CHARACTERIZATION OF ELEMENT CONCENTRATIONS AND COMPARISON TO LITERATURE
The sex ratio of collected muskrats was 13M:23F and concentrations did not differ based on sex for Cu (t = -0.01, p = 0.99), nor for Zn (t = 0.784, p = 0.44). The concentrations (based on wet weight in mg/kg) for each element in muskrat livers are as follows: arsenic ranged from 0.005–0.02, cadmium 0.002–0.02, chromium 0.01–0.5, copper 1.34–3.65, mercury 0.002–0.02, nickel 0.08–0.19, lead 0.0007–0.02, and zinc 19.37–30.98 (Fig. 2). Copper and zinc, both essential metals, were the only elements accumulating in muskrat livers at levels that were noticeably higher than the instrument detection limits. While no experimental heavy metal/metalloid toxicology studies have been conducted on muskrats, such studies have been conducted with similar species. Mink with Zn liver levels up to 212 mg/kg wet weight experienced no adverse effects (Aulerich et al., 1991), which is far above the 30.98 mg/kg observed in muskrats on the refuge. For Cu, excretory capacity in rats was not overwhelmed until liver concentrations exceeded 20 mg/kg wet weight (Milne & Weswig, 1968), far exceeding the maximum concentration found in the present study of 3.65 mg/kg (Fig. 2). In an observational study, concentrations of essential metals Cu and Zn in livers of muskrats collected near a smelter were comparable to the present study, ranging from 1–2.6 and 18.4–27.4 mg/kg wet weight respectively, while concentrations of non-essential metals Hg (non-detectable – 0.22) and Pb (0.27–0.96 mg/kg wet weight) were higher than in the present study (Blus et al., 1987). Based on these comparisons, neither Zn, Cu, nor any of the other elements studied are presumed to be at levels harmful for muskrats on the refuge.
The concentrations (based on dry weight in mg/kg) for each element in hybrid cattail are as follows: arsenic ranged from 0.19–11.5, cadmium 0.03–1.06, chromium 0.05–2.64, copper 1.82–59.25, mercury 0.006–0.02, nickel 0.2–4.67, lead 0.83–42.9, and zinc 2.01–53.66 (Fig. 3). Levels found in T. x glauca roots tended to be lower than in other Typha species (T. latifolia, T. angustifolia, T. domingensis) growing in contaminated wetlands (Bonanno & Cirelli, 2017). For instance, during their first collection period, average concentrations in roots of T. latifolia growing at the inflow point of a constructed wetland near a landfill had higher mean root concentrations than in the present study for the following elements in mg/kg dry weight: As (39.22 ± 2.68), Cr (9.00 ± 1.02), Cu (38.3 ± 4.66), Ni (12.6 ± 1.99) and Zn (126 ± 10.68) (Salem et al., 2014). As Typha species tend to have generally high resilience to heavy metals and metalloids and since concentrations in our hybrid cattail roots tended to be lower than its parent species in contaminated wetlands, there is no presumed harm to this species on the refuge (Bonanno & Cirelli, 2017).
The concentrations (based on dry weight in mg/kg) for each element in sediment were as follows: arsenic ranged from 1.27–11.5, cadmium 0.16–2.83, chromium 2.58–19.68, copper 5.51–25.83, nickel 4.54–22.31, lead 3.28–42.42, and zinc 6.8–82.17 (Fig. 4). We compared the sediment concentrations to two different EPA thresholds, the protective soil screening levels for plants, and mammals, which were given on a dry weight basis in mg/kg (U.S. Environmental Protection Agency, 2005). Where data did not exist from the EPA, we used the Texas Commission on Environmental Quality’s soil screening benchmarks for the protection of soil invertebrates or plants given in the same units (Texas Commission on Environmental Quality, 2021). Concentrations exceeded the soil screening level for plants for Cr only, of which all 36 sediment samples exceeded the threshold of 1 mg/kg. Concentrations in some sediment samples exceeded the soil screening levels for mammals for Cd and Zn. For Cd, 30 of 36 sediment samples exceeded the threshold of 0.36 mg/kg, while for Zn just one sediment sample exceeded the threshold of 79 mg/kg in MPN at 82.17 mg/kg in MPN. Overall, based on these thresholds, Zn is of low concern as only one sample exceeded the mammal threshold and only by a small margin. However, many samples exceeded the safe threshold for Cr in plants and for Cd in mammals.
Chromium does not have any known biological role in plant physiology and can impede growth and metabolic processes (Sharma et al., 2020). However, wetland plants have been shown to have high tolerance to metal contamination by sequestering them in the vacuoles of root cells thus preventing translocation to the aerial parts of the plant (Sharma et al., 2020). In the present study, a significant, but moderate positive correlation between Cr in the sediment and Cr in hybrid cattail roots (r = 0.56, p < 0.001) provides evidence that some but not all Cr in the sediment is bioavailable for uptake by T. x glauca. However, the Cr concentrations observed in hybrid cattail roots are not at levels presumed to cause harm. One study found no evidence of adverse effects in T. latifolia with maximum concentrations of 6.75 ± 1.20 mg/kg Cr in its roots, which is about three times higher than the maximum concentration we observed in T. x glauca roots at 2.64 mg/kg (Bonanno & Cirelli, 2017). Therefore, although the EPA suggests negative effects to plants at the current Cr levels in sediment, we do not suspect harm to hybrid cattail, which may extend to other resilient wetland plant species.
Cadmium has no known biological function and effects of Cd toxicity in mammals include reduction of food and water intake, growth depression, renal dysfunction, osteoporosis, hypertension, anemia, bleaching of incisors, and cancers (Cooke, 2011). While most of the Cd concentrations seen in soil samples in the current study surpassed the EPA’s safe level for mammals of 0.36 mg/kg, levels in muskrat livers were negligible, ranging from just 0.002–0.02 mg/kg. While the liver is a main organ of accumulation for many metals, Cd has been shown to be 2–8 times higher in kidneys than in livers of small mammals (Cooke, 2011). Nevertheless, even at levels eight times higher than observed, liver concentrations from muskrats in the current study are still far below what is considered a sub-lethal effect level of 10 mg/kg (wet weight) in livers of vertebrates (Peakall & Burger, 2003). While other mammals could possibly be experiencing detrimental effects of Cd, since muskrats are generalist feeders, we would expect them to be fairly representative at least of other non-carnivorous wetland mammalian species.
Bioconcentration factors (BFs) between sample types were calculated to provide insight into the potential trophic transfer pathway from sediment to hybrid cattail roots to muskrat livers. Typically, higher BF values imply a greater capacity for bioaccumulation, with BF values exceeding 1 indicating that a species could act as a hyperaccumulator of trace elements (Zhang et al., 2002). The elements showed the following decreasing trends in BFs (mean values), with asterisks representing those with BFs greater than 1:
BF (root:sediment): Cu* > Pb > Zn > As > Cd > Hg > Ni > Cr
BF (liver:root): Zn* > Cu*
BF (liver:sediment): Zn* > Cu
Between root and sediment, the only element with an average BF exceeding one was Cu (1.01), indicating that T. x glauca was taking up at least as much Cu as was present in the surrounding sediment. Other notably high BFs over 0.5 included Pb (0.86) and Zn (0.81). Bioconcentration factors for the other elements decreased in the following order: As 0.49, Cd 0.41, Hg 0.17, Ni 0.12, and Cr 0.08. The results of BFs between root and sediment contrast with a previous study, which showed BFs in T. domingensis, T. latifolia, and T. angustifolia in a different order as follows: Hg > Ni > Cd > Zn > As > Cr > Pb > Cu, with Hg and Ni showing BFs greater than one (Bonanno & Cirelli, 2017). However, it is noteworthy that the ranges found in our sediment samples were fairly small, especially for As (1.27 to 11.5), Cd (0.03–1.06 mg/kg), and Hg (0.06 to 0.11), which may not have been sufficient to cause detectable differences in hybrid cattail roots (Fig. 4). The difference in BFs between the studies may also reflect differences in environmental conditions that would impact bioavailability of the studied elements for uptake by hybrid cattail. Alternatively, our results may provide evidence that T. x glauca differs in its potential to take up heavy metals and metalloids compared to other species in its genus. These results should encourage further lab and field research, as a true difference in metal uptake capacity between the hybrid cattail and other members of its genus has implications for metal cycling dynamics in wetlands as T. x glauca achieves dominance.
Between muskrat livers and hybrid cattail roots, average BFs for both Cu (1.45) and Zn (8.03) exceeded 1, indicating that muskrats were accumulating more Cu and Zn in their livers than were present in the roots. Between muskrat livers and sediment, the average BF exceeded 1 for Zn (4.41), while the BF for Cu was also high (0.79). Together, these results provide some evidence that consumption of hybrid cattail and/or incidental ingestion of sediment may be sources of Cu and Zn, both essential elements, in muskrat liver tissue. That no other elements besides Cu and Zn were accumulating in notable quantities in muskrat livers likely indicates a greater bioavailability of essential elements (Cu, Zn) over non-essential (Cd, Hg, Pb) or comparatively more toxic elements (As, Cr, Ni).
In sediment, differences in concentrations between plots were found for all elements except Hg and Pb, while between sublots differences were not seen for Cd, Hg nor Pb (Table S5). In hybrid cattail roots, there were fewer significant differences between plots and subplots. Concentrations differed for As among subplots and Cd, Cr, and Ni between plots. Post-hoc tests revealed that among plots, it was always one of the two Main Pool sites that were higher than Radke or Teal, an expected result as Radke and Teal flow into Main Pool (Table S4). Overall, the high degree of patchiness in sediment between plots, and to a lesser degree between subplots was an expected result given the complexities of metal and metalloid mobility and bioavailability. Less variability in cattail roots indicates that the plants are potentially regulating the amount of metals and metalloids that they are taking up. For instance, plants are known to exert influence on the pH of the rhizosphere by 2–3 units, which could release essential metals for them to take up (Nason et al., 2018).
GIS analysis revealed that Cd, Cr, Cu, Ni, and Zn shared a hotspot in the sediment in the northernmost subplot in MPN with average concentrations within the subplot as follows in mg/kg dry weight: Cd 2.27, Cr 18.57, Cu 22.93, Ni 19.63, Zn 75.18 (Fig. 5). In T. x glauca roots, only Cr showed a hotspot in the same location, which aligns with the predicted effect of accumulation in plants based on sediment concentrations in this location exceeding the EPA’s threshold of 1 mg/kg by about 20 times (sediment average of 18.6 mg/kg) (Fig. 5). Two additional hotspots were located. A hotspot for Cu in cattail in the middle subplot in MPS (average 36.6 mg/kg), but not in the sediment suggests that T. x glauca may be actively taking up the essential metal, Cu, at that location (Fig. 6). The other hotspot was As in sediment in the leftmost subplot in MPS (average 5.3 mg/kg) (Fig. 6). Although these hotspots represent locations where concentrations were higher than other sampling sites, concentrations still remained under EPA thresholds, besides Cd and Cr, which exceeded thresholds for mammals and plants, respectively, as previously discerned from the overall averages. Coldspots were found in sediment in Teal for As (average 1.65 mg/kg) and in Radke for Cr (average 4.2 mg/kg), and Ni (average 5.2 mg/kg) (Fig. 5, Fig. 6).
We suspect that the observed sediment hotspot for Cd, Cr, Cu, Ni, and Zn in MPN may be anthropogenically influenced from runoff from the agricultural fields that border it to the east as all of these elements are known components of sewage sludge, which is often used as an agricultural amendment (Canet et al., 1998). We are less inclined to believe it is naturally caused as the underlying bedrock primarily composed of dolostone was the same for the hotspot location as most other sampling sites that did not display elevated concentrations (Fig. 5). In addition, we would not expect environmental conditions such as pH, dissolved oxygen or soil composition, which can affect retention of heavy metals and metalloids in sediments, to be drastically different in the hotspot compared to the other nearby subplots within MPN. While the hotspots in MPS are less easily explained, anthropogenic factors may have influenced these concentrations as well given that these sampling locations are located along a road and receive input from the west branch of the Rock River that could carry in contaminants. The coldspots located at the north end of the refuge provide evidence that potential metal and metalloid inputs contributing to the hotspots are less likely to be coming from the north.
COMPARISON OF SEDIMENT DATA 1990 TO 2021
Concentrations of elements in sediment were generally similar between 2021 and 1990 in both the northern and southern sampling sites with some exceptions (Warner, 2012). In the northern site, concentrations were significantly lower in 2021 on average for Cu (22.93 mg/kg in 2021, 31.33 mg/kg in 1990, t = 4.3, p = 0.02), Cr (18.57 mg/kg in 2021, 27.67 mg/kg in 1990, t = 8.7179, p = 0.002) and Zn (75.18 mg/kg in 2021, 90.33 mg/kg in 1990, t = 3.21, p = 0.04) (Fig. 7). No significant differences were found between years in the southern sampling site (Fig. 7). The observed decrease over three decades may be due to reduction in anthropogenic input or from natural changes in environmental conditions. Potential point source pollutants for Cu, Cr, and Zn include agricultural inputs, wastewater, or industry waste, which are all found in the vicinity of the refuge (Alloway, 2013). Therefore, reductions in agricultural applications or stronger regulations surrounding metal waste from wastewater treatment plants and/or industries over the last three decades may account for the observed changes. Reductions in inputs from non-point source pollutants including atmospheric deposition from mining, metal smelting and refining, manufacturing processes, transport, or waste incineration may also be a feasible explanation for the differences between years. The observed decrease may also be a natural phenomenon. Many environmental factors promote the release of metals from sediments, which would lead to lower detection in sediments. One such factor is lowered pH, which has an inverse relationship with dissolved oxygen and water temperature (Clark, 2017; Li et al., 2013). However, due to some known state regulatory changes regarding wastewater treatment plants as well as farming practice improvements, especially to the north of the refuge, we find more support for the anthropogenic improvement hypothesis.
It should be noted that the differences in sampling methods between 1990 and 2021 may have impacted these results. In 1990 samples were measured via ICP-OES (C. Dahman, personal communication, May 1, 2022), versus samples in the present study, which were measured with HR-ICP-MS. In addition, the 1990 samples were collected at a shallower depth of 0 -4 centimeters (K. James, personal communication, May 26, 2021), making them more susceptible to short term biological influences whether from human pollution or environmental such as changes in oxidation state. In addition, those samples were likely to contain more organic debris to which more metals and metalloids may have adhered. Nevertheless, while fine differences could be impacted by these method differences, both methods should give an accurate overall picture of total metal concentrations.