ARTICLE pubs.acs.org/est
Influence of Forest Canopies on the Deposition of Methylmercury to Boreal Ecosystem Watersheds Linnea D. Mowat,† Vincent L. St. Louis,*,† Jennifer A. Graydon,† and Igor Lehnherr† †
Department of Biological Sciences, University of Alberta, Edmonton, Alberta Canada T6G 2E9
bS Supporting Information ABSTRACT: Although it has been previously shown that forest canopies significantly increase the total deposition of Hg to watersheds, sources and fates of atmospherically deposited MeHg in particular remain poorly understood. In this study, net loadings of MeHg to a watershed were quantified, and the retention and (photo)reduction of MeHg on foliage were measured using unique stable Hg isotope experiments. Annual loadings of MeHg in throughfall (0.34 ( 0.01 to 0.60 ( 0.16 mg ha1 yr1) and litterfall (0.77 ( 0.07 to 0.97 ( 0.34 mg ha1 yr1) were collectively 34 times higher under different forest canopies than loadings of MeHg in the open (0.41 mg ha1 yr1), suggesting dry deposition of MeHg to forest canopies. Using Me199Hg, we found that a portion of MeHg wet deposited to forest canopies is retained on foliage over time, eventually contributing to MeHg in litterfall. Average half-lives (t1/2) of Me199Hg on spruce, jack pine, and birch foliage were 204 ( 66, 187 ( 101, and 8 ( 3 days, respectively. We also found using Me199Hg that following wet deposition, MeHg is rapidly (photo)reduced to 199Hg(0) on canopy foliage, which then evades to the atmosphere. We were unable to quantify concentrations of particulate-bound MeHg (p-MeHg) in the air using vacuum pumps and quartz microfiber air sampling filters, despite the possibility that p-MeHg does exist in small quantities. As a result, the source of dry deposited MeHg remains partially elusive.
’ INTRODUCTION Methylmercury (MeHg) is a potent vertebrate neurotoxin and a contaminant of global concern. Several recent studies, including the whole-ecosystem Mercury Experiment To Assess Atmospheric Loading In Canada and the U.S. (METAALICUS) at the Experimental Lakes Area (ELA) in northwestern Ontario, Canada, have demonstrated that increased anthropogenic loading of mercury (Hg) to watersheds increases bioaccumulation of MeHg in top predatory organisms such as fish.1 Whereas deposition of inorganic Hg(II) and subsequent microbial methylation in anaerobic lake sediments 2 and wetlands 3 is likely the primary source of MeHg to aquatic food webs,4 direct atmospheric deposition of MeHg to watersheds might also be an important source of MeHg in fish.5,6 To determine the relative importance of each source in contributing to MeHg concentrations in fish, and to accurately mass-balance MeHg fluxes for a given lake, it is important to understand and consider all internal and external sources of MeHg to the water column. Atmospheric deposition of MeHg to the watershed includes not only MeHg directly deposited to the lake surface but also deposition to the surrounding terrestrial compartments as well. Terrestrial compartments of watersheds typically receive higher atmospheric Hg loadings than do their aquatic counterparts due to the large surface area available for efficient scavenging of atmospheric Hg by foliage in forest canopies.79 In fact, many studies have shown that Hg deposition under forest canopies is much higher than that in the open.7,913 The dry deposition of Hg to forest canopies is often determined by subtracting wet deposition loadings in open areas from the sum of throughfall r 2011 American Chemical Society
(precipitation passing through the canopy) and litterfall (senescent foliage) loadings.14 However, recent research demonstrating photoreduction and emission to the atmosphere of a portion of newly wet-deposited Hg from foliage suggests that this mathematical approach underestimates the contribution of the canopy to total dry loadings of Hg to watersheds.15 Whereas forest canopies significantly increase the total deposition of Hg to watersheds, sources and fates of atmospherically deposited MeHg in particular remain poorly understood.16 Measurement of wet deposition of total Hg (THg, all forms of Hg in a sample) and MeHg in open areas has been ongoing at the ELA since 1992 and was expanded in 1998 to include measurements of THg and MeHg under forest canopies for estimation of dry deposition,11 yielding the longest record of Hg deposition of any remote site worldwide.10,17 In this study, we used enriched stable Hg isotopes to examine the fate of MeHg on canopy foliage following deposition, including long-term retention of MeHg on foliage and (photo)reduction of MeHg to Hg(0). We also attempted for the first time to measure particulate-bound MeHg (p-MeHg) in the atmosphere to determine if p-MeHg contributes to MeHg dry deposition. This new information will allow researchers to better constrain measures of MeHg dry deposition to watersheds in remote boreal regions.
Received: December 29, 2010 Accepted: May 3, 2011 Revised: May 1, 2011 Published: May 25, 2011 5178
dx.doi.org/10.1021/es104377y | Environ. Sci. Technol. 2011, 45, 5178–5185
Environmental Science & Technology
’ METHODS General Site Description. Our research was conducted in a number of different watersheds throughout the ELA in the remote Boreal ecoregion of northwestern Ontario, Canada (Figure S1 of the Supporting Information). This area is underlain by Precambrian Shield geology with upland forests dominated by stands of jack pine (Pinus banksiana), black spruce (Picea mariana), balsam fir (Abies balsamea), and white birch (Betula papyrifera) of varying age due to fire succession. Wetland regions typically contain mixed black spruce, jack pine, and tamarack (Larix laricina) canopies with alder (Alnus rugosa) shrub understory. Sources and Fates of MeHg on Canopy Foliage. Annual Loadings of MeHg in Open Area Precipitation, Throughfall, and Litterfall. In 2007, we continued long-term monitoring of MeHg and THg deposition in open area precipitation and in throughfall and litterfall under four canopy types at the ELA.10 The experimental design and sampling method remained unchanged from those in the past, detailed descriptions of which can be found in Graydon et al.10 Details of MeHg and THg analytical techniques and QA/QC are presented at the end of the Methods section. Retention of MeHg on Canopy Foliage. To better understand the fate of MeHg following deposition, we quantified the longterm retention of MeHg on foliage and compared it to that of inorganic Hg(II). On July 18, 2006, nine trees (3 each of birch, jack pine, and black spruce) in the upland portion of the Lake 302 watershed (Figure S1 of the Supporting Information) were each sprayed using a hand-held plastic garden sprayer with 1 L of isotopically enriched Me199Hg and 198Hg(II) solution (∼0.4 μg L1 and ∼1 μg L1 as Hg, respectively), such that foliage was dripping following spray. Me199Hg tracer was synthesized from 199HgO (92% purity, Trace Sciences International) using methylcobalamin.18 198Hg(II) tracer was obtained by oxidizing 198Hg(0) (90.5% purity, Trace Sciences International) using concentrated HNO3 and diluting with 5% HCl. Each spray solution was mixed using acidified oligotrophic lake water (pH 4 using trace metal grade HCl) to more closely mimic natural speciation by providing ligands to which the Me199Hg and 198Hg(II) could bind before application to the trees. The procedures are similar to those employed for the METAALICUS loading experiment.19 After complete drying (∼0.5 h), a foliage sample was clipped from each tree and bagged using clean-hands, dirty-hands protocol. Each sample consisted of foliage clipped from several locations on the tree to account for potential intratree variability following spray application. All nine trees were sampled again 1, 3, and 6 days post spray, and weekly thereafter to determine rate constants and half-lives for the retention of Me199Hg and 198 Hg(II) on foliage. Sampling of spruce and jack pine lasted for a total of 35 days postspray, while the birch were not sampled beyond 2 weeks postspray due to drought-induced early senescence. Additionally, the spruce and jack pine were sampled again in May 2007 for a longer-term comparison of Hg retention. A description of Me199Hg and 198Hg(II) analyses on foliage is provided at the end of the Methods section. To compare Me199Hg and 198Hg(II) retention among the nine trees, we first standardized for intertree variability in postspray Me199Hg and 198Hg(II) concentrations of canopy foliage by expressing the foliar concentration at each time point as a proportion of the initial concentration. Rate constants (k) for the decline of foliar Me199Hg and 198Hg(II) were calculated for
ARTICLE
each tree assuming (pseudo)first-order chemical kinetics: lnðspiket =spike0 Þ ¼ kt
ð1Þ
where spiket and spike0 are the concentrations of either Me199Hg or 198Hg(II) at time t (t > 0) and time 0, respectively. A mixed effects model (Proc Mixed) in SAS v. 9.2 (SAS Institute Inc. 2007) was used to determine if decline rates varied between Me199Hg and 198Hg(II) and/or among foliage types. In this analysis, k obtained for each tree was the response variable, with tree ID as the random effect, and foliar species and Hg species as the fixed effects. Average kMeHg and kHg(II) values for each tree species were used to calculate half-lives (t1/2) of Me199Hg and 198 Hg(II) retention: t 1=2 ¼
lnð2Þ k
ð2Þ
(Photo)reduction of MeHg and Evasion as Hg(0). To determine if (photo)reduction with subsequent evasion as Hg(0) is a mechanism of loss of newly deposited MeHg from foliage, we sprayed a spruce, jack pine, and birch tree each with 1 L of isotopically enriched Me199Hg solution (∼1 μg L1 as Hg, prepared as described above) and measured rates of 199Hg(0) flux using a Teflon dynamic flux chamber (DFC) that allowed enclosed foliage to maintain normal physiological function (Figure S2 of the Supporting Information). The DFC consisted of a rectangular PFA Teflon bag (Welch Fluorocarbon Inc.; 15 15 40 cm, ∼9 L) mounted on an external stainless steel frame with an outlet sampling port located at the bottom center of the chamber and four inlet holes evenly spaced on the front of the chamber (Figure S2 of the Supporting Information). PFA Teflon was chosen for the chamber bag because it is inert to Hg binding and allows high transmissivity of UV radiation (Figure S3 of the Supporting Information). Two Brailsford TD-2NA vacuum pumps connected to the outlet port via Teflon tubing and a straight Teflon compression fitting created air flow through the chamber. The flow rate through the DFC was measured using a GFM17S Aalborg stainless steel thermal mass flow controller (0.1 L min1 precision). Flows were maintained between 7 and 9 L min1, establishing a complete exchange of air within the chamber approximately every 1 min. As a result, there was no visible condensation buildup on the inside of the chamber, nor any signs of leaf wilting. A Brailsford TD-2NA vacuum pump subsampled the outlet line between 0.5 and 2 L min1 through a gold-coated glass bead trap (gold trap) that stripped Hg(0) completely from the air. A second sampling stream was used to quantify spike Hg(0) concentrations entering the chamber through inlet holes. Aalborg GFM17 thermal mass flow meters (0.01 L min1 precision) and TOT10 totalizers measured total air flow through the gold traps. Soda lime traps (814 mesh, Alfa Aesar) were used upstream of gold traps to remove humidity (Figure S2 of the Supporting Information). Flux measurements lasted 1 h and were performed twice on the day trees were sprayed and repeated 1, 6, and 17 days postspray. A new chamber was used for each measurement (n = 15); however, six were run as blank chambers prior to use on tree foliage. Prior to each flux measurement, all sampling lines upstream of the gold traps were switched out with new or acidwashed replacements. All flux measurements on a given tree were performed on a designated portion of the same branch. Branches on which the flux measurements were performed were chosen so 5179
dx.doi.org/10.1021/es104377y |Environ. Sci. Technol. 2011, 45, 5178–5185
Environmental Science & Technology
ARTICLE
as to minimize any relative variability in basic characteristics (e.g., age of tree, height of branch above ground, condition of foliage) and any potential differences in radiative exposure (e.g., direction in relation to the sun, shading from nearby foliage) that might arise between the branches over the 17 day study period. All gold traps were heated to 400 °C in a stream of UHP N2 prior to flux measurements to desorb any residual Hg, then sealed using Teflon plugs and tape, and individually bagged. Gold traps were stored under positive pressure in sealed acid-cleaned glass mason jars purged with UHP N2 both before and after use. Travel blank gold traps (n = 3) accompanied sample traps to and from our field site. Hg(0) was desorbed from sample traps at 400 °C into UHP Ar carrier gas and delivered to an inductively coupled plasma mass spectrometer (ICP-MS) for detection (also see below). To ensure the quality and consistency of peak signals among samples, sample traps were first desorbed onto an analytical trap, which was then desorbed into the ICP-MS. Spike Hg(0) concentrations were determined to be above detection limit when the amount of 199Hg(0) was greater than 0.25% of the ambient Hg(0) desorbed from a trap.20 Rates of 199Hg(0) flux (F, ng g1 h1 or ng m2 h1) from foliage were calculated using the following equation: F ¼
ð199 Hgð0Þo =βo 199 Hgð0Þi =βi Þ fm W
ð3Þ
where 199Hg(0)o = total spike 199Hg(0) on the outlet gold trap (ng), βo = total air flow through the outlet gold trap (L), 199 Hg(0)i = total spike 199Hg(0) on the inlet gold trap (ng), βi = total air flow through the inlet gold trap (L), fm = volume of air exiting the chamber (L hr1), and W = freeze-dried mass (g) or total leaf area (m2) of foliage inside the chamber. As 199Hg(0) was below detection in air exiting all blank chambers and was never detected on travel blank traps, it was not necessary to include these potential background sources in the flux rate calculation. To best compare relative rates of (photo)reduction and 199Hg(0) evasion among the three tree species, the 199Hg(0) flux from each tree was expressed as a proportion of the initial flux. Sources of Dry Deposited MeHg. We sampled for particlebound MeHg (p-MeHg) in air using sampling units consisting of a quartz microfiber air sampling filter (Whatman QM-A 47 mm diameter, >99.95% retention of particles g0.3 μm21) housed in an acid-washed Teflon filter pack connected to a constant flow high-volume air sampling pump (SKC Inc., 224-PCXR8 Universal Sample Pump). The filter packs were set 1.5 m above the ground and ∼20 m3 of air was pulled through each filter at a rate of 5 L min1 for ∼3 days. The inlet port of each filter pack was capped with Nytex screening (200 μm mesh opening) to prevent insects from entering the sampling stream and collecting on filters. Plastic domes were mounted above filter packs to prevent filters from being wetted during precipitation events, although we attempted whenever possible to sample during rain-free intervals when atmospheric concentrations of particulate matter should have been highest. Air sampling units were deployed within the METAALICUS watershed (2 under forest canopies and one in the open) and in the open at the ELA meteorological site (Figures S1 and S4 of the Supporting Information). Sampling units were deployed in this manner because we hypothesized that the efficiency of canopy foliage at scrubbing and binding particles from the atmosphere
would result in higher concentrations of p-MeHg in open air than in air under the canopy. Prior to use, all filters were precombusted at 450 °C for 6 h, placed individually in acid-washed plastic filter cases sealed with Parafilm, and stored double bagged under positive pressure in sealed acid-cleaned glass mason jars purged with UHP N2. Following collection from the field, filters were stored in the same manner and frozen until MeHg analysis (details below). Ambient and Isotopic Hg Analyses. All sample analyses were completed in the University of Alberta Biogeochemical Analytical Laboratory (Edmonton, AB, Canada) using an Elan DRC-e, PerkinElmer Sciex ICP-MS. Because the ICP-MS measures amounts of individual Hg isotopes, it was possible to distinguish and calculate concentrations of both ambient Hg and experimentally applied Hg spikes in samples. To calculate concentrations of ambient MeHg and THg, an isotope that was not used as a spike in a particular experiment was used as an ambient Hg surrogate.20 For samples collected within the METAALICUS watershed, 199Hg was used as the ambient Hg surrogate. When Me199Hg or 198Hg(II) were applied as spike, 202Hg was selected as the ambient Hg surrogate instead. Precipitation and Throughfall. For MeHg analysis, 100 mL of precipitation/throughfall sample was distilled at 125 °C for approximately 5 h. Volatile Hg species were then purged from the distillate and trapped onto Tenax following ethylation using NaBEt4. Trapped Hg species were thermally desorbed and separated using a gas chromatograph glass column packed with 15% OV-3 Chromosorb (60/80 Mesh) heated to 105 °C and delivered to the ICP-MS for detection. Me201Hg, proportional to the amount expected to be found in each sample, was added as an internal standard prior to distillation to correct for procedural recoveries. Concentrations of ambient and spike MeHg were calculated from the measured isotope ratios as described in Hintelmann and Ogrinc.20 Precipitation and throughfall samples were analyzed for THg using an automated Tekran 2600 total Hg analyzer interfaced with the ICP-MS. Briefly, all Hg species in a sample were oxidized to Hg(II) by the addition of 0.2% BrCl (v/v), reduced to Hg(0) using SnCl2, purged onto gold traps, and analyzed by ICP-MS after thermal desorption. Foliage and Filters. MeHg content of foliage and filters was quantified in a similar manner as the precipitation/throughfall samples; however, prior to distillation, samples were digested overnight in 500 uL of 9 M H2SO4, 200 uL of 20% KCl, and 40 mL of Milli-Q water in 50 mL Teflon vials. For litterfall samples, ∼250 mg of ground litterfall was digested, whereas for spray tree foliage only ∼100 mg of ground tissue was digested. In the case of filters, the entire filter was digested. Me201Hg was added prior to the digestion as an internal standard. For THg analysis, ∼250 mg of ground foliage was digested in 60 mL Teflon bombs using 7 mL of 7:3 (v/v) HNO3/H2SO4. Bombs were heated in a vented oven for two hours at 125 °C. Samples were allowed to cool and 19 mL of Milli-Q water and 1 mL of BrCl were added. Bombs were closed and heated overnight at 60 °C. A 0.5 mL subsample of the digest was diluted with Milli-Q water to a final volume of 50 mL and 0.04% (v/v) hydroxylamine hydrochloride was added to neutralize excess BrCl. Sample reduction, delivery, and detection were as described above for precipitation/throughfall samples. Successful interlaboratory comparisons of duplicate water and foliage samples were conducted10 with Flett Research Ltd. (Winnipeg, MB) and the Trent University Worlsford Water 5180
dx.doi.org/10.1021/es104377y |Environ. Sci. Technol. 2011, 45, 5178–5185
Environmental Science & Technology
ARTICLE
Table 1. Decline Rate Constants (k) and Half Lives (t1/2) of Me199Hg and 198Hg(II) on Foliage of 3 Tree Species Based on 35 Days (Short-Term) and 295 Days (Long-Term) of Monitoring Following a Simulated Wet-Deposition Event Short-term tree species
k (day1)
t1/2 (days)
Long-term k (day1)
t1/2 (days)
Me199Hg birch
0.085
8(3
jack pine
0.035
20 ( 7
0.0037
187 ( 101
spruce
0.025
28 ( 5
0.0034
204 ( 66
198
Hg(II)
Figure 1. Retention of Me199Hg and 198Hg(II) on spruce (n = 3), jack pine (n = 3), and birch (n = 3) foliage in 2006. Data points are ln of the ratio of initial postspray Hg spike concentration remaining on foliage at the time of sampling (mean ( S.D.). Data points from the single sampling event in 2007 for spruce and jack pine are included here for reference but were not included in the regression analyses.
Quality Center (Peterborough, ON). Limits of detection (LODs) for ambient MeHg and THg analyses were calculated as 3 SD of analytical blanks and were between 0.01 to 0.03 ng L1 and 0.05 to 0.3 ng L1 respectively for precipitation/ throughfall and 3 pg g1 and 0.4 ng g1 respectively for litterfall. The LODs for the experimentally applied Hg isotopes varied with the precision of the isotope ratio measurement, ranging between 0.5 and 2.0% of the ambient Hg concentration of a sample.
’ RESULTS AND DISCUSSION Sources and Fates of MeHg on Canopy Foliage. Annual Loadings of MeHg in Open Area Precipitation, Throughfall, and Litterfall. In 2007, as in previous years,10 the majority (83%) of
open area MeHg loading (0.41 mg ha1) occurred during the icefree season (May 5 to October 31), primarily due to lower water deposition during the winter (142 mm) than during the ice-free season (647 mm). Average annual throughfall loadings of MeHg did not differ among the various canopy types (wetland = 0.44 mg ha1, jack pine = 0.34 ( 0.01 mg ha1, deciduous = 0.35 mg ha1), except for under the old growth canopy where deposition was somewhat higher (0.60 ( 0.16 mg ha1, Figure S5 of the Supporting Information). Although the difference was small, throughfall MeHg loadings at the jack pine and deciduous sites
birch
0.10
7(1
jack pine
0.045
15 ( 4
0.0046
151 ( 75
spruce
0.028
24 ( 5
0.0043
161 ( 49
were lower than MeHg loading in the open, suggesting that some portion of newly deposited MeHg may remain bound to foliage or returned to the atmosphere following deposition. The mean annual mass flux of litterfall at all of our sites was 2770 ( 910 kg ha1, 70 ( 14% of which occurred before the end of October. The resulting litterfall MeHg loadings were slightly higher under the old growth (0.97 ( 0.34 mg ha1) and wetland (0.94 ( 0.34 mg ha1) canopies than under the jack pine (0.80 ( 0.16 mg ha1) and deciduous (0.77 ( 0.07 mg ha1) canopies. On average, litterfall loadings in 2007 resulted in a 23 fold increase in MeHg deposition at forested sites compared to in the open, as has been found at the ELA in the past (ref 10 and Figure S5 of the Supporting Information). Combined net throughfall and litterfall loadings of MeHg under the forest canopy (1.1 ( 0.07 1.6 ( 0.50 mg ha1) ranged from 3 to 4 times higher than loadings in the open (0.41 mg ha1). Retention of MeHg on Canopy Foliage. Whereas combined throughfall and litterfall loadings suggest that forest canopies are net sources of MeHg to watersheds, not all newly deposited MeHg remains bound to foliage. Concentrations of Me199Hg and 198Hg(II) initially retained on foliage following spray application were similar for both species of Hg and for all trees (Me199Hg = 6.3 ( 1.5 ng g1, 198Hg(II) = 5.7 ( 1.1 ng g1), and declined exponentially over the course of the study, consistent with a first-order process (Figure 1). Average kMeHg and kHg(II) for each tree species ranged from 0.025 to 0.085 day1 and 0.028 to 0.10 day1 respectively with corresponding t1/2 of spike retention on foliage of 28 ( 5.0 to 8.2 ( 3.2 days for Me199Hg and 24 ( 4.8 to 6.8 ( 1.4 days for 198Hg(II) (Table 1). We found no significant difference between the loss rates of Me199Hg and 198 Hg(II) from foliage (p = 0.10); however, reduction of foliar concentration of both Me199Hg and 198Hg(II) occurred significantly faster on birch foliage than on either jack pine or spruce needles (p < 0.001). These results are consistent with previous studies that found Hg(II) spike in the METAALICUS watershed was retained longer in coniferous canopies than in deciduous canopies.15,22 By the end of the 2006 sampling season in late August, 34 days after the initial application of Me199Hg to the trees, 34 ( 11% and 26 ( 12% of the applied Me199Hg remained on spruce and jack pine foliage, respectively. A similar portion, 26 ( 14%, remained on birch foliage at the time of leaf drop 2 weeks prior. Retention of 198Hg(II) was similar with 26 ( 5%, 14 ( 8%, and 5181
dx.doi.org/10.1021/es104377y |Environ. Sci. Technol. 2011, 45, 5178–5185
Environmental Science & Technology 31 ( 6% remaining on spruce, jack pine, and birch foliage, respectively. Early in the 2007 growing season, the percentage of applied Me199Hg remaining on the coniferous foliage had decreased slightly (spruce = 27 ( 9%, jack pine = 16 ( 6%), but not significantly so (paired t test, n = 6, p = 0.081), suggesting that loss of MeHg from the canopy during the winter is low or even negligible. The amount of 198Hg(II) remaining on foliage had similarly decreased (spruce = 18 ( 7%, jack pine = 8 ( 3%); however, the difference in the amount of 198Hg(II) retained on foliage between the end of the 2006 growing season and that retained in early 2007 was significant (paired t test, n = 6, p = 0.022). When the 2007 data are included in our regression models, rate constants decreased and half-lives increased substantially (Table 1). Long-term half-lives for Me199Hg were 204 ( 66 and 187 ( 101 days for spruce and jack pine respectively and for 198 Hg(II) were 161 ( 49 and 151 ( 75 days for spruce and jack pine, respectively. These values are consistent with previously reported long-term half-lives for Hg(II) spike retention on foliage (180 ( 40 days and 110 ( 30 days on coniferous and deciduous canopies, respectively; ref 22). However, unlike results presented in Graydon et al.,22 which sampled foliage from the same trees over multiple growing seasons, here birch foliage was only collected over the period of a single growing season. Birch foliage was not sampled in 2007 along with the coniferous foliage because it was expected that the majority of Me199Hg and 198 Hg(II) remaining on foliage would be lost at the time of leaf abscission; however, Graydon et al.’s22 findings suggest otherwise, likely due to the retention of Hg on branches analyzed with foliage in that study. It should be noted that the long-term rate constants represent an average rate for summer and winter loss processes. The prolonged retention of Me199Hg and 198Hg(II) on foliage following our initial summer sampling season is likely a result of seasonal changes in environmental conditions. During winter, not only are low temperatures likely to reduce biological and chemical activity on and within plant tissue, and therefore any chemical transformations of MeHg or Hg(II), but the reduced amount of solar radiation incident on foliage (from both lower sun angles and snow cover) would also decrease (photo)reduction rates of these Hg species.23 Although we were unable to measure loss of MeHg during the winter in this study, if summer-time rates were applied on an annual basis or to foliage retained in the canopy for multiple growing seasons, the amount of MeHg lost from the canopy via this process would be grossly overestimated. Thus, the long-term rates reported here represent the best annual rates currently available, and, for both coniferous and deciduous tree species, use of a long-term rate constant to determine retention and/or loss of MeHg from canopy foliage seems most appropriate. (Photo)reduction of MeHg and Evasion as Hg(0). As has been previously determined for Hg(II),15 we detected fluxes of 199Hg(0) from all three trees sprayed with Me199Hg, showing that a portion of newly wet-deposited MeHg is returned almost immediately to the atmosphere. 199Hg(0) evasion rates were highest immediately after spray application and declined rapidly thereafter, leveling off to near detection limits within one week postspray (Figure 2). Flux rates of 199Hg(0) from the birch branch decreased more rapidly than those from either jack pine or spruce (Figure 2). Although our ability to make inferences about any differences associated with foliage type is limited by our small sample size, it does appear that foliage type has an effect on the rate of demethylation of MeHg and subsequently evasion
ARTICLE
Figure 2. Relative fluxes of 199Hg(0) from the foliage of spruce, jack pine, and birch trees sprayed with Me199Hg to simulate a wet deposition event. Each flux rate is presented as a ratio of the initial flux rate for easy comparison of the flux pattern among the trees over the two week sampling period. The flux rate measured from the spruce tree on day 17 was below detection limit and is indicated as such by an open, as opposed to shaded, symbol.
as Hg(0). This was not unexpected given the vastly different structures of coniferous needles and deciduous leaves. For example, potential differences in the thickness of the waxy cuticle and/or the microenvironments on foliage surfaces may affect MeHg binding. Additionally, foliage structure can affect the amount of direct light received by foliar surfaces. The greater sustained 199Hg(0) flux measured from the coniferous species, in comparison to the birch, could be a result of the complex and densely packed needle structure of coniferous trees, which likely provides more shade from direct radiation, effectively reducing the initial loss of MeHg. If low levels of Hg(0) evasion are sustained, such as those measured at and beyond 6 days post spray (Figure 2), this mechanism of loss may account for some of the observed continuous long-term declines of MeHg we observed on foliage (Figure 1). However, as the magnitude of 199Hg(0) fluxes declined more rapidly, and tapered off more quickly, than the observed decline in Me199Hg on foliage in the previous experiment, it is plausible that mechanisms of loss besides (photo)reduction and evasion might also be removing MeHg from the canopy pool. As precipitation and wind pass through forest canopies, mechanical weathering could remove some portion of the MeHg and Hg(II) bound to foliar surfaces, some of which would likely contribute to throughfall loadings. Unfortunately, a more rigorous comparison between the two experiments could not be conducted because the destructive nature of sampling prevented us from quantifying the amount of Me199Hg that was retained on the flux sample branch over the course of our study. This study did not examine mechanisms behind the observed demethylation of Me199Hg and subsequent volatilization of 199Hg(0) from plant foliage; however, the results of similar studies suggest that a photochemical reaction is largely responsible.15,24 For example, MeHg in lake water is degraded primarily by UV radiation (280400 nm; 24,25,4). MeHg photodemethylation in water is thought to be mediated by photosensitizing species such as iron, nitrate, and dissolved organic matter 2527 and the same could be true on wet foliage surfaces following spray application. 5182
dx.doi.org/10.1021/es104377y |Environ. Sci. Technol. 2011, 45, 5178–5185
Environmental Science & Technology Regardless of the mechanism, the use of stable isotopes allowed us to clearly demonstrate that one of the end products of (photo)demethylation of MeHg on canopy foliage was Hg(0), thereby making the overall process reductive. Because Hg(0) is rapidly evaded, (photo)reduction of MeHg represents a loss of newly deposited MeHg from the canopy. Considering that 83% of annual MeHg wet deposition loading occurs during the icefree period when (photo)reduction and loss from the canopy is highest, this has important implications for quantifying gross and net deposition of atmospheric MeHg to forested landscapes. Furthermore, the impact of the canopy on atmospheric deposition of MeHg to watersheds and the amount of potential dry deposition occurring are underestimated when the common mathematical approach of subtracting wet deposition loadings in open areas from the sum of throughfall and litterfall loadings is used to estimate dry deposition. Sources of Dry Deposited MeHg. Of the 54 filters deployed to collect p-MeHg over a series of 16 sampling intervals, not one collected detectable levels of MeHg (LOD = 5 pg filter1, calculated as 3 * SD of travel blanks, n = 33). This finding does not mean that p-MeHg does not exist; only that it was either not detectable using our current sampling protocols or that it exists in concentrations lower than we were able to quantify. Given that ∼20 m3 of air was pulled through each filter, atmospheric concentrations of p-MeHg could be as high as 0.25 pg m3 and remain at or below the 5 pg analytical detection limit for our filters. It is reasonable to assume that this might be the case as summertime p-Hg concentrations measured in the open by a Tekran air Hg speciation unit at the ELA are low, often between 5 and 20 pg m3 (V. St Louis/J. Graydon, unpublished data), and MeHg is thought to represent only a small fraction of THg in the atmosphere (∼0.1% of the global atmospheric budget of Hg; ref 28). Assuming that this is a suitable method for quantifying p-MeHg, and using our detection limit as an upper limit of atmospheric concentrations, we conclude that p-MeHg accounts for less than 5% of all particulate-bound Hg species in air at the ELA. Although little work has been published on the deposition of p-Hg, deposition velocities (Vd) are thought to be in the range of 0.022.0 cm s1 and are consistent with theoretical and field studies conducted for other fine particle species.29 Assuming that Vd for p-MeHg is similar, and that p-MeHg concentrations are at most 0.25 pg m3, we estimate that the annual dry depositional flux of p-MeHg is unlikely to exceed ∼0.0151.5 mg ha1. At the upper limits of this flux range, dry deposition of p-MeHg could partially account for the differences observed between loadings of MeHg in the open and under forested sites in the METAALICUS watershed (0.71.1 mg ha1). Lending support to the notion of dry deposited p-MeHg, Lamborg et al.28 and Hall et al.30 proposed that the association of a portion of MeHg with particles in rainwater samples is evidence of atmospheric wash-out of p-MeHg. However, Bloom and Watras31 failed to observe a decreasing trend in the concentration of MeHg in consecutive samples collected during a single rain event, a phenomenon that is typically observed during washout events. Additionally, as Mason et al.32 and Graydon et al.10 were both unable to determine a correlative trend between MeHg concentrations and associated rainfall amounts, the argument for a proposed washout of atmospheric p-MeHg is weakened further. Alternatively, Hammershmidt et al.33 suggested abiotic methylation of reactive Hg(II) species (HgR) as a likely source of MeHg in rain, especially HgR associated with particulate matter. If the latter is true, then
ARTICLE
p-MeHg in rainwater may simply be the result of aqueous phase methylation of p-Hg, and it is increasingly unlikely that washout of atmospheric p-MeHg is occurring or that atmospheric p-MeHg exists in general. Clearly, further effort and more rigorous testing are required to fully elucidate the source(s) of MeHg in wet deposition and subsequently the dry-deposition of MeHg to watersheds that is enhanced by forest canopies. Although dry deposition of Hg to forest canopies occurs via several processes and can result in Hg both on the surface and inside of canopy foliage, all potential sources of MeHg in particular are not yet fully understood. Stomatal uptake of Hg(0) can be a significant source of Hg dry deposition;14 however, it is unlikely that such a source would have a significant effect on MeHg loadings because gaseous MeHg, to the best of our knowledge, is virtually nonexistent. In the case of MeHg, p-MeHg is still considered to be the principal dry source enhancing deposition at forest sites. As with wet deposited MeHg, some portion of dry deposited p-MeHg may remain bound to foliage or retained in the wax layer of leaves and needles due to the strong affinity of MeHg for organic matter.5 Because p-MeHg has potential to be washed from foliage surfaces and end up in throughfall, adsorption to foliage would need to be very efficient for this source to account for the measured litterfall loadings being higher than their associated throughfall loadings.8 In addition to atmospheric deposition of MeHg, other sources of Hg to the canopy have been proposed, although they too are unlikely to account for all of the enhanced deposition of MeHg observed at forested sites. For example, previous studies have indicated that, except in areas with heavily Hg contaminated soils,34 roots act as a barrier preventing uptake and subsequent sequestration of Hg in foliage and root uptake likely accounts for less than 3% of MeHg measured in litterfall.35,36 In-canopy methylation of Hg(II), which could then either be rinsed off in throughfall or be retained on foliage and contribute to litterfall, has been speculated as another potentially important source of MeHg at forested sites.6,8 However, in throughfall samples collected under canopies in the METAALICUS watershed, Graydon et al.10 found that less than 0.5% of spike Hg(II) had been methylated, whereas up to 3% of ambient THg was in the methylated form. Therefore, whereas it is clear that deposition of MeHg is enhanced at forested sites compared to open areas, other than the potential for drydeposition of p-MeHg as discussed above, additional sources of MeHg to canopy foliage remain somewhat elusive. Determining the source and magnitude of all internal and external fluxes of MeHg to watersheds is imperative for effective management of MeHg contamination in fish. Given that measurements of internal MeHg production are considered more difficult to measure than external sources,4,6 it is of utmost importance to identify and accurately report the magnitude of all external inputs. Dry deposition of MeHg, whether it occurs via p-MeHg or some yet to be determined source, can occur directly to the lake surface or to the surrounding terrestrial compartments of the watershed, where it is enhanced by canopy foliage, eventually making its way into the lake. As a result, this source (which we have demonstrated to be larger than previously estimated) will continue to have an effect on the mass balancing of MeHg fluxes into lakes1 and needs to be better understood. The findings reported here will allow researchers to better constrain measures of MeHg dry deposition to watersheds; however, much work remains to be done to improve our understanding of the sources and fates of atmospherically deposited MeHg. 5183
dx.doi.org/10.1021/es104377y |Environ. Sci. Technol. 2011, 45, 5178–5185
Environmental Science & Technology
’ ASSOCIATED CONTENT
bS
Supporting Information. Five figures illustrating methods and results. This material is available free of charge via the Internet at http://pubs.acs.org.
’ AUTHOR INFORMATION Corresponding Author
*Phone: (780) 492-9386; fax: (780) 492-9234; e-mail: vince.
[email protected].
’ ACKNOWLEDGMENT Contribution No. 41 of the Mercury Experiment to Assess Atmospheric Loading in Canada and the United States (METAALICUS). We greatly appreciate the field support from K. Sandilands and M. Tate, laboratory support from S. Berkel, P.Y. Chan, M. Hawkins, C. Lee, J. Kirk, and J. Zhu, and statistical input from J. Roland and A. Clason. This study was funded by the Natural Sciences and Engineering Research Council of Canada, Canadian Circumpolar Institute, and Department of Biological Sciences, University of Alberta. ’ REFERENCES (1) Harris, R. C.; Rudd, J. W. M.; Amyot, M.; Babiarz, C. L.; Beaty, K. G.; Blanchfield, P. J.; Bodaly, R. A.; Branfireun, B. A.; Gilmour, C. C.; Graydon, J. A.; Heyes, A.; Hintelmann, H.; Hurley, J. P.; Kelly, C. A.; Krabbenhoft, D. P.; Lindberg, S. E.; Mason, R. P.; Paterson, M. J.; Podemski, C. L.; Robinson, A.; Sandilands, K. A.; Southworth, G. R.; Louis, V. L. S.; Tate, M. T. Whole-ecosystem study shows rapid fishmercury response to changes in mercury deposition. Proc. Natl. Acad. Sci. U.S.A. 2007, 104, 16586–16591. (2) Gilmour, C. C.; Henry, E. A.; Mitchell, R. Sulfate stimulation of mercury methylation in fresh-water sediments. Environ. Sci. Technol. 1992, 26, 2281–2287. (3) Louis, V. L., St.; Rudd, J. W. M.; Kelly, C. A.; Beaty, K. G.; Bloom, N. S.; Flett, R. J. Importance of wetlands as sources of methyl mercury to boreal forest ecosystems. Can. J. Fish. Aquat. Sci. 1994, 51, 1065–1076. (4) Sellers, P.; Kelly, C. A.; Rudd, J. W. Fluxes of methylmercury to the water column of a drainage lake: The relative importance of internal and external sources. Limnol. Oceanogr. 2001, 46, 623–631. (5) Hultberg, H.; Iverfeldt, Å.; Lee, Y. H. Methylmercury input/ output and accumulation in forested catchments and critical loads for lakes in southwestern Sweden. In Hg Pollution: Integration and Synthesis; Watras, C. J., Huckabee, J., Eds.; Lewis Pub.: Boca Raton, FL, 1994. (6) Rudd, J. W. M. Sources of methyl mercury to freshwater ecosystems: A review. Water, Air, Soil Pollut. 1995, 80, 697–713. (7) Iverfeldt, Å. Mercury in forest canopy throughfall and its relation to atmospheric deposition. Water, Air, Soil Pollut. 1991, 56, 553–564. (8) Munthe, J.; Hultberg, H.; Iverfeldt, Å. Mechanisms of deposition of methylmercury and mercury to coniferous forests. Water, Air, Soil Pollut. 1995, 80, 363–371. (9) Lindberg, S. E. Forests and the global biogeochemical cycle of mercury: the importance of understanding air/vegetation exchange processes. In Global and Regional Mercury Cycles: Sources, Fluxes and Mass Balances; Baeyens, W., Ebinghaus, R., Vasiliev, O., Eds.; NATO ASI Series, Vol. 21, Kluwer Academic Publishers: Dordrecht, The Netherlands, 1996; pp 359. (10) Graydon, J. A.; Louis, V. L., St.; Hintelmann, H.; Lindberg, S. E.; Sandilands, K. A.; Rudd, J. W. M.; Kelly, C. A.; Hall, B. D.; Mowat, L. D. Long-term wet and dry deposition of mercury in the remote Boreal ecoregion of Canada. Environ. Sci. Technol. 2008, 42, 8345–8351. (11) Louis, V. L., St.; Rudd, J. W. M.; Kelly, C. A.; Hall, B. D.; Rolfhus, K. R.; Scott, K. J.; Lindberg, S. E.; Dong, W. Importance of the
ARTICLE
forest canopy to fluxes of methyl mercury and total mercury to boreal ecosystems. Environ. Sci. Technol. 2001, 35, 3089–3098. (12) Munthe, J.; Hultberg, H.; Lee, Y.-H.; Parkman, H.; Iverfeldt, Å.; Renberg, I. Trends of mercury and methylmercury in deposition, run-off water and sediments in relation to experimental manipulations and acidification. Water, Air, Soil Pollut. 1995, 85, 743–748. (13) Rea, A. W.; Keeler, G. J.; Scherbatskoy, T. The deposition of mercury in throughfall and litterfall in the Lake Champlain watershed: A short-term study. Atmos. Environ. 1996, 30, 3257–3263. (14) Lindberg, S.; Bullock, R; Ebinghaus, R.; Engstrom, D.; Feng, X.; Fitzgerald, W.; Pirrone, N.; Prestbo, E.; Seigneur, C. A synthesis of progress and uncertainties in attributing the sources of mercury in deposition. Ambio 2007, 36, 19–32. (15) Graydon, J. A.; Louis, V. L., St.; Lindberg, S. E.; Hintelmann, H.; Krabbenhoft, D. P. Investigation of mercury exchange between forest canopy vegetation and the atmosphere using a new dynamic chamber. Environ. Sci. Technol. 2006, 40, 4680–4688. (16) Schroeder, W. H.; Munthe, J. Atmospheric mercury an overview. Atmos. Environ. 1998, 32, 809–822. (17) Louis, V. L., St.; Rudd, J. W. M.; Kelly, C. A.; Barrie, L. A. Wet deposition of methyl mercury in northwestern Ontario compared to other geographic locations. Water, Air, Soil Pollut. 1995, 80, 405–414. (18) Hintelmann, H. Comparison of different extraction techniques used for methylmercury analysis with respect to accidental formation of methylmercury during sample preparation. Chemosphere 1999, 39, 1093–1105. (19) Sandilands, K. A.; Rudd, J. W. M.; Kelly, C. A.; Hintelmann, H.; Gilmour, C. C.; Tate, M. T. Application of enriched stable mercury isotopes to the Lake 658 watershed for the METAALICUS project, at the Experimental Lakes Area, Northwestern Ontario, Canada. Can. Tech. Rep. Fish. Aquat. Sci. 2005, 2597. (20) Hintelmann, H.; Ogrinc, N. Determination of stable mercury isotopes by ICP/MS and their application in environmental studies. In Biogeochemistry of Environmentally Important Trace Elements; Cai, Y., Braids, C. O., Eds.; ACS Symposium Series Vol. 835; American Chemical Society: Washington, DC, 2003. (21) Whatman QMA - filters for use in US EPA PM10 ambient air monitoring; Technical note technical specifications and performance data for Whatman products; Whatman Inc.: Clifton, NJ, 2002; http:// www.whatman.com/References/TechnicalNotePM10Filters.pdf (22) Graydon, J. A. Influence of vegetation on mercury deposition. Ph.D. Thesis, University of Alberta, Edmonton, 2009. (23) Grigal, D. F. Inputs and outputs of mercury from terrestrial watersheds: A review. Environ. Rev. 2002, 10, 1–39. (24) Lehnherr, I.; Louis, V. L., St. Importance of ultraviolet radiation in the photodemethylation of methylmercury in freshwater ecosystems. Environ. Sci. Technol. 2009, 43, 5692–5698. (25) Hammerschmidt, C. R.; Fitzgerald, W. F. Iron-mediated photochemical decomposition of methylmercury in an arctic Alaskan lake. Environ. Sci. Technol. 2010, 44, 6138–6143. (26) Chen, J.; Pehkonen, S. O.; Lin, C. J. Degradation of monomethylmercury chloride by hydroxyl radicals in simulated natural waters. Water Res. 2003, 37, 2496–2504. (27) Zheng, W.; Hintelmann, H. Isotope fractionation of mercry during its photochemical reduction by low-molecular-weight organic compounds. J. Phys. Chem. A 2010, 114, 4246–4253. (28) Lamborg, C. H.; Fitzgerald, W. H.; Vandal, G. M.; Rolfhus, K. R. Atmospheric mercury in northern Wisconsin: sources and species. Water Air Soil Pollut. 1995, 80, 189–198. (29) Zhang, L.; Wright, L .P.; Blanchard, P. A review of current knowledge concerning dry deposition of atmospheric mercury. Atmos. Environ. 2009, 43, 5853–5864. (30) Hall, B. D.; Manolopoulos, H.; Hurley, J. P.; Schauer, J. J.; Louis, V. L., St.; Kenski, D.; Graydon, J.; Babiarz, C. L.; Cleckner, L. B.; Keeler, G. J. Methyl and total mercury in precipitation in the Great Lakes region. Atmos. Environ. 2005, 39, 7557–7569. (31) Bloom, N. S.; Watras, C. J. Observations of methylmercury in precipitation. Sci. Total Environ. 1989, 87/88, 199–207. 5184
dx.doi.org/10.1021/es104377y |Environ. Sci. Technol. 2011, 45, 5178–5185
Environmental Science & Technology
ARTICLE
(32) Mason, R. P.; Lawson, N. M.; Sullivan, K. A. The concentration, speciation and sources of mercury in Chesapeake Bay precipitation. Atmos. Environ. 1997, 31, 3541–3550. (33) Hammerschmidt, C. R.; Lamborg, C. H.; Fitzgerald, W. F. Aqueous phase methylation as a potential source of methylmercury in wet deposition. Atmos. Environ. 2007, 41, 1663–1668. (34) Lindberg, S. E.; Jackson, D. R.; Huckabee, J. W.; Janzen, S. A.; Levin, M. J.; Lund, J. R. Atmospheric emission and plant uptake of mercury from agricultural soils near Almaden Mercury Mine. J. Environ. Qual. 1979, 8, 572–578. (35) Bishop, K. H.; Lee, Y.; Munthe, J.; Dambrine, E. Xylem sap as a pathway for total mercury and methylmercury transport from soils to tree canopy in the boreal forest. Biogeochemistry 1998, 40, 101–113. (36) Graydon, J. A.; Louis, V. L., St.; Hintelmann, H.; Lindberg, S. E.; Sandilands, K. A.; Rudd, J. W. M.; Kelly, C. A.; Tate, M. T.; Krabbenhoft, D. P.; Lehnherr, I. Investigation of uptake and retention of atmospheric Hg(II) by Boreal Forest plants using stable Hg isotopes. Environ. Sci. Technol. 2009, 43, 4960–4966.
5185
dx.doi.org/10.1021/es104377y |Environ. Sci. Technol. 2011, 45, 5178–5185
