7/26/2019 Controls of Trace Metals in Seawater
1/25
6.02
Controls of Trace Metals in Seawater
K. W. Bruland and M. C. Lohan
University of California at Santa Cruz, CA, USA
6.02.1 INTRODUCTION 236.02.1.1 Concentrations 246.02.1.2 Distributions 27
6.02.1.2.1 Conservative-type distributions 276.02.1.2.2 Nutrient-type distributions 286.02.1.2.3 Scavenged-type distributions 306.02.1.2.4 Hybrid distributions 316.02.1.2.5 Mixed distributions 32
6.02.2 EXTERNAL INPUTS OF TRACE METALS TO THE OCEANS 326.02.2.1 Rivers 326.02.2.2 Atmosphere 336.02.2.3 Hydrothermal 35
6.02.3 REMOVAL PROCESSES 356.02.3.1 Active Biological Uptake in the Surface Waters 35
6.02.3.1.1 Lessons from laboratory studies 356.02.3.1.2 Non-Redfieldian assimilation 36
6.02.3.2 Passive Scavenging 376.02.3.2.1 Adsorption/desorption processes 376.02.3.2.2 Lessons from radionuclides 37
6.02.4 INTERNAL RECYCLING 396.02.4.1 Recycling within the Water Column 396.02.4.2 Benthic Inputs 40
6.02.5 COMPLEXATION WITH ORGANIC LIGANDS 406.02.5.1 Copper 416.02.5.2 Iron 426.02.5.3 Zinc 43
REFERENCES 44
6.02.1 INTRODUCTION
Since the early 1970s, marine chemists have
gained a first-order understanding of the concen-
trations, distributions, and chemical behaviors oftrace metals in seawater. Important factors initiat-
ing this quantum leap in knowledge were major
advances in modern analytical chemistry and
instrumentation, along with the development andadoption of clean techniques. An instrumental
development in the mid-1970s that spurred theearly research on trace metals was the availability
of the sensitive graphite furnace as the sampleintroduction system to an atomic absorption
spectrometer. More recently, the appearance
of inductively coupled plasma (ICP) massspectrometers has provided an even more sensi-tive and powerful instrumental capability to thearsenal of marine chemists. In addition to theseinstruments back in shore-based laboratories,there has been the development of sensitiveshipboard methods such as stripping voltammetryand flow injection analysis (FIA) systems witheither chemiluminescence or catalyticallyenhanced spectrophotometric detection. Along
with the development of these highly sensitiveanalytical techniques came a recognition andappreciation of the importance of handlingcontamination issues by using clean techniquesduring all phases of sampling and analysis. This is
23
7/26/2019 Controls of Trace Metals in Seawater
2/25
necessary due to low concentrations of tracemetals in seawater relative to the ubiquitousnessof metals on a ship or in a laboratory (e.g., dust,steel hydrowire, rust, paint with copper and zincantifouling agents, brass fittings, galvanizedmaterial, sacrificial zinc anodes, etc.). As a result,
seawater concentrations of most trace metals havenow been accurately determined in at least someparts of the oceans, and their oceanic distributionshave been found to be consistent with oceano-graphic processes.
The concentrations and distributions of tracemetals in seawater are controlled by a combi-nation of processes. These processes includeexternal sources of trace metals delivered byrivers along ocean boundaries, by wind-blowndust from arid and semi-arid regions of thecontinents, and by hydrothermal circulation atmid-ocean ridges. Processes removing tracemetals from seawater include active biologicaluptake or passive scavenging onto either living ornonliving particulate material. Much of thisparticulate material (along with its associatedtrace metals) is internally recycled either in thewater column or in surficial sediments. Theultimate sink of trace metals is generally marinesediments. These various sources and sinks aresuperimposed on the general circulation andmixing of the oceans, resulting in the character-istic distributions of each trace metal. One of thefirst examples of the emergence of oceanographi-cally consistent vertical profiles was for the trace-metal cadmium (Boyle et al., 1976;Martin et al.,1976; Bruland et al., 1978a). These studiesdemonstrated that the distribution of dissolvedcadmium in the sea follows a pattern similar tothat of the nutrients phosphate and nitrate.Sparked by these surprising results, severalinvestigators during the following two decadeswere able to obtain excellent data sets on a widevariety of trace metals. This chapter will attemptto provide a basic overview of what is knownabout the controls of the concentrations and
distributions of trace metals in the open ocean.Subtleties in their distributions will not bepresented. The distributions of trace metals incoastal regions are more dynamic and compli-cated and will not be discussed in this chapter.
The bulk of the data for vertical profiles of tracemetals in seawater are from papers published inthe 1980s and 1990s and most of the profiles arefrom either the North Pacific or North Atlantic.There is a paucity of vertical profiles from theSouth Atlantic and South Pacific. It has recentlybeen argued that a new GEOSECS-type trace-
metal program needs to be in place in order toprovide appropriate global coverage of tracemetals. Much of the impetus for such a programcomes from the recognition of iron as an importantmicronutrient influencing global biogeochemical
cycles in the oceans (Moore et al., 2002) and thepotential role of other trace metals such as zinc. Inparticular, there is a pressing need for anexpansion of the global database of dissolvediron distributions in the oceans. These measure-ments are needed to both initiate and verify
models and to identify processes not contained inexisting models.There have been a number of reviews of trace
elements in seawater that form a foundationfor this chapter. Among them are: Bruland(1983)on oceanographically consistent data sets;Burton and Statham (1990) on trace metals inseawater; andDonat and Bruland (1995)on traceelements in oceans. There are two reviews thatdeal with more of the biological role of tracemetals: Bruland et al. (1991) on interactiveinfluence of bioactive trace metals on biologicalproduction in ocean waters; and Hunter et al.(1997) on biological roles of trace metals innatural waters. A highly complementary chapterin this Treatise that deals with the influence ofessential trace metals on biological processes hasbeen written by Morel et al. (Chapter 6.05).Turning to on-line sources of information,Nozaki has done an excellent job perusing theavailable literature and compiling vertical profilesfrom the North Pacific for each element in aperiodic table that makes an excellent figure(http://www.agu.org/eos_elec/97025e.html). KenJohnson, a marine chemist at the Monterey Bay
Aquarium Research Institute (MBARI), has a website with a periodic table of the elements contain-ing a brief review of information on each element(http://www.mbari.org/chemsensor/pteo.htm).
6.02.1.1 Concentrations
Concentrations of trace metals in seawaterfall within a range bounded by ,10 mmol kg21
at the upper end and extending through fmol kg21
at the lower end (m(micro) 1026, n (nano)
1029, p (pico) 10212, and f (fempto) 10215).The concentrations of metals in seawater rangeover 15 orders of magnitude from sodium, themost abundant cation, at a concentration close to0.5 mol kg21 to iridium at a concentration aslow as 0.5 fmol kg21. A concentration of10 mmol kg21 (,1 ppm by weight) is chosen asthe concentration separating trace metals fromthe major and minor metals in seawater. Themean concentration of a trace metal is stronglyinfluenced by its deep-water value, particularlythat found in the deep waters of the Pacific Ocean
with its large volume. Table 1presents the rangeof concentrations and their mean for many of theelements in seawater along with an estimate ofthe major inorganic form or species found inseawater.
Controls of Trace Metals in Seawater24
http://www.agu.org/eos_elec/97025e.htmlhttp://www.mbari.org/chemsensor/pteo.htmhttp://www.mbari.org/chemsensor/pteo.htmhttp://www.agu.org/eos_elec/97025e.html7/26/2019 Controls of Trace Metals in Seawater
3/25
Trace metals in seawater can exist in a varietyof physical and chemical forms. The simplestphysical distinction is particulate versus dissolvedforms. This is somewhat of an operationaldefinition with 0.4 mm or 0.2 mm pore size filtersgenerally providing this separation. Particulate
forms include those metals adsorbed onto particlesurfaces, incorporated within particles of biogenicorigin and incorporated in the matrix of
aluminosilicate minerals or co-precipitated inother authigenic minerals. Dissolved metalsinclude various soluble complexes of the tracemetals and potential colloidal forms. The redoxchemistry of the particular metal and its environ-ment dictate the oxidation state and form of the
parent species. For trace metals, the parent speciescan be the simple mono-, di-, or trivalent cationsuch as Zn2, or for metals existing in higher
Table 1 A periodic table of the elements in seawater indicating the element number, the dominant inorganic speciespredicted to be found in the oceans, the range of concentrations observed in the open ocean, and an estimate of theelements mean concentration. The mean concentration is shown on the left-hand column in parenthesis and on the
right-hand column as a thick line. Hydrogen, noble gases, lanthanides, and elements after lead are omitted.
Introduction 25
7/26/2019 Controls of Trace Metals in Seawater
4/25
oxidation states, it can be an oxycation such as
UO22 or an oxyanion such as MoO224 . For a trace
metal such as mercury, the parent species can be
Hg2, CH3Hg, or Hg0. Cationic parent species
can form complexes with a variety of both
inorganic and organic ligands in seawater. For
example, Hg2 will complex with chloride and
exist primarily as HgCl22
4 , and UO22 willcomplex with carbonate and exist as the uranyl
carbonate complex (UO2CO3423 ).
The inorganic speciation presented in Table 1
is primarily taken from the compilations of
Turner et al. (1981) and Byrne et al. (1988).
Turner et al. (1981) used a database of stability
constants for more than 500 metal complexes to
calculate the inorganic speciation for 58 traceelements in model seawater at pH 8.2, 25 8C, and
1 atm.Byrne et al. (1988)extended this work by
considering the influence of temperature and pH
on speciation. The free hydrated divalent cationdominates the dissolved inorganic speciation of
Zn(II) and the first transition series metals
Mn(II), Co(II), and Ni(II). Strongly hydrolyzed
trace metals include Be(II), Al(III), Fe(III),
Controls of Trace Metals in Seawater26
7/26/2019 Controls of Trace Metals in Seawater
5/25
and Ga(III). Trace metals whose dissolvedspeciation is dominated by chloride complexationinclude Cd(II), Hg(II), Ag(I), and Pd(II).
The complexation or chelation of trace metalswith organic ligands will be discussed separatelyat the end of this chapter. As we will see, the
chemistry and behavior of many trace metals inthe water column is dominated by complexation,biological assimilation at uptake sites on cell
surfaces, and adsorption on surface sites ofsuspended particles. All three of these processes,
which are particularly important in surface waterswhere biological activity is most intense, are
controlled by similar coordination mechanisms(Hering and Morel, 1990).
Starting at the beginning of the periodic table,beryllium (element #4) is the first trace metal, and
is the only trace metal with an atomic number less
than 12. Beryllium has a concentration range in theoceans of 4 30 pmol kg21 (Measures andEdmond, 1982) and exists as Be(II) with the
hydrolysis species Be(OH) and BeOH02 as themajor inorganic species. Aluminum (element #13)
is the next trace metal encountered in the periodictable. Although aluminum is very abundant in theEarths crust, it is a trace metal in the openocean with concentrations ranging between
0.3 nmol kg21 and 40 nmol kg21 (Hydes, 1983;Orians and Bruland, 1986). Dissolved aluminum
exists as Al(III), with the hydrolysis species
AlOH03 and AlOH
2
4 dominating its inorganicspeciation.
All of the first-row transition metals are tracemetals with concentrations ranging by a factor
of 104from a low of 4 pmol kg21 up to36 nmol kg21. The least abundant first-row tran-
sition metal is cobalt, thought to exist as Co(II),with a concentration range of 4300 pmol kg21,while the most abundant is vanadium, existing asthe vanadate oxyanion (HVO4
22) at a concentration
between 30 nmol kg21 and 36 nmol kg21. Inter-estingly, the most abundant transition metal in the
oceans is molybdenum (element #42), existing at,105 nmol kg21 (Sohrin et al., 1989) as Mo(VI)
in the form of the molybdate oxyanion (MoO422).
The least abundant trace metal in seawater
(excluding radioactive elements) is thought to bethe platinum group metal iridium (element #77).Iridium is extremely rare in the Earths crust and itsseawater concentration is reported to be between
0.5 fmol kg21 and 1 fmol kg21 (Anbar et al.,1996). It is thought to exist as Ir(III) primarily as
the hydrolysis species IrOH03. The analyticalchallenges of determining the concentration of an
element at sub-femptomolar concentrations(,10215 M) are impressive indeed. In this case itwas determined by isotope dilution using thermal
ionization mass spectrometry after concentrationwith anion exchange resins.
6.02.1.2 Distributions
Trace metals exhibiting a relatively narrowrange of concentrations in seawater tend to existeither as oxyanions (MoO224 , WO
224 , ReO
224 ) or as
larger monovalent cations (Cs, Rb). Some ofthese metals such as molybdenum have a
significant biological requirement; however, theytend to exist in seawater at relatively highconcentrations relative to their requirement bythe biota. Those trace metals exhibiting thegreatest range in concentrations tend to beintimately involved in the major biogeochemicalcycles and are actively assimilated by phyto-plankton in surface waters. These include tracemetals such as iron, zinc, and cadmium. Tracemetals have been grouped into three principalcategories reflecting their distributions and chemi-cal behavior in seawater: conservative, nutrient,
and scavenged.
6.02.1.2.1 Conservative-type distributions
Conservative-type trace metals interact onlyweakly with particles, have oceanic residencetimes greater than 105 yr (much greater than themixing time of the oceans), and have concen-trations that maintain a relatively constant ratio tosalinity. Trace metals with conservative-type
distributions in seawater such as molybdenum,antimony, tungsten, rhenium, caesium, and rubi-dium are involved in the major biogeochemicalcycles of particle formation and destruction, butthis is negligible relative to their concentrationin seawater. Molybdenum is probably the bestexample of a conservative-type trace metal.It exists at an average concentration of 105 nmol kg21 as the oxyanion molybdate,MoO224 , and has an oceanic residence timeof ,8 10
5 yr (Emerson and Huested, 1991).It exhibits an almost uniform distribution in theoceans with only a slight depletion at the surface.
Although molybdenum is required as an essentialmetal co-factor in a number of enzymes such asnitrogenase, this requirement is small relative tothe amount of molybdenum available and does notimpact its distribution appreciably. In addition,negatively charged anions such as MoO224 have arelatively low particle affinity at the slightly basicpH of seawater. Other trace metals existing asoxyanions that exhibit relatively conservative-type behavior include tungstate (WO224 ) andperrhenate (ReO24).
There is interest in the use of some of these
oxyanions as potential paleogeochemical proxiesof the oxygen content or the redox state of deepwaters (Emerson and Huested, 1991;Morford andEmerson, 1999). For example, under anoxicconditions molybdenum can be reduced from
Introduction 27
7/26/2019 Controls of Trace Metals in Seawater
6/25
the6 oxidation state as an oxyanion to insolubleMoS2(s) or converted to particle-reactive thiomo-lybdates (Vorlick and Helz, 2002).
Other examples of conservative-type tracemetals include caesium and rubidium. Caesi-um(I) exists at an average concentration of
2.2 nmol kg
21
(Brewer e t al., 1972) as therelatively unreactive monovalent cation Cs. Itsoceanic residence time has been estimated tobe ,3 105 yr (Broecker and Peng, 1982).Rubidium exists as the monovalent cation Rb
at
a concentration of 1.4 mmol kg21 (Spenceret al.,1970) with an oceanic residence time estimated tobe 3 Myr.
6.02.1.2.2 Nutrient-type distributions
Trace metals with nutrient-type distributionsare significantly involved with the internal cyclesof biologically derived particulate material. Theirdistributions are dominated by the internal cycleof assimilation by plankton in surface waters andthe export production or transport of part of thismaterial out of the surface layer followed byoxidation and remineralization of the bulk of thismaterial in deeper waters. Consequently, theirconcentrations are lowest in surface waters wherethey are assimilated by phytoplankton and/or
adsorbed by biogenic particles, and increase inthe subsurface waters as sinking particlesundergo decomposition or dissolution. Inaddition, nutrient-type metals exhibit a relativelylow level of scavenging in the deep sea and thustheir concentrations increase along the flow pathof water in the worlds oceans as the water ages.Oceanic residence times of nutrient-type,recycled elements are intermediate (a fewthousand to one hundred thousand years).
Zinc is perhaps the most striking example of atrace metal with a nutrient-type distribution in theoceans. Bruland et al. (1978b) reported the first
accurate zinc distribution in seawater and demon-strated its strong correlation with silicic acid.Subsequently, Bruland (1980), Bruland andFranks (1983),Martinet al. (1989,1993),Morleyet al. (1993), Bruland et al. (1994), and Lohanet al. (2002) have provided other consistentprofiles. Characteristic vertical profiles of dis-solved zinc and silicic acid from high latitudes ofthe North Atlantic and North Pacific are presentedin Figure 1, demonstrating the strong interbasinfractionation that exists for this nutrient-type tracemetal. The concentration data available for deep
waters of the oceans yield a linear relationshipbetween dissolved zinc (nmol kg21) and silicicacid (mmol kg21): Zn 0:05H4SiO4 0:8.Figure 2(a) presents the distribution of silicicacid at 3,000 m depth in the worlds oceans (from
the NODC data set) and, using the aboverelationship between silicic acid and dissolvedzinc, Figure 2(b) illustrates the estimated dis-solved zinc concentration at 3,000 m depth. Silicicacid increases by a factor of 10 from a concen-tration of 20 mmol kg21 at a depth of 3,000 m at
h igh latit ud es o f th e No rth Atl ant ic to200 mmol kg21 at high latitudes of the NorthPacific. Similarly, dissolved zinc increases by afactor of 5 from 2 nmol kg21 in the young watersof the western North Atlantic to 10 nmol kg21 in
Figure 1 Vertical profiles of (a) silicic acid and(b) dissolved zinc observed at high latitudes of theNorth Atlantic ( ) (598 300 N, 208 450 W; data fromMartin et al., 1993) and the North Pacific (X) (508 N,
1458 W; data fromMartin et al., 1989).
Controls of Trace Metals in Seawater28
7/26/2019 Controls of Trace Metals in Seawater
7/25
Figure 2 Horizontal gradients of the annual mean concentration of (a) silicic acid (mmol kg21) at 3,000 m depthin the worlds oceans (source NODC), and (b) dissolved zinc (nmol kg21) at a depth of 3,000 m based upon thedeep-water relationship ([Zn] 0.05[H4SiO4] 0.8) between silicic acid and dissolved zinc from stations in theNorth Atlantic (Bruland and Franks, 1983; Martin et al., 1993), North Pacific (Bruland, 1980; Martin et al., 1989;
Brulandet al., 1994), and Southern Ocean (Martin et al., 1990).
Introduction 29
7/26/2019 Controls of Trace Metals in Seawater
8/25
the old deep waters found at high latitudes of theNorth Pacific. Similar figures can be produced forthe nutrient-type trace metal cadmium using thestrong correlation between cadmium and phos-phate or nitrate in deep waters of the oceans.Barium also exhibits a nutrient-type distribution
with concentrations ranging from 35 nmol kg
21
insurface waters to 56 nmol kg21 at 3,000 m depthin the North Atlantic, and increasing along theflow path of deep water to concentrations of150 nmol kg21 in the deep North Pacific (Chanet al., 1976,1977).
Silver provides another interesting example of anutrient-type trace metal with a strong interbasinfractionation. The first accurate data were reportedby Martin et al. (1983) for the eastern NorthPacific where they observed values of 0.4 pmol kg21 in surface waters increasing to
23 pmol kg
21
at a depth of 2,300 m. Figure 3presents dissolved silver depth profiles from highlatitudes of the North Atlantic (Rivera-Duarteet al., 1999) and high latitudes of the NorthwestPacific (Zhang et al., 2001). Vertical profiles ofdissolved silver exhibit a strong similarity tosilicic acid and the deep waters of the NorthwestPacific are enriched by slightly more than a factorof 10 for silicic acid and slightly less than a factorof 10 for silver (Figure 3). This interbasinfractionation for silver represents an even greaterfractionation than that observed for zinc. It isunclear why silver would exhibit such a nutrient-type profile, since there is no known biologicalrequirement for silver. Perhaps the nutrient metalacquisition sites are not selective enough to
discriminate against silver and it is mistakenlyassimilated into phytoplankton by the zinc (orsome other metal) uptake system. Alternatively,silver may be passively adsorbed to selectedsurface sites of biogenic particles and transportedto depth where it is remineralized as the biogenic
particulate material undergoes oxidation or dis-solution. Nutrient-type trace metals have beenused as paleoproxies for nutrient concentrations orages of deep waters in the geologic past. Gooddirect fossil records of the phosphate, nitrate, orsilicic acid content of deep waters do not exist.Thus, the fossil record of nutrient-type tracemetals such as cadmium or zinc whose concen-trations mimic and are strongly correlated with themacronutrients phosphate, nitrate, or silicic acid,can serve as indirect proxies of past nutrientconditions. The best known example involves theuse of cadmium (Boyle, 1988), where the Cd/Caratio in benthic foraminifera in sediment cores canbe used to infer the past concentrations ofcadmium in deep seawater, and by correlation,phosphate in the overlying deep waters. Thisapproach is complementary to the use of morestandard tracers such as d13C. The nutrient-typedistribution of zinc has also been used as apaleoproxy for the age and nutrient content ofdeep waters (Marchitto et al., 2002). Theseresearchers used increases in benthic foraminiferalZn/Ca and Cd/Ca ratios as evidence for a greatlyincreased presence of nutrient-rich Southern
Ocean water in the glacial North Atlantic versusrelatively nutrient-poor North Atlantic deep water.
6.02.1.2.3 Scavenged-type distributions
Trace metals with scavenged-type distributionshave strong interactions with particles and shortoceanic residence times (,100 1,000 yr), resi-dence times that are less than the ventilation ormixing time of the oceans. Their concentrationstend to be maximal near major sources such as
rivers, atmospheric dust, bottom sediments, andhydrothermal vents. Concentrations decrease withdistance from the sources and, in general, theconcentrations of the scavenged metals tend todecrease along the flow path of deep water due tocontinual particle scavenging.
Aluminum is the best illustration of a tracemetal with a scavenged-type distribution in theoceans. The major external input of aluminum isfrom the partial dissolution of atmospheric dustdelivered to the surface ocean. Vertical profiles inthe Mediterranean, the North Atlantic, and the
North Pacific are presented inFigure 4.Extremelyelevated concentrations of dissolved aluminumare observed in the Mediterranean Sea (Hydeset al., 1988), a region that receives a highatmospheric input of dust. Concentrations in
Figure 3 Vertical profiles of dissolved silver in theNorth Atlantic ( ) (composite of two stations 54.58 N,48.58 W, and 52.78 N, 358 W; data fromRivera-Duarteet al., 1999) and the western North Pacific (X) (408 N,
1458 W; data fromZhanget al., 2001).
Controls of Trace Metals in Seawater30
7/26/2019 Controls of Trace Metals in Seawater
9/25
the North Atlantic (Hydes, 1983) are maximal inthe surface waters and elevated throughout therelatively young deep waters of the NorthAtlantic. The surface waters of the North Pacific,
a region that receive less dust input than theAtlantic, exhibit aluminum concentrations that arecorrespondingly lower (Orians and Bruland, 1985,1986). The old deep waters of the North Pacifichave dissolved aluminum concentrations that are(840)-fold lower than in the North Atlantic and,100-fold lower than those observed in theMediterranean. This decrease along the deep-water flow path is consistent with an oceanicresidence time for dissolved aluminum of only,200 yr (Orians and Bruland, 1986). This markeddifference between the North Pacific and North
Atlantic deep waters is the reverse of that shownby the nutrient-type trace metals and the greatestinterbasin fractionation of any trace metal.Unfortunately, it does not appear that this markedinterbasin fractionation observed for dissolvedaluminum can be utilized as a paleogeochemicaltracer.
6.02.1.2.4 Hybrid distributions
Some trace metals, such as iron and copper, have
distributions that are strongly influenced by bothrecycling and relatively intense scavenging pro-cesses. Like nutrient-type elements, dissolved ironis observed to be depleted in remote oceanic surfacewaters such as high-nutrient, low-chlorophyll
(HNLC) regimes, and appears to be regeneratedwith depth (Figure 5)(Martin and Gordon, 1988;Johnson et al., 1997). In less productive waters ofthe oligotrophic central gyres, particularly in areas
of high dust inputs, dissolved iron can exhibitsurface-water maxima more indicative of sca-venged elements (Bruland et al., 1994; Measureset al., 1995;Johnsonet al., 1997). While nutrient-type metals, with their relatively long oceanic
Figure 4 Vertical profiles of dissolved aluminum in the Mediterranean Sea (W) (348 180 N, 208 020 W; data fromHydeset al., 1988), the North Atlantic ( ) (408 510 N, 648 100 W; data fromHydes, 1979), and the North Pacific (X)
(288 150 N, 1558 070 W; data fromOrians and Bruland, 1986).
Figure 5 Vertical profiles of dissolved iron from highlatitudes of the North Atlantic ( , S) (598 300 N, 208 450
W and 478 N, 208 W; data from Martin et al., 1993)and the North Pacific ( , W) (508 N, 1458 W and 458 N,
1428 520 W; data fromMartin et al., 1989).
Introduction 31
7/26/2019 Controls of Trace Metals in Seawater
10/25
residence times, tend to increase in concentrationin the deep waters of the ocean as the latter age, theresidence time of iron in deep waters is estimated tobe ,200 yr and does not exhibit this trend. Figure 5presents vertical profiles of dissolved iron inremote high-latitude regions of the North Atlantic
and North Pacific oceans. The concentration of ironat depths greater than 1,000 m is not significantlydifferent, which is in marked contrast to profiles ofnutrients or nutrient-type trace metals. Instead, thedeep-water concentration of dissolved iron appearsto be controlled by a balance of remineralizationfrom the rain of particulates from above andparticulate scavenging (Johnsonet al., 1997).
6.02.1.2.5 Mixed distributions
There are also trace metals that exist in more
than one chemical form with substantially differ-ing distributions. A fascinating example involvesthe trace element germanium. Germanium(element #32) is located just beneath silicon inthe periodic table. Inorganic germanic acidbehaves similar to silicic acid in seawater(Froelich and Andreae, 1981). It is assimilated ata molar ratio Ge : Si of ,0.7 1026 into thesiliceous tests of diatoms and other planktonicorganisms that make tests of biogenic opal. Whenthe tests dissolve, the germanium is released in thesame ratio and as a result there is a tight
correlation between the distribution of germanicacid (H4GeO4) and silicic acid (H4SiO4) inseawater (Froelich and Andreae, 1981). Unlikesilicon, however, germanium is also found to existas the methylated forms CH3GeOH
03and CH32
GeOH02 (Lewis et al., 1989), that appear sostable to degradation that they have been calledthe Teflon of the sea. The remarkable stabilityof these methylated species is reflected in theirconservative vertical profiles (concentrations ofdimethylgermanium are 100 pmol kg21 andmonomethylgermanium are 310330 pmol kg21)
(Figure 6). This conservative distribution is inmarked contrast to the nutrient-type distribution ofgermanic acid (2 120 pmol kg21). There are anumber of other examples of methylated com-pounds that comprise a significant fraction of tracemetals and metalloids in seawater (e.g., arsenic,selenium, mercury, tin).
6.02.2 EXTERNAL INPUTS OF TRACEMETALS TO THE OCEANS
6.02.2.1 Rivers
For the major ions in seawater, the input fromrivers is generally the dominant source. Thehistorical approach to estimate the river flux of
major elements is to measure their concentrationsin both dissolved and particulate forms in the riverand multiply these concentrations by the riverdischarge rate, thus arriving at the input of bothforms of the elements. For trace metals, however,estimating the river flux is more difficult. Thereare major problems due to under-sampling ofrepresentative river systems. Although largerivers dominate the global river input to the
oceans, such rivers are located in remote regionsand are insufficiently sampled to allow adequatefluxes of dissolved trace-metal concentrations inrelation to season and flow to be determined(Jickells, 1995). Most of the historical river datafor trace metals are not accurate, and thedevelopment of trace-metal clean techniquesalso needs to be applied to river sampling (Shillerand Boyle, 1991;Windomet al., 1991;Kimet al.,1999).
Not only are accurate data for trace metals inrivers sparse, there are complications that exist at
the riversea interface. The increase in salinityoccurring at the riversea water interface, with itsconcomitant increase in the concentrations of themajor seawater cations, can lead to flocculationand sedimentation of trace metals such as iron(Boyle et al., 1978; Sholkovitz and Copeland,1983) or to desorption from suspended riverineparticles of trace metals such as barium (Edmondet al., 1978). In organic-rich rivers a majorfraction of dissolved trace metals can exist inphysiochemical association with colloidal humicacids. Sholkovitz and Copeland (1983) used
product-mode mixing experiments on filteredScottish river water, and observed that ironremoval was almost complete due to the floccula-tion of strongly associated iron-humic acidcolloids in the presence of the increased
Figure 6 Vertical profiles of dissolved germaniumspecies from the North Pacific: inorganic germanium(X) (258 N, 1708 050 E; data fromFroelich and Andreae,1981); methyl-germanium ( , W) (data from Lewis
et al., 1985).
Controls of Trace Metals in Seawater32
7/26/2019 Controls of Trace Metals in Seawater
11/25
concentrations of Ca2
and Mg2
found inestuaries. Copper and nickel were also removedto an appreciable extent. This removal withinestuarine mixing zones is not as important inrivers with lower dissolved organic carbon con-tent. Nonconservative behavior within estuaries
makes it difficult to obtain realistic estimates ofthe actual river input of trace metals to the oceans.
6.02.2.2 Atmosphere
Using a sparse network of field measurementsof atmospheric aerosols,Duce and Tindale (1991)were able to provide some of the first globalestimates of atmospheric dust input to the oceans.More recently, satellites have provided estimatesof the global distribution of atmospheric aerosolsallowing model estimates of dust deposition(Tegen and Fung, 1995; Mahowald et al., 1999).Figure 7 presents the results of such models(Moore et al., 2002). The atmospheric input oftrace metals varies markedly spatially and term-porally, and is of a similar magnitude as theriverine input. However, aeolian fluxes impactdirectly on the oceanic euphotic zone, whilefluvial inputs are subjected to considerablemodification in estuarine and coastal waters(Jickells, 1995).
Aluminum is a major component of continentalmaterials and is present in seawater at low
concentrations in regions devoid of large dustdeposition due to the short residence time ofaluminum in surface waters (35 yr; Orians andBruland, 1986). Therefore, dissolved aluminum
concentrations are an excellent tracer of atmos-pheric inputs to the ocean. A study by Vink andMeasures (2001), using both a model of dustdeposition (Measures and Brown, 1996) anddissolved aluminum concentrations in the Atlan-tic, has demonstrated that surface-water aluminum
concentrations can be utilized to study the spatialand interannual variations of aeolian input.Concomitant with the aluminum concentrations,dissolved iron measurements were also carriedout. At the interface between the canary currentand the south equatorial current, a similar trend ofmaxima and minima in both the iron andaluminum concentrations was observed, implyinga common atmospheric source. Although ironconcentrations in the oceans are influenced byatmospheric inputs, low iron values in the surfacewaters of South Atlantic were observed,where aluminum concentrations suggestedhigh dust deposition indicating that otherfactors such as, solubility and biological removalare mor e impo rtant in contr ol ling iro ndistributions.
The maximum concentration of iron in oceanicsurface seawater is controlled by the solubility ofinorganic forms and the availability of organiccomplexing ligands to promote higher solubilities(Zhuet al., 1997;Jickells and Spokes, 2001;Vinkand Measures, 2001). Perhaps the greatest uncer-tainty in estimating the impact of this atmosphericinput on the oceans is the estimate of the
percentage of trace metals associated with dustthat is soluble upon entering the ocean. For iron,recent estimates are between 1% and 10%.Jickells and Spokes (2001) in a review suggest
Figure 7 Modeled global estimates of aeolian iron deposition from: (a) Tegen and Fung (1994, 1995) and(b)Mahowaldet al. (1999)(sourceMooreet al., 2002).
External Inputs of Trace Metals to the Oceans 33
7/26/2019 Controls of Trace Metals in Seawater
12/25
the mean solubility of iron from atmospheric dustis ,2%. New estimates indicate that ,40% of the
worlds oceans are limited by the trace metal iron.
These regions are primarily located in remoteregions where the supply of nutrients is high. Inthe remote subtropical gyres of the oceans, the
source of nutrients and iron from vertical mixingor upwelling is small and, therefore, the atmos-
pheric supply of iron is the dominant input. In theNorth Pacific, the input of Asian dust is at a
maximum in the spring and this input may
significantly increase the primary productionduring this time. The atmospheric input of iron
to the central gyres can also be important withrespect to nitrogen fixation, a process that requires
iron as a metal co-factor (see Chapter 6.05).
The variability in atmospheric input of tracemetals such as iron can be extremethe varia-
bility ranges from rapid day-to-day changes as aresult of dust storms in Asia, to seasonal changes,to decadal changes (Jickells, 1995; Zhu et al.,
1997). Less is known about the mean solubility ofother biologically important trace metals such as
zinc and manganese.
An excellent example of the atmosphericinput of a trace metal strongly influencing its
surface-water concentration is lead. Tetraethyl
lead, an anti-knock gasoline additive, was usedextensively in the 1960s and 1970s, with its usage
peaking in the late 1970s and then markedlydeclining in the 1980s and 1990s as a result of
actions taken under the Clean Air Act of 1970(Nriagu, 1989). This extensive use of leaded
gasoline resulted in a large anthropogenic leadsignal in atmospheric dust, particularly downwind
of industrialized nations. Pioneering work ofPatterson and co-workers (Schaule and Patterson,
1981, 1983; Flegal and Patterson, 1983) on thedistribution of lead in the major central gyres of
the Atlantic and Pacific oceans graphically point
out how this anthropogenic atmospheric lead inputmarkedly perturbed the distribution of lead in the
different ocean basins at the peak of its input in
the late 1970s and early 1980s (Figure 8). Due tothe prevailing wind pattern, the North Atlantic
received the brunt of the US lead input to theoceans and contained markedly elevated concen-trations of lead in the upper 1,000 m. In contrast,
the remote South Pacific central gyre hadsurface lead concentrations over an order-of-
magnitude lower. Subsequent to this work,
Boyle and co-workers (Wu and Boyle, 1997)presented results from a 16-year time series of
lead concentrations in the western North Atlantic
Figure 8 Vertical profiles of dissolved lead in the central North Atlantic (348 150 N, 668 170 W; data fromSchauleand Patterson, 1983), the central North Pacific (328 410 N, 1458W; data fromSchaule and Patterson, 1981), and thecentral South Pacific (208S, 1608W; source Flegal and Patterson, 1983). Estimates of the atmospheric input atthe time of sampling and in ancient times prior to the large anthropogenic lead input are also shown (Flegal and
Patterson, 1983).
Controls of Trace Metals in Seawater34
7/26/2019 Controls of Trace Metals in Seawater
13/25
showing that lead concentrations decreased mark-edly during the decade of the 1980s and that thisdecrease can be attributed to the phasing out ofleaded gasoline in the US.
6.02.2.3 Hydrothermal
It has been estimated, using arguments basedupon the 3He anomaly and heat flux, that the entireocean mixes through hydrothermal vent systems,undergoing high-temperature interaction withfresh oceanic basalt every 8 10 Myr, leading tothe production of high temperature (,350 8C),acidic (pH,3.5), reducing, sulfide- and metal-richhydrothermal fluids (Edmond et al., 1979; vonDammet al., 1985). This ridge crest hydrothermalactivity has proven to be the major oceanic sink
for the major ions, magnesium, and sulfate, and tobe a major source for trace metals such as iron andmanganese (see Chapter 6.08). Iron and manga-nese concentrations in the 350 8C vent waters canbe a million-fold higher than in the surroundingseawater. Much of the iron, however, is rapidlyprecipitated; initially either as iron sulfides andthen oxidized to iron oxyhydroxide precipitates,or rapidly oxidized from the soluble Fe(II) form toinsoluble Fe(III), and deposited as sediments overthe mid-ocean ridges. The dissolved manganesecan advect further away from the vent source priorto its microbially mediated oxidation and precipi-tation (Cowen et al., 1998). The hydrothermalinput of iron and manganese, however, isessentially all scavenged and removed in thedeep sea prior to having a chance to mix back intothe surface waters.
6.02.3 REMOVAL PROCESSES
6.02.3.1 Active Biological Uptake inthe Surface Waters
6.02.3.1.1 Lessons from laboratory studiesA great deal of insight has been gained from
well-defined laboratory studies of the effects oftrace metals on phytoplankton growth rates,which in turn has provided knowledge on thecontrol of trace metals in the upper watercolumn by biological processes (see Chapter6.05). Culture media have been designed inwhich the concentration and speciation of tracemetals are controlled by the use of strongchelating ligands such as EDTA (Morel et al.,1979; Price et al., 1988/89; Sunda, 1988/89).
The rate of uptake of a trace metal (M) isusually proportional to its free-metal concen-tration [M2] or its unchelated concentration,[M0] (defined as the sum of the kinetically labileinorganic species of M). The [M0] and Mn are
related by their inorganic side reaction coeffi-cient a M0=Mn. The EDTA chelatedmetal can act as a metal ion buffer thatmaintains [Mn] and [M0] at constant values inthe media. By judiciously varying the EDTAand total metal concentrations, the experimental
[M
2
] and [M
0
] can be controlled over a widerange.Diatoms are particularly important in biogeo-
chemical cycles because of their role asmajor players in new and export production(Smetacek, 1999). Along with the assimilationand export of carbon, nitrogen, phosphorus, andsilicon, diatoms also play an important role inthe export of bioactive trace metals fromsurface waters. Sunda and Huntsman (1995a)have carried out extensive laboratory studieson iron uptake by coa stal and oceanicdiatoms. Figure 9 presents the cellular Fe/C(mmol mol21) ratio in diatoms as a functionof the estimated [Fe0]. These diatoms exhibit anincreasing cellular Fe/C ratio as the [Fe 0] in themedia increases. At elevated [Fe0] in the media,the diatoms exhibit luxury uptake with the twocoastal species reaching cellular Fe/C valuesthat were 20 30 times higher than thoseestimated to be required for maximum growth(Sunda and Huntsman, 1995a). It has beensuggested that this high uptake rate and storagecapacity in the diatoms at elevated [Fe0] allowsthese species to accumulate excess iron during
periods of high iron availability, that can thenbe passed on to their progeny and utilized whenthe dissolved iron may later be drawn down toconcentrations limiting growth rates (Sunda andHuntsman, 1995a). This can be an effectivestrategy for diatom blooms in coastal upwellingregimes.
Figure 9 The relationship between the intracellularFe/C ratio as a function of the inorganic iron concen-tration, [Fe0], for three diatoms species, Thalassiosiraoceanica (W) an oceanic species ( ), and two coastalspecies Thalassiosira weissflogii, and Thalassiosirapseudonana( ) (sourceSunda and Huntsman, 1995a).
Removal Processes 35
7/26/2019 Controls of Trace Metals in Seawater
14/25
6.02.3.1.2 Non-Redfieldian assimilation
Figure 9is important with respect to removal ofiron from the surface ocean. Nitrogen andphosphorus are assimilated and removed fromthe surface ocean at a ratio within about a factor of2 of the Redfield ratio (Falkowski, 2000; Karl,
2002). Differences are observed at time-seriesstations in the North Pacific and North Atlantic(HOTS and BATS) that vary depending upon thesource of nitrogen; however, they generally varyby less than a factor of 2. In contrast, there is noconstant Redfield ratio of Fe/C; the Fe/C ratio indiatoms can vary by a factor of 100 dependingupon the availability of iron and whether they areoceanic or coastal species. Interestingly, oceanicspecies have evolved to get by at lower ironconcentrations than coastal species (Sunda andHuntsman, 1995a). Diatoms are also the type of
phytoplankton most responsive to episodicchanges in iron and nutrient inputs as demon-strated in each of the novel mesoscale iron-enrichment experiments (Martin et al., 1994;Coale et al., 1996; Boyd et al., 2000). Theirability to maximize iron uptake and to store luxuryuptake, can be a major factor in controlling ironconcentrations in surface seawater.Bruland et al.(2001) examined this process in coastal diatomsunder upwelling conditions and argued that ironis removed preferentially to nitrate in coastalupwelling regimes and tends to drive the system
towards iron limitation. The amount of uptake ofsilicic acid by diatoms relative to the assimilationof nitrate and phosphate has been shown to dependon iron availability (Hutchins and Bruland, 1998;Takeda, 1998). Under iron-replete conditions, thesilicon and nitrogen uptake is roughly equal, whileunder low-iron conditions diatoms exhibit Si/Nuptake ratios of ,3. Under low iron concen-trations the suppression of iron-containingenzymes such as nitrate or nitrite reductase hasbeen confirmed for the chain-forming diatoms inthe laboratory (de Baar et al., 2000), resulting in
reduced uptake of nitrogen relative to phosphorusand N/P ratios of 46 as compared to ,1214 atadequate iron supply.
Studies have also been carried out on otheressential trace metals such as zinc (Sunda andHuntsman, 1992, 1995b), and similar relationshipsbetween Zn/C and [Zn2] have been observed. Instudies with the diatom Thalassiosira oceanica,the Zn/C (mmol mol21) ratio varied from 0.2 at[Zn2] of 10213 M to ,40 at [Zn2] of 1029 M(Figure 10). As with iron, there is more than 100-fold variation in the Zn/C ratio in diatoms as a
function of [Zn2
]. The uptake of cadmium bydiatoms is somewhat unique and is dependent notonly upon the [Cd2], but also on [Zn2] (Sundaand Huntsman, 1998). At elevated [Zn
2], cad-
mium assimilation and Cd/C ratios in diatoms
are low. Under zinc depletion, however, theuptake of cadmium and Cd/C ratios markedlyincrease. Morel and co-workers (Price and Morel,1990; Lee and Morel, 1995) have shown thatcadmium can functionally replace zinc and thatdiatoms have a carbonic anhydrase enzymeutilizing cadmium instead of zinc as a metal co-factor. Thus, under low-zinc conditions, theassimilation of cadmium markedly increases and
results in the depletion of cadmium in surfacewater. This is an example of where the cellularuptake of one metal, cadmium, responds to acomplex matrix of other metals such as zinc.Cullen et al. (2003) provided evidence that theeffect of iron limitation on resident diatoms in theSouthern Ocean is to decrease growth rates,leading to elevated cellular cadmium content. Inthis case the assimilation of carbon, nitrogen, andphosphorus was markedly decreased as a result ofiron limitation, but cadmium assimilation con-tinued, leading to enhanced cellular Cd/P ratios.
For copper, the cellular metal/carbon ratios varyin more of a sigmoidal fashion, with what appearsto be a region of varying [Cu2] with the Cu/C ratiosomewhat constant and regulated (Sunda andHuntsman, 1995c). The Cu/C ratio in diatoms,however, still varies by roughly a factor of 100over a wide range of concentrations of [Cu2] andhas been implicated as an important factor incontrolling the distribution of copper in the oceans(Sunda and Huntsman, 1995c).
Trace metals for which active biologicalassimilation may be an important factor in
controlling surface-water concentrations and dis-tributions include the first-row transition metalsiron, zinc, manganese, copper, nickel, and cobalt,along with cadmium. Bruland et al. (1991)compiled data on the composition of plankton in
Figure 10 The relationship between the intracellularZn/C ratio as a function of the free ionic zincconcentration, [Zn2], for an oceanic diatom species
Thalassiosira oceanica (source Sunda and Huntsman,1992).
Controls of Trace Metals in Seawater36
7/26/2019 Controls of Trace Metals in Seawater
15/25
the Pacific taken from two sources (Martin andKnauer, 1973; Martin et al., 1976) and excludedonly data with an aluminum content .4 mmol g21
(.100 mg g21) dry weight in an attempt tominimize the contribution from aluminosilicatesminerals. Metal/carbon ratios for iron were
close to 50mmol mol
21
, Zn/C ranged from8 mmol mol21 to 17 mmol mol21, while Mn/Cratios averaged 3.6 mmol mol21. These M/Cvalues for iron and zinc in diatom samples fromthe field off central California lie within the rangeexpected for these region based upon laboratorystudies. Biological assimilation of these metalsat such M/C ratios with subsequent export of afraction of this material to the deeper watercolumn is particularly important in influencingthe oceanic distributions of iron and zinc. Theseare two of the trace elements that exhibit markedsurface depletion due to their involvement in thisbiological cycle.
6.02.3.2 Passive Scavenging
6.02.3.2.1 Adsorption/desorption processes
In addition to the role of active assimilation ofrequired trace metals by phytoplankton, there isalso passive scavenging of trace metals onto thewide variety of relatively high affinity surfacesites on both living and dead particulate material
existing in the surface waters. The combinedprocess of surface adsorption, followed by particlesettling, is termed scavenging (Goldberg, 1954;Turekian, 1977). Such binding is effectivelypassive, in contrast with the active uptake ofessential trace metals. Examples of trace metalsimplicated in such scavenging from surface watersinclude lead, aluminum, gallium, and the radio-active isotopes of thorium.
6.02.3.2.2 Lessons from radionuclides
Thorium, with its four different radioactiveisotopes of greatly differing half-lives, is an excell-ent tracer which provides insight into the rates andthe process of scavenging. Thorium isotopes andtheir half-lives are: 232Th, t1=2 1:4 10
10 yr(essentially 232Th can be considered a stableisotope); 230Th, t1=2 7:54 10
4 yr with 234Uas a parent; 228Th, t1=2 1:91 yr with
228Ra asa parent; and 234Th, t1=2 24:1 d with
238U as aparent. Two conceptual models that have beenused to model particle reactive thorium data arepresented inFigure 11.Figure 11(a)incorporates
reversible exchange and remineralization of theparticles, whileFigure 11(b)uses a net scavengingrate constant to examine the net scavenging ofthorium.Figure 12presents the distribution of thenet scavenging rate constant of 234Th in surface
waters of the Pacific. The scavenging intensity of234Th varies dramatically between oligotrophicgyres of the North Pacific and South Pacific andproductive regions such as the subarctic Pacific,the upwelling regime off central California and theequatorial Pacific. Values of the net scavenging
rate constant in the surface mixed layer of theoligotrophic gyres were between 0.009 d21 and0.003 d21 (yielding a mean life of dissolved 234Thwith respect to particle scavenging of 100 300 d).In contrast, in productive regions such as thesubarctic Alaskan Gyre and intense coastalupwelling regimes off central California, the netscavenging rate constants were 0.11 0.10 d21
(mean life of 9 10 d) and 0.175 0.125 d21
(mean life of 6 8 d). The net scavenging rateconstant has been argued to be proportional to newor export production (Coale and Bruland, 1987)and has been observed to be most intense inregimes of high particle production with substan-tial export of this material from the surface waters,and least intense in regions of lower primaryproduction where regenerated productiondominates.
Aluminum and gallium are two trace metalswith hydrolysis chemistry similar to that ofthorium and the affinity of a metal cation toform these hydrolysis species has been suggestedas an important parameter in models for adsorp-tion and scavenging processes. Orians andBruland (1986, 1988) have presented data foraluminum and gallium (Figure 13) at some of thesame stations where net thorium scavenging rateswere determined (Figure 12). The surface-waterconcentrations of aluminum and gallium shouldbe a function of the magnitude of sources relativeto their scavenging removal rate. Both of thesemetals exhibit higher concentrations in the sur-face waters of the oligotrophic central gyres, withmuch lower values observed towards the northand eastern boundaries where rates of exportproduction and 234Th scavenging are higher.Undoubtedly, since Fe(III) is another trace
metal with similar hydrolysis chemistry, suchscavenging would also take place for it. It is,however, difficult to separate such passivescavenging from active assimilation that alsooccurs with iron.
Thorium isotopes have also been used to gaininsight into scavenging in deep waters. Using acombination of thorium isotopes, investigatorshave been able to determine the dynamics ofthorium scavenging (Bacon and Anderson, 1982;Nozakiet al., 1987;Clegget al., 1991;Murname,1994; Roy-Barman et al., 1996). These studies
have provided evidence for a dynamic systemwhereby thorium isotopes appear to be reversiblyscavenged from the deep sea onto fine particles,these particles are packaged into larger particlesthat sink and then either disaggregate or are
Removal Processes 37
7/26/2019 Controls of Trace Metals in Seawater
16/25
remineralized, and the thorium is either desorbed
or released back into solution. In this manner, anindividual thorium isotope might spend a total
of 2050 yr in the deep sea before burial in
the sediments (see Chapter 6.09) and during this
time undergo numerous reversible adsorption/desorption exchanges with particles that are conti-
nually aggregating into larger particles, sinkingand disaggregating or being remineralized.
Estimates of residence times of thorium in
various forms within the deep sea suggest that
an isotope might spend close to a year in thedissolved form prior to adsorption onto the
surface of a small particle where it may reside
for another four months prior to aggregating to a
larger, more rapidly sinking particle. It mightspend a few days sinking a few hundred meters
prior to disaggregation and then a few monthsresiding on the small particle prior to desorption
or remineralization back into the dissolved phase
Radioactive
decay
Small
particulate
Dissolved
234Th
Large
particulate
Export flux(a)
234Th
234Th
APThA
DTh
PTh
k*
2APThk
*
1ADTh
AUTh
ADTh Th APTh Th
Radioactive
decay
Small
particulate
Dissolved
234Th 234Th
Large
particulate
Export flux
234Th
APTh
ADTh
PTh
k1ADTh
k1APTh
k2 APTh
AUTh
ADTh Th A
PTh Th
k2ADPTh
ADPTh
APTh
(b)
Figure 11 Conceptual models of thorium scavenging (Coale and Bruland, 1985;Bruland and Coale, 1986; Clegget al., 1991). (a) The surface water 234Th net scavenging model. This model incorporates two different size classes ofparticles, small suspended particles and large sinking particles with the various sources and sinks for the activity (A)of 234Th depicted.lThis the decay constant of
234Th,kp1 is the net rate transfer of 234Th from dissolved to suspended
particles and kp
2 is the net rate of transfer of 234
Th from small suspended particles to large sinking particles. (b) Areversible scavenging model including desorption, particle disaggregation and remineralization for the deep sea.230Th and 234Th can be both modeled to yield estimates of rate constants. Rate constants: k1, adsorption onto smallsuspended particles andk21;desorption from small suspended particles; k2;aggregation or packaging rate of smallsuspended particles into large sinking particles and k22 is the disaggregation of large sinking particles and g
remineralization of carrier phases including respiration of organic matter or dissolution.
Controls of Trace Metals in Seawater38
7/26/2019 Controls of Trace Metals in Seawater
17/25
(Murname, 1994). Similar dynamic reversibleexchange processes in the deep sea may beoccurring for other trace metals such as iron,aluminum, gallium, and titanium.
6.02.4 INTERNAL RECYCLING
6.02.4.1 Recycling within the Water Column
Internal recycling is particularly relevant for
nutrient-type trace metals that, like the macro-nutrients nitrate, phosphate, and silicic acid,undergo multiple cycles of assimilation intobiogenic particulate material within surfacewaters and release or remineralization at depth.In this manner an element can undergo manyinternal cycles within the ocean prior to ultimateburial in sediments. For example, imagine anindividual zinc ion delivered to the surface watersby upwelling or vertical mixing. It can beassimilated into a phytoplankton cell, whereit might reside a day or two prior to being
grazed. In the open ocean the bulk of thisphytoplankton zinc and other nutrients will beremineralized in the surface waters and availableto undergo perhaps 5 or 10 such assimilation/remineralization cycles within the surface layer
(Hutchins et al., 1993; Hutchins and Bruland,1994,1995) prior to removal as an export flux inthe form of a fecal pellet excreted from azooplankter or a larger aggregate of particles.
This particulate zinc can be remineralized in thedeep sea as the fecal pellet serves as a source offood and nutrition for heterotrophic organisms inthe deep sea. Remineralization of particulate zinccan occur at depth by the degradation of organicmatter or the dissolution of the inorganic carrierphases (metal oxides, opal, calcite). Then perhapsa few hundred to a thousand years later this samezinc ion can be mixed back up into the surfacewaters. Numerous such cycles can occur during its(1050) 10
4 yr odyssey in the oceans. Oncefinally removed via burial in the sediments, the
zinc atom may have to wait ,100 Myr to betectonically uplifted onto a continent and thenanother 100 Myr before it is exposed to con-tinental weathering and makes its way via rivers ordust input back to the ocean.
60N
50
180W30S
20
10
0
10
20 Hawaii
PacificL
atitude
30
40
160 140
Longitude
Scavenging rate constant
k*1(d1)
0 .01 .02 .03 .04 .05
120 100 80
Ocean
Tahiti
Santa Cruz
Figure 12 The net scavenging rate constant,kp1 (d21),
of 234Th from the surface mixed layer of the PacificOcean (sourcesBruland and Coale, 1986;Bruland and
Beals, unpublished).
180W 140 100120160
20
2 nm
Dissolved Al
40
60N
180W 160 140 120 100
Ocean
Hawaii
Pacific
Santa Cruz
Dissolved Ga
5 pm
20
40
60N
Hawaii
Pacific
Ocean
Santa Cruz
(a)
(b)
Figure 13 Concentrations of (a) dissolved gallium(data fromOrians and Bruland, 1988) and (b) dissolvedaluminum (data fromOrians and Bruland, 1986) in thesurface waters of the eastern North Pacific at stationsfor which 234Th net scavenging rate data exist
(seeFigure 12).
Internal Recycling 39
7/26/2019 Controls of Trace Metals in Seawater
18/25
6.02.4.2 Benthic Inputs
Upon settling to the surface sediments, tracemetals can be recycled back into the dissolvedphase and act as a source to the deep ocean. Pore-water concentrations of trace metals can besignificantly higher than that observed in the
overlying water column. Elevated pore-waterconcentrations suggested a potential for benthicfluxes of dissolved metals out of the sediment,which has been verified by direct measurements(Elderfield et al., 1981;Westerlund et al., 1986).
Trace metals in marine sediments are frequentlyassociated with iron and manganese hydroxidesand changes in the sediment redox chemistry nearthe sediment water interface can lead to alternat-ing periods of reductive dissolution and oxidationof these phases. Depth profiles in the shelf watersoff the coast of the Falkland Islands and SW
Africa indicate a significant supply of dissolvediron (2 38 nmol kg21) to the overlying watercolumn through reductive benthic processes(Bowie et al., 2002). Iron concentrations justabove the sediment interface were 12-fold higherthan those observed at the surface. Benthic inputsof iron, cobalt, and manganese have beenobserved in the highly productive coastal watersof the North Sea (Tappin et al., 1995). Trace-metal recycling through benthic inputs has thepotential to supply trace metals back into surfacewaters. The dominant source of iron to waters off
the coast of California is through sedimentresuspension followed by upwelling (Johnsonet al., 1999;Bruland et al., 2001).
6.02.5 COMPLEXATION WITH ORGANICLIGANDS
Studies using electrochemical techniques havedemonstrated that in surface seawater a majorfraction of many trace metals, particularly thebioactive trace metals such as iron, zinc, copper,
cobalt, and cadmium, are present as chelates withstrong metal-binding organic ligands. Theseelectrochemical methods employ sensitive strip-ping voltammetric analysis. Anodic strippingvoltammetry (ASV), using a thin mercury film(TMF), rotating glassy carbon disk electrode(RGCDE) in the differential pulse (DP) mode,has been used in open ocean studies of copper(Coale and Bruland, 1988, 1990), zinc (Bruland,1989), cadmium (Bruland, 1992), and lead(Capodaglio et al., 1990). This method directlymeasures the kinetically labile [M0] (inorganic
complexes and free metal), while the metalchelated with strong organic ligands is kineticallyinert with respect to being detected by thismethod. ASV involves a deposition or concen-tration step that can be ,10 min or 20 min in
duration where M0 is continually reduced andconcentrated into the mercury amalgam. The DPstripping step involves ramping the potential in apositive direction and the measurement of theanodic stripping current as the metals are oxidizedback into solution. Titrations of a sample with the
metal of interest and determining [M
0
] at eachtitration point allow both the concentrations ofthe M-binding ligands and binding strengths(conditional stability constants) of the metalligand complexes to be determined. The numberof trace metals that can be determined by ASV,however, is limited to those that can be reduced(and reoxidized) at appropriate potentials and thatare soluble in a mercury amalgam.
A second powerful voltammetric approach thatis amenable to a far broader group of trace metalsis adsorptive cathodic stripping voltammetry(AdCSV). The application of AdCSV to specia-tion studies involves the addition of a well-characterized added ligand (AL) that sets up acompetitive equilibrium with the natural ligandsfor the metal of interest (van den Berg, 1988).Most methods involve formation of a neutralbiscomplex with the AL, M(AL)2
0 (Bruland et al.,2000). The ALs generally form planer biscom-plexes with the metal of interest that have a strongtendency to adsorb on the surface of a hangingmercury drop electrode, whereas the naturalmetalligand complexes do not. After an appro-priate adsorption or accumulation time period, thepotential is ramped in a negative direction (ineither linear, DP or Osteryoung square wave (SW)mode) and the cathodic stripping current ismeasured as the metal (and sometimes also theAL) is reduced at the electrode surface. The use ofthese AdCSV methods does not require the metalto be soluble in a mercury amalgam and thus canbe used for a wide variety of trace metals. Metaltitrations and the determination of the M(AL)2
0
concentration at each titration point allows thedetermination of natural, strong metal-bindingligand concentrations and their conditional stab-
ility constants. The use of AdCSV methods hasbeen applied to studies of the speciation of iron(van den Berg, 1995; Rue and Bruland, 1995;Bruland and Rue, 2001), copper (van den Berg,1984; Moffett et al., 1990; Donat and van denBerg, 1992), zinc (van den Berg, 1985;Donat andBruland, 1990;Ellwood and van den Berg, 2000;Lohanet al., 2002), and cobalt (Saito and Moffett,2001;Ellwood and van den Berg, 2001). It shouldbe noted that detection windows of voltam-metric techniques examine stability constants ofmetal-binding ligands specified by the technique,
and that natural metal-binding organic ligands arenot a single entity but rather classes of ligandswith average values assigned.
Results of such voltammetric studies havedemonstrated that complexation of trace metals
Controls of Trace Metals in Seawater40
7/26/2019 Controls of Trace Metals in Seawater
19/25
with relatively specific and strong metal-bindingorganic ligands is important in oceanic surfacewaters. Greater than 99% of Fe(III) in surfacewaters is complexed with strong Fe(III)-bindingorganic ligands existing at sub-nanomolar con-centrations in slight excess of the dissolved iron
(Rue and Bruland, 1995;van den Berg, 1995;Wuand Luther, 1995;Powell and Donat, 2001;Boyeet al., 2003). Greater than 99% of copper exists asorganic complexes (Coale and Bruland, 1988;Moffett, 1990). Approximately 98% of dissolvedzinc in surface waters is complexed with organicligands (Bruland, 1989;Donat and Bruland, 1990;Ellwood and van den Berg, 2000; Lohan et al.,2002). Greater than 90% of cobalt exists com-plexed to strong cobalt-binding organic ligands(Saito and Moffett, 2001; Ellwood and van denBerg, 2001) and ,80% of cadmium in surfacewaters is complexed with organic ligands(Bruland, 1992).
We know little about the chemical structure orarchitecture of the organic ligands involved inbinding metals in seawater. There have beenrecent advances, however, into the structure andfunction of marine siderophores. The conditionalstability constants of the marine siderophores sofar examined are similar to the stability constantsof the natural Fe(III)-binding ligands found inseawater (Barbeau et al., 2001, 2003) and side-rophores appear to constitute a significant fractionof the natural Fe(III)-binding organic ligands inseawater (Macrelliset al., 2001). Siderophores aredefined as low molecular weight organic chelatorswith a very high and specific affinity for Fe(III),the biosynthesis of which is regulated by ironlevels, and whose function is to mediate ironuptake by microbial cells.
The selectivity of siderophores for Fe(III) isachieved through optimal selection of metal-binding groups, the number of binding units, andtheir stereochemical arrangement (Boukhalfa andCrumbliss, 2002). Most siderophores are hexa-dentate and incorporate hydroxamate, catecholate,
and/or a-hydroxy carboxylate binding subunitsarranged in different architectures. Barbeau et al.(2003)have presented a summary of many of themarine siderophores groups produced by bothheterotrophic and photosynthetic marine bacteriaand characterized the photochemical reactivity ofthe different Fe(III)-binding functional groups.Hydroxamate groups are photochemically resist-ant regardless of Fe(III) complexation. Catecho-lates are susceptible to photo-oxidation in theuncomplexed form, but stabilized against photo-oxidation when ferrated. a-Hydroxy carboxylate
groups are stable as the uncomplexed acid, butwhen coordinated to Fe(III) these moietiesundergo light-induced ligand oxidation andreduction of Fe(III) to Fe(II). These photochemi-cal properties appear to determine the reactivity
and fate of Fe(III)-binding siderophores in oceansurface waters (Barbeau et al., 2003).
Other possible candidates of natural Fe(III)-binding ligands in seawater are porphyrin-type ligands released as degradation productsof cytochrome-containing systems (Rue and
Bruland, 1997). As yet, little is known about thispossibility. In addition, little is known about thestructure, functional groups, or architecture ofother metal-binding organic ligands in seawater.What is known is that there appear to be smallconcentrations of strong and relatively specificmetal-binding organic ligands that play an import-ant role in the chemical speciation of quite a fewof the bioactive trace metals in the sea.
6.02.5.1 Copper
Copper provides an interesting example of atrace metal that is an essential, required element,but that can be toxic at relatively low concen-trations. It can be considered the Goldilocksmetal. Surface-water concentrations in the openocean are ,1 nmol kg21. Without organic com-plexation, the free copper concentration would beapproximately a factor of 20 lower than thetotal dissolved concentration, with Cu(CO3)
0
predicted to be the dominant species. This wouldyield [Cu2] concentrations ,0.5 10210 M or
10
210.3
M. This concentration would be toxic tomany oceanic phytoplankton, particularly theprokaryotic photosynthetic bacteria such as syne-chococcus (Brand et al., 1986). Figure 14(b)presents the actual vertical distribution of [Cu2]in the upper 500 m of the Northeast Pacific. Coaleand Bruland (1988,1990)observed a slight excessof a strong copper-binding class of organicligands, called L1. This class of strong copper-binding ligands was found to occur in surfacewaters at concentrations of 12 nmol kg21 and itspresence led to greater than 99.8% of the copperbeing chelated to this class of ligands. As a result,
[Cu2] was reduced by close to a factor of 1,000and exists at concentrations ,10213 M, which is aconcentration not toxic to phytoplankton.Moffettand Brand (1996)have shown that cyanobacteriawhen stressed with slightly elevated [Cu2] canproduce a ligand with a similar conditionalstability constant. It appears that, somehow, thephytoplankton of the open ocean, particularlythe prokaryotic phytoplankton, are controlling theexternal concentration of free copper by producinga strong copper-binding ligand that reduces the[Cu2] to levels that are no longer toxic. As a
result, the [Cu2] in surface waters is buffered by
the L1 class of ligands at a concentration that isnot too little and not too much, but justright; thus, the Goldilocks example. Thisbuffering of [Cu2] also influences its distribution
Complexation with Organic Ligands 41
7/26/2019 Controls of Trace Metals in Seawater
20/25
and, because this chelated form is unavailablebiologically, the dissolved copper is not depletedto a great degree in oceanic surface waters.
6.02.5.2 Iron
Dissolved iron provides an example of abioactive, essential trace metal that is depleted inoceanic surface waters to such an extent that ithas been estimated to be the limiting nutrient in,40% of the worlds oceans (Mooreet al., 2002).
In particular, iron has been shown to bethe limiting nutrient in the HNLC regions of theSouthern Ocean, the equatorial Pacific, and thesubarctic Pacific (Martin et al., 1994). Not onlydoes iron exist at extremely low concentrations in
surface waters, its chemical speciation is domi-nated by complexation with Fe(III)-bindingorganic ligands. The initial evidence thatdissolved iron is strongly chelated by naturalFe(III)-binding organic ligands in seawater wasprovided by van den Berg (1995), Rue and
Bruland (1995), and Wu and Luther (1995).Results from surface waters of the equatorial andNorth Pacific oceans (Rue and Bruland, 1995,1997) indicate that two classes of Fe(III)-binding ligands exist. There appears to be0.30.4 nmol kg21 of particularly strong ligands(an L1 class) with a conditional stability constant(with respect to Fe(III)0) of 1012.5 1013 M21. Inaddition, there is a more variable class of weakerligands (an L2class) existing at concentrations of0.21.5 nmol kg21 with a conditional stabilityconstant of 10
11.5 10
11.8M21
. This is a casewhere it appears that microorganisms are essen-tially carrying out chemical warfare in theirattempts to acquire this metal.
Marine siderophores produced by both hetero-trophic and photosynthetic bacteria that havebeen so far examined have conditional stabilityconstants consistent with the classes of naturalorganic ligands observed in seawater (Barbeauet al., 2001,2003). For marine siderophores to becompetitive for binding Fe(III) in surface sea-water, they would have to be this strong. Barbeauet al. (2001) have also shown that photo-degradation products of the original siderophoreshave weaker conditional stability constants simi-lar to the L2class of ligands observed byRue andBruland (1995, 1997). It is known that manymicroorganisms are able to not only utilize theirown siderophores, but in addition, can assimilatenumerous other bacterias siderophores (Wilhelmand Trick, 1994). Eukaryotic diatoms are thoughtto not have the receptor sites to assimilate Fe(III)siderophores directly. There is evidence, however,that diatoms can utilize cell-surface reductasesystems to reduce the Fe(III) bound to thesiderophore, and Fe(II) can dissociate and become
available either as Fe(II) or can be reoxidized toFe(III) and become available for assimilation(Maldonado and Price, 1999). Barbeau et al.(2001) have shown that iron associated with themore weakly held photoproduct of marine side-rophores is more readily available than theoriginal Fe(III) siderophore. Hutchins et al.(1999) have also addressed the issue of theavailability of Fe(III) bound to various side-rophores and other ligands and have observeddifferences in availability. Fe(III) porphyrincomplexes seem to be more readily available
to the eukaryotic diatoms, while Fe(III)siderophores are more readily available by theprokaryotic phytoplankton community.
In the iron-limited regions of the ocean, iron iscycled through the planktonic community so
Figure 14 Concentration profiles of (a) total dissolvedcopper, [CuT], and strong copper-binding organicligands, [L1], and (b) free [Cu
2] and [CuT] in theupper 500 m of the North Pacific (source Coale and
Bruland, 1988).
Controls of Trace Metals in Seawater42
7/26/2019 Controls of Trace Metals in Seawater
21/25
rapidly that an individual iron atom may be in adifferent form each day. It may exist dissolved asan Fe(III) siderophore one day, be part of theintracellular photosynthetic machinery of aphotosynthetic bacteria the next day, be regener-ated as an ironporphyrin cell lysis product the
next day, and then rapidly reassimilated by adiatom. In this case, the iron that is biologicallyavailable is changing on a day-to-day basis. Ofinterest to this chapter is the eventual export ofiron from the surface water as the diatom is grazedby a copepod and a part of the undigested residueof the diatom and its associated iron is removedfrom the surface layer as a fecal pellet to betransported into the deep sea where it can beremineralized and either scavenged or eventuallymixed back into the surface layer once again.Interestingly, it has been argued that Fe(III)-binding organic ligands observed in the deep sea(Rue and Bruland, 1995) play an important role inallowing the dissolved iron in the deep sea to existat concentrations on the order of a nmol kg21
(Johnson et al., 1997). This is a higher concen-tration than would be expected from estimates ofinorganic solubility (Liu and Millero, 2002).
6.02.5.3 Zinc
Field studies have revealed that there arenanomolar concentrations of strong zinc-binding
organic ligands in surface waters (Bruland, 1989;Donat and Bruland, 1990; Ellwood and van denBerg, 2000; Lohan et al., 2002) that play animportant role in chelating zinc in surface waters.Figures 15(a) and (b)presents data from the NorthPacific (Bruland, 1989) and it is apparent that,98% of the dissolved zinc in surface watersexists in a chelated form with organic ligands.Although we have insight into the concentrationsand conditional stability constants of theseligands, we know little or nothing about theirfunctional character or molecular architecture.There are no immediate solubility constraints on
the zinc concentration and it does not undergoredox cycling so that the benefits conferred bycomplexation upon iron availability do not applyto zinc. Unlike copper, zinc is not toxic tophytoplankton at concentrations observed in theopen ocean. There appears to be no immediateadvantage to the phytoplankton community inreducing [Zn2] through organic complexation.
One interesting idea is that this might be anexample of smart banking by the planktoncommunity. By having the bulk of the zincchelated and presumably less bioavailable, the
removal rate of zinc by the biota will be less (seeFigure 10). As discussed previously in Section6.02.3.2.1,the Zn/C ratio in diatoms is dependentupon the [Zn2]. Therefore, by markedly decreas-ing [Zn2] with the production of a zinc-binding
organic ligand, the removal rate of zinc will be
decreased and its residence time in the surfacewater increased, e.g., the capital is preserved.Conversely, some of the zinc-binding ligands maybe a zinc siderophore-type compound wherebycertain microorganisms are attempting to gain anadvantage for this potentially biolimiting element.
Voltammetric techniques have provided apowerful tool for measuring the speciation oftrace metals at low concentrations. Most of thebioactive trace metals have now been analyzedusing these techniques. Knowledge of the inor-ganic, organic, and free metal forms in the
dissolved phase is critical as the different formsare involved in very different biological andgeochemical interactions and hence the cyclingof trace metals within the ocean. Studies oftrace-metal speciation have influenced ideas
Figure 15 Concentration profiles of (a) total dissolvedzinc, [ZnT], and strong zinc-binding organic ligands, [L]and (b) free [Zn2] and [ZnT] in the upper 600 m of the
North Pacific (sourceBruland, 1989).
Complexation with Organic Ligands 43
http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-7/26/2019 Controls of Trace Metals in Seawater
22/25
about the role of trace metals in biologicalsystems. The previous conceptions that tracemetals bound to these ligands were not bioavail-able has undergone a complete change since thediscovery that iron bound to siderophores can bereadily accessed by phytoplankton. It may be that
the strong cobalt-binding ligands observed bySaito and Moffett (2001)andEllwood and van denBerg (2001)are a type of cobalophore and playa role similar to siderophores with iron. Theproduction of highly specific metal-bindingligands by phytoplankton has led to manyinteresting questions which are yet to be resolved.
New analytical techniques are emerging suchas electrospray ionization mass spectroscopy(ESI-MS) that may prove useful in gaining insightinto the structure and functional groups of metal-binding organic ligands in seawater. First, how-ever, the low concentrations of metal-bindingligands must be concentrated from the seawatermatrix and isolated or separated from the rest ofthe dissolved organic matter. This will hopefullyallow a better mechanistic understanding of theproduction and regulation of these ligands inthe upper water column and hence increase theunderstanding of the cycling of dissolved tracemetals in the oceans.
In summary, we have presented a generaloverview of the major controls of trace metals inseawater developed from the extensive researchon trace metals over the last few decades. We
wanted the reader to gain a first order under-standing and not present a comprehensive reviewon the distribution of each trace metal. Each of thetrace metals discussed will undoubtedly prove tohave unique characteristics and subtle differencesfrom this simplified version, yet the comparisonwith these simplified characteristics will serve as agood spring board to a more complete under-standing. We have gained an insight into many ofthe processes affecting trace-metal cycling withinthe oceans, which has revealed many interestingquestions still to be answered. Developments in
isolating and characterizing metal-binding organicligands should elucidate further the controlsof trace metals in the upper water column. Whilein the mesopelagic zone the characterization oftrace-metal fluxes and the processes involved inremineralization will greatly enhance our under-standing of trace-metal cycling both within thisregion and exchanges between the benthos and theupper photic zone.
REFERENCES
Anbar A. D., Wasserburg G. J., Papanastassiou D. A., andAnderson P. S. (1996) Iridium in natural waters. Science273, 15241528.
Bacon M. P. and Anderson R. F. (1982) Distribution of thoriumisotopes between dissolved and particulate forms in thedeep-sea.J. Geophys. Res. 87, 20452050.
Barbeau K., Rue E. L., Bruland K. W., and Butler A. (2001)Photochemical cycling of iron in the surface ocean mediatedby microbial iron(III)-binding ligands. Nature 413,409413.
Barbeau K., Rue E. L., Trick C. G., Bruland K. W., and ButlerA. (2003) The photochemical reactivity of siderophoresproduced by marine heterotrophic bacteria and cyanobac-teria, based on characteristic iron(III)-binding groups.
Limnol. Oceanogr.48, 10691078.Boukhalfa H. and Crumbliss A. L. (2002) Chemical aspects of
siderophore mediated iron transport. Biometals 15,325339.
Bowie A., Maldonado M., Frew R. D., Croot P. L., Achterberg
E. P., Mantoura R. F. C., Worsfold P. J., Law C. S., and BoydP. W. (2001) The fate of added iron during a mesoscalefertilisation experiment in the Southern Ocean. Deep-Sea
Res. II48, 27032743.Boyd P. W., Watson A. J., Law C. S., Abraham E. R., Trull T.,
Murdoch R., Bakker D. G. E., Bowie A., Buesseler K. O.,Chan H., Charette M. A., Croot P. L., Downing K., Frew R.,Gall M. P., Hadfield M., Hall J., Harvey M., Jameson J.,LaRoche J., Liddicoat M., Ling R., Maldonado M. T.,
McKay R. M., Nodder S., Pickmere S., Pridmore R., RintoulS., Safi K., Sutton P., Strzepek R., Tanneberger K., TurnerS., Waite A., and Zeldis J. (2000) A mesoscale phytoplank-ton bloom in the Polar Southern Ocean stimulated by ironfertilization.Nature 407, 695702.
Boye M. B., Aldrich A. P., van den Berg C. M. G., deJong J. T.M., Veldhuis M. J. W., and de Baar H. J. W. (2003)Horizontal gradient of the chemical speciation of iron insurface waters of NE Atlantic Ocean. Mar. Chem. 50,129143.
Boyle E. A. (1988) Cadmium: chemical tracer of deep water
oceanography.Paleoceanography3, 471 489.Boyle E. A., Sclater F. R., and Edmond J. M. (1976) On the
marine geochemistry of cadmium. Nature 263, 42 44.Boyle E. A., Edmond J. M., and Sholkovitz E. R. (1978) The
mechanism of iron removal in estuaries. Geochim. Cosmo-chim. Acta41, 1313.Brand L. E., Sunda W. G., and Guillard R. R. L. (1986)
Reduction in marine phytoplankton reproduction rates bycopper and cadmium.J. Exp. Mar. Biol. Ecol. 96, 225 250.
Brewer P. G., Spenser D. W., and Robertson D. E. (1972) Traceelement profiles from the GEOSECS II test station in theSargasso Sea. Earth Planet. Sci. Lett.16, 111116.
Broecker W. S. and Peng T. H. (1982) Tracers in the Sea.Eldigo Press, Palisades, NY.
Bruland K. W. (1980) Oceanographic distributions of Cd, Zn,Cu, and Ni in the North Pacific. Earth Planet. Sci. Lett. 47,176198.
Bruland K. W. (1983) Trace elements in sea water. InChemical
Oceanography(eds. J. P. Riley and R. Chester). Academic
Press, London, vol. 8, pp. 157220.Bruland K. W. (1989) Complexation of zinc by natural organic
ligands in the central North Pacific. Limnol. Oceanogr. 34,269285.
Bruland K. W. (1992) Complexation of cadmium by naturalorganic ligands in the central North Pacific. Limnol.Oceanogr.37, 10081017.
Bruland K. W. andCoaleK. H. (1986)Surface water 234Th/238U
disequilibria: spatial and temporal variations of scavengingrates within the Pacific Ocean. In Dynamic Processes inthe Chemistry of the Upper Ocean, NATO ConferenceSeries IV: Marine Sciences (eds. J. D. Burton, P. G. Brewer,and R. Chesselet). Plenum, New York, pp. 159 172.
Bruland K. W. and Franks R. P. (1983) Mn, Ni, Cu, Zn, and Cd
in the western North Atlantic. In Trace Metals in Seawater
(eds. C. S. Wong, E. A. Boyle, K. W. Bruland, J. D. Burton,and E. D. Goldberg). Plenum, New York, pp. 395414.
Bruland K. W. and Rue E. L. (2001) Analytical methods fordetermination of concentrations and speciation of iron. In
The Biogeochemistry of Iron in Seawater(eds. D. R. Turnerand K. A. Hunter). Wiley, Chichester, pp. 255289.
Controls of Trace Metals in Seawater44
7/26/2019 Controls of Trace Metals in Seawater
23/25
Bruland K. W., Knauer G., and Martin J. (1978a) Cadmium in
Northeast Pacific waters. Limnol. Oceanogr. 23, 618 625.
Bruland K. W., Knauer G., and Martin J. (1978b) Zinc in
Northeast Pacific waters. Nature 271, 741 743.
Bruland K. W., Donat J. R., and Hutchins D. T. (1991)
Interactive influences of bioactive trace metals on biological
production in oceanic waters. Limnol. Oceanogr. 36,
15551577.
Bruland K. W., Orians K. J., and Cowen J. P. (1994) Reactivetrace metals in the stratified central North Pacific. Geochim.
Cosmochim. Acta58, 31713182.
Bruland K. W., Rue E. L., Donat J. R., Skabal S., and Moffett
J. W. (2000) An intercomparison of voltammetric
approaches to determine the chemical speciation of
dissolved copper in a coastal seawater sample.Anal. Chim.
A 405, 99113.
Bruland K. W., Rue E. L., and Smith G. J. (2001) The influence
of iron and macronutrients in coastal upwelling regimes off
central California: implications for extensive blooms of
large diatoms. Limnol. Oceanogr.46, 16611674.
Burton J. D. and Statham P. J. (1990) Trace metals in seawater.
In Heavy Metals in the Marine Environment (eds. P. S.
Rainbow and R. W. Furness). CRC Press, Boca Raton, FL.Byrne R. H., Kump L. R., and Cantrell K. J. (1988) The
influence of temperature and pH on trace metal speciation in
seawater. Mar. Chem. 25, 163181.
Capodaglio G., Coale K. H., and Bruland K. W. (1990) Lead
speciation in surface waters of the eastern North Pacific.
Mar. Chem.29, 221 238.
Chan L. H., Edmond J. M., Stallard R. G., Broecker W. S.,
Chung Y., Weiss R. F., and Ku T. L. (1976) Radium and
barium at GEOSECS stations in the Atlantic and Pacific.
Earth Planet. Sci. Lett. 32, 258 267.
Chan L. H., Drummond D., Edmond J. M., and Grant B. (1977)
On the barium data from the Atlantic GEOSECS expedition.
Deep-Sea Res. 24, 613 649.
Clegg S. L., Bacon M. P., and Whitfield M. (1991) Application
of a generalized scavenging model to thorium isotope andparticle data at equatorial and high latitude sites in the
Pacific Ocean. J. Geophys. Res.96, 2066520670.
Coale K. H. and Bruland K. W. (1985) 234Th/238U disequilibria
within the California Current.Limnol. Oceanogr.30, 2233.
Coale K. H. and Bruland K. W. (1987) Oceanic stratified
euphotic zone as elucidated by 234Th/238U disequilibria.
Limnol. Oceanogr.32, 189 200.
Coale K. H. and Bruland K. W. (1988) Copper complexation in
the northeast Pacific. Limnol. Oceanogr. 33, 10841101.
Coale K. H. and Bruland K. W. (1990) Spatial and temporal
variability in copper complexation in the North Pacific.
Deep-Sea Res. 37, 317 336.
Coale K. H., Johnson K. S., Fitzwater S. E., Gordon R. M.,
Tanner S., Chavez F. P., Ferioli L., Sakamoto C., Rogers P.,
Millero F., Steinberg P., Nightingale P., Cooper D., Cochlan
W. P., Landry M. R., Constantinou J., Rollwagen G.,
Trasvina A., and Kudela R. (1996) A massive phytoplankton
bloom induced by a ecosystem-scale iron fertilisation
experiment in the equatorial Pacific Ocean. Nature 383,
495501.
Cowen J. P., Betram M. A., Baker G. T., Feely R. A., Massoth
G. J., and Summit M. (1998) Geomicrobial transformations
of manganese in Gorda Ridge event plumes. Deep-Sea Res.
II45, 27132737.
Cullen J. T., Chase Z., Coale K. H., Fitzwater S. E., and
Sherrell R. M. (2003) Effect of iron limitation on the
cadmium to phosphorus ratio on natural phytoplankton
assemblages f