S8 Forestry and mercury: Defining the connection

Research from the boreal-nemoral zone indicates a strong connection between forestry operations and the input of mercury/methylmercury to aquatic ecosystems. A synthesis of research available in 2006 concluded that between 9 and 23% of the mercury in the fish of forested Swedish lakes was released into the aquatic ecosystem in conjunction with forest operations. As more studies have been completed, however, large differences are appearing in the magnitude of the Hg response to forest operations (harvest, site preparation, drainage, drain-blocking) with respect to leakage and bioaccumulation of methylmercury. This presentation will synthesize results from a dozen recently completed and ongoing studies in Finland, Sweden, Norway and Canada. Both intensive catchment-scale studies and larger, synoptic surveys are included. In some cases mercury concentrations in waters and the biota were largely unchanged, while in other cases, manifold increases persisting for up to a decade have been observed. Understanding why there are differences in catchment sensitivity can improve management strategies to mitigate the contribution of forest operations to the mercury problem in freshwater fish. Alternative hypotheses will be considered, including the possibility that it is the increases in water tables associated with forest harvest that inundate previously oxic soils explains some of the more extreme harvest effects on methylmercury.

Groundwater and surface water fluxes of Methyl Hg (MeHg) and Total Hg (THg) were measured for five years in a 26 ha zero order Precambrian Shield hillslope-wetland complex on the Precambrian Shield in northwestern Ontario, Canada. Unlike results from earlier forest harvest studies from Finland and Sweden we measured no detectible change in MeHg or Total Hg concentrations in surface waters draining the logged catchment. Hg fluxes from the logged watershed approximately doubled and this is attributed to the increased water flux resulting from decreased evapotranspiration and interception. Post logging internal catchment fluxes of Hg changed dramatically with the hillslope THg flux increasing four times over pretreatment conditions. A significant drought episode two years after logging resulted in a threefold increase in THg upon rewetting of catchment soils but a sevenfold increase was observed in the reference catchment. Our results demonstrate the importance of both natural processes (drought) and anthropogenic land use change (logging) in modifying the terrestrial processing and transport of Hg in the Precambrian Shield environment. The enhanced flux of THg from the reference catchment in relation to the logged catchment suggests that higher soil moisture and water tables in the logged catchment may have reduced soil mineralization rates and subsequent THg release upon catchment rewetting. Our results also contribute to the growing body of measurements illustrating the variability in the Hg response to forest management activities in the boreal forest region.

Forestry operations have been reported to increase the concentrations and loads of mercury (Hg) in surface waters. In a series of catchment-scale experiments we investigated how the export of total mercury (THg) and methyl mercury (MeHg) to aquatic ecosystems was affected by logging and subsequent site preparation as well as by stump harvesting.

Stream water draining catchments subjected to forestry operations was analyzed for THg, MeHg, organic carbon and particle concentrations. The effects of logging and site preparation were studied using a paired-catchment approach in three boreal catchments (one of them an untreated control) in Balsjö in northern Sweden. The effect of stump harvesting relative to traditional site preparation was investigated by using a similar setup in three catchments in Örebro in southern Sweden.

In Basjö, the first two years after logging (before site preparation) resulted in about 15% increase of the THg concentrations. Site-preparation caused an additional 25% increase in the THg concentration relative to post-harvest conditions. However, no significant response in the MeHg concentrations was observed neither after logging nor site preparation relative to pre-logging conditions.

Although the concentrations of THg and MeHg at the Örebro site were exceptionally high compared to other sites in Scandinavia and North America, the results indicate that stump harvesting has not caused increased concentrations of THg or MeHg relative to traditional site preparation or the untreated control. The very high concentrations in this area seem to be controlled by factors such as organic carbon, hydrology, temperature, reduced microenvironments and the initial logging rather than by the stump harvest or the site preparation.

In neither Balsjö nor Örebro did the studied forestry operations result in effects on the Hg export as dramatic as reported in some earlier studies. This could be either due to good forestry practice and/or a result of varying sensitivity to forestry among sites. Ongoing research will therefore aim at investigating factors controlling the relative sensitivity of a catchment to forestry operations from a mercury perspective.

In accordance with current Finnish energy policy, tree tops, branches and stumps produced at harvesting should also be removed and used as biofuel. However, compared to traditional stem-only harvesting (SOH) the impacts of such whole-tree harvesting (WTH) on biogeochemical cycling and leaching, are largely unknown, especially concerning heavy metals and drained peatland forests. In this study we compare the effects of SOH and WTH of drained peatland forests on the concentrations of heavy metals, especially mercury (Hg) and methyl mercury (MeHg), in drainage waters using a "paired catchment" approach. We hypothesized that the greater degree of disturbance and increase in the amount of standing water associated with WTH would result in increased Hg and MeHg concentrations in drainage waters. Since the release and transport of heavy metals from soils is often associated with organic matter complexes and particulates we also hypothesized that Hg and MeHg concentrations would be correlated with dissolved organic carbon (DOC) and suspended solids. Eight small (0.5–3 ha) drained peatland catchments, located in eastern Finland, were selected for study. The forests were dominated by Scots pine (Pinus sylvestris) and Norway spruce (Picea abies). In 2007, study sites were established. In 2009, two catchments were harvested using SOH and four catchments harvested using WTH; the remaining two catchments were left unharvested (controls). Sampling has been carried out between April 2008 and October 2010 during the snow free period. Samples for Hg and MeHg analyses were collected into FLPE bottles, and analyzed using CVAFS. In addition concentrations of heavy metals, nutrients, DOC and suspended solids were determined. During the calibration period (2008), catchment mean Hg concentrations varied between 3.6 and 8.1 ng L^-1 and MeHg concentrations varied between 0.18 and 3.19 ng L^-1. After harvesting, catchment mean Hg concentrations increased in both the WTH and SOH treated catchments, WTH catchments on average by 26% and SOH catchments by 5 %. MeHg concentrations in the WTH catchments increased (32%) after harvesting, but decreased (20%) in the SOH catchments. Concentrations of Hg correlated significantly with DOC and MeHg concentrations with suspended solid concentrations. Both DOC and suspended solid concentrations increased after harvesting. MeHg concentrations showed a clear seasonal pattern, with concentrations the highest in the end of summer. Hg concentrations did not show a seasonal pattern.

Studies have shown that watershed soil Hg concentrations influence fish mercury concentrations. Watersheds with O and A-horizon soils that have higher mercury concentrations tend to have fish with higher mercury concentrations. A number of factors influence the amount of Hg in soils including deposition and management factors. Logging has been shown to influence mercury in soils, runoff and the aquatic food chain. Recent studies indicate that emissions from fire can influence both short-term soil storage and long-term accumulation. Fire has also been implicated in increasing fish Hg concentrations. Although rarely tested, blowdown has the potential to accumulate soil mercury, especially in the forest floor.

In this study we took advantage of a combination of blowdown, fire and salvage logging treatments in the sub-boreal region of northern Minnesota. In 1999 severe winds damaged 200,000ha of the Superior National Forest. Over the next several years the US Forest Service contracted salvage logging in the windthrown area to reduce fire risk. In May 2007 a fire burned 15,000ha through this same landscape. The patchiness of these disturbances created various combinations of blowdown, salvage logging, and wildfire: Blowdown-Salvage-Fire (BSF), Blowdown-Fire (BF), Fire (F), Blowdown (B), and Control (C). Control areas were unimpacted mature forests. In 2009, we randomly chose six study sites within each of five treatments and established plots along 40-m grid intervals. Within each plot we collected forest floor and mineral soils to 10 cm. Forest floor and soil samples were analyzed for Hg, total carbon and total nitrogen.

Although no treatment effects were found for mineral soils, we did see significant effects on both concentrations and mass/area for forest floor. Forest floor concentrations and mass/area for the BSF treatment was significantly lower than all other treatments. Also the F treatment was less than the B treatment for both concentrations and mass/area. In addition, the BF and C treatments had lower mass/area than the B treatment. Fairly strong relationships were found between Hg and total carbon and nitrogen in both forest floor and mineral soils. Cumulatively our data indicate that fire and salvage logging lead to decreases in forest floor Hg and blowdown leads to greater forest floor Hg.

Forest harvest may stimulate MeHg leaching from catchments with consequences for MeHg levels in aquatic food webs. We designed a paired-catchment study in two small boreal headwater catchments with ca 20% peat in southern Norway to test effects of harvest on Hg and MeHg in streamwater and invertebrates. The pre-treatment period was June 2008 - January 2009. Forest in one catchment was harvested in January 2009, according to national forestry practices. The soil was not frozen and the machines made deep and visible tracks in the soil. Water chemistry, discharge and invertebrates were monitored in both catchments before and after harvest.

Average streamwater chemistry in the reference and the treated catchment in the preharvest period were 0.09 and 0.19 ng/L MeHg, 14 and 22 mg/L TOC and pH 4.8 and 4.5, respectively. Effects on water chemistry after harvest so far are increased concentrations of K, totP and inorganic N species. No increase in MeHg has been observed until now. The harvested catchment shows seasonal peaks of MeHg (ca 0.5 ng/L) during winter (starting prior to harvest), dropping to <0.1 ng/L during snowmelt. Such seasonality is lacking in the reference catchment. We do not understand the driver of the winter peaks of MeHg but speculate that the snowpack may simulate hydrological connection between the peatland and the stream in the harvested catchment.

Both streams had similar, short food chains. We sampled biota in the autumn of 2008 and the spring of 2009. Biofilm MeHg was unexpectedly low in both streams. MeHg in primary consumers – stoneflies and blackflies - was highest in the MeHg-rich stream, while trophic enrichment of MeHg did not differ. Thus, exposure to MeHg at the bottom of the aquatic food chain appeared to control differences in MeHg in predator invertebrates (caddisflies). MeHg in winter-biota in both streams was significantly related to retention of bacterial fatty acids, suggesting a link between dietary uptake of bacteria and MeHg accumulation at the base of the food chain. In the summer-biota, MeHg was negatively related to retention of algal fatty acids, suggesting that ingestion of algae diluted MeHg bioaccumulation.

A better understanding of catchment processes driving MeHg in streamwaters is needed to allow prediction of MeHg leaching, exposure of food webs to Hg, and ultimately, of Hg levels in fish.

Data mining in data bases of Québec Ministry of Environment, Hydro-Québec, COMERN strategic network as well as new data collected under Environment Canada’s CARA research program allowed us to reconstruct the historical variations of mercury (Hg) levels in two fish species most sought after and consumed in Québec (Canada): northern pike (Esox lucius) and walleye (Sander vitreus). In order to compare different years and lakes among each others, we calculated fish flesh Hg levels at standardized lengths using von Bertalanffy growth models, 675 mm and 375 mm for northern pike and walleye respectively. We present results covering the period of 1979 to 2010 for a series of 10 large lakes of mid-northern Québec frequently fished by sport fishers. Oddly enough, fish caught in two of the three lakes where intense mining activities have been occurring over the past decades, Chibougamau and Matagami lakes, presented the lowest Hg levels at standardized length for both species. Fish caught in mine impacted lakes exhibited marked decrease in Hg concentrations since the 1980’s. For the remaining seven lakes, we followed the intensity of logging activities and wild fires thanks to satellite images covering the past three decades. In all five lakes with watersheds most significantly impacted by logging activities since 1979, i.e. with a rate of logging equivalent to at least 20% of the watershed surface every 10 years, we observed much higher Hg levels in northern pikes and walleye at standardized length. In all these lakes, Hg levels reached peak values in the 1980’s and significantly decreased since then. We could not relate recent decreased Hg levels in fish to remediation techniques in mining activities nor to new logging practices replacing clear cuttings as fish caught in unperturbed lakes also exhibited decreased Hg levels since the 1980’s. We are proposing that Hg levels in fish in large lakes of mid-northern Québec have been decreasing for a few years because of a conjunction of factors such as decreased atmospheric Hg loadings, decreased acid rain deposition and average temperatures rises.

In the Amazonia, as in other forest ecosystems, mercury is present as a result of natural mineral deposits, atmospheric deposition, and regional anthropogenic sources of pollution, including gold mining. Intact forests and soils within these systems serve as a sink for mercury, diminishing or delaying its movement in the environment and in the human food chain. In the Amazonian region deforestation rate was around 1.7 x 10⁶ ha yr^-1 between 2000 and 2010 and biomass burning is generally used as the first step of deforestation. How this high rate of deforestation and biomass burning can impact mercury cycle was the aim of several studies conducted in different places of the Amazonian region (Negro river basin and different areas of the states of Amapá, Mato Grosso and Acre) by our laboratory over a period of 10 years. Mercury concentrations were determined in many compartments (air, rainwater, throughfall, river water, sediment, fresh vegetation, litterfall, soil, fish and human) allowing to calculate and to compare Hg fluxes at the atmosphere/vegetation, vegetation/soil, soil/atmosphere (deposition and emission) and soil/water interface in forests and deforested areas. The factors involved in the mercury immobilization/mobilization processes observed in forests and deforested areas were also studied. This paper aims to present a synthesis of these studies.