CONTENTS

Chapter I. 11

The Vertical Distribution of Mercury Species in Wisconsin lakes:

Accumulation in Plankton layers

Carl J. Watras and Nicolas S. Bloom

Abstract ..................................................................................................................................... 137 I. Introduction ........................................................................................................................ 138

II. Methods .............................................................................................................................. 139 A. Study Sites ................................................................................................................... 139 B. Bio-Optical Profiles .................................................................................................... 139 C. Mercury Analysis ........................................................................................................ 140 D. Ancillary Chemistry ...................... : ............................................................................. 140

ill. Results ................................................................................................................................ 140 A. Hg and Methyl-Hg ..................................................................................................... 140

1. Little Rock Lake .................................................................................................... 140 2. Pallette Lake ........................................................................................................... 142 3. Mary Lake .............................................................................................................. 143

B. Ancillary Parameters .................................................................................................. 144 1. Iron and Manganese .............................................................................................. 144 2. Chlorophyll-a and Bacteriochlorophylls ............................................................. 144 3. Dissolved Oxygen, Organic Carbon, Sulfate and Sulfide ................................. 144

IV. Discussion .......................................................................................................................... l48 Acknowledgments .................................................................................................................... l50 References ................................................................................................................................. 150

ABSTRACT: The aquatic cycling of highly bioreactive trace elements such as Hg is potentially influenced by layers of physiologically active plankton which develop in the thermocline and hypolimnia of lakes. Here we examine the distribution of waterborne mercury species (Hg and methyl-Hg in dissolved and particulate forms) with respect to the vertical distribution of plankton layers in three quite different Wisconsin lakes. We also examine corresponding distributions of waterborne Fe, Mn, dissolved oxygen, organic carbon, sulfate, and sulfide. To resolve fine-scale spatial features of the plankton layers, the watercolumn was mapped using computer-based, bio-optical techniques with an absolute resolving power of about 2 em.

Mercury accumulated in all the plankton layers we observed, reaching maximum concentrations of 45 ng Hg!L (compared to 1 ng!L in the epilimnion),

there were distinct differences in the distributions of Hg and methyl-Hg. accumulated primarily in plankton layers below the oxic/anoxic (0/A)

, reaching 12 ng!L in anoxic layers (> 100-fold higher than [methyl-Hg] epilimnion). Spatially and seasonally, [Hg] tended to track the concentration

(eukaryotic phytoplankton and cyanobacteria); [methyl-Hg]

137

138 MERCURY POLLUTION: INTEGRATION AND SYNTHESIS

tracked the distribution of bacteriochlorophylls (phototrophic sulfur bacteria). Maxima of bacteriochlorophyll and methyl-Hg both occurred near the sulfide/ sulfate transition zone below the oxic/anoxic boundary. Unlike Fe and Mn, the concentration of Hg and methyl-Hg tended to decrease in the region between the lowest plankton layers and the sediment surface. While the orthograde distributions of Fe and Mn indicate redox-driven diffusion from sediments into the overlying watercolumn, more complex processes appear to determine the observed vertical distribution of mercury species.

Given the strong affinity of Hg for biotic particles (log Kd = 5 to 6), either Hg scavenging/settling from above or Hg diffusion/sorption to plankton layers could account for the elevated [Hg] observed in all plankton layers. Both mechanisms could be involved if Hg was released from settling detritus and then resorbed by cells within the layers. Since [methyl-Hg] maxima were confined to anoxic plankton layers, either diffusion/sorption from below (i.e., sediments) or de novo production of methyl-Hg within these layers is implied. Our data tend to support the hypothesis that methyl-Hg is formed within anoxic layers via conversion of Hg derived from the upper watercolumn. We explore this hypothesis and its implications.

I. INTRODUCTION

With growing reports of methyl-Hg contamination in freshwater fisheries, there is an in­creasing need for accurate data on the concentration and distribution of waterborne mercury species in lakes and streams. Although the transformation of inorganic mercury to methylated Hg species is known to occur in freshwater, estuarine, and marine sediments,1- 3 there is also evidence of methylmercury production in the watercolumn4-<i dating back to the calculations of Topping and Davies7 which indicated that the annual production of methylmercury in the water and sediments of the world oceans was similar. There remains substantial uncertainty about:

1. The relative importance of sedimentary vs. watercolumn methylation; 2. The identity of methylating microbes; 3. Rates of methyl-Hg formation and destruction in natural ecosystems; and 4. The environmental dependencies of these rates.

These uncertainties stem, in part, from a lack of reliable data on the distribution of waterborne Hg species.

Recent analytical advances now permit the determination of Hg and alkyl-Hg species in natural waters with high precision and reproducibility.8•9 Using trace metal clean sampling techniques, we have determined the ambient concentration and distribution of Hg species in pristine lake water, rain, air, sediments, porewaters, and the biota of several Wisconsin lakes. HHS In the surface waters of these lakes, [HgT] generally ranged from about 0.5 to 5 ng Hg!L and [methyl-Hg] ranged from 0.05 to 0.5 ng methyl-Hg!L. However, much higher concentrations of both Hg and methyl-Hg have been observed at depth in some lakes, par­ticularly in metalirnnia and hypolimnia where dense populations of plankton accumulate in discrete layers. 12

Plankton layers typically comprise either phytoplankton or bacterioplankton along with assemblages of protozoans and small metazoans. 19 Guerrero and Mas20 coined the acronym MPM (multilayered planktonic microbial communities) to describe plankton layers and they documented similarities between laminated microbial communities in the watercolurnn and those observed in benthic mats. Gasol and Pedros-Alio21 have demonstrated that plankton layers form via in situ growth rather than the passive accumulation of settling cells.

VERTICAL DISTRIBUTION OF MERCURY IN WISCONSIN LAKES

Table 1 Characteristics of the three study lakes in north-central Wisconsin.

139

Surface ZmaxLake Type area (ha) (m) pHa Ref.

Pallette Seepage, dimictic 70 18.2 7.3 33Little Rock" Seepage, dimictic 9.8 10.3 5.0 26, 34Mary Drained, meromictic 1.2 21.7 5.9 31,35,36

a Epilimnetic.b Treatment Basin.

Although relationships between the physiological activity of plankton layers and the cy­cling of carbon and sulfur have been studied,22 their roles in trace element cycling remainlargely unknown. Our early work on mercury distributions in Wisconsin lakes suggested thatplankton layers were potentially important sites for biologically driven mercurytransfonnations.12

Here, we report data on the distribution of Hg and alkyl-Hg species relative to the distri­bution of plankton layers and other physicochemical gradients in the watercolumn of threesmall lakes in northern Wisconsin. These data were obtained using bio-optical instrumenta­tion to map the fine-scale, vertical distribution of planktonic communities23.24 combined withthe "clean" analytical and sampling techniques used for low-level mercury determinations.We also examine hypotheses relating plankton layers and sulfur cycling to methyl-Hgfonnation.

II. METHODS

A. STUDY SITESThe three study sites are small, remote lakes located in the Northern Highlands Lake Districtof Wisconsin, near the Trout Lake Biological Station (46"N, 89"W). This region is heavilyforested and sparsely populated. The area contains hundreds of seepage and drainage lakes

in thick glacial tills and outwash sands at an elevation of about 500 m.25 The threelakes were selected because their stratification regimes result in very different patterns

planidton layering. In Little Rock Lake, plankton layers form close to the sediment/waterIn Pallette Lake, they fonn at mid-water depths; and in Mary Lake they fonn inwaters well above the sediments. By comparing Hg accumulation in layers that

at such different locations, we hope to evaluate potential Hg sources and accumulationhanisms.

elakes were also chosen because they represent both holomictic and meromictic waters.Rock Lake and Pallette Lake are dimictic, clearwater seepage lakes that stratify ther­soon after ice-out in spring (April/May) and remain stratified until the autumn mixis

ber/October). Mary Lake is dystrophic (color 100 PCU) and meromictic. Additionalthese lakes are listed on Table 1.

-OPTICAL PROFILESer-based, bio-optical profiling unit was used to measure temperature, dissolved

th, optical density, right-angle light scattering, in vivo chlorophyll-a fluorescence,elling irradiance at 1 to 2 cm depth intervals throughout the watercolumn of each

housing was lowered at 1 cm s-l using an electric winch deployed from a(Boston Whaler). The profiler design was based on prototypes described

and Steenbergen et aJ.24 Temperature was measured with a Ther-thelmistor; dissolved oxygen with a continuously stirred, temperature­

polarographic electrode; depth with a Druck PTX l601D depth trans­with a 25-cm, folded-beam, Martek XMS transmissometer; right-angle

140 MERCURY POLLUTION: INTEGRATION AND SYNTHESIS

scatter with a Turner Designs Model 10 field nephelometer; in vivo fluorescence with aTurner Designs Model 10-005R field fluorometer with the standard Chl-a filter set suppliedby Turner Designs; and downwelling irradiance with a Licor underwater PAR cosine col­lector. Water was pumped through the deck-mounted fluorometer and the nephelometer usinga submersible pump. The pump and all other sensors were aligned with the transmissometerbeam at depth. All data were logged on a portable computer, corrected for pumping rate,and displayed real-time during descent.

C. MERCURY ANALYSISWater samples for Hg analysis were collected by extending a I m acrylic outrigger to thesensor housing of the bio-optical profiler and then pumping water to the surface throughrigorously cleaned and blanked tubing using a nonmetallic pump. The sample intake portwas aligned with the transmissometer beam. Water samples were collected on deck intoprecleaned, Teflon® bottles that were hermetically sealed and double-bagged for transport.Hg-clean technique was followed throughout the process of sample collection and handling. 10

Total Hg, (HgT), was detennined on unfiltered samples after BrCl wet-oxidation by SnChreduction, dual gold-column amalgamation, followed by thermal desorption and atomic flu­orescence detection.8 Dissolved Hg, (HgD), was detennined the same way on lakewatersamples filtered through Hg-free, 0.4 J.lm, 47 rum disposable filtration units. Particulate Hg(HgP) was calculated by difference. Operational detection limits for these Hg species wereabout 0.05 ng HgIL in a 100 mL sample.

Total rnonomethylmercury, (methyl-HgT), was detennined by extracting unfiltered sam­ples from an HClIKCI matrix into CH2Ch, then back-extracting into pure water by solventevaporation. The aqueous phase was ethylated with NaB(C2Hs)4 and the volatile ethyl analogswere quantified via pre-trapping by gas purging onto graphitized carbon, cryogenic GCseparation, pyrolysis to HgO, followed by atomic fluorescence detection.9 Dissolved methyl­Hg (methyl-HgD) was detennined the same way on filtered samples. Particulate methyl-Hg(methyl-HgP) was detennined by difference. Detection limits for these species were about0.01 ngIL as Hg in a 50 mL sample.

C. ANCILLARY CHEMISTRYIn addition to the in situ oxygen probe, dissolved oxygen was analyzed at selected depthsusing the Pomeroy-Kirschner modification of the Winkler titration. Sulfate was detenninedby suppressed-column ion chromatography. Dissolved sulfide was detennined after SAOBIladdition using an. Orion model 94-16BN silver/sulfide electrode and an Orion double­junction reference electrode (detection limit circa 100 nM).27 Fe and Mn were determinedon unfiltered samples using ICP. Dissolved organic carbon (DOC) was detennined on 0.4J.lm filtered samples by wet oxidation with sodium persulfate followed by heating (2000 C)and non-dispersive CO2 detection. Photosynthetic pigments were detennined by reverse­phase HPLC, following the method of Hurley and Watras28 for bacteriochlorophylls. Sus­pended particulate mass (SPM) was detennined on oven-dried, preweighed filters.

III. RESULTS

A. Hg AND METHYL-Hg1. little Rock LakeThe development of two deep plankton layers in Little Rock Lake began soon afterand the layers reached peak density in mid-August (Figures 1 and 2). The locationlayers changed as the thermocline deepened and anoxia developed in the hypolimnion.upper layer comprised Chl-a-containing phytoplankton and the lower layer comprisedd-containing bacterioplankton (Figure 9). The optical density trace went off-scale indue to intense scattering associated with the small particles (Figure 2).

VERTICAL DISTRIBUTION OF MERCURY IN WISCONSIN LAKES

Fluorescence (Rfu)o 1 2 3 4 5

141

Figure 1 Seasonal development of plankton layers in the thermocline and hypolimnionof Little Rock Lake, as measured by in vivo chlorophyll fluorescence with a spatial reso­lution of 2 cm (RFU, relative fluorescence units). Profiles were similar at multiple stations.

Seasonal changes in [HgT] paralleled changes in [CW-a] within the upper plankton layer(Figure 3). At peak stratification, [HgT] reached rougWy 45 ng/L in this region. This rep­resents an increase of 40- to 50-fold over the average epilimnetic [HgT] of about 1 ng/L.

Near the O/A boundary, there were pronounced changes in the chemical speciation andphysical form of waterborne mercury (Figure 4). As noted above, maximum concentrationsof total Hg occurred within the upper plankton layer which spanned the O/A boundary andInost (94%) of the HgT was particulate (Figure 4a). Both [HgT] and [HgP] decreased fromthis point to the sediment surface. The dissolved Hg fraction (HgD) was relatively constant(8.95 to 8.97 ng/L) in the three bottom samples from 9.0 to 9.5 M.

Maximum concentrations of methyl-Hg occurred within the lower (anoxic) plankton layer(.figure 4b). [Methyl-Hg] reached 11.6 ng/l in this layer, over 100 times greater than epilim­

tic concentrations. As observed with [HgP], [methyl-HgP] was highest near the planktoner peak, where it comprised >50% of the methyl-HgT. Concentrations of methyl-HgTdmethyl-HgD decreased between the layer and the sediment surface.

142 MERCURY POLLUTION: INTEGRATION AND SYNTHESIS

Optical Density (11M)o 5 10 15 20 25 30 35

Figure 2 Seasonal development of plankton layers in Little Rock Lake as measured bytransmissometry.

2. Pallette LakeIn Pallette Lake we also observed two major plankton layers, but they were broader, morewidely spaced, and located further from sediments than those in Little Rock Lake (Figure5). The upper layer was associated with a dissolved oxygen maximum indicating oxygenicphotosynthesis in the metalimnion. Microscopic inspection indicated that this upper planktonlayer was dominated by small, filamentous cyanobacteria. The lower (anoxic) plankton layerwas sharper and it was rich in Bchl-e, characteristic of brown-green (Chlorobiaceae) pho­totrophic sulfur bacteria (PSB). Unlike Little Rock Lake, where the PSB layer was close tothe sediment surface, these PSB were 4-6 m above the bottom.

[HgT] peaked within both the oxic and anoxic plankton layers of Pallette LakeSa). Near the top of both plankton layers, [HgP] constituted up to 96% of the [HgP]5b). Concentrations of dissolved mercury were highest just below the particulate Hg l1l~llo.lll%

but [HgD], [HgP], and [HgT] all decreased toward sediments.As in Little Rock Lake, maximal [methyl-Hg] occurred in anoxic water just

second [Hg] maximum (Figure 6). Methyl-Hg constituted roughly 30% of the HgTlower layer and about half of the methyl-Hg was in the particulate form. Due to anaJytii¢!lldifficulties, we were not able to resolve the dissolved/particulate fractionation above orthe layers; but, as observed with [Hg], concentrations of methyl-HgT decreasedsediments.

VERTICAL DISTRIBUTION OF MERCURY IN WISCONSIN LAKES 143

(ng/L)

45403530252015.';' 10

.~t: 5

o

(ng/L)

45403530252015105o

(1J9/L

100908070605040302010o

MJ J AS 0 N

Month

0-.-----""""''''''1 HgP

23456789 +---.--.,.--""

M A

0123456789-

01 HgT

2

~3

.c 41i 5Q)

6Q

789

Mary LakeLake is meromictic and during August the thermocline was very close to the surfacelake. The mixed layer was confined to the upper I m (Figure 7). Dissolved oxygen

lcelltratiOlls decreased rapidly below the narrow mixed layer and the O/A boundary wasat about 2 m depth, roughly 17 m above the profunda! sediments. A distinct plankton

in this region.eth1{H-jLg concentrations were greatest in the plankton layer at 2 m (Figure 7a). At this

40% of the HgT was methylated (Figure 7b). In surface waters and at depth,of the Hg was methylated. A secondary methyl-Hg maximum was located

of the monimolimnion. Below 10 m, the concentration of methyl-Hg was verydec:reased with depth to the sediment surface.

Seasonal and spatial development of chlorophyll-a, total mercury (HgT), andparticulate mercury (HgP) maxima in Little Rock Lake.

144 MERCURY POLLUTION: INTEGRATION AND SYNTHESIS

6Bio-optical Data

2 3 4 5o5rr--~~~~~-~~~-~--,

6Bio-optical Data

2 3 4 5o5rr---~~~--'-~--~~"""-1

6 7' 6

10 20 30Mercury (ng/l)

*Hgt ~Hgp

40 50

\(

i:~_~~ 0\.~••~.~~nnn

9 _,.,",,' ..._p7f~' /v"- \

mmHQ'd ~mHgt FI10,--~--------~~_-,

o 4 8 12Methylmercury (ng/l)

• methyl-Hgt cO methyl-Hgd

Figure 4 Vertical distribution of mercury species in Little Rock Lake at maximum strat­ification compared to em-scale profiles of in vivo chlorophyll fluorescence (FI, relative units)and right-angle light scatter (Sc, relative units). Hgt, total mercury; Hgp, particulate mercury(>0.4 Jlm); methyl-Hgt, total monomethylmercury; methyl-Hgd, dissolved monomethyl­mercury «0.4 Jlm). Dotted line indicates approximate depth of the oxic/anoxic boundary(O/A boundary) roughly 1 m above profundal sediments. Note that depth scale begins at5 m to improve spatial resolution of near sediment zone.

B. ANCILLARY PARAMETERS1. Iron and ManganeseDistributions of Fe and Mn in all three lakes indicated a strong redox response to anoxicconditions in the hypolimnion (Figure 8). Increases in [Fe] and [Mn] below the O/A boundarysuggested diffusive flux from sediments into the overlying water. The behavior of theseredox-sensitive elements contrasts with that observed with Hg and methyl-Hg species, whichdecreased in concentration in the region between plankton layers and sediments.

2. Chlorophyll-a and BacteriochlorophyllsRather than following [Fe] and [Mn], the distribution of mercury species more closely fol­lowed the distribution of photosynthetic pigments in these lakes (Figure 9). In general,maxima of [ChI-a] and [Hg] tended to co-occur, as did maxima of [Bchl] and [methyl-Hg].

3. Dissolved Oxygen, Organic Carbon, Sulfate and SulfideIncreased dissolved oxygen near peaks in Chl-a indicated active oxygenic photosynthesis inall the lakes (Figure 9). All profiles were made near midday on sunny afternoons, when onewould expect high photosynthetic rates in nutrient-replete algae. DOC also tended to increasein these regions, perhaps as the result of leaking photosynthate.

In all cases, the methyl-Hg maximum occurred in anoxic water near the sulfatt~/slllfi(ie

transition zone. The vertical distributions of S04 and total sulfide were similar in alllakes, although absolute concentrations varied (Figure 9). [S04], for example, wastently higher in Little Rock Lake due to experimental additions during the aci,dificatiiol1years.26 Within the thermocline, [S04] assumed an S-shaped distribution indicating alterna1til1g

VERTICAL DISTRIBUTION OF MERCURY IN WISCONSIN LAKES 145

3

15

0.8

1 2Mercury (ng/L)

Bio-optical Data5 10

>-~:~Qmethyl-Hg ~"

<)

Bio-optical Data5 10 15---I

II

!

0.2 0.4 0.6Methylmercury (ng/L)

o

Tr

FI

15

5

oo

20L----1,_-'-_--'-_.1----I._--"-_-'-_--'

o

15

20L----~-~---'--L------' 20L--'---_~_~ ____'o 1 2 3 4 0

Mercury (ng/L)

Bio-optical Datao 5 10Or-----,------,---~r

I

5 i l-------_-<>

fO~~;::L::-:~f~~15 I -'

FI I <)'

Figure 5 Vertical distribution of Hg species in Pallette Lake relative to centimeter-scaleprofiles of transmissivity (Tr = optical density m-1), and in vivo chlorophyll fluorescence(FI, relative units). H9h total mercury; Hgp particulate mercury (>0.4 11m); H9d, dissolvedmercury «0.04 11m). Dotted line indicate O/A boundary roughly 5.5 m above profundalsediments.

Figure 6 Vertical distribution ofmonomethylmercury (methyl-Hg)in Pallette Lake, WI, in relation to8Emtimeter-scale profiles of trans­rtlissometry (Tr = optical densityrn-1), and in vivo chlorophyll fluo­rescence (FI, relative units). Dottedlire shows 0/A boundary roughly.5.5 m above sediments.

146 MERCURY POLLUTION: INTEGRATION AND SYNTHESIS

Profiler Data Profiler Data0 5 10 15 0 5 10 15 20

0<3l"~"''''''''''

0

-0:w- --..fl-q?__4i>---'2

\ , Ii \i \ I

5 i \ 5 I\ ,

b

1\'\ II

~ ° , ~ I

/~' mmHgtI

£;10 £;10Il'

\a. ! \ a. :\ :Q) :1 \ Q)

0 II mmHgd 0II I ',1,

I o,'mmHg I15

* 115 ~

%

h I

il I

Ii Tr ITr

9~ t

2011 20 d

0 1 2 3 4 5 0 20 40 60 80 100Methylmercury (ng/L) Methylmercury Fraction (%)

Figure 7 Vertical distribution of methyl-Hg species in meromictic Mary Lake relative toem-scale profiles of dissolved oxygen (02, mglL), transmissometry (Tr = optical densitym-1), and temperature (0C). Methyl-Hg fraction: % methyl-Hg = ([methyl-Hgt]/[Hgt] x 100).

-+- Manganese (Il9/L)

0 50 100 150 0 50 100 50 1000 0

PALLRT

25 5

~4

.<= 10 10i5.Q)

60

15 158

10 20 200 5 10 15 0 50 100 0 2 4 6

__ Iron (mg/L)

Figure 8 Vertical distribution of Fe and Mn during stratification in Little Rock (LRT),Pallette (PAL) and Mary lakes. Dotted line indicates O/A boundary.

VERTICAL DISTRIBUTION OF MERCURY IN WISCONSIN LAKES 147

0 6 12 18 24 Terrp, , , , , , , ,

9 2,0 qo , 8p , 190 Chi a 9 , 0,4 0.8 , 1;2 1;6 , 2;0 So., .9 20 ~O 8p 190 Bohl e 9 0;2 0;4 0.6 0.8 , 1 jO L Sulfide, , ,

04 5 6 7 8 9 DOC

0

4 4

~ 8..c:liQ)

012

2PAL PAL

2

0 6 12 18 24 Terrp, , , , , ,9 , 1QO , 2QO • 390 , 490 , 590 Chi a 9 2 4 6 8 1,0 So., . , ,9 , lQO , 2QO , 390 , 490 , 590 Bohl d 9 1 ~ 1,2 1,6 50 L Sulfide

12 16 00, 1 2 3 DOC0

LRL 0LRL

2 2

~ 4 ~ 4..c: ..c:li liQ) Q)

0 6 0 6

8 8

0 6 12 18 24 Terrp, , , , , ,

8p 100 Chi a 0, , .8p 100 Bohl e 0, ,

50 a.0

4 4

~ 8 ~ 8..c: ..c:li liQ) Q)0 12 012

MY2 2

II1II Terrperature ('C) II1II So. (mglL)o Chlorophyll a (j.lgll) o r Sullide (mgJl)

o Baeteriochloropyll (llglL) o DOC(mgil)lID Dissolved oxygen (mglL)

9 Vertical distribution of chlorophyllous pigments (chlorophyll-a and bacterioch­1".r.nh"I1~\ dissolved organic carbon (DOC), dissolved oxygen (02), temperature, sulfate,

sulfide during stratification in Little Rock (LRT), Pallette (PAL), and Mary (MY)

148 MERCURY POLLUTION: INTEGRATION AND SYNTHESIS

zones of depletion-enrichment-depletion. Below the thennocline, [S04] was low but relativelyconstant. Vertical changes in the distribution of sulfide mirrored the distributions of sulfate.

IV. DISCUSSION

The distribution of photosynthetic pigments, dissolved oxygen, and sulfur species indicatethat the plankton layers observed in these lakes were not simply passive accumulations ofmaterial settling from the epilimnion. Microscopy continued that they comprised active pop­ulations of oxygenic and anoxygenic photosynthetic organisms, along with heterotrophicbacteria, protozoans, and metazoans. Although epilimnetic particles probably settled into thelayers and were remineralized there, the degradation products of chlorophyllous pigmentswere not principal constituents of the layers.28

Although more intensive process-oriented investigations are needed, the settling and re­mineralization of epilimnetic particulate could provide a potential pathway for recycling Hgat depth. Sediment trap studies in these lakes indicate that atmospherically derived Hg isscavenged in the upper waters, transported downward with settling particulate, and subse­quently remineralized.'7 Hg mass balances indicate that the dominant Hg source to the lakesis direct atmospheric deposition. 10.29 Mass balances for methyl-Hg indicate that in-lake pro­duction is an important source for this Hg species.29 Although the mechanism for in situ Hgmethylation is unknown, we speculate that some may occur in anoxic plankton layers.

Ancillary data allow preliminary evaluation of alternative ways in which methyl-Hg-richplankton layers could arise in anoxic waters. Diffusion of methyl-Hg into the layers frombelow (i.e., sediments) is one possible pathway, but the observed decreases in dissolvedmethyl-Hg with depth below the layers are not consistent with this accumulation route. Lateraldiffusion from littoral sediments is also possible, but it seems an unlikely common cause forsimilar distributions in such morphometrically and hydrodynamically different lakes. Thesettling of prefonned particulate methyl-Hg also seems an unlikely source for anoxic layers,since one would need to invoke relatively high rates of demethylation in oxic waters toaccount for the lack of methyl-Hg-rich layers there. If demethylation did occur rapidly inoxic plankton layers, we are then left without a source for downward transport to anoxiclayers.

By default, we entertain the hypothesis that methyl-Hg is fonned de novo within thelayers, perhaps via the same mechanisms that govern Hg methylation in sediments. Hgmethylation in anoxic sediments appears to be linked to sulfur cycling (perhaps sulfate re­duction),3.3o and the observations that (1) sulfate reduction occurs in the anoxic watercolumnof these lakes31 and (2) waterborne [methyl-Hg] maxima occur near the sulfate/sulfide tran­sition zone provide a preliminary basis for hypothesizing that a common mechanism governsin situ methyl-Hg production in both the sediments and the watercolumn.

Vertical profiles of particle enrichment in methyl-Hg are also consistent with this hy­pothesis of in-layer methyl-Hg production. Suspended particulate in deep layers is richer inmethyl-Hg than particulate above and below the layers (Figure 10). This pattern would beexpected if methyl-Hg were fonned within the cells of the layer and then released duringsenescence and sinking. Conversely with Hg, particle enrichment is highest in surfacewaters-as expected for an element scavenged in the epilirnnion and remineralized at depth(Figure 10).

We have calculated hypothetical rates of in situ methylation within anoxic plankton layersbased on a rough, seasonal time-course of methyl-Hg build-up (Table 2). Interestingly, thecalculated rates of methyl-Hg production in layers are remarkably consistent (50--100 ng/m2d) even though maximum concentrations vary almost lOO-fold between lakes. Further­more, these calculated rates are similar to estimates of methylation rates in other ecosystems,such as an oxic estuarine watercolumn (140 to 1400 ng/m2d)4 and flow-through, sediment-

VERTICAL DISTRIBUTION OF MERCURY IN WISCONSIN LAKES 149

-0- Total Hg

-- Methy!·Hg

a 0.2 0.4 0.6a

3

4

5

6

7

8

9

10.1.---.,,---,,---,--'October

0.5aO-r.--~-"......-\

2

3

4

5

6

7

8

9

10.1.--...-...,----'August

0.5

2

3

4

5

6

7

8

9

10.1.-----,-----'July

0.40.2

Particle Enrichment (~g/g)

aOt;--~-"7i

7

8

9

10.1.-----,--,,----'April

2

3

~ 45 5c­O>o 6

Figure 10 Seasonal pattems of mercury enrichment (Ilg!g, dry weight) in suspendedparticulate (>0.8 11m) of Little Rock Lake. Data collected the year prior to bio-opticalprofiling, hence the coarse spatial resolution.

Table 2 Estimated net rate of methylmercury production in anoxic planktonlayers of Wisconsin lakes.

509254

Netllux(nglm2/d)

100100100

Time(d)

8.0 x l()3

3.5 x l()5

1.3 x 104

Surface areaof layer

(m2)

4.0 x lQ73.2 X 109

7.0 X 107

Methyl-Hgaccumulated

(ng)

Note: LRT: Little Rock Lake-treatment basin; PAL: Pallette Lake.

Lake

water microcosms (10 to 40 ng/m2d),37 For Little Rock Lake, the calculated methyl-Hg fluxis sufficient to account for about two thirds of the methyl-Hg annually accumulated by fish(assuming annual fish production of 30%). This would also account for a significant fractionof the methyl-Hg estimated to be cycling through all of the biota annually.29 Thus, theplanktonic production of methyl-Hg is potentially an important input to the lake. However,we emphasize that direct estimates of methylation are needed to confirm these calculations.

Previous work on the composition and physiological activity of microbial communitiesnear O/A boundaries in sediments, benthic mats, and the watercolurnn of lakes2o.32 indicatesthat some biogeochemical processes may be facultatively benthic or planktonic, dependingon the location of the O/A boundary and the physical structure of the watercolurnn. Sincethe pioneering work of Jemelov and colleagues,38.39 most research on methylation in aquaticenvironments has focused on sedimentsP If benthic microbial consortia tended to migrate

sediments and spread out into anoxic water, as illustrated on Figure 11, then importantrnethy'latilon sites might move with them. If so, anoxic microbial layers could serve as useful

systems for direct methylation studies, eliminating some of the complications involvedstu(iyirlg more tightly packed sediment communities.40

LRTPALMary

150 MERCURY POLLUTION: INTEGRATION AND SYNTHESIS

A. Non-Stratified Lake

~~~~~

~Compressed Microbial

loyers

B. Stratified Lake

~~~~~~

Figure 11 Theoretical comparison of microbial distributions in surficial sediments and ina layered plankton community, relative to the location of the O/A boundary. (Adapted fromGuerrero, R. and Mas, J., Microbial Mats: Physiological Ecology of Benthic MicrobialCommumlies, Cohen, V. and Rosenberg, E., Eds., American Society for Microbiology,Washington DC, 1989, and Sieburth, J. MeN., Microbes in the Sea, Sleigh, M.A., Ed.,Horwood, Chichester, England, 1987.)

ACKNOWLEDGMENTS

We thank Ken Morrison, Steve Claas, Tris Hoffman, Jim Hurley, and Lian Liang for assis­tance in the field and laboratory. Discussions with John Sieburth, Cindy Gilmour, RichardBartha, Norbert Pfenning, John Stoltz, and Matti Verta helped to shape this manuscript.Research was supported by the Electric Power Research Institute (RP202Q-1O) and the Wis­consin Department of Natural Resources. The is a contribution from the Trout Lake FieldStation, Center for Limnology, University of Wisconsin-Madison.

REFERENCES

1. Berman, M. and Bartha, R, Levels of chemical versus biological methylation ofin sediments, Bull. Environ. Contam. Toxicol., 36,401, 1986.

2. Winfrey, M.R. and Rudd, J.W.M., Environmental factors affecting the formation ofylmercury in low pH lakes, Environ. Toxicol. Chem., 9, 853, 1990.

3. Gilmour, C.C. and Henry, E.A., Mercury methylation in aquatic systems affected bydeposition, Environ. Pollut., 71, 131, 1991.

4. Ramlal, P.S., Rudd, J.W.M., Furitani, A., and Xun, L., The effect of pH on methylproduction and decomposition in lake sediment, Can. J. Fish. Aquat. Sci., 42,

5. Xun, L. Campbell, N.E.R., and Rudd, J.W.M., Measurements of specific rates ofylmercury production in the watercolumn and surface sediments of acidified andneutral lakes, Can. J. Fish Aquat. Sci., 44,750, 1987.

VERTICAL DISTRIBUTION OF MERCURY IN WISCONSIN LAKES 151

6. Steffan, R.I., Korthals, E.T., and Winfrey, MR, Effects of acidification on mercury meth­ylation, demethylation and volatilization in sediments from an acid susceptible lake, Appl.Environ. Microbiol., 54, 2003, 1988.

7. Topping, G. and Davies, I.M., Methylmercury production in the marine watercolumn,Nature, 290,243, 1981.

8. Bloom, N.S. and Fitzgerald, W.F., Detennination of volatile mercury species at the pi­cogram level by low temperature gas chromatography with cold vapor atomic fluorescencedetection, Anal. Chim Acta., 208, 151, 1988.

9. Bloom, N.S., Detennination of picogram levels of methylmercury by aqueous phase eth­ylation, followed by cryogenic gas chromatography with cold vapor atomic fluorescencedetection, Can. J. Fish. Aquat. Sci., 46, 1131, 1989.

10. Fitzgerald, W.F. and Watras, C.I., Mercury in the surficial waters of rural Wisconsin lakes,Sci. Total Environ., 87/88,223, 1989.

11. Bloom, N.S. and Watras, C.I., Observations of methylmercury in precipitation, Sci. TotalEnviron., 87/88, 199, 1989.

12. Bloom, N.S. and Watras, C.I., Seasonal and vertical variability in the mercury speciationof an acidified Wisconsin lake, in: Heavy Metals in the Environment: Proceedings of theSeventh Intemational Conference, Vol. 2, Vernet, J.-P., Ed., CEP Consultants, Edinburgh,U.K., 1989,349.

13. Watras, C.l, Bloom, N.S., Fitzgerald, W.P., Hurley, lP., Krabbenhoft, D.P., Rada RG.,and Wiener, J.G., Mercury in temperate lakes: a mechanistic field study, Verh. Int. Verein.limnol., 24,2199, 1991.

14. Wiener, lG., Fitzgerald, W.F., Watras, C.I., and Rada, RG., Partitioning and bioavaila­bility of mercury in an experimentally acidified Wisconsin lake, Environ. Toxicol. Chem.,9,909, 1990.Bloom, N.S., Watras, C.I., and Hurley, J.P., Impact of acidification on the methylmercurycycling of remote seepage lakes, Water Air Soil Pollut., 56,477, 1991.Fitzgerald, W.F., Mason, RP., and Vandal, G.M., Atmospheric cycling and airwaterexchange of mercury over mid-continental lacustrine regions, Water Air Soil Pollut., 56,745, 1991.Hurley, lP., Watras, C.I., and Bloom, N.S., Mercury cycling in a northern Wisconsinseepage lake: the role of particulate matter in vertical transport, Water Air Soil Pollut., 56,543, 1991.Watras, C.I. and Bloom, N.S., Mercury and methylmercury in individual zooplankton:implications for bioaccumulation, Limnol. Oceanogr., 37, 1313, 1991.Lindholm, T., Ecology of Photosynthetic Procaryotes with Special Reference to Merom­ictic Lakes and Coastal Lagoons, Abo Academy Press, Abo, Finland, 1987.Guerrero, R and Mas, J., Multilayered microbial communities in aquatic ecosystems:growth and loss factors, in: Microbial Mats: Physiological Ecology ofBenthic MicrogialCommunities, Cohen, Y. and Rosenberg, E., Eds., American Society for Microbiology,Washington, D.C., 1989, chap. 4.Gasol, J.M. and Pedros-Alio, C., On the origin of deep algal maxima: the case of LakeCiso, Verh. Int. Verein. Limnol., 24, 1024, 1991.van Gemerden, H., Montesinos, E., Mas, J., and Guerrero, R, Diel cycle of metabolismof phototrophic purple sulfur bacteria in Lake Ciso (Spain), Limnol. Oceanogr., 30, 932,1985.Watras, C.I. and Baker, A.L., The spectral distribution of downwelling light in northernWisconsin lakes, Arch. Hydrobiol., 112,481, 1988.Steenbergen, C.L.M., Korthals, H.l, Baker, A.L., and Watras, C.I., Microscale verticaldistribution of algal and bacterial plankton in Lake Vechten (The Netherlands), FEMSMicrobiol. Ecol., 62,209, 1989.

r

152 MERCURY POLLUTION: INTEGRATION AND SYNTHESIS

25. Magnuson, lJ., Bowser, C.J., and Kratz, T.K, Long-term ecological research (LTER) onnorth temperate lakes of the United States, Verh. Int. Vereill. Limnol., 22,533, 1984.

26. Watras, e.J. and Frost, T.M., Little Rock Lake: perspectives on an experimental ecosystemapproach to seepage lake acidification, Arch. Environ. Contam. Toxicol., 18, 157, 1989.

27. van Gemerden, Competition between purple sulfur bacteria and green sulfur bacteria: roleof sulfur, sulfide and polysulfides, in: Ecology ofPhotosynthetic Procaryotes with SpecialReference to Meromictic Lakes and Coastal Lagoons, Lindholm, T., Ed., Abo AcademyPress, Abo, Finland, 1987, 13.

28. Hurley, J.P. and Watras, C.J., Identification ofbacteriochlorophylls in lakes using reverse­phase HPLC, Limnol. Oceanogr., 36,307, 1991.

29. Watras, C.J., Bloom, N.S., Hudson, R.J.M., Gherini, S., Wiener, lG., Fitzgerald, W.P.,and Porcella, D.B., Sources and fates of mercury and methylmercury in Wisconsin lakes,in: Mercury Pollution: Integration and Synthesis, Watras, C.J. and Huckabee, lW., Eds.,Lewis Publishers, Chelsea, MI, 1994, chap 1.12.

30. Compeau, G.C. and Bartha, R., Sulfate-reducing bacteria: principal methylators of mercuryin anoxic marine sediment, Appl. Environ. Microbiol., 50,261, 1987.

31. Parkin, T.B. and Brock, TD., The role of phototrophic bacteria in the sulfur cycle of ameromictic lake, Limnol. Oceanogr. 26, 880, 1981.

32. Sieburth, J. MeN., Contrary habitats for redox-specific processes: methanogenesis in oxicwaters and oxidation in anoxic waters, in Microbes in the Sea, Sleigh, M.A., Ed., Horwood,Chichester, England, 1987.

33. Engle, S.S., Utilization of Food and Space by Cisco, Yellow Perch and Introduced CohoSalmon, With Notes on Other Species, in Pallette Lake, WI, Ph.D. Dissertation, Universityof Wisconsin-Madison, 1972.

34. Brezonik, P.L., Baker, L.A, Eaton, J.R., Frost, T.M., Garrison, P., Kratz, T.K, Magnuson,J.J., Rose, W.J., Shepard, B.K, Swenson, W.A., Watras, C.J., and Webster, KE., Theacidification of Little Rock Lake, Water Air Soil Pol/ut., 31, 115, 1986.

35. Weimer, W.e. and Lee, G.P., Some considerations of the chemical limnology of mer­omictic Mary Lake, Limnol. Oceanogr., 18,414, 1973.

36. Hutchinson, G.E.,A Treatise on Limnology, Vol. 1, John Wiley & Sons, New York, 1957.37. Wright, D.R. and Hamilton, R.D., Release of methylmercury from sediments: effects of

mercury concentration, low temperature and nutrient addition, Can. J. Fish. Aquat. Sci.,39, 1459, 1982.

38. Jemelov, A, Release of methylmercury from sediment with layers containing inorganicmercury at different depths, Limnol. Oceanogr., 15, 958, 1970.

39. Fagerstrom, T. and Jeme1ov, A, Formation of methylmercury from pure mercuricin anaerobic organic sediment, Water Res., 5, 121, 1971.

40. Sweerts, J.-P.R.A., Oxygen Consumption Processes, Mineralization and Nitrogenat the Sediment-Water Interface of North Temperate Lakes, Ph.D. Thesis, Rijksllni'veJ'site:itGroningen, Holland, 1990.