Avian mercury exposure and toxicological risk across western North America: A synthesis

2.2. Data acquisition: literature review

A thorough literature review of all peer-reviewed journal articles and published reports documenting mercury concentrations in birds in western North America was conducted. Literature searches were conducted in Web of Science^™ and Google Scholar^™. For each study, the following information was extracted: bird species, tissue type, location (latitude and longitude), year, mean mercury concentration, units of measurement (including if data were reported in wet weight or dry weight), and sample size. Sometimes, year was reported as a range and, in these cases, the midpoint was used. When year was not reported, the publication year minus one was applied. Similarly, when sample sizes were reported as ranges, the midpoint was used as the sample size. When composite samples were used in a study, the number of composite samples was multiplied by the number of individual samples within the composites to calculate the effective sample size that was used in the study to produce the grand mean. When mean mercury concentrations were obtainable only from figures, rather than as values in a table or the text, the mean mercury concentration was visually approximated within the figure. Within the same study, mean mercury concentrations were kept separate for each species and location when possible; this often resulted with a single publication contributing multiple mean mercury concentrations, one mean for each species and location within the study. When location data were not reported within the study, study site descriptions (e.g., county and lake names) were used to assign approximate latitudes and longitudes using Google Earth^™.

2.3. Assigning bird taxonomy, foraging guilds, and habitats

For both the original and literature-review data, each species was assigned to a foraging guild and general habitat type. Taxonomy was based on the seventh edition of the American Ornithologists’ Union’s Checklist of North and Middle American Birds (retrieved Aug ust13, 2013 from http://checklist.aou.org/). Bird species were assigned to foraging guilds following DeGraaf et al. (1985) with the following modifications: (1) when a bird species occurred in multiple foraging guilds, such as piscivore and crustaceovore for several coastal seabirds, the primary foraging guild was used, and (2) when foraging guild differed by season (breeding, non-breeding, or year round), the foraging guild for the breeding season was used because most of the mercury data were from eggs or adults during the breeding season. Foraging guilds were categorized as piscivore, carnivore, insectivore, crustaceovore, molluscovore, vermivore, omnivore, granivore, or herbivore. Bird species were assigned to the following general habitats: ocean, coastal, salt marsh, both fresh and brackish water, freshwater, terrestrial-canopy, terrestrial-lower canopy, and terrestrial-ground. Habitats were assigned using DeGraaf et al.’s (1985) classifications as well as the Birds of North America series’ (http://bna.birds.cornell.edu/bna/) habitat descriptions. All avian taxa (including order, family, and species), foraging guilds, and habitats are summarized in Table S1.

2.4. GIS data layers

Geographic Information Systems (GIS; ArcGIS 10.2, Environmental Research Systems Institute, Redlands, CA, USA) were used to attribute each sample location with landscape variables, including ecoregion and 100-km × 100-km grid cell. The U.S. Environmental Protection Agency’s ecoregion level one category (Commission for Environmental Cooperation, 1997), which separates North America into 15 distinct ecoregions, was used and two additional categories were added: one for samples collected in the Pacific and Arctic Oceans (including various small islands and atolls), and one for samples collected on the Hawaiian Islands, for a total of 17 possible ecoregions. The Create Fishnet geoprocessing tool (ArcGIS 10.2) was used to create a grid of cells, each measuring 100 km × 100 km, across the extent of the sample locations in western North America, and then the Spatial Join geoprocessing tool (ArcGIS 10.2) was used to attribute each data point with the ecoregion and grid cell it occupied. Distribution maps of mercury concentrations in birds throughout western North America were produced using ArcGIS 10.2, and overlaid on a physical base layer provided by the U.S. National Park Service.

2.5. Data transformations and assumptions

Numerous data assumptions and transformations were necessary to consolidate, organize, and convert various tissue types and concentration units into similar values. First, only the following tissues were included : whole blood, eggs, muscle, liver, kidney, and fully grown feathers. These tissues represented >98% of the available data, and they are more readily comparable to one-another than the other available tissues, such as whole carcass. Second, data from any laboratory dosing or artificial studies that did not represent data from wild birds were excluded. Data from hazard assessments were included, because the impetus for a large number of studies was a known or suspected hazard (especially within the ECDMS dataset). Third, data were included only for eggs, adult tissues, or post-fledged juveniles. Samples collected from pre-fledged juveniles were excluded, because chicks undergo rapid changes in mercury concentrations in internal tissues as they grow and age (Ackerman et al., 2011) making any comparisons difficult. Fourth, data for both total mercury and methylmercury were included. All mercury in eggs (Ackerman et al., 2013), whole blood (Rimmer et al., 2005), muscle (Scheuhammer et al., 1998), and feathers (Thompson and Furness, 1989) was assumed to be in the methylmercury form, and, therefore, total mercury and methylmercury concentrations were used to represent methylmercury concentrations in birds. A significant proportion of the mercury in liver and kidney can be in the inorganic form (Eagles-Smith et al., 2009b; Scheuhammer et al., 1998; Thompson and Furness, 1989). Very few data (<1%) were available for these tissues as methylmercury concentrations, but, for those limited data, methylmercury concentrations were transformed into equivalent total mercury (THg) concentrations by using an adjustment of 88% of THg being in the methyl mercury form in liver (Eagles-Smith et al., 2009b). This assumption was justified because most data occurred below the 8.5 μg/g dry weight (dw) liver threshold where demethylation begins, above which a smaller proportion of THg as methylmercury would be expected in the liver (Eagles-Smith et al., 2009b). No adjustments were necessary for methylmercury concentrations in kidney, because THg concentrations were always available when methylmercury concentrations in kidneys were reported. Fifth, to make the mercury data comparable across bird tissues, all tissue concentrations were converted into blood-equivalent THg concentrations (μg/g) in wet weight (ww) using multiple equations from Eagles-Smith et al. (2008) and Ackerman et al. (2016a) detailed below. Before using these equations, it was necessary to convert THg concentration data from each tissue compartment into the same units, and thus all muscle, liver, kidney, and feather data were converted into dry weight THg concentrations (μg/g dw) using the reported percent moisture in the sample. Likewise, blood data were converted into wet weight THg concentrations (μg/g ww) using the reported percent moisture in the sample in the few instances (<1%) where blood was reported in dry weight. When moisture content was not reported, an average moisture content of 79% in blood, 67% in liver, 70% in muscle, and 74% in kidney was used (Eagles-Smith et al., 2008). For eggs, it is important to report mercury concentrations on a fresh wet weight (fww) basis (Ackerman et al., 2013; Stickel et al., 1973); however, the necessary egg morphometrics to make these adjustments were not available in many of the raw datasets and this made the conversion to fresh wet weight not possible. Therefore, when egg morphometric data were unavailable, egg THg concentrations (μg/g) were converted on a dry weight basis into a wet weight basis using the reported percent moisture in the individual egg or, when moisture content was not reported, an average egg moisture content of 75% was used (Ackerman et al., 2013). When egg morphometric data were available (i.e., authors’ data), THg concentrations on a fresh wet weight basis (μg/g fww) were used and calculated following Ackerman et al.(2013). In one instance, only albumen THg concentrations (μg/g ww) were reported, and the albumen THg concentration was converted into a whole-egg THg concentration (μg/g fww) using the predictive equation in Stebbins et al. (2009), before conversion into a blood-equivalent THg concentration. Hereafter, all egg THg concentrations are reported as simply μg/g ww. For most analyses, data points that were derived from the same bird, but in a different tissue, were excluded. Priority was given to tissues from the same bird in the following order: whole blood, eggs, muscle, liver, kidney, and fully grown feathers (see Table 2).

To convert THg concentrations in bird tissues into THg concentrations in blood, the following equations (eqs. 1–4) from Eagles-Smith et al. (2008), which were developed from >600 birds of 4 bird species with a broad range of tissue THg concentrations, were used. For the feather equation, the predictive equation for breast feathers, rather than head feathers, was used because most of the feathers sampled are typically body feathers and this differentiation among feather types was not usually reported.

To convert THg concentrations in eggs into equivalent THg concentrations in blood, the following equation (eq. 5) from Ackerman et al. (2016a), that was developed using 83 females and their full clutches for 3 species with a broad range of tissue THg concentrations, was used:

These tissue conversion equations (1–5) were developed for multiple species and used the largest sample sizes currently available, and therefore represent the best available conversion equations for multiple species. However, these equations were developed for four species in the order Charadriiformes and may not be representative of all bird species.

2.6. Statistical analysis

Linear mixed-effects models were used to examine factors influencing blood-equivalent THg concentrations in birds. Separate analyses were conducted for the two types of datasets: original raw data and the literature-review data. This separation ensured that data were not pseudoreplicated within the analyses because some of the original raw datasets were used to publish journal articles and reports that were summarized in the literature review dataset. THg concentrations in birds were log_e-transformed (natural log denoted as ln in equations) to improve normality. Back-transformed least squares means are reported with standard errors that were estimated using the delta method (Seber, 1982).

For the original raw dataset, log_e-transformed blood-equivalent THg concentration was the dependent variable; foraging guild (9 guilds), habitat (8 habitats), and ecoregion (11 ecoregions) were fixed factors; and grid cell (432 grid cells each 100 km × 100 km), year (29 years: 1982–2015), and species (225 species) were random factors. To compare THg concentrations among species without the inclusion of habitat and foraging guild, a separate analysis was conducted where log_e-transformed blood-equivalent THg concentration was the dependent variable; species was a fixed factor ; and grid and year were random factors. To examine the spatial distribution of THg in birds without the inclusion of habitat and foraging guild, an additional analysis was conducted where log_e-transformed blood-equivalent THg concentration was the dependent variable; grid was a fixed factor; and species and year were random factors. This same analysis was repeated for each guild with sample sizes >5000 (within a guild) to specifically examine the distribution of THg in birds in the piscivore (n=10243), insectivore (n=8464), and omnivore (n=6685) guilds.

For the literature-review dataset, log_e-transformed mean blood-equivalent THg concentration was the dependent variable; foraging guild (8 guilds), habitat (8 habitats), and ecoregion (15 ecoregions) were fixed factors; and grid cell (313 grid cells), year (46 years: 1968–2013), and species (176 species) were random factors. For this analysis, blood-equivalent mean THg concentrations were weighted by the square-root of the study’s effective sample size (i.e., the number of individuals used to estimate the mean), which placed more emphasis on the mean estimates that were derived from larger sample sizes. To compare mean THg concentrations among species without the inclusion of habitat and foraging guild, an analysis was conducted where log_e-transformed mean blood-equivalent THg concentration was the dependent variable; species was a fixed factor; and grid and year were random factors. To examine the distribution of mean THg concentrations in birds without the inclusion of habitat and foraging guild, an additional analysis was conducted where log_e-transformed mean blood-equivalent THg concentration was the dependent variable; grid was a fixed factor; and species and year were random factors.

3.1. Factors influencing bird mercury: original raw data

Bird blood-equivalent THg concentrations differed among foraging guilds (F_8,192.3=11.72, p<0.0001; Figure 2a) and habitat types (F_7,349.1=12.69, p<0.0001; Figure 2b), but did not differ among ecoregions (F_10,949.1=0.93, p=0.50). Piscivores (0.33±0.05 μg/g ww) and carnivores (0.32±0.10 μg/g ww) exhibited the greatest blood-equivalent least squares mean THg concentrations, whereas herbivores (0.03±0.01 μg/g ww) and granivores (0.02±0.01 μg/g ww) exhibited the lowest blood-equivalent least squares mean THg concentrations. These results are consistent with other studies that have found that birds foraging at higher trophic levels often have higher THg concentrations due to the biomagnification of methylmercury through food chains (Anderson et al., 2009; Blévin et al., 2013). In contrast, birds foraging on plants and seeds at the base of the food chain had substantially lower THg concentrations. Although these results were expected based on the ability of methylmercury to biomagnify, this is the first study to demonstrate differences in THg concentrations among such a wide range of foraging guilds.

Open in a separate window

Figure 2

Mean blood-equivalent total mercury (THg) concentrations in birds across western North America based on data derived from a literature review (n=1,712 means, representing n=19,998 individual samples). All mean data points are shown, with lower mean THg concentrations as larger symbols in the background and higher mean THg concentrations as smaller symbols in the foreground.

Bird blood-equivalent least squares mean THg concentrations were greatest in ocean (0.49±0.22 μg/g ww) and salt marsh (0.31±0.07 μg/g ww) habitats and lowest in terrestrial-ground habitats (0.04±0.01 μg/g ww; Figure 2b). Aquatic environments have biogeochemical conditions that are more conducive to methylation and methylmercury is more prevalent in aquatic than terrestrial ecosystems (Ullrich et al., 2001); therefore, it was not surprising that THg concentrations in birds would be lower in terrestrial than aquatic environments. However, some terrestrial birds can receive substantial aquatic subsidies of methylmercury through emergent aquatic insects and the associated food web (Cristol et al., 2008; Jackson et al., 2011b), so some terrestrial species can be exposed to higher methylmercury levels than would be assumed based upon their terrestrial foraging habits. Ocean and estuary environments tended to have birds with higher THg concentrations than those in freshwater environments. This difference could be due to several mechanisms, including differences in bioavailable methylmercury (such as differences in biogeochemical conditions, inorganic mercury availability, and methylmercury production; Ullrich et al., 2001), generally more complex food web structures and longer food-chain lengths in oceans and estuaries compared to smaller freshwater ecosystems (Post, 2002), or the ecology of bird species in these different habitats.

Blood-equivalent THg concentrations also differed among bird species (F_224,22076=130.79, p<0.0001; Figure 3; Figures S1–S7; Table S8). In particular, Forster’s terns had the highest least squares mean blood-equivalent THg concentrations of any species with sample sizes ≥60 (Figure 3), which is the approximate sample size necessary to estimate a population’s mean THg concentration with 10% accuracy (Ackerman et al., 2016b). Blood-equivalent geometric mean THg concentrations were 2.35 μg/g ww in Forster’s terns (5^th to 95^th percentile: 0.87–6.39 μg/g ww; Table S8). For comparison, common loons in the west, another piscivore that is well studied throughout North America, had a blood-equivalent geometric mean THg concentration of 0.89 μg/g ww (5^thto 95^th percentile: 0.25–3.90 μg/g ww; Table S8). Some other species with notably high blood-equivalent geometric mean THg concentrations were pigeon guillemots (2.08 μg/g ww), Caspian terns (1.58 μg/g ww), least terns (1.15 μg/g ww), black skimmers (0.90 μg/g ww), Clark’s grebes (0.83 μg/g ww), and black-necked stilts (0.79 μg/g ww; Table S8).

Open in a separate window

Figure 3

Least squares (LS) mean ± standard error blood-equivalent total mercury (THg) concentrations in birds among (A) foraging guilds and (B) habitats in western North America using original data at the individual level (black-filled bars; n=27,629 individual samples) and mean data derived from a literature review (hatched bars; n=1,712 means, representing n=19,998 individual samples). LS mean blood-equivalent THg concentrations were estimated separately for each dataset from models with foraging guild, habitat, and ecoregion as fixed effects, and grid cell, year, and species as random effects. Different letters next to bars denote significant (p<0.05) differences between means for the raw dataset (capital letters) and literature-review dataset (lower case letters).

Blood-equivalent THg concentrations of individual birds were above common toxicity benchmarks (Table 1) in many areas throughout western North America (Figure 1a). In particular, multiple individuals exhibited THg concentrations above 3.0 μg/g ww in San Francisco Bay, California; Central Valley, California; Carson River watershed, Nevada; Great Salt Lake, Utah; northeastern Washington; northeastern Montana; multiple sites along the Missouri River; southern Arizona; the Gulf Coast of Texas; Alaska’s North Slope; and the Aleutian Archipelago. These individuals typically were from species belonging to upper trophic level guilds, such as piscivores and carnivores.

To examine spatial variation in mercury exposure of birds that accounted for differences in THg concentrations among species, the distribution of blood-equivalent THg concentrations in birds also were mapped using model-estimated least squares means within 100-km × 100-km grid cells across western North America. As expected, bird blood-equivalent THg concentrations differed among grid cells (F_431,26126=20.67, p<0.0001; Figure 4a). Model-estimated mean THg concentrations were greatest in coastal California, western Nevada, and Alaska’s North Slope (Figure 4a). Other apparent hotspots, such as those in other parts of Alaska, British Columbia, Hawaiian Islands, and the western contiguous United States, had high THg concentrations but low sample sizes (typically <15; Figure 4b) and high coefficients of variation (>25%; Figure 4c) making interpretation at these sites more difficult. The analysis was repeated separately for each guild with a sample size >5000 and similar results were generally found for the piscivore (Figure 5a), insectivore (Figure 5b), and omnivore guilds (Figure 5c). THg concentrations were compared among guilds when they overlapped in the same grid cell. The strength of the correlations between guild-specific least squares mean blood-equivalent THg concentrations within grid cells varied among guilds, although the relationships were always positive (Pearson correlation s; omnivore vs insectivore: n=79 grid cells, r=0.47, p<0.0001; omnivore vs piscivore: n=56 grid cells, r=0.24, p=0.08; piscivore vs insectivore: n=69 grid cells, r=0.16, p=0.19).

Open in a separate window

Figure 4

Least squares (LS) mean ± standard error blood-equivalent total mercury (THg) concentrations among bird species in western North America using original data at the individual level. Only species with sample sizes 60 are displayed; see Figures S2–S7 for a complete listing of species by taxanomic order. LS mean blood-equivalent THg concentrations were estimated from a model with species as a fixed effect, and grid cell and year as random effects.

Open in a separate window

Open in a separate window

Figure 5

Blood-equivalent total mercury (THg) concentrations in birds across western North America using raw data (n=27,629 individual samples). Each grid cell is 100 km × 100 km. (A) The large map on the opposite page displays grid cells by their percentile of least squares (LS) mean THg concentration relative to the entire dataset, such that 20% of all grid cells are represented by each color. LS mean THg concentrations were estimated from a model with grid cell as a fixed effect, and species and year as random effects. (B) Displays the sample size in each grid cell. (C) Displays the coefficient of variation (as a percentage) for the model-estimated LS mean THg concentration in each grid cell. The three maps can be used in combination to evaluate the confidence in the estimated blood-equivalent THg concentration in individual grid cells. The darker graduations indicate (B) smaller sample sizes and (C) greater coefficients of variation which denote lower confidence in the model-estimated LS mean THg concentrations in those grid cells.

3.2. Factors influencing bird mercury: literature review

Bird blood-equivalent least squares mean THg concentrations differed among foraging guilds (F_7,167.1=16.01, p<0.0001; Figure 2a), habitat types (F_7,211=2.86, p=0.01; Figure 2b), and ecoregions (F_14,283.3=2.08, p=0.01; Figure 6). Carnivores (0.37±0.15 μg/g ww) and piscivores (0.31±0.09 μg/g ww) exhibited the greatest blood-equivalent least squares mean THg concentrations, whereas herbivores (0.01±0.01 μg/g ww) exhibited the lowest blood-equivalent least squares mean THg concentrations. As observed in the raw dataset, bird blood-equivalent least squares mean THg concentrations were highest in salt marsh (0.35±0.28 μg/g ww) and ocean (0.23±0.08 μg/g ww) habitats and lowest in terrestrial-ground habitats (0.04±0.01 μg/g ww). Among ecoregions, bird blood-equivalent least squares mean THg concentrations were greatest in tropical dry forests (0.22±0.13 μg/g ww) and tundra habitats (0.22±0.06 μg/g ww) and lowest in temperate Sierras (0.06±0.03 μg/g ww) and southern semi-arid highlands (0.04±0.02 μg/g ww), but pair-wise comparisons suggested few statistically significant differences among ecoregions (Figure 6).

Open in a separate window

Figure 6

Bird blood-equivalent total mercury (THg) concentrations in (A) piscivores (n=10,243 individual samples), (B) insectivores (n=8,464 individual samples), and (C) omnivores (n=6,685 individual samples) across western North America using raw data. Each grid cell is 100 km × 100 km. Maps display grid cells by their percentile of least squares (LS) mean THg concentration relative to the entire dataset, such that 20% of all grid cells in each foraging guild are represented by each color. LS mean THg concentrations in each foraging guild were estimated from a model with grid cell as a fixed effect, and species and year as random effects.

Similar to the raw dataset, bird blood-equivalent mean THg concentrations differed among species (F_177,1427=10.61, p<0.0001; Figures S8–S14) and grid cells (F_312,1195=4.17, p<0.0001; Figure 7a). Model-estimated mean bird THg concentrations based on the literature data also were highest in central and coastal California, western Nevada, Alaska’s North Slope, and the Aleutian Islands (Figure 7a). Additional hotspots were present throughout the west, although several of these additional sites had low sample sizes (typically <15; Figure 7b) and high coefficients of variation (>25%; Figure 7c). To directly compare the raw data (432 grid cells) to the literature data (313 grid cells), model-estimated mean bird THg concentrations within the 165 grid cells (100 km²) that contained both raw data and literature-review data were correlated. Least squares mean blood-equivalent THg concentrations were positively correlated between the two separate datasets, although the strength of the correlation was moderate (Pearson correlation; r=0.34; p<0.0001).

Open in a separate window

Figure 7

Least squares (LS) mean ± standard error blood-equivalent total mercury (THg) concentrations in birds among ecoregions in western North America using data derived from a literature review (n=1,712 means, representing n=19,998 individual samples). LS mean blood-equivalent THg concentrations were estimated from a model with foraging guild, habitat, and ecoregion as fixed effects, and grid cell, year, and species as random effects. Different lowercase letters next to bars denote significant (p<0.05) differences between means. Literature-derived bird THg concentrations were available for 15 of the possible 17 ecoregions in western North America.

3.3. Hotspots of bird mercury contamination in western North America

From the raw and literature-review data analyses, hotspots were identified in western North America for mercury contamination in birds. Several of these identified hotspots were common to both the raw and literature-review datasets, including the western Aleutian Islands, Alaska’s North Slope, Great Basin (especially western Nevada), and San Francisco Bay and Central Valley of California (Figures 4 and ​and7).7). To facilitate visualization of avian mercury exposure risk across western North America, a comprehensive map (Figure 8) was produced by combining the maps developed from the raw data and the literature-review data. When a grid cell contained THg concentration estimates from both analyses, priority was given to the estimate derived from the raw data and excluded the literature review-derived estimate for that grid cell. All grid cells that contained least squares mean blood-equivalent THg concentrations that were above the 80^th percentile of the entire dataset were considered to be potential hotspots for bird mercury contamination. Using this approach, 101 grid cells were identified that can be considered to be hotspots for avian mercury contamination in western North America (red grid cells in Figure 8). These hotspots included locations in the Aleutian Islands; the North Slope of Alaska; east-central Alaska; southeastern Alaska; northern Nunavut, Canada; Puget Sound, Washington; Great Basin (especially northern Idaho, and western and northern Nevada); San Francisco Bay and Central Valley, California; southern Arizona; the Gulf Coast of Texas; and the Hawaiian Islands (Figure 8).

Open in a separate window

Open in a separate window

Figure 8

Blood-equivalent total mercury (THg) concentrations in birds across western North America based on data derived from a literature review (n=1,712 means, representing n=19,998 individual samples). Each grid cell is 100 km × 100 km. (A) The large map on the opposite page displays grid cells by their percentile of least squares (LS) mean THg concentration relative to the entire dataset, such that 20% of all grid cells are represented by each color. LS mean THg concentrations were estimated from a model with grid cell as a fixed effect, and species and year as random effects. (B) Displays the effective sample size in each grid cell. (C) Displays the coefficient of variation (as a percentage) for the model-estimated LS mean THg concentration in each grid cell. The three maps can be used in combination to evaluate the confidence in the estimated blood-equivalent THg concentration in individual grid cells. The darker graduations indicate (B) smaller sample sizes and (C) greater coefficients of variation which denote lower confidence in the model-estimated LS mean THg concentrations in those grid cells.

Among the grid cell hotspots identified from the combination of the raw and literature-review datasets, many were characterized by low sample sizes (<15 samples; n=1 grid cell), high coefficients of variation (>25%; n=7 grid cells), or both (n=71 grid cells). Thus, additional sampling in these locations would help to determine if they are hotspots for bird mercury contamination. On the other hand, 22 of the identified hotspots were well sampled (>15 samples) and had relatively low coefficients of variation (<25%). These identified hotspots (red grid cells with bolded black borders in Figure 8) included the North Slope of Alaska; the western Aleutian Islands; Puget Sound; southwestern Idaho; western Wyoming; northern Montana; North Dakota and South Dakota along the Missouri River; central Arizona; the Gulf Coast of Texas; western Nevada ; and San Francisco Bay, California. Similar hotspots of mercury contamination were observed at some sites for freshwater fishes, especially in western and northern Nevada and central Arizona (Eagles-Smith et al., submitted to this issue). Avian mercury hotspots on the North Slope of Alaska may reflect recent trends in increased mercury exposure observed in piscivorous birds in the Arctic (Evers et al., 2014; Rigét et al., 2011), which are thought to be related to atmospheric deposition (Blum et al., 2013; Sunderland et al., 2009) and warmer Arctic temperatures associated with climate change potentially releasing inorganic mercury within snowpack, permafrost, and sea ice, and enhancing methylmercury production (AMAP, 2002; Brooks et al., 2006). In the Aleutian Islands, several studies have demonstrated high THg concentrations in birds with concentrations sometimes increasing westward across the island chain (Anthony et al., 2007; Ricca et al., 2008). In Washington’s Puget Sound, surf scoters exhibited THg concentrations similar to those of surf scoters in San Francisco Bay, California (Henny et al., 1991; Ohlendorf et al., 1987) and mercury concentrations of both surf scoters and western grebes increased as they over-wintered in Puget Sound (Henny et al., 1991, 1990). The hotspot in the Gulf Coast of Texas included Lavaca Bay, a designated mercury superfund site. Finally, San Francisco Bay estuary, California and western Nevada, have a long history of mercury contamination due to the legacy of mining (Conaway et al., 2008; Singer et al., 2013) and have widespread mercury contamination of biota (Ackerman et al., 2008, 2007; Eagles-Smith and Ackerman, 2014; Eagles-Smith et al., 2009a; Henny et al., 2007, 2002). San Francisco Bay, California; western Nevada ; and other Great Basin areas are of particular concern for methylmercury exposure to birds in western North America, and would benefit from inclusion in continental contaminant monitoring programs (Mason et al., 2005).

This work was conducted as a part of the Western North American Mercury Synthesis Working Group supported by the John Wesley Powell Center for Analysis and Synthesis, funded by the U.S. Geological Survey with additional support from the U.S. Geological Survey’s Ecosystems Mission Area and Contaminant Biology Program. We thank the U.S. Fish and Wildlife Service’s Environmental Contaminants Data Management System (ECDMS) managed by staff at the Analytical Control Facility, and the hundreds of scientists which have contributed to the ECDMS and STAMP datasets. We thank Charles Henny of U.S. Geological Survey for access to osprey data. We also thank Sarah Lemelin, Laura Young, Michelle Boyles, and Branden Johnson for help with data compilation; Michelle Lutz and Michael Tate for GIS layers; and Julie Yee for statistical advice. The use of trade, product, or firm names in the publication is for descriptive purposes only and does not imply endorsement by the U.S. Government.