THE ROLE OF THE AMAZON RAINFOREST IN REMOVING ATMOSPHERIC MERCURY
Atmospheric mercury is effectively trapped in forest canopies by deposition on leaves surface. Mercury compounds that accumulate on foliage are then transferred to the soil through litterfall and its decomposition, and via throughfall. The Amazon rainforest covers approximately 5.5 million km2, representing over half of the planet's remaining rainforests. Nevertheless, the role of the Amazon in the balance between Hg emissions and depositions is still a field of research with high uncertainties evolved. The aim of this work was to quantify the amount of Hg removed from the atmosphere and stocked in this forest every year. The impact of deforestation on the deposition/emission mass balance is also discussed.
Total Hg deposition was calculated as the sum of Hg deposited by litterfall (HgLitterfall) and by throughfall (HgThroughfall). Calculations were performed based on a large set of data obtained from the literature. The removing rate of atmospheric Hg would be of 121 g km-2 y-1, with 49 g km-2 y-1 due to HgLitterfall. Dry deposition was also estimated to 103 g km-2 y-1. Combining this data with the Amazon rainforest area (5,475,737 km2), results in a total rate of mercury immobilization of ~662 ton y-1. In a global scale, the total atmospheric Hg deposition to land has been estimated to 3,200 ton y-1. The Amazonian rainforest would therefore be responsible for the immobilization of, at least, ~21% of the total atmospheric Hg deposited to land. This high efficiency to remove atmospheric Hg can be related to many parameters, such as: large forest area, high wet deposition, high leaf area index and high leaf lifetime (both increasing Hg concentration in foliage/litter) and high litterfall rate.
On the other hand, the Amazonian forest suffers constant deforestation, which is mainly conducted by forest burning. This process results in Hg volatilization stocked in vegetation and superficial soil; we also showed that it also significantly enhances soil Hg emissions in deforested area. Considering the annual rate of deforestation of the Brazilian Amazonian forest (11,075 ± 7,177 km2 y-1 for the 2004-2014 period) we estimated that ~6 ton Hg can be reemitted to the atmosphere each year due to deforestation. One aspect that has never been included in this mass balance is that deforestation also results in a loss of the forest capacity to remove atmospheric Hg. When considering the deforestation area for the 10 last years and the dry deposition rate, we estimated that the removing loss was ~97 ton Hg.
EXCHANGE PROCESSES OF GASEOUS MERCURY WITH THE CLOSED ECOLOGICAL SYSTEM OF THE MASOALA RAINFOREST, ZOO ZURICH
The nature and relative importance of exchange processes of elemental mercury (GEM)) between the terrestrial ecosystem and the atmosphere is still subject to debate in the research comunity. Even on the question, whether the global terrestrial ecosystem serves as a net sink or source to the atmosphere, one has not yet agreed upon. Thereby the uncertainty regarding the leaf-atmosphere exchange of GEM plays a pivotal point. Two measurement techniques have primarily been used to study GEM fluxes: dynamic flux chambers and micro-meteorological measurements. While both methods have their benefits, their shortcomings introduce substantial uncertainties. A novel approach has been used in this study to overcome some of the deficiencies. As a study site, we used the closed ecological system of the Masoala Rainforest, Zoo Zurich, a project within the framework of the zoo's conservation strategy. On an area of 10’856 m² a dense plantation reproduces a piece of Madagascan rainforest. A hall constructed from transparent ethylene tetrafluoroethylene foil provides a tropical climate, which is regulated by a ventilation system. This setup, resembling a scaled-up flux chamber, presents a unique opportunity to study the exchange processes of a terrestrial ecosystem with the atmosphere. Gaseous elemental mercury was therefore measured in the ventilation system of the Masoala Rainforest. Depending on the ventilation regime different situations could be studied. Generally, we observed GEM concentrations inside the Masoala hall to be lower by about 30% compared to outdoor air. During periods where the air was merely recirculating inside the hall (i.e. with no influx of fresh outdoor air), GEM levels indoors steadily declined, indicating that the Masoala hall is a net sink for GEM. The analysis of our time series allowed the derivation of an uptake rate of mercury within this rainforest ecosystem. Ongoing work includes the analysis of plant material, soil, and water from the Masoala Rainforest. We believe our work can provide a new approach to assess mercury fluxes between terrestrial ecosystems and the overlying atmosphere. With our study, we contribute to the existing knowledge gap regarding mercury exchange fluxes and help to improve the global mercury mass balance.
MERCURY EMISSION TO THE ATMOSPHERE DOMINATES ANNUAL MASS BALANCE OF A BOREAL PEATLAND: TIME TO RETHINK TIMELINES FOR RECOVERY?
To estimate the potential of different ecosystems as sinks or sources for atmospheric Hg, reliable quantification of land-atmosphere exchange of gaseous elemental Hg (GEM) is crucial. We have made the first annual Hg budget based on continuous measurements of peat-atmosphere exchange of GEM using a novel relaxed eddy accumulation (REA) system. The annual Hg mass balance was dominated by net GEM emission (10.2 µg m-2) due to substantial evasion between May and October. The annual wet bulk deposition of Hg was 3.9 µg m-2. The annual discharge export of Hg from the peatland area (1.9 km2) amounted to 1.3 µg m-2. The GEM evasion rate, a factor of 2.6 higher than wet bulk deposition, can be explained by the recent reduction in the atmospheric Hg concentration to a value below the compensation point for this peatland, turning it from a sink into a source of Hg emission back to the atmosphere after decades of Hg accumulation. This is consistent with the Hg concentration gradients in the superficial peat which decline from a Hg concentration peak at about 30 cm depth (110 ng g-1, corresponding to Hg emission peaks during the 1950s) towards the surface (24 ng g-1). Under the assumptions that environmental conditions remain stable and that catchment runoff is dominated by Hg from the uppermost peat layers, it will take around 80 years to deplete the entire pool of legacy Hg in the uppermost 34 cm to a background concentration level of 20 ng THg g-1. We suggest that the strong Hg evasion demonstrated in this study means that open boreal peatlands and thus downstream ecosystems may recover more rapidly from past atmospheric Hg deposition than previously assumed.
Given the current international efforts to protect human health and the environment from the adverse effects of mercury in accordance with the United Nation’s 2013 Minamata Convention on Mercury, we believe our findings contribute to a better understanding of how emission reductions will influence mercury cycling in northern peatlands and their role for the mercury status of fresh water fish in the Northern Hemisphere. Our findings also raise the question as to whether recent reductions of atmospheric mercury concentrations have led to net evasion from other ecosystems as well, such as forests and oceans. We believe that the REA technique could be applied more widely to define the balance between new emissions and re-emission for different ecosystems.
EFFECT OF PH AND DOC ON PARTITIONING OF MERCURY ON MINERAL FORESTED SOIL SURFACE
Acid deposition has acidified northeastern forest ecosystems. Regulation of emissions and subsequent decreases in acid deposition are reversing acidification. However, recent studies indicate that concentrations of dissolved organic carbon (DOC) have increased, apparently in response to decreases in acid deposition and/or changing climate. This study was conducted to evaluate how changes in acidity and concentrations of DOC can affect soil adsorption/desorption of mercury (Hg). We initiated a series of adsorption experiments using soil collected from mineral horizon at Honnedaga Lake-watershed a forested site in the Adirondack region, NY. The DOC stock solution for this experiment was obtained by soaking freeze-dried Oa horizon in deionized water for 3 days and the suspension was filtered and stored at 4C. We equilibrated 50 ml of either 0, 3.5, 8.1, 25.2, 41.9 and 56.4 mg/l solutions of DOC with 1 g of sieved, freeze-dry mineral soil on a shaker for 24 hr. Either nitric acid or sodium hydroxide was added to the initial experimental solutions to adjust pH. Particulate matter was removed from the final solutions by centrifugation and filtration. The supernatant was measured for Hg and DOC concentration and pH. The results of the adsorption experiment demonstrated that changes in solution pH have a large effect on Hg and DOC release from forest mineral soil. Note the mass of total Hg in soil was much greater than the mass of initial Hg in solution (approximately 230 ng on soil vs 0 to 1.3 ng in solution in each trial tube), therefore we were not able to accurately evaluate the adsorption behavior of Hg, rather desorption behavior of soil was examined. In the pH range 4-4.5, there was adsorption of Hg and minimal desorption of DOC, but as the pH increased above 4.5 there was a gradual increase in desorption in both DOC and Hg. Hg in solution reached a maximum percentage of soil Hg of approximately 2% at pH 5.9. A strong correlation was observed between desorbed DOC and Hg over the pH range of our experiment (R2 = 0.91) which suggests that functional groups of dissolved organic matter strongly complex Hg. Consequently Hg desorption is driven by the release of dissolved organic matter from soil. Results of this experiment can be applied in characterization of a complexation model of Hg, organic compound, and soil surface.
QUANTIFYING MERCURY IN LEAVES, BARK AND WOOD OF EIGHT TREE SPECIES ACROSS FOUR NORTHEASTERN FORESTS USING APPROPRIATE METHODS FOR SAMPLE PREPARATION AND ANALYSIS
Mercury deposition affects remote areas such as forests, but the amount of Hg in trees is not well known, in part because concentrations of Hg in wood are below detection limits of some methods. For example, ICP-OES requires a liquid sample, which in the case of wood, might be 1/100th of the concentration of the tissue sample prior to digestion and dilution. Solid samples can be directly analyzed by thermal decomposition, catalytic conversion, amalgamation, and atomic absorption spectrophotometry through a Total Mercury Analyzer, giving much lower detection limits, but questions remain about sample preparation, such as whether air-dried samples would be suitable for analysis. We examined the effects of drying temperature during sample preparation using wood samples at the Hubbard Brook Experimental Forest, New Hampshire, USA. Samples that were freeze-dried or oven-dried at 65 ˚C were suitable for analysis of Hg, whereas oven-drying at 103 ˚C resulted in Hg losses, and air-drying resulted in Hg gains, presumably due to sorption from the indoor atmosphere. Having established suitable methods, we analyzed Hg in wood, bark, and foliage of eight tree species across four sites (Huntington Forest, NY; Sleepers River, VT; Hubbard Brook, NH; Bear Brook, ME) in the northeastern USA to determine the importance of Hg in trees. Foliar concentrations of Hg averaged 16.3 ng g-1 among the hardwood species, namely American beech (Fagus grandifolia Ehrh.), white ash (Fraxinus americana L.), yellow birch (Betula alleghaniensis Britt.), sugar maple (Acer saccharum Marshall.), and red maple (Acer rubrum L.). Foliage of conifers, namely red spruce (Picea rubens Sarg.), balsam fir (Abies balsamea (L.) Mill.) and white pine (Pinus strobus L.) averaged 28.6 ng Hg g-1), significantly higher than the hardwoods (p < 0.001). Similarly, bark concentrations of Hg were lower (p < 0.001) in hardwoods (7.7 ng g-1) than conifers (22.5 ng g-1). Species also differed significantly in Hg concentration of foliage (p = 0.02) and bark (p < 0.001). For wood, concentrations of Hg were highest in yellow birch (2.5 ng g-1) compared with all the other species (mean of 1.4 ng g-1) (p < 0.0001). Sites differed significantly in Hg concentrations of foliage and bark (p = 0.02) but not wood (p = 0.60). The Hg content of trees, estimated from modeled biomass and measured concentrations at each site, was higher in wood than foliage. Wood is important to Hg budgets in spite of low concentrations, because of its large mass.
LANDSCAPE INFLUENCES ON MERCURY CYCLING AND BIOAVAILABILITY IN VERNAL POOLS
Vernal pools are temporary seasonal water bodies found in Northeastern forests. They are subject to mercury (Hg) contamination via throughfall, leaf litter and snowmelt, and to conditions of high organic matter and low pH which support methylation of Hg to the more toxic and bioavailable species, methylmercury (MeHg). Currently, few studies have reported Hg and MeHg in vernal pools, but levels of Hg in the forest floor and streams have been shown to vary widely with landscape characteristics such as forest type, canopy cover, and land-use over small spatial scales. A suite of 21 pools were examined for Hg and MeHg levels and bioavailability across different canopy covers, with additional temporal sampling at six of the pools. Water column dissolved MeHg concentrations ranged between 0.02 to 2.76 ng/L, and from 5 - 52% of total Hg. Concentrations of total Hg in pools were correlated to dissolved organic carbon (DOC) across sites, whereas MeHg had a temporal relationship with water temperature between snow melt and drying. While coniferous sites had lower pH, canopy type alone did not predict MeHg concentrations. Wood frog and salamander embryos from pools surrounded by deciduous forest had higher concentrations of MeHg than those from coniferous sites, suggesting forest type affects bioavailability.
ELEVATIONAL AND SEASONAL PATTERNS IN METHYLMERCURY INPUTS AND PRODUCTION FORESTS ACROSS A MONTANE ELEVATION GRADIENT
Mercury contamination in remote regions is generally derived from atmospheric deposition. While many studies have examined patterns in mercury deposition and processing in aquatic ecosystems, less is known about the fate of mercury in terrestrial systems and particularly in montane environments. We examined soil samples collected across an elevational gradient on Whiteface Mountain in the Adirondack Region of New York to determine spatial patterns of methylmercury concentrations across the forested montane landscape. We also investigated sources of total and methylmercury inputs (throughfall, precipitation, and litterfall), as well as mercury uptake into avian species. We found that soil methylmercury concentrations were highest in the mid-elevation coniferous zone (0.39 ± 0.58 ng/g), compared to the alpine (0.28 ± 0.36 ng/g) and deciduous zones (0.17 ± 0.19 ng/g), while the percent mercury as methylmercury in soils decreased linearly with elevation. Additionally, methylmercury concentrations and percent mercury as methylmercury were highest in the Oa layer (0.30 ± 0.30 ng/g, 0.12 ± 0.16%) compared to the Oe/Oi layer (0.18 ± 0.27 ng/g, 0.07 ± 0.10%). Finally, we found litterfall mercury inputs to exhibit relationships with soil methylmercury concentrations and percent mercury as methylmercury, with the highest concentrations found in the coniferous zone (0.052 ± 0.017 ng/g, 0.10 ± 0.018%) compared to the alpine (0.039 ± 0.014 ng/g, 0.061 ± 0.023%) and deciduous zones (0.031 ± 0.020 ng/g, 0.10 ± 0.060%). Soil methylmercury concentrations and fluxes vary seasonally and appear to be driven by internal processing of ionic Hg, as opposed to atmospheric deposition of methylmercury to the forest floor. These findings are consistent with methylmercury patterns in terrestrial bird species and suggest that future declines in mercury emissions could be important in reducing litterfall mercury inputs to terrestrial systems and thus concentrations of mercury in montane avian species.
MERCURY CYCLING AND ISOTOPIC FRACTIONATION IN FOREST ECOSYSTEMS: A MODELING STUDY
Mercury (Hg) is subject to long-range atmospheric transport. Forest ecosystems cover >30% of the Earths land surface and are considered as an important Hg reservoir on a global scale. Recent assessments on Hg removal through vegetative Hg uptake followed by litterfall and accumulation in soil, in addition to wet deposition, suggest that global forests are a major sink of atmospheric mercury. However, the processes driving Hg cycling in forest ecosystems have not been fully understood. Forest characteristics, meteorological conditions, and surface terrain are important drivers of Hg deposition in forested area. In the past few years, data of stable Hg isotope measurement for the air, soil, biomass, and litter samples at multiple forest sites offer new insights to Hg cycling processes in forest ecosystem. In this study, we build a hybrid mass balance and isotopic fractionation model for simulating Hg transformation, translocation and isotopic composition changes in forest ecosystems. The model parameterizes mercury exchanges between forest and atmosphere, the transformation processes of deposited Hg through litterfall on forest floor, and the isotopic massdependent and independent fractionations involved in these processes. The model results are verified with the measurements of Hg concentrations and isotopic compositions of air, biomass, litter (fresh and degraded) and soil samples, as well as air-surface (air-soil and air-foliage) exchanges of elemental Hg vapor at a subtropical evergreen forest site located in Mt. Ailao Nature Reserve in Yunnan Province, Southwest China. Based on the model results, the mass budget, cycling and the forcing of Hg isotopic compositions of at the experimental forest site will be elucidated. Implications of the modeling assessment will be discussed in terms of the role of forest ecosystems in global Hg budget and its impact on the isotopic composition of atmospheric Hg.