in coastal waters is the possibility that some subset of the older literature on copper in near-shore waters may have been correct. How- ever, since many of the same authors also reported high open- Depth Si NO3 Cu Ni Cd ocean concentrations, it is probably impossible to sort out the 25 1.8 0.27 valid data from erroneous data. 75 13.2 29.00

125 19.6 32.68 Nickel 175 22.2 33.75

Nickel variability reflects that of the labile nutrients. The 325 28.6 31.84 lowest nickel concentrations---2 nmol/kg--are the same in 450 39.8 29.32 575 52.0 30.75 both oceans. The equatorial Pacific enrichment in nickel-- 750 63.0 43.33 about I nmol/kg•ccurs along with a phosphate enrichment 1000 78.6 45.63 of I/•mol/kg (Figure 9b). There were no phosphate data for 1500 118.2 45.76 the second leg of this cruise, but Ni does show a good correla- 1750 133.9 40.02 2000 157.6 41.39 tion with surface temperature in this area (Figure 9a). The co-

pointed out by $clater et al. [1976], the distribution in a vertical profile can then be represented by an equation of the form

TABLE 8. Profile From USNS DeSteiguer, November 20, 1979, 12ø54.4'N, 100ø3.89'W

1.02, 1.03 1.33 3.02, 2.92 0.42, 0.37

1.13, 1.16 4.29, 4.58 0.60, 0.62 1.26 5.13, 5.06 0.63, 0.67

1.45, 0.99 5.34, 5.42 0.75, 0.74 1.33, 1.39 6.47, 6.39 0.89, 0.95 1.20, 1.30 7.46, 6.76 1.06, 0.95 2.02, 1.67 8.37, 8.02 1.60, 1.15 1.83, 2.05 9.17, 9.05 1.11, 1.12

2.34 9.74, 9.80 1.11 2.42, 2.54 9.51, 9.68 1.04, 1.06 3.35, 3.71 10.19, 9.99 1.03, 0.98


TABLE 9. Data From GEOSECS Stations 219, 293

Potential Salinity, Silicate, Phosphate, Depth, m Temperature, øC %0 /•m/kg /•m/kg

Nickel, nm/kg

Station 219, Oct. 8, 1973, 177ø17.5'W, 53ø6.6'N, Depth 3734 m 5 7.107 33.069 32.2 1.60

160 3.533 33.524 74.2 2.42 349 3.548 33.886 100.8 2.89 460 3.417 34.027 113.7 2.97 599 3.294 34.159 128.0 3.03 642 3.226 34.177 131.0 2.89 792 3.047 34.266 141.9 2.89 943 2.834 34.337 150.9 3.04

1189 2.531 34.413 163.3 2.95 1440 2.249 34.480 176.5 3.05 1667 1.983 34.536 190.1 3.01 1988 1.741 34.585 202.7 2.98 2287 1.594 34.613 209.5 2.91 2583 1.482 34.635 214.6 2.89 3480 1.310 34.665 226.9 2.80 3711 1.281 34.669 221.7 2.74

Station 293, March 1, 1974, 175ø5.0'W, 52ø46.0'S, Depth 5454 rn 3 11.328 34.419

216 7.65 34.419 458 6.46 34.342 685 5.27 34.268 961 4.07 34.31'2

1143 3.30 34.347 1290 2.83 34.395 1587 2.45 34.525 1733 2.37 34.577 2029 2.19 34.659 2326 2.018 34.708

2621 1.78 34.732 2916 1.52 34.736 3065 1.422 34.735 3509 1.05 34.723 4019 0.77 34.711 4474 0.586 34.705 4926 0.48 34.700 5271 0.446 34.699

6.1 6.5 7.4

8.1

7.2

8.4 7.6

8.5 10.0 9.6

10.6 8.6

11.3 11.2 10.4 10.8

2.7 0.72 4.1

5.2 1.28 4.5, 4.7 6.6 1.40 5.0

13.1 1.63 5.9

38.0 2.00 5.8, 6.2 44.8 2.13 7.0 53.8 2.24 7.4 68.9 2.28 8.0

73.4 2.27 7.2, 7.1 80.3 2.19 6.9 83.7 2.10 7.8

7.0 7.1

88.5 2.06 6.9 95.8 2.08 6.3 98.3 2.08 7.8

108.7 2.10 7.3 116.6 2.13 7.9 120.5 2.15 7.4

124.5 2.16 8.2, 7.8 125.3 2.17 7.2

Ni -- A + B(P) + C(S0, (Table 10). Bruland [1980] confirmed and that the coefficients are not globally valid. For example, this dual covariance for the eastern North Pacific for three sta- in Table 9 and Figure 11, previously unpublished GEOSECS tions between Monterrey and Hawaii (Table 10). It is impor- nickel profiles from the Pacific Antarctic and from the Bering tant to emphasize that these correlations are not mechanistic Sea are presented [Boyle, 1976 ]. These profiles also show sig-

'I ""'•'"_' ""l"ø'M'a'sW•i""'n'•xø'n'!fi_'___.' "'" '"? [• t 'e _!piti_ace 70 _____ 30•-- ........... ? BRULanD (1980).-* -- .... I I •XN, •• --[ •100Re (1978)

.............................. ' ............... -r7 x•-w'••t -• 'I ............. • [• DeSTeiGuer '791 Kno?r '70 0ceanus '78.'• •o ......................... I • --• ....... l,t i ......... ........ '-1 øI ..... ....... I-- - .... "i - ...... ::'"'

180 1:10 60

Fig. 2. Surface water sample locations from this study and from Moore [1978] and Bruland [1980].

8058 BOYLE ET AL.: CU, Ni, AND Cd IN N. HEMISPHERE SURFACE WATERS

LONCITUDœ

øE øW

lq0 160 180 160 lq0 120

To•:yo ........ So;,, Diego

20f•- ••p•atu•e. øC ,

3'1,.5

3',1,.0

33.5;

]2

O.q

0.2

0.0

Nickel. nmol/kg

. i

200

100

0

Fig. 3a. Data from Thomas Washington 1976.

g6

gOl

0œ1-

07.1

Ogl ggl o61 g61 oog go2

gig ogg

g22 oœ2 gœ2.

ot:,2 gt:,2

ogg

g92 -- OZ2--

gg2 062

gl o•,•

ot,œ --

ogœ

0 0 0 0 0 0 0 0 0 0

o 8 o ø g g o o

(w) 4ldec]


RECIFE

Temperature

DAKAR LISBON

ß ß - • I ............. t20 37

36

35

34

$ !ini y " .... "' a t

, . 11 . II

37

,35

ß ' ' il ' ,I ' ' ' i Silicate

II II ..............

Copper II ' II '

Lat., øS Lon•., øW

2

7 11 15 19 23 27 31 35 •t.øN

•11 ' '

Nickel ,

j - - - is - - it .... Cadmium

i

loo i

i

i

Fig. 4. Data from Oceanus 1978; average concentrations plotted.

ioo

5o

nificant dual covariances with phosphate and silicate (Table a subsequent increase in the regenerative flux of silicate rela- 10) but the silicate cofactors are significantly lower than those tive to nickel. Similar differences in linear element-silicate observed in the eastern North Pacific. These differences are correlations have been observed for barium [Chan et al., due to the excess of diatom productivity in high latitudes and 1977]. As an example of how even small changes in location


LATITUDE, ON 4 6 $ LATITUDE, øN

• • • • i • 4 & 8 I0 12

Temperature

Salhfity. :31 Silicate a

I 2.5

ß 1.5

__ ,m tit '/

•Nickei

ß ß ß .

m m m ß m m ß ß ß

Nitrate

PhosPhate • r _

28

27

26

Copper - - I -

_ _

Nickel -3

0

200 Cadmium 200

1OO I

1OO

o ,0.5 ' ß 0

Fig. 6. Data from Knorr 1979; individual analyses plotted.

o.o

o.5

o.o

Fig. 5. Data from Gillis 1979; individual analyses plotted.

The intercept is higher and the P coefficient is lower than those obtained by using the full data set. Clearly these regression equations must be treated as statistical properties of lim- ited data sets and not as global constants for the world ocean.

Cadmium

Cadmium variability also reflects that of the labile nutrients. It should be noted that our choice of a lower pre- concentration ratio (35: 1) compared to that of Bruland [1980] (300: t) precludes our detection of the lowest cadmium levels reported in the North Pacific central gyre (ca. t pmol/kg), which we accept as accurate. In profile, cadmium reflects

can affect the coefficients obtained from such regressions, the rapid regeneration in the upper waters similar to processes op- Ni-P-Si covariances were recalculated from the data of Bru- erating on phosphorus and nitrogen (Figure t0) [Boyle et al., land [1980] eliminating the one station near Hawaii and re- 1976; Martin et al., 1976; Bruland et al., 1978; Bruland, 1980]. taining the two stations in the California current (Table t0). When the surface water data is examined in detail, however,


LATITUDE, ON 35 40

Temperature

-25

2O

15

. Copper

'3

' Nickel

200

ioo

Fig. 7. Data from Pierce 1979; individual analyses plotted.

there are significant differences between the distributions of cadmium, phosphate, and nitrate. Bruland [1980] showed that cadmium was 'exhausted' after nitrate but before phosphate resulting in a linear regression with a negative intercept:

Cd = -0.07 + 0.35 (P) nmol/kg pmol/kg

In Figure 12a this line is plotted along with his observed points (open circles); the regression obviously is driven by the deep samples in that the observed Cd concentrations do not begin to increase until phosphate is greater than 0.3 pmol/kg [Collier, 1981]. In Figure 12a the equatorial Pacific (DeStei- guer 1979 leg 1) data are also plotted and it is clear that many of the points (where P > 0.4 pmol/kg) fall below Bruland's data set.

This behavior may be due to either (option 1) organisms preferentially removing cadmium relative to phosphate in the

equatorial Pacific, or (option 2) the data reflecting the outcropping of upper thermocline isopycnals where phosphorus is regenerated more rapidly than cadmium. The surface data is insufficient to resolve these mechanisms. Option 1 is discussed later in this paper, and Collier [1981] presents evidence for the second possibility.

In Figure 12b the cadmium-phosphate relation for the North Pacific (Thomas Washington 1976) data also is different from that of Bruland [1980]. Although most points fall along a line with a slope of about 4 x 10 -4 Cd/P, the concentrations are all higher than in Bruland's data set so that the line would pass through the origin. Since these samples are unfiltered, we cannot distinguish between dissolved and particulate cadmium, and our interpretation must be somewhat equivocal. But since most of the high Cd and P points are from the com- plex eddy structure of the western North Pacific [Kenyon, 1978, Figure 3], we propose that this Cd-P relation reflects the mixing of (1) outcropping nutrient-enriched thermocline water (that has not yet been affected by biological removal) and (2) low-nutrient gyre water.

These data prove that there are significant differences between the surface distributions of cadmium and phosphate. Determining the mechanisms producing these differences will contribute significantly to understanding of the biological uptake and recycling of trace metals.

MECHANISMS CONTROLLING THE TRACE ELEMENT

CONCENTRATION OF SURFACE OCEAN WATERS

While many trace elements (Cu, Ni, Cd, Ba, Se, As, Cr) show clear evidence of involvement in the biological cycle of near-surface removal and regeneration at depth, the degree of depletion in the surface waters varies enormously between these elements. Table 11 gives the relative depletion factors for these elements in the North Pacific Ocean. By what mechanisms do these inter-element differences arise? A complete understanding will of course require information on mechanisms for biological uptake and the subsequent transforma- tion into particulates that remove these elements from the mixed layer. However, these differences can be understood in a general sense through a basic parameterization. In its sim- plest form, the model assumes that primary producers (PP) remove the dissolved trace metals (X) by a fractionation factor a relative to dissolved phosphorus (P):

(X/P)pp a -- (1)

Grazers will then feed on this material and recycle a portion of the material back into solution and eliminate the rest of the

material as fast-sinking fecal material (Fro) that will remove the elements from the mixed layer. This introduces a further fractionation (fi) between X and P:

#= (X/P)pp (2)

Of course, a• will vary between populations and depend on environmental conditions. We choose phosphorus as the reference element since it is clear that trace element and phosphorus uptake continue long after nitrate is exhausted [Bru- land, 1980]. This parameterization is in no way intended to imply a direct link between phosphorus and the trace element or to assign them to any particular common phase or site.


SAN DIEGO ....... ii

....... ii

ACAPULCO

- - , ..... II , , II ß , II

PANAMA ' ii -

.11 .

- ii -

Salinity

il

ii - ii - ii ' ii -

II , , II I1 . .

Silicate

ICoPper ...... ,,- .... !

o N ø W Lat. La•.

........ ii - - !1 - ' II ....... ii -

ß ß ß ß ß

i i i . - - . . . . i . i . . . i .

o N øs o w o s o N o w øN [.at. Long. [.at. Long. Lat.

........ l| - ' II ' ' II ....... II ' t4

. II . . II . . II - . ..... II - t 0 Cadmium

IO

Nitrate

Fig. 8. Data from DeSteiguer 1979; individual analyses plotted.

BOYLE ET AL..' Cu, Ni, AND Cd IN N. HEMISPHERE SURFACE WATERS 8063

ß De•teig•er.•/9, ,•ø N' ' .

n II I,

10' ' ' 2'0 .... 3• ' ' Temperature, øC

Fig. 9a. Nickel versus, temperature for DeSteiguer 1979 surface samples north of the equatorial upwelling zone. Individual analyses plotted.

If water is upwelled in a batch with initial concentrations Xo and Po, and biological removal takes place without further addition, then

if a/• is constant, then

dX

dP -•) (3)

•oo [PoI (4) This behavior is illustrated for various values of aft in Figure 13. If aft -- 1, then there is a strict linear covariance between the trace element and phosphorus passing through the origin. If aft < 1, the relation shows that phosphorus is removed more rapidly than the trace element (within the scatter of real data it might appear to be a linear correlation) until the ratio be- comes high enough to overcome the effect of the discrimination factor. Substantial amounts of the trace element can re-

main then for large relative depletions of an essential nutrient. If aft > 1, then the element is preferred over phosphate; nitrate may be taken as an example of this case. In reality, organisms that live in nutrient-depleted waters probably are adapted to have higher discrimination factors for essential elements relative to potentially toxic trace elements such as arsenic or copper (e.g., see the hypothesis suggested by Benson quoted in Maugh [1979]). This variation of aft with X/P would extend the linear portion of the curves to nearer the intercept and would explain the residual quantities of many trace elements

Ni,

nmol/kg

3

ß

ß

ß

%

0.0 0.5 ' 1.0

Phosphate, pmol/kg Fig. 9b. Ni versus P, leg l, DeSteiguer 1979. P data interpolated

from cruise report.

at very low nutrient concentrations. In some cases (Cu, As) the variation in a/• with X/P may be necessary because of the potential toxicity problem, whereas for other elements the variation may be dependent on a biologically coincidental common uptake mechanism with a toxic element (i.e., if nickel uptake occurs by the same mechanism as copper uptake). Alternatively, organisms may fractionate the elements during uptake in order to maintain a biologically optimum chemical composition [Collier, 1981]. If these ideas have any validity, then it should be possible to reconcile the observed surface distributions of trace elements with the biological and chemical properties of these elements.

An estimate of the combined fractionation a/• can be ac- complished by considering the Bruland[1980] data on Cu and Ni in surface waters between Monterrey and Hawaii. The data were grouped into closely similar concentration ranges, averaged, and the ratio (AX/AP)/(Xavg/Pavg) plotted as a func- tion of surface P (Figure 14a) and Xavg/P• (Figure 14b). This approach treats the oceans as if batch upwelling occurs fol- lowed by closed-system removal of trace elements and phosphorus and neglects the effect of vertical and horizontal diffusion, so it must be considered as a very rough approximation. But it does demonstrate the qualitative decrease in a/• with increasing X/P. Collier [1981] made independent estimates of the fractionations during uptake (taken from bulk plankton composition) and regeneration (as based on plankton decomposition experiments); these are plotted in Figure 14b as open symbols; for copper the agreement is remarkably good considering the coarseness of the assumptions; for nickel the agreement is good in low productivity waters but poorer at

S•hcate. umol/kg

o o . t9o

ß

NitroLe, umol/kg •0 0 o

Copper', nmol/kg 2

' o 12

Nickel, nmol/kg ß

Cadmium, nmol/kg 0 !

ß

Fig. 10. Profiles from DeSteiguer 1979 station at 13øN, 100øW. Individual analyses plotted.

8064 BOYLE ET AL..' Cu, Ni, AND Cd IN N. HEMISPHERE SURFACE WATERS

TABLE 10. Ni (P, Si) Multiple Linear Regressions for Various Data Sets

Ni - A + B(P) + C(Si)

Data Set/Reference A B C

Sclater et al. [1975] 3.5 1.07 0.033 Bruland [1980] 2.7 0.95 0.033 GEOSECS station 219 4.8 0.38 0.020

(Berin• Sea) GEOSECS station 293 2.5 1.88 0.009

(Pacific Antarctic) Bruland [1980] 3.7* 0.64 0.033

stations CI, CII only

* Note that this intercept is close to that reported by Sclater et al. for their stations in the California current.

lOW Ni/P, although both estimates show the same trend of less discrimination at lower Ni/P. The situation for Cd is less

clear. While Cd does show more discrimination in low productivity zones, the element/phosphorus ratio actually decreases in the Pacific central gyre.

This model also makes specific predictions about the ap- proximate magnitude and qualitative trends that should be observed in the near-surface particle flux experiments that are becoming popular. None of this data is published, but the re- suits from two deep (26?0 and 536? m) experiments have been published that can be compared to our estimated fractionation factors for nickel (copper cannot be considered because of the deep water scavenging observed for this element [Boyle et al., 1977]). In the Sargasso Sea, where GEOSECS station 30 reports about 0.03 tunol/kg phosphate and for which we can estimate 2 nmol/kg nickel from this work, the particle flux data [Spencer et al., 1978] can be used to estimate an aft factor of 0.06 at Ni/P - 50 x 10 -3. In the eastern equatorial Pacific, where phosphate is about 0.5 tunol/kg [Bishop et al., 1981] and nickel can be estimated as 3 nmol/kg, the trap data of Cobbler and Dymond [1980] implies a• = 5 at Ni/P -- 6 x 10 -3. The Sargasso Sea trap estimate is remarkably similar to the surface water based estimate, and while the eastern equatorial Pacific trap estimate is outside the calibration range of the surface water estimate, it clearly is consistent with the no- tion that nickel is more discriminated against in waters of high Ni/P. Further data on the elemental composition of fast-sinking particles should refine these estimates.

Since this model only applies to the removal of trace elements from the mixed layer, it implies nothing about covariances due to regeneration. Thus this model assigns a• values greater than 1 to both cadmium and zinc in nutrient-rich waters, even though their different regeneration depths make their oceanic distributions resemble phosphate and silicate, respectively.

TABLE 11. Relative Surface Depletions for Trace Elements

Surface Deep Minimum, Maximum, Minimum/

Element nmol/kg nmol/kg Maximum

Cd 0.001 1.1 0.001 Zn 0.07 9 0.008 Cu 0.5 6 0.08 Ni 1.5 11 0.14 Ba 30 150 0.20 Se 0.55 2.3 0.24 Cr 1.2 3 0.4 As 1.1 1.9 0.6

M•I. nmollk 8

2 ß

3 ß

ß ß

Fig. 11. Nickel profiles from Pacific GEOSECS stations in the Ber- ing Sea (219) and Circumpolar Current (293).

Equatorial Pacific

./: ß

' ........ ' o'.s ' ' . Phosphate, •Jmol/kg

Fig. 12a

I• 2oo

North Pacific

i loo

Phosphate,/jmol/kg Fig. 12b

Fig. 12. Cd versus P; (a) equatorial Pacific (P data interpolated from cruise data report); (b) North Pacific. Open circles denote data from Bruland [1980]; solid line denotes his regression equation.


X/Xo

Fig. 13. Plot of trace element concentration (X/Xo) versus phosphate (P/Po) for closed system removal conforming to equation (3) for various values of aB.

CONCLUSIONS

1. Copper is significantly enriched in coastal water over continental shelves in all areas studied to date. This enrich-

ment can be greater than 2 nmol/kg and may be due to a nutrient-trapping of diagenetically remobilized copper from mildly reducing terrigenous sediments.

2, Copper-enriched coastal waters can be transported into the open ocean and retain this signature for considerable dis- tances from the source.

3. Copper and nickel also show small (0.4 and I nmol/kg, _

respectively) increases in cool eastern boundary currents re- suiting from the outcropping of cool nutrient-rich waters.

4. The removal of trace elements from the surface waters of

the ocean can be explained by a discrimination factor model in which the trace elements are taken up less effectively at higher element/phosphorus ratios. This process results in significant residual quantities of some trace elements in nutrient- depleted waters.

2-

Phosphate, ymol/kg

Fig. 14a. Estimates of the fractionation factor a• versus surface phosphate concentrations.

1.0 ' '. I ,

O• • ' '! ' •

log(X/P) Fig. 14b. Estimates a/• versus the element X/P ratio in surface

waters, derived from surface water data [Bruland, 1980].

.4cknowledgments. Dave Drummond collected the Thomas l•ashington 1976 samples, Barry Grant, Russ McDuff and Chris Measures helped collect the Gillis 1979 samples, Trudy Hall collected the Pierce 1979 samples, and Michael Bacon collected the Knon, 1979 samples. We thank these people for collecting the samples with great care. Rich Hatbison and Larry Madin allowed us to participate in the Oceanus 1978 cruise. AI Zirino made our participation in the DeSteiguer 1979 cruise possible. Barbara Mangum offered useful comments on the analytical procedure. Discussions with Bob Collier were particularly helpful in refining the ideas presented here. Ken Bruland and another 'anonymous' reviewer overlooking Narragansett Bay provided con- structive comments on the manuscript. We thank the officers and crews of the R/• Thomas l•ashington, R/V Oceanus, R/V Knorr, R/ V Gillis, USNS DeSteiguer, and G. I•. Pierce for their assistance. The Perkin-Elmer analytical equipment was purchased through a grant fro m the Sloan foundation. This research was supported by ONR contract N00014-80-C-0273.

REFERENCES

Andreae, M. O., Arsenic speciation in seawater and interstitial waters.' The influence of biological-chemical interaction on the chemistry of a trace element, Limnol. Oceang.i., 24, 440-452, 1979.

Bender, M. L., and C. L. (3agner, Dissolved copper, nickel, and cadmium in the Sargasso Sea, J. Ma•. Res., 34, 327-339, 1976.

Berner, R. A., Early Diagenesis, .4 Theoretical .4pproach, Princeton University Press, Princeton, N. 7., 1980.

Bishop, J., R. Collier, J. M. Edmond, and D. Ketten, The chemistry, biology, and vertical flux of particulate matter in the upper 1500 m of the Panama Bas'm, Deep Sea Res., 27.4, 615-640, 1980.

Bothnet, M. H., P. J. Aruscavage, W. M. Ferrebee, and P. A. Bae- decker, Trace metal concentration in sediment cores from the continental sheff off the South-eastern United States, Estuarine Coastal Mar. $cL, 10, 523-541, 1980.

Boyle, E. A., The marine geochemistry of trace metals, Ph.D. thesis, MIT/WHOI Joint Program, Cambridge, Mass., 1976.

Boyle, E. A., Copper in natural waters, in Copper in the Environment, edited by J. R. Nriagu, p. 77-87, John Wiley, New York, 1979.

Boyle, E. A., and J. M. Edmond, Determination of trace metals in aqueous solution by APDC chelate co-precipitation, in .4nab, tical Methods in Oceanography, .4C$ .4dr. Chem. Set. 147, edited by T. R. Gibb, Jr., p. 44-55, American Chemical Society, Washington, D.C., 1975a.

Boyle, E. A., and J. M. Edmond, Copper in surface waters sou• of New Zealand, Nature, 2õ$, 107-109, 1975b.

Boyle, E. A., and J. M. Edmond, Determination of copper, nickel, and cadmium in sea water by APDC chelate coprecipitation and flameless atomic absorption spectometry, .4hal. Chim. .4 eta, 91, 189-197, 1977.

Boyle, E. A., F. Sclater, and J. M. Edmond, On the marine geochemistry of cadmium, Nature, 265, 42-44, 1976.

Boyle, E. A., F. R. Sclater, and J. M. Edmond, The distribution of dissolved copper in the Pacific, Earth Planet. SoL Left., $7, 38-54, 1977.

Bruland, K. W., Oceanographic distributions of cadmium, zinc, nickel, and copper in the north Pacific, Earth Planet. $cL Left., 47, 176-198, 1980.

8066 BOYLE EL AL.: Cu, Ni, AND Cd IN N. HEMISPHERE SURFACE WATERS

Bruland, K. W., G. Knauer, and J. Martin, Cadmium in northeast Pacific waters, Limnol. Oceanogr., 23, 618-625, 1978.

Butman, B., M. Nobel, and D. W. Folger, Long-term observations of bottom current and bottom sediment movement on the mid-Atlan-

tic continental shelf, J. Geophys. Res., 84, 1187-1205, 1979. Chan,. L., D. Drummond, J. M. Edmond, and B. Grant, On the bar-

iu TM data from the Atlantic GEOSECS Expedition, Deep Sea Res., 24, 613-649, 1977.

Cobbler, R., and J. Dymond, Sediment trap experiment on the Gala- pagos spreading center, Equatorial Pacific, Science, 209, 801-803, 1980.

Collier, R., The trace element geochemistry of marine biogenic particulate matter, Ph.D. thesis, MIT/WHOI, Cambridge, Mass., 1981.

Cranston, R. E., and J. W. Murray, The determination of chromium species in natural waters, Anal. Chim. Acta, 99, 275-282, 1978.

Froelich, P. N., G. P. Klinkhammer, M. L. Bender, N. A. Luedtke, G. R. Heath, D. Cullen, P. Dauphin, D. Hammond, B. Hartman, and V. Maynard, Early oxidation of organic matter in pelagic sediments of the eastern equatorial Atlantic: Suboxic diagenesis, Geochim. Cosmochim. Acta, 43, 1075-1090, 1979.

Gibbs, R. J., Transport phases of transition metals in the Amazon and Yukon Rivers, Geol. Soc. Am. Bull., 88, 829-843, 1977.

Guileher, A., Estuaries, deltas, shelf, slope, in The Sea, edited by M. N. Hill, p. 620-654, Interscience, New York, 1968.

Halliwell, G. R., Jr., and C. N. Mooers, The space-time structure and variability of the shelf water-slope water and Gulf Stream surface temperature fronts and associated warm-core eddies, J. Geophys. Res., 84, 7707-7725, 1979.

Jenkins, W. J., Tritium and He-3 in the Sargasso Sea, J. Mar. Res., 38, 533-569, 1980.

Kaufman, A., R. M. Trier, and W. S. Broecker, Distribution of Ra- 228'in the World Ocean, J. Geophys. Res., 78, 8827-8848, 1973.

Kenyon, K. E., The surface layer of the eastern North Pacific in win- ter, J. Geophys. Res., 83, 6115-6122, 1978.

Klinkhammer, G., Early diagenesis in sediment from the eastern equatorial Pacific, II, Pore water metal results, Earth Planet. Sci. Lett., 49, 81-101, 1980.

Knauss, K. G., T. L. Ku, and W. S. Moore, Radium and thorium iso- topes in the surface waters of the east Pacific and coastal southern California, Earth Planet, Sci. Lett., 39, 235-249, 1978.

Landing, W. M., and K. W. Bruland, Manganese in the North Pacific, Earth Planet. Sci. Lett., 49, 45-56, 1980.

Martin, J. H., K. W. Bruland, and W. W. Broenkow, Cadmium transport in the California Current, in Marine Pollutant Transfer, edited by H. Windom and R. Duce, p. 159-184, Lexington Books, Lexing- ton, Mass., 1976.

Mayer, D. A., D. V. Hansen, and D. A. Ortman, Long-term current and temperature observations on the middle Atlantic shelf, J. Geophys. Res., 84, 1776-1792, 1979.

Measures, C. I., R. E. McDuff, and J. M. Edmond, Selenium redox chemistry at GEOSECS I, Earth Panet. Sci. Lett., 49, 102-108, 1980.

Moore, R. M., The distribution of dissolved copper in the eastern At- lantic Ocean, Earth Planet. Sci. Lett., 41, 461-468, 1978.

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.

Sclater, F. R., E. Boyle, and J. M. Edmond, On the marine geochemistry of nickel, Earth Planet. Sci. Lett., 31, 119-128, 1976.

Smith, R. L., Upwelling, Oceanogr. Mar. Biol. Ann. Rev., 6, 11-46, 1968.

Spencer, D. W., P. G. Brewer, A. Fleer, S. Honjo, S. Krishnaswami, and Y. Nozaki, Chemical fluxes from a sediment trap experiment in the deep Sargasso Sea, J. Mar. Res., 36, 493-523, 1978.

Strickland, J. D. H., and T. R. Parsons, A Practical Handbook of Aeawater Analysis, Fisheries Research Board of Canada, Ottawa, 1968

Stumm, W., and J. J. Morgan, Aquatic Chemistry, p. 535, Interscience, New York, 1970.

Sverdrup, H. U., M. W. Johnson, and R. H. Fleming, The Oceans, p. ,124. Prentice-Hall. Englewood Cliffs, N.J., 1942.

Waugh, T. H., It isn't easy being king, Science, 203, 637, 1979. Wyrtki, K., Surface currents of the eastern equatorial Pacific Ocean,

Inter Am. Trop. Tuna Comm. Bull., 9, 270-304, 1965. Yeats, P. A., B. Sundby, and J. M. Bewers, Manganese recycling in

coastal waters, Mar. Chem., 8, 43-55, 1979.

(Received October 13, 1980; revised April 27, 1981;

accepted April 20, 1981.)

on the distribution of copper, nickel, and cadmium in the surface...

Documents