2d-2: Mercury fate in aquatic and terrestrial food webs

Fish in freshwater ecosystems constitute an important exposure pathway of Hg to humans and wildlife and thus are considered as critical receptors of long-range transboundary atmospheric transport of mercury (Hg). Fish Hg levels even in remote areas commonly exceed environmental quality standards (EQS) set by the WHO/FAO (0.5 1.0 mg kg-1 ww) and the EQS set by the water framework directive (0.02 mg kg-1 ww) is exceeded in all water bodies across Fennoscandia. Effects of atmospherically transported air pollutants and climate change on fish Hg levels have been studied from a long-term perspective in a Fennoscandian network. We have assembled fish Hg data and associated explanatory variables on fish, climate, deposition and catchment characteristics in a database, through a Nordic cooperation supported by the International Cooperative Programmes on Assessment and Monitoring Effects of Air Pollution on Rivers and Lakes (ICP Waters) and Integrated Monitoring of Air Pollution Effects on Ecosystems (ICP IM), the bodies under the UNECE Convention on Long-range Transboundary Air Pollution (CLRTAP). Individual fish specimen with reported Hg data were retrieved from Swedish (n = 39967), Norwegian (n = 6580), Finnish (n = 19692), and Russian (n = 225) databases. Monitoring of Hg levels in freshwater fish in Fennoscandia has been done since the 1960s, resulting in fish Hg data from a wide range of climatic, depositional, and land cover gradients typical for boreal and subarctic ecosystems. The Fennoscandinavian fish Hg database comprise a variety of fish species (n : perch pike >> i.a, brown trout, arctic char, roach) with a variation in fish species composition within and between lakes. Fish species, size and trophic level is used for the analysis but we also use additional data on water chemistry, climate and deposition to test hypotheses on environmental change on Hg in fish. High Hg levels in fish are usually associated with lakes with high concentrations of dissolved organic matter. Currently, surface waters are browning in boreal and subarctic Nordic ecosystems, possibly impacting Hg levels in fish. However, to understand variation in fish Hg levels many factors that affect Hg cycling and bioaccumulation (e.g., catchment characteristics, water quality, trophic structure, and climate) in addition to atmospheric deposition of Hg come into play. The Fennoscandian database is used to assess effects of long-range transboundary air pollution and climate on Hg in fish, through evaluation of temporal and spatial patterns in fish Hg levels. The results will be communicated to international policy bodies focusing on air pollution and mercury contamination.

Fisheries in highly productive, nutrient-rich lakes in the Northern Great Plains face Hg consumption advisories. This is counterintuitive since high pH, eutrophic lakes generally have lower fish mercury concentrations than expected based on comparisons to similar sized lakes with lower pH and less productivity. It is well understood that oligotrophic lakes with low pH, high dissolved organic carbon concentrations, and high proportions of wetland area have elevated fish Hg concentrations. Less understood are the environmental factors controlling fish Hg concentrations in eutrophic lakes.

Sources of methylmercury (MeHg) to fish are largely dietary with evidence coming from both experimental and stable isotope studies. Thus, invertebrates are important vectors of Hg from sites of Hg methylation to fish. We investigated the role of Leptodora kindtii (hereafter Leptodora), large (2-14mm), predatory invertebrates, common prey for plaktivorous fish in eutrophic systems, compared to smaller zooplankton, in the trophic transfer of MeHg to fish. Leptodora have been shown to exhibit pronounced size specific diel vertical migration (DVM) in eutrophic lakes where adults remain at or on the sediments by day moving into the water column by night. As a consequence of DVM, adult Leptodora may be an important dietary item for fish feeding in low light conditions common in eutrophic systems, provided planktivores have non-visual means of detecting prey items. We tested the hypothesis that migratory Leptodora act as a vector of MeHg from the sites of methylation to fish communities as an explanation for unexpected high fish Hg concentrations in eutrophic systems.

We measured MeHg concentrations in Leptodora and bulk zooplankton dominated by small copepods collected from a eutrophic prairie lake at mid-morning and mid-night in 20 m of water in 3 m intervals. During the day, the majority of MeHg in water column invertebrates was in the bulk zooplankton in areas less than 9 m deep. However, at depth, the majority of MeHg was in Leptodora. At night, MeHg in the Leptodora pool increased to between 67-96% of MeHg in invertebrates at all depths including shallow regions. Thus, the availability of MeHg in invertebrates in the water column at night was 10X greater than during the day and these invertebrates occupied the entire water column. Since Leptodora are available to fish using mechano-receptive feeding mechanisms, this is consistent with our hypothesis that Leptodora act as a conduit of MeHg from sites of methylation to fish due to upwards migration from sediments at night.

Mercury (Hg) contamination of fish is a global concern due to the deleterious health effects in humans and wildlife associated with ingesting fish with elevated concentrations. This study was conducted to assess Hg concentrations in biota of the Finger Lakes (New York, USA), a region where fisheries are an important economic driver but where no comprehensive assessment of food web Hg dynamics in lakes and streams has been completed to date. This region is of interest for the study of Hg accumulation in biota because the dominant land cover in this region is agriculture, which can affect lake trophic status and thus the bioavailability of methyl Hg (MeHg). Furthermore, there is a point source of Hg in the region from an active coal-fired power plant. The study objectives were 1) to determine if fish Hg concentrations were of concern to human and fish-consuming wildlife, 2) to assess differences in Hg accumulation among lakes and determine predictors of fish Hg concentrations, 3) to evaluate the predictive power of lower trophic level MeHg concentrations on fish Hg concentrations, and 4) to evaluate the influence of dissolved organic carbon and land cover on observed biota Hg accumulation patterns.

Between May October 2015, zooplankton and benthos were sampled monthly in five of the Finger Lakes (Honeoye, Canandaigua, Seneca, Cayuga, and Owasco Lakes). Fish were sampled once over the summer and species targeted from all trophic levels. Two species of stream fish (Blacknose Dace, Rhinichthys atratulus, and Creek Chub, Semotilus atromaculatus) were collected in three tributaries of each lake. Benthic macroinvertebrates representing various feeding groups and periphyton were collected for MeHg determination. Samples for water quality and dissolved organic carbon were also taken in both lakes and streams.

Results for top predatory fish, including Lake Trout (Salvelinus namaycush), Largemouth Bass (Micropterus salmoides), and Walleye (Sander vitreus) showed significant differences among lakes and concentrations above consumption guidelines (300 ng/g wet weight). No clear pattern among lakes was evident in lower trophic level fishes such as Yellow Perch (Perca flavescens) and Golden Shiner (Notemigonus crysoleucas) but concentrations were low. Zooplankton MeHg concentrations varied significantly among months within a lake and among months across all lakes. Lake morphometry, land cover, and water chemistry, including dissolved organic carbon, and lower trophic level MeHg concentrations will be included in a predictive model of fish Hg concentrations.

The bioaccumulation of neurotoxic methylmercury (MeHg) in freshwater ecosystems is thought to be mediated in complex ways by carbon (e.g., dissolved organic carbon [DOC]), the forms of mercury (Hg) present, and trophic structure, including the diet composition of heterotrophs. To better understand the combined effects of these factors on Hg bioaccumulation in stream invertebrates and fish, we studied tributaries across a range of DOC and dissolved mercury concentrations in the watershed of Lake Sunapee, New Hampshire, USA. We measured total mercury (THg) and methylmercury (MeHg) in three benthic invertebrate and three fish taxa,. Dissolved THg and MeHg concentrations in streamwater increased linearly with DOC, however, mercury concentrations in fish and invertebrates responded non-linearly to increased DOC, suggesting that MeHg bioavailability may be reduced at high levels of DOC. Methylmercury concentrations in invertebrates increased with trophic position. In addition, fish THg concentrations (and to a lesser extent invertebrate MeHg) increased with increased reliance on autochthonous diet items (as indicated by depleted del13C), suggesting that enhanced Hg bioaccumulation may be associated with a pelagic-based diet. Overall, our results suggest that mercury bioaccumulation in stream food webs is mediated both by streamwater characteristics and trophic structure.

Mercury contamination is a global issue owing to its long-range transport and toxicity. Variation in Hg accumulation among food webs is well studied, but less attention has been paid to how individual organisms, species traits, and food web interactions affect the movement of Hg within food webs and across ecosystem boundaries. We quantified Hg flux in food webs of the Colorado River in Grand Canyon by coupling measures of animal production, gut content analysis, and Hg concentrations over two years before and after an experimental flood. Mercury fluxes from basal food resources to animals were dominated by consumption of amorphous detritus and diatoms, accounting for ~90% of Hg fluxes among sites and years. Consumption of blackfly larvae was the dominant pathway of Hg flux to fishes, accounting for ~80% of Hg fluxes to fish at all sites. Large spatial and temporal discontinuities in Hg fluxes resulted from among-site differences in animal production and rates of resource consumption, species traits, ecotrophic efficiency (the proportion of total invertebrate production consumed by fishes), and food web stability (the degree to which food webs responded to the flood). For example, pre-flood Hg fluxes to invertebrates in the tailwater section of the river below Glen Canyon Dam were ~50 times higher than fluxes to fish because of a large imbalance between invertebrate secondary production and invertebrate consumption by rainbow trout (the only fish occurring in the tailwater). Invasive New Zealand mudsnails were extremely productive in the tailwater, accounting for 41% of total Hg fluxes to invertebrates, but these predator-resistant taxa were a sink for Hg as they were rarely eaten by trout. The gap between total Hg fluxes to invertebrates and trout narrowed significantly in the tailwater post-flood because mudsnail production declined by 10X, shrinking the pool of Hg that was unavailable to trout. Patterns in Hg flux were much different at downstream sites where food webs were less productive, more complex, more efficient, and more resilient to flood disturbance. Apex predatory fish affected the fate of Hg (moving from river to riparian food webs) by controlling densities of emerging aquatic insects. Food web dynamics affect fluxes of contaminants within food webs and across ecosystem boundaries, and understanding these dynamics is a critical step toward assessing vulnerable populations, managing ecosystems, and mitigating the effects of contaminants in the environment.

Methylmercury (MeHg) bioaccumulation in estuarine fauna is driven by local and regional scale variables that affect mercury inputs, bioavailability, and trophic transfer. Many of these variables, including organic carbon inputs, water temperature, and salinity, are predicted to change with changing global climate. In order to investigate the influence of these variables on MeHg fate, 12 estuarine sites spanning a range of environmental conditions were sampled during summer 2015. Sites were chosen in Delaware, Connecticut, and Maine in order to span a range of average growing season temperatures. In each region, two sets of paired high and low sediment organic carbon sites were chosen, aiming for one set to have relatively higher salinity than the other. We hypothesized that bioaccumulation would be relatively higher at the sites with warmer growing season temperatures and lower sediment organic carbon. At each of the 12 sites, sediment, surface water, and biotic samples were obtained and analyzed for mercury concentration and appropriate ancillary parameters (e.g. chlorophyll a, dissolved organic carbon, %loss on ignition, etc.). Methylmercury concentration data from both abiotic (sediment, dissolved water fraction, particulate water fraction) and biotic (Menidia menidia, Fundulus heteroclitus, and Palaemonetes sp.) samples will be presented. In addition, conclusions drawn from this and past fieldwork regarding the importance of each variable as a local or regional driver of MeHg fate in Northeast estuaries will be discussed.

Mercury contamination has been documented in Acadia National Park for many decades. The resultant data set of 20 years of studies provides an opportunity to synthesize what we know about mercury in the park. Our objectives were to describe how mercury concentrations vary across space within the park, explain species-level variation in mercury, and quantify how habitat influences the relationship between abiotic and biotic mercury levels. To do this, we created a database of mercury studies in Acadia National Park from published scientific papers, reports, and databases maintained by the park and other organizations. We found sufficient metadata that detailed when and where samples were collected for 675 records of abiotic samples (mercury in surface water) and 1035 records of biotic samples (mercury in blood, feather, egg, muscle, fur, and whole body samples, ranging from invertebrates to eagles). Then we grouped the data by watershed and built two spatially linked statistical models in a Bayesian modeling framework. The first is a model that predicts abiotic mercury using stream- or lakewater pH, dissolved organic carbon (DOC), time of year, watershed, and habitat surrounding the sample site. The second is a model that predicts biotic mercury using the type of sample collected, the species sampled, time of year, watershed, and the surrounding habitat. After fitting the model using Markov chain Monte Carlo sampling, we found that abiotic and biotic mercury were not correlated at the watershed scale. Abiotic mercury levels were strongly correlated with water body chemistry (pH, DOC) and the habitat surrounding each of the samples. Biotic mercury was strongly related to time of year and the type of sample taken, but few species were significantly higher or lower than the mean for all species. Moreover, surrounding habitat did not seem to strongly influence biotic mercury levels. These results suggest that mercury exposure is consistent (and relatively high) across the biota we sampled in the park. The lack of correlation between abiotic and biotic mercury could be due to a lack of connection between some animal mercury exposure and mercury in the surface waters, a mismatch between the spatial scale that the biota represent within their environment, or the scale at which we measured the habitat (e.g., local versus watershed scale). Further, the data document the ubiquity of mercury contamination across the many habitats in Acadia National Park.

Anthropogenic manipulation of aquatic habitats can profoundly alter mercury (Hg) cycling and bioaccumulation. The impoundment of flowing rivers is among the most common habitat manipulations and can result in increased fish Hg concentrations immediately following impoundment. However, differences in Hg concentrations between reservoirs and other habitats are still not well understood at larger spatial scales. We address this question at two scales. First, we evaluated total Hg (THg) concentrations in 64,386 fish from 883 reservoirs and 1,387 lakes located across the western United States and Canada to assess differences between reservoirs and lakes. In a second analysis, we examined differences in Smallmouth Bass (Micropterus dolomieu) THg concentrations between free-flowing and impounded segments of the Snake River in Idaho, USA, a semi-arid river with 22 impoundments constructed between the early 1900’s and 1980’s along its course. Across western North America, fish THg concentrations were 1.5-fold higher in reservoirs (0.134 ± 0.011 µg/g ww ± standard error) than in natural lakes (0.087 ± 0.006 µg/g ww), though this difference varied among ecoregions. Specifically, fish THg concentrations in reservoirs from the North American Deserts, Northern Forests, and Mediterranean California ecoregions were 1.5- to 2.6-fold higher than in lakes, whereas the two habitats did not differ in four other ecoregions. Along the Snake River, preliminary analyses indicate that Smallmouth Bass THg concentrations in reservoirs (0.141 ± 0.003 µg/g ww) and river segments directly below reservoirs (0.143 ± 0.005 µg/g ww) were 1.7-fold higher than in free-flowing segments not directly below reservoirs (0.082 ± 0.003 µg/g ww). Further, in free-flowing river segments, bass THg concentrations declined with distance from the nearest upstream dam, suggesting that these reservoirs can influence mercury bioaccumulation in fishes far downstream. Possible causes for elevated fish THg concentrations in reservoir habitats could include increased methylmercury production due to thermal stratification, effects of water level fluctuations, or differences in food webs between free-flowing and impounded reaches. Together, these data indicate that elevated fish THg concentrations can persist in reservoirs well past the initial impoundment phase and, coupled with the abundance of impoundments in the west, suggest that reservoirs may play an important role in determining Hg exposure to fish, wildlife, and humans. However, because reservoirs are often heavily managed, they may also provide unique opportunities to mitigate these risks.