Mercury exposure to red-winged blackbirds (Agelaius phoeniceus) and dragonfly (Odonata: Aeshnidae) nymphs in Prairie Pothole wetlands

During the spring and summer surveys, aeshnid dragonfly nymphs were collected using dip nets from four of the eight wetlands (Table S2). Because of the relatively high unfiltered water Hg concentrations observed during spring in site B03 compared with other wetlands, dragonflies from site B03 were also collected in summer. Samples were collected near the shoreline, identified to suborder, rinsed, divided into two to three pools per site, and stored in Ziploc^® bags. Nymphs were frozen at −20 °C for storage and freeze-dried prior to mechanical homogenization using an acid-washed mortar and pestle.

RWBLs were caught using mist nets between 0500 h and 0830 h at four of the eight wetlands (Table S1). Adult birds could not be caught at all wetlands and thus effort was placed on those with higher populations and easier access. Eggs were collected from additional wetlands to increase sample distribution for this proxy (see below). Birds were also surveyed at site B03 during summer, although total RWBL abundance in wetland A35 decreased in the summer due to invasion by yellow-headed blackbirds (YHBLs) (Xanthocephalus xanthocephalus (Bonaparte, 1826)). Although RWBLs establish breeding territories by late March (Orians and Willson 1964), later-arriving YHBLs are known to usurp resident RWBLs (Orians and Willson 1964; Willson 1966). To minimize detection by birds, nets were placed near vegetation and sheltered from wind. A playback of male and female RWBL mating calls was used to attract individuals to the nets.

Blood samples (∼20 μL) were collected from the cutaneous ulnar vein in Heparinized Mylar^®-wrapped 75 mm Hematocrit capillary tubes, with a maximum of 3–4 tubes per bird. Vials were capped with Critocaps Disposable Micro-Hematocrit tube closures, placed in Ziploc^® bags, stored on ice, and frozen at −20 °C within 24 h. Blood was dried in an oven at 60 °C prior to analysis and percent moisture calculated. Percent moisture was 77.8% ± 2.2% (mean ± SD). The right and left second secondary flight feathers were pulled from each bird and stored in Ziploc^® bags until analysis. Blood THg concentrations represent MeHg exposure within the last few days or weeks (Evers et al. 2005; Edmonds et al. 2010). As MeHg is bound to keratin and sequestered in feathers during active growth, feather THg represents MeHg exposure during feather formation (Jaspers et al. 2004). RWBL molt generally occurs at the end of summer in August and thus THg concentrations indicate body burden at the end of the previous breeding season at the latest. Prior to release, each bird was marked to prevent resampling.

Egg collection occurred during mid-May, which was approximately 3–5 d postlaying. One egg per nest from each wetland was collected to a maximum of 12 eggs per wetland. Although official incubation stage data were not collected during initial surveying, using observational data we approximated the egg ages at approximately 3–5 d. Laboratory observations noted that embryo formation was not supported, which was consistent across the eggs collected. Eggs were randomly selected from each nest and placed in Ziploc^® bags, and stored in a cooler for transportation. Samples were frozen at −20 °C within 24 h until analysis. Prior to analysis, eggs were freeze-dried and homogenized. Percent moisture was 83.6% ± 1.0%. Collection of animal tissues was performed using procedures approved by the Canadian Council on Animal Care under appropriate permits obtained from the University of Regina (Animal Utilization Protocol 11-04; Canadian Council on Animal Care (CCAC-CCPA) 2010).

Water chemistry data were collected during spring and summer surveys. Parameters such as temperature (°C), conductivity (mS), pH, and dissolved oxygen (mg·L⁻¹) were measured using an YSI 556 MPS multiprobe. We sampled surface whole water for the analysis of THg and MeHg concentrations. Unfiltered water was collected in precleaned glass bottles using clean sampling techniques (US EPA 1630, 2001; US EPA 1631, 2002). Samples were preserved by acidification (THg samples acidified to 10% HCl and MeHg to 20% HCl) and stored at 4 °C until analysis. Sulfate (SO₄) concentrations were also determined in surface water using standard methods (Stainton et al. 1977).

We used THg as a proxy for nymph and RWBL tissue MeHg concentrations. Previous studies on dragonfly nymph show that 75%–88% of bodily THg is MeHg (Tremblay et al. 1996; Bates and Hall 2012; Edmonds et al. 2012). Likewise, the majority of THg (>90%) in avian blood, feathers, and eggs is in the form of MeHg (Evers et al. 2005; Rimmer et al. 2005; Wada et al. 2010; Ackerman et al. 2013), so the use of THg to approximate MeHg concentrations is justified.

A Milestone Direct Mercury Analyzer (DMA)-80 equipped with a gold amalgamator quantified THg concentrations in nymph and RWBL tissues using the US EPA Method 7473 (2007) at the University of Saskatchewan. A preweighed sample amount was added to a nickel boat and combusted. The DMA-80 was calibrated with SPEX Certiprep (PLHG2-1AY) liquid Hg standard using a gradient of Hg amounts from 0.5 to 1000 ng. THg emissions were measured by absorbance at 254 nm. Certified reference materials (CRM) (DORM-4, TORT-3, and IAEA-85) for secondary check standards were included during each batch. Sample duplicates were analyzed after 10 samples to ensure replication between 10% error. Recovery of CRM was 98% ± 2% (mean ± SD) for each set analyzed, and sample duplicates were within 10% error. The lower detection limit of the DMA-80 was approximately 0.04 ng Hg. Concentrations were expressed as dry weight (dw) for nymph, fresh weight (fw) for feathers, and wet weight (ww) for blood and eggs.

Water THg concentrations were analyzed using a Tekran^® Series 2600 Cold Vapour Atomic Fluorescence Spectrophotometer (CVAFS) at the University of Regina Mercury Lab, using protocols outlined by the US EPA Method 1631 (2002). Quality assurance and quality control (QA/QC) procedures were performed regularly, including Hg standards, duplicates, and spike recoveries. Duplicates were within 10% of each other, and the recovery range for spiked samples was between 88% and 115%. For MeHg determination, water was distilled using Tekran^® Series 2750 Mercury Distillation System at the University of Regina Mercury Lab using protocols in the US EPA Method 1630 (2001). Ammonium 1-pyrrolidine carbodithionate was added to each vial to dissociate Hg bound to organic matter and then distilled over 2 h. Samples were shipped overnight for analysis at the University of Western in Ontario by CVAFS. Machine stability measures were similar for MeHg CVAFS as those described for THg determination, with spike recoveries within 94%–101%.

Dragonfly THg concentrations ranged between 6.06 and 445.12 ng·g⁻¹, with lowest and highest levels in nymphs collected from sites A25 and A26, respectively. Mean nymph THg concentrations in spring and summer were 123.36 ± 117.26 and 136.24 ± 132.87 ng·g⁻¹, respectively. These concentrations were not significantly different between seasons (t = 0.22, df = 20.89, and p = 0.83) (Fig. 2A).

THg concentrations in RWBL blood ranged between 17.06 and 247.05 ng·g⁻¹ THg (Fig. 2B) and were significantly lower in spring (70.70 ± 57.11 ng·g⁻¹) than in summer (173.54 ± 44.18 ng·g⁻¹) (t = −5.18, df = 16.03, and p < 0.05). There was no significant difference between sexes in blood THg concentrations (Fig. S1). A few individuals from wetland A35 exhibited higher concentrations (>0.17 ng·g⁻¹) compared with other sites, which contributed to greater variation in spring THg concentrations; however, no birds were caught at this site during summer. Only one bird was caught at B03 in summer, and it had similar THg concentration as birds from A25 and A26.

Feather THg concentrations ranged from 99.26 to 1432.60 ng·g⁻¹ (Fig. 2C). Mean feather THg concentrations were 755.89 ± 402.07 ng·g⁻¹ in spring and 611.15 ± 467.37 ng·g⁻¹ in summer. There were no significant differences in feather THg concentrations between seasons (t = 1.06, df = 13.20, and p = 0.31).

The relationship between RWBL blood and feather THg concentrations was related to season. GLM analysis indicated that season, feather THg concentration, and the interaction between these variables were predictive of blood THg concentration (pseudo-R² = 0.59) (Table 1). Specifically, feather THg concentrations were positively related to blood THg concentration in spring (Fig. 3); however, this relationship decoupled later in summer when no relationship between tissues was observed.

Mean egg concentration was 7.60 ± 4.87 ng·g⁻¹ and ranged from 0.78 to 24.65 ng·g⁻¹. Eggs from sites A26 and B01 had the highest mean concentrations at 11.17 and 11.24 ng·g⁻¹, respectively, whereas those from B02 had the lowest at 2.60 ng·g⁻¹ (Fig. 4). Only one egg was collected from A35, which had a THg concentration below detection limit. There was a significant difference in egg THg concentrations among sites (F_(6,32) = 2.49; p = 0.05).

THg concentrations were highest in feathers followed by blood and then eggs. The ratio of THg concentrations in feather:blood:egg was 7:1:0.1.

The majority of dragonfly nymphs in this survey had THg concentrations below 150 ng·g⁻¹. These concentrations are lower than aeshnid nymphs collected farther north in the PPR (∼200 ng·g⁻¹ dw; Bates and Hall 2012) and in boreal Canada (∼190 ng·g⁻¹ dw; Hall et al. 1998). Values are also lower than those from odonate nymphs in the northeastern Great Lakes (230–1200 ng·g⁻¹ dw; Sinclair et al. 2012; Haro et al. 2013), but comparable with nymphs collected from ephemeral and semi-permanent ponds in Texas (120 ng·g⁻¹ dw MeHg; Chumchal et al. 2017). Moreover, mean THg concentrations in nymphs from our sites were <500 ng·g⁻¹ dw, suggesting that organisms feeding on dragonfly nymphs are not at great risk of accumulating high Hg concentrations. However, studies show that odonates contribute a large proportion of MeHg flux from ponds in the Great Plains (Chumchal and Drenner 2015). This flux can be greater in fishless ponds, such as in our study (compared with fish consuming invertebrates; Chumchal et al. 2017), which may result in some risk to more vulnerable organisms such as passerine nestlings (Williams et al. 2017).

Surveyed RWBLs from the PPR were indicative of background to low-risk Hg exposure. The maximum blood concentration observed was 247 ng·g⁻¹, substantially below the benchmark value of high-risk exposure for passerines (>3000 ng·g⁻¹; Ackerman et al. 2016b). Likewise, these concentrations were generally below values known to cause adverse ecological effects such as a 10% reduction in nesting success with blood concentrations >∼700 ng·g⁻¹ (Jackson et al. 2011; Lane et al. 2012). Egg THg concentrations were also substantially below the lethal concentration for passerines (tree swallows (Tachycineta bicolor (Vieillot, 1808)): LC₅₀ = 250 ng·g⁻¹ ww; Heinz et al. 2009) as well as levels associated with a 10% reduction in nest success (110 ng·g⁻¹; Jackson et al. 2011). However, two birds had blood concentrations above 200 ng·g⁻¹, which may indicate moderate physiological and behavioural alterations such as oxidative stress (Custer et al. 2000) and impaired reproductive success (Ackerman et al. 2016b). There are few studies examining feather, blood, and egg THg concentrations in RWBLs that we found to occur as a ratio of 7:1:0.1. Similar THg concentrations of feather:blood:egg ratios have been reported for the common loon (Gavia immer (Brunnich, 1764)) (6:1:0.3; Evers et al. 2005) and tree swallow (9:1:0.4; Brasso and Cristol 2008). Mean blood (113 ng·g⁻¹) and feather (697 ng·g⁻¹) THg concentrations were within the range found in marshes of northeastern USA (blood: 23–279 ng·g⁻¹; feather: 93–826 ng·g⁻¹; Tsipoura et al. 2008; Lane et al. 2012). Likewise, Bicknell’s thrush (Catharus bicknelli (Ridgway, 1882)) within boreal forests has similar blood (90 ng·g⁻¹ ww) and feather (700 ng·g⁻¹ fw) concentrations (Rimmer et al. 2005). Assuming 80% moisture, egg concentrations in the PPR were similar to mean concentrations observed in RWBL eggs from Voyageurs National Park, Minnesota, USA (∼13.2 ng·g⁻¹ ww; Tyser et al. 2016), but they were much lower than those from urban wetlands in New Jersey, USA (48 ng·g⁻¹ ww; Tsipoura et al. 2011).

Trophic level position affects biomagnification of Hg in passerines. Hg concentrations in RWBLs were much lower than those for rusty blackbirds (Euphagus carolinus (Statius Muller, 1776)) in the Acadian Forest (blood: 940 ng·g⁻¹, feather: 8260 ng·g⁻¹), which are almost exclusively invertivorous (Edmonds et al. 2010). Despite feeding mainly on insects during the breeding season (McNicol et al. 1982), omnivory by RWBLs may reduce their overall biomagnification potential. Indeed, in a synthesis of passerine Hg concentrations across North America, RWBL blood concentrations fall within similar values as those of other omnivores (Jackson et al. 2015). Overall, based on three tissues, ecological impediments to the surveyed RWBLs within the PPR as a result of Hg exposure were likely minimal.

Unfiltered water MeHg concentrations within the wetlands were similar to those found in other prairie systems, which have been found to be elevated compared with other systems (0.02 to over 9 ng·L⁻¹; Sando et al. 2007; Hall et al. 2009; Bates and Hall 2012; Hoggarth et al. 2015). Of particular concern, wetland B03 had significantly higher spring THg (15.9 ng·L⁻¹) and MeHg (9.3 ng·L⁻¹) concentrations compared with other wetlands (∼4.2 and 1.3 ng·L⁻¹, respectively). We did not find a relationship between summer unfiltered water MeHg and blood concentrations, even in B03 where high whole water MeHg concentrations were not reflected by similarly elevated concentrations in nymph or RWBL tissue compared with other wetlands (Table S2). The lack of relationship could be due to spatial and temporal variation in factors such as temperature, sediment redox potential, organic content, and sulfate concentration, which influence Hg methylation and hence accessibility to biota in shallow, PPR wetlands (Lavoie et al. 2013; Hoggarth et al. 2015; Paranjape and Hall 2017). The lack of relationships could also be because we examined relationships with MeHg concentrations in whole water and not those in the dissolved phase. Likely the lack of relationship of unfiltered water concentrations to other tissues reflects complexity in Hg biogeochemistry, spatiotemporality, and ecology (e.g., diet shifts and competition) that influences tissue THg concentrations (Edmonds et al. 2012; Paranjape and Hall 2017). As such, further research is needed to determine how these conditions affect the extent of Hg bioaccumulation through food webs.

Our results show that egg THg concentration was significantly related to spring whole water MeHg concentrations, whereas RWBL blood and feathers were not (Table 2). These results support previous indications that data from eggs may be useful indices for Hg risk assessment (Tsipoura et al. 2008; Peck et al. 2016; Tyser et al. 2016). In their review, Ackerman et al. (2016b) placed priority on using blood and egg THg concentrations as proxies for recent Hg exposure. This is partially because collecting eggs is relatively easy and because maternal THg blood concentrations are correlated with those in eggs (Ackerman et al. 2016a). As a result, eggs provide insight into Hg risk of passerines during the breeding season, a sensitive life history event. Clutch initiation date influences egg THg concentrations due to current Hg body burden and diet as well as subsequent depuration into eggs (Tyser et al. 2016). As spring whole water MeHg was correlated with egg THg and overall concentrations were relatively low, we argue that early clutches are not at risk of elevated Hg exposure resulting from greater biomagnification potential in breeding RWBLs. However, further research is needed to investigate the impacts of prolonged exposure to Hg-contaminated food sources of RWBLs on later clutches in the PPR, which may increase with clutch initiation date (e.g., Tyser et al. 2016).