2j: Mitigating Mercury Bioaccumulation in Contaminated Aquatic Ecosystems

The main Polish sources of mercury atmospheric emissions are combustion of fossil fuels, mainly coal, cement production and industrial production processes, in particular the mercury cell chlor-alkali processes for production of chlorine and caustic soda, smelting of ferrous and non-ferrous metals, especially zinc and copper as well as consumption-related discharges. When the critical limit for Hg in case of Hg contaminated soils is taken into consideration (30 ppm), the areas inside the chlor-alkali plants (CAP) are characterized by Hg concentration range with the median value from 40 ppm up to 557 ppm. Gaseous emission of Hg from CAPs was considered to be one of the most important point sources of anthropogenic mercury in Poland.

Not only Europe needs safe and cost-effective solutions for the permanent storage of elemental mercury inside chlor-alkali plants. The solution is expected to be environmentally sound and economically viable. Phytoremediation, an approach that is possible to use plants physiological properties for soil pollutant degradation, stabilization or extraction is taken into consideration also in relation to mercury problem.

Presented investigations are related to the changes in mercury cycling after soil amendment application. We hypothesized, that soil appropriate additives have positive impact on concentrations of bioavailable mercury in soil solutions, bioaccumulation and volatilization. Biogeochemical processes were analyses based on lab scale treatability study (lysimeter experiment) and plot scale experiment located in natural conditions on the selected area inside one of the CAP. Experiments were planned to test phytostabilization and aided phytostabilization of mercury contaminated soil.

Soil and soil solution samples were analyzed for their pH, EC, content of bioavailable fractions (soil water-soluble and exchangeable mercury fractions). Selected plant species (dicots and monocots) and soil additives (sulphur, zeolites) were tested for their phytostabilization effectiveness.

Additives were also analyzed as the potential soil factors stimulating local biodiversity and natural succession from outskirts modulate dominant plant species, rhyzosphere microorganisms, plant cover, plant abundance and Hg bioaccumulation on treated plots. Changes in mercury evaporation (Lumex) and soil respiration (LCpro+ system) related to the plant species were found.

The studies confirmed the possibility of using plants as efficient remediation tools in phytostabilization of mercury-contaminated soil. Mercury volatilization depends strictly on extend of plants coverage and species composition. Soil additives (granular sulfur and zeolites) as soil stabilizers reduced Hg bioaccumulation, volatilization and evaporation.

Wetlands are critical for wildlife and healthy ecosystem function, and yet are also important areas for methylmercury (MeHg) production, bioaccumulation, and export to adjacent water bodies. Therefore, preserving or restoring wetland habitat, while limiting MeHg export and risk to local food webs, presents a unique challenge to wetland managers. We examined how a combination of wetland design and water management influenced MeHg export and bioaccumulation in order to minimize biological MeHg exposure on-site, and transport downstream. The study area consisted of eight managed wetlands (approximately 10 hectares) that were flooded annually (4-20 cm depth) during mid-September thru late April to provide habitat for winter migratory waterfowl. Four of the wetlands were treatment wetlands, and were constructed with two distinct cells separated with a narrow berm. Eighty percent of the wetland was maintained shallow for water-bird habitat, and the remaining 20% of the area (~2 hectare) at the outflow end was excavated to a depth of ~1 m. The four control wetlands were shallow and relatively uniform in depth. Treatment wetlands were operated in a semi-continuous flow-through mode, while control wetlands were operated in a fill-and-hold mode, where water was only released at the end of the managed flood period. We quantified three MeHg removal mechanisms (particulate settling, benthic demethylation, and photo-demethylation) in the deep-water treatment cells. Over two years, mean whole-water MeHg load (by mass) decreased ~40% from the in-flow to the outflow in the deep cells on average. Photodegradation accounted for ~7% of the mass (18% of loss), while particle flux to the benthos accounted for ~24% (60% of loss) of the MeHg budget. Benthic MeHg degradation was exceeded by MeHg production indicating an unknown loss mechanism accounting for the remaining 9% of the mass (22% of loss) in the MeHg budget. While deep-cells within the treatment wetlands served as net sinks for MeHg, overall the net MeHg export from the flow-through treatment wetlands exceeded export from the fill-and-hold control wetlands. Despite the effectiveness of the deep cell in lowering MeHg export, Hg concentrations in biosentinel fish (Gambusia affinis) showed no change from inlet to outlet. Although the use of deep cells in wetland design were effective in lowering MeHg exports under flow-through conditions, further optimization of wetland management is needed prior to widespread implementation to decrease biotic exposure.

In situ Activated Carbon (AC) amendments offer an attractive, potentially low impact approach for reducing contaminant bioavailability in ecologically sensitive environments. Over the last few years, we have carried out several field and laboratory trials to evaluate the efficacy of activated carbon (AC) as a remediation tool for mercury (Hg) contaminated marsh soils. Our focus has been on field trials of thin-layer placement of AC on marsh soils, and on laboratory studies to evaluate the geochemical factors that influence the impact of AC on Hg bioavailability, net MeHg production and MeHg bioavailability. Our studies show that AC can successfully reduce inorganic mercury (Hg) and methylmercury (MeHg) concentrations in porewater, and bioavailability to benthic organisms. However, the effectiveness of AC in reducing Hg and MeHg bioavailability varies among sites and soils. Additionally, AC amendments can change the balance of MeHg production and degradation.

Field trials of thin-layer AC placement have been carried out at the plot scale in three Hg-contaminated tidal locations: a salt marsh in Maine, a Phragmites marsh in, NJ, and a tidal creek bottom in the Chesapeake. Study designs varied, but in all cases, AC amendment plots were compared to untreated control plots, and sediment and pore water biogeochemistry was followed in detail over time. The Maine and NJ plot studies ran for 2 years. We will discuss the efficacy of AC in each of the field studies, in relation to site geochemistry, and over time. We will also present laboratory studies on the impact of sediment chemistry and microbial activity on AC efficacy, and the impact of AC amendments on net MeHg production. Our goal is to develop an empirical model that will allow end-users to evaluate the potential efficacy of AC as a Hg-remediation tool for specific sites. Specifically, we will discuss how soil sulfide and organic matter content impact AC efficacy; the impact of pore water dissolved organic matter on Hg and MeHg partitioning to AC; and the impact of AC on microbial activity and net MeHg production.

In Situ amendments of Activated Carbon (AC) have been used as a low impact method to remediate contaminants in ecologically sensitive environments. Laboratory experiments and field trials have shown that AC is successful in reducing inorganic mercury (Hg) and methylmercury (MeHg) concentrations in porewater. Yet, the effectiveness of AC varied in different soils, with AC reducing MeHg porewater concentrations between 30% and 90% depending on the soil. These data indicate that site geochemistry might dictate the efficacy of AC use for Hg and MeHg remediation. In the environment, Hg partitioning is effectively a competition between solid phase environmental ligands and dissolved phase ligands. Thus, partitioning can be profoundly affected by high concentrations of ligands, such as dissolved organic matter (DOM) or sulfide. In this study, we evaluated the impact of DOM concentration on MeHg and Hg partitioning to the solid phase in Hg-contaminated sediment/AC mixtures

The experiment was conducted over 21 days in anaerobic sediment/AC slurry microcosms that were amended 5% dry weight AC and Suwannee River Humic Acid at a range of concentrations. Marsh soils were collected from a Phragmites marsh in Berry’s Creek, NJ. The microcosms were destructively sampled at 5 time points to monitor Hg and MeHg soil:water partitioning (K_D).

Results showed that the addition of SRHA did not significantly impact the partitioning of either ambient MeHg or a fresh Me¹⁹⁹Hg spike onto AC. However, DOM did reduce inorganic Hg partitioning (both ambient Hg and fresh ²⁰¹Hg spike) to AC in soils, in a concentration-dependent manner. DOM is known to slow the formation of HgS nanoparticles, and we hypothesize that DOM impacts the efficacy of AC for inorganic Hg removal in porewater by holding Hg-sulfide species in solution. AC efficacy in MeHg sorption was not impacted by DOM up to 60 mg/L, suggesting that AC may be valuable in reducing MeHg risk even in high DOM soils, like marshes. This study highlights the need to evaluate AC efficacy for remediation in the matrix being considered for treatment and the need to develop models of AC efficacy based on site biogeochemistry.

The Guadalupe River Watershed in Santa Clara County, California, is contaminated with mercury waste from the former New Almaden Mining District: North America’s oldest and most productive mercury mine, and the fifth largest in the world. Though active mining halted by 1970, waste rock and contaminated sediment persist as sources of mercury to the watershed. The Santa Clara Valley Water District (District) manages four water bodies affected by historical mining operations: Almaden, Calero, and Guadalupe reservoirs, and Almaden Lake. These eutrophic water bodies stratify seasonally, creating anoxic conditions that facilitate the bacterial conversion of mercury to bioavailable methylmercury. In 2005, Tetra Tech calculated 5.8 mg Hg/ kg for standardized 40 cm largemouth bass in Guadalupe Reservoir. Signage throughout the watershed warns anglers to not consume fish. The Guadalupe River Watershed Mercury TMDL was adopted in 2008.

Beginning in 2006, the District installed solar-powered circulators in Almaden Lake and hypolimnetic line-diffuser oxygenation systems in the reservoirs to suppress anoxia and methylmercury production. After several years of intermittent operation, continuous operation began in 2016. The effectiveness of the control systems is analyzed using water quality profiles as well as mercury and nutrient samples collected monthly to twice- monthly. Biannually, young largemouth bass and prey fish are sampled and analyzed for mercury body burden.

Results thus far have indicated that Almaden Lake’s solar circulators, while failing to improve dissolved oxygen and redox potential, have reduced methylmercury and nutrient concentrations in the water column. The hypolimnetic oxygenation systems in Almaden and Guadalupe reservoirs have improved dissolved oxygen concentrations, leading to decreased methylmercury production and reduced nutrient efflux from bottom sediments. In 2016, seasonal peak hypolimnetic total methylmercury concentrations were reduced compared to pre-oxygenation in Almaden Reservoir from 10.5 ng/L to 1.4 ng/L; in Guadalupe Reservoir from 60 ng/L to 1.3 ng/L; and previously in Calero Reservoir from 13.2 ng/L to 0.46 ng/L (Calero did not efficiently retain added oxygen in 2016). However, epilimnetic methylmercury concentrations increased or were unchanged. Though fish data collected during continuous operation of the oxygenation systems is limited, results have not indicated a decline in fish tissue mercury concentrations. Calero Reservoir, the least contaminated of the reservoirs downstream of New Almaden, exhibited lower fish mercury concentrations than Stevens Creek Reservoir, which is located outside of the watershed and used as a control site. This phenomenon may be due to Calero’s larger size, higher algal productivity, and more diverse fish assemblage.

The anaerobic sediments of stratified lakes are particularly active zones for methylation of inorganic mercury (Hg²⁺) and can be an important source of methylmercury (MeHg) to the water column during summer anoxia and fall turnover. Production of MeHg is promoted by anaerobic conditions and the supply of Hg²⁺, sulfate (SO₄²⁻), and labile organic carbon. Hg-contaminated Onondaga Lake, NY is naturally SO₄²⁻ rich, and its eutrophic condition contributed an ample supply of organic carbon and low redox potentials. These conditions resulted in accumulations of MeHg in the hypolimnion during the summer stratification interval, followed by entrainment of MeHg into the epilimnion with the approach to fall turnover. Elevated MeHg concentrations during this period represented a potentially important exposure pathway for aquatic organisms.

During 2011–2013, a whole-lake nitrate (NO₃^–) addition pilot test was conducted with the objective of limiting release of MeHg from the pelagic sediments to the hypolimnion through maintenance of NO₃^––N concentrations >1 mgN/L. A liquid calcium-nitrate (Ca(NO₃)₂) solution was added to the hypolimnion from a barge as a neutrally buoyant plume approximately two to three times per week during the summer stratification interval. Water from the surface of the lake was used to dilute the Ca(NO₃)₂ solution and achieve neutral buoyancy at 1-2 meters above the sediments. Following a successful pilot study, the program was adopted as part of the long-term remediation program for the lake. The annual dosing of NO₃^– ranged from 5.6×10⁴ to 8.8×10⁴ kg during the first six years of the program (2011–2016).

Maximum concentrations of MeHg in the hypolimnion remained <0.5 ng/L during the six years of the NO₃^– addition program, a decrease of >94% from 2009 levels. Increased sorption of MeHg to Fe and Mn oxyhydroxides in surficial sediments was identified as an important regulating mechanism. The hypolimnetic NO₃^– supply explained 95% of the interannual variations in hypolimnetic accumulations of MeHg over the 1992–2016 interval. Increased MeHg concentrations in the upper waters during fall turnover, which had been a generally recurring pattern, did not occur during 2011–2016. Lower MeHg concentrations in the water column resulted in lower MeHg concentrations in zooplankton, which in turn may result in lower exposure of fish to MeHg.

This presentation discusses a two year assessment of spatial and temporal patterns of methylmercury (MeHg) and ancillary redox-related constituents (dissolved oxygen, nitrate, iron and manganese) in Occoquan Reservoir, a large run-of-the-river drinking water reservoir in Virginia, USA. A tributary to the reservoir receives input of nitrate-rich tertiary treated wastewater that enhances the oxidant capacity of bottom water and increases the safe water yield of the reservoir. Multiple lines of evidence supported the hypothesis that the presence of nitrate and/or oxygen in bottom water correlated with low MeHg in bottom water by enhancing redox in surfacial sediment. Bottom water MeHg was significantly lower in a nitrate-rich tributary (annual mean of 0.05 ng/L in both 2012 and 2013) compared to a nitrate-poor tributary (annual mean of 0.58 ng/L in 2012 and 0.21 ng/L in 2013). The presence of nitrate and oxygen also corresponded with significantly lower bottom water MeHg at an upstream station in the main reservoir. In 2012 the reservoir exhibited a longitudinal gradient with nitrate and oxygen decreasing and MeHg and manganese increasing downstream. In both study years, there was a clear threshold of oxygen equivalent (3-5 mg/L), a metric that combines the oxidant capacity of nitrate and oxygen, above which MeHg was negligible (< 0.05 ng/L) and Mn and Fe were low (< 0.5 mg/L). The presentation concludes with a comparisons of the relative advantages of using nitrate versus oxygen, another oxidant that can potentially be used to manage redox conditions and mercury cycling in the bottom of reservoirs.

Environmental risk from mercury-contaminated sediments derives mainly from methylmercury production, exposure and bioaccumulation. Methylmercury is produced predominantly by heterotrophic sulfate-reducing bacteria. Our research is evaluating the ability of redox-buffering mineral-based amendments to suppress mercury methylation by inhibiting microbial sulfate reduction near the sediment surface, with the objective of reducing methylmercury exposure and food web bioaccumulation. Laboratory sediment-water aquarium microcosms were prepared with homogenized sediment from Nevertouch Marsh in the Berrys Creek Study Area and site water. Manganese(IV) oxide minerals (pyrolusite or birnessite) were either directly mixed into the upper 5 cm of sediment or applied in a thin-layer cap. Overlying water and porewater were monitored. Direct addition of Mn(IV) oxide amendment or inclusion in a thin sand layer resulted in a 66% to 69% reduction in net methylation for pyrolusite and an 81% to 89% reduction for birnessite (measured as %MeHg/THg in porewater at 0 to 5 cm). A thin sand layer alone resulted in 65% reduction in net methylation. CO2 respirometry experiments showed that the amendments stimulated microbial activity. Microbial community census by PCR and DNA sequencing indicated that the addition of Mn(IV) oxides did not significantly alter the indigenous sediment microbial community structure, although a small increase in abundance of iron and manganese reducers was observed after a 2 week incubation period. The mechanism of methylmercury suppression therefore most likely involved a shift from sulfate reduction to manganese reduction as the energetically favorable redox process, which was also confirmed by microelectrode voltammetry profiling of the sediment microcosms. Manganese X-ray Absorption Spectroscopy (XAS) of amended sediment documented the gradual conversion of Mn(IV) oxide amendments to Mn(III) oxides (bixbyite), mixed-valence Mn(II/III) oxides (hausmannite) and Mn(II)-carbonate (rhodochrosite) and aqueous Mn2+ over time. This solid phase assemblage is expected to continue to buffer redox to inhibit sulfate reduction and suppress mercury methylation in the surficial sediment. The retention of the added manganese in sediment solid phases also has interesting implications for in situ self-regeneration of Mn(IV) oxides in dynamic settings where soils and sediments experience periodic water level and redox fluctuations (e.g. in intertidal zones, tidal marshes, seasonal wetlands, reservoirs), as this would prolong the effective lifetime of the amendments.