Impact of Forestry on Total and Methyl-Mercury in ... - ACS Publications

Mar 25, 2014 - The combined effects of logging and site preparation increased total and ... indicated many-fold increases in the global deposition of ...
2 downloads 0 Views 3MB Size
Article pubs.acs.org/est

Impact of Forestry on Total and Methyl-Mercury in Surface Waters: Distinguishing Effects of Logging and Site Preparation Karin Eklöf,*,†,▼ Jakob Schelker,‡ Rasmus Sørensen,† Markus Meili,§ Hjalmar Laudon,‡ Claudia von Brömssen,∥ and Kevin Bishop†,⊥ †

Department of Aquatic Sciences and Assessment, Swedish University of Agricultural Sciences, Box 7050, SE-75007, Uppsala, Sweden Department of Forest Ecology and Management, Swedish University of Agricultural Science, SE-901 83, Umeå, Sweden § Department of Applied Environmental Science, Stockholm University, SE-106 91 Stockholm, Sweden ∥ Department of Economics, Unit of Applied Statistics and Mathematics, Swedish University of Agricultural Sciences, Box 7013, SE-750 07, Uppsala, Sweden ⊥ Department of Earth Sciences, Box 256, SE-751 05,Uppsala University, Uppsala, Sweden ‡

S Supporting Information *

ABSTRACT: Forestry operations can increase the export of mercury (both total and methyl) to surface waters. However, little is known about the relative contribution of different forestry practices. We address this question using a pairedcatchment study that distinguishes the effects of site preparation from the antecedent logging. Runoff water from three catchments, two harvested and one untreated control, was sampled biweekly during one year prior to logging, two years after logging, and three years after site preparation. The logging alone did not significantly increase the concentrations of either total or methyl-mercury in runoff, but export increased by 50−70% in one of the harvested catchments as a consequence of increased runoff volume. The combined effects of logging and site preparation increased total and methyl-mercury concentrations by 30−50% relative to preharvest conditions in both treated catchments. The more pronounced concentration effect after site preparation compared to logging could be related to site preparation being conducted during summer. This caused more soil disturbance than logging, which was done during winter with snow covering the ground. The results suggest that the cumulative impact of forest harvest on catchment mercury outputs depends on when and how forestry operations are implemented.



organic carbon (DOC), and other nutrients.10,11 Furthermore, strong increases in concentrations and loadings of the more bioavailable and toxic form of Hg, methyl-mercury (MeHg) have been observed in runoff water after forestry operations in catchments in Finland12 and Sweden.13,14 As a consequence, Hg concentrations in downstream fish,15 zooplankton,16 and periphyton17 have been found to increase after harvest in some Canadian catchments. On the basis of previous studies in hemiboreal catchments in Scandinavia and Canada, Bishop et al.9 suggested that 9−23% of the Hg in the fish of Swedish inland waters could be attributed to forest harvesting activities. Productive forest management involves a number of different forestry practices and thereby different ecosystem disturbances. Logging itself and subsequent site preparation could have

INTRODUCTION High concentrations of mercury (Hg) in fish are a major concern in surface waters of many forested landscapes in the hemiboreal zone. In Sweden, Hg concentrations in tissues of freshwater fish exceed the European Union threshold of 0.02 mg of Hg kg−11 in most of the country’s hundred thousand lakes.2 Such excesses are also common in Finland, Canada, and the U.S. This large-scale contamination of freshwater fish is mainly attributed to anthropogenic emissions of Hg deposited after long-range transport.3 Studies of ice cores4 and sediment cores5,6 in North America have indicated many-fold increases in the global deposition of mercury since preindustrial times. In Europe, Hg deposition has declined slightly since the 1960s.7 However, given that soils still contain large amounts of stored Hg8 there is a risk that forestry operations will increase the mobilization of Hg into aquatic ecosystems. Forestry operations in hemiboreal areas in Scandinavia and North America have been identified to increase aquatic loading and concentration of Hg,9 as well as sediments, dissolved © 2014 American Chemical Society

Received: Revised: Accepted: Published: 4690

November 1, 2013 March 22, 2014 March 25, 2014 March 25, 2014 dx.doi.org/10.1021/es404879p | Environ. Sci. Technol. 2014, 48, 4690−4698

Environmental Science & Technology

Article

tree removal. (4) Responses will increase with harvest intensity but be reduced by buffer zones.

different influences on the movement of terrestrial Hg into aquatic ecosystems, as well as the net methylation of Hg. Site preparation refers to the disturbance of soils prior to planting of new trees which is a common procedure to increase survival of newly planted seedlings.18 In Sweden, during recent decades, this mechanical soil treatment is often performed by disctrenching, which exposes the mineral soil and forms mounds or ridges where the new seeds or seedlings can be planted. A previous study conducted by Sørensen et al.19 in the Balsjö paired catchment study in northern Sweden that investigated the effects on Hg after logging but prior to site preparation reported only moderate increases in total Hg (THg) concentrations (around 15%) and no significant changes in MeHg concentrations. The lack of any strong effects on Hg concentrations at Balsjö was suggested to be the result of minimal soil disturbance because the logging was conducted during winter when snow protected the ground. This study also raised the question of whether subsequent site preparation operations are the agent of strong increases of Hg concentrations19 that have been reported in some other studies.12,9 However, the study by Sørensen et al.19 noted an increase in export of THg and MeHg of up to 63%, mainly caused by an increase in runoff. Decreased transpiration loss due to removal of vegetation,20 as well as greater snow accumulation in open areas compared to those covered by forest,21 result in increased groundwater recharge and runoff after logging.22,23 More superficial flow paths through the upper organic, Hg-rich soil horizons usually result in high concentrations of organic carbon as well as Hg in streams.24 Increased groundwater levels will not only increase the export of Hg, but the wetter and more anoxic or suboxic environments can also provide favorable conditions for the bacterial transformation of inorganic Hg to MeHg if proper substrates for methylation are available. Methylation is mainly carried out by sulfate reducing bacteria (SRB),25,26 but other bacterial communities, such as iron reducing bacteria (IRB), could contribute as well.27 Forest harvest might also change the net balance between methylation and demethylation, in favor of methylation, through the addition of fresh organic carbon sources from decomposing logging residues working as electron donors19 and increased soil temperatures due to decreased shading by the forest canopy.11 In addition to the hydrological effects of tree removal, mechanized forestry operations, during both logging and site preparation, will also cause soil compaction and affect soil structure. However, the magnitude of the disturbance caused by compaction varies with conditions such as climate, soil properties and management practices.28 Although previous studies suggest some strong Hg effects, the contributions from different types of forestry operations have not yet been carefully assessed. It is possible that site preparation will have a more pronounced effect than that from logging only. We are not aware of any previous study that attempted to separate the effect caused by logging from that of the subsequent site preparation. The aim of this study was to test four hypotheses: (1) Logging and subsequent site preparation will both result in increased THg and MeHg concentrations in runoff water. (2) The concentration response of MeHg and THg will differ because of differences in the response to soil disturbance and water saturation changes. (3) Flux responses will be more influenced by logging than site preparation due to flow increases after logging following the



SITE DESCRIPTION The Balsjö paired catchment experiment is situated in northern Sweden (N 64° 1′37″ E 18° 55′43″) approximately 70 km west of the Baltic Sea coast (Supporting Information, Figure S1). Mean annual precipitation in the area is 623 mm and mean annual air temperature is 1.8 °C for the period 2005−2011. The minimum mean monthly temperature (2005−2011) varied between −5.2 °C and −14.4 °C (December-March). The maximum mean monthly temperature varied between 12.8 and 17.0 °C (July-August). The forest in the area is typical Scandinavian hemiboreal forest dominated by Norway spruce (Picea abies) in the lower wetter areas and Scots pine (Pinus sylvestris) on the higher, well-drained soils. Birch (Betula sp.) is also present in the wetter areas, for example, along stream channels. The highest marine shoreline passes through the area, where Ref-S and North are mainly above and CC is mainly below this line. The bedrock consists of pegmatite with aplitic granite and aplite. This bedrock is covered by glacial till that is unwashed by the sea in the areas above the highest marine shoreline. The dominant soil type is orthic podzol, with histosols in the wetter areas in the valleys and along the streams. For more detailed information about the study area see Löfgren et al.18 The Balsjö paired catchment study was designed to investigate the influence of forestry operations on water quality. The first stage of the study focused on the influence of logging on different aspects of water quality, such as Hg,19 DOC,29 and other chemical variables,18 as well as hydrology.23 The experimental design consists of three catchments (Supporting Information, Figure S1) that have been monitored since 2005 for Hg. Two of the catchments (CC and North) were logged in March 2006. The logging was done using accepted standards for good forestry practices.30 The harvest was carried out during winter on soils covered with snow. Wetlands were not harvested and harvesting within 10 m of the stream was avoided. All the logged areas were site prepared by disctrenching when the soil was free of snow and soil frost in May 2008. The disc-trencher was fitted with sowing equipment. In this way, eleven seeds of Pinus sylvestris were sown per meter. One of the catchments was a 24 ha reference catchment (RefS) where the forest cover (less than 5% clear-cut) was left throughout the study period. Ref-S was a subcatchment of the North catchment, where the lower parts of the catchment were logged with a small buffer strip of trees left along the stream (∼5−10 m wide on each side of the stream). Approximately 35% of the 40 ha North catchment was logged. The third catchment (CC) was logged with a higher intensity and without a riparian buffer strip. The CC catchment covers 41 ha of which 64% was logged. This difference in treatment between CC and North was designed to study the influence of harvesting intensity and protective buffer zones on the treatment effects of THg and MeHg. More information about the catchments, including the basal area of removed logs during harvest, can be found in Table S1 in Supporting Information.



METHODS Field Sampling, Chemical Analyses, and Hydrological Measurements. Stream water grab samples were taken every two weeks beginning in March 2005 during the ice-free season 4691

dx.doi.org/10.1021/es404879p | Environ. Sci. Technol. 2014, 48, 4690−4698

Environmental Science & Technology

Article

Figure 1. Boxplots (5th, 25th, 50th, 75th, and 95th percentile) of the THg and MeHg concentrations and exports from the CC, North, and Ref-S catchments during the prelogging period, after logging, and after site preparation.

and once a month during the winter (typically Nov−March). This sample schedule is termed “general sampling” (around 115 samples for each parameter and catchment). During high-flow events additional samples were collected (around 35 samples for each parameter and catchment). The field sampling procedures are detailed in Supporting Information, text S1. Concentrations of both THg and MeHg were measured on unfiltered samples. Concentrations of THg were analyzed at the Stockholm University, Department of Applied Environmental Science, using cold vapor atomic fluorescence spectroscopy after oxidation by BrCl and reduction to Hg(0) with SnCl2 following the U.S. EPA (Environmental Protection Agency) standards, method-EPA 1631 (2002). The detection limit was 0.3 ng L−1 and the analytical precision was ±3% for THg in a concentration range of 5−50 ng L−1. MeHg analyses were performed at the Department of Chemistry at Umeå University by species-specific isotope dilution followed by mass spectrometry.31 The detection limit was 0.04 ng L−1 and the analytical precision determined as ±6% for concentrations of 0.25 ng L−1. The general chemical analyses included, among others, TOC and total suspended solids (TSS) according to documented standards (Supporting Information, text S1). Hourly discharge values were derived from each catchment by water level measurement, at a 90 degree V-notch weir (further information in Supporting Information, text S1). Fluxes of chemical variables were calculated from daily discharge and daily concentrations. The daily discharge data were calculated as the mean of the hourly discharges. Daily concentrations were derived by linear interpolation of measured concentrations. Data Analyses and Statistics. The site preparation period was extended from two years to three years because of contamination problems for MeHg that arose during sample storage in the MeHg laboratory for half a year in 2009/2010. Approximately two years of MeHg data thereby exist during the period after site preparation, whereas other chemical parameters include three years of data from this period.

Because of high seasonal variation and non-normality in the data, tested by a Shapiro−Wilk test for normality (p < 0.05), nonparametric statistical methods and resampling procedures were used. A randomized intervention analysis (RIA)32 was used to determine if the median values of the difference in concentration between the catchments at each sample occasion changed significantly after treatment. Daily exports were used when evaluating the forestry effects on fluxes of chemical variables and hydrology. RIA simulated the distribution of the median values by resampling the observed time series of each variable and is therefore distribution-free. Instead of randomizing the observed values without restrictions as in the standard RIA method, the resampling of the observed values was done within the specified season in order to retain the seasonal variation as in Löfgren et al.18 and Sörensen et al.19 Each RIA was based on 2000 simulations. The general sampling schedule consisting of biweekly samples was used in the RIA analyses of changes in concentrations. All other statistical analyses, including the RIA analyses of effects on fluxes, were based on the whole sampling schedule that included more frequent sampling during high-flow events. The percentage changes given in this paper are calculated in similar way as RIA analysis calculated the significance of the treatment effect. The change in the median value of the difference in concentrations or fluxes between a treated and reference catchment for every sample occasion (concentrations) or day (fluxes and hydrology) after each treatment was compared to the median concentration or flux in the treated catchment before the treatment. The significance level used in the RIA analysis was set to p < 0.05. Throughout the paper, whenever we speak about a change occurring, or a significant effect, we are always referring to an effect that has been tested for this level of significance. Correlations between variables were tested by Spearmańs ρ in JMP 10, using a significance level of p < 0.05. 4692

dx.doi.org/10.1021/es404879p | Environ. Sci. Technol. 2014, 48, 4690−4698

Environmental Science & Technology

Article

Figure 2. Relationships of THg and MeHg with total organic carbon (TOC), total suspended solids (TSS) and discharge (Q), respectively. Data from the entire sampling period, 2005−2011.



during the period before logging and 20.6 mg L−1 in the period after site preparation. Possible changes after treatment were also investigated for the ratios of THg/TOC, MeHg/TOC, and MeHg/THg. None of these ratios changed significantly after logging. The ratio of MeHg/THg increased by 28% in CC relative to North after site preparation when compared with the prelogging period. The ratio of THg/TOC increased slightly (13%) in North relative to Ref-S after site preparation compared with the prelogging period. Treatment Effects on Fluxes. Changes in exports can be driven to a large extent by changes in discharge. Logging increased discharge by 47% in CC and 11% in North relative to Ref-S (Supporting Information, Tables S1 and S2). The site preparation did not have any additional effect on the discharge compared to logging only. Discharge remained greater in the period after site preparation compared to the period before logging by 38% in CC and 21% in North relative to Ref-S. As there was a higher percentage of logged area in CC (64% logged) compared to North (35% logged), it is reasonable that the flow increase was around 30% more in CC than in North both after logging and site preparation, compared to the period before logging. Logging effects on discharge in the Balsjö catchments are further discussed in Sørensen et al.23 in which the flow between 2004 and 2008 was evaluated. Median daily THg export in the period before logging was 2.7 ng m−2 day−1 from CC and 2.2 ng m−2 day−1 from North (Figure 1 and Supporting Information, Table S2). After logging, exports of THg increased by 54% from CC and 21% from North relative to the export from Ref-S (Supporting Information, Table S3). Median daily THg export during the 2 years following logging was 3.7 ng m−2 day−1 from CC and 3.1 ng m−2 day−1 from North. After site preparation, THg export from CC increased by an additional 17% relative to Ref-S when compared to logging only. When combining the effects of logging and site preparation, median THg export increased by

RESULTS Treatment Effects on Concentrations. The median concentrations of THg during the period before logging (March 2005−March 2006) were 3.8 ng L−1 in CC and 4.3 ng L−1 in North (Figure 1 and Supporting Information, Table S2). After logging, the concentrations did not increase significantly in either of these logged catchments relative to the concentrations in the reference catchment, Ref-S. After site preparation (May 2008−May 2011) the concentrations of THg increased to a median of 4.8 ng L−1 in CC and 5.6 ng L−1 in North. This corresponded to an increase by 16% in both treated catchments relative to Ref-S (Supporting Information, Table S3). The concentrations after site-preparation were also 31% higher in CC and 27% higher in North relative to Ref-S when compared with the period before logging. The MeHg concentration also did not change significantly after logging in CC and North. The site preparation, when compared with the period after logging, also did not cause any additional change of the MeHg concentration. When the effect of site preparation was combined with that of logging, by comparing the concentrations before logging with those after site preparation, the MeHg concentrations increased in the more intensively treated CC catchment by 49% relative to RefS. The median concentration of MeHg in CC was 0.24 ng L−1 in the period before logging and 0.44 ng L−1 in the period after site preparation. Possible treatment effects were also investigated for concentrations of TOC and TSS. Possible treatment effects on TSS concentrations are described in Supporting Information, text S2. Logging did not significantly change the concentrations of TOC. The combined effect of logging and site preparation increased the TOC concentrations, in one of the treated catchments, CC, by 21% relative to Ref-S. However, when comparing the TOC concentrations after site preparation to the period after logging there were no significant changes. The median concentrations of TOC in CC were 15.7 mg L−1 4693

dx.doi.org/10.1021/es404879p | Environ. Sci. Technol. 2014, 48, 4690−4698

Environmental Science & Technology

Article

Figure 3. Concentrations of THg and MeHg, as well as daily mean air temperature and discharge for the entire sampling period, 2005−2011.

from CC and 11.2 mg m−2 day−1 from North. Possible treatment effects on exports of TSS are described in Supporting Information, text S2. Temporal Controls of THg and MeHg Concentrations. The annual variation of THg and MeHg concentrations was strong, and the concentrations of THg and MeHg, followed each other well between the catchments both before and after the treatments (Figure 3). In fact, the intraannual variation of the concentrations was greater than variation between the catchments due to treatment. The drivers for the temporal variation appear to differ between THg and MeHg. Organic Matter and Particles As Drivers for THg and MeHg Concentrations. There were strong positive correlations between THg and TOC (THg-TOC) at all sites (Spearmańs ρ between 0.80 and 0.92, p < 0.001) (Figure 2). When dividing the data set by sites and treatment period (before logging, after logging and after site preparation) the median values of the THg/TOC ratios varied between 0.23 × 10−6 (CC before logging) and 0.30 × 10−6 (North after site preparation). The fact that the highest ratio was found in North after site preparation is consistent with the finding of a significant change in the THg/TOC ratio identified after site preparation in the RIA analysis. Correlations between MeHg and TOC (MeHg−TOC) were weak, but significant and positive in CC and North (p < 0.05). They were not significant in Ref-S. The correlations in CC and North were also weaker (Spearmańs ρ 0.19 in CC and 0.18 in North) than those for THg−TOC. The median values of the

51% from CC and 44% from North since the prelogging period. The increase was also 35% higher from CC as compared to North. Median daily export in the 3 years following site preparation was 4.0 ng m−2 day−1 from CC and 2.9 ng m−2 day−1 from North. Median daily MeHg export in the period before logging was 0.19 ng m−2 day−1 from both CC and North. Export increased after logging by 66% from CC relative to Ref-S. The increase was also 41% higher from CC than from North. Site preparation did not change the MeHg exports significantly compared with the period after logging, but when compared with the period before logging the exports increased by 83% from CC. No significant change was detected from the North catchment, relative to the situation either before or after logging. Median daily exports from the CC catchment increased to 0.31 ng m−2 day−1 after logging and 0.40 ng m−2 day−1 after site preparation. In the prelogging period, median daily export of TOC was 11.4 mg m−2 day−1 from CC and 9.7 mg m−2 day−1 from North. After logging the export of TOC increased by 47% from CC relative to Ref-S. Median daily export of TOC in the period after logging was 15.6 mg m−2 day−1 in CC. Site preparation did not cause any significant additional changes in TOC exports. However, when assessing the effect of logging and site preparation together, TOC export increased by 48% from CC and 21% from North relative to Ref-S. The increase in TOC export was also 36% higher from CC relative to North after both logging and site preparation. Median daily export of TOC in the 3 years following site preparation was 15.1 mg m−2 day−1 4694

dx.doi.org/10.1021/es404879p | Environ. Sci. Technol. 2014, 48, 4690−4698

Environmental Science & Technology

Article

MeHg/TOC ratios varied between 1.3 × 10−8 (Ref-S after logging) and 2.0 × 10−8 (North after site preparation). There were significant but weak negative correlations (Spearman’s ρ between −0.36 and −0.52) between THg and TSS in all sites when using data from the whole sampling period (2005−2011). The correlation between MeHg and TSS for the same period was weak but significantly positive (Spearman’s ρ between 0.22 and 0.40). Discharge and Temperature as Drivers for THg and MeHg Concentrations. Discharge (Q) was found to be an important driver for the temporal variation of THg. There were significant positive correlations between THg and Q (Spearman’s ρ between 0.76 and 0.79) at all sites during the whole sampling period (2005−2011). Furthermore, the time series of THg did not show any seasonal pattern, such as an influence caused by variation in air temperature (Figure 3). In contrast, MeHg concentrations varied more periodically and peaked every year during the warmest periods of the summer when flow was low. This occurred in all three catchments and during all periods. However, as these concentration peaks were only sampled by the regular sampling schedule (biweekly) there is a risk that we did not catch the very peak concentrations in all catchments. Therefore the concentrations during these peaks did not lend themselves to speculation on treatment effects. There was also a negative but weak correlation between MeHg and Q (Spearman’s ρ between −0.28 and −0.47). The snowmelt period, occurring every year between April first and May 31st, was identified as a time with a large impact on the annual flux of THg and MeHg. The two months of snowmelt season (17% of the year) accounted for around 25− 56% of the annual flux of THg (Supporting Information, Table S4). The discharge during spring flood accounted for 24−50% of the total annual runoff. In contrast, the loads of MeHg were lower, in some cases similar to, the average load for the entire year (16−35%).

buffer strips, it was also not possible to clearly differentiate the effect of the buffer strip from the lower proportion of harvest in North relative to the situation in CC where more of the catchment area was impacted. Buffer strips along the stream are one way to prevent forest machinery driving in the near stream zone. Driving in the near stream zone could increase the compaction and formation of driving tracks that will increase the hydrological connectivity between methylation hot-spots in the treated area and adjacent surface waters. Driving tracks in the near stream zone were present in some areas in the CC catchment. However, the natural variation of THg and MeHg in runoff water is large and the mobilization and methylation mechanisms are generally recognized to be complex. Among the mechanisms that could explain the stronger response of stream Hg concentrations to site preparation compared to logging alone, one could hypothesize a delay in the response of logging that only became evident after site preparation. Such a delay could be induced by weather conditions modifying the runoff of Hg or the establishment of methylating bacteria such as SRB in newly formed discharge areas and flooded soils. In this regard, it is worth noting that site preparation started immediately after a discharge peak, whereas logging was conducted during late-winter low-flow conditions (Figure 3). Another factor that could influence the magnitude of forestry effects originating from site preparation compared to logging is that logging was carried out during the winter when snow covered the soil, protecting it from a more severe disturbance effect.19 Site preparation was carried out in the early summer, causing more soil disturbance11 and presumably also soil compaction compared to winter logging. This suggests that if forestry operations are planned to protect water quality, practices such as performing logging on snow during winter when soil disturbance will be limited, could reduce Hg loads to receiving waters. However, such precautions are difficult for site preparation operations, because their explicit purpose is to disturb the upper soil. Climate warming is also shortening the period of snow cover in Sweden and other hemiboreal regions, reducing the opportunities for logging on snow covered ground. Although site preparation could never be done without soil damage, a successful site preparation will help re-establish the forest faster, reducing the time until the runoff starts going back down to preharvest levels. Although minimal disturbance during logging operations could be a factor that caused the minimal concentration effects of logging observed in this study, there are examples from similar forest types where stump harvest33 and logging34 have caused extensive site disturbances without significant effects on Hg concentrations. Furthermore, although concentrations responded more strongly after site preparation, at Balsjö the exports of both THg and MeHg increased more after logging than after site preparation (Figure 1 and Supporting Information, Table S3). The changes in median daily exports of THg (54% in CC and 21% in North) and MeHg (64% in CC) after logging were dominated by increases in water flow (47% in CC and 11% in North) rather than concentration changes. A large number of studies have observed increased water fluxes after forest harvest.20 The forest clearing in the Hubbard Brook experimental forest demonstrated in the 1970s that forest harvest caused increases in discharge (40% during the first year after logging).35 The study from Hubbard Brook35 found that increases in runoff were more pronounced during the growing season (June−September), similar to findings in the Balsjö catchments.23 This occurred even though snow



DISCUSSION Treatment Effects. The results of this study indicated that logging and site preparation can have different influences on Hg concentrations and exports. Site preparation generally had a stronger effect on concentrations. Logging had a greater effect on exports since logging increased flows, while site preparation did not alter runoff levels relative to the situation created by logging. THg concentrations increased after site preparation but not after logging alone. The fact that the MeHg concentrations increased significantly in the CC catchment only after the combined treatment of logging and site preparation but not after logging alone also highlights the importance of site preparation for Hg chemistry in boreal streams, as well as the importance of evaluating these treatment effects separately. The concentration response on THg from logging and site preparation together was similar between the CC and North catchment, while MeHg concentrations responded only in the CC catchment. The combined treatment effects on MeHg after both logging and site preparation were also more pronounced than that on THg in the CC catchment. In the CC catchment the stream was not protected by a buffer strip of vegetation, and a higher proportion of the catchment was treated relative to the North catchment. The lower intensity harvest in North may thereby be a reason for the lack of response on the MeHg concentrations in this catchment. While there is great hope for buffer strips, and this study could not disprove the value of 4695

dx.doi.org/10.1021/es404879p | Environ. Sci. Technol. 2014, 48, 4690−4698

Environmental Science & Technology

Article

thereby be an indication of increased activity of SRB in this site after the disturbance. THg concentrations and exports tended to be strongly correlated to discharge whereas temperature seems to be a more important driver for MeHg concentrations. Both discharge23 and soil temperature11 increased after logging in the treated areas of the Balsjö study site. Because of the strong correlation between THg and discharge, as well as MeHg and temperature, a treatment effect on the concentrations could be hypothesized after logging. However, soil disturbance was low during the winter logging operations,19 and it seems that the additional disturbance during site preparation was needed to cause significant effects on THg and MeHg concentrations. This study showed that driving damage caused by site preparation has a strong effect on both THg and MeHg concentrations in runoff, although the discharge volume will not be changed by this treatment. It is thereby of importance to avoid driving damages where possible. This study thereby suggests that the timing of forestry operations, for example during winter or summer conditions, as well as the magnitude of disturbance might be of great importance for the treatment effects on THg and MeHg.

accumulation during winter (Dec−April) increased after logging in Balsjö.22 However, the increased snow accumulation only affected snowmelt runoff during some years, as the direct evaporation from the snow surface was an important factor in other years.22 The variation in discharge is certainly of importance for fluxes of THg and MeHg. In this study, we can only speculate on whether increases in discharge during the growing season are of special importance for the MeHg in runoff because of the importance of temperature and water saturation for methylation. When the effect of site preparation was added to that of logging, the magnitude of the treatment effect on the THg concentrations in CC and North was around 30%, while for MeHg the treatment effect was 48% in CC relative to the prelogging situation. This was similar to that seen in a synoptic study in Sweden covering 54 catchments subjected to logging/ site preparation, logging/stump harvest and untreated reference conditions.36 A more severe forestry effect was found in Finland, where Porvari et al.12 reported approximately 1.9 times higher MeHg concentrations after logging and site preparation had been conducted relative to levels before the logging. The exports of THg and MeHg in the same study were several times higher in the 3 years following logging and site preparation than before these treatments.12 A strong forestry effect on MeHg was also documented in southwest Sweden where forest machines driving across a catchment disturbed the soil. The MeHg concentrations downstream from this disturbance increased by at least a factor of 3.13 Similar to our study, this indicated that soil disturbance caused by the driving of forestry machinery could result in significant forestry effects on Hg. Munthe and Hultberg13 suggested that increased MeHg concentrations were a consequence of changed water flow pathways that mobilized MeHg in the soil pool. Although the site preparation did not change the amount of discharge, driving of forestry machinery during the site preparation may change the hydrological pathways by soil compaction that reduces soil macroporosity and lowers the soil permeability.37 The infiltration capacity of water into the soil might decrease as a result of compaction.37 This could result in water ponding in driving tracks and other artificial cavities, possibly resulting in hot-spots for Hg methylation.24 The disturbance caused by forestry machinery operations has also been observed to enhance the potential for soil erosion37 and may thereby further increase the mobility of Hg bonded to soil particles. Organic Carbon, Discharge, and Temperature as Drivers for THg and MeHg Concentrations and Exports. TOC changes were closely related to the THg changes (Figure 2), which reflects the strong binding of THg to TOC/DOC, for example, refs 38−44. The ratio of THg/TOC might, however, change as a result of changed hydrological and reduction− oxidation conditions after forestry operation. The production of hydrogen sulfide (HS−) by SRB could result in increased solubility and mobility of Hg in newly waterlogged soils.14 Increased mobility of THg caused by factors other than TOC, such as increased HS− concentrations, would then increase the ratio of THg/TOC. This could explain the significant increase of the THg/TOC ratio in North that appeared after site preparation compared with the prelogging period. However, the lack of an increase of the MeHg/THg ratio argues against any increased activity of SRB after site preparation in this area, as this ratio could be used as an indicator of elevated net methylation rates.14 The significant increase of the MeHg/THg ratio in CC (relative to North) after site preparation could



ASSOCIATED CONTENT



AUTHOR INFORMATION

S Supporting Information *

Additional descriptions of the field sampling, chemical analyses and hydrological measurements (text S1) and a description of the treatment effects on TSS (text S2) as well as figure showing the study catchments (Figure S1), information about the catchments (Table S1), median concentrations and fluxes of THg, MeHg, TOC, and TSS before and after treatments (Table S2), statistical results for the change in THg, MeHg, TOC, and TSS concentrations and fluxes, as well as in the ratio of THg/ TOC, MeHg/TOC, and MeHg/THg after logging and site preparation (Table S3), and the sum and the percent yield of THg and MeHg export, as well as discharge and precipitation during the snowmelt season (Table S4) is presented in the supporting information. This material is available free of charge via the Internet at http://pubs.acs.org. Corresponding Author

*E-mail: [email protected]. Present Address

▼ Karin Eklöf: Penn State Institutes of Energy & Environment, Pennsylvania State University, 233 Forest Resource Building, University Park, PA 16802, USA

Author Contributions

The manuscript was written through contributions of all authors. All authors have given approval to the final version of the manuscript. Funding

This study was founded by Formas, ForWater, and MISTRA Future Forests. Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS The forest in this study is owned by Holmen Skog. The Balsjö catchment study is administered by Skogforsk, and special thanks are owed to Eva Ring and Lars Högbom from that organization. We thank Peder Blomkvist, Lenka Kuglerova, Ida 4696

dx.doi.org/10.1021/es404879p | Environ. Sci. Technol. 2014, 48, 4690−4698

Environmental Science & Technology

Article

(18) Löfgren, S.; Ring, E.; von Brömssen, C.; Sørensen, R.; Högbom, L. Short-term effects of clear-cutting on the water chemistry of two boreal streams in northern Sweden: A paired catchment study. Ambio 2009, 38 (7), 347−356. (19) Sørensen, R.; Meili, M.; Lambertsson, L.; von Brömssen, C.; Bishop, K. The effect of forest harvest operations on mercury and methylmercury in two boreal streams: Relatively small changes in the first two years prior to site preparation. Ambio 2009, 38 (7), 364−372. (20) Bosch, J. M.; Hewlett, J. D. A review of catchment experiments to determine the effect of vegetation changes on water yield and evapotranspiration. J. Hydrol. 1982, 55 (1−4), 3−23. (21) Murray, C. D.; Buttle, J. M. Impacts of clearcut harvesting on snow accumulation and melt in a northern hardwood forest. J. Hydrol. 2003, 271 (1−4), 197−212. (22) Schelker, J.; Kuglerová, L.; Eklöf, K.; Bishop, K.; Laudon, H. Hydrological effects of clear-cutting in a boreal forest: Snowpack dynamics, snowmelt and streamflow responses. J. Hydrol. 2013, 484, 105−114. (23) Sørensen, R.; Ring, E.; Meili, M.; Högbom, L.; Seibert, J.; Grabs, T.; Laudon, H.; Bishop, K. Forest harvest increases runoff most during low flows in two boreal streams. Ambio 2009, 38 (7), 357−363. (24) Shanley, J. B.; Bishop, K. H. Hg cycling in terrestrial watersheds. In Mercury in the Environment: Pattern and Process; Bank, M. S., Ed.; University of California Press: Oakland, CA, 2012; pp 117−139. (25) Compeau, G. C.; Bartha, R. Sulfate-reducing bacteria: Principal methylators of mercury in anoxic estuarine sediment. Appl. Environ. Microbiol. 1985, 50 (2), 498−502. (26) Gilmour, C. C.; Henry, E. A.; Mitchell, R. Sulfate stimulation of mercury methylation in freshwater sediments. Environ. Sci. Technol. 1992, 26 (11), 2281−2287. (27) 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. 2005, 72 (1), 457−464. (28) Greacen, E.; Sands, R. Compaction of forest soils. A review. Soil Res. 1980, 18 (2), 163−189. (29) Laudon, H.; Hedtjärn, J.; Schelker, J.; Bishop, K.; Sørensen, R.; Ågren, A. Response of dissolved organic carbon following forest harvesting in a boreal forest. Ambio 2009, 38, 281−286. (30) Normark, E.; Aretorn, M.; Eriksson, J.; Hedman, Y.; Hägglund, D.; Klingström, L.; Kårén, O.; Nilsson, M.; Norgren, O.; Rantaniemi, L.; Rothpfeffer, C.; Stridsman, D.; Triumf, H.; Tolblad, A.; Tretten J.; Ö sterman, B. Riktlinjer för uthålligt skogsbruk (In Swedish, “Guidelines for sustainable forestry”) Holmen Skog, 2011. For more information, see the following link: http://www.holmen.com/ Global/Holmen%20documents/Publications/Handledningar/ Riktlinjer%20f%c3%b6r%20uth%c3%a5lligt%20skogsbruk%202011. pdf?663447. (31) Lambertsson, L.; Björn, E. Validation of a simplified fieldadapted 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 (7), 871− 875. (32) Carpenter, S. R.; Frost, T. M.; Heisey, D.; Kratz, T. K. Randomized intervention analysis and the interpretation of wholeecosystem experiments. Ecology 1989, 70 (4), 1142−1152. (33) Eklöf, K.; Meili, M.; Åkerblom, S.; von Brömssen, C.; Bishop, K. Impact of stump harvest on run-off concentrations of total mercury and methylmercury. For. Ecol. Manage. 2013, 290, 83−94. (34) Allan, C. J.; Heyes, A.; Mackereth, R. J. Changes to groundwater and surface water Hg transport flollowing clearcut logging: a Canadian case study. In Does forestry contribute to mercury in Swedish fish?, Royal Swedish Academy of Agriculture and Forestry (KSLA) Report; Royal Swedish Academy: Stockholm, Sweden, 2009; Vol. 148, pp 50−54. (35) Hornbeck, J. W.; Pierce, R. S.; Federer, C. A. Streamflow changes after forest clearing in New England. Water Resour. Res. 1970, 6 (4), 1124−1132. (36) Eklöf, K.; Kraus, A.; Weyhenmeyer, G. A.; Meili, M.; Bishop, K. Forestry influence by stump harvest and site preparation on

Taberman, Viktor Sjöblom, Marsha Hanson, Pia Kärrhage, Staffan Åkerblom, Lars Lambertsson, and Erik Björn for great help in the field and the laboratory. We also thank the reviewers and Brian Branfireun for valuable comments on the manuscript.



REFERENCES

(1) EU. Directive 2008/105/EC of the European Parliament and of the Council of 16 December 2008 on environmental quality standards in the field of water policy. Off. J. Eur. Union 2008, 51/L348, 84−97. (2) Munthe, J.; Wängberg, I.; Rognerud, S.; Fjeld, E.; Verta, M.; Porvari, P.; Meili, M. Mercury in Nordic Ecosystems, Tema Nord/IVL Report B1761; Nordic Council of Ministers (NMR): Copenhagen, 2007; pp. 1−44. (3) Munthe, J.; Bodaly, R. A.; Branfireun, B. A.; Driscoll, C. T.; Gilmour, C. C.; Harris, R.; Horvat, M.; Lucotte, M.; Malm, O. Recovery of mercury-contaminated fisheries. Ambio 2007, 36 (1), 33− 44. (4) Schuster, P. F.; Krabbenhoft, D. P.; Naftz, D. L.; Cecil, L. D.; Olson, M. L.; Dewild, J. F.; Susong, D. D.; Green, J. R.; Abbott, M. L. Atmospheric mercury deposition during the last 270 years: A glacial ice core record of natural and anthropogenic sources. Environ. Sci. Technol. 2002, 36 (11), 2303−2310. (5) Lorey, P.; Driscoll, C. T. Historical trends of mercury deposition in adirondack lakes. Environ. Sci. Technol. 1999, 33 (5), 718−722. (6) Swain, E. B.; Engstrom, D. R.; Brigham, M. E.; Henning, T. A.; Brezonik, P. L. Increasing rates of atmospheric mercury deposition in midcontinental North America. Science 1992, 257 (5071), 784−787. (7) Munthe, J.; Kindbom, K.; Kruger, O.; Petersen, G.; Pacyna, J.; Iverfeldt, Å. Examining source−receptor relationships for mercury in Scandinavia; Modelled and empirical evidence. Water Air Soil Pollut. Focus 2001, 1, 299−309. (8) Johansson, K.; Bergbäck, B.; Tyler, G. Impact of atmospheric long-range transport of lead, mercury and cadmium on the Swedish forest environment. Water Air Soil Pollut. Focus 2001, 1, 279−296. (9) Bishop, K.; Allan, C.; Bringmark, L.; Garcia, E.; Hellsten, S.; Högbom, L.; Johansson, K.; Lomander, A.; Meili, M.; Munthe, J.; Nilsson, M.; Porvari, P.; Skyllberg, U.; Sörensen, R.; Zetterberg, T.; Åkerblom, S. The effects of forestry on Hg bioaccumulation in nemoral/boreal waters and recommendations for good silvicultural practice. Ambio 2009, 38 (7), 373−380. (10) Kreutzweiser, D. P.; Hazlett, P. W.; Gunn, J. M. Logging impacts on the biogeochemsitry of boreal forest soils and nutrient export to aquatic systems: A review. Environ. Rev. 2008, 16, 157−179. (11) Schelker, J.; Eklöf, K.; Bishop, K.; Laudon, H. Effects of forestry operations on dissolved organic carbon concentrations and export in boreal first-order streams. J. Geophys. Res. 2012, 117 (G1), No. G01011. (12) Porvari, P.; Verta, M.; Munthe, J.; Haapanen, M. Forestry practices increased mercury and methyl mercury output from boreal forst catchments. Environ. Sci. Technol. 2003, 37 (11), 2389−2393. (13) Munthe, J.; Hultberg, H. Mercury and methylmercury in runoff from a forested catchmentConcentrations, fluxes, and their response to manipulations. Water Air Soil Pollut. Focus 2004, 4, 607−618. (14) Skyllberg, U.; Westin, M. B.; Meili, M.; Björn, E. Elevated concentrations of methyl mercury in streams after forest clear-cut: A consequence of mobilization from soil or new methylation? Environ. Sci. Technol. 2009, 43 (22), 8535−8541. (15) 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. (16) Garcia, E.; Carignan, R.; Lean, D. R. S. Seasonal and interannual variations in methyl mercury concentrations in zooplankton from boreal lakes impacted by deforestation or natural forest fires. Environ. Monit. Assess. 2007, 131, 1−11. (17) Desrosiers, M.; Planas, D.; Mucci, A. Short-term responses to watershed logging on biomass mercury and methylmercury accumulation by periphyton in boreal lakes. Can. J. Fish. Aquat. Sci. 2006, 63 (8), 1734−1745. 4697

dx.doi.org/10.1021/es404879p | Environ. Sci. Technol. 2014, 48, 4690−4698

Environmental Science & Technology

Article

methylmercury, total mercury and other stream water chemistry parameters across a boreal landscape. Ecosystems 2012, 15 (8), 1308− 1320. (37) Kozlowski, T. T. Soil compaction and growth of woody plants. Scand. J. For. Res. 1999, 14 (6), 596−619. (38) Meili, M.; Iverfeldt, Å.; Håkanson, L. Mercury in the surface water of Swedish forest lakesConcentration, speciation and controlling factors. Water Air Soil Pollut. 1991, 56, 439−453. (39) Mierle, G.; Ingram, R. The role of humic substances in the mobilization of mercury from watersheds. Water Air Soil Pollut. 1991, 56, 349−356. (40) Driscoll, C. T.; Blette, V.; Yan, C.; Schofield, C. L.; Munson, R.; Holsapple, J. The role of dissolved organic carbon in the chemistry and bioavailability of mercury in remote Adirondack lakes. Water Air Soil Pollut. 1995, 80, 499−508. (41) Watras, C. J.; Morrison, K. A.; Host, J. S.; Bloom, N. S. Concentration of mercury species in relationship to other site-specific factors in the surface waters of nortern Wisconsin lakes. Limnol. Oceanogr. 1995, 40 (3), 556−565. (42) Scherbatskoy, T.; Shanley, J. B.; Keeler, G. J. Factors controlling mercury transport in an upland forested catchment. Water Air Soil Pollut. 1998, 105, 427−438. (43) Shanley, J. B.; Kamman, N. C.; Clair, T. A.; Chalmers, A. Physical controls on total and methylmercury concentrations in streams and lakes of the northeastern U.S.A. Ecotoxicology 2005, 14, 125−134. (44) Eklöf, K.; Fölster, J.; Sonesten, L.; Bishop, K. Spatial and temporal variation of THg concentrations in run-off water from 19 boreal catchments, 2000−2010. Environ. Pollut. 2012, 164, 102−109.

4698

dx.doi.org/10.1021/es404879p | Environ. Sci. Technol. 2014, 48, 4690−4698