Invertebrate composition effects on primary production
To evaluate the effects of fiddler crabs, marsh crabs, and mussels on benthic algae and cordgrass production, the dietary sources for fiddler and marsh crabs, respectively27,28, we measured benthic diatom biomass and cordgrass stem density every 4–6 weeks and quantified cordgrass biomass and grazing damage at the conclusion of the experiment in August 2017. Diatom biomass was enhanced in enclosures with mussels and/or marsh crabs relative to enclosures with only fiddler crabs or no invertebrates, and relative to all ambient plots (F36, 200 = 1.5; P = 0.04; Tukey’s HSD, all P < 0.015; Fig. 2a). In addition to providing evidence of diatom consumption by fiddler crabs, these results also suggest that mussels and marsh crabs stimulate diatom production, likely through their biodeposition of nutrient-rich pseudofeces and soil aeration activities, respectively29.
Inverttebrate composition effects on primary producer biomass, grazing damage, and crab diet. Diatom biomass recorded over the 5-month field experiment (a), and above ground cordgrass biomass (b), percent of cordgrass leaf area grazed by marsh crabs in all enclosures without marsh crabs (white), all enclosures with marsh crabs (teal) and all ambient plots (black) (c), and the δ13C and δ15N isotopes of fiddler and marsh crabs (d) recorded at the end of the experiment. Data are shown as the mean ± standard error of 3–4 measurements per plot per date in a; 6 replicate plots in (b); 23 replicate enclosures without marsh crabs, 12 replicate enclosures with marsh crabs, and 12 replicate ambient plots exposed to background levels of marsh crabs in (c); and 3 replicate enclosures per crab per treatment in (d). In (d), grey and purple shading denote the δ13C range previously reported for cordgrass and benthic algae, respectively, along the Georgia coast30. Colors and patterns denote treatment as defined in the legends for panels (a, b, d). Statistically significant findings (black star) are identified in (b, c). Summaries of statistical tests are shown as insets to each panel. Invertebrate icons were obtained through the Integration and Application Network or Google Imaged (labeled for reuse with modification).
While cordgrass stem density was not affected by treatment (P = 0.2; Supplementary Fig. S1), cordgrass biomass was 18% lower in all enclosures with marsh crabs (F1,38 = 21.2; P < 0.0001; Fig. 2b. Similarly, while mussels increased cordgrass biomass (Tukey’s HSD, P = 0.03), this effect was lost when marsh crabs were present (F9, 47 = 3.4; P = 0.003; Tukey’s HSD, P > 0.60; Fig. 2b). Grazing damage (percent of leaf area lost due to marsh crab shredding and consumption) indicated that marsh crab grazing pressure was 17-times higher in enclosures with marsh crabs and in ambient plots accessible to this species compared to enclosures lacking this species (F1,38 = 13.5; P = 0.0007; Tukey’s HSD, P < 0.001; Fig. 2c), reflecting both the effectiveness of our treatment design in controlling marsh crab distribution and the intensity with which this species grazes cordgrass. The low levels of grazing recorded in “no marsh crabs” plots likely occurred due to juvenile marsh crabs entering and consuming cordgrass within these enclosures given their ability to crawl through the caging mesh.
Marsh crab and fiddler crab diet
Isotopic carbon (δ13C) and nitrogen (δ15N) analyses demonstrated that neither marsh crabs nor fiddler crabs shifted their diet (i.e. δ13C, Kruskal–Wallis chi-square ≥ 1.51, P ≥ 0.6) or trophic position (i.e. δ15N, Kruskal–Wallis chi-square ≥ 1.22, P ≥ 0.3) in response to the experimental treatments. Fiddler crab δ13C values (− 16.2 to − 16.5) were significantly lower than those of marsh crabs (− 13.6 to − 14.5, Kruskal–Wallis chi-square = 21.3, P < 0.0001), indicating that this species primarily consumes benthic algae [benthic algae δ13C: − 16.2 to − 17.930]. In contrast, marsh crab δ13C values, although slightly lower, suggested that their diet primarily consisted of cordgrass [cordgrass δ13C: − 12.3 to − 13.630 (Fig. 2d)]. The lower marsh crab δ13C values may be due to trophic discrimination factors (e.g. differences in the isotopic ratio between the consumer and its diet) which have been shown to range widely for Sesarmid species31. Prior work27,28 and our own analyses of the gut contents of > 100 fiddler and > 200 marsh crab further supported these results, indicating neither species consumes mussels or detectable quantities of other invertebrates. In addition to cordgrass, we also observed sediment particles in marsh crab stomachs. As the average δ13C for intertidal sediment sampled from a Georgia saltmarsh ranged from − 14.3%. to − 20.0%32, we suspect that sediments may contribute to the lower δ13C values of Sesarma reticulatum relative to cordgrass. Using δ15N values to calculate trophic level33, we confirm that both fiddler crabs (1.8 ± 0.04) and marsh crabs (2.2 ± 0.03) functioned as primary consumers and maintained consistent trophic pathways for PCB accumulation across treatments.
Mussel effects on pseudofeces and sediment PCB concentrations
To assess the hypothesis that mussels enhance PCB contamination and bioavailability via their deposition of PCB-enriched pseudofeces, we collected: (1) mussel pseudofeces recently deposited on marsh surface, (2) the top 0–5 cm of root-bound sediment, (3) marsh crabs, and (4) fiddler crabs from mussel and non-mussel experimental enclosures. For each sample, we quantified the concentrations of a total of 100 PCB congeners (listed in Supplementary Table S1), hereafter referred to as PCBT.
PCBT concentrations of pseudofeces [mean ± SE; 121.1 ± 6.0 ng/g dw (33.4 ± 1.7 ng/g ww)] and sediment [384.3 ± 54.4 ng/g dw (101.4 ± 17.7 ng/g ww)] in mussel enclosures were 3.7- and 11.6-times higher, respectively, than those of sediment in no-mussel enclosures [33.1 ± 8.3 ng/g dw (9.4 ± 3.0 ng/g ww); F6,14 = 71.5; P < 0.0001; all Tukey’s HSD P < 0.01; Fig. 3a].
Mussel effects on PCBT and PCB homolog distribution. PCBT, reported as the sum of 100PCB congeners in ng/g wet weight (a), and the percent concentration of PCB homologs Cl4–Cl10 (b) in pseudofece, sediment, fiddler crabs, and marsh crabs (left to right), collected from enclosures without mussels (solid bars) and with mussels striped bars. Data are shown as the mean ± standard error of 3 replicate enclosures per treatment. In (a), the total concentration of the same 100 PCBs measured in our study were previously recorded in stripped mullet, whiting, Atlantic croaker, silver perch, spot, and spotted seatrout filets (gray arrows; TL = Trophic Level34), common fish collected from the TRBE24.
The lower PCBT concentrations of the pseudofeces compared to the underlying surficial sediment in mussel enclosures is likely due to intensive crab burrowing within and immediately surrounding mussel aggregations, and the effects of this burrowing on microbial activity and organic matter loss29. Specifically, crab burrowing aerates the soil, stimulating microbial activity and the resulting loss of organic material from the pseudofeces29. Since PCBs preferentially bind to organic material, we suspect that, as organic matter is lost through microbial decomposition, these contaminants gradually concentrate in the remaining, reduced volume of organic sediment particles. Over time, this material gets pushed into deeper soil depths via crab burrowing.
The pronounced difference in sediment PCBT between mussel and non-mussel enclosures supports our hypothesis that mussels rapidly elevate PCB concentrations via their filtration of PCB-enriched particulates from the tidewater and deposition of this material to the marsh surface (Fig. 4a,b).
Trophic and non-trophic pathways of PCB biomagnification in saltmarsh food webs. Mussels enhance PCB concentrations in surface sediments through their filtration of PCB-laced particulates from the water column and deposition of pseudofeces (a). Marsh crab bioturbation, by aerating and mixing the pseudofeces into underlying marsh sediment, cause PCB concentrations to both increase and ‘sink’; and, through their sustained interaction with PCB-enriched sediment, the marsh crabs accumulate PCBs primarily through epidermal/gill absorption (b). Marsh crabs, as prey for nekton, then serve as vectors for PCB biomeagnification up the saltmarsh food web (c). Arrows signify the flow of PCBs, green and yellow circles identify trophic and non-trophic pathways, respectively, and blue, purple, and red circles denote relatively low, intermediate, and high PCB concentrations (d). Icons were obtainted through the Integration and Application Network or Google Imaged (labeled for reuse with modification).
These findings, together with surveys showing that blue mussel (Mytilus edulis) biodeposits increase PCB sedimentation by 50% in the Baltic Sea18, highlight that suspension-feeding bivalves may generate ‘hotspots’ of PCB bioavailability in coastal sediments where tidal flow velocities are low enough to enable the settling of their PCB-enriched biodeposits. By summing concentrations of 65 congeners previously reported in the region24 in our samples, we also found that sediment PCB concentrations in mussel enclosures (238 ± 35.7 ng/g dw) were nearly 3-times greater [and in non-mussel enclosures 5-times less (16.3 ng/g dw)] than the average total PCB concentration reported for the top 2-3 cm of archived saltmarsh sediment collected near the LCP facility in 1996, the year it was placed on the National Priorities List (79.3 ± 2.47 ng/g dw)24. Thus, mussels enhanced PCB concentrations in our enclosures to “superfund site-levels” within 5 months, highlighting the magnitude of their localized effects on enhancing PCB bioavailability.
Mussel effects on crab PCB biomagnification
After 5-months on Blythe Island, fiddler crabs accumulated PCBT concentrations approximately 30-times greater (33.0 ± 6.8 and 28.4 ± 5.1 ng/g ww in no-mussel and mussel enclosures, respectively) than those collected directly from our reference site, Sapelo Island (PCBT: 1.1 ± 0.02 ng/g ww, Fig. 3a, Supplementary Fig. S2). However, fiddler crab PCBT did not differ across treatments (P > 0.9, Fig. 3a). Thus, fiddler crabs responded to the enhanced PCB exposure on Blythe Island relative to our reference site but were insensitive to the localized mussel-induced enhancement of PCBs in pseudofeces and sediment.
In contrast, marsh crabs from mussel enclosures (217.9 ± 8.8 ng/g ww) were significantly greater with concentrations on average 43-times higher-than their baseline values from Sapelo Island (5.1 ± 4.3 ng/g ww) and over 10-times higher than marsh crabs from no-mussel enclosures (20.8 ± 7.4 ng/g ww, Tukey’s HSD, all P < 0.001, Fig. 3a, Supplementary Fig. S2). Further, marsh crab PCBT from mussel enclosures was comparable to, or far exceeded, the average values reported for several higher-trophic level fish in the TBRE24,34 (Fig. 3a). As marsh crabs are important prey for many of these fish and other predators (e.g. birds35, red drum36), our data suggest that this primary consumer, when associated with mussel mounds, has the potential to be an important vector sustaining PCB biomagnification in this coastal food web (Fig. 4c,d).
PCB homolog composition analyses
To confirm the principle role of mussels in driving increased sediment and marsh crab PCBT concentrations via their pseudofeces, we compared homolog (Cl) profiles. Homologs are classes of PCBs based on their number of chlorine atoms, which range from one to ten. Homolog one (Cl1) consists of PCBs with one chlorine atom whereas Cl10 consists of PCB 209, the only congener with 10 chlorine atoms. Highly chlorinated congeners are more persistent and more prone to biomagnification37. While all samples were dominated by the highly chlorinated homologs (Cl7–Cl10) associated with Aroclor 12687,24,38, the pseudofeces exhibited a distinct profile, which was highly dominated by Cl9 and had the highest proportion of Cl10, relative to other sampled media. Within mussel enclosures, sediment and marsh crabs exhibited identical profiles (Tukey’s HSD, P > 0.40), being similarly dominated by Cl9 and high proportions of Cl10. Sediment and marsh crabs from no-mussel enclosures and all fiddler crabs were also indistinguishable (Tukey’s HSD, P > 0.30), and were instead dominated by Cl8 and exhibited relatively low proportions of Cl10 (MANOVA; F5,12 = 7.3, P < 0.0001; Fig. 3b).
Collectively, these homolog profiles suggest that the particulate-bound PCBs in tidewater that mussels filter and deposit as pseudofeces are more chlorinated than the PCBs present in marsh surficial sediments, which likely express lower chlorination due to environmental degradation processes37. Through the biodeposition of these highly chlorinated PCBs, mussels not only increase PCBT, but also redistribute highly chlorinated PCBs to surficial sediments. Moreover, the identical homolog profiles of marsh crabs and sediments in both mussel and no-mussel enclosures indicate that this species primarily accumulates PCBs via epidermal or gill absorption from contaminated sediment, a non-trophic pathway, rather than through trophic transfer (i.e. cordgrass consumption) (Fig. 4c). If marsh crabs were accumulating PCBs solely from cordgrass consumption, one would expect their homolog profile to be less chlorinated than the sediment profile because cordgrass selectively accumulates less chlorinated PCB congeners relative to the sediment in which it grows39. Thus, marsh crab bioaccumulation of highly chlorinated PCBs most likely occurs through the non-trophic mechanism of absorption from the sediment rather than through trophic transfer.
Our finding that fiddler crabs exhibit similar homolog profiles regardless of mussel presence suggest that this species, unlike marsh crabs, did not interact strongly with the pseudofeces or sediment adjacent to mussel aggregations (i.e. where sediment was sampled). Indeed, prior research has shown that fiddler crabs of the size transplanted into our experiment rarely burrow immediately on or near mussel aggregations22. The fiddler crabs’ lack of attraction to mussel aggregations and the marsh crabs’ strong affinity for burrowing adjacent to mussels29 highlights that niche differences among functionally similar organisms can dramatically influence which species are most prone to biomagnification via this non-trophic pathway.
Impact of mussels on risk of adverse biological effects
Finally, to determine whether the PCBT recorded in mussel versus non-mussel enclosures are high enough to induce adverse biological effects, we first compared our sediment concentrations with effects range low (ERL) and median (ERM) values. Concentrations equal to or below the ERL (22.7 ng/g dw) rarely result in adverse biological effects whereas those equal to or above the ERM (180 ng/g dw) are likely to induce deleterious biological effects40. While 100% of pseudofeces and sediment from mussel enclosures surpass the ERM value, only 30% of the sediment sampled in no-mussel enclosures exceeded this threshold, suggesting that mussels increased the risk of adverse biological effects among the benthic and surrounding estuarine community in only a few months’ time.
Impact of mussels on local dioxin-like toxicity
To further explore the potential adverse health impacts of PCB concentrations on marsh and fiddler crabs, we summed the concentration of the 12 most harmful PCBs whose toxicity is similar to polychlorinated dibenzo-p-dioxins41. Of these Dioxin-Like PCBs only five (105, 118, 126, 169, 189, hereafter referred to as PCBDL) were detected above our analyte-specific method detection limit and thus included in our PCBDL values. Marsh crab PCBDL from mussel enclosures (4.7 ± 0.24 ng/g ww) was significantly greater compared to all other treatments and mediums sampled (F6,14 = 18.7, P < 0.0001), which were indistinguishable from each other with PCBDL ranging from 0.03 (Sapelo Island marsh crabs) to 1.8 ng/g ww (sediment from mussel enclosures) (Supplementary Table S2).
Dioxin-like toxicity is commonly expressed by a toxic equivalency value (TEQ), which is the sum of the concentration of each dioxin-like compound detected in the sample multiplied by its corresponding toxic equivalent factor (TEF)41. TEFs are a ratio of the half maximal effective dose (ED50) for Tetrachlorodibenzo-p-dioxin (TCDD), a highly toxic dioxin, to the ED50 for the dioxin-like compound of interest41,42. Using standard TEFs41, we found that TEQs were 4- and 13-times greater in sediment and marsh crabs from mussel relative to no-mussel enclosures (Supplementary Table S2). Further, marsh crab TEQs increased 18-times in no-mussel and 157-times in mussel enclosures on Blythe Island compared to their baseline TEQs from Sapelo Island (Supplementary Table S2) and were significantly higher than all other sampled media (F6,14 = 23.26, P < 0.0001). Fiddler crab TEQs also increased 30- and 20-fold in no-mussel and mussel-enclosures relative to their Sapelo Island baseline values (Supplementary Table S1).
Since no TEQ data was found in the literature for either crab species in this study, we compiled PCB TEQs from five similar invertebrates in the literature (Supplementary Table S2) to contextualize our results. These PCB TEQs ranged from 0.2 pg/g ww (blue crab, Callinectes sapidus from Pensacola, FL43) to 39.6 pg/g ww (Chinese mitten crab, Eriocheir sinensis from Thames River, UK, where historically there have been high PCB inputs44,45). The average marsh crab TEQs from mussel enclosures surpassed those reported for the Chinese mitten crab, a burrowing omnivore, by threefold (131.9 pg/g ww) while the rest of our study’s TEQs fell within the range reported in the literature.
Collectively, our ERL, ERM, and TEQ values, all standardized metrics of risk for PCB-associated adverse health effects, highlight that mussels, through their filtering and deposition of highly chlorinated PCBs from tidewaters to the marsh surface, are dictating where PCB concentrations are continuing to reach biologically-harmful levels on Blythe Island and, likely, in saltmarshes throughout the TRBE and surrounding region more generally.
Source: Ecology - nature.com
