Elevated Concentrations of Methyl Mercury in Streams after Forest

Oct 5, 2009 - Elevated Concentrations of Methyl Mercury in Streams after Forest Clear-Cut: A Consequence of Mobilization from Soil or New Methylation?...
0 downloads 0 Views 636KB Size
Download PDF
Environ. Sci. Technol. 2009, 43, 8535–8541

Elevated Concentrations of Methyl Mercury in Streams after Forest Clear-Cut: A Consequence of Mobilization from Soil or New Methylation? U L F S K Y L L B E R G , †,* ¨ RKMAN WESTIN,† MATTIAS BJO ¨ RN§ MARKUS MEILI,‡ AND ERIK BJO Department of Forest Ecology and Management, Swedish University of Agricultural Sciences, SE-90183 Umeå, Department of Applied Environmental Science, Stockholm University, SE-10691 Stockholm, Sweden, and Department of Chemistry, Umeå University, SE-90187 Umeå, Sweden

Received April 21, 2009. Revised manuscript received September 2, 2009. Accepted September 3, 2009.

Concentrations of inorganic, mercuric mercury (HgII), methyl mercury (MeHg) and ancillary chemistry measured in firstorder streams draining 0-4 (N ) 20) and 4-10 (N ) 27) yearold clear-cuts of former Norway Spruce Picea abies (Karst.) forest stands were compared with concentrations in streams draining >70 year-old Norway Spruce reference stands (N ) 10). Concentrations of MeHg, and ratios of MeHg TOC-1 and HgII TOC-1, were significantly (p < 0.01) elevated in 0-4 year-old clearcuts, as compared to references. The only ancillary variable showing a significant elevation for 0-4 year-old clear-cuts was Mn (p < 0.02). The 4-10 year-old clear-cuts showed intermediate concentrations with nonsignificant differences as compared to references. pH, nitrate, sulfate, Ca, Fe, TOC, TON, and the aromaticity of TOC (SUVA254 nm) showed nonsignificant differences between clear-cut age classes and references. Assuming that MeHg and HgII are mobilized from soil to stream to a similar relative extent as a consequence of clear-cutting, a calculation showed that 1/6 of the elevated MeHg concentration was due to enhanced mobilization from soil and 5/6 was due to new methylation of HgII 0-4 years after clearcut. New methylation after clear-cut is suggested to be stimulated by an increased availability of electron donors for methylating bacteria, as a consequence of degradation of logging residue (“slash”) and soil organic matter. A subdivision of sites situated above and below the highest postglacial coastline (HC) revealed a significant elevation of MeHg, MeHg TOC-1 and HgII TOC-1 (p < 0.05) beyond their references in 0-4 year-old clear-cuts above (but not below) the HC. This suggests that postglacial deposits of FeS(s) and FeS2(s) were not an important factor for elevation of MeHg after clearcut.

* Corresponding author e-mail: [email protected]; phone: +46(0)90-7868460; fax: +46(0)90-7868163. † Swedish University of Agricultural Sciences. ‡ Stockholm University. § Umeå University. 10.1021/es900996z CCC: $40.75

Published on Web 10/05/2009

 2009 American Chemical Society

Introduction Long-range transported Hg, emitted to the atmosphere as a consequence of combustion of fossil fuels, is diffusively deposited over the landscape as mainly inorganic, mercuric HgII. In regions dominated by wetlands and forest ecosystems, HgII is substantially transformed to the neurotoxin methyl mercury (MeHg), which bioaccumulates in organisms. Mass balance studies have shown that a net production of MeHg takes place under anoxic conditions in wetlands (1). In a field experiment, addition of sulfate was shown to enhance the total concentration of MeHg in a sediment (2), which was interpreted as a net production of MeHg (“methylation”) mediated by the activity of sulfate reducing bacteria (SRB). This ability of SRB has been verified in controlled laboratory experiments (3). More recently, iron(III) reducing bacteria (FeRB) also have proven to mediate production of MeHg in laboratory experiments (4-6). Despite major impact by forestry activities in temperate regions, very few studies have covered the role of forest harvesting on the mobilization and net production of MeHg. Garcia and Carignan (7, 8) conclusively showed that both zooplankton and fish (Esox lucius) contained significantly (p < 0.01) higher concentrations of MeHg and Hg in lakes within watersheds having in average 43% of the area covered by 1-year-old clear-cuts, as compared to reference watersheds unaffected by forestry activities during at least 40 years. In a follow-up study, elevated Hg accumulation in phytoplankton was shown to persist three years after clear-cut (9). Possible processes explaining these results, such as mobilization and/or formation of MeHg as a consequence of clearcut, were not investigated. In another study, significantly higher concentrations (and masses after correction for flux) of total Hg and MeHg were observed in the runoff from a small Norway spruce (Picea abies) catchment 1-7 years after clear-cutting and soil scarification (10, 11), as compared to the 3-year pretreatment monitoring period. Some of the Hg and MeHg export was explained by mobilization with DOC. After one year, DOC concentrations returned to pretreatment levels, whereas concentrations of Hg and MeHg remained elevated. This may indicate an enhanced mobilization of Hg and MeHg (with other agents than DOC) from soil and/or possibly an increased net formation of MeHg in soils after clear-cut. In contrast, increased export of MeHg and Hg after clear-cut in an ongoing study in Canada (12) was merely an effect of increased runoff (concentrations of Hg and MeHg remained unchanged). Based on the above studies, it was estimated that 9-23% of Hg currently accumulated in fish of forested, high-latitude landscapes can be attributed to forest harvesting practices (13). As discussed by Skyllberg (14), increased concentrations of MeHg in stream and soil/sediment pore waters, e.g., as a consequence of changes in land-use, may either be caused by a net production and/or by a mobilization of MeHg desorbed from soil or sediment particles. Because a mobilization of MeHg (and Hg) will inevitably occur if concentrations of strong complexing agents such as organic thiols, inorganic mono-, bi-, or polysulfides increase in solution, a true net formation of MeHg can only be demonstrated if the total amount of MeHg in the system is determined, or if proper corrections are made for shifts in equilibria leading to a dissolution of MeHg. It is also possible that Hg and MeHg are transported in association with mobile forms of colloidal FeS(s) and HgS(s). O’Driscoll and co-workers (28) showed that FeS(s) colloids were associated with DOC (passing a 0.45 µm filter) in oxic water, and suggested an VOL. 43, NO. 22, 2009 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

8535

increased mobilization linked to forest logging activities. It has also been shown that HgS(s) nanoparticles (passing 0.2 µm filter) are stabilized by humic substances (15). In this study, the release of Hg and MeHg as a consequence of clear-cutting was determined by measuring concentrations in streams draining young (0-4 year-old) and older (4-10 year-old) clear-cuts, as well as mature (>70 year-old) reference stands of Norway spruce in northern Sweden. In order to investigate possible effects of different sulfur and iron geochemical conditions, sites were selected to be situated above and below the highest postglacial coastline (HC). Historically, periods with marine and brackish water conditions have provided input of sulfur into the soils below this boundary (16, 17). In subsoils (protected from erosion and oxidation), solid phases of FeS(s) and FeS2(s) still persist. When oxidized, these phases produce electron acceptors [FeIII and sulfate] available to FeRB and SRB.

Material and Methods Site Descriptions. A total of 47 forest stands dominated (>70% of tree biomass) by Picea abies and subjected to clear-cutting during the period 1998-2007, and 10 mature (>70 years old) Picea abies stands, were selected as treatments and as reference sites, respectively (Supporting Information (SI) Figure S1). The sites were selected to be dominated by upland areas on glacial till, with Vaccinium shrubs and feather mosses dominating the field and bottom layers. No site had any significant contribution from wetlands. A criterion for site selection was that a first-order stream was generated and collected water from a significant area within the borders of the reference or clear-cut site and that runoff contributions from sources outside the site were negligible. The absolute area contributing to the runoff in each sampled stream was not explicitly determined, but by selecting streams of a similar size (having a width of 0.5-1.0 m at the sampling point, located at the downstream boundary of the site), the area contributing to the runoff was considered to be of similar order of magnitude for all sites. The area of reference stands was in average 4.9 ( 1.6 ha above the HC and 8.5 ( 5.4 ha below the HC. All reference sites were drained by one major stream (which were sampled), but the stream in several cases did not drain the whole area of the stand. Because of economical considerations, several mature Picea abies stands differing somewhat in age may be taken together to a clearcut. Therefore the area of the clear-cuts varied much more in size (range: 1.8-52.7 ha, average ( SD: 22.4 ( 13.5 ha above the HC; range: 4.7-44.5 ha, average: 14.3 ( 10.5 ha below the HC) than reference stands. For larger clear-cuts, which were drained by multiple streams, one stream in the center of the site was selected. Sites were selected well above and below the highest postglacial coastline (HC), which is located between 200-230 meters-above-sea-level (m.a.s.l.) in the study area. Most sites selected below this boundary were affected by marine conditions during the Litorina Sea period 7000-4000 BP. Twenty-two of the clear-cuts were situated above (315-491 m.a.s.l.) and 25 clear-cuts were situated below (13-193 m.a.s.l) the HC. Of the 10 reference stands, four were situated above (308-435 m.a.s.l.) and six below (14-192 m.a.s.l.) the HC. A complete list of all sites is presented in Table S1 (Supporting Information). The mean annual temperature in the study area was about +3 °C and the mean annual precipitation was 650 mm during the period 1961-1990 (18). The bedrock of the area is dominated by sedimentary gneiss with 6% Fe, mainly in biotite (19). Glacial till deposits dominate above the HC. Below the HC the finest material is lost from the till by wave action, but silty material is commonly found in protected pockets along slopes and in valley bottoms. Below the HC, solid phases of FeS(s) and FeS2(s) are found in reduced horizons of silty sediments (19). 8536

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 43, NO. 22, 2009

Sampling of Stream Water and Chemical Analyses. All streams were sampled once during a two-week period in the beginning of August 2007. During this period no major precipitation event (>5 mm day-1) occurred in the study area. Following standard clean-sampling protocols for mercury, 150 mL of streamwater was collected in acid washed Teflon PFA (Nalgene) bottles kept in double-plastic zip-bags and transported in a cooler. Within one day, subsamples for total Hg (Hgtot ) HgII + MeHg) analyzes were transformed to acid washed HDPE bottles, and acidified to pH 1 with HNO3. According to ref 20, filtration of water samples should be done immediately under nonoxidized conditions, or otherwise better avoided (because of losses of HgII and MeHg adsorbed to newly precipitated Fe-oxy/hydroxides). In this study the generally low-turbidity conditions in the small streams facilitated sampling without particle contamination. Comparison of a subset of filtered and unfiltered samples showed no significant differences in MeHg, Hgtot or TOC (data not shown), and therefore only chemical analyses determined on unfiltered samples are reported. Field blanks were prepared every sixth sample by exposing Milli-Q water for the same treatment (the bottled was kept open during sampling) as the stream samples. None of the 10 blanks showed Hgtot or MeHg concentrations above the detection limit (data not shown). For ancillary chemistry analyses, another 100 mL of streamwater was collected in PET bottles. Concentrations of MeHg were determined using isotope dilution analysis. To account for possible losses during storage, sample preparation and analyses, an isotope enriched (97.7%) Me202Hg-standard was added to the sample in tightly sealed Falcon polypropylene tubes at 4 °C within a day after sampling. After storage for maximum two weeks in a darkness at 4 °C, samples were ethylated by NaB(C2H5)4, and the derivatized MeHg was purged and trapped on Tenax adsorbent columns (21). Ethylated MeHg was desorbed and analyzed in a GC-ICPMS system (Agilent 7500 ICPMS, Agilent 6890N GC) (22). The limit of detection, calculated as 3 × SD of blanks, was 0.035 ng L-1. Concentrations of Hgtot were determined using cold vapor atomic fluorescence spectroscopy (CVAFS, PSA Millenium Mercury Analyzer 10.035), after oxidation of organic matter in the sample by BrCl (U.S. Environmental Protection Agency standard method 1631). The limit of detection was estimated to be 0.3 ng L-1. Concentrations of inorganic HgII were calculated as the difference between concentrations of Hgtot and MeHg. Total concentrations of S, Fe, Mn, Na, and Ca were determined on ICP-MS (Elan 6100DRC, PerkinElmer), using Rh and Sc (Referensmaterial AB, Ulricehamn, Sweden) as internal standards, and total concentrations of Cl, NO3, PO4 and SO4 were determined by ion chromatography (Dionex 4000i). Concentrations of total organic carbon (TOC) and total inorganic carbon (TIC) were determined on a Shimadzu TOC-5000 analyzer. Total N was determined using flowinjection-analyses and spectrophotometric detection after oxidation of organic and inorganic N by peroxidisulfate. Total organic N (TON) was calculated as the difference between total N and nitrate. For samples below nitrate detection (0.01 mg L-1), a value corresponding to half of the detection limit (0.005 mg L-1) was used in the calculation of TON. Ammonium was not determined given the generally very low concentration of ammonium ions in streams typical for the study area (23). Ultraviolet absorbance at 254 nm was determined using a GBC UV/vis 920 spectrophotometer (GBC Scientific Equipment, Dandenong, Australia). Specific UVA (SUVA254 nm, L mg-1 m-1) was calculated by a division of the absorbance by TOC, after correction for the UV absorbance by Fe in accordance with Weishaar et al. (24). Statistical Analyses. The sampling scheme was designed to (1) compare each of two clear-cut age classes (0-4 years: N ) 20, 4-10 years: N ) 27) with reference sites (N ) 10) and

FIGURE 1. Mean values ( standard errors for variables showing significant differences (ANOVA, SI Table S4) among the two age classes of clear-cuts (0-4 years; N ) 20, 4-10 years; N ) 27) and reference stands (N ) 10). Significant differences (p < 0.05*, p < 0.01**) as determined by Dunnett’s two-tailed test for comparison between each of the two clear-cut age classes and the reference stands (SI Table S5). (2) compare each of the two clear-cut age classes with reference sites situated above (0-4 years: N ) 13, 4-10 years: N ) 9, reference sites: N ) 4) and below (0-4 years: N ) 7, 4-10 years: N ) 18, reference sites: N ) 6) the HC. In each of these two cases a one-way ANOVA was used to test the null-hypothesis that subsets of data were part of the same population (no significant differences). When ANOVA revealed a significant difference (p < 0.05, one-tailed F-test), a two-tailed Dunnett’s test (25) was used for pairwise comparisons of each of the two clear-cut age classes with the control (reference stands). Based on previous findings (7-9), showing significant effects the first three years, and in order to get a fairly equal number of sites, the clear-cuts were divided into 0-4 and 4-10 year-old age classes. Because the sample size of controls and reference sites were not equal, a more conservative estimate of the sample standard error, and thus the test quantity (the q-value), was used in agreement with ref 26.

Results and Discussion Stream Water Chemistry at Sites Above and Below the Highest Postglacial Coastline. The purpose of selecting sites above and below the HC was to explore possible effects of postglacial sulfur deposits on the mobilization of HgII and mobilization/production of MeHg after clear-cut. Consequently, the only ancillary variable showing a significant difference between sites above and below the HC was sulfate (SI Tables S2 and S3). Average concentrations of sulfate at all sites (irrespective of their status as clear-cuts or references) were significantly (ANOVA; p ) 0.004) higher in streams below (0.80 ( 0.40 mg L-1) as compared to above (0.54 ( 0.22 mg L-1) the HC. This reflects a higher abundance of amorphous FeS(s) and FeS2(s) in soils below the HC (19, 27). These postglacial sulfur deposits are continuously oxidized to sulfate

and Fe(III) oxy-hydroxides as a consequence of iso-static uplift (4-9 mm year-1 in the region) and drainage. Clear-Cut Effects on HgII and MeHg. A comparison of each of the two age classes of clear-cuts with the reference stands (ANOVA, SI Table S4, followed by Dunnett’s test, SI Table S5), revealed that concentrations of MeHg, as well as the MeHg TOC-1 ratio, were significantly (p < 0.01) elevated in streams draining the 0-4 year-old clear-cuts (Figure 1). The ratio HgII TOC-1 (p < 0.01), but not absolute concentrations of HgII (nor %MeHg), was significantly enhanced in streams of the 0-4 year-old clear-cuts. In streams draining the older clear-cuts, MeHg, MeHg TOC-1, HgII TOC-1, and %MeHg showed intermediate values between the younger clear-cuts and the reference stands, but in comparison with the reference stands no statistic differences (p < 0.05) were observed. When each of the two age classes of clear-cuts were compared with references sites within subsets either above or below the HC, the pattern was similar as for the total data set (even if differences were less significant owing to a smaller number of sites). Younger clear-cuts above the HC showed significantly elevated concentrations of MeHg, MeHg TOC-1, and HgII TOC-1 (p < 0.05), as compared to reference stands (ANOVA, SI Table S6, and Dunnett’s test SI Table S7). No significant differences were observed for the older clearcuts. Differences between clear-cut age-classes and references followed the same pattern below the HC, but no differences were significant at the 5% level (ANOVA, SI Table S8, and Dunnett’s test SI Table S9). Clear-Cut Effects on Organic Carbon, Nitrogen, Manganese, Iron, and Sulfur. Streams draining the youngest clear-cuts had lower concentrations of sulfate and higher concentrations of Fe and nitrate than references, but differences were nonsignificant (SI Table S4). Neither did VOL. 43, NO. 22, 2009 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

8537

FIGURE 2. Relationships between TOC, sulfate, Fe TOC-1, and HgII and MeHg. Open symbols represent 0-4 year-old clear cuts, black, filled symbols represent 4-10 year-old clear-cuts and gray, filled circles represent reference stands. Triangles represent clear-cut sites above and diamonds clear-cut sites below the HC. Illustrated relationships with TOC are linear and relationships with sulfate and Fe TOC-1 are on the form y ) axb and y ) aebx, respectively. TOC, TOC TON-1 ratio, or organic matter aromaticity (SUVA254 nm) differ significantly between any of the clear-cut age classes and references. This is in line with the study by O’Driscoll et al. (28), in which no changes in the overall aromaticity of DOC were detected in 23 lakes subjected to upstream clear-cut, as compared to lakes unaffected by clearcutting. Possible changes in stream organic matter quality induced by clear-cutting might simply be diluted in the large pool of aromatic C associated to more persistent, macromolecular (humic) material. In contrast, concentrations of Mn showed a clear trend in our study, decreasing in the order 0-4 year-old clear-cuts > 4-10 year-old clear-cuts > references. In the full data set, concentrations of Mn were significantly (p < 0.02) elevated in streams of the youngest clear-cuts (61 µg L-1), as compared to references (16 µg L-1), Figure 1. In subsets of sites above and below the HC, the elevation of Mn in streams of the youngest clear-cuts, relative to references, was almost significant (SI Tables S7 and S9). Elevated concentrations of Mn may be due to redox-processes (reduction of Mn (III, IV) to Mn(II)), but since no significant solubility effects were seen for Fe after clear-cutting (an increase in total Fe would indicate formation of Fe(II)), we suggest that the elevation of Mn mainly was due to a release of Mn2+ ions from new clear-cut “slash” and enhanced degradation of soil organic matter (SOM). It is well-known that Mn2+ (similar to K+ and Cs+, (29)) is readily released (leached) from fresh litter and degraded SOM as a consequence of clear-cutting (30-32). Thus, elevated concentrations Mn in streams of younger 8538

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 43, NO. 22, 2009

clear-cuts can be interpreted as an indication of an increased availability of easily degradable organic matter. Degradation of SOM (initially resulting in a loss of mainly C) combined with a production of new microbial biomass (bacteria having a C N1- ratio of 5-8; (33)), may explain a lower (yet not statistically significant) TOC TON-1 ratio in streams of younger clear-cuts (SI Table S5). In a study in Finland, no significant time trend was seen on the DOC DON-1 ratio (49-63) 1-6 years after clear-cutting Picea abies stands on former peatland (34), but no data were reported for uncut reference stands. Are Elevated Concentrations of MeHg after Clear-Cut Due to Mobilization from Soil or New Production of MeHg? The concentration of HgII was significantly, positively correlated to TOC across all sites (R2 ) 0.73, p < 0.001, relationship not shown), as well as in subsets of clear-cut age classes and reference stands (Figure 2a). This observation indicates that HgII in streams mainly was mobilized and transported from soil to stream associated to dissolved or colloidal natural organic matter (NOM). The strong correlation is in agreement with several previous studies and is explained by the high affinity of HgII for thiol groups in NOM (35). The relationship across all sites was somewhat stronger between HgII and the UV absorbance at 254 nm (y ) 4.29x + 0.54, R2 ) 0.78, p < 0.001), in agreement with ref 36. In the latter study, the slope and intercept of relationships between Hgtot (the contribution from MeHg was not reported) and UV absorbance did not show any statistical differences (p > 0.05) among three different streams draining forested

watersheds in Vermont and New Hampshire, despite the fact that data covered both base and high flow periods with highly varying concentrations of DOC and Hgtot. Given that runoff at different seasons reflects different soil depths and compartments in the landscape, this indicates that the concentration of Hg in relation to DOC (or its UV absorbing moieties), within a restricted area (having similar atmospheric loads of Hg), is quite independent of spatial and depth-related variations in SOM contents and quality. The relationship was strong enough (R2 ) 0.92, p < 0.001) for the authors to suggest UV absorbance as a good proxy for Hgtot. Also in our study, the slopes of the HgII v.s. TOC (or UV absorbance) relationships were not significantly different for the three data sets. The intercept, however, was significantly higher (q ) 2.3, q0.05(2),50,2 ) 2.01, p < 0.05, Dunnett’s test) in the 0-4 year-old clear-cuts as compared to the reference stands. Thus, the relationship between HgII and TOC was significantly, shifted to a higher level for the 0-4 year-old clear-cuts. This is another way to illustrate the significantly enhanced HgII TOC-1 ratio (Figure 1c). This finding suggests that in addition to TOC, other factors contributed to the release of HgII into the streams of young clear-cuts. Because HgII is fairly conservative in the sense that it is not substantially produced or degraded in soil, an elevation of the HgII TOC-1 ratio may be taken as an indication of mobilization with other agents than TOC. Since HgII is known to associate exclusively with reduced sulfur, dissolved and/ or colloidal inorganic sulfides are the only reasonable candidates. It is well-know from experimental studies (37) and theoretical modeling (14) that the solubility of HgII increases substantially with an increase in concentrations of HS-, also under the formation of HgS(s). This is due to the formation of soluble Hg(SH)20, HgS2H-, and HgS22- complexes. Under suboxic conditions, soluble Hg-polysulfides may form, which will further enhance the HgII solubility. It is also possible that HgII is mobilized as HgS(s) or in association to colloidal FeS(s). In order to prove this, measurements need to be taken in soil pore waters and along hydrological pathways in soil (which was not part of this study). Because of rapid oxidation, it would not be meaningful to measure dissolved sulfides in the stream. The Hg-sulfide complexes may be more stable, but their concentrations are too low for proper separation and detection with methods currently available. In contrast to HgII, linear relationships between MeHg and TOC were nonsignificant; across all sites (R2 ) 0.06, N ) 57, plot not shown), as well as for subsets of clear-cut age classes and reference stands (Figure 2b). Together with the significantly elevated concentrations of MeHg and MeHg TOC-1 ratio (Figure 1), this implies that additional processes than mobilization of MeHg-TOC complexes from soil to stream contributed to the elevated MeHg concentrations in streams draining 0-4 year-old clear-cuts. Similar to HgII, MeHg may be mobilized from soil to stream with dissolved (MeHgSH and MeHgS-) and colloidal inorganic sulfides (adsorbed to FeS(s) and other metal sulfides). Because the effect of HS- on the solubility of HgII and MeHg differs and varies with environmental conditions like formation of HgS(s), Hg-polysulfides etc., a correct estimate would require thermodynamic calculations (14). However, in lack of relevant data from the soils generating the stream runoff, a rough estimate could be made by assuming a similar enhanced mobilization of HgII and MeHg by other agents than TOC after clear-cut. If the difference of 60 ng g-1 in HgII TOC-1 ratio between 0-4 year-old clear-cuts (204 ng g-1) and reference stands (144 ng g-1) is attributed to an increased mobilization of HgII from soil to stream as a consequence of clear-cutting, this corresponds to a 42% enhancement in relation to the reference stands. A similar relative enhancement of MeHg caused by mobilization would result in a

corrected MeHg TOC-1 ratio of 8.1 ng g-1 (reference stands ) 5.7 ng g-1 + 42% increase of 2.4 ng g-1). The remaining difference of 10.8 ng g-1 (between 0-4 year-old clear-cuts )18.9 ng g-1 and the corrected value of 8.1 ng g-1) could be attributed to a new methylation of MeHg as a consequence of clear-cut. This simple calculation suggests that roughly 1/6 (2.4/(18.9-5.7)) of the increase in MeHg 0-4 years after clear-cut is due to an increased mobilization of MeHg from soil and 5/6 is due to new methylation of HgII. Methylation of HgII is well-known to be mediated by FeRB and SRB, and the new methylation suggested to be caused by clear-cutting should be attributed to these two groups of bacteria. The activity of these bacteria is commonly measured as Fe(II) and sulfide production, respectively. The insignificant differences in concentrations of sulfate and Fe in streams draining clear-cuts and references suggest that possible changes in soil redox-conditions after clear-cutting could not be easily traced in the runoff. On the other hand, a significant, negative relationship obtained between MeHg and sulfate (R2 ) 0.16, N ) 57, p < 0.001, exponential relationship not shown) and significant, positive relationships with total Fe (R2 ) 0.27, N ) 57, p < 0.001, exponential relationship not shown) in the streams across all sites may indicate that redox processes involving sulfate and Fe were related to methylation and mobilization of MeHg into streams. In Figure 2c and 2d relationships between MeHg and sulfate and between MeHg and Fe TOC-1 are shown for clearcut age-classes and references. Because sulfides have a higher tendency to be adsorbed and precipitated (as FeS) in the soil than the more mobile sulfate ion, decreased concentrations of sulfate in the draining stream may be taken as a rough, indirect measure of sulfate reduction. Total Fe in stream can be considered as a sum of soil mobilized Fe(II) (dominated by the free Fe2+-ion under the acidic conditions in most streams) and Fe(III) (Fe(OH)3-nn+ ions complexed by NOM and possible inorganic Fe-oxyhydroxide colloids). Under the assumption that Fe(III) mainly is complexed by (or in the case of colloids, associated with) TOC, an increase in the Fe TOC-1 ratio in stream can be considered a better estimate of Fe2+ production in soils than total concentrations of dissolved Fe. While the relationships in panels c and d of Figure 2 may be invoked as an indication that variations in MeHg concentrations within subsets of clear-cut age classes and references in part can be related to the activities of SRB and FeRB (note, however, that this interpretation relies heavily on the assumption that available sulfate and Fe(III) TOC-1 are roughly equal across sites), the most striking pattern that we want to illustrate is the vertical, parallel shifts of the relationships. The ranges of TOC, sulfate, and Fe TOC-1 are similar for the three subsets, with 0-4 year-old clear-cuts shifted to higher MeHg concentrations than older clear-cuts, which in turn are shifted to higher concentrations than the references. This is another way to illustrate the previously shown significant (p < 0.01) enhancement of concentrations of MeHg (c.f. Figure 1), and lack of significant differences in TOC, sulfate, and Fe, between the young clear-cuts and references. This suggests that the enhancement of MeHg after clear-cut (illustrated by the parallel shifts of relationships in Figure 2b-d) is mainly an effect of another factor. We suggest this factor is availability of easily degradable organic matter; the energy source for SRB and FeRB. As discussed above, this interpretation is supported by the significant enhancement in the concentration of Mn in streams draining the youngest clear-cuts. Why are MeHg Concentrations in Stream Draining Young Clear-Cuts more Elevated Above than Below the Highest Postglacial Coastline? Sites were selected above and below the HC in order to evaluate possible effects of VOL. 43, NO. 22, 2009 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

8539

postglacial sulfur deposits, and their release of sulfate, on the production of MeHg after clear-cut. Because the elevation of MeHg after clear-cut was less pronounced below the HC, it may suggest that availability of electron-acceptors (sulfate) for SRB was not a limiting factor for the production of MeHg. The topography in general is more hilly above than below the HC, where, at least at some sites, conditions for Fe(III) and sulfate reduction could be assumed more common also in the lowers slopes of mature forest stands. We may therefore speculate that top-soils changing from oxic to iron(III) reducing conditions were of larger importance for the clearcut effect above the HC. Some support for this is given in the stream data, even if differences were not statistically significant. Both Fe concentrations and the Fe TOC-1 ratio were elevated (50%) in streams of young clear-cut above the HC, as compared to reference stands (p < 0.13-10.6, ANOVA, SI Table S6), and no major differences were seen between clearcuts and references for sulfate. Below the HC differences in sulfate or Fe (or Fe TOC-1) between young clear-cuts were highly insignificant (p < 0.60, ANOVA, SI Table S8). It should, however, be noted that higher abundance of sulfur in soils below the HC may blur the indirect link between SRB activity on sulfate in streams. Studies of S isotopes in streams situated below the HC in the area of the study showed that dissimilatory sulfate reduction is an important process affecting streamwater (27). Follow-up, detailed processoriented studies in soils along hydrological pathways of clearcuts and references, as well as experimental studies, are needed in order to quantify the contribution from mobilization and methylation processes, and in particular to discern the role of FeRB and SRB. Given that the forest productivity generally decreases with altitude (slightly cooler climate and shorter growing season), the mass of organic debris left after clear-cut, if any difference, in average should be higher below the HC. Even if no obvious differences were observed, the composition and development of plants like shrubs, herbs, and grasses after clear-cut may possibly differ slightly in relation to the HC. This also includes possible differences in decomposition of below-ground litter (like dead fine roots) and root-exudation of low molecular mass organic molecules. The effects of clear-cut on degradation of soil OM and organic debris after clear-cut needs to be addressed in follow-up studies. Because the average quality of DOM (SUVA254 nm) in draining stream does not seem to change after clear-cut (this study and ref 28), focus in future studies should be on production of low molecular mass organic substances (readily available for FeRB and SRB) as a consequence of clear-cut. This study does not give any information about the actual meso- or microsites in which anoxic conditions and subsequent production and mobilization of MeHg is supposed to occur. The proportion of upland, discharge, and wetland areas was not explicitly determined for each site. While no obvious differences were observed among sites situated above and below the HC, there could be differences in quantity and quality of, e.g., patches of small wetlands that are of importance for MeHg production after clear-cut.

Acknowledgments Ola Kåren, Lars Carlsson, and Ulf Niederbach, Holmen Skog AB, are gratefully acknowledged for their invaluable help during initiation of this project and selection of sites. Mats Ho¨gstro¨m is acknowledged for help with Figure S1. A special thanks to Ida Fredriksson, Anna Karlsson, and Pia Kärrhage for assistance in the lab. Funding was provided by Holmen Skog AB, the Kempe Foundation and by the Forest Faculty, SLU, Umeå. 8540

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 43, NO. 22, 2009

Supporting Information Available Map with site locations, tables with output from calculations of ANOVA and Dunnett’s test. This material is available free of charge via the Internet at http://pubs.acs.org.

Literature Cited (1) St. Louis, V. L.; Rudd, J. W. M.; Kelly, C. A.; Beaty, K. G.; Flett, R. J.; Roulet, N. T. Production and loss of methylmercury and loss of total mercury from boreal forest catchments containing different types of wetlands. Environ. Sci. Technol. 1996, 30, 2719– 2729. (2) Gilmour, C. C.; Henry, E. A.; Mitchell, R. Sulfate stimulation of mercury methylation in freshwater sediments. Environ. Sci. Technol. 1992, 26, 2281–2287. (3) Benoit, J. M.; Gilmour, C. C.; Mason, R. P. The influence of sulfide on solid-phase mercury bioavailability for methylation by pure cultures of Desulfobulbus propionicus (1pr3). Environ. Sci. Technol. 2001, 35, 127–132. (4) Kerin, E. J.; Gilmour, C. C.; Roden, E.; Suzuki, M. T.; Coates, J. D.; Mason, R. P. Mercury methylation by dissimilatory ironreducing bacteria. Appl. Environ. Microbiol. 2006, 72, 7919– 7921. (5) Fleming, E. J.; Mack, E. E.; Green, P. G.; Nelson, D. C. Mercury methylation from unexpected sources: Molybdate-inhibited freshwater sediments and an iron-reducing bacterium. Appl. Environ. Microbiol. 2006, 72, 457–464. (6) Schaefer, J. K.; Morel, F. M. M. High methylation rates of mercury bound to cysteine by Geobacter sulfurreducens. Nat. Geosci., 2009, 2, 123-126, DOI: 10.1038/NGEO412. (7) Garcia, E.; Carignan, R. Impact of wildfire and clear-cutting in the boreal forest on methyl mercury in zooplankton. Can. J. Fish. Aquat. Sci. 1999, 56, 339–345. (8) Garcia, E.; Carignan, R. Mercury concentrations in northern pike (Esox lucius) from boreal lakes with logged, burned, or undisturbed catchments. Can. J. Fish. Aquat. Sci. 2000, 57, 129– 135. (9) Garcia, E.; Carignan, R. Lean, D. R. S. Seasonal and inter-annual variations in methyl mercury concentrations in zooplankton from boreal lakes impacted by deforestation and natural forest fires. Environ. Monit. Assess. 2007, 131, 1–11. (10) Porvari, P.; Verta, M.; Munthe, J.; Haapanen, M. Forestry practices increase mercury and methyl mercury output from boreal forest catchments. Environ. Sci. Technol. 2003, 37, 2389–2393. (11) Porvari, P.; Verta, M.; Linjama, J.; Munthe, J. Forestry practices cause extreme mercury and methyl mercury output from boreal forest catchments. J. Roy. Swed. Acad. Agric. For. 2009, 148 (1), 34–37. (12) Allan, C. J.; Heyes, A.; Mackereth, R. J. Changes to groundwater and surface water Hg transport following clearcut logging: a Canadian case study. J. Roy. Swed. Acad. Agric. For. 2009, 148 (1), 50–54. (13) Bishop, K.; Allan, C. J.; Bringmark, L.; Garcia, E.; Hellsten, S.; Ho¨gbom, L.; Johansson, K.; Lomander, A.; Meili, M.; Munthe, J.; Nilsson, M.; Porvari, P.; Skyllberg, U.; Sorensen, R.; Zetterberg, T.; Åkerblom, S. Forestry’s contribution to Hg bioaccumulation in freshwaters: assessment of available evidence. J. Roy. Swed. Acad. Agric. For. 2009, 148 (1), 9–23. (14) Skyllberg, U. Competition among thiols and inorganic sulfides and polysulfides for Hg and MeHg in wetland soils and sediments under suboxic conditions: Illumination of controversies and implications for MeHg net production. J. Geophys. Res. 2008, 113, G00C03, DOI: 10.1029/2008JG000745. (15) Deonarine, A.; Hsu-Kim, H. Precipitation of mercuric sulfide nanoparticles in NOM-containing water: implications for the natural environment. Environ. Sci. Technol. 2009, 43, 2368– 2373. (16) Boman, A.; Åstro¨m, M.; Fro¨jdo¨, S. Sulfur dynamics in boreal acid sulfate soils rich in metastable iron sulfide - the role of artificial drainage. Chem. Geol. 2008, 255, 68–77. ¨ born, I. Geochemistry and partitioning of trace (17) Sohlenius, G.; O metals in acid sulphate soils in Sweden and Finland before and after sulphide oxidation. Geoderma 2004, 122, 167–175. (18) Alexandersson, H.; Karlstro¨m, C.; Larsson-McCann, S. Temperature and Precipitation in Sweden 1961-1990; Reference normals. The Swedish Meteorological and Hydrological Institute: Norrko¨ping, Sweden, 1991; Vol. 81, p 87. (19) Jansson, M.; Ivarsson, H. Causes of acidity in the River Lillån in the coastal zone of central northern Sweden. J. Hydrol. 1994, 160, 71–87.

(20) Drott, A.; Lambertsson, L.; Bjo¨rn, E.; Skyllberg, U. Effects of oxic and anoxic filtration on determined methyl mercury concentrations in sediment pore water. Mar. Chem. 2007, 103, 76–83. (21) Lambertsson, L.; Bjo¨rn, E. Validation of a simplified field-adapted procedure for routine determinations of methyl mercury at trace levels in natural water samples using species-specific isotope dilution mass spectrometry. Anal. Bioanal. Chem. 2004, 380, 871–875. (22) Larsson, T.; Bjo¨rn, E.; Frech, W. Species specific isotope dilution with on line derivatisation for determination of gaseous mercury species. J. Anal. At. Spectrom. 2005, 20, 1232–1239. (23) Petrone, K.; Buffam, I.; Laudon, H. Hydrologic and biotic control of nitrogen export during snow melt: A combined conservative and reactive tracer approach Water Resour. Res 2007, 43, W06420, DOI: 10.1029/2006WR005286. (24) Weishaar, J. L.; Aiken, G. R.; Bergamaschi, B. A.; Fram, M. S.; Fujii, R.; Mopper, K. Evaluation of specific ultraviolet as an indicator of the chemical composition and reactivity of dissolved organic carbon. Environ. Sci. Technol. 2003, 37, 4702–4708. (25) Dunnett, C. W. A multiple comparison procedure for comparing several treatments with a control. J. Am. Stat. Assoc. 1955, 50, 1096–1121. (26) Zar, J. H. Biostatistical Analysis, 2nd ed.; Prentice-hall, Inc.: Engelwood Cliffs, NJ, 1984; p 718. (27) Mo¨rth, C.-M.; Laudon, H.; Mellqvist, E.; Torssander, P.; Giesler, R. Sources of stream water sulfate during the spring snowmelt in boreal streams: Evidence from 34δS isotope measurements. J. Geophysical Res 2008, 113, G01005, DOI: 10.1029/ 2007JG000457. (28) O’Driscoll, N. J.; Siciliano, S. D.; Peak, D.; Carignan, R.; Lean, D. R. S. The influence of forestry activity on the structure of dissolved organic matter in lakes: Implications for mercury photoreactions. Sci. Total Environ. 2006, 366, 880–893.

(29) Berthelsen, B. O.; Steinnes, E.; Naeumann, R. Distribution of 137Cs in surface soils as affected by forest clear-cutting. J. Environ. Rad. 1999, 42, 39–49. (30) Fuller, R. D.; Simone, D. M.; Driscoll, C. T. Forest clearcutting effects on trace metal concentrations: Spatial patterns in soil solutions and streams. Wat. Air Soil Pollut. 1988, 40, 185–195. ¨ rlander, G. Effect of conventional (31) Gronflaten, L. K.; Steinnes, E.; O and whole-tree clear-cutting on concentrations of some micronutrients in coniferous forest soil and plants. For. Stud. 2008, 48, 5–16. (32) Heal, K. V. Manganese and land-use in upland catchments in Scotland. Sci. Total Environ. 2001, 265, 169–179. (33) Fenchel, T.; King, G. M.; Blackburn, T. H. Bacterial Biogeochemistry: The Ecophysiology of Mineral Cycling; Elsevier Academic Press: Amsterdam, 1998; p 307. (34) Nieminen, M. Export of dissolved organic carbon, nitrogen and phosporus following clear-cutting of three Norway Spruce forests growing on drained peatlands in southern Finland. Silva Fennica 2004, 38 (2), 123–132. (35) Skyllberg, U.; Bloom, P. R.; Qian, J.; Lin, C.-M.; Bleam, W. F. Complexation of mercury (II) in soil organic matter: EXAFS evidence for linear two-coordination with reduced sulfur groups. Environ. Sci. Technol. 2006, 40, 4147–4180. (36) Dittman, J. A.; Shanley, J. B.; Driscoll, C. T.; Aiken, G. R.; Chalmers, A. T.; Towse, J. E. Ultraviolet absorbance as a proxy for total dissolved mercury in streams. Environ. Pollut. 2009, 157, 1953– 1956. (37) Jay, J. A.; Morel, F. M. M.; Hemond, H. F. Mercury speciation in the presence of polysulfides. Environ. Sci. Technol. 2000, 34, 2196–2200.

ES900996Z

VOL. 43, NO. 22, 2009 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

8541