3c-2: Risk assessment of mercury exposure to fish, birds, and wildlife

Simplified screening approaches for the neurotoxin mercury (Hg) may be helpful in risk assessments of natural resources in aquatic ecosystems. We explored the development of a simple screening approach in the Adirondack region of New York, USA, a known Hg hot spot with previously demonstrated high levels of bioaccumulation. Thirty-six small, perennial streams were sampled and analyzed for major solutes and Hg concentrations in fish tissue. These streams were broadly representative of a population of 391 Adirondack streams derived from previous investigations. Data analyses were based on length-normalized site median Hg concentrations in juvenile and adult Brook Trout, Creek Chub, Blacknose Dace, and Central Mudminnow and on median multi-species z-scores. Several landscape and chemical metrics known to be strongly related to methyl Hg concentrations in aquatic biota were explored for their strength of association with fish Hg concentrations. Surprisingly, none of the landscape metrics, including percent wetland area and mean basin slope, were significantly related to regional variation in fish Hg concentrations across the 36 streams. In contrast, several chemical metrics including dissolved organic carbon (DOC) concentrations, sulfate concentrations (SO4), pH, ultra-violet absorbance at 254 nm. (UV254), and specific ultra-violet absorbance were significantly related to regional variation in fish Hg concentrations. A cluster analysis based on the chemical metrics identified three distinct groups of streams, but all three groups were significantly different only for pH and not the other measures. Further analysis combined the data into two clusters that were most self-similar for further analysis. Among these two clusters, DOC, SO4, pH, UV254, and fish z-scores were all significantly different (p<0.05). Screening values of DOC>6.9 mg/L, SO4<2.8 mg/L, pH<6.6, and UV254>0.31/cm were suggested as thresholds to identify Adirondack stream sites likely to have the highest fish Hg concentrations. These values were especially effective predictors of high Hg streams likely to have at least one fish species with fish Hg concentrations greater than 100 ng/g, a level of concern for fish consumption by wildlife. A projection of Hg risk based on DOC concentrations to the larger population of 391 streams showed that about 28% were likely to have high wildlife risk, and these streams were located mainly in the western part of the Adirondacks. Finally, fish Hg/DOC ratios in this data set suggest less efficient uptake of Hg as stream DOC concentrations increase, consistent with the results of several recent studies.

Multiple studies have reported that fish in the Laurentian Great Lakes region are contaminated with elevated concentrations of methylmercury. However, little is known of concomitant levels of seleniuman element that binds with mercury and may ameliorate mercury toxicity when molar concentrations of selenium exceed those of mercury (i.e., molar ratios greater than 1.0). We measured mercury and selenium in axial muscle and liver of Northern Pike (Esox lucius), Smallmouth Bass (Micropterus dolomieu), and Largemouth Bass (Micropterus salmoides) sampled in spring 2008 2012 from lakes in Indiana Dunes National Lakeshore, Voyageurs National Park, Isle Royale National Park, Sleeping Bear Dunes National Lakeshore, and Pictured Rocks National Lakeshore. Not all species were obtained at all locations in all parks. Concentrations of mercury in axial muscle tissue of some predatory fish in all lakes exceeded the USEPA tissue residue criterion of 300 ng/g wet weight for methylmercury, a criterion established to protect the health of humans who eat noncommercial, wild fish. Selenium concentrations were negatively correlated with mercury in axial muscle of Northern Pike, but not Smallmouth or Largemouth Bass. Molar ratios of selenium to mercury ranged from 0.18 to 25, but were less than 1.0 in approximately 50% of fish with mercury concentrations greater than 300 ng/g wet weight. Concentrations of selenium and mercury in livers were positively correlated and molar ratios of selenium to mercury always exceeded 1.0, ranging from 1.8 to 85. However, there was evidence of cellular damage in livers with increasing concentrations of mercury. Lipofuscin, a pigment that accumulates in the liver of fish as a result of cell damage, can be measured spectrophotometrically and was positively correlated with liver mercury. In addition to predatory fish, prey fish were sampled in these five parks and from three streams in Grand Portage National Monument to assess the risk of methylmercury to fish-eating birds. Maximum concentrations of mercury exceeded 180 ng/g in whole body of prey fish--threshold effects levels associated with reduced reproduction in fish-eating birdsin nine of twenty-three water bodies in the six national park. Whole body molar concentrations of selenium exceeded mercury in all prey fish sampled. These results suggest that fish, humans who eat fish, and piscivorous birds are at risk from methylmercury exposure at a substantive number of locations within these parks and this risk may not be ameliorated by selenium.

Mercury (Hg) and selenium (Se) can threaten wildlife and impede ecosystem restoration efforts across a range of habitats. Selenium also interacts with Hg by impacting both biogeochemical processes in the environment and physiological processes within organisms. In particular, when Se in tissues is present in molar quantities that equal or exceed Hg, it may bind the methylmercury rendering it unavailable, thereby reducing its toxicity. Total Hg (THg) concentrations in fish from streams across the western United States have shown to inversely correlate with Se concentrations in tissues, with most fish having a molar THg:Se ratio <1 (excess Se), indicating that these fish may be less at risk for Hg toxicity than previously thought. Yet, the specific mechanisms controlling molar THg:Se ratios in nature are far from clear and we lack understanding of the importance of physiology, food web energetics, aquatic biogeochemistry and underlying geology in driving these ratios. We evaluated the relationships between Hg and Se concentrations in fish from across western North America to better understand the spatial variability in these relationships. Importantly, the expansive dataset compiled in this synthesis effort spans large gradients in both Hg and Se exposure, includes measurements in both muscle and whole body tissues, and includes a large number of species representing many feeding guilds. Molar Hg:Se ratios in more than 8,000 whole body fish samples ranged from 0.003 to 42 with a mean molar ratio of 0.54 (median = 0.24). Similar to prior published studies, more than 97% of the fish sampled had THg:Se molar ratios <1 (excess Se). Molar ratios in tissues varied among foraging guilds and to a lesser degree, geographical regions. Piscivores tended to have higher THg:Se molar ratios than benthivore or generalist species, possibly due to them feeding higher in the food chain where Hg biomagnifies. Mercury concentrations did not appear to influence the slope of the relationship between molar THg:Se and THg concentrations in muscle tissues when individuals were grouped into high (>1.5 µg/g dry weight) and low THg concentrations. But, a higher slope was observed between molar THg:Se ratio and THg concentrations in whole body samples (all foraging groups). Selenium excess over THg was greater in whole body samples compared to muscle tissues in all foraging groups. These results indicate that molar THg:Se ratios are affected by multiple interdependent factors that should be considered when assessing the potential protective role of Se in mitigating the toxic effects of methylmercury.

In the Canadian Arctic, mercury (Hg) concentrations in the tissues of non-anadromous (landlocked) Arctic char are elevated with ~30% of the sampled populations exceeding toxicity thresholds. In 2011, 2012, 2015, and 2016 we collected tissues (liver, muscle, brain) from Arctic char (n=227) from four lakes (Small, North, 9-Mile, and Amituk) on Cornwallis Island, Nunavut. The lakes sampled span a gradient of Hg contamination, allowing for the comparison of biological endpoints in char with low Hg concentrations to char with high Hg concentrations. The objectives of this research were to (1) measure total Hg and Hg speciation in Arctic char livers, brains and subcellular components, (2) determine correlations between Hg concentrations and biomarkers of oxidative stress in livers and brains, and (3) assess potential histological changes in livers and brains in relation to Hg exposure. Results thus far show a wide range in total Hg concentrations in Arctic char among lakes (0.04 to 2.6 parts-per-million wet weight in muscle), with about 36 % of individuals exceeding a toxicity threshold of 0.5 parts-per-million wet weight. Methylmercury was the predominate form of total Hg in liver (~80 %), where concentrations were two to three times higher than in muscle (to a maximum concentration of 6.5 parts-per-million wet weight). Total Hg was predominately found in the sensitive subcellular pools (mitochondria, microsomes and lysosomes, and heat-denatured proteins, including enzymes) of low- and high-Hg char, suggesting that increasing Hg exposure does not result in an activation of detoxification mechanisms in the liver. This may explain histological changes in the livers of fish from the most contaminated lake. Furthermore, a significant fraction of Hg may exit the liver and reach the brain, as concentrations in the two organs were similar. This research goes beyond documenting Hg concentrations in fish and will provide critical knowledge concerning fish health status.

The toxic effects of methylmercury (MeHg) on wildlife are well known, but impacts on the health and sustainability of global fisheries have not been well-characterized. Global tuna fisheries account for a large fraction of human exposure to methylmercury in many countries but are also an important food source of protein and essential nutrients for many populations. We have developed a framework for connecting our existing knowledge of mercury emissions and biogeochemistry to its impact on fisheries. In prior work, we have developed a modeling framework for simulating the global fate and transport of mercury in the atmosphere, terrestrial ecosystems and the oceans, including conversion to the bioaccumulative species methylmercury and uptake at the base of the oceanic food web. Here we link this simulation to a dynamic model of growth, reproduction and biomass of fish stocks and their interactions with the environment and fishing (Dynamic Bioclimate Envelope Model). We use a food-web bioaccumulation model to provide information on both the magnitude and timing of changes in fish MeHg levels relative to seawater. We specify dose-response relationships for methylmercury based on a review of the ecotoxicological literature on exposure impacts on growth, reproduction and body size of fish. When forced by the all-time historical emissions and future emissions scenarios developed by Argonne National Laboratory, we project a 30-50% increase from present day in seawater MeHg concentrations under business-as-usual future emissions and a 0-10% decrease under future mercury controls. Pacific tuna concentration changes show small lags (less than a decade) relative to seawater concentrations, with the lag depending on the size/age of the fish. We find that the long-term fish-specific MeHg trends are significantly affected by inter-annual variability in water temperature. We will discuss the impact of MeHg on maximum sustainable yields for fisheries of these commercially important species.

Differences in the impacts of stressors such as mercury toxicity and overfishing among fish species are not adequately understood, but are likely to have important consequences on the quality and supply of commercial fish for human consumption. Mercury can bioaccumulate to potentially toxic levels in fish, and fishing pressure driven by global fish consumption is on a steady rise. Both factors can increase stress on wild fish populations. However, the demographic impacts of these stressors will vary across fish species due to differences in their resilience, or ability to withstand stress. Here we examine relationships between fish mercury concentrations and resilience across a broad range of fish species. We hypothesized that fish with the lowest resilience, indicated by multiple measures of population productivity, are also at greater risk of having elevated, potentially toxic mercury content due to shared, underlying relationships with larger body size, and slower growth. We combined and analyzed existing data on total mercury concentrations from the Seafood Mercury Database, and life history characteristics from Fishbase, for approximately 400 species of finfish. We found that 83 (21%) species had total mercury levels exceeding the conservative toxicity threshold of 0.5 ppm (w.w.) associated with suppressed reproduction, growth, and behavior in freshwater fish. These species were dominated by large-bodied, marine pelagics, such as tunas and sharks. Of the 83 species with elevated mercury content, 34 (42%) are currently classified as vulnerable, near threatened, threatened, endangered, or critically endangered by the IUCN. Across all species in the analysis, both mercury content and resilience were strongly, positively related to body size, length at first maturity, maximum age, and negatively related to growth rate (measured as the Von Bertalanffy rate to reach asymptotic length). Thus, large, slow growing fish species are both at higher risk of having elevated, potentially toxic mercury levels, and are less likely to recover from population decreases due to stressors in general. Future studies should examine the combined and relative risks of mercury toxicity, overfishing, and other pressures, and prioritize their relevance to large, slow growing, commercially important fish.

The Pinchi Lake Mercury Mine in northern British Columbia, Canada produced elemental mercury (Hg) from 1940 to 1944 (historical operation) and from 1968 to 1975 (modern operation). During historical operations, mercury-contaminated calcines (primarily as cinnabar [HgS]) were deposited directly into Pinchi Lake (55 km2). Investigations conducted over the last two decades focused on characterizing the extent and magnitude of mercury (and other contaminants) contamination and associated environmental risks on the mine site and throughout the Pinchi Lake aquatic receiving environment. Results of these efforts informed the development of a closure plan for the upland mine site (implemented in 2010 to 2012) as well as a long-term aquatic monitoring program for Pinchi Lake. As part of continued monitoring, a comprehensive study was conducted in 2016 to document the current status of Hg in the Pinchi Lake ecosystem relative to a nearby reference lake (Tezzeron). In addition, a detailed sediment coring program was conducted, repeating a 1997 investigation, to provide updated information on natural recovery.

The 2016 results provide an updated context to the long-term monitoring data set. Pinchi Lake fish mercury concentrations remain elevated relative to nearby reference lakes and are attributable to historical mine operations, nearly 60 years earlier. Mercury concentrations in sediment cores remain high and show measurable, but slow natural recovery due to burial. While Hg/MeHg concentrations in water and zooplankton are low and similar historically and to many lakes elsewhere in Canada, they are still elevated relative to nearby Tezzeron Lake. Although fish mercury concentrations in lake trout standardized to 55 cm have declined significantly since the 1970s (~5 ppm), the trend in declining concentrations is quite slow (~1 ppm in 2016). Despite reductions in fish mercury concentrations, they are sufficiently elevated to potentially cause effects to fish and fish-eating wildlife based on literature studies. However, field investigations have not detected adverse ecological effects on fish, nor to local piscivorous birds (red-necked grebe, bald eagle) and mammals (river otter).

Legacy mercury (Hg) contamination is a complex environmental challenge in river systems, requiring substantial resource to adequately understand the problem and collaboration among multiple stakeholders for a meaningful and pragmatic solution. A regulatory-driven, watershed-level baseline ecological risk assessment (BERA) for the South River (Virginia, USA) provides a unique example of the scale, complexity, and collaboration involved to address legacy Hg contamination in a river system. Mercury release to the South River occurred from a former DuPont plant between 1929 and 1950 when it was used in acetate flake and yarn production. This caused potential impacts to the aquatic and riparian terrestrial systems (including the areas within the 62-year floodplain) along approximately 25 miles of the river downstream of the former plant. Legacy Hg in the potentially impacted areas has been the subject of numerous studies over the last four decades. The South River Science Team (SRST), a multi-stakeholder collaborative program, was established in 2001 to investigate the Hg issue in the study area. The BERA integrated extensive chemical, physical, and biological data to assess the potential ecological risk and was performed as a part of the US Resource Conservation and Recovery Act (RCRA) Facility Investigation (RFI). This presentation will focus on the collaborative approach to the BERA process and the implications of the BERA results on the remedial approach.