2g-1: Legacy site assessment and management

The Berrys Creek Study Area (BCSA) Superfund site is a 4.2 km2 tidal tributary to the Hackensack River in Bergen County, New Jersey. The site includes 4.2 km of waterways, surrounded by 3.1 km2 of intertidal marshes dominated by Phragmites australis. Mercury is among the chemicals of potential concern (COPCs) at the site. We designed a phased, conceptual site model (CSM)-oriented investigation that integrated information on the physical, chemical, and biological conditions of the site and the region to assess site-specific uptake.

Multiple lines of evidence collected during the remedial investigation indicate that mercury uptake in the BCSA is mediated via a detrital-based food web. Empirical relationships developed for mercury in shallow sediments and surface water across a range of spatial scales indicate the importance of sediment as the source of mercury for biological uptake. Mercury residues in detritus at the surface of the marshes and in the topmost layer of waterway surface sediments are low relative to deeper sediments, thereby limiting availability of mercury for biouptake; the flat trophic structure of this urban, low-salinity estuary limits biomagnification. In both marsh and waterway sediments, the organic carbon levels, sulfides, and other physical/ chemical characteristics also limit mercury bioavailability.

High resolution, optically-based monitoring and other measurements demonstrate that marsh-derived detritus settles to the waterway sediment bed during slack tides to form a thin, benthic layer of low-density organic material on which mercury and methyl mercury can be sorbed. Resuspension of these materials during high tidal velocities and storm flows is a primary mechanism for mercury and methyl mercury transport to surface water. Empirical relationships developed for mercury in shallow sediments and surface water across a range of spatial scales indicate the importance of sediment as the source of mercury for biological uptake.

The Androscoggin River near Berlin, NH flows past a Superfund site that, prior to 1962, was a chlor-alkali facility. Leakage of elemental Hg from the site directly into the river was reported during site characterization in the late 1990s and had likely been occurring for decades. Despite remediation efforts that concluded in 1999, as recently as 2010 elevated levels of Hg were observed in surface sediments and in the water directly adjacent to the site. The objectives of this study were to: 1) evaluate the extent/concentration of Hg in the downstream floodplain, and 2) evaluate the fate of legacy floodplain Hg after deposition. Evidence for Hg releases during operation of the facility were confirmed by dating an upland (non-floodplain) soil ~4 km downstream using short-lived fallout radionuclides, 7Be and 210Pb. Mercury concentrations in the soil increased significantly during chlor-alkali production in the early 1900’s and peak just before decommissioning in 1962 – most likely from Hg release to the atmosphere from the facility and localized atmospheric deposition. Downstream floodplain sediments contain significantly elevated levels of Hg relative to upland soils and this is most likely due to fluvial input of particle-bound Hg from the site. Mercury concentrations are highest in bankfull floodplain sediments, reaching up to 10 µg/g as far as 5 km downstream of the site and remain elevated relative to background levels up to ~15 km downstream. The concentration of Hg in sediment profiles varies as a function of depth as well as distance from the river channel as shown by 13 floodplain transects. Generally, Hg inventories increase from the channel edge with increasing distance from the channel until the edge of the bankfull floodplain where Hg inventories peak and then slowly decrease with increasing distance from the channel. Sequential extractions indicate that Hg in near channel sediments consists of organically-bound Hg(II) as well as a significant fraction comprised of inorganically-bound Hg. In all transects, the proportion of inorganically-bound mercury increases with depth. Soil pore water sampling also indicates a strong temporal variation in the floodplain; pore water Hg concentrations spike during snowmelt runoff and then decay through the summer. Our results suggest that the potential for contaminated floodplains to be a source of Hg to downstream ecosystems will strongly depend on hydrological and biogeochemical processes occurring in the floodplain. This diversity in potential may also indicate similar diversity in future risk of Hg mobilization due to climate change.

The Flin Flon (Manitoba, Canada) copper-zinc smelter, constructed in the late 1930s, was a major mercury (Hg) and other metals emitter, e.g., an estimated 20 tonnes/yr of Hg were released to the atmosphere in the late 1970s and early 1980s. Studies conducted at that time showed that Hg, copper (Cu), cadmium (Cd), and zinc (Zn) exceeded CCME Probable Effects Levels for sediments in several lakes within 8 km of the smelter and did not approach baseline concentrations until ca. 70 km distant. However, Hg concentrations were exceptionally low (0.04-0.10 µg/g) in 525-mm Northern Pike (Esox lucius) in lakes closest to the smelter with highest concentrations (0.70-0.73 µg/g) in two of four lakes 75-85 km distant; these lakes were slightly acidic (ph 5.9-6.1). Low Hg concentrations in pike close to the smelter were attributed to the inhibition of Hg methylation by metals at toxic concentrations and/or selenium (Se) interference with Hg uptake. Metal emissions subsequently declined with improvements in smelter design and then closure in 2010. Our study, conducted over 2008-2015, is investigating Hg biomagnification in lake food webs, with a focus on Northern Pike. Despite reductions in metal emissions, Cd, Hg, Cu, Zn, and Se concentrations in lake sediments close to the smelter are 1.4-2 higher and in distant lakes 1.1-2.0 times higher than in the early 1980s; immense metal reserves that remain on the landscape continue to be a metal source to these lakes through re-emission and runoff. Hg concentrations in 525-mm Northern Pike in two lakes close to the smelter were higher (0.15-0.49 µg/g) than in 1982 possibly because metals have become less toxic with chemical weathering of metals bound to particulates in the watershed and lake. Percent methyl Hg, a measure of Hg methylation rates, was 2-3% and similar to that observed in lakes (2-6%) in the Thompson smelter area to the east where sediments are not contaminated with metals. Hg concentrations have shown smaller increases (to 0.72-0.92 µg/g) in 525-mm pike in more distant lakes with increases possibly related to warming trends and/or changing water chemistry; pH has increased (mean 7.5) possibly suggesting a reduction in acidfying emissions. Pike condition factor was similar in 1982 (0.68±0.13) and 2008-2016 (0.68±0.08) as was age (5-7 yr) with no differences between metal contaminated and distant lakes suggesting no or little smelter impact on pike growth rates and longevity. Overall, our study shows that despite reductions in smelter mercury emission rates, mercury concentrations in Northern Pike show no evidence of decrease.

Mercury bioavailability and bioaccumulation was investigated in the St. Louis River Estuary (SLRE), a shallow (1-3 m) freshwater estuary at the mouth of the largest tributary to Lake Superior. Fish and other benthic invertebrates in the SLRE contain some of the highest levels of methylmercury (MeHg) in the region and are significantly elevated compared to the upstream river and Lake Superior. The upstream river receives large sulfate loads from iron mining, while the lower estuary contains some sediment with elevated mercury from legacy contamination. Also, the SLRE is one of many great lakes coastal wetlands that experiences significant seiche activity, which dilutes river flow in the lower estuary with low-MeHg water from Lake Superior and increases mixing of riparian backwater embayments with river water. The combination of these factors creates a unique setting in which the importance of surface water and sediment as sources of MeHg to the food web cannot be easily discerned. MeHg concentrations in benthic macroinvertebrates were compared with paired sediment and water samples from estuary sites across a spatial gradient of sediment THg and MeHg concentration. Carbon and nitrogen stable isotopes in predator fish tissue show unambiguously that fish feeding in the upper reaches of the estuary have higher MeHg than those feeding in the lower portions of the estuary or Lake Superior. Smaller Hexagenia had higher %MeHg than large Hexagenia and length-corrections were used to investigate spatial patterns and relations with sediment. Linear regressions comparing MeHg in Hexagenia larvae with MeHg in sediment did not reveal a significant relationship and a classification and regression tree model was used to interpret relationships between Hexagenia MeHg and sediment MeHg, carbon, and sulfur. Surface water MeHg concentrations varied seasonally, consistent with patterns in the upstream river, but showed some spatial patterns, likely related to demethylation, seiche-driven dilution and internal MeHg loading. Preliminary results suggest that legacy sediments and internal loading contribute a relatively small fraction of MeHg loads to surface waters of the SLRE, particularly under high flow conditions; however, the importance of near-shore, high carbon MeHg production zones and legacy mercury mobility needs to be more fully understood to effectively target management efforts aimed at minimizing human and ecosystem Hg exposure.

The historical use of mercury (Hg) by an industrial plant in the town of Visp, Switzerland has resulted in Hg contamination of soils and sediments in this region. From 1930 to 1976, Hg was utilized as a catalyst for acetaldehyde production, and the Hg-containing wastewater was discharged into an open channel canal which flows along an agricultural floodplain and then converges with the Rhone River. Current estimates indicate that approximately 50 to 60 tons of Hg were released into the canal. Following this, canal sediments were dredged and re-distributed as soil conditioner for agricultural fields and private gardens, resulting in widespread Hg contamination in soils and canal sediments within the affected floodplain. To better understand the distribution and fate of this Hg, soils and sediments within this region were collected and analyzed for total Hg and methylmercury (MeHg) concentrations. Total Hg was quantified by acid digestion and ICP-MS. MeHg was quantified using the following procedure: (i) a combined hydrochloric acid-dichloromethane extraction to release Hg from the soil and sediment matrices and to extract MeHg species to the organic dichloromethane phase; (ii) an aqueous back extraction step to transfer the MeHg to a L-Cysteine solution (0.1% (w/v) L-Cysteine); and (iii) further separation of inorganic Hg(II) and MeHg species using HPLC and quantification of MeHg using ICP-MS (detection limit: ~1 µg/kg). Total Hg concentrations ranging up to 28 mg/kg were detected in agricultural soils adjacent to the canal. Total Hg concentrations decreased with increasing distance from the canal (up to 40 m), and with increasing soil depth (0 to 50 cm). MeHg concentrations up to 7.8 µg/kg were quantified in the soil samples. Our data indicated that MeHg concentrations ranged from < 1 to 3.2 µg/kg, and did not change significantly with distance from the canal or with soil depth, except in one case (depth: 20 cm, distance from canal: 5 m) where 7.8 µg/kg MeHg was detected. MeHg/total Hg ratios for all samples were ≤ 0.08 %. Ongoing analyses include quantifying MeHg concentrations in sediment grab samples as well as in sediment cores along the length of the canal, and total organic carbon, total carbon, nitrogen and sulphur analyses to determine the source of MeHg and soil/sediment properties which may be correlated with MeHg and Hg methylation processes.

Hydroelectric power accounts for 83% of global electricity generation in Brazil. It is estimated that in 2030, electricity consumption in Brazil will be between 950 and 1,250 TWh/year, much higher than the current consumption of around 400 TWh/year. A great expand capacity is underway and it will be concentrated in the Amazon Basin where the high mercury load in the ecosystem is known and distinct social and environmental conditions need to be carefully analyzed, due to the potential to cause environmental, social and health impacts of great magnitude. The scale of the impacts is related to the type of technology used by the hydroelectric plant. In our case, the Madeira hydroelectric power plants used run-of-river type without formation of large reservoirs, consequently, with smaller flooded areas. The main aim of the present study was to assess the health impact of mercury exposure of the riverine communities in the phase of pre and post-impoundment of the hydroelectric reservoir in the Madeira River in the Amazon Basin. Traditional diet of riverine communities in Amazon basin is rich in fish that is an important source of protein. It is well known by the literature that flooded reservoir provide significant MeHg production peaks within the first 5 years in the aquatic environment and gradually declined back to baseline over several decades. In our case study of the hydroelectric power plant in the Madeira River, the health impact assessment of the mercury exposure of riverine communities had start before the pre-impoundment (2009-2011). For the phase of the post-impoundment (2014 to 2016) a longitudinal and a transversal studies were carried out for the influence area of the hydroelectric power plant. The results showed that the hair Hg levels were different for the upstream and downstream riverine communities for the two periods (pre and post- impoundment). The results of the post-impoundment (2009-2011) showed that the group of adults (16 years of age or older) the average concentration of Hg in hair was 9.4 ± 10.6 μg/g, with a minimum of 0.3 μg/g and a maximum of 66.8 μg/.g. In the period of post-impoundment 2014-2016, the average concentration of mercury in hair was 5.9 ± 6.1 μg/g, with a minimum of 0.1 μg/g and a maximum of 32.5 μg/g. The median in both periods was 5.7μg/g-1 and 4.2μg/g, respectively. The difference between the periods was statistically significant (p-value = 0.008).

The saying mad as a hatter stems from the medical condition of hatmakers who commonly used mercury nitrate in the preparation of felt for hats. Danbury, Newtown and Norwalk were centers of hatmaking in Connecticut in the 1800-1900s, with production of > 1million hats/year around the turn of the century. Much of that Hg has remained in the local uplands near the former hatmaking plants (68 plants in Danbury alone) and in the sediment of the rivers draining these areas (Still River, Housatonic River, Norwalk River). Sediment cores from a cove along the Still River carry up to 100 ppm Hg and river bed sediment of the Still River and Housatonic River has several ppm Hg. Sediment cores from river coves and marshes are analyzed for 210Pb, 14C, Hg, bulk dry density, and TOC and these dated core records provide accumulation rates of Hg for the last 100-200 years. The onset of Hg contamination coincides broadly with the beginnings of the local hatmaking industry and most profiles show a decline since the mid 1900s. Mercury use in hatmaking was halted in 1941. Ultimately, much of the hatting Hg ended up in Long Island Sound (LIS). Mass balancing the Hg input from the Housatonic River into LIS suggests that 25 to 35% of all Hg in LIS stems from the hatmaking industry. This Hg is spread through tidal and current activity widely throughout the Sound. During periods of intense rainfall, Hg-rich sediment is carried through the rivers directly into the Sound, leaving thin Hg-enriched event layers in the stratigraphic record. We compare the LIS Hg records with lake records from Block Island, RI, which we interpret as representing largely in situ atmospheric deposition. Marshes in central LIS remote from the source area and river inputs still carry a strong imprint of this legacy Hg from the Western CT hatmaking industry. This old Hg can become reduced in LIS and then potentially vented back into the atmosphere as Hg(o).

Mercury is a globally important contaminant which is of particular concern in stream systems. Because the consumption of contaminated fish is the most significant source of mercury (Hg) exposure to humans, regulatory guidelines that focus on the protection of human health require monitoring of both aqueous and fish tissue Hg concentrations at contaminated sites. However, because Hg is predominantly accumulated in fish via dietary rather than aqueous exposure, the link between aqueous Hg concentrations and concentrations in fish is not always straightforward, confounding remediation efforts at contaminated sites. Benthic macroinvertebrates can often comprise a major portion of fish diets and can therefore be important in the trophic transfer of bioaccumulative contaminants like Hg to fish. Benthic macroinvertebrate communities are commonly monitored in streams as an indicator of water quality, and while these community surveys do not provide direct information on Hg bioaccumulation, they can provide information relevant to understanding food web dynamics, and therefore Hg bioaccumulation, within a given system. Here, we examined changes in the invertebrate community over 30 years at two sites in East Fork Poplar Creek, a Hg-contaminated stream in East Tennessee, to explain why Hg concentrations in resident fish have not decreased in response to remediation activities that have been successful in decreasing aqueous Hg concentrations. We used the density and biomass of the major functional feeding groups in the invertebrate community, in conjunction with current efforts to quantify mercury inventories within the biota of this stream, to create a food web model to predict the most important factors leading to elevated Hg concentrations in fish. Bioconcentration factors (BCFs) were calculated for macroinvertebrate functional feeding groups using Hg concentrations in macroinvertebrates relative to aqueous Hg concentrations from 2013-2016. The BCFs were then applied to historical aqueous Hg concentrations to back-calculate Hg concentrations in the macroinvertebrate community over time. The diet of rock bass, the target fish species, was estimated at a given time period based on relative biomass and densities of macroinvertebrates at that time to calculate a weighted concentration for Hg trophic transfer to this species. Our results suggest that observed increases in the benthic macroinvertebrate community species diversity over time (i.e. longer food chain lengths) could explain why Hg bioaccumulation in fish has not changed despite substantial reductions in Hg loading into this stream.