Physicochemical and geomorphological parameters were similar among sites (Table S1). No pairwise comparison of the sites (mussel/non-mussel or above/below Sardis) had significantly different mean values of these parameters, except for depth which was slightly greater at mussel than non-mussel sites (Table S2). However, the 0.07 m mean difference in depth observed between sites is unlikely to be biologically meaningful.
Elevated mercury was found in macroinvertebrate consumers that comprise the base of the food web (Table 1). Elmidae (riffle beetle) and Plecoptera (stonefly, genus Eccoptura) larvae had low mercury concentrations relative to other groups sampled while Ephemeroptera (mayfly) and Trichoptera (caddisfly) larvae had relatively higher mercury concentrations (Figure 2). There were differences within the latter two groups as well. Ryacophilidae, a caddisfly family of mainly predators, had the highest overall mercury concentration. Tortopus (Polymitarcyidae), a genus of burrowing mayflies and Stenoema (Heptageniidae), a genus of grazing mayflies, also had higher concentrations than other groups sampled. While Polymytarcidae’s close association with aquatic sediments, where mercury is methylated, might explain the higher concentrations in these larvae, it is less clear why Heptageniidae were also elevated (Table S3).
In addition to serving as a food source for aquatic predators, larval forms of aquatic insects complete their lifecycle by emerging from aquatic ecosystems into terrestrial adult forms. When these emergent insects leave their aquatic homes they transport the mercury in their bodies from aquatic (Gerrard & St Louis, 2001; Menzie, 1980; Raikow, Walters, Fritz, & Mills, 2011; Tweedy, Drenner, Chumchal, & Kennedy, 2013; Walters, Fritz, & Otter, 2008) to terrestrial food webs (Speir et al., 2014). Larval insects with a terrestrial adult form in the Kiamichi River ranged from approximately 30-300 ng/g total mercury. The same concentrations in spiders were found to pose a health risk to young songbirds in another study (Gann, Powell, Chumchal, & Drenner, 2015).
We found elevated mercury concentrations in all fish species sampled. At the base of the food web, stoneroller minnows (Campostoma spadiceum) had the lowest mercury concentrations of the fishes but were still higher than mussels. Though largely herbivorous, stoneroller minnows also consume invertebrates and detritus (Evans-White, Dodds, Gray, & Fritz, 2001), which could account for their elevated mercury concentration relative to mussels, another primary consumer (Figure 4).
We also observed surprisingly high mercury concentrations (Table 2) in darters (Percina maculata, Percina copelandi, Percina caprodes, Etheostoma maculatum, Percina phoxocephala, Lepomis humilis, Etheostoma radiosum, Etheostoma nigrum, Etheostoma spectabile, Percina sciera, and Crystallaria asprella), consistent with the only other study to have examined this group (Riva-Murray et al., 2011). The mercury concentrations of some darter individuals were as high as that of much larger piscivorous fish such as gar and bass, and they were comparable to or higher than centrarchids (Centrarchidae) and catfish (Ictaluridae) of a similar or larger size (Figure 4). The diet of darters (Percidae) is a possible explanation. A past study of the stomach contents of darters and sunfish stomach found benthic invertebrates, including mayflies, made up a high proportion of darters’ diets while sunfish consumed far fewer mayflies and consumed more terrestrial insects (W.J. Matthews, personal communication). In the Kiamichi River mayfly larvae had some of the highest mercury concentrations in our system (Figure 3). Preferential consumption of mercury-rich mayflies may contribute to elevated mercury concentrations in darters. Age may also play a role in this relationship. Many darters live up to 5 years (Paine, 1990), while sunfish in the 5 cm length range are young of year (Cargnelli & Gross, 1996). Many of the darters we sampled likely had several years to accumulate mercury in their bodies. While darters are not consumed by humans, they are an important prey item for aquatic predators such as bass (Schlosser, 1987). Additionally terrestrial predators, such as birds, are important top level predators in many aquatic systems (Steinmetz, Kohler, & Soluk, 2003) and are likely consuming darters as well. If darters are accumulating more mercury than other prey items and are preferred prey for top level predators, they may be disproportionately responsible for mercury body burden in top level predators. Because darters tend to only occur in rivers and streams (Miller & Robinson, 2004), this may be a unique factor that contributes to mercury contamination in rivers and not lakes.
At higher trophic levels, we found several fish taxa with individuals at or above the EPA mercury consumption advisory limit (Table 2). The majority of smallmouth bass over 10 cm in total length were over the EPA consumption advisory limit as were some sunfish and catfish. These results likely represent a conservative estimate of the extent of mercury contamination in the river. We were only able to electroshock wadeable portions of the river. This likely biased our samples to smaller fish. For instance, the catfish sampled were all under 20 cm standard length and the largest bass was only 22.6 cm. There are certainly larger fish in deeper habitats in the river. These larger fish likely have mercury concentrations higher than smaller fish. Thus, our results represent a conservative estimate of the upper limit of mercury concentrations in fish in the river. Our findings strongly suggest that the Kiamichi River, and others like it, represent a risk to human health with regards to the consumption of fish from the river (Table S4).
No significant difference in mercury concentrations was detected for fishes between sites with and without dense mussel assemblages. We were unable to test for differences in invertebrate taxa due to several sites where the biomass of invertebrates was too low for mercury analyses.
Our data show that a mid-size river in Oklahoma has mercury concentrations comparable to those found in large lakes and reservoirs. However, while these are monitored for human health threats, smaller order streams such as the Kiamichi River, and rivers rarely receive the same attention with regards to monitoring. The mercury concentrations found in fauna in the Kiamichi River is high enough to pose a health risk to both humans and wildlife, like the two impoundments on the river. While monitoring all smaller order streams seems unfeasible, further research exploring the relationship between reservoir mercury (which is monitored) and stream mercury could help in the issuing of more comprehensive and effective mercury advisories.