Notice: Wiley Online Library will be unavailable on Saturday 27th February from 09:00-14:00 GMT / 04:00-09:00 EST / 17:00-22:00 SGT for essential maintenance. Apologies for the inconvenience.
 The distribution of silver in the North Pacific, determined using a suite of surface and subsurface samples, was found to be controlled by both natural and anthropogenic processes. Within the water column, silver was distributed as a typical nutrient-type element, as was previously observed in other oceanic waters. However, surface water concentrations of silver near Japan were the highest ever measured for the open ocean (up to 12 pM) and are considered evidence of anthropogenic contamination. Since prevailing wind flow during this study was westerly, this enrichment is attributed to aeolian fluxes of Asian industrial aerosols to North Pacific surface waters. Additional evidence of the impact atmospheric inputs of silver have on the North Pacific water column is shown by comparisons of intermediate depth waters analyzed during this study and during a study carried out 20 years previously that indicate temporal increases in industrial silver fluxes to these waters. Measurements from this study also confirm that silver is significantly correlated with dissolved silica, although the correlations are measurably different between the Atlantic and Pacific basins, and that the ratio of silver to copper is diagnostic of individual water masses in the Pacific and can be used as a geochemical tracer of ocean circulation. Overall, data obtained during this study enable a complete preliminary survey of silver in the World Ocean.
If you can't find a tool you're looking for, please click the link at the top of the page to "Go to old article view". Alternatively, view our Knowledge Base articles for additional help. Your feedback is important to us, so please let us know if you have comments or ideas for improvement.
 With the acquisition of data from the North Pacific, this study completes a preliminary survey of silver in the World Ocean. Previous silver data in the Pacific were limited to one site in the South Pacific [Murozumi, 1981], two sites in the eastern North Pacific [Martin et al., 1983], and three sites near Japan [Zhang et al., 2001]. This study has more than doubled the amount of silver data for the North Pacific using an extensive suite of surface and subsurface samples collected on a transect from Japan to Hawaii in the northwest and central North Pacific. These samples constitute the most comprehensive data set of silver for any one region of the World Ocean, and when combined with previous measurements of silver in the Atlantic and Southern oceans [Flegal et al., 1995; Ndung'u et al., 2001; Rivera-Duarte et al., 1999], enable a preliminary description of silver cycling in the World Ocean.
 This study was conducted as part of the fourth Intergovernmental Oceanographic Commission (IOC IV) Global Investigation on Pollution in the Marine Environment (GIPME), which had two main goals. The first was to investigate the role of atmospheric dust deposition on surface water biogeochemistry of the North Pacific, and the second was to complete a baseline contaminant survey for this region of the world ocean. (For a more thorough description of cruise objectives and background, please see in this theme C. Measures et al., Hydrographic observations during the 2002 IOC Contaminant Baseline Survey in the western Pacific Ocean, submitted to Geochemistry, Geophysics, Geosystems, 2004; hereinafter referred to as Measures et al., submitted manuscript, 2004). The following data were specifically gathered to provide a more detailed characterization of the biogeochemical cycling of silver and to determine whether atmospheric fluxes measurably influenced that cycling in the North Pacific as hypothesized in the Atlantic [Ndung'u et al., 2001; Rivera-Duarte et al., 1999].
2.1. Water Samples and Analyses
 All IOC IV samples were collected aboard the R/V Melville between May 1 and June 3, 2002, along a cruise track that extended north-east from Osaka, Japan, then south and east to Honolulu, Hawai'i (Figure 1). Subsurface waters were obtained from nine vertical profile stations using Teflon®-coated and acid-cleaned 30 L Go-Flo® sampling bottles. Each Go-Flo® was deployed to depth along a Kevlar® hydrographic cable and triggered with Teflon® messengers. To assess whether each Go-Flo® was tripped at the proper depth, salinity and nutrient samples from each bottle were checked against CTD salinity data and nutrient samples gathered from a rosette; data from bottles that obviously leaked or were contaminated were discarded. Surface waters were obtained from outside the ship's wake using a modified bathythermograph (a “fish”) towed from a boom extended about 8 m outboard from the port aft quarter of the ship. Water was pumped from the “fish” through Teflon® tubing into an onboard trace metal clean laboratory under (Class 100) HEPA filtration. All samples (n = 84) were collected in acid-cleaned 2 L polyethylene (LPDE) bottles using trace metal clean techniques [e.g., Flegal et al., 1991] and transported back to the laboratory at the University of California, Santa Cruz, where they were acidified to <pH 2 with ultrapure HCl (Optima® grade, Fisher Scientific) and then stored for at least six months before analysis.
 Samples were prepared and analyzed for silver (K. Ndung'u et al., manuscript in preparation, 2004) and copper [Ndung'u et al., 2003] concentrations using on-line direct flow-injection methods coupled with a Finnigan Element magnetic-sector high-resolution inductively-coupled plasma mass spectrometer (ICPMS). As a necessary pre-cursor to these methods samples were first digested with ultraviolet radiation in order to liberate any metal bound in refractory organic complexes, then introduced into online minicolumns loaded with a chelating resin (For silver, Dowex 1-X8, and for copper, Toyopearl AF Chelate 650 M). The resin was next washed of salts with Millipore® ultrapure (<18 MΩ cm−1) water, and the metals directly eluted into the ICPMS using ultrapure (Optima® grade, Fisher Scientific) HNO3. Samples, standards, reference materials, and blanks were treated identically and analyzed concurrently. Figures of merit for these analyses are given in Table 1.
Table 1. Analytical Figures of Merit for Silver and Copper Analyses Performed as Part of IOC IV in the North Pacific Ocean From May to June 2002a
Precision, % RSD
Accuracy was assessed with National Research Council of Canada certified reference materials (CRMs) for trace metals CASS-4 (Nearshore Seawater), NASS-4 (Open Ocean Seawater), and SLEW-3 (Estuarine Water). Although these CRMs are not certified for silver, SLEW-3 has an informational value of 28 pM. For copper, CASS-4 is certified at 9.32 ± 0.87 pM, and NASS-4 is certified at 3.59 ± 0.17 pM. Recoveries of copper in this study average 110 ± 4% and 109 ± 4% for CASS-4 and NASS-4, respectively. Precision was assessed as percent relative standard deviation (% RSD) on replicate analyses within an analytical session and was consistently <4%. Detection limits were calculated as three times the standard deviation of the blanks. Dashed lines indicate that SRM not measured for a particular analyte.
 The use of ultraviolet radiation to oxidize samples prior to analysis is unique to IOC IV methodologies; however, these new data are also considered comparable to previous data. Analyses of reference materials from this study (Table 1) are slightly higher than the certified values, but when the standard deviation of each measurement is included in the comparison there is no significant difference between the certified value and this study's results. Overall, a cautious assessment of these seemingly disparate methodologies indicates that a first-order comparison of silver data sets is indeed acceptable.
2.2. Site Hydrography
 The cruise track was designed to cover a variety of surface water regimes and current systems of the North Pacific, including the Kuroshio and Oyashio boundary currents along the western margin of the basin, and the subpolar and subtropical gyres. Station 1 was positioned within the Kuroshio Current, while Stations 2 (also Japanese JGOFS time series station KNOT) and 3 were located farther north in the subpolar gyre. Station 3, the northernmost profile, was situated in the middle of a high-nutrient low-chlorophyll region. Station 4 was located in a mixed water region that contained eddies from branches of both the warm-water Kuroshio and cold-water Oyashio currents. Station 5 was within an eastward-propagating extension of the Kuroshio Current, and was chosen to compare to the other vertical profile station in that system (Station 1). Stations 6, 7, 8, and 9 (also U.S. JGOFS time series station ALOHA) were all located within the North Pacific subtropical gyre.
 Vertical profile stations were located in order to sample specific water masses and to investigate the entrainment of any dust signal into the thermocline. Since this latter goal narrowed the cruise focus to primarily intermediate waters, subsurface sampling at depths >1500 m only occurred at Stations 2, 7, and 9. For Station 2 however, due to malfunctioning Go-Flo® bottles, there were no reportable silver data from depths >1200 m.
 Two specific subsurface water masses below the mixed layer were sampled: North Pacific Intermediate Water (NPIW) and Pacific Deep Water (PDW). NPIW is identified by a salinity minimum and is formed along the northwestern edge of the North Pacific [Talley, 1993] and then spreads out to the eastern and southern North Pacific [Talley et al., 1995]. Ventilation of NPIW in both the subtropical and subpolar gyres is by inputs created each winter as brine exclusion during sea ice formation in the Sea of Okhotsk [Shcherbina et al., 2003]. Below ∼3°C (1,000–1300 m) samples came from the nearly-uniform Pacific Deep Water, which has its source in Antarctic Bottom Water created in the South Pacific. A more detailed discussion of the physical hydrography observed during IOC IV is given by Measures et al. (submitted manuscript, 2004).
 As stated earlier, IOC IV aimed to examine the effects of a major dust storm on surface ocean biogeochemistry, and accordingly the cruise was timed to occur in the spring during the Asian high-dust season [Prospero et al., 1989]. Unfortunately, no large dust storm was encountered during the cruise, although smaller “dust events” were noted from high levels of particulate iron and aluminum in aerosol samples on May 3–5, May 9–13, May 19 and May 31. For further information about dust deposition during IOC IV please see C. Buck (manuscript in preparation, 2004), and for details about silver and other trace element deposition please see M. Ranville (manuscript in preparation, 2004).
3. Results and Discussion
3.1. Surface Waters
 Not including coastal areas and estuaries [Sañudo-Wilhelmy and Flegal, 1992; Sañudo-Wilhelmy et al., 2002; Squire et al., 2002; Wen et al., 1997], the highest oceanic surface water concentrations of silver ever measured were found during this study (Table 2). The greatest concentrations were near the Asian mainland (up to 12 pM), and although higher than any previous open ocean surface water data, were consistent with data from surface waters around Japan (4–8 pM) obtained by Zhang et al. . Near the central part of the North Pacific gyre, concentrations dropped to levels of 1–2 pM. The resulting spatial distribution of surface water concentration data, which were highest on the western margin of the basin and steadily decreased eastward (Figure 2), corresponded to prevailing westerly wind flow from the Asian mainland.
Table 2. Silver Concentrations From Surface Waters Obtained During IOC IV From May to June 2002 in the North Pacific Ocean
Date Collected, UTC
 This correspondence suggests that atmospheric transport of mineral and industrial aerosols over the North Pacific [e.g., Arimoto et al., 1989, 1996; Duce et al., 1991] is responsible for elevated silver concentrations in North Pacific surface waters. Increased surface water concentrations as a consequence of atmospheric deposition has been suggested in other studies; for instance surface water concentrations in the Atlantic (up to 4.5 pM) modestly elevated above background (0.25 pM) were tentatively attributed to deposition of Saharan dust and industrial aerosols [Ndung'u et al., 2001; Rivera-Duarte et al., 1999]. If baseline levels of 0.25 pM in the Atlantic are considered to be roughly the same for the Pacific, then some surface water silver concentrations measured during this study are as much as 50-fold enriched over that proposed background value.
 Over the North Pacific, atmospheric silver inputs are believed to predominantly originate anthropogenically as opposed to transport of mineral aerosols, since atmospheric silver is primarily associated with high-temperature industrial emissions such as coal-burning and copper and zinc metal refining [Flegal et al., 1997; Purcell and Peters, 1998]. To test this assumption, the relationship was examined between surface water concentrations of silver and aluminum, which is considered a conservative marker for atmospheric inputs of mineral aerosols. Utilizing a simple linear correlation, silver and aluminum were found to not significantly covary, but instead be weakly and negatively correlated (n = 29, Ag = −1.3 Al + 7.6, R2 = 0.46). This weak, negative correlation contrasts with the significant, positive correlation (simple linear regression, R2 = 0.91) between silver and aluminum concentrations in the North and South Atlantic, that was attributed to the predominance of atmospheric inputs mineral aerosols in that region [Ndung'u et al., 2001]. But for this study, the lack of covariance between silver and aluminum lends support to the hypothesis that elevated silver concentrations in North Pacific surface waters are the result of atmospheric fluxes of industrial emissions rather than natural processes.
 That hypothesis was then tested by calculating a seawater enrichment factor for silver relative to aluminum, using surface water concentrations of silver and aluminum as compared to their upper continental crustal values [Taylor and McLennan, 1995]:
This measure indicated that surface water silver concentrations were, on average, enriched 5800-fold over crustal values relative to those of aluminum. For comparison, using the same upper crustal values used by Taylor and McLennan , silver in the Eastern Atlantic was only enriched sevenfold [Flegal et al., 1995]. Therefore the enrichment of silver over aluminum in surface seawater, the lack of correlation between the two elements in surface waters, and the spatial distribution of silver concentrations all suggest that silver is substantially elevated in North Pacific surface waters by atmospheric inputs that are predominantly industrial in origin.
3.2. Vertical Profiles
 The distribution of silver within the water column was consistent with the oceanic conveyor belt circulation model, which holds that at depth older Pacific waters are enriched in nutrient-type elements in comparison to younger Atlantic waters [Broecker and Peng, 1982]. All nine vertical profiles of silver acquired during IOC IV exhibited nutrient-type behavior (Figure 3), showing depletion at the surface and substantial enrichment at depth. Deep water concentrations exceeded 85 pM at Station 9 and were the highest ever measured in open ocean waters (Table 3). Data from this study, an earlier North Pacific study, and from previous IOC cruises in the Atlantic Ocean were combined in Figure 4 to show a range of silver concentrations for each ocean basin. The effects of oceanic circulation are clearly demonstrated, as ranges increased from the North Atlantic to the South Atlantic, and then dramatically to the North Pacific.
Table 3. Concentrations of Silver and Dissolved Silica, and Ag:Si Molar Ratios for Stations 1–9 Obtained During IOC IV in the North Pacific Ocean From May to June 2002a
Dashes indicate measurement was below detection limits. Dissolved silica measurements provided by K. Morse (unpublished data, 2003).
Station 1 (34°28′N, 146°59′E)
Station 2 (44°00′N, 155°00′E)
Station 3 (50°00′N, 167°00′E)
Station 4 (39°21′N, 170°34′E)
Station 5 (33°45′N, 170°35′E)
Station 6 (30°30′N, 170°35′E)
Station 7 (24°15′N, 170°20′E)
Station 8 (26°00′N, 175°00′W)
Station 9 (22°45′N, 158°00′W)
 In addition to global ocean circulation, vertical profile data from this study suggest an additional source of silver to the North Pacific. Silver data from the other Pacific studies, while within the general concentration pattern established by global oceanic circulation, are lower than concentrations found during this study (Figure 5). Zhang et al.  published data from three profiles obtained in 1998 near Japan, including one at ∼40°N, 145°E that was located near this study's Station 1 (∼34°N, 147°E). At 1483 m they found silver concentrations of 34 pM, while at 1342 m this study measured silver concentration of 59 pM. Moreover, there was an even greater inconsistency between silver concentrations that we measured at comparable depths as compared to two eastern North Pacific profiles obtained in 1981 by Martin et al. . Both of their profiles (at 18°N, 108°W, and 36.5°N, 123°W) had silver concentrations of 23 pM at 2,440 m, as compared to a silver concentration of 88 pM at 2508 m measured at Station 9 during this study. Consequently, silver concentrations measured at depth in this study are four-fold higher than those measured at other locations in the North Pacific in the previous two decades.
 Some of these disparities might be explained by differences in sampling protocols or hydrographic conditions. For example, Zhang et al.  analyzed filtered samples (0.04 μm) while this study analyzed unfiltered water, and they characterized the water mass they sampled as part of the Oyashio Current, while this study's Station 1 was located within the Kuroshio Current. Other possible reasons for the variability between data sets, including the larger discrepancy between Pacific data from Martin et al.  and this study are unknown.
 But it is proposed that the discrepancies in silver concentrations at intermediate depths between this study and the eastern North Pacific [Martin et al., 1983] reflect temporal increases in atmospheric inputs of contaminant silver aerosols associated with recent increase in Asian industrial emissions. Those emissions have risen dramatically over the last 20 years [e.g., Jacob et al., 1999; Streets et al., 2000; van Aardenne et al., 1999] due to increases in industrial capacity and energy production [Energy Information Administration, 2004]. Further, it is speculated that much of the silver contamination in the North Pacific is derived from deposition of Asian industrial aerosols in the Sea of Okhotsk.
 The Sea of Okhotsk is in close proximity to the Asian mainland and directly in the path of prevailing westerly winds, and each winter brine-rejection creates dense water that sinks and passes through straits in the Kuril Islands to ventilate NPIW [Shcherbina et al., 2004]. Silver concentrations in a profile from the Sea of Okhotsk obtained by Zhang et al.  range from 4.2 pM (surface) to 47 pM (∼3000 m), which is considerably higher than the range of 1.3 pM (surface) to 23 pM (∼2500 m) observed in the eastern North Pacific by Martin et al. . In NPIW, evidence of ventilation by water formed in the Sea of Okhotsk is seen to a density of 27.6 σθ [Talley, 1991], which corresponds to a depth of ∼1500 m for most profiles collected during this study. Furthermore, by using chlorofluorocarbons (CFCs) as geochemical tracers, Warner et al.  estimated that NPIW in the subtropical gyre is ventilated on a timescale of 20–25 years, and that the western side of the subtropical gyre is ventilated to a greater depth (∼1200–1800 m) than the eastern side (∼600–800 m). Similarly, Fine et al.  also used CFCs to calculate water mass ages along isopycnals in the North Pacific and found the ages increased from west to east (e.g., ages increased from 15 to 35 years along 27.2. σθ).
 These studies support the proposal that recent temporal increases in silver fluxes to the western Pacific are incorporated into intermediate waters. Ventilation of NPIW to intermediate depths, which begins on the western edge of the basin and continues to the east, can occur on decadal timescales and thus could explain the discrepancy between vertical profiles collected in this study and ones collected in 1981 in the eastern Pacific [Martin et al., 1983]. To conclusively test this proposal a series of deep profiles in the eastern North Pacific would need to be sampled and analyzed using the same methodologies employed during IOC IV.
 More specifically, within the Pacific water column silver remobilization seems to be somewhat delayed as compared to the Atlantic, although the reasons for this are unclear. An interbasin fractionation of 12 exists for silver between North Pacific data from this study and data from the far North Atlantic (i.e., 6.9 pM at 3000 m [Rivera-Duarte et al., 1999]). This silver fractionation is more than the 10-fold interbasin fractionation observed for dissolved silica [Bruland and Lohan, 2004], again suggesting that for unknown reasons silver is more slowly remineralized than silica.
 Examination of Ag:Si molar ratios (Table 3) in the Atlantic and Pacific oceans also show a buildup of silver at depth in the Pacific Ocean as compared to biogenic silica. As illustrated in Figure 7, the Ag:Si molar ratio at depth is greater (0.5, ≥1500 m) for North Pacific vertical profiles than for Atlantic profiles (e.g., ∼0.2 in the eastern Atlantic [Flegal et al., 1995]). These ratios again indicate that silver is regenerated more slowly than biogenic silica, although it is unclear if slower regeneration is a phenomenon unique to the Pacific Ocean or a global oceanic circulation feature that only becomes evident in deep Pacific waters. Alternatively, the non-linearity of Ag:Si plots in the Pacific could include an additional source of silver in deep Pacific waters, such as remineralization of bottom sediments, or more efficient transfer of silver out of surface waters on sinking particle aggregates [Fisher and Wente, 1993].
 Although Ag:Si ratios in oceanic surface waters are much more variable, differences in those ratios between the Atlantic and Pacific provide further evidence of external inputs of silver to the North Pacific. In the eastern Atlantic [Flegal et al., 1991] and the south-western Atlantic [Ndung'u et al., 2003] the Ag:Si correlation intercepts are 0.5 and 1.7 (Figure 6), respectively, while for the far North Atlantic Ocean [Rivera-Duarte et al., 1999] the intercept is considerably higher (2.3). Rivera-Duarte et al.  hypothesized that elevated surface water concentrations in the North Atlantic were the result of external inputs, although it was unclear whether this elevation is due to atmospheric or terrestrial sources. The Ag:Si intercepts further increase to 3.1 (IOC IV) and 4.3 [Zhang et al., 2001] in the North Pacific, where elevated surface water concentrations are more convincingly interpreted as evidence of atmospheric deposition. In conclusion, examination of silver and dissolved silica correlations in oceanic waters leads to qualitative conclusions about the existence of external inputs and biogeochemical recycling of silver within the water column.
3.4. Covariance With Copper
 Silver is also correlated with copper in the World Ocean, which has led to the proposal that Ag:Cu ratios of could be used as a geochemical tracers of water masses [Sañudo-Wilhelmy et al., 2002]. Using data from the Atlantic [Flegal et al., 1995], the Pacific [Martin et al., 1983], and the Southern oceans, Sañudo-Wilhelmy et al.  observed that the Ag:Cu ratio increased along the lines of global ocean circulation from the Atlantic to the Pacific while the intercept of that correlation became more negative. Data from this study are in agreement with this pattern and exhibit a Ag:Cu ratio of 24 and an intercept of −17 (Figure 8).
 Although the overall Ag:Cu ratio for the North Pacific is in agreement with the global pattern, individual water masses sampled during this study exhibited different Ag:Cu ratios, suggesting that biogeochemical cycling of silver and copper differ between water masses within a single basin (Figure 9). Ratios in surface waters (≤175 m) and deep waters (>1000 m) have positive intercepts, which contrasts with the Ag:Cu correlations observed in other ocean basins and the North Pacific as a whole. The slightly positive intercept in surface waters is interpreted as signaling relatively greater atmospheric inputs of silver than copper to those waters, while in deep waters the positive correlation is tentatively attributed to more efficient biogenic recycling of copper relative to silver in the North Pacific. In contrast, the Ag:Cu correlation of the intermediate waters (200–1000 m) is more similar to the whole North Pacific correlation. This gradient could be the result of simple mixing of water masses from above and below, or could instead reflect different regeneration characteristics of intermediate-depth waters. The Ag:Cu correlation in North Pacific intermediate waters is also similar to the Weddell Sea correlation ([Ag] = 11[Cu] − 5) given by Sañudo-Wilhelmy et al. , although the significance of this is unknown.
 Data gathered during IOC IV complete a preliminary survey of silver in the World Ocean. Nine vertical profiles confirm that silver is distributed within the water column as a nutrient-type element, and in accordance with general ocean circulation models, is substantially enriched at depth in the Pacific as compared to the Atlantic. Silver was again correlated with dissolved silica, but correlations were found to vary between the Pacific and the Atlantic oceans, indicating that silver is remobilized more slowly in the deep Pacific than in the deep Atlantic or refractory silver is being advected from the Atlantic to the Pacific. Additionally, silver was also positively, but variably, correlated with copper, confirming that Ag:Cu ratios can be used as geochemical tracer of water masses.
 Most notably, our measurements of silver concentrations of up to 12 pM in surface waters were the highest ever mentioned in the open ocean, and our measurements of silver concentrations (e.g., 88 pM) in intermediate waters (2400–2500 m) were approximately four-fold greater than those (e.g., 23 pM) previously measured in eastern North Pacific waters [Martin et al., 1983]. This apparent increase in silver concentrations in the North Pacific is tentatively attributed to the corresponding increase in Asian industrial emissions and their subsequent incorporation into the water column. It is proposed those emissions have elevated silver concentrations in surface waters as much as 50-fold above theoretical background values. If substantiated, this apparent enrichment indicates that a substantial anthropogenic perturbation of the ocean biogeochemical cycling of silver is now occurring, and that silver is one of the most (if not the most) contaminated elements in the World Ocean.
 The authors would like to thank Chief Scientists Chris Measures, William Landing, and Greg Cutter for the opportunity to participate in IOC IV. The authors would also like to thank the crew of the R/V Melville, Keola Morse of the University of Hawaii for the dissolved silica data, Lynne Talley of Scripps Institute of Oceanography, and Rob Franks, Kuria Ndung'u, Christopher Conaway, Genine Scelfo, Kristen Buck, and Helen Cole at UCSC. Comments and suggestions from two anonymous reviewers enhanced the final version of this manuscript.