Environ. Sci. Technol. 2001, 35, 3089-3098
Importance of the Forest Canopy to Fluxes of Methyl Mercury and Total Mercury to Boreal Ecosystems V I N C E N T L . S T . L O U I S , * ,† JOHN W. M. RUDD,‡ CAROL A. KELLY,‡ BRITT D. HALL,† KRISTOFER R. ROLFHUS,§ KAREN J. SCOTT,| STEVE E. LINDBERG,⊥ AND WEIJIN DONG⊥ Department of Biological Sciences, University of Alberta, Edmonton, Alberta, T6G 2E9 Canada, Department of Fisheries and Oceans, Freshwater Institute, 501 University Crescent, Winnipeg, Manitoba, R3T 2N6 Canada, Water Chemistry Program, University of WisconsinsMadison, 660 North Park Street, Madison, Wisconsin 53706, Department of Microbiology, University of Manitoba, Winnipeg, Manitoba, R3T 2N2 Canada, and Environmental Sciences Division, Oak Ridge National Laboratory, P.O. Box 2008, Oak Ridge, Tennessee 37831-6038
The forest canopy was an important contributor to fluxes of methyl mercury (MeHg) and total mercury (THg) to the forest floor of boreal uplands and wetlands and potentially to downstream lakes, at the Experimental Lakes Area (ELA), northwestern Ontario. The estimated fluxes of MeHg and THg in throughfall plus litterfall below the forest canopy were 2 and 3 times greater than annual fluxes by direct wet deposition of MeHg (0.9 mg of MeHg ha-1) and THg (71 mg of THg ha-1). Almost all of the increased flux of MeHg and THg under the forest canopy occurred as litterfall (0.141.3 mg of MeHg ha-1 yr-1 and 110-220 mg of THg ha-1 yr-1). Throughfall added no MeHg and approximately 9 mg of THg ha-1 yr-1 to wet deposition at ELA, unlike in other regions of the world where atmospheric deposition was more heavily contaminated. These data suggest that dry deposition of Hg on foliage as an aerosol or reactive gaseous Hg (RGM) species is low at ELA, a finding supported by preliminary measurements of RGM there. Annual total deposition from throughfall and litterfall under a fireregenerated 19-yr-old jack pine/birch forest was 1.7 mg of MeHg ha-1 and 200 mg of THg ha-1. We found that average annual accumulation of MeHg and THg in the surficial litter/fungal layer of soils since the last forest fire varied between 0.6 and 1.6 mg of MeHg ha-1 and between 130 and 590 mg of THg ha-1 among sites differing in drainage and soil moisture. When soil Hg accumulation sites were matched with similar sites where litterfall and throughfall were collected, measured fluxes of THg to the forest floor (sources) were similar to our estimates of longterm soil accumulation rates (sinks), suggesting that the Hg in litterfall and throughfall is a new and not a recycled * Corresponding author phone: (780)492-9386; fax: (780)492-9234; e-mail:
[email protected]. † University of Alberta. ‡ Freshwater Institute. § University of WisconsinsMadison. | University of Manitoba. ⊥ Oak Ridge National Laboratory. 10.1021/es001924p CCC: $20.00 Published on Web 06/28/2001
2001 American Chemical Society
input of Hg to forested ecosystems. However, further research is required to determine the proportion of Hg in litterfall that is being biogeochemically recycled within forest and wetland ecosystems and, thus, does not represent new inputs to the forest ecosystem.
Introduction Litterfall is a major component of energy flow and nutrient recycling in terrestrial ecosystems (1). The elemental composition of litter comes not only from the composition of plant-derived substances but also from elements that are scavenged out of the atmosphere by the forest canopy (2). Thus, litterfall carries new inputs of elements from the atmosphere to the forest floor as well as recycled elements incorporated from the soil into plant structures. Not all of the scavenged elements reach the forest floor as litterfall; some portion is rinsed off during rain events and is found in throughfall (1). Throughfall is precipitation that passes through the forest canopy, and it may have higher concentrations of elements than precipitation in open areas (2). Both litterfall and throughfall, therefore, can function as additional sources of elements to ecosystems as well as routes of recycling from soil back to soil. Of particular interest to this study is the contribution of the forest canopy to the flux of mercury (Hg) in boreal forests and potentially to downstream lakes where it could accumulate to toxic levels in fish. Depending on where the Hg in the forest canopy originated, the Hg reaching the forest floor can be considered a new or recycled input (Figure 1). New inputs are ones that originate from the atmosphere and include deposition onto foliage as dry-particulate deposition (e.g., ref 3) or possibly by direct adsorption of reactive gaseous Hg (RGM) species to foliage (4) (Figure 1). Atmospheric Hg can also be assimilated as gaseous Hg(0) through leaf stomata (e.g., ref 5). In addition, Hg may also enter the forest canopy by recycling processes within a forest stand. For example, some of the Hg taken up through the leaf stomata can be Hg(0) that was released from the soils directly below the canopy through volatilization. Vascular plants transport Hg from soils and soil waters through their roots to foliage where it accumulates (e.g., refs 6 and 7), although many studies have found little uptake unless concentrations of Hg were high in soils (e.g., refs 8 and 9). Therefore, litterfall and throughfall potentially contribute to both the biogeochemical recycling of Hg within forest stands and the deposition of new Hg stripped from the atmosphere by various mechanisms. This study had two primary goals, both intended to improve our understanding of Hg in forested ecosystems. First, we compared the relative importance of wet deposition, throughfall, and litterfall as fluxes of Hg to the forest floor, and because our study site was in an area of relatively low atmospheric Hg deposition (10), we also compared these fluxes to previously published studies conducted in regions where atmospheric deposition was more heavily contaminated. Second, because we were able to determine rates of Hg accumulation in soils over a known period since a previous forest fire, we compared present-day fluxes of Hg to the forest floor via throughfall and litterfall to longer-term accumulation of Hg in soils and to rates of Hg export from forested watersheds to lakes.
Methods General Study Site. Our study was conducted at the Experimental Lakes Area (ELA) located in the boreal region VOL. 35, NO. 15, 2001 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
3089
FIGURE 1. Diagrammatic representation of Hg cycling in a forest canopy. of northwestern Ontario (Figure 2). We measured inputs of methylmercury (MeHg) and total mercury (THg) to a fireregenerated upland forest dominated by a dense stand of jack pine (Pinus banksiana) interspersed with birch (Betula papyrifera) and to a forested wetland that was dominated by an overstory of tamarack (Larix laricina) and black spruce (Picea mariana) trees, with an understory of leatherleaf (Chamedaphne calyculata), Labrador tea (Ledum groenlandicum), and alder (Alnus rugosa). The last forest fire passed through the ELA region in 1980. Further details of specific sampling sites are given in the text. Direct Wet Deposition of Mercury. We measured direct deposition of Hg in open areas at ELA. From 1998 to 1999, direct wet deposition was collected at the ELA meteorological site (Figure 2) and analyzed for MeHg and THg as described by St. Louis et al. (10). In general, wet deposition samples were collected in hot HNO3-washed 0.5-L wide-mouthed Teflon jars set out on acid-cleaned plastic trays mounted on wooden posts approximately 1.5 m above the ground. Collectors were set out just prior to or within 15 min of the beginning of a precipitation event. Just following the end of a precipitation event, samples were transferred slowly to acidcleaned 125-mL Teflon sample bottles and acidified with concentrated trace metal grade HCl equal to 2% of the sample volume. On two occasions, we collected bulk snow deposition samples by scooping up the upper layer of snow from the center of the frozen surface of Lake 239 (Figure 2) with clean 2-L Teflon jars. Stringent clean-hands, dirty-hands protocol was followed throughout the setup and collection process (11). All unfiltered samples were analyzed as in St. Louis et al. (10) at Flett Research Ltd. (Winnipeg, Manitoba) using cold 3090
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 15, 2001
vapor atomic fluorescence spectrophotometry (CVAFS) (1214). Direct wet deposition of MeHg and THg (mg ha-1 yr-1) was estimated by multiplying the overall volume-weighted mean concentration of MeHg and THg collected in wet deposition by the mean annual wet deposition of 721 mm for the years 1998 (668 mm) and 1999 (774 mm). Throughfall Fluxes of Hg. Throughfall samples were collected in 1998-1999 for MeHg and THg analyses from four upland jack pine forests sites interspersed with birch, all in close proximity at ELA (Figure 2). All four sites were burnt during the 1980 forest fire, resulting in a dense regeneration forest of similar age and composition. Three of the sites were 0.7-ha subcatchments subsequently flooded as part of the FLooded Upland Dynamics EXperiment (FLUDEX; 1998 preflood), whereas the fourth site used in 1999 was 0.5 km away at the ELA meteorological station (Figure 2). Duplicate collection stations were set up at each of the four sites by attaching wooden eavestrough holders to the trunks of trees. Just prior to or usually within 15 min of the beginning of a precipitation event, acid-washed Teflon-lined eavestroughing (12.5 cm wide by 75 cm long) was set out on the holders. Eavestroughing extended out under the canopy and far enough away from the small diameter trunks of the regenerating trees (average diameter at breast height: 3.2 ( 1.5 cm; 15) to prevent collection of stemflow. Throughfall drained into a hot HNO3-washed wide-mouthed 1-L Teflon jar through acid-washed 500-µm nytex screening to eliminate large particles. On three sampling occasions, the total throughfall volume exceeded that of the Teflon collection jar. For two of the events, the initial portion of the event was sampled prior to the jar overflowing, biasing toward high
FIGURE 2. Location of the litterfall collectors (sites U1-U6, W1-W3, and S1-S2), throughfall samplers, direct wet deposition samplers, soil samples, and upland catchment used to measure Hg export at the Experimental Lakes Area (ELA), northwestern Ontario. concentrations because we presumably sampled the initial scrubbing of the atmosphere and canopy and eliminated the dilute end of the event. For a third occasion, the jar was set back out after the initial portion of the event was subsampled. Samples were slowly transferred from the jars to acid-cleaned 125-mL Teflon sample bottles and acidified with concentrated trace metal grade HCl equal to 2% of the sample volume. Duplicate MeHg and THg samples were taken when sufficient deposition volume was available from a sample station. Unfiltered samples were analyzed at Flett Research Ltd. using CVAFS as described above for direct wet deposition. Again, clean-hands, dirty-hands protocol was followed. For each deposition event, we averaged concentrations of MeHg and THg in throughfall first for a given site and then for all sites. Throughfall deposition of MeHg and THg (mg ha-1 yr-1) was estimated by multiplying an overall volumeweighted mean concentration of MeHg and THg calculated
from all throughfall events in 1998-1999 by the mean annual throughfall volume. Throughfall volume was measured using standard rain gauges set out under the canopies and compared with volume collected in standard rain gauges stationed in open areas nearby. Throughfall volume at the sites was 55.4 ( 2.4% (ranging between 43% and 69%) of direct wet deposition due to interception by the dense canopy and evapotranspiration. This percent of direct wet deposition was very similar to the average of 59% of direct wet deposition previously measured by Allan et al. (16) over a 2-yr period during the May-October growing season under mature jack pine forests at ELA (52% in 1989 and 65% in 1990). We estimate that annual throughfall volume during the period of study was 400 mm or 55.4% of the 721 mm that fell in the open at the ELA meteorological site (Figure 2). It should be noted that throughfall samples as snow were not collected during the winter for concentration analyses, nor was throughfall volume measured in the winter when there may have been VOL. 35, NO. 15, 2001 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
3091
less interception by the forest canopy and evapotranspiration than during the growing season. Because concentrations of MeHg and THg in snow tend to be lower than in rain (this study; 10), possibly the overestimation of snow Hg concentrations balanced the underestimation of throughfall volume in winter. Regardless, during the 1998-1999 period that throughfall was collected, only 19% of the annual wet deposition occurred as snow. Litterfall Fluxes of Hg. Square litterfall collectors were constructed with an internal dimension of 30 cm by 30 cm using 4 cm by 20 cm untreated lumber. The frame of the collectors was cut 5.0 cm from the bottom, and a piece of 500-µm nitex screening was laminated between the upper and the lower portions of the frame. A pressboard bottom was attached to the frame to prevent contamination of the litter with soil once it had fallen. Holes to allow drainage of rain were drilled into the sides of the frame at the base of the collectors. We dug six collectors into the soils of upland forests and five collectors into the peatland of the basin wetland in early June 1995 for a total of 11 sites (Figure 2). Litter remained in the collectors only for a short period of time prior to collection to minimize leaching of Hg from the fallen tissues and accumulation of Hg on the litter from throughfall. Although litter fell onto the nitex screening, allowing for drainage of water, it was impossible to keep litterfall isolated from all moisture because litter often falls during precipitation events. However, there was very little litterfall prior to the senescence of shrubs and deciduous trees in the autumn of 1995, and samples were collected soon after in October of that year. Conifer needles and small tree branches that broke under the weight of snow fell directly onto the snow during winter where they remained frozen until they were deposited into collectors during spring melt. Litter was collected again shortly after spring melt in May 1996. At all times, cleanroom gloves were worn during collection of litter to prevent contamination. The litter was stored frozen at -20 °C in Ziploc bags. Litter was freeze-dried, and fall and spring samples from each site were completely combined to calculate an annual mass flux. The combined sample was homogenized in a stainless steel coffee grinder rinsed initially with deionized water and stored in Ziploc bags. Cleaned stainless steel coffee grinders were found not to contaminate samples with Hg, even when tissue masses were much smaller than those used in this study (e.g., ref 17). MeHg concentrations were determined by distillation extraction, aqueous-phase ethylation, and CVAFS detection [modification of Horvat et al. (14)] following a digestion at 95 °C of approximately 0.2 g of litter in a mixture of 45 mL of deionized water and 800 µL of 9 M H2SO4 saturated with KCl. Flett Research Ltd. performed all MeHg analyses. The average concentration (( SD) of 2 of the 11 samples run in either duplicate or triplicate was 0.23 ( 0.03 ng g-1. THg concentrations were determined by cold vapor atomic absorption spectrophotometry (CVAAS) following an overnight digestion of approximately 0.2 g of litter in 5 mL of 4:1 H2SO4:HNO3 at 185 °C and reduction to Hg(0) with (NH2OH)2H2SO4-SnCl2-NaCl (18). THg analyses were completed in the Hg laboratory at the Freshwater Institute, Winnipeg. Certified mercury reference solution for atomic absorption was used for standard series (0, 10, 25, and 50 ng of THg; r 2 ) 0.999). NRC dogfish muscle (Dorm-1) certified reference material was analyzed according to the same procedures as the samples. Recovery for reference material was 92% of the certified value. THg analyses were performed in duplicate. The overall average concentration (( SD) for all site samples was 42 ( 2.2 ng g-1. Annual fluxes of Hg via litterfall (mg ha-1 yr-1) were calculated by multiplying the 3092
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 15, 2001
dried mass of litter for each site by the mean concentration measured in the litter at the same site. Accumulation of Hg in Upland Forest Soils. Rates of MeHg and THg accumulated in the surficial litter/fungal layer of upland forest soils since the 1980 fire were estimated from 9 and 14 soil cores, respectively, collected from the same three 19-yr-old jack pine/birch forests that throughfall was collected in 1998 (Figure 2). Although the tree species composing the canopy of these three sites was similar, the sites varied widely in ground vegetation, soil moisture, and organic matter content due to different elevations and drainages among the subcatchments. Therefore, replicate cores were collected from zones of contrasting vegetation cover within each site to account for spatial heterogeneity. Soil profiles were sampled in May 1999 using a stainless steel barrel corer lined with acid-cleaned plastic sleeves. Cores were sectioned, and the upper organic surficial litter/fungal layer, which developed since the 1980 fire and was clearly delineated by a distinct charcoal layer below, was stored in a 10% HNO3-cleaned polypropylene cup and immediately frozen. Soil samples were lyophilized at -45 °C and homogenized using an acid-cleaned glass mortar and pestle prior to analyses for MeHg and THg. MeHg in soils was determined by distillation extraction, followed by aqueous-phase ethylation and CVAFS detection. Distillation consisted of a 0.10.3-g sample heated to 135 °C in 80 mL of deionized water, 1 mL of 8 N H2SO4, 1 mL of 1 M CuSO4, and 0.5 mL of saturated KCl for 6-7 h until approximately 85% of the liquid had been distilled. Triplicates were analyzed in 11% of samples, with a mean CV of 18%. MeHg spike recoveries averaged 100 ( 25% for 13 spiked samples. Total Hg was determined using CVAFS detection following microwave digestion in Teflon screw-cap bombs, which consisted of 0.1-0.5 g of soil in 10 mL of a 7:3 mixture of HNO3:H2SO4. A total of 2 mL of BrCl was added to each bomb, followed by 12 h of heating at 50 °C and dilution with 10 mL of deionized water. Triplicates were analyzed in 14% of the samples (mean CV of 13%), and the National Research Council Canada standard reference materials Tort-2 (lobster hepatopancreas) and Mess-2 (marine sediments) averaged 110% and 99% of their certified values, respectively. Total Hg spike recoveries averaged 102 ( 4% for 16 spiked samples. Annual accumulation rates of MeHg and THg in the surficial litter/fungal layer of soils (mg ha-1) were calculated by multiplying the concentration of Hg in the surficial litter/ fungal layer of the core by a mean soil bulk density for each sampling site and dividing by 19, the number of years since the last fire. Mean soil bulk densities were 0.99, 0.29, and 0.39 g cm-2 at the three sites, respectively (19). Our calculated soil accumulation rates excluded MeHg and THg stores in and below the charcoal layer remaining from the previous forest fire. We assumed that the Hg in those layers was deposited prior to new forest growth.
Results and Discussion Direct Wet Deposition of Hg. Mean concentrations of MeHg in direct wet deposition collected during 1998-1999 ranged between 0.02 and 0.17 ng L-1 (Figure 3), with an overall volume-weighted average of 0.12 ng L-1 (Table 1). Concentrations of THg ranged between 1.3 and 26 ng L-1 (Figure 3), with an overall volume-weighted average of 9.8 ng L-1 (Table 1). Volume-weighted average wet deposition concentrations measured in 1998-1999 were about double those measured in direct wet deposition between 1992 and 1995 at the same site using the same collection and analytical techniques (0.05 ng of MeHg L-1, 4.0 ng of THg L-1; 10). Thus, there may be a trend of increasing Hg deposition at ELA. St. Louis et al. (10) showed that precipitation events at ELA with high THg concentrations came from U.S. industrial regions to the
FIGURE 3. Concentrations of MeHg and THg in direct wet deposition (white bars) and throughfall (gray bars) collected under four 18-yr-old jack pine/birch forests sites at ELA in 1998-1999. Standard error bars are presented on mean throughfall concentrations measured for individual throughfall events at the three FLUDEX sites in 1998 (Figure 1).
TABLE 1. Mean Annual Deposition Volume in the Open and under a 19-yr-old Jack Pine/Birch Forest Canopy, and Mean (( SD) Concentration and Annual Flux of MeHg and THg in Direct Open Deposition and Throughfall
direct open throughfall
concentration
mean annual deposition (1998-1999) (mm)
MeHg (ng L-1)
721 400
0.12 ( 0.06 0.22 ( 0.11
southeast, while events with high concentrations of MeHg originated to the west of ELA. Western sources of MeHg to precipitation were not known (10). Fluxes of Hg in direct wet deposition were estimated using the average volume-weighted concentrations calculated from all samples taken in 1998-1999 and the volume of wet deposition estimated for the same 2-yr period. Fluxes were on average 0.9 mg of MeHg ha-1 yr-1 and 71 mg of THg ha-1 yr-1 (Table 1). Throughfall Fluxes of Hg. Mean concentrations of MeHg in throughfall collected in 1998-1999 under upland jack pine/ birch forests ranged between 0.06 and 0.43 ng L-1 (Figure 3), with an overall volume-weighted average of 0.22 ng L-1 (Table 1). Mean concentrations of THg ranged between 4.2 and 42 ng L-1 (Figure 3), with an overall volume-weighted average of 20 ng L-1 (Table 1). During 1998-1999, for those events when wet deposition and throughfall were collected coincidentally, average volumeweighted concentrations of MeHg (0.22 ng L-1, n ) 9) and THg (20 ng L-1, n ) 10) in throughfall were 2.2 and 2.0 times higher than average volume-weighted concentrations of MeHg (0.10 ng L-1) and THg (9.8 ng L-1) in precipitation collected in the open. Concentrations of MeHg and THg in net throughfall (concentrations in throughfall minus concentrations measured in direct wet deposition) were not
n
THg (ng L-1)
18 11
10 ( 7 20 ( 12
annual Hg flux
n
MeHg (mg ha-1)
THg (mg ha-1)
20 10
0.9 0.9
71 80
significantly correlated with concentrations in wet deposition measured in the open (Figure 4), suggesting that something other than direct wet deposition drives fluxes of Hg in throughfall, at least for THg. When the one outlier was removed from the MeHg regression analysis (Figure 4), there was a significant relationship between concentrations of MeHg in direct wet deposition and net throughfall (r 2 ) 0.5, df ) 7, p ) 0.05). The fire-regenerated forests where throughfall inputs were measured were extremely dense (24 500 ( 17 000 stems ha-1) (15). This density was similar to that measured in young jack pine forests in the boreal region near Thompson and Nelson House, Manitoba (5700-42 000 stems ha-1) but much larger than that measured on mature jack pine forests in the same area (1300-3500 stems ha-1) (20). The high canopy density most likely contributed to the low average throughfall volume measured at our sites (55% of that measured in the open) as compared to that measured under an old-growth forest at ELA (59%; 16) and elsewhere (e.g., 65-85% of precipitation under a range of mature coniferous and deciduous tree species at the Acadia Forest Experiment Station, New Brunswick, Canada; 21). Fluxes of Hg in throughfall were 0.9 mg of MeHg ha-1 yr-1 and 80 mg of THg ha-1 yr-1 (Table 1). These fluxes of MeHg and THg were respectively comparable to or slightly higher VOL. 35, NO. 15, 2001 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
3093
FIGURE 4. Relationship of MeHg and THg measured in wet deposition measured in the open with that measured in net throughfall. MeHg regression analysis: r 2 ) 0.06, df ) 8, p ) 0.5; THg regression analysis: r 2 ) 0.15, df ) 9, p ) 0.3.
TABLE 2. Concentrations and Annual (1995-1996) Mass Flux of MeHg and THg in Litterfall to Various Boreal Forest Uplands and Wetlands at the Experimental Lakes Areaa concentration
annual Hg flux
litter collection site
MeHg (ng g-1)
THg (ng g-1)
% MeHg
litterfall annual mass flux (kg ha-1)
MeHg (mg ha-1)
THg (mg ha-1)
Upland Forests (Age) U1, jack pine (old growth, >75 yr) U2, balsam fir stand (22 yr) U3, jack pine/birch/Alnus crispa (22 yr) U4, jack pine/birch (22 yr) U5, jack pine/birch (15 yr) U6, jack pine (15 yr) average ( SD
0.24 0.06 0.19 0.15 0.06 0.40 0.18 ( 0.12
79 53 29 30 25 33 42 ( 19
0.3 0.1 0.7 0.5 0.2 1.2 0.5 ( 0.4
2730 2410 4670 3510 4740 3270 3550 ( 890
0.66 0.14 0.89 0.51 0.26 1.30 0.63 ( 0.39
220 130 130 110 120 110 140 ( 40
Wetland Trees W1, tamarack/spruce W2, tamarack/spruce W3, tamarack/spruce average ( SD
0.50 0.26 0.39 0.38 ( 0.10
69 48 35 51 ( 14
0.7 0.6 1.1 0.8 ( 0.2
1250 1350 1840 1480 ( 260
0.62 0.35 0.72 0.56 ( 0.16
86 64 65 72 ( 10
Wetland Shrubs S1, leatherleaf S2, Alnus rugosa average ( SD
0.35 0.55 0.45 ( 0.10
30 34 32 ( 1.7
1.2 1.6 1.4 ( 0.2
940 4340 2640 ( 1700
0.33 2.40 1.36 ( 1.03
29 150 88 ( 59
a
The site listed describes the plant species that primarily contributed to the litterfall sample collected. Please see Figure 2 for collector locations.
than those in direct wet deposition, suggesting that dry deposition of mercury at ELA is relatively low. The relatively low THg fluxes of throughfall at ELA (a 15% enhancement) are different than measurements made at other more contaminated sites (see below). Lindberg et al. (22) have suggested that net throughfall is a good estimate of dry deposition of mercury to ecosystems. Furthermore, Rea et al. (23) found using foliar washing experiments that dry deposition readily washes off foliage. Preliminary measurements at ELA during August 2000 found mean daily RGM concentrations in the range of 1-2 pg m-3 (24), very low in comparison to concentrations measured in more industrialized regions of the United States (30-50 pg m-3; 25). We can crudely estimate an RGM dry deposition flux using these measurements, expected average RGM dry deposition velocities (26), and assuming a leaf area index (LAI) of about 3 under a forest canopy at ELA (24). These calculations suggest that dry deposition of RGM at ELA could be on the order of about 4-8 mg of THg ha-1 yr-1, close to the 9 mg of THg ha-1 yr-1 dry deposition that we estimate from our net throughfall measurements. Litterfall Fluxes of Hg. Concentrations of MeHg were on average highest in the litter from plants growing in the wetland (Table 2). Litter from woody plants in dry upland forests had lower MeHg concentrations than litter from wetland trees and wetland shrubs (Table 2). These results correspond well with observations that low-lying, wetland areas of watersheds tend to produce MeHg, while drier upland 3094
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 15, 2001
areas do not (11, 27, 28). Thus, MeHg produced in the wetland appears to be taken up by plants growing there (6, 7), although this cannot be unequivocally demonstrated by these data. The pattern of concentrations of THg in the litter types was different from MeHg. Litter from old-growth jack pine contained the highest concentrations of THg, followed by litter containing needles from tamarack, balsam fir, and spruce trees (Table 2). As a result, the percent Hg in tissues that was MeHg increased from on average 0.5% in litter from upland forests to on average 1.4% in wetland shrubs (Table 2). Where does the Hg in plant foliage originate? Rasmussen (29) found that THg concentration in needles from balsam fir and spruce increased by about 5-10 ng g-1 yr-1 over a 3-yr period, suggesting that Hg readily accumulates in plant foliage over time. Vascular plants transport Hg from soils and soil waters through their roots to foliage where it accumulates (e.g., refs 6 and 7). Atmospheric Hg can also be taken up as Hg(0) through leaf stomata (e.g., ref 4). Leaves and needles may scavenge Hg from the atmosphere as dry particulate Hg (e.g., ref 3) or RGM species (e.g., ref 4). Some of our highest concentrations of THg were found in litter composed of conifer needles, which have higher surface area to volume ratios than deciduous leaves. However, all or a large portion of this Hg is believed to wash off foliage surfaces in throughfall (23), indicating that most of the Hg is incorporated in foliage via either root or foliar uptake.
TABLE 3. Average Annual Direct Wet Deposition, Litterfall, and Throughfall of MeHg and THg at the Experimental Lakes Area (This Study) and Other Sites in the United States and Scandinavia MeHg (mg ha-1 yr-1)
direct direct wet through- litter- canopy total wet through- litter- canopy total deposn fall fall (% wet deposn) deposn fall fall (% wet deposn) ref
site This Study ELAa
0.9
0.9
0.8
1.7 (190%)
United States Marcell Experimental Forest, MNb Lake Champlain Basin, VTc Walker Branch Watershed, TNd Scandinavia Lake Gardsjon Watershed, S. Swedene Svartberget Research Forest, N. Swedenf Paronikorpi Forest, Finlandf a
Upland sites U5 and U6; Table 1.
THg (mg ha-1 yr-1)
b
1.2 0.8 1.0
1.6 1.7
6.0 3.0 6.4
Spruce/aspen/birch. c Mixed hardwood.
Litterfall biomasses were much greater in upland forested areas than in forested wetland areas (Table 2). Thus, flux of both MeHg and THg via litterfall was driven primarily by the annual biomass flux of litter, despite differences in concentrations of MeHg and THg among the different litter types (Table 2). Average annual flux of MeHg under upland and wetland trees was 70% and 62%, respectively, of the 0.9 mg ha-1 measured in direct wet deposition in open areas at ELA. Litterfall fluxes of MeHg under Alnus rugosa were, however, 2.7 times higher than the average annual direct wet deposition due primarily to the large biomass flux. Flux of THg via litterfall was either similar to (72 mg of THg ha-1 yr-1 under wetland trees) or on average higher than the 71 mg of THg ha-1 yr-1 flux measured in direct wet deposition (Table 2). For example, annual litterfall inputs of 220 mg of THg ha-1 to the old-growth jack pine forest floor (site U1; Table 2) were 3.1 times higher than the average annual direct wet deposition. Total Flux of Hg under the Forest Canopy. Our data clearly show that the forest canopy provides, primarily via litterfall, an important additional flux of both MeHg and THg to forest ecosystems. In fact, total flux of MeHg and THg under a standard 15-19-yr-old jack pine/birch forest canopy at ELA is 1.9 and 2.9 times the flux of MeHg and THg in wet deposition in the open (Table 3). On the basis of the very small net amounts of Hg seen in throughfall at ELA as compared to other more polluted sites, the Hg content in throughfall appears to be related to the average Hg concentrations in the local atmosphere. For example, during the 1990s, rates of direct wet deposition of MeHg measured in open areas at the Lake Gardsjon Watershed in southern Sweden were about 1.3-fold higher than at ELA, and THg deposition was about 1.6 times higher (Table 3; 32). The combined throughfall/litterfall fluxes of MeHg and THg at the Lake Gardsjon Watershed were also higher (by 4.5 and 2.3 times, respectively) than in the 19yr-old jack pine/birch forest at ELA (Table 3; 32). At the Svartberget Research Forest, in a more remote northern region of Sweden where MeHg and THg fluxes in the open were similar to those at ELA, combined throughfall/litterfall fluxes of MeHg and THg were still 2.8 and 1.5 times higher, respectively, than at ELA (Table 3; 32). In other regions of Scandinavia and in the United States that are more industrialized than at ELA, deposition of THg was 1.5-2 times higher in throughfall and up to 5 times higher in litterfall than that at ELA, even when direct deposition of THg in the open was similar to or lower than that measured at ELA (Table 3; 4, 30, 31, 33). Thus, areas with greater atmospheric pollution can be expected to have greater fluxes of Hg to soils via throughfall and litterfall.
7.6 (630%) 4.7 (590%) d
70
80
120
200 (290%)
65 79 100
130 117 140
123 130 300
253 (390%) 247 (310%) 440 (440%)
30 31 4
100 70 51
230 150
230 180 595
460 (460%) 330 (470%)
32 32 33
Oak/hickory/pine. e Spruce. f Pine/spruce.
Is the Hg in Throughfall and Litterfall a New or Recycled Input to Ecosystems? One approach to evaluate whether MeHg and THg in throughfall and litterfall might represent new inputs (atmosphere-derived) to the ecosystem is to see if wet deposition by itself is a large enough source to account for the sinks of Hg. These sinks include the accumulation in the surficial litter/fungal layer of soils, the export of MeHg and THg in runoff, and the volatilization of Hg via Hg(II) reduction and MeHg demethylation in soils. We have developed a conceptual model of Hg fluxes in a 19-yr-old boreal jack pine/birch forest. This model uses throughfall inputs measured under four 18-19-yr-old jack pine/birch forest sites (Table 1, Figure 2) and litterfall fluxes at the 15-yr-old jack pine/birch forests (U5 and U6) (Table 2, Figure 2). These litterfall sites were chosen because they are in the same burned area where throughfall and soil accumulation rates were measured. However, litterfall fluxes measured at these burned sites were similar to fluxes measured at the majority of the other upland forest sites (Table 2). Export of Hg in runoff was quantified at a purely upland catchment with no wetland areas (Figure 2). The largest sink term was the rate of accumulation in soils, and this rate differed among the upland sites with varied ground vegetation, soil moisture, and organic matter content due to elevation and drainage (Table 4). The average annual accumulation rates of THg in the surficial litter/fungal layer of soils was 590, 200, and 130 mg ha-1 at sites 1-3, respectively, suggesting that small-scale heterogeneity occurs in inputs, the subsequent biogeochemical processing of the litter, and/or storage of Hg. By contrast, annual accumulation rates of MeHg in the surficial litter/fungal layer of soils were smaller and more uniform at the three sites, averaging only 1.6, 0.6, and 0.9 mg ha-1. The high measured accumulation rate of MeHg in site 1 may be due to Hg methylation in low-lying saturated regions of the subcatchment dominated by Sphagnum spp. (34). The overall average annual accumulation rates (( SD) of MeHg and THg in the surficial litter/fungal layer of soils at the sites were 1.0 ( 0.5 mg of MeHg ha-1 and 320 ( 230 mg of THg ha-1 (Table 4). To estimate an annual sink of Hg in a forested upland catchment, we used annual Hg accumulation rates of 0.6 mg of MeHg ha-1 and 200 mg of THg ha-1 for soils at site 2 because its ground cover was most similar to that at the sites where litterfall was collected. St. Louis et al. (27) calculated average (1990-1993) exports for a purely upland catchment at ELA to be 0.1 ( 0.01 mg of MeHg ha-1 and 16 ( 4.5 mg of THg ha-1 annually (Table 5). Thus, the total estimated annual sink in the forest soils [excluding MeHg demethylation and Hg(0) evasion] was 0.7 mg of MeHg ha-1 and 216 mg of THg ha-1 (Table 5). When the sources (inputs) of MeHg are VOL. 35, NO. 15, 2001 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
3095
TABLE 4. Mean Concentration (( SD) of MeHg and THg in the Surficial Litter/Fungal Layer of Soil Cores Collected from Three Sites at ELA (Figure 2), the Areal Mass of Hg in the Layer, and the Annual Accumulation Rate of Hg in the Layer over a 19-yr Period Since the Last Fire concentration soil collection sitea
MeHg (ng g-1)
n
THg (ng g-1)
site 1, wet soils with areas of Sphagnum 0.3 ( 0.1 3 110 ( 34 spp. growth site 2, dry soils primarily covered with 0.4 ( 0.1 3 130 ( 18 jack pine and birch litter site 3, thin, well-drained ridgetop soils 0.5 ( 0.1 3 61 ( 7 all sites combined a
Hg areal mass
annual Hg accumulation
n
surface soil bulk density (g dw cm-2)
5
0.99
0.3 ( 0.01 110 ( 34 1.6 ( 0.2
590 ( 180
5
0.29
0.1 ( 0.02
38 ( 5
0.6 ( 0.1
200 ( 28
4
0.39
0.2 ( 0.05
24 ( 3
0.9 ( 0.3
130 ( 14
0.2 ( 0.1
60 ( 44 1.0 ( 0.5
0.4 ( 0.1 9 104 ( 37 14
MeHg (ng cm-2)
THg (ng cm-2)
MeHg (mg ha-1)
THg (mg ha-1)
320 ( 230
Also see Figure 2.
TABLE 5. Comparison of Measured Rates of MeHg and THg Accumulation and Accumulation Rates Estimated from the Difference between Hg Inputs and Hg Export from Various ELA Watershedsa MeHg (mg ha-1 yr-1) source: average flux in throughfall under 18/19-yr-old jack pine/birch forests in 1998/99 (Table 1) source: average flux rates in litterfall from 15-yr-old jack pine/birch forests (sites U5 and U6; Table 2) in 1995 total sources sink: average export from an upland catchment at ELA (27) sink: measured accumulation in the surficial litter/fungal layer of dry upland forest soils primarily covered with jack pine and birch litter (site 2; Table 4) sink: estimated soil evasion from upland soils at ELA (24) total sinks sinks minus sources a
THg (mg ha-1 yr-1)
0.9
80
0.8
120
1.7
200
0.1 0.6
16 200
0.7
10-20 230-240
-1.0
30-40
Sources represent inputs measured from 1995 to 1999, and sinks represent the mean soil accumulation rate from 1981 to 1998.
compared with the sinks (the mean soil accumulation rate), we calculated that litterfall and throughfall can more than explain all the MeHg accumulated in soils and exported (Table 5). In fact, it appears that approximately 1 mg of MeHg ha-1 is annually lost from forest soils, possibly through demethylation. MeHg is more difficult to interpret in this type of analysis than THg because it can be produced and subsequently demethylated within soils. Annual fluxes of THg in direct wet deposition were approximately 30% of the combined sinks for THg (Table 5), making it likely that throughfall and litterfall inputs are at least partly new inputs that are necessary to explain the accumulation of Hg in the surficial litter/fungal layer of soils and the export in runoff. If these inputs are considered to be completely new inputs to the upper soil layer that accumulated since the last forest fire, we calculate that accumulation rates and runoff are very similar to the estimated inputs from throughfall and litterfall. Our calculations suggest an annual input deficit of only about 20 mg of THg ha-1 is needed to sufficiently explain the accumulation of THg in the surficial litter/fungal layer of soils (Table 5); however, this deficit would be larger if we include Hg(0) evasion as a sink of THg (35). Detailed soil evasion studies have been published for temperate forests (36) and for boreal nonforested soils (37). Mean summertime Hg evasion rates at these sites ranged from about 3 to 7 ng m-2 h-1, and the annual Hg evasion at the temperate site was estimated to be 100 mg ha-1 (38). We can make a rough estimate of soil evasion at ELA from recent, limited measurements (using dynamic flux chambers operated at ∼1 L/min flow rates for 0.5-2 h, collecting atmospheric Hg by standard gold-trap methods), which yielded a summertime average Hg evasion rate of about 1-2 ng m-2 h-1 over these boreal soils (24). These data and seasonal soil temperatures at ELA suggest an annual Hg evasion around 10-20 mg ha-1 (Table 5). 3096
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 15, 2001
If the deficit is real, what are other possible inputs to soils not measured here? Fluxes of Hg in stemflow could contribute to some of the Hg deficit. Kolka et al. (39), for example, found that in a black spruce, aspen (Populus tremuloides), and birch watershed, stemflow contributed about 4% additional THg to that measured in throughfall alone. If we assume an equivalent percentage flux in our jack pine/birch forest via stemflow, this would increase our annual inputs by 3 mg of THg ha-1, reducing our deficit by only a small amount. Hg could possibly also come from below the soil organic horizon. However, concentrations of MeHg and THg in the charcoal layer of the soils were approximately half the concentrations found in the overlying surficial litter/fungal layer (40). These data suggest that elevated relative concentrations in the surface litter layers are not due to diffusion of Hg in soil water from deeper charcoal and mineral layers where Hg(II) is probably tightly bound. Weathering of Hg from bedrock could also contribute to some of the Hg deficit. We estimated this potential contribution at ELA using the net export of two minerals (Si and Mg) whose primary source is weathering in the watershed, and assuming that Hg is weathered from bedrock at the same rate as Si and Mg, the ratio of Hg to each mineral in bedrock at ELA. The concentration of Hg in granite was found to range between 7 and 200 ng g-1, with a median value of 62 ng g-1 (41). Concentrations of Si and Mg in bedrock at ELA are 337 and 3 mg g-1, respectively (42). The average annual net export of Si and Mg from four upland forest catchments at ELA in 1988-1990 was 5.2 and 0.14 kg ha-1, respectively (16). Using the ratio of Hg to Si and Hg to Mg in bedrock, we estimated average annual weathering rates of Hg to be approximately 1 and 3 mg ha-1, respectively. These weathering rates are therefore also small sources of Hg to soil runoff. These considerations leave two possible additional Hg sources to these soils that could explain the higher sink
strength. One possible source of Hg is direct soil uptake (deposition) of Hg volatilized from nearby forests via fires (43). On the long-term in the boreal region of Canada, 0.4% of the forest area is burned annually (44). A more likely explanation is a recent decrease in atmospheric deposition to these soils. If regional Hg fluxes were higher in the early 1980s after the fire at ELA, then the 19-yr mean soil accumulation rate would reflect these inputs. For example, the Hudson Bay Mining and Smelting Company Ltd. in Flin Flon, Manitoba, which is approximately 700 km northwest of ELA, accounted for 30% of all anthropogenic Hg emissions in Canada until a few years ago. These emissions have now been reduced by over 50% due to a change in the refining process (45). Since these potential sources were in addition to the throughfall and litterfall deposition rates we measured from 1995 to 1999, we may not expect a better mass balance than computed here. Certainly, the deficit of 30-40 mg of THg ha-1 is within the errors of our different source and sink estimates. In either case, our data support an external source for Hg in throughfall and litterfall. Implications of Hg Deposition via Throughfall and Litterfall. Our study has several implications for the biogeochemical cycling of Hg in boreal forests. The upland and wetland forest canopies provided a substantial flux of MeHg and THg via litterfall when compared with inputs from direct wet deposition but a relatively small flux in throughfall. This suggests that the contribution from dry deposition is much lower than at other more polluted sites. We suspect that this THg represents a new input to the forest floor. Because Hg deposited on foliage as particulate Hg or as RGM species is most likely rinsed off in throughfall, the THg accumulated in foliage (and hence litterfall) must come via metabolic processes (e.g., soil or direct foliar uptake). If the Hg in litterfall is a new input to ecosystems, then the Hg must originate from stomatal uptake of Hg(0) in the atmosphere (e.g., ref 5). If the Hg is being recycled, the Hg must come from stomatal uptake of Hg(0) emitted from soils below the canopy and/or via root uptake of Hg in mineral soils below the surficial litter/fungal layer to account for the estimated soil accumulation rate of THg. However, these pathways would further decrease the net soil input and increase the difference in our source/sink mass balance, suggesting that the flux in litterfall is external. If all the Hg in litterfall is new, this would have important implications for the proposed imposition of critical loads of Hg to ecosystems, which now only consider direct wet deposition inputs to ecosystem. It is important to resolve this issue. Further studies are required to separate the “new” proportion of MeHg and THg in litterfall and throughfall from the proportion of MeHg and THg that is being internally recycled. These studies should consider controlled laboratory studies and field isotope addition experiments in which stable isotopes of MeHg and THg are added to either the canopy of a forest or forest floor soils, and subsequently followed through plants to determine the origin of the majority of Hg in the canopy. One such study is now underway at ELA (46).
Acknowledgments We are indebted to several people who made important contributions to this study. Paul Humenchuk helped build and install litterfall collectors. Jennifer Shay identified Alnus spp. Robert Flett and Danesh Kannangara analyzed litter for MeHg. Allan Wiens helped analyze litter for THg. Ken Beaty provided wet deposition volumes from ELA meteorological site. Steve Brooks assisted with RGM measurements. We thank Manitoba Hydro, Ontario Hydro, Hydro Quebec, Department of Fisheries and Oceans Canada, the U.S. Department of Energy, and NSERC for financial support. This paper is Contribution No. 47 of the Experimental Lakes Area Reservoir Project (ELARP), Contribution No. 1 of the
FLooded Upland Dynamics EXperiment (FLUDEX), and Contribution No. 1 of the Mercury Experiment to Assess Atmospheric Loadings in Canada and the United States (METAALICUS).
Literature Cited (1) Schlesinger, W. H. Biogeochemistry: An Analysis of Global Change, 2nd ed.; Academic Press: San Diego, 1997. (2) Trudgill, S. T. Soil and Vegetation Systems, 2nd ed.; Oxford University Press: Oxford, 1988. (3) Iverfeldt, A. Water Air Soil Pollut. 1991, 56, 553-564. (4) Lindberg, S. E. In Global and Regional Mercury Cycles: Sources, Fluxes and Mass Balances; Baeyens, W., Ebinghaus, R., Vasliliev, O., Eds.; NATO-ASI Series 21; Kluwer Academic Publishers: Dordrecht, The Netherlands, 1996; pp 359-380. (5) Hanson, P. J.; Lindberg, S. E.; Tabberer, T. A.; Owens, J. G.; Kim, K.-H. Water Air Soil Pollut. 1995, 80, 373-382. (6) Godbold, D. L.; Huttermann, A. Physiol. Plant. 1988, 74, 270275. (7) Bishop, K. H.; Lee, Y.-H.; Munthe, J.; Dambrine, E. Biogeochemistry 1998, 40, 101-113. (8) Lindberg, S. E.; Jackson, D. R.; Huckabee, J. W.; Janzen, S. A.; Levin, M. J.; Lund, J. R. J. Environ. Qual. 1979, 8, 572-578. (9) Cocking, D.; Rohrer, M.; Thomas, R.; Walker, J.; Ward, D. Water Air Soil Pollut. 1995, 80, 1113-1116. (10) St. Louis, V. L.; Rudd, J. W. M.; Kelly, C. A.; Barrie, L. A. Water Air Soil Pollut. 1995, 80, 405-414. (11) St. Louis, V. L.; Rudd, J. W. M.; Kelly, C. A.; Beaty, K. G.; Flett, R. J.; Roulet, N. T. Environ. Sci. Technol. 1996, 30, 2719-2729. (12) Bloom, N. S.; Crecelius, E. A. Mar. Chem. 1987, 14, 49-59. (13) Bloom, N. S. Can. J. Fish. Aquat. Sci. 1989, 46, 1131-1140. (14) Horvat, M.; Bloom, N. S.; Liang, L. Anal. Chim. Acta 1993, 281, 135-152. (15) Huebert, D. Experimental Lakes Area Upland Flooding Experiment Vegetation Analysis. Winnipeg, Manitoba, 2000, unpublished report. (16) Allan, C. J.; Roulet, N. T.; Hill, A. R. Biogeochemistry 1993, 22, 37-79. (17) Hall, B. D.; Rosenberg, D. M.; Wiens, A. Can. J. Fish. Aquat. Sci. 1998, 55, 2036-2047. (18) Malley, D. F.; Stewart, A. R.; Hall, B. D. Environ. Toxicol. Chem. 1996, 15, 928-936. (19) Schiff, S. University of Waterloo, Ontario, Canada, unpublished results. (20) Chen, J. M.; Rich, P. M.; Gower, S. T.; Norman, J. M.; Plummer, S. J. Geophys. Res. 1997, 102, 29, 429-443. (21) Mahendrappa, M. K. For. Ecol. Manage. 1990, 30, 65-72. (22) Lindberg, S. E.; Owens, J. C.; Stratton, W. J. In Mercury as a Global Pollutant; Watras, C. J., Huckabee, J. W., Eds.; Lewis Publishers: Chelsea, MI, 1994; pp 261-272. (23) Rea, A. W.; Lindberg, S. E.; Keeler, G. J. Environ. Sci. Technol. 2000, 34, 2418-2425. (24) Lindberg, S. E. Oak Ridge National Laboratory, Oak Ridge, TN, unpublished results. (25) Stratton, W. J.; Lindberg, S. E. Water Air Soil Pollut. 1995, 80, 1269-1278. (26) Lindberg, S. E.; Stratton, W. J. Environ. Sci. Technol. 1998, 32, 49-57. (27) St. Louis, V. L.; Rudd, J. W. M.; Kelly, C. A.; Beaty, K. G.; Bloom, N. S.; Flett, R. J. Can. J. Fish. Aquat. Sci. 1994, 51, 1065-1076. (28) Hurley, J. P.; Benoit, J. M.; Babiarz, C. L.; Shafer, M. M.; Andren, A. W.; Sullivan, J. R.; Hammond, R.; Webb, D. E. Environ. Sci. Technol. 1995, 29, 1867-1875. (29) Rasmussen, P. E. Water Air Soil Pollut. 1995, 80, 1039-1042. (30) Grigal, D. F.; Kolka, R. K.; Fleck, J. A.; Nater, E. A. Biogeochemistry 2000, 50, 95-109. (31) Rea, A. W.; Keeler, G. J.; Scherbatskoy, T. Atmos. Environ. 1996, 30, 3257-3263. (32) Lee, Y.-H.; Bishop, K. H.; Munthe, J. Sci. Total Environ. 2000, 260, 11-20. (33) Lee, Y.-H.; Bishop, K. H.; Munthe, J.; Iverfeldt, A.; Verta, M.; Parkman, H.; Hultberg, H. Biogeochemistry 1998, 40, 125-135. (34) Allan, C. J.; Heyes, A.; Roulet, N. T.; St. Louis, V. L.; Rudd, J. W. M. Biogeochemistry 2001, 52, 13-40. (35) Lindberg, S. E.; Kim, K.-H.; Munthe, J. Water Air Soil Pollut. 1995, 80, 383-392. (36) Kim, K.-H.; Lindberg, S. E.; Meyers, T. P. Atmos. Environ. 1995, 27, 267-282. (37) Poissant, L.; Casimir, A. Atmos. Environ. 1998, 32, 883-893. (38) Lindberg, S. E.; Hanson, P. J.; Meyers T. P.; Kim, K.-H. Atmos. Environ. 1998, 32, 895-908. VOL. 35, NO. 15, 2001 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
3097
(39) Kolka, R. K.; Nater, E. A.; Grigal, D. F.; Verry, E. S. Water Air Soil Pollut. 1999, 113, 273-294. (40) Rolfhus, K. R. University of WisconsinsMadison, Madison, WI, unpublished results. (41) Jonasson, I. R.; Boyle, R. W. In Proceedings of the Special Symposium on Mercury in Man’s Environment, February 1516, 1971; Royal Society of Canada: 1971; pp 5-21. (42) Brunskill, G. J.; Povoledo, D.; Graham, B. W.; Stainton, M. P. J. Fish. Res. Board Can. 1971, 28, 277-294. (43) Veiga, M. M.; Meech, J. A.; On ˜ ate, N. Nature 1994, 368, 816817. (44) Weber, M. G.; Stocks, B. J. Ambio 1998, 27, 545-550.
3098
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 15, 2001
(45) Environment Canada. http://www.ec.gc.ca/pp/english/stories/ hbmsstoe.html, 2000 (accessed November 26, 2000). (46) Mercury Experiment to Assess Atmospheric Loading in Canada and the United States (METAALICUS). http:// www.biology.ualberta.ca/metaalicus/metaalicus.htm, 2000 (accessed November 26, 2000).
Received for review November 30, 2000. Revised manuscript received May 7, 2001. Accepted May 8, 2001. ES001924P
